Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2010 Mar 22.
Published in final edited form as: Cell. 2000 Oct 27;103(3):481–490. doi: 10.1016/s0092-8674(00)00139-2

MMP-9 Supplied by Bone Marrow–Derived Cells Contributes to Skin Carcinogenesis

Lisa M Coussens 1,3,6,7, Christopher L Tinkle 1, Douglas Hanahan 2,4,6, Zena Werb 5,6
PMCID: PMC2843102  NIHMSID: NIHMS184997  PMID: 11081634

Summary

The matrix metalloproteinase MMP-9/gelatinase B is upregulated in angiogenic dysplasias and invasive cancers of the epidermis in a mouse model of multistage tumorigenesis elicited by HPV16 oncogenes. Transgenic mice lacking MMP-9 show reduced keratinocyte hyperproliferation at all neoplastic stages and a decreased incidence of invasive tumors. Yet those carcinomas that do arise in the absence of MMP-9 exhibit a greater loss of keratinocyte differentiation, indicative of a more aggressive and higher grade tumor. Notably, MMP-9 is predominantly expressed in neutrophils, macrophages, and mast cells, rather than in oncogene-positive neoplastic cells. Chimeric mice expressing MMP-9 only in cells of hematopoietic origin, produced by bone marrow transplantation, reconstitute the MMP-9-dependent contributions to squamous carcinogenesis. Thus, inflammatory cells can be coconspirators in carcinogenesis.

Introduction

Epithelial tumors are composed of both transformed, genetically altered epithelial cells and a variety of “normal” cells, including endothelial cells forming the tumor vasculature, fibroblasts, and inflammatory cells (lymphocytes, macrophages, mast cells, neutrophils). As tumors evolve in a stepwise fashion, malignant progression of neoplastic cells is driven by intrinsic events, such as activation of oncogenes and/or loss of tumor suppressor genes; increasingly, evidence is accumulating that extrinsic factors present in the local microenvironment are also influential in tumorigenesis.

Proteinases play pivotal roles in altering local microenvironments during embryonic development and growth as well as in tissue remodeling (physiologic and pathologic) processes. In particular, matrix metalloproteinases (MMPs) such as gelatinase B/MMP-9, gelatinase A/MMP-2, stromelysin-1/MMP-3, and matrilysin/MMP-7 have emerged as regulators of development, angiogenesis, and tumor progression (Coussens and Werb, 1996; Wilson et al., 1997; Cockett et al., 1998; Vu et al., 1998; Sternlicht et al., 1999; Stetler-Stevenson, 1999). Among these, MMP-9 has received relatively little attention as a significant factor in cancer despite studies with MMP-9-expressing tumor cells correlating its activity with malignant growth potential. MMP-9 can be expressed by epithelial and nonepithelial cell types, including fibroblasts, endothelial cells, and cells of hematopoietic origin (reviewed in Vu and Werb, 1998). Interestingly, recent studies have suggested a role for inflammatory cells, which are a source of MMP-9, in cancer phenotypes (Dvorak, 1986; Cordon-Cardo and Prives, 1999; Coussens et al., 1999; Fosslien, 2000).

We have studied a transgenic mouse model of epithelial carcinogenesis that involves expression of the human papillomavirus type 16 (HPV16) early region genes in basal keratinocytes (Arbeit et al., 1994); the K14-HPV16 mice reproducibly show multistage development of invasive squamous cell carcinoma (SCC) of the epidermis (Coussens et al., 1996). Animals are born phenotypically normal; by one month of age, with 100% penetrance, epidermal hyperplasias appear, and advance focally into angiogenic dysplasias between 3 and 6 months. Angiogenic dysplasias are characterized by intense mast cell infiltration of the reactive stroma, and increased density and altered architecture of capillaries (Smith-McCune et al., 1997; Coussens et al., 1999). By one year of age, 50% of transgenic mice develop invasive SCCs, of which ~20% metastasize to regional lymph nodes (Coussens et al., 1996). SCCs most frequently develop on ear (25%) and truncal (72%) skin, and, less frequently, on the head (2%) and appendages (1%); the tumors emerge out of dysplastic skin lesions. In this report, we have evaluated the role of MMP-9 during carcinogenesis in K14-HPV16 transgenic mice. The data indicate that MMP-9 expressed by inflammatory cells is functionally involved in distinct processes of epithelial carcinogenesis: regulation of oncogene-induced keratinocyte hyperproliferation, progression to invasive cancer, and end-stage malignant grade.

Results

MMP-9 Is Upregulated Early during Neoplastic Progression in K14-HPV16 Transgenic Mice

Extensive extracellular matrix (ECM) remodeling, characterized by collagen fibril degradation and epithelial basement membrane thinning, occurs early in neoplastic progression prior to frank tumor cell invasion in HPV16 transgenic mice (Coussens et al., 1996). To characterize proteolytic activities associated with this ECM remodeling, we used gelatin substrate zymography to reveal temporal changes in gelatinolytic activities in lysates of staged neoplastic tissue (Figure 1). The proenzyme form of MMP-9 was not detectable in normal skin, but was incrementally upregulated in hyperplastic, dysplastic, and tumor tissue. Active MMP-9 was detected at low levels in some hyperplasias and was abundant in 100% of dysplastic and tumor tissue lysates (Figure 1). Gelatinase A/MMP-2 (latent and active) was also not detected in normal skin, but the proenzyme form was present in hyperplastic, dysplastic, and tumor lysates. Notably, active MMP-2 was detected only in lysates of tumors (Figure 1). A serine proteinase, mast cell protease-4, was also detected by gelatin zymography in dysplastic and some tumor tissue lysates (Coussens et al., 1999).

Figure 1. MMP-2 and MMP-9 Are Activated during Epithelial Carcinogenesis in HPV16 Transgenic Mice.

Figure 1

Open in a new tab

Gelatin-substrate zymography of skin biopsies from ear skin of negative litter mates (norm) and K14-HPV16 transgenic mice with hyperplastic (hyp), dysplastic (dys) and Grade II SCCs. Arrows indicate location of enzymatic activity corresponding to pro- and active forms of MMP-9 and MMP-2. Molecular weight markers (kDa) are shown at left.

