• View in gallery
    Figure 1—

    Representative results of western blot analysis of MMP-9 expression in neoplastic and clinically normal canine tissues. The 4 lanes included a molecular weight marker (lane 1), a sample of uterus from a clinically normal dog (lane 2), a sample of liver from a clinically normal dog (lane 3), and a sample of mammary gland carcinoma from a dog (lane 4). Notice the presence of the active form of MMP-9 (82 kDa) in all 3 tissues (although weakly in the liver). The band presenting the proenzyme (proMMP-9 [92 kDa]) was least marked in the mammary gland carcinoma tissue and most marked in the hepatic tissue.

  • View in gallery
    Figure 2—

    Representative photomicrographs of sections of a mammary gland tubulopapillary carcinoma (A, B, and C) and a mammary gland solid carcinoma (D, E, and F) obtained from 2 dogs following immunohistochemical staining for MMP-9 (A and D), uPA (B and E), and MIB-1 (C and F). A—Fibroblasts adjacent to the tubulopapillary carcinoma (arrowheads) have high MMP-9 expression. B—Expression of uPA in stromal fibroblasts (arrowheads) is high in this section. C—Nuclear labeling for MIB-1 is evident in the tubulopapillary carcinoma cells (arrows). D—Fibroblasts adjacent to the solid carcinoma (arrowheads) have low MMP-9 expression, and a few neoplastic cells are weakly immunoreactive for MMP-9 (arrows). E—Expression of uPA is scattered among a few stromal cells adjacent to the solid carcinoma (arrowheads). F—Nuclear labeling for MIB-1 is evident in a few solid carcinoma cells (arrows). Immunohistochemical stains; in all panels, bar = 50 μm.

  • View in gallery
    Figure 3—

    Kaplan-Meier overall survival time curves following surgical removal of 70 MMTs from 70 dogs and grouping on the basis of MMP-9 expression in stromal cells (24 of the 94 dogs were excluded because the follow-up period was not yet completed for 12 dogs, 7 dogs were lost to follow-up, and 5 dogs died of unknown reasons within the immediate postoperative period). Tumors were classified as having low (n = 45) or high (25) MMP-9 expression in stromal cells (ie, < 50% [solid line] or ≥ 50% [dashed line] of stromal cells with cytoplasmic staining, respectively). Vertical marks on each line represent censored dogs. The mean ± SE overall survival time for dogs with tumors with high MMP-9 expression in stromal cells (15.13 ± 1.7 months) was significantly (P < 0.001) lower than the mean overall survival time (22.22 ± 0.83 months) for dogs with tumors with low stromal MMP-9 expression.

  • View in gallery
    Figure 4—

    Kaplan-Meier disease-free interval curves following surgical removal of 74 MMTs from 74 dogs and grouping on the basis of MMP-9 expression in stromal cells (20 of the 94 dogs were excluded because the follow-up period was not yet completed for 8 dogs although they remained free of detectable metastases, 7 were lost to follow-up, and 5 died of unknown reasons within the immediate postoperative period). Tumors were classified as having low (n = 45) or high (29) MMP-9 expression in stromal cells (ie, < 50% [solid line] or > 50% [dashed line] of stromal cells with cytoplasmic staining, respectively). Vertical marks on each line represent censored dogs. The mean ± SE disease-free interval for dogs with tumors with high MMP-9 expression in stromal cells (12.94 ± 1.94 months) was significantly (P < 0.001) shorter than the mean disease-free interval for dogs with tumors with low stromal cell MMP-9 expression (20.55 ± 1.13 months).

  • 1.

    Singer CF, Kronsteiner N, Marton E, et al. MMP-2 and MMP-9 expression in breast cancer-derived human fibroblasts is differentially regulated by stromal-epithelial interactions. Breast Cancer Res Treat 2002; 72:6977.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 2.

    Pupa SM, Ménard S, Forti S, et al. New insights into the role of extracellular matrix during tumor onset and progression. J Cell Physiol 2002; 192:259267.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3.

    Sommerville RPT, Oblander SA, Apte SS. Matrix metalloproteinases: old dogs with new tricks. Genome Biol 2003; 4:216.1216.11.

  • 4.

    Sternlicht MD, Werb Z. How matrix metalloproteinases regulate cell behaviour. Ann Rev Cell Dev Biol 2001; 17:463516.

  • 5.

    Duffy MJ, Maguire TM, Hill A, et al. Metalloproteinases: role in breast carcinogenesis, invasion and metastasis. Breast Cancer Res 2000; 2:252257.

  • 6.

    Del Casar JM, González LO, Alvarez E, et al. Comparative analysis and clinical value of the expression of metalloproteinases and their inhibitors by intratumor stromal fibroblasts and those at the invasive front of the breast carcinomas. Breast Cancer Res Treat 2009; 116:3952.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7.

    Vizoso FJ, González LO, Corte MD, et al. Study of matrix metalloproteinases and their inhibitors in breast cancer. Br J Cancer 2007; 96:903911.

  • 8.

    González LO, Corte MD, Junquera S, et al. Expression and prognostic significance of metalloproteinases and their inhibitors in luminal A and basal-like phenotypes of breast cancer. Human Pathol 2009; 40:12241233.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9.

    Lana SE, Ogilvie GK, Hansen RA, et al. Identification of matrix metalloproteinases in canine neoplastic tissue. Am J Vet Res 2000; 61:111114.

  • 10.

    Yokota H, Kumata T, Taketaba S, et al. High expression of 92 KDa type IV collagenase (matrix metalloproteinase-9) in canine mammary adenocarcinoma. Biochym Biophys Acta 2001; 1568:712.

    • Search Google Scholar
    • Export Citation
  • 11.

    Hirayama K, Yokota H, Onai R, et al. Detection of matrix metalloproteinases in canine mammary tumours: analysis by immunohistochemistry and zymography. J Comp Pathol 2002; 127:249256.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12.

    Loukopoulos P, Mungall BA, Straw RC, et al. Matrix metalloproteinase-2 and −9 involvement in canine tumors. Vet Pathol 2003; 40:382394.

  • 13.

    Kawai K, Uetsuka K, Doi K, et al. The activity of matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) in mammary tumors of dogs and rats. J Vet Med Sci 2006; 68:105111.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14.

    Nowak M, Madej JA, Podhorska-Okolow M, et al. Expression of extracellular matrix metalloproteinase (MMP-9), E-cadherin and proliferation-associated antigen Ki-67 and their reciprocal correlation in canine mammary adenocarcinomas. In Vivo 2008; 22:463470.

    • Search Google Scholar
    • Export Citation
  • 15.

    Vinothini G, Balachandran C, Nagini S. Evaluation of molecular markers in canine mammary tumors: correlation with histological grading. Oncol Res 2009; 18:193201.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16.

    Amălinei C, Căruntu I, Bălan RA. Biology of metalloproteinases. Rom J Morphol Embryol 2007; 48:323334.

  • 17.

    Brooks SA, Lomax-Browne HJ, Carter TM, et al. Molecular interactions in cancer cell metastasis. Acta Histochem 2010; 112:325.

  • 18.

    Kessenbrock K, Plaks V, Werb Z. Matrix metalloproteinases: regulators of tumor microenvironment. Cell 2010; 141:5267.

  • 19.

    Jodele S, Blavier L, Yoon JM. Modifying the soil to affect the seed: role of stromal-derived matrix metaloproteinases in cancer progression. Cancer Metastasis Rev 2006; 25:3543.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20.

    Pellikainen JM, Ropponen KM, Kataja VV, et al. Expression of matrix metalloproteinase (MMP-2) and MMP-9 in breast cancer with a special reference to activator protein-2, HER-2, and prognosis. Clin Cancer Res 2004; 10:76217628.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21.