Absence of MMP-9 Decreases the Incidence of Carcinomas

To assess the functional significance of elevated MMP-9 activity during neoplastic progression in HPV16 transgenic mice, we adopted a genetic approach utilizing mice carrying a homozygous disruption of the MMP-9 gene (Vu et al., 1998). HPV16/MMP-9 +/– mice, bred four generations into FVB/n (n = 76), were phenotypically and histopathologically indistinguishable from control HPV16 transgenic mice (n = 133). With 100% penetrance, HPV16/MMP-9 +/– mice develop hyperplastic skin that is visually discernible at weaning, followed by development of angiogenic dysplasias by 4 months of age (data not shown). HPV16/MMP-9 +/– mice exhibited a similar latency and incidence of SCC development (46%) by 12 months of age as HPV16/MMP-9 +/– mice (50%), with the earliest cancers also arising at ~3 months (Figure 2A).

Figure 2. HPV16/MMP-9 –/– Mice Develop Fewer, but More Malignant, SCCs.

Figure 2

Open in a new tab

(A) Percentage of SCC-free transgenic mice in three cohorts: HPV16/MMP-9 +/+, n = 133 (FVB/n N14-N20; yellow triangles), HPV16/MMP-9 +/–, n = 76 (FVB/n N4; pink squares), and HPV16/MMP-9 –/–, n = 137 (FVB/n N4; blue diamonds). Both HPV16/MMP-9 +/+ and HPV16/MMP-9 +/– mice exhibit an overall SCC incidence of ~50% (50% and 47% respectively). In contrast, the incidence of SCCs in HPV16/MMP-9 –/– mice was reduced, to ~25% (Fisher's Exact Test, p = 0.0004).

(B) Immunostaining of intermediate filaments (brown staining) reveals degree of keratinocyte differentiation retained in various grades of SCC. Well-differentiated Grade I SCCs exhibit hallmark keratin pearl structures (asterisks) and express the suprabasal keratin K10 while not expressing the simple keratin K8. Grade II SCCs appear less differentiated and express both K10 and K8. In contrast, poorly differentiated Grade III SCCs lose terminal differentiation capacity, do not express K10, while maintaining expression of K8. Grade IV SCCs attain a spindle cell morphology, exhibit limited K10 expression and primarily express the mesenchymal intermediate filament protein, vimentin. Bar: 50 μm.

(C) A shift in tumor grade in MMP-9 –/– mice. SCCs were collected from the three cohorts of mice shown in panel (A) (HPV16/MMP-9 +/+, 133 mice, 67 mice with invasive SCCs, 80 SCCs total; HPV16/MMP-9 +/–, 76 mice, 35 mice with invasive SCCs, 43 SCCs total; HPV16/MMP-9 –/–, 137 mice, 37 mice with invasive SCCs, 42 SCCs total) and graded based on the criteria in panel (B). Both MMP-9-sufficient/HPV16 (+/+ and +/–) cohorts exhibited a similar spectrum of SCC grades, with no incidence of Grade IV SCCs. In contrast, MMP-9-deficient (–/–) transgenic mice exhibited an altered spectrum of SCCs biased toward less differentiated, more malignant cancers (p < 0.0001, Wilcoxon Score for Variable Grade).

In contrast, HPV16/MMP-9 –/– mice (n = 137) showed a marked delay in development of the characteristic hyperplastic and dysplastic phenotypes. At weaning, the HPV16/MMP-9 –/– mice did not show hyperplasias and were not phenotypically distinguishable from their MMP-9 –/– non-HPV16 littermates (data not shown). However, by 2 months of age, 100% of HPV16/MMP-9 –/– mice showed a mildly hyperplastic epidermis, characterized by a 2-fold increase in all keratinocyte layers with retention of complete terminal differentiation capacity, minimal stromal remodeling, and inflammatory cell infiltration, more typical of 3- to 4-week-old HPV16 controls (data not shown). By 4 months of age, only ~20% of HPV16/MMP-9 –/– manifested dysplasias characterized by loss of keratinocyte terminal differentiation, increased proliferation of keratinocytes, inflammatory cell infiltration, and intense angiogenesis. By 7 months of age, 90% had progressed to dysplasia; however, the lesions that did develop were more focal and limited in the extent of skin involvement compared to age-matched controls. The HPV16/MMP-9 –/– mice also exhibited a reduced incidence of malignant conversion. Only 27% of HPV16/MMP-9 –/– mice (p < 0.0001; Fisher's Exact Test) developed tumors, in contrast to the ~50% incidence in MMP-9-proficient/HPV16 mice (Figure 2A). The average number of SCCs per mouse (1.06% ± 0.12) and their anatomic locations (ear, 23%–29%; trunk, 51%–54%; head, 5%–16%; appendages, 10%–12%) in all three cohorts (+/+, +/–, –/–) did not change significantly.

Tumors Arising in HPV16/MMP-9 –/– Mice Are More Malignant

Tumors that arise in MMP-9-proficient/HPV16 mice represent a spectrum of carcinoma grades based on characteristic epithelial differentiation markers, ranging predominantly from well-differentiated (Grade I) carcinomas to infrequent poorly differentiated (Grade IIII) carcinomas and rare Grade IV cancers (Figures 2B and 2C). Because the SCCs that arose in MMP-9-deficient/HPV16 did so with lower frequency and extended latency, our expectation was that they would tend toward a more terminally differentiated cancer phenotype. To assess this, we graded tumors based on the expression of keratin intermediate filament isoforms (Broders, 1932; Lane and Alexander, 1990; Coussens et al., 1996). The most well-differentiated tumors expressed keratins 10 and 14, but not the simple epithelial keratin K8, and were classified as Grade I SCCs (Figure 2B). Tumors expressing K10, K14, and K8 were classified as Grade II SCCs, whereas poorly differentiated tumors no longer expressing K10, while retaining expression of K14 and K8, were classified as Grade III SCCs (Figure 2B). We defined carcinomas with a spindle cell appearance as Grade IV SCCs, consistent with an epithelial-to-mesenchymal transition that suppressed expression of keratin intermediate filaments and increased expression of vimentin, a mesenchymal cell marker (Sternlicht et al., 1999; Figure 2B).

Remarkably, the distribution of tumors in HPV16/MMP-9 –/– mice was biased toward carcinomas that expressed less differentiated or embryonic phenotypes representative of higher malignant grades, compared to tumors from HPV16/MMP-9 +/– and +/+ controls. Well-differentiated Grade I cancers decreased in the absence of MMP-9 from 57.5% (MMP-9-proficient/HPV16 mice) to 21.4% in MMP-9 null mice. In contrast, the percentage of Grade II, III, and IV cancers all increased in the absence of MMP-9 (Grade II, 35% to 42.8%; Grade III, 7.5% to 28.5%; Grade IV, 0% to 7.1%; p < 0.0001, Wilcoxon Score for Variable Grade; Figure 2C).