    Matos AJF, Lopes C, Faustino AMR, et al. MIB-1 indices according to clinico-pathological variables in canine mammary tumours: a multivariate study. Anticancer Res 2006; 26:18211826.

    • Search Google Scholar
    • Export Citation
  • 22.

    Misdorp W, Else RW, Helmén E, et al. Histological classification of the mammary tumours of the dog and the cat. In: Shulman FL, ed. World Health Organization international histological classification of tumours of domestic animals. 2nd ed. Washington DC: Armed Forces Institute of Pathology, 1999;1629.

    • Search Google Scholar
    • Export Citation
  • 23.

    Elston CW, Ellis IO. Assessment of histological grade. In: Rosen PP, ed. Rosen's breast pathology. 3rd ed. Philadelphia: Lippincott-Raven Publishers, 1998;365382.

    • Search Google Scholar
    • Export Citation
  • 24.

    Hsu S, Raine L, Fanger H. The use of antiavidin antibody and avidin-biotin-peroxidase complex in immunoperoxidase techniques. Am J Clin Pathol 1981; 75:816821.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25.

    Santos A, Lopes C, Marques RM, et al. Immunohistochemical analysis of urokinase plasminogen activator and its prognostic value in canine mammary tumours. Vet J 2011; 189:4348.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 26.

    Agresti A. Inference for contingency tables. In: Balding DJ, Bloomfield P, Cressie NAC, eds. Categorical data analysis. 2nd ed. Hoboken, NJ: Wiley, 2002;9197, 369373.

    • Search Google Scholar
    • Export Citation
  • 27.

    Klein M, Kleinbaum DG. Kaplan-Meier survival curves and the log-rank test. In: Gail M, Krickeberg K, Samet J, et al, eds. Survival analysis—a self-learning text. 2nd ed. New York: Springer, 2005;4582.

    • Search Google Scholar
    • Export Citation
  • 28.

    Köhrmann A, Kammerer U, Kapp M, et al. Expression of matrix metalloproteinases (MMPs) in primary human breast cancer and breast cancer cell lines: new findings and review of the literature. BMC Cancer [serial online]. 2009; 9:188. Available at: www.biomedcentral.com/1471-2407/9/188. Accessed Oct 10, 2010.

    • Search Google Scholar
    • Export Citation
  • 29.

    Owen JL, Iragavarapu-Charyulu V, Gunja-Smith Z, et al. Up-regulation of matrix metalloproteinase-9 in T lymphocytes of mammary tumour bearers: role of vascular endothelial growth factor. J Immunol 2003; 171:43404351.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30.

    Tětu B, Trudel D, Wang CS. Proteases des cellules stromales réactionnelles du cancer: une cible thérapeutique attrayante. Bull Cancer 2006; 93:944948.

    • Search Google Scholar
    • Export Citation
  • 31.

    MacDougall JR, Matrisian LM. Contributions of tumor and stromal matrix metalloproteinases to tumor progression, invasion and metastasis. Cancer Metastasis Rev 1995; 14:351362.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 32.

    Micke P, Östman A. Tumour-stroma interaction: cancer-associated fibroblasts as novel targets in anti-cancer therapy? Lung Cancer 2004; 45:S163S175.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 33.

    Klein T, Bischoff R. Physiology and pathophysiology of matrix metalloproteinases. Amino Acids 2011; 41:271290.

  • 34.

    Kim HJ, Park C, Park BW, et al. Expression of MT-1 MMP, MMP2, MMP9 and TIMP2 mRNAs in ductal carcinoma in situ and invasive ductal carcinoma of the breast. Yonsei Med J 2006; 47:333342.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 35.

    Jobim FC, Xavier NL, Uchoa DM, et al. Prevalence of vascular-endothelial growth factor, matrix metalloproteinases and tissue inhibitors of metalloproteinases in primary breast cancer. Braz J Med Biol Res 2009; 42:979987.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 36.

    Peña LL, Nieto AI, Pérez-Alenza D, et al. Immunohistochemical detection of Ki-67 and PCNA in canine mammary tumors: relationship to clinical and pathologic variables. J Vet Diagn Invest 1998; 10:237246.

    • Crossref
    • Search Google Scholar
    • Export Citation

Advertisement

Matrix metalloproteinase-9 expression in mammary gland tumors in dogs and its relationship with prognostic factors and patient outcome

Andreia A. SantosDepartment of Veterinary Clinics, University of Porto, 4099-003 Porto, Portugal.
Abel Salazar Biomedical Sciences Institute and the Multidisciplinary Unit for Biomedical Research, University of Porto, 4099-003 Porto, Portugal.

Search for other papers by Andreia A. Santos in
Current site
Google Scholar
PubMed
Close
 DVM
,
Célia C. LopesDepartment of Veterinary Pathology, University of Porto, 4099-003 Porto, Portugal.

Search for other papers by Célia C. Lopes in
Current site
Google Scholar
PubMed
Close
 BSc
,
Raquel M. MarquesDepartment of Anatomy, University of Porto, 4099-003 Porto, Portugal.

Search for other papers by Raquel M. Marques in
Current site
Google Scholar
PubMed
Close
,
Irina F. AmorimDepartment of Veterinary Pathology, University of Porto, 4099-003 Porto, Portugal.

Search for other papers by Irina F. Amorim in
Current site
Google Scholar
PubMed
Close
 DVM
,
Maria F. GärtnerDepartment of Veterinary Pathology, University of Porto, 4099-003 Porto, Portugal.
Institute of Molecular Pathology and Immunology, University of Porto, 4200-465 Porto, Portugal.

Search for other papers by Maria F. Gärtner in
Current site
Google Scholar
PubMed
Close
 DVM, PhD
, and
Augusto J. F. de MatosAbel Salazar Biomedical Sciences Institute and the Multidisciplinary Unit for Biomedical Research, University of Porto, 4099-003 Porto, Portugal.

Search for other papers by Augusto J. F. de Matos in
Current site
Google Scholar
PubMed
Close
 DVM, PhD

Abstract

Objective—To immunohistochemically evaluate matrix metalloproteinase (MMP)-9 expression in benign and malignant mammary gland tumors (MMTs) in dogs and relate expression to prognostic factors and patient outcome.

Animals—118 female dogs with naturally occurring mammary gland tumors and 8 dogs without mammary gland tumors.

Procedures—24 benign mammary gland tumors and 94 MMTs (1/affected dog) were obtained during surgical treatment; control mammary gland tissue samples were collected from unaffected dogs after euthanasia for reasons unrelated to the study. Tumors were evaluated for proliferation, invasive growth, histologic grade, and metastatic capacity; expression of MMP-9 was determined immunohistochemically, and its relationship with clinical and histologic findings was investigated. For dogs with MMTs, follow-up continued for 2 years; data were used to compute overall survival time and disease-free interval and construct survival curves.

Results—MMTs had significantly higher MMP-9 expression in stromal cells and in neo-plastic cells than did the benign neoplasms. Stromal MMP-9 expression was also higher in highly proliferative tumors and in tumors with invasive growth, high histologic grade, and metastatic capacity. Furthermore, tumors from patients with shorter overall survival times and disease-free intervals had higher expression of MMP-9 in stromal cells.

Conclusions and Clinical Relevance—In dogs with MMTs, level of MMP-9 expression by stromal cells was related to factors of poor prognosis and shorter overall survival times and disease-free intervals. These results suggested that MMP-9 produced by tumor-adjacent stromal cells contributed to MMT progression in female dogs and that assessment of MMP-9 expression may be a valuable prognostic factor.

Abstract

Objective—To immunohistochemically evaluate matrix metalloproteinase (MMP)-9 expression in benign and malignant mammary gland tumors (MMTs) in dogs and relate expression to prognostic factors and patient outcome.

Animals—118 female dogs with naturally occurring mammary gland tumors and 8 dogs without mammary gland tumors.

Procedures—24 benign mammary gland tumors and 94 MMTs (1/affected dog) were obtained during surgical treatment; control mammary gland tissue samples were collected from unaffected dogs after euthanasia for reasons unrelated to the study. Tumors were evaluated for proliferation, invasive growth, histologic grade, and metastatic capacity; expression of MMP-9 was determined immunohistochemically, and its relationship with clinical and histologic findings was investigated. For dogs with MMTs, follow-up continued for 2 years; data were used to compute overall survival time and disease-free interval and construct survival curves.

Results—MMTs had significantly higher MMP-9 expression in stromal cells and in neo-plastic cells than did the benign neoplasms. Stromal MMP-9 expression was also higher in highly proliferative tumors and in tumors with invasive growth, high histologic grade, and metastatic capacity. Furthermore, tumors from patients with shorter overall survival times and disease-free intervals had higher expression of MMP-9 in stromal cells.

Conclusions and Clinical Relevance—In dogs with MMTs, level of MMP-9 expression by stromal cells was related to factors of poor prognosis and shorter overall survival times and disease-free intervals. These results suggested that MMP-9 produced by tumor-adjacent stromal cells contributed to MMT progression in female dogs and that assessment of MMP-9 expression may be a valuable prognostic factor.

Tissue remodeling is a key factor in the invasion and metastasis of malignant tumors.1 Invasion of the surrounding tissues by tumor cells requires degradation of the ECM and basement membrane components, which is mediated by several proteolytic systems, including zinc-dependent MMPs.2

The MMP family is composed of several substrate-specific subfamilies; gelatinases are one of the subfamilies frequently related to tumor invasion and metastasis and include MMP-2 (gelatinase A) and MMP-9 (gelatinase B).3 Gelatinases degrade type IV collagen of the basement membrane, thereby breaking the structural barriers against invasion and contributing to tumor growth and spread.4 In several human breast cancer studies, high MMP-9 expression has been associated with MMT progression5,6 and metastasis.7,8 In dogs with MTs, MMP-9 activity or expression has been associated with tumor malignancy,9–13 proliferation rate,14 and histologic grade.15

Matrix metalloproteinases are produced in the inactive zymogen form (as pro-MMPs), and their in vivo proteinase action depends on the activation by several proteases, such as plasmin and other MMPs.16 Urokinase-type plasminogen activator is one of the most frequently implicated proteases in the activation of several pro-MMPs, including pro-MMP-9.16,17

The multiple effects of MMPs on neoplastic diseases are attributable to the MMPs' ability to regulate the tumor microenvironment by cleaving various substrates, including structural components of the ECM, growth factor-binding proteins and growth factor precursors, receptor tyrosine kinases, cell adhesion molecules, and other proteases.17 Thus, MMPs are instrumental in the molecular crosstalk between tumor and stromal cells.18

The exact location of the production of these enzymes in tumors is still debated. There is a growing body of evidence that, in humans, MMP-9 originates essentially from tumor-adjacent stromal components of many cancers, including breast carcinoma.19 In patients with breast carcinoma, high stromal MMP-9 expression is associated with poor prognosis.1,8,20

To date, immunohistochemical investigations11,14 of MMP-9 in MTs in dogs have focused on its expression by neoplastic cells, with no evaluation of stromal cell expression. Hence, the influence of stromal cell-derived MMP-9 on the biological behavior of MTs in dogs is still unknown. Furthermore, to our knowledge, there are no reports of survival studies of dogs with MTs in which tumor expression of MMP-9 was investigated. Therefore, the main purpose of the study reported here was to evaluate the expression of MMP-9 in BMTs and MMTs in dogs and relate MMP-9 expression to prognostic factors and patient outcome.

Materials and Methods

Dogs and specimen collection—One hundred eighteen client-owned sexually intact or spayed female dogs with naturally occurring MTs were included in the study. For inclusion in the study, each dog had to have ≥ 1 MTs with no evidence of distant metastasis at the time of diagnosis, have no previous history of tumor disease, be free of tumors other than MT at the time of diagnosis, and have an owner who consented to surgery with curative intents but declined postoperative adjuvant treatments. Owners also consented to postoperative evaluation of their dog for a period of 2 years.

During surgical treatment, MTs were removed and 1 entire tumor from each dog was selected for purposes of this investigation. Regional lymph nodes were collected from 72 dogs with malignant tumors. In nearly half of those cases, intramammary lymph nodes (1 to 7 nodes/dog) were identified and examined.

In addition, samples of mammary gland tissues, uterus, and bone were obtained from 8 dogs without MTs that had been euthanized. Euthanasia was performed via slow IV injection of pentobarbital sodium solution at a municipal kennel as part of the national stray dog control program.

The study protocol was performed in compliance with institutional guidelines for research on animals and approved by the Animal Care and Ethics Committee of the University of Porto. All diagnostic or therapeutic procedures were in the best interest of the patients and approved by the owners. No procedures were conducted for purposes of the study alone.

Characterization of MTs—All tumors and clinically normal mammary gland tissues were fixed in neutral-buffered 10% formalin for 48 hours. The largest cross-sectional diameter of each tumor was recorded and categorized as either < 3 or ≥ 3 cm. Tumors ≤ 1 cm in diameter were paraffin embedded in 1 block, and larger tumors were cut sequentially at 5-mm intervals to provide a series of tissue blocks representative of the entire lesion. The samples of clinically normal mammary glands were also cut sequentially at 5-mm intervals and paraffin embedded.

After dehydration and embedding in paraffin wax, 3-μm sections were cut from each block. For diagnostic purposes, 1 section/block was stained with H&E stain. One adjacent section to the H&E-stained section from each block was selected for immunohistochemical analysis; for tumors > 1 cm, a section from each of 2 representative blocks was collected. When available, local and regional lymph nodes were processed and examined as previously described.21

The H&E-stained sections were evaluated independently by 2 observers (MFG and IFA) according to the criteria of the World Health Organization for the histologic classification of mammary gland tumors of domestic animals.22 The histologic classification of the MTs was performed as part of this study. Histologic grading of each tumor was performed according to the Nottingham grading method for human breast tumors,23 and tumors were classified as grade I (well differentiated), grade II (moderately differentiated), or grade III (poorly differentiated). The mode of growth of each tumor was assessed and classified as expansive (cohesive and well-delimited growth of the tumor mass pushing normal surrounding tissue) or invasive (infiltrative growth or lymphatic or blood vessel invasion). In dogs with > 1 MT, the tumor with more aggressive clinicopathologic features was selected for analysis.

Immunohistochemical analysis of MMP-9 expression—Tumors and clinically normal mammary glands sections adjacent to those that underwent H&E staining were analyzed immunohistochemically by use of the modified avidin-biotin-peroxidase complex method.24 After sections were dewaxed and rehydrated, endogenous peroxidase activity was blocked by treating the sections with 3% hydrogen peroxide in methanol for 10 minutes. Slides were then incubated with rabbit seruma for 20 minutes at room temperature (20°C) and then with the anti–MMP-9 (C-20) goat polyclonal antibodyb (diluted 1:200 in Tris-buffered saline solution with 5% bovine serum albumin), overnight at 4°C in a humid chamber. Sections were then incubated with biotinylated rabbit anti-goat antibodyb (diluted 1:100) for 30 minutes, followed by incubation with avidin-biotin-peroxidase complexc for a further 30 minutes. Color was developed with a solution of 3,3′-diaminobenzidine, and sections were then counterstained with hematoxylin, dehydrated, and mounted. In negative controls, the primary antibody was replaced with nonimmune goat immunoglobulin to confirm the specificity of the immunohistochemical staining. Positive controls consisted of sections of human breast cancer tissue (samples from an invasive ductal carcinomad) known to express MMP-9 and clinically normal canine tissues (uterus and bone) collected from the euthanized stray dogs).12

For each tumor, MMP-9 expression was evaluated in a semiquantitative manner. Each section was examined microscopically by 2 independent observers (AAS and AJFD), who each estimated the percentage of tumor-adjacent stromal cells (fibroblasts) with cytoplasmic staining and the percentage of neoplastic cells with cytoplasmic staining. Tumors were classified as having low or high MMP-9 expression in stromal cells (ie, < 50% or ≥ 50% of stromal cells with cytoplasmic staining, respectively) and low or high MMP-9 expression in neoplastic cells (ie, < 25% or ≥ 25% of neoplastic cells with cytoplasmic staining, respectively). In tumors > 1 cm, both sections were examined and the consensus estimation, in a global perspective, was considered instead of the 2 different individual estimations. When the estimations of the 2 observers were dissimilar (< 5% of the cases), a consensus was obtained from the observers by use of a multiheaded microscope.