Keratinocyte Hyperproliferation Is Restricted in MMP-9-Deficient/HPV16 Mice

A tumor's aggressiveness is commonly graded in terms of the extent to which it has lost normal differentiated characteristics, and by the proportion of proliferating cells. To address the latter parameter, keratinocyte proliferation was analyzed by bromodeoxyuridine (BrdU) incorporation, revealing marked reductions in HPV16/MMP-9 –/– mice, at all time points and histologic stages, compared to HPV16/MMP-9 +/– controls (Figure 3). Whereas progression to characteristic dysplasia and angiogenesis were merely delayed ~3–5 months by the absence of MMP-9 in HPV16 mice, keratinocytes never fully attained characteristic stage-specific increases in proliferation, even when histologically similar tissues were compared (albeit at different ages), e.g., hyperplasia versus hyperplasia, dysplasia versus dysplasia, or in similarly graded tumors. Thus, despite the propensity toward a higher malignant grade, keratinocyte proliferation in MMP-9-deficient SCCs remained suppressed compared to controls (Figure 3).

Figure 3. Reduced Epithelial Hyperproliferation at All Stages of Carcinogenesis in HPV16/MMP-9 –/– Mice.

Figure 3

Open in a new tab

Percentage of bromodeoxyuridine (BrdU)-positive keratinocytes in age-matched normal FVB/n negative littermates (-LM; [1], MMP-9 –/– [1]), and HPV16/MMP-9 +/– and HPV16/MMP-9 –/– mice at 1 and 2 months [2], 3 and 4 months [3], greater than 5 months of age [4], and from Grade II SCCs [5]. BrdU-positive keratinocytes were counted from a minimum of five high power (40×) fields from five mice per age group. Results are shown as mean percentages, ± standard error of the mean. P values (unpaired, nonparametric Mann-Whitney) for the respective groups are 0.66 [1], 0.11 [2], 0.73 [3], 0.03 [4], and 0.06 [5], respectively.

The initial induction and increasingly intense angiogenesis that characterizes the premalignant stages in the presence of MMP-9 in this pathway (Smith-McCune, 1997; Coussens et al., 1999) is also affected: angiogenesis is attenuated much as is keratinocyte proliferation in the hyperplasias and dysplasias in the absence of MMP-9. However, while keratinocyte proliferation never reaches wild-type plateaus, angiogenesis eventually does, but with a delay of many months (data not shown).

Reactive Stromal Cells Supply MMP-9 during Neoplastic Progression

Increased proliferation and upregulation of MMP-9 in keratinocytes at the edges of squamous epithelial wounds have been suggested to provide a migratory advantage for those keratinocytes (Madlener et al., 1998). Perhaps the diminished proliferative capacity of keratinocytes in MMP-9-deficient/HPV16 mice is specifically due to a lack of MMP-9 in HPV16-positive keratinocytes. Accordingly, we sought to determine the cellular sources of MMP-9 in HPV16 transgenic skin (Figure 4).

Figure 4. MMP-9 in Neoplastic Tissue Is Present in Reactive Stromal Cells.

Figure 4

Open in a new tab

(A–C) In situ hybridization analysis of neoplastic tissue from normal and HPV16 transgenic skin for expression of MMP-9 mRNA. Dark-field photographs of ear-tissue sections from normal nontransgenic ear skin (A), HPV16 dysplastic ear skin (B), and a Grade II SCC (C). The epidermal (e), dermal (d), cartilagenous (c), malignant tumor keratinocytes (t), and tumor stroma (s) regions of the tissues are indicated. The epithelial basement membrane is marked with a dashed line. Bar: 50 μm (A); 100 μm (B and C).

(D–I) Immunolocalization of MMP-9 (brown staining) in normal and HPV16 transgenic skin sections counterstained with methyl green from normal nontransgenic ear (D), dysplastic ear skin (E), and a Grade II SCC (F). High power microscopy reveals presence of MMP-9 in mast cell granules in dysplasias (G, arrows), consistent with our previous observations (Coussens et al., 1999). In addition, MMP-9 is found in neutrophils with characteristic bilobed nuclei in both dysplasias and carcinomas (H, arrows). In tumors, MMP-9 is also found in macrophages (I, arrow) with characteristic membranous folds and finger-like projections, as well as in the ECM (F, arrowheads). The epidermal (e), dermal (d), cartilagenous (c), malignant tumor keratinocytes (t), tumor stroma (s), and capillaries (*) are indicated. The epithelial basement membrane is marked with a dashed line. Bar: 50 μm (D); 75 μm (E); 20 μm (F); 10 μm (G); 7 μm (H, I).

MMP-9 mRNA was not present in the epithelial or dermal compartments of normal nontransgenic skin as determined by in situ hybridization analysis on paraffinembedded tissue biopsies (Figure 4A). In characteristic hyperplastic skin, MMP-9 mRNA was not detected, except in isolated basal keratinocytes abutting skin abrasions and in some hair follicles (data not shown). In contrast, MMP-9 mRNA expression in dysplastic skin was broadly observed in reactive cells residing in the dermis (Figure 4B). In tumors, mRNA expression was primarily detected in cells localized around capillaries and in a subset of tumor-associated fibroblasts (Figure 4C). Thus, MMP-9 mRNA showed a dual localization in neoplastic skin: It was predominantly restricted to cells in the stroma, rather than the evolving cancer cells; sporadic basal keratinocytes in hair follicles also expressed MMP-9.

Since reactive inflammatory cells can deliver proMMP-9 to lesional tissue in preformed granules, we also assessed the presence of pro- and active MMP-9 by immunolocalization with MMP-9-specific antibodies in HPV16 neoplastic tissue (Figure 4). MMP-9 protein was not detected in normal nontransgenic skin (Figure 4D). In hyperplastic skin, MMP-9 was only observed in isolated capillary-associated cells in the stromal compartment (data not shown). In contrast, MMP-9 in dysplastic skin and in carcinomas was detected in many cells in the stromal compartment (Figures 4E and 4F). At higher magnification, the nuclear and cytoplasmic granule morphology indicated that mast cells (Figure 4G) and neutrophils (Figure 4H) were the predominant sources of MMP-9 in premalignant tissue. In tumors, MMP-9 was also found in neutrophils associated with capillaries (Figure 4F), in isolated macrophages (Figure 4I), and the ECM in the core of the tumors (Figure 4F). Consistent with our previous observations (Coussens et al., 1999), we did not detect mast cells expressing MMP-9 in the core of solid tumors, although MMP-9-positive mast cells were associated with invasive tumor fronds at the periphery. Thus, MMP-9 was not detected in the oncogene-positive neoplastic cells; rather, it was present in the ECM and in reactive inflammatory cells populating the stroma in the premalignant and malignant stages of the SCC pathway.

Inflammatory Cells Expressing MMP-9 Can Restore Characteristic HPV16 Phenotypes

The data document the functional involvement of MMP-9 in epithelial hyperproliferation, angiogenesis, and malignant potential, and reveal the predominant cells expressing it to be inflammatory cells in the stroma of dysplasias and tumors. Macrophages, neutrophils, and mast cells all arise from hematopoietic stem cells in the bone marrow (Weissman, 2000). If inflammatory cells are a key supplier of MMP-9, then reconstitution of MMP-9-deficient/HPV16 bone marrow with MMP-9 +/+ bone marrow-derived (BM-d) cells should restore characteristic neoplastic phenotypes, as seen previously for endochondral ossification in MMP-9 –/– mice (Vu et al., 1998). Accordingly, we lethally irradiated (7.5 Gy) 1-month-old HPV16/MMP-9 –/– mice and reconstituted their bone marrow with either wild-type (MMP-9 +/+) or MMP-9 –/– cells (Figures 5A and 5B). Wild-type bone marrow (Figure 5B) was successfully engrafted into HPV16/MMP-9 –/– mice. MMP-9 was detected by zymography by eight weeks posttrans-plantation (3-month-old mice; compare Figure 5C with Figure 1). By 16 weeks posttransplantation (5-month-old mice), MMP-9 levels (both pro- and active forms) were typical for dysplastic lysates from HPV16 control mice (compare Figure 5C with Figure 1).

Figure 5. Restoration of MMP-9 Activity in MMP-9-Deficient/HPV16 Mice by Bone Marrow Transplantation.

Figure 5

Open in a new tab

(A) Scheme of bone marrow transplantation (BMT) protocol where 1-month-old mice (HPV16/MMP-9 +/–, HPV16/MMP-9 –/–) were lethally irradiated (7.5 Gy) and transplanted with BM-d cells (1 × 106) from nontransgenic donor mice (MMP-9 +/+ or MMP-9 –/–). Each cohort contained a minimum of ten animals. Ear biopsies were taken at 8 and 16 weeks post-BMT when animals were 3 and 5 months of age. All animals were studied either until tumors appeared, or animals developed health-compromising phenotypes, or until 12 months of age. (B) Gelatin-substrate zymography of BM-d cells (1 × 106) from MMP-9 +/+ (+/+, lane 1) and MMP-9 –/– (–/–, lane 2) donor mice. An arrow indicates enzymatic activity corresponding to proMMP-9.

(C) Gelatin-substrate zymography of tissue biopsies from HPV16/MMP-9 –/– mice reconstituted with either wild-type (+/+) or MMP-9-deficient (–/–) BM-d cells (1 × 106) at 8 and 16 weeks post-BMT. Presence of enzymatic activity corresponding to pro- and active MMP-9, pro- and active MMP-2, as well as an uncharacterized gelatinase are indicated. Molecular weight markers (kDa) are shown at right.

After transplantation with MMP-9 +/+ BM-d cells (Figure 5B), HPV16/MMP-9 –/– animals were restored to the characteristic phenotypes of neoplastic progression (from hyperplasia to dysplasia by 4 months of age), analogous to the control HPV16/MMP-9 +/– cohort transplanted with MMP-9 +/+ BM-d cells (data not shown). Enhanced keratinocyte hyperproliferation in tumors, as assessed by the cell cycle regulated proliferating cell nuclear antigen (PCNA), was restored in the chimeric HPV16/MMP-9 –/– mice transplanted with MMP-9 +/+ BM-d cells (Figure 6A). By contrast, the keratinocyte proliferation index remained low in HPV16/MMP-9 –/– mice transplanted with MMP-9 –/– BM-d cells (Figure 6A).

Figure 6. MMP-9 from Bone Marrow-Derived Cells Is Sufficient to Restore Wild-Type Neoplastic Phenotype in HPV16/MMP-9 –/– Mice.

Figure 6

Open in a new tab

(A) Percentage of PCNA-positive keratinocytes present in Grade I and/or Grade II SCCs from nonirradiated, nontransplanted, HPV16/MMP-9 +/+ and –/– mice, and from HPV16/MMP-9 +/– and –/– mice that were lethally irradiated and transplanted with either wild-type (+/+) or MMP-9-deficient (–/–) BM-d cells. PCNA-positive keratinocytes were counted from a minimum of five high power (40×) fields/tumor analyzed. Results shown are derived from 5 SCCs each from nonirradiated, nontransplanted HPV16/MMP-9 +/– [46.6% ± 1.9] and –/– [32.3% ± 2.3] mice, 5 SCCs from irradiated/transplanted HPV16/MMP-9 +/–, BMT +/+ [45.6% ± 5.0] and HPV16/MMP-9 –/–, BMT –/– [45.9% ± 4.6] mice, and from the 1 SCC-bearing HPV16/MMP-9 –/–, BMT –/– [29.9% ± 5.1] mouse (p < 0.05, ANOVA [one-way analysis of variance]). Results are shown as mean percentages, ± standard error of the mean.

(B) Keratinocyte proliferation, as measured by immunoreactivity for proliferating cell nuclear antigen (PCNA), in Grade II SCCs from four cohorts of mice: (a) HPV16/MMP-9 +/– (b) HPV16/MMP-9 –/–, (c) HPV16/MMP-9 –/–, irradiated and transplanted with MMP-9 +/+ BM-d cells, and (d) HPV16/MMP-9 –/–, irradiated and transplanted with MMP-9 –/– BM-d cells. Note in (a) that proliferation is heterogeneous throughout the malignant clusters, whereas in (b), proliferation is restricted to within 3–4 cell layers proximal to stroma (s). Transplantation of MMP-9 +/+ BM-d cells restores the characteristic pattern of proliferation to keratinocytes (c). Bar: 100 μm (a–d).

(C) Incidence of SCCs in animals from BMT study (HPV16/MMP-9 +/–, BMT +/+,n = 10 mice, 7 with SCCs; HPV16/MMP-9 –/–, BMT +/+, n = 10 mice, 5 with SCCs; HPV16/MMP-9 –/–, BMT –/–, n = 11 mice, 1 with an SCC) was determined in comparison to nonirradiated, nontransplanted transgenic animals (HPV16/MMP-9 +/+, n = 133, 67 mice with SCCs; HPV16/MMP-9 –/–, n = 137 mice, 37 with SCC).