Immunohistochemical analysis of MIB-1 and uPA expressions—Sequential tumor sections of the MMTs were also immunostained and evaluated for MIB-1 labeling index (Ki-67 expression in neoplastic cells) and for uPA expression (in stromal cells), according to the previously described methods.21,25 The MIB-1 labeling index was determined by counting the number of immunopositive cells in 1,000 neoplastic cells (at 400× magnification) in the area of highest labeling, and the index was expressed as a percentage. Tumors were classified as having low or high MIB-1 expression in neoplastic cells (ie, < 43% or ≥ 43% of neoplastic cells with MIB-1-specific cytoplasmic staining). The expression of uPA was assessed in a semiquantitative manner. Sections from MMT were examined microscopically, and the percentage of stromal cells (fibroblasts) with cytoplasmic staining for uPA was estimated. Tumors were classified as having low or high uPA expression in stromal cells (ie, < 10% or ≥ 10% of stromal cells with uPA-specific cytoplasmic staining). In tumors > 1 cm, both sections were examined and the consensus estimation, in a global perspective, was considered instead of the 2 individual estimations.

SDS-PAGE and western blot analysis of MMP-9—To confirm the specificity of the antibody for canine MMP-9 protein, canine tissues samples (mammary gland carcinoma and clinically normal uterine and hepatic tissues collected from the euthanized stray dogs) were evaluated by western blot analysis. Cytosolic fractions of samples were obtained by homogenization and incubation in a lysis buffere containing protease inhibitorsf at room temperature over a period of 10 minutes. Lysates were centrifuged at 10,000 × g at 4°C for 5 minutes. Supernatants were then collected and stored at −80°C.

For western blot analysis, cytosolic fractions (55 mg of total protein/well) contained on a 10% SDS gel were blotted for 3.5 hours (40 V and 300 mA) into a polyvinylidene difluoride membrane.g The membrane was incubated overnight in a buffer of PBS-Tween 20 solution containing 3% bovine serum albumin to block nonspecific binding and further incubated for 3 hours at room temperature with the anti-MMP-9 (C-20) goat polyclonal IgGb diluted 1:75 in PBS-Tween 20 solution. After the membrane was washed in PBS-Tween 20 solution (three 5-minute washes and two 10-minute washes), it was incubated for 45 minutes with horseradish peroxidase–coupled polyclonal donkey anti-goat IgGb diluted 1:5,000 in PBS-Tween 20 solution. Bound antibodies were detected by use of a 3,3′,5,5′ tetramethlylbenzidineh liquid substrate system for membranes.

Follow-up assessment of dogs with MMTs—Dogs were routinely examined prior to surgery, 3 weeks after surgery, and every 3 months thereafter for a 2-year period if no clinical signs were detected. Owners were instructed to report and discuss with the researchers any detected abnormalities, even if those abnormalities were not obviously related to the MTs, at any time. Each examination consisted of a complete physical examination, thoracic radiography (3 views), and complete abdominal ultrasonographic examination. In case of the development of new skin or mammary gland nodules, lymph nodes alterations, or any other organ abnormalities, appropriate additional examinations (eg, cytologic examination of fine-needle aspiration specimens, histologic examination of excisional biopsy specimens, or skeletal radiography) were performed with the owner's consent to rule out other neoplasms or confirm metastatic disease. Complete necropsies were performed in all dogs that died or were euthanized (via IV administration of sodium pentobarbital) during the 2-year follow-up period. Metastases were histologically confirmed.

Statistical analysis—The Fisher exact test was used to analyze differences in stromal and neoplastic cell expressions of MMP-9 between benign and malignant MTs and to evaluate the relationship between MMP-9 expression in MMTs and tumor size, mode of growth, MIB-1 labeling index, regional lymph node status, and distant metastasis.26 The relationship between MMP-9 expression and histologic grade of the tumors was analyzed by use of the Pearson χ2 test.26

For dogs with MMTs, overall survival time was calculated from the date of tumor removal to the date of the dog's death or euthanasia due to tumor metastasis. Disease-free interval among dogs with MMTs was calculated from the date of surgery to the date of detection of the first local recurrence or development of distant metastases. The Kaplan-Meier method was used to compute overall survival time and disease-free interval and to construct the survival curves.27 A log-rank test was used to analyze the significance of differences between groups.27 In the overall survival time assessment, dogs were censored when they died of causes unrelated to MTs, were lost to follow-up, or were alive 2 years after surgery. In the disease-free interval assessment, dogs were censored when they were lost during follow-up, died for causes unrelated to MTs before developing signs of metastatic disease, or were free of distant metastases 2 years after surgery. Values of P < 0.05 were considered significant. Statistical analysis was performed by use of statistical software.i

Results

Dogs and MTs—Among the 118 dogs with MTs, 24 dogs had only benign tumors. Of those 24 dogs, 8 had 1 tumor, 5 had 2 tumors, 7 had 3 tumors; 1 dog each had 4, 5, 8, or 9 tumors. Ninety-four dogs had malignant tumors. Of those 94 dogs, 62 had 1 tumor, 23 had 2 tumors, 6 had 3 tumors, 2 had 4 tumors, and 1 had 5 tumors. One tumor from each dog was selected and analyzed in the study. The selected BMTs included 6 simple adenomas, 11 complex adenomas, and 7 benign mixed tumors. The selected MMTs included 31 solid carcinomas, 21 complex carcinomas, 20 tubulopapillary carcincomas, 2 micropapillary carcinomas, 2 mucinous carcinomas, 2 anaplastic carcinomas, 1 spindle cell carcinoma, 13 carcinosarcomas, and 2 carcinomas in benign tumors (a malignant MT type with foci or distinct nodules of malignant cells in complex adenomas or benign mixed tumors). Because of the low number of some histologic types, MMTs were grouped as follows: complex carcinomas (n = 21), simple carcinomas (solid, tubulopapillary, anaplastic, and micropapillary carcinomas [55]), and others (carcinosarcomas, carcinomas in benign tumor, spindle cell carcinomas, and mucinous carcinomas [18]). Among the BMTs, 19 were < 3 cm in diameter and 5 were ≥ 3 cm in diameter. Among the MMTs, 48 were < 3 cm in diameter and 45 were ≥ 3 cm in diameter.

Sections of each MMT were examined histologically to determine mode of growth and histologic grade. Among the 94 MMTs, 24 were considered expansive and 70 were considered invasive. Histologic grades I, II, and III were assigned to 21, 43, and 30 MMTs, respectively. Presence or absence of regional lymph node metastases at the time of surgery was also determined histologically for 72 dogs with MMTs; lymph nodes were not obtained from 22 dogs with MMTs.

Clinically normal mammary gland tissue samples—Following euthanasia of 8 dogs without any palpable mammary gland masses, 8 apparently normal mammary glands were collected. Histologic examination of H&E-staining sections of those tissues confirmed the absence of tissue abnormalities.

Immunohistochemical and western blot analysis of MMP-9 expression—Sections of canine bone and uterus (positive controls) had intense MMP-9 immunoreactivity in bone osteoclasts and in the myometrium, as expected. Sections of tumors that were treated with nonimmune goat immunoglobulin instead of the anti–MMP-9 (C-20) goat polyclonal antibody (negative controls) did not have any MMP-9–specific staining. Western blot analysis revealed the presence of an 82-kDa band mostly in mammary gland carcinomas and uterus, whereas hepatic tissue had weak bands at 82 and 92 kDa (Figure 1). These bands corresponded to the active (82 kDa) and inactive (proenzyme) forms of canine MMP-9.9,11,12

Figure 1—
Figure 1—

Representative results of western blot analysis of MMP-9 expression in neoplastic and clinically normal canine tissues. The 4 lanes included a molecular weight marker (lane 1), a sample of uterus from a clinically normal dog (lane 2), a sample of liver from a clinically normal dog (lane 3), and a sample of mammary gland carcinoma from a dog (lane 4). Notice the presence of the active form of MMP-9 (82 kDa) in all 3 tissues (although weakly in the liver). The band presenting the proenzyme (proMMP-9 [92 kDa]) was least marked in the mammary gland carcinoma tissue and most marked in the hepatic tissue.

Citation: American Journal of Veterinary Research 73, 5; 10.2460/ajvr.73.5.689

In clinically normal mammary gland samples, only keratinocytes from epidermis and the smooth muscle layer of vascular structures were immunopositive for MMP-9; the epithelial and myoepithelial cells and the stromal fibroblasts were negative for MMP-9 expression. The staining of both keratinocytes and vascular smooth muscle were used as internal positive controls.

The expression patterns of MMP-9 in tumor-adjacent stromal cells (fibroblasts) and in neoplastic cells were cytoplasmic (Figure 2). Malignant MTs had significantly higher MMP-9 expression than did BMTs (Table 1); in each BMT, < 50% of the stromal cells were MMP-9 positive, which was the rationale for use of 50% as a cutoff in the grouping of MMTs on the basis of stromal expression. Neoplastic cell reactivity was not as remarkable (both in percentage and staining intensity) as was stromal cell expression in both BMT and MMTs. The cutoff of 25% for neoplastic cell reactivity was considered to be representative of the detected differences between BMTs and MMTs.

Figure 2—
Figure 2—

Representative photomicrographs of sections of a mammary gland tubulopapillary carcinoma (A, B, and C) and a mammary gland solid carcinoma (D, E, and F) obtained from 2 dogs following immunohistochemical staining for MMP-9 (A and D), uPA (B and E), and MIB-1 (C and F). A—Fibroblasts adjacent to the tubulopapillary carcinoma (arrowheads) have high MMP-9 expression. B—Expression of uPA in stromal fibroblasts (arrowheads) is high in this section. C—Nuclear labeling for MIB-1 is evident in the tubulopapillary carcinoma cells (arrows). D—Fibroblasts adjacent to the solid carcinoma (arrowheads) have low MMP-9 expression, and a few neoplastic cells are weakly immunoreactive for MMP-9 (arrows). E—Expression of uPA is scattered among a few stromal cells adjacent to the solid carcinoma (arrowheads). F—Nuclear labeling for MIB-1 is evident in a few solid carcinoma cells (arrows). Immunohistochemical stains; in all panels, bar = 50 μm.

Citation: American Journal of Veterinary Research 73, 5; 10.2460/ajvr.73.5.689

Table 1—

Expression* of MMP-9 (determined semiquantitatively via immunohistochemical staining) in tumor-adjacent stromal and neoplastic cells in BMTs and MMTS removed surgically from 118 dogs.

  No. (%) of tumors with MMP-9–positive stromal cellsNo. (%) of tumors with MMP-9–positive neoplastic cells
Tumor typeTotal No. of tumorsLow expressionHigh expressionP valueLow expressionHigh expressionP value
Characteristic   < 0.001  0.023
 BMT2424 (100.0)0 20 (83.3)4 (16.7) 
 MMT9458 (61.7)36 (38.3) 56 (59.6)38 (40.4) 

Sections of tumors were stained immunohistochemically for MMP-9 and examined microscopically (1 section was examined in tumors < 1 cm and 2 sections were examined in tumors > 1 cm; in tumors > 1 cm, both slides were examined, and the global estimation was considered instead of the 2 individual estimations). Tumors were classified as having low or high MMP-9 expression in stromal cells (ie, < 50% or ≥ 50% of stromal cells with cytoplasmic staining, respectively) and as having low or high MMP-9 expression in neoplastic cells (ie, < 25% or ≥ 25% of neoplastic cells with cytoplasmic staining, respectively).

Matrix metalloproteinase-9 expression by stromal cells was not significantly associated with histologic type (Table 2), although complex carcinomas had lower (P = 0.033) stromal cell MMP-9 expression than did the other MMT groups considered together. In contrast, there was a significant association between histologic type and neoplastic cell expression of MMP-9. Complex carcinomas had a lower (P = 0.004) neoplastic cell expression of MMP-9, compared with findings for the other MMT groups considered together.

Table 2—

Relationships between MMP-9 expression* (determined semiquantitatively via immunohisto-chemical staining) in stromal and neoplastic cells in MMTs removed surgically from 94 dogs and tumor characteristics (histologic type, mode of growth, histologic grade, and MIB-1 labeling index), regional lymph node and distant metastases, and expression of uPA in tumor-adjacent stromal cells.

  No. (%) of tumors with MMP-9–positive stromal cells 
CharacteristicTotal No. of tumorsLow expressionHigh expressionP valueLow expressionHigh expressionP value
Tumor histologic type   NS  0.004
 Complex carcinoma2117 (81.0)4 (19.0) 18 (85.7)3(14.3) 
 Simple carcinoma5531 (56.4)24 (43.6) 32 (58.2)23(41.8) 
 Other1810 (55.6)8 (44.4) 6 (33.3)12(66.7) 
Tumor diameter   0.001  NS
 < 3 cm4837 (77.1)11 (22.9) 30(62.5)18(37.5) 
 ≥ 3 cm4520 (44.4)25 (55.6) 25(55.6)20 (44.4) 
Mode of growth   0.034  NS
 Expansive2419 (79.2)5 (20.8) 13(54.2)11 (45.8) 
 Invasive7039 (55.7)31 (44.3) 43(61.4)27 (38.6) 
Histologic grade   0.005  NS
 I2119 (90.5)2 (9.5) 15(71.4)6(28.6) 
 II4325 (58.1)18 (41.9) 27 (62.8)16(37.2) 
 III3014 (46.7)16 (53.3) 14(46.7)16(53.3) 
Regional lymph node   0.032  0.006
 metastases§       
 Absent4936 (73.5)13 (26.5) 36 (73.5)13(26.5) 
 Present2311 (47.8)12 (52.2) 9(39.1)14(60.9) 
Distant metastases   0.014  NS
 Absent5236 (69.2)16 (30.8) 29 (55.8)23 (44.2) 
 Present239 (39.1)14 (60.9) 14(60.9)9(39.1) 
MIB-1 labeling index   0.010  NS
 Low5439 (72.2)15 (27.8) 34 (63.0)20 (37.0) 
 High3918 (46.2)21 (53.8) 22 (56.4)17(43.6) 
uPA expression in stromal cells#   < 0.001  NS
 Low expression2624 (92.3)2 (7.7) 33 (75.0)11(25.0) 
 High expression6834 (50.0)34 (50.0) 43(58.1)31 (41.9) 

Because of the low number of some histologic types, MMTs were grouped as follows: complex carcinomas, simple carcinomas (solid, tubulopapillary, anaplastic, and micropapillary carcinomas), and others (carcinosarcomas, carcinomas in benign tumors, spindle cell carcinomas, and mucinous carcinomas).