Significantly, the tumors in HPV16/MMP-9 –/– mice showed a distinct difference in their pattern of proliferating keratinocytes, in addition to a reduced proliferation index. Whereas proliferating keratinocytes (as revealed by the S-phase markers BrdU and PCNA) in tumors of HPV16/MMP-9 +/+ and +/– mice were found throughout malignant clusters (Figure 6B, panel a), proliferating keratinocytes in tumors from HPV16/MMP-9 –/– mice were restricted to layers of epithelia in close apposition to stroma (Figure 6B, panel b). The diffuse pattern of keratinocyte proliferation was restored in HPV16/MMP-9 –/– mice by transplantation of +/+ BM-d cells (Figure 6B, panel c), but not by transplantation of MMP-9 –/– BM-d cells (Figure 6B, panel d). These data provide evidence that neoplastic cell hyperproliferation is not controlled solely by factors intrinsic to neoplastic cells such as E6/E7 oncogene expression, but is also regulated extrinsically by factors emanating from the stroma, under the control of MMP-9.

Furthermore, the chimeric MMP-9/HPV16 mice, where the sole source of MMP-9 was from BM-d cells, showed an identical incidence of SCCs (50%) to wild-type (Figure 6C). In contrast, HPV16/MMP-9 –/– mice transplanted with MMP-9 –/– BM-d cells showed delayed neoplastic progression (data not shown) and less frequent cancers (Figure 6C). Intriguingly, HPV16/MMP-9 +/– mice lethally irradiated and transplanted with +/+ BM-d cells exhibited an increased incidence of SCCs (70%). The increased incidence was not a simple consequence of lethal whole-body irradiation since HPV16/MMP-9 –/– mice lethally irradiated and transplanted with MMP-9 –/– BM-d cells retained the ~20% incidence of tumors observed in their nonirradiated, nontransplanted counterparts. This effect deserves future investigation.

Discussion

This work has uncovered regulatory capabilities for MMP-9 during squamous carcinogenesis. We found that MMP-9 increases the rate and broadens the distribution of hyperproliferation of oncogene-expressing keratinocytes, enhances malignant conversion of dysplasias into frank carcinomas, and affects differentiation characteristics of emergent tumors. Furthermore, MMP-9 imparts these regulatory capabilities on oncogene-positive neoplastic cells as a paracrine factor, originating from reactive inflammatory cells conscripted to support neoplastic growth and progression.

Paracrine Cell Growth Regulation by MMP-9

MMPs, including MMP-2 and MMP-9, are expressed in a wide variety of human and animal cancers, and their roles have historically been associated with invasive and metastatic phenotypes, particularly in bioassays using cultured tumor cells (Nelson et al., 2000). In contradistinction, the use of transgenic mice lacking MMP-9 has clearly demonstrated a much different role for this MMP in a de novo squamous carcinogenesis model induced by the HPV16 oncogenes. HPV16 mice lacking MMP-9 have restricted keratinocyte hyperproliferation and delayed neoplastic progression, from hyperplasia to dysplasia to invasive cancer. The biphasic induction and amplification of angiogenesis in the dermis underlying hyperplasias and dysplasias is delayed. The overall incidence of cancers was reduced from 50% in HPV16 mice to 27%. Remarkably, the less frequent cancers, containing fewer proliferating keratinocytes, were more malignant, as assessed by classical squamous cancer grading criteria based on epithelial differentiation markers. Thus, during tumor progression, MMP-9 serves as a paracrine regulator, stimulating proliferation and malignant conversion of oncogene-positive keratinocytes, enhancing angiogenesis, and yet restricting tumor progression to more aggressive, poorly differentiated stages.

How can these effects be rationalized? Based on current knowledge, we can envisage several plausible scenarios. The pattern and percentage of proliferating keratinocytes in MMP-9-deficient/HPV16 mice suggests that MMP-9 could be mobilizing a poorly diffusible growth factor. In the absence of MMP-9, epithelial cell proliferation is restricted to one or two layers of cells most proximal to the stroma, both in dysplasias and in fronds of cancer cells, whereas with functional MMP-9, proliferating cells are distributed throughout both lesions. Perhaps growth factors tethered to the cell surface or stored in the matrix are released by MMP-9, enabling them to diffuse more widely, producing the observed pattern of uniform cell proliferation. Indeed, there is precedence for growth factor storage and release by proteinases (Werb, 1997; Bergers and Coussens, 2000; Yu and Stamenkovic, 2000). We anticipate that MMP-9 also regulates angiogenesis by altering the local balance of bio-available angiogenic growth factors. In another mouse model, of islet carcinoma of the pancreas (Hanahan, 1985), we have recently demonstrated that MMP-9 again helps regulate angiogenesis, and that it does so in part by rendering VEGF, an angiogenic inducer molecule, more available to its receptors on endothelial cells (Bergers et al., 2000). Notably, VEGF is also expressed in the HPV16 model (Smith-McCune et al., 1997). In both models, the absence of MMP-9 delays angiogenesis and reduces the frequency of tumors; since angiogenesis is eventually activated and tumors do form, cryptic compensatory mechanisms are implicated. This is clearly not the case for proliferation of the neoplastic keratinocytes since neither their proliferation index nor their pattern is restored with ensuing progression.

An alternative explanation to lack of growth factor mobilization for the reduced proliferation and altered differentiation seen in the absence of MMP-9 could lie in the nature of the target cell. Perhaps the oncogene-positive keratinocytes undergoing neoplastic progression and malignant conversion in a MMP-9 null environment arise from a more primitive squamous stem/progenitor cell, which might have different responses to the HPV16 oncogenes and distinct requirements for stromal support, with consequentially altered proliferation and differentiation profiles. Stem cells are believed to proliferate more slowly than their progeny and to retain embryonic, undifferentiated features (Weissman, 2000). Markers delineating epidermal stem/progenitor cells versus committed basal keratinocytes are still incompletely characterized (Jensen et al., 1999) but should allow this possibility to be addressed in the future.