In 1 case, tumor size was impossible to measure because it was multifocal throughout the mammary gland chain.

Lymph nodes from 22 dogs were not submitted for examination (1 regional lymph node/dog).

The presence of distant metastases was assessed via physical examination, 3-view thoracic radiography, and abdominal ultrasonography, and confirmed via cytologic or histologic analysis. At the end of the study, 12 dogs were alive and still undergoing follow up; 7 dogs were lost to follow-up.

Tumor sections were immunostained for MIB-1 labeling (assessment of Ki-67 expression) in neoplastic cells and evaluated microscopically. The MIB-1 labeling index was determined by counting the number of immunopositive cells among 1,000 neoplastic cells (at 400× magnification) in the area of highest labeling, and the index was expressed as a percentage. Tumors were classified as having low or high MIB-1 labeling index in neoplastic cells (ie, < 43% or ≥ 43% of neoplastic cells with MIB-1–specific cytoplasmic staining). Because of technical reasons, the percentage of labeled neoplastic cells was not determined for 1 tumor.

Tumor sections were immunostained for uPA expression and evaluated microscopically (1 section was examined in tumors < 1 cm and 2 sections in tumors > 1 cm; in tumors > 1 cm, both slides were examined and the global estimation was considered instead of the 2 individual estimations). The percentage of stromal cells (fibroblasts) with cytoplasmic staining for uPA was estimated. Tumors were classified as having low or high uPA expression in stromal cells (ie, < 10% or ≥ 10% of stromal cells with uPA–specific cytoplasmic staining).

NS = Not significant (P ≥ 0.05).

See Table 1 for remainder of key.

Stromal MMP-9 expression was significantly higher in MMTs that were ≥ 3 cm in diameter, compared with MMTs that were < 3 cm in diameter (P = 0.001); in histologic grade III tumors, compared with grade II or I tumors (P = 0.005); in tumors with invasive growth, compared with tumors with expansive growth (P = 0.034); in tumors with a high MIB-1 labeling index, compared with tumors with a low MIB-1 labeling index (P = 0.010); and in tumors with a high expression of uPA in stromal cells, compared with tumors with a low expression of uPA in stromal cells (P < 0.001; Table 2). There were also significant associations between high stromal expression of MMP-9 and regional lymph node and distant metastases (P = 0.032 and 0.014, respectively). The expression of MMP-9 in neoplastic cells was not related to the aforementioned characteristics, with the exception of regional lymph node metastases (P = 0.006).

Overall survival times and disease-free intervals—The assessment of overall survival times involved 70 dogs with MMTs (for 12 dogs, the follow-up period was not yet completed; another 7 dogs were lost to follow-up, and another 5 dogs died because of unknown reasons during the immediate postoperative period), and the assessment of disease-free intervals involved 74 dogs with MMTs (for 8 dogs, the follow-up period was not yet completed although they remained free of detectable metastases at the last examination performed; another 7 dogs were lost to follow-up, and another 5 dogs died for unknown reasons during the immediate postoperative period). The mean ± SE overall survival time for dogs with tumors with high MMP-9 expression in stromal cells (15.13 ± 1.7 months) was significantly (P < 0.001) lower than the mean overall survival time (22.22 ± 0.83 months) for dogs with tumors with low stromal cell MMP-9 expression (Figure 3). Furthermore, there was a significant (P < 0.001) relationship between high expression of MMP-9 in stromal cells and shorter disease-free interval (mean ± SE, 12.94 ± 1.94 months), compared with tumors with low stromal cell MMP-9 expression (20.55 ± 1.13 months; Figure 4). Matrix metalloproteinase-9 expression in neoplastic cells was not significantly associated with either overall survival time or disease-free interval. Dogs that had MMTs with a low MIB-1 labeling index (< 43% MIB-1–positive cells) had a significantly (P < 0.001) longer overall survival time (22.13 ± 0.86 months) and disease-free interval (20.32 ± 1.20 months) than did dogs that had MMTs with a high MIB-1 labeling index (mean overall survival time, 15.95 ± 1.64 months; mean disease-free interval, 13.62 ± 1.87 months).

Figure 3—
Figure 3—

Kaplan-Meier overall survival time curves following surgical removal of 70 MMTs from 70 dogs and grouping on the basis of MMP-9 expression in stromal cells (24 of the 94 dogs were excluded because the follow-up period was not yet completed for 12 dogs, 7 dogs were lost to follow-up, and 5 dogs died of unknown reasons within the immediate postoperative period). Tumors were classified as having low (n = 45) or high (25) MMP-9 expression in stromal cells (ie, < 50% [solid line] or ≥ 50% [dashed line] of stromal cells with cytoplasmic staining, respectively). Vertical marks on each line represent censored dogs. The mean ± SE overall survival time for dogs with tumors with high MMP-9 expression in stromal cells (15.13 ± 1.7 months) was significantly (P < 0.001) lower than the mean overall survival time (22.22 ± 0.83 months) for dogs with tumors with low stromal MMP-9 expression.

Citation: American Journal of Veterinary Research 73, 5; 10.2460/ajvr.73.5.689

Figure 4—
Figure 4—

Kaplan-Meier disease-free interval curves following surgical removal of 74 MMTs from 74 dogs and grouping on the basis of MMP-9 expression in stromal cells (20 of the 94 dogs were excluded because the follow-up period was not yet completed for 8 dogs although they remained free of detectable metastases, 7 were lost to follow-up, and 5 died of unknown reasons within the immediate postoperative period). Tumors were classified as having low (n = 45) or high (29) MMP-9 expression in stromal cells (ie, < 50% [solid line] or > 50% [dashed line] of stromal cells with cytoplasmic staining, respectively). Vertical marks on each line represent censored dogs. The mean ± SE disease-free interval for dogs with tumors with high MMP-9 expression in stromal cells (12.94 ± 1.94 months) was significantly (P < 0.001) shorter than the mean disease-free interval for dogs with tumors with low stromal cell MMP-9 expression (20.55 ± 1.13 months).

Citation: American Journal of Veterinary Research 73, 5; 10.2460/ajvr.73.5.689

Discussion

Currently, cancer researchers have widened their focus from neoplastic cells alone to neoplastic cells and the surrounding tumor stroma. Tumor-adjacent stromal cells are considered to have a role in the process of tumor invasion. In recent years, investigation of MMPs has amplified the interest in stromal cell involvement in tumor progression and metastasis.19

The present study evaluated the expression of MMP-9 in stromal and neoplastic cells of MTs in dogs. The MMP-9 expression in tumor-adjacent fibroblasts in MMTs was significantly higher than that in BMTs or clinically normal mammary gland tissues, as previously reported by human breast cancer researchers.28 By use of zymography and western blot analysis, Yokota et al10 confirmed significantly higher MMP-9 activity and expression in MTs in dogs, compared with clinically normal mammary gland tissues, and determined that the activity and expression was higher in carcinomas than in BMTs. These findings were attributed to increased MMP-9 gene transcription. However, in that study,10 tumor tissue homogenates were analyzed and it was not possible to identify which cells (neoplastic or stromal) were responsible for the MMP-9 activity. It has also been demonstrated immunohistochemically that the expression and gelatinolytic activity of MMP-9 was higher in mammary gland carcinomas of dogs than in mammary gland adenomas.11,13 Although the expression of MMP-9 by tumor-adjacent stromal cells is considered an important prognostic factor for humans with breast tumors, such expression and its clinical relevance was not investigated in studies in dogs.1,8,20 It has been postulated that, in epithelial cancers, the upregulation of MMPs occurs in stromal cells in response to stimulation by chemokines and cytokines produced by neoplastic cells.1,4,19,29 It is believed that neoplastic cells may acquire the ability to control the expression of certain molecules by neighboring cells as part of the carcinogenic process. Several molecules (eg, transforming growth factor-β, platelet-derived growth factor, and ECM metallo-proteinase inducer) produced by neoplastic cells upregulate the production of stromal proteases (eg, MMP-9), which mediate signaling events essential for neoplastic cell migration and invasion.19,30 The high expression of MMP-9 in stromal cells detected in the present study supports that theory.