Inflammatory Cells as Coconspirators in Carcinogenesis

A remarkable result of this study has come in the demonstration that bone marrow-derived inflammatory cells are a sufficient source of MMP-9 to manifest the contributions of this proteinase to the HPV16 squamous carcinogenesis pathway. It would not have been unreasonable to predict that MMP-9 is expressed by the developing cancer cells so as to control their own destiny, in the form of induction of angiogenesis and achievement of a diffuse hyperproliferative capability. But the data reveal otherwise. Although keratinocytes in cell culture can demonstrably express MMP-9 (reviewed in Coussens and Werb, 1996; Chambers and Matrisian, 1997), we found that MMP-9 was not made by keratinocytes in vivo; rather, cells in the stroma of hyperplasias, dysplasias, and cancers expressed MMP-9 mRNA and protein. Transfer of MMP-9 +/+ BM-d cells into MMP-9-deficient/HPV16 mice restored all phenotypes ascribed to this proteinase in squamous carcinogenesis: diffuse keratinocyte proliferation at all stages, abundant angiogenesis, and progression of invasive cancer in 50% of animals by 1 year; moreover, the predominance of low grade cancer was also restored. Thus, inflammatory cells are the critical suppliers of MMP-9 to this carcinogenesis pathway. But which bone marrow–derived cell type? In the present study, we observed that neutrophils were abundant in dysplasias and carcinomas. Mast cells were prevalent in hyperplasias, dysplasias, and at the fronts of invasive cancer, while macrophages were detected in tumors; all three cell types express MMP-9. T-lymphocytes, which also express MMP-9 (Weeks et al., 1993), were also present in neoplastic lesions in these mice (D. Daniels, L. M. C., and D. H., unpublished data). Each of these BM-d cells potentially contributes to the supply of MMP-9; conditional genetic ablation of individual cell types should enable their contributions to be assessed.

It is clear from our data that a major role for these inflammatory cells is to deliver MMP-9 to lesional tissue. This single gene product then orchestrates multiple parameters necessary for full manifestation of the malignant phenotype initiated by HPV16 oncogenes. Although these inflammatory cells may not be “transformed” by oncogenes or lose tumor suppressor genes, they clearly serve the developing neoplasm by amplifying both angiogenesis and neoplastic cell proliferation, and by affecting progression to malignant cancer (this work, and Coussens et al., 1999). It is important to appreciate, however, that BM-d cells are not the only paracrine players in the tumor microenvironment. Tumor fibroblasts can also contribute to cancer phenotypes (Boudreau and Bissell, 1998; Masson et al., 1998; Olumi et al., 1999), and endothelial cells are widely investigated as viable anti-cancer targets. The signals eliciting the inflammatory response in the neoplastic lesions remain to be elucidated.

The important contribution of inflammatory cells expressing MMP-9 revealed by this study raises the question of whether the cells or the factor represents a more tractable target for anti-cancer therapy. If our results, that MMP-9 enhances angiogenesis and tumor growth and is supplied by inflammatory cells, can be generalized to human cancer, we envisage that inflammatory cell types may come to be another important class of tumor-associated cells targeted by anti-cancer drugs. While MMP-9 is certainly a key gene product of these BM-d cells, it is likely not the only one. A caveat, of course, is the possibility that pharmacological suppression of inflammatory cells supplying MMP-9 function may also select for progression to more poorly differentiated cancers. That notwithstanding, it seems plausible that the efficacy of anti-inflammatory agents such as COX-2 inhibitors (Fosslien, 2000) being ascribed to angiogenesis inhibition and cancer chemoprevention is due to impairment of recruitment and/or function of inflammatory cells that supply MMP-9.

Thus a new perspective has emerged, that tumors are complex multicellular enterprises, wherein ostensibly normal cells are conscripted by transformed cancer cells to make significant contributions to the tumor phenotype (Boudreau and Bissell, 1998; Hanahan and Weinberg, 2000). In particular, this study has demonstrated that inflammatory cells, especially those expressing MMP-9, can be accomplices to neoplastic cells during squamous carcinogenesis.

Experimental Procedures

K14-HPV16 Transgenic and MMP-9 Null Mice and Tissue Preparation

The K14-HPV16 transgenic mice (Arbeit et al., 1994), the characterization of neoplastic stages based on keratin intermediate filament expression, and preparation of tissue sections for histologic examination have been previously reported (Coussens et al., 1996). Generation of MMP-9 homozygous null animals has been reported (Vu et al., 1998). MMP-9 +/– mice were bred four generations into FVB/n before intercrossing with HPV16 transgenic mice (N21 FVB/n), in turn producing MMP-9-deficient (–/–; n = 133) and MMP-9-proficient (+/–; n = 76) HPV16 mice.

Substrate Zymography

Tissue samples representing distinct histological stages of neoplastic progression, as verified by histological analysis of paraffin-embedded sections from adjacent skin, were weighed and then homogenized (1:4 weight to volume) in lysis buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% SDS. Soluble and insoluble extracts were separated by centrifugation and subsequently stored at –20°C. Equivalent amounts of soluble extract were analyzed by gelatin zymography (Herron et al., 1986a, b) on 10% SDS-polyacrylamide gels copolymerized with substrate (1 mg/ml of gelatin) in sample buffer (10% SDS, 0.25 M Tris-HCl, 0.1% Bromphenol Blue, pH 6.8). After electrophoresis, gels were washed twice for 15 min in 2.5% Triton X-100, incubated for 16 hr at 37°C in 50 mM Tris-HCl, 10 mM CaCl2 (pH 7.6), and then stained in 0.5% Coomassie Blue and destained in 50% methanol. Negative staining indicates the location of active protease bands. Exposure of proenzymes within tissue extracts to SDS during gel separation procedure leads to activation without proteolytic cleavage (Talhouk et al., 1991). For inhibition of MMP proteolytic activities, substrate gels were incubated in substrate buffer with 4 mM 1,10-phenanthroline (Sigma), as described (Adler et al., 1990). Data shown in Figure 1 are representative of results obtained following examination of 107 tissue pieces representing various stages of neoplastic progression in K14-HPV16 transgenic mice. Data shown in Figure 5C are representative of lysates generated from biopsies removed from individual mice (n = 10) from each respective cohort.

In Situ Hybridization

In situ hybridization analysis on 5 mm paraffin-embedded sections of normal and transgenic skin was carried out as previously described (Coussens et al., 1996). A linearized plasmid (pSP65- 92 KD-2) containing a 323 base pair fragment of the murine MMP-9 cDNA (Reponen et al., 1994) labeled with 35S was used as a probe. Slides were washed at a final stringency of 65°C in 2× SSC, dipped in photoemulsion, and exposed for 12 and 26 days. Slides were counterstained with hematoxylin and eosin.

Immunohistochemistry

Tissue pieces from transgenic animals were immersion-fixed in 3.75% paraformaldehyde and phosphate buffered saline (PBS) followed by dehydration through graded alcohols and xylene, and embedded in paraffin. 5–μm–thick paraffin sections were cut using a Leica 2135 microtome. Sections were deparaffinized and subjected to immunohistochemical staining as previously described (Coussens et al., 1996). Dilution used for rabbit anti-MMP-9 (generated as reported in Behrendtsen et al., 1992) antibody was 1:5000, 1:100 for mouse anti-BrdU (CalTag, Clone IU-4), 1:100 for mouse anti-PCNA (BioGenex, Clone # 19A2), 1:2500 for rabbit anti-keratin 10 (BabCo, Antibody MK10), 1:2500 for rabbit anti-keratin 14 (BabCo, Antibody MK14), 1:50 for rat anti-K8 (TROMA 1, a gift of Dr. Rolf Kemler, Germany), and 1:50 for rabbit anti-vimentin (DAKO, Clone Vim 3B4) in blocking solution containing PBS pH 7.4, 2.5% goat serum and 1% bovine serum albumin (BSA). Incubation with primary antibody was overnight at 4°C. Following incubation with a biotinylated secondary antibody (goat anti-rabbit IgG, goat anti-rat IgG, or goat anti-mouse IgM, 1:200, Pierce Chemical) for 30 min at ambient temperature, antigens were revealed with 3,3′-diaminobenzidine (DAB; Sigma,) according to manufacturer's instructions. Sections were counterstained in 1% methyl green (1 min), dehydrated in graded alcohols (70%, 95%, 100% ethanol), mounted in Cytoseal 60 (Stephens Scientific), and visualized with Nomarski optics. All immunolocalization experiments were repeated three times on multiple tissue sections and included negative controls for determination of background staining, which was negligible. All images were digitally captured on a Leica DMR microscope equipped with a Hamamatsu digital camera and imaged utilizing Improvision Open-Lab software.

Lethal Irradiation and Bone Marrow Transplantation

HPV16/MMP-9 +/– and HPV16/MMP-9 –/– mice were lethally X-irradiated with 7.5 Gray (Gy). 24 hr later, bone marrow–derived cells from either wild-type MMP-9-sufficient (+/+) mouse or MMP-9-deficient (–/–) mice were obtained by flushing the cavity of freshly dissected syngeneic femurs and tibias with PBS. Flushed cells were dispersed by pipetting, washed, and resuspended in PBS. Nucleated cells (1 × 106) in 50 ml were transplanted retro-orbitally into lethally irradiated animals, 10–12 animals per cohort. Neomycin at 2 mg/ml was added to the water of the irradiated mice. To verify engraftment, small (~1 mm2) punch biopsies were removed from ear skin at 3 and 5 months of age and cut in half. One half of each punch biopsy was subjected to gelatin zymographic analysis to test for MMP-9 activity. The other half of each punch biopsy was embedded in paraffin (see above) and used for histopathology.

Acknowledgments

We wish to thank Ole Behrendtsen, Alexis Scherl, C. Alex Hartman, Lidiya Korets, and Helen Capili for technical assistance; Dylan Daniels and Nicole Meyer-Morse for assistance with bone marrow transplantation; Alex McMillan for statistical advice; and Gabriele Bergers and Yves DeClerck for critical discussions and insightful comments on the manuscript. This work was supported by grants from the National Cancer Institute (CA72006 to Z. W. and CA 47632 and R37-CA37395 to D. H.) and the American Cancer Society (IRG-9715001 to L. C.); L. C. gratefully acknowledges startup support provided by the UCSF Comprehensive Cancer Center.