As previously reported,11 BMTs had lower MMP-9 expression in neoplastic cells than did MMTs. However, this association was not as marked as that determined for stromal cells. Complex carcinomas had significantly lower expression of MMP-9 than did other malignant histologic types, which may be related to the less aggressive behavior of the former.

In the present study, there was a significant association between larger tumors (≥ 3 cm in diameter) and increased stromal MMP-9 expression. In agreement with data for human breast carcinomas,6 high concentrations of MMP-9 in stromal cells were also significantly related to higher MIB-1 indexes, suggesting that neoplastic cell proliferation is stimulated by stromal MMP-9. Stromal-derived MMPs may facilitate tumor growth through the activation of latent growth factors or inactivation of growth inhibitory molecules and by providing space for the expanding tumor mass.2,31 One important signaling pathway for tumor growth involves transforming growth factor-β, which can be activated by MMP-9.18,32 Matrix metalloproteinase-9 is also an important promoter of tumor growth by increasing the bioavailability of angiogenic factors.18,19,33

The expression of MMP-9 by neoplastic cells was not significantly associated with MIB-1 index in the present study, in contrast to the results of Nowak et al.14 Differences in field of view selection and counting procedures used to determine MIB-1 may explain the discordant results, but this is uncertain because there are no details of the counting protocol in that other report.14

In human breast cancers, MMP-9 expression by tumor-associated fibroblasts has been related with histologic grade of the tumor.20 In the present study, grade III tumors had a significantly higher expression of MMP-9 in stromal cells, compared with findings for grade II or I tumors. By use of western blot analysis, Vinothini et al15 detected a positive association between high concentrations of MMP-9 and less differentiated MTs in dogs. In addition to the upregulation of mitogenic signaling via release and activation of cytokines and growth factors by MMP-9, this could also be explained by the capacity of MMP-9 to modify integrin-mediated anchorage of epithelial cells (through proteolysis of ECM and cleavage of adhesion molecules such as cadherins), which contributes to the disruption of tissue architecture.2,18

Recently, we determined that uPA expression by fibroblasts in the tumor vicinity is associated with poor prognostic factors and poor outcome in dogs with MMTs.25 Urokinase-type plasminogen activator, a serine protease, participates in the activation cascade of MMP-9.16,33 To our knowledge, the present study is the first to investigate the possible relationship between uPA expression and MMP-9 expression in MTs in dogs. We found a significant association between high expressions of both proteins, suggesting that one of the mechanisms by which uPA promotes malignancy of MTs in dogs may be through an increase in MMP-9 activity in tumor-associated fibroblasts.

Although there is a wide range of biological functions of MMPs in neoplastic processes, the proteolytic degradation of the endothelial basement membrane and other matrix components by type IV collagenases is crucial for the tissue infiltration and intravasation of neoplastic cells into circulation.5,18 Matrix metalloproteinase-9 is also an important molecule for metastatic niche formation, thereby establishing a metastasis-supportive microenvironment.18,19 In the present study, stromal expression of MMP-9 in MMTs with invasive growth and regional lymph node and distant metastases was significantly higher than the stromal expression in MMTs with expansive growth and without regional lymph node and distant metastasis, which reinforces the hypothesis that stromal MMP-9 facilitates cancer cell migration and contributes to the development of a metastatic phenotype. Similar results have been reported for human breast cancers.6,7 Loukopoulos et al12 found that in tumors of dogs, metastatic malignancies produced higher concentrations of MMP-9 than did nonmetastatic tumors, although this association did not reach significance.

In the present study, high MMP-9 expression by tumor cells was significantly related to regional lymph node metastasis but not to tumor size, mode of growth, histologic grade, stromal cell expression of uPA, or distant metastasis. With regard to human breast cancers, Kim et al34 and Jobim et al35 reported the absence of significant associations between tumor cell MMP-9 expression and poor prognostic factors, such as tumor size, histologic grade, or metastasis. After a thorough search in the literature, we were unable to locate any reports of immunohistochemical studies investigating the association of MMP-9 expression and any of the aforementioned factors in MTs in dogs.

As reported previously by our group21 and in accordance with findings of other researchers,36 high MIB-1 indexes were negatively associated with overall survival time and disease-free interval in dogs with MMTs. Furthermore, dogs with MMTs that had high stromal cell expression of MMP-9 had significantly shorter disease-free intervals and overall survival times, compared with those of dogs with MMTs with low stromal cell expression of MMP-9, similar to findings in human breast cancer studies.6–8,20 It is possible that higher MMP-9 activity in stromal cells translates into a more invasive and metastatic tumor.19 In contrast, the MMP-9 expression in neoplastic cells of MMTs was not significantly associated with survival time in dogs of the present study. To our knowledge, this is the first study of MMP-9 expression in MMTs and survival time in dogs.

In dogs with MTs, MMP-9 expression by tumor-associated fibroblasts appears to be significantly related to poor prognostic factors and shorter disease-free intervals and overall survival time. The results of the present study have suggested that tumor-adjacent stromal cells contribute to progression of MTs in dogs and that the stromal cell MMP-9 expression may be of prognostic value.

ABBREVIATIONS

BMT

Benign mammary gland tumor

ECM

Extracellular matrix

MIB-1

Mindbomb homolog-1

MMP

Matrix metalloproteinase

MMT

Malignant mammary gland tumor

MT

Mammary gland tumor

uPA

Urokinase-type plasminogen activator

a.

Dako, Glostrup, Denmark.

b.

Santa Cruz Biotechnology Inc, Santa Cruz, Calif.

c.

Vector, Burlingame, Calif.

d.

Provided by Dr. Carlos Lopes, Diagnostic Pathology Service, Centro Hospitalar do Porto, Porto, Portugal.

e.

Protein Extraction Reagent kit, Thermo Scientific, Rockford, Ill.

f.

Complete Mini Protease Inhibitor Cocktail Tablets, Roche Diagnostics, Indianapolis, Ind.

g.

Amersham Biosciences, GE Healthcare, Carnaxide, Portugal.

h.

Sigma Aldrich Corp, St Louis, Mo.

i.

PASW Statistics, version 18.0, SPSS Inc, Lisbon, Portugal.

References

  • 1.

    Singer CF, Kronsteiner N, Marton E, et al. MMP-2 and MMP-9 expression in breast cancer-derived human fibroblasts is differentially regulated by stromal-epithelial interactions. Breast Cancer Res Treat 2002; 72:6977.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 2.

    Pupa SM, Ménard S, Forti S, et al. New insights into the role of extracellular matrix during tumor onset and progression. J Cell Physiol 2002; 192:259267.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3.

    Sommerville RPT, Oblander SA, Apte SS. Matrix metalloproteinases: old dogs with new tricks. Genome Biol 2003; 4:216.1216.11.

  • 4.

    Sternlicht MD, Werb Z. How matrix metalloproteinases regulate cell behaviour. Ann Rev Cell Dev Biol 2001; 17:463516.

  • 5.