References

  1. Adler RR, Brenner CA, Werb Z. Expression of extracellular matrix-degrading metalloproteinases and metalloproteinase inhibitors is developmentally regulated during endoderm differentiation of embryonal carcinoma cells. Development. 1990;110:211–220. doi: 10.1242/dev.110.1.211. [DOI] [PubMed] [Google Scholar]
  2. Arbeit J, Munger K, Howley P, Hanahan D. Progressive squamous epithelial neoplasia in K14-human papillomavirus type 16 transgenic mice. J. Virol. 1994;68:358–368. doi: 10.1128/jvi.68.7.4358-4368.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Behrendtsen O, Alexander CM, Werb Z. Metalloproteinases mediate extracellular matrix degradation by cells from mouse blastocyst outgrowths. Development. 1992;114:447–456. doi: 10.1242/dev.114.2.447. [DOI] [PubMed] [Google Scholar]
  4. Bergers G, Coussens LM. Extrinsic regulators of epithelial tumor progression: metalloproteinases. Curr. Opin. Genet. Dev. 2000;10:120–127. doi: 10.1016/s0959-437x(99)00043-x. [DOI] [PubMed] [Google Scholar]
  5. Bergers G, Brekken R, McMahon G, Vu TH, Itoh T, Tamaki K, Tanzawa K, Thorpe P, Itohara S, Werb Z, Hanahan D. Matrix metalloproteinase-9 triggers the angiogenic switch during carcinogenesis. Nat. Cell Biol. 2000;2:737–744. doi: 10.1038/35036374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Boudreau N, Bissell MJ. Extracellular matrix signaling: integration of form and function in normal and malignant cells. Curr. Opin. Cell Biol. 1998;10:640–646. doi: 10.1016/s0955-0674(98)80040-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Broders A. Practical points on the microscopic grading of carcinoma. NY State J. Med. 1932;32:667–671. [Google Scholar]
  8. Chambers AF, Matrisian LM. Changing views of the role of matrix metalloproteinases in metastasis. J. Natl. Cancer Inst. 1997;89:1260–1270. doi: 10.1093/jnci/89.17.1260. [DOI] [PubMed] [Google Scholar]
  9. Cockett MI, Murphy G, Birch ML, O'Connell JP, Crabbe T, Millican AT, Hart IR, Docherty AJ. Matrix metalloproteinases and metastatic cancer. Biochem. Soc. Symp. 1998;63:295–313. [PubMed] [Google Scholar]
  10. Cordon-Cardo C, Prives C. At the crossroads of inflammation and tumorigenesis. J. Exp. Med. 1999;190:1367–1370. doi: 10.1084/jem.190.10.1367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Coussens LM, Werb Z. Matrix metalloproteinases and the development of cancer. Chem. Biol. 1996;3:895–904. doi: 10.1016/s1074-5521(96)90178-7. [DOI] [PubMed] [Google Scholar]
  12. Coussens LM, Hanahan D, Arbeit JM. Genetic predisposition and parameters of malignant progression in K14–HPV16 transgenic mice. Am. J. Pathol. 1996;149:1899–1917. [PMC free article] [PubMed] [Google Scholar]
  13. Coussens LM, Raymond WW, Bergers G, Laig-Webster M, Behrendtsen O, Werb Z, Caughey G, Hanahan D. Inflammatory mast cells up-regulate angiogenesis during squamous epithelial carcinogenesis. Genes Dev. 1999;13:1382–1397. doi: 10.1101/gad.13.11.1382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dvorak HF. Tumors: wounds that do not heal. N. Engl. J. Med. 1986;315:1650–1659. doi: 10.1056/NEJM198612253152606. [DOI] [PubMed] [Google Scholar]
  15. Fosslien E. Molecular pathology of cyclooxygenase-2 in neoplasia. Ann. Clin. Lab. Sci. 2000;1:3–21. [PubMed] [Google Scholar]
  16. Hanahan D. Heritable formation of pancreatic β-cell tumors in transgenic mice harboring recombinant insulin/simian virus 40 oncogenes. Nature. 1985;315:115–122. doi: 10.1038/315115a0. [DOI] [PubMed] [Google Scholar]
  17. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100:57–70. doi: 10.1016/s0092-8674(00)81683-9. [DOI] [PubMed] [Google Scholar]
  18. Herron GS, Werb Z, Dwyer K, Banda MJ. Secretion of metalloproteinases by stimulated capillary endothelial cells. I. Production of procollagenases and prostromelysin exceeds expression of proteolytic activity. J. Biol. Chem. 1986a;261:2810–2813. [PubMed] [Google Scholar]
  19. Herron GS, Banda MJ, Clark EJ, Gavrilovic J, Werb Z. Secretion of metalloproteinases by stimulated capillary endothelial cells. II. Expression of collagenase and stromelysin activities is regulated by endogenous inhibitors. J. Biol. Chem. 1986b;261:2814–2818. [PubMed] [Google Scholar]
  20. Jensen UB, Lowell S, Watt FW. The spatial relationship between stem cells and their progeny in the basal layer of human epidermis: a new view based on whole mount labeling and lineage analysis. Development. 1999;126:2409–2418. doi: 10.1242/dev.126.11.2409. [DOI] [PubMed] [Google Scholar]
  21. Lane E, Alexander C. Use of keratin antibodies in tumor diagnosis. Semin. Cancer Biol. 1990;1:165–179. [PubMed] [Google Scholar]
  22. Madlener M, Parks WC, Werner S. Matrix metalloproteinases (MMPs) and their physiological inhibitors (TIMPs) are differentially expressed during excisional skin wound repair. Exp. Cell Res. 1998;242:201–210. doi: 10.1006/excr.1998.4049. [DOI] [PubMed] [Google Scholar]
  23. Masson R, Lefebvre O, Noel A, Fahime ME, Chenard MP, Wendling C, Kebers F, LeMeur M, Dierich A, Foidart JM, et al. In vivo evidence that the stromelysin-3 metalloproteinase contributes in a paracrine manner to epithelial cell malignancy. J. Cell Biol. 1998;140:1535–1541. doi: 10.1083/jcb.140.6.1535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Nelson AR, Fingleton B, Rothenberg ML, Matrisian LM. Matrix metalloproteinases: biologic activity and clinical implications. J. Clin. Oncol. 2000;18:1135–1149. doi: 10.1200/JCO.2000.18.5.1135. [DOI] [PubMed] [Google Scholar]
  25. Olumi AF, Grossfeld GD, Hayward SW, Carroll PR, Tlsty TD, Cunha GR. Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium. Cancer Res. 1999;59:5002–5011. doi: 10.1186/bcr138. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Reponen P, Sahlberg C, Munaut C, Thesleff I, Tryggvason K. High expression of 92-kD type IV collagenase (gelatinase B) in the osteoclast lineage during mouse development. J. Cell Biol. 1994;124:1091–1102. doi: 10.1083/jcb.124.6.1091. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Smith-McCune K, Zhu Y-H, Hanahan D, Arbeit J. Cross-species comparison of angiogenesis during the premalignant stages of squamous carcinogenesis in the human cervix and K14–HPV16 transgenic mice. Cancer Res. 1997;57:1294–1300. [PubMed] [Google Scholar]
  28. Sternlicht MD, Lochter A, Sympson CJ, Huey B, Rougier J-P, Gray JW, Pinkel D, Bissell MJ, Werb Z. The stromal proteinase MMP3/stromelysin-1 promotes mammary carcinogenesis. Cell. 1999;98:137–146. doi: 10.1016/s0092-8674(00)81009-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Stetler-Stevenson WG. Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. J. Clin. Invest. 1999;103:1237–1241. doi: 10.1172/JCI6870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Talhouk RS, Chin JR, Unemori EN, Werb Z, Bissell MJ. Proteinases of the mammary gland: Developmental regulation in vivo and vectorial secretion in culture. Development. 1991;112:439–449. doi: 10.1242/dev.112.2.439. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Vu T, Werb Z. Gelatinase B: Structure, regulation and function. In: Parks WC, Mecham RP, editors. Matrix Metalloproteinases. Academic Press; San Diego, CA: 1998. pp. 115–148. [Google Scholar]
  32. Vu TH, Shipley JM, Bergers G, Berger J, Helms JA, Hanahan D, Shapiro SD, Senior R, Werb Z. MMP-9/gelatinase B is a key regulator of growth plate angiogenesis and apoptosis of hypertrophic chondrocytes. Cell. 1998;93:411–422. doi: 10.1016/s0092-8674(00)81169-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Weeks BS, Schnaper HW, Handy M, Holloway E, Kleinman HK. Human T lymphocytes synthesize the 92 kDa type IV collagenase (gelatinase B). J. Cell. Physiol. 1993;157:644–649. doi: 10.1002/jcp.1041570326. [DOI] [PubMed] [Google Scholar]
  34. Weissman IL. Stem cells: units of development, units of regeneration, and units in evolution. Cell. 2000;100:157–168. doi: 10.1016/s0092-8674(00)81692-x. [DOI] [PubMed] [Google Scholar]
  35. Werb Z. ECM and cell surface proteolysis: regulating cellular ecology. Cell. 1997;91:439–442. doi: 10.1016/s0092-8674(00)80429-8. [DOI] [PubMed] [Google Scholar]
  36. Wilson CL, Heppner KJ, Labosky PA, Hogan BL, Matri-sian LM. Intestinal tumorigenesis is suppressed in mice lacking the metalloproteinase matrilysin. Proc. Natl. Acad. Sci. USA. 1997;94:1402–1407. doi: 10.1073/pnas.94.4.1402. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Yu Q, Stamenkovic I. Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-β and promotes tumor invasion and angiogenesis. Genes Dev. 2000;14:163–176. [PMC free article] [PubMed] [Google Scholar]

RESOURCES