    Duffy MJ, Maguire TM, Hill A, et al. Metalloproteinases: role in breast carcinogenesis, invasion and metastasis. Breast Cancer Res 2000; 2:252257.

  • 6.

    Del Casar JM, González LO, Alvarez E, et al. Comparative analysis and clinical value of the expression of metalloproteinases and their inhibitors by intratumor stromal fibroblasts and those at the invasive front of the breast carcinomas. Breast Cancer Res Treat 2009; 116:3952.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7.

    Vizoso FJ, González LO, Corte MD, et al. Study of matrix metalloproteinases and their inhibitors in breast cancer. Br J Cancer 2007; 96:903911.

  • 8.

    González LO, Corte MD, Junquera S, et al. Expression and prognostic significance of metalloproteinases and their inhibitors in luminal A and basal-like phenotypes of breast cancer. Human Pathol 2009; 40:12241233.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9.

    Lana SE, Ogilvie GK, Hansen RA, et al. Identification of matrix metalloproteinases in canine neoplastic tissue. Am J Vet Res 2000; 61:111114.

  • 10.

    Yokota H, Kumata T, Taketaba S, et al. High expression of 92 KDa type IV collagenase (matrix metalloproteinase-9) in canine mammary adenocarcinoma. Biochym Biophys Acta 2001; 1568:712.

    • Search Google Scholar
    • Export Citation
  • 11.

    Hirayama K, Yokota H, Onai R, et al. Detection of matrix metalloproteinases in canine mammary tumours: analysis by immunohistochemistry and zymography. J Comp Pathol 2002; 127:249256.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12.

    Loukopoulos P, Mungall BA, Straw RC, et al. Matrix metalloproteinase-2 and −9 involvement in canine tumors. Vet Pathol 2003; 40:382394.

  • 13.

    Kawai K, Uetsuka K, Doi K, et al. The activity of matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) in mammary tumors of dogs and rats. J Vet Med Sci 2006; 68:105111.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14.

    Nowak M, Madej JA, Podhorska-Okolow M, et al. Expression of extracellular matrix metalloproteinase (MMP-9), E-cadherin and proliferation-associated antigen Ki-67 and their reciprocal correlation in canine mammary adenocarcinomas. In Vivo 2008; 22:463470.

    • Search Google Scholar
    • Export Citation
  • 15.

    Vinothini G, Balachandran C, Nagini S. Evaluation of molecular markers in canine mammary tumors: correlation with histological grading. Oncol Res 2009; 18:193201.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16.

    Amălinei C, Căruntu I, Bălan RA. Biology of metalloproteinases. Rom J Morphol Embryol 2007; 48:323334.

  • 17.

    Brooks SA, Lomax-Browne HJ, Carter TM, et al. Molecular interactions in cancer cell metastasis. Acta Histochem 2010; 112:325.

  • 18.

    Kessenbrock K, Plaks V, Werb Z. Matrix metalloproteinases: regulators of tumor microenvironment. Cell 2010; 141:5267.

  • 19.

    Jodele S, Blavier L, Yoon JM. Modifying the soil to affect the seed: role of stromal-derived matrix metaloproteinases in cancer progression. Cancer Metastasis Rev 2006; 25:3543.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20.

    Pellikainen JM, Ropponen KM, Kataja VV, et al. Expression of matrix metalloproteinase (MMP-2) and MMP-9 in breast cancer with a special reference to activator protein-2, HER-2, and prognosis. Clin Cancer Res 2004; 10:76217628.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21.

    Matos AJF, Lopes C, Faustino AMR, et al. MIB-1 indices according to clinico-pathological variables in canine mammary tumours: a multivariate study. Anticancer Res 2006; 26:18211826.

    • Search Google Scholar
    • Export Citation
  • 22.

    Misdorp W, Else RW, Helmén E, et al. Histological classification of the mammary tumours of the dog and the cat. In: Shulman FL, ed. World Health Organization international histological classification of tumours of domestic animals. 2nd ed. Washington DC: Armed Forces Institute of Pathology, 1999;1629.

    • Search Google Scholar
    • Export Citation
  • 23.

    Elston CW, Ellis IO. Assessment of histological grade. In: Rosen PP, ed. Rosen's breast pathology. 3rd ed. Philadelphia: Lippincott-Raven Publishers, 1998;365382.

    • Search Google Scholar
    • Export Citation
  • 24.

    Hsu S, Raine L, Fanger H. The use of antiavidin antibody and avidin-biotin-peroxidase complex in immunoperoxidase techniques. Am J Clin Pathol 1981; 75:816821.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25.

    Santos A, Lopes C, Marques RM, et al. Immunohistochemical analysis of urokinase plasminogen activator and its prognostic value in canine mammary tumours. Vet J 2011; 189:4348.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 26.

    Agresti A. Inference for contingency tables. In: Balding DJ, Bloomfield P, Cressie NAC, eds. Categorical data analysis. 2nd ed. Hoboken, NJ: Wiley, 2002;9197, 369373.

    • Search Google Scholar
    • Export Citation
  • 27.

    Klein M, Kleinbaum DG. Kaplan-Meier survival curves and the log-rank test. In: Gail M, Krickeberg K, Samet J, et al, eds. Survival analysis—a self-learning text. 2nd ed. New York: Springer, 2005;4582.

    • Search Google Scholar
    • Export Citation
  • 28.

    Köhrmann A, Kammerer U, Kapp M, et al. Expression of matrix metalloproteinases (MMPs) in primary human breast cancer and breast cancer cell lines: new findings and review of the literature. BMC Cancer [serial online]. 2009; 9:188. Available at: www.biomedcentral.com/1471-2407/9/188. Accessed Oct 10, 2010.

    • Search Google Scholar
    • Export Citation
  • 29.

    Owen JL, Iragavarapu-Charyulu V, Gunja-Smith Z, et al. Up-regulation of matrix metalloproteinase-9 in T lymphocytes of mammary tumour bearers: role of vascular endothelial growth factor. J Immunol 2003; 171:43404351.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30.

    Tětu B, Trudel D, Wang CS. Proteases des cellules stromales réactionnelles du cancer: une cible thérapeutique attrayante. Bull Cancer 2006; 93:944948.

    • Search Google Scholar
    • Export Citation
  • 31.

    MacDougall JR, Matrisian LM. Contributions of tumor and stromal matrix metalloproteinases to tumor progression, invasion and metastasis. Cancer Metastasis Rev 1995; 14:351362.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 32.

    Micke P, Östman A. Tumour-stroma interaction: cancer-associated fibroblasts as novel targets in anti-cancer therapy? Lung Cancer 2004; 45:S163S175.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 33.

    Klein T, Bischoff R. Physiology and pathophysiology of matrix metalloproteinases. Amino Acids 2011; 41:271290.

  • 34.

    Kim HJ, Park C, Park BW, et al. Expression of MT-1 MMP, MMP2, MMP9 and TIMP2 mRNAs in ductal carcinoma in situ and invasive ductal carcinoma of the breast. Yonsei Med J 2006; 47:333342.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 35.

    Jobim FC, Xavier NL, Uchoa DM, et al. Prevalence of vascular-endothelial growth factor, matrix metalloproteinases and tissue inhibitors of metalloproteinases in primary breast cancer. Braz J Med Biol Res 2009; 42:979987.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 36.

    Peña LL, Nieto AI, Pérez-Alenza D, et al. Immunohistochemical detection of Ki-67 and PCNA in canine mammary tumors: relationship to clinical and pathologic variables. J Vet Diagn Invest 1998; 10:237246.

    • Crossref
    • Search Google Scholar
    • Export Citation

Contributor Notes

Supported by grants from the Fundação para a Ciência e Tecnologia (project SFRH/BD/42363/2007).

Address correspondence to Dr. de Matos (ajmatos@icbas.up.pt).