Immunohistochemical expression of vascular endothelial growth factor, matrix metalloproteinase 2, and matrix metalloproteinase 9 in cutaneous melanocytic lesions
Article first published online: 25 OCT 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 9, pages 1963–1970, 1 November 2002
How to Cite
Simonetti, O., Lucarini, G., Brancorsini, D., Nita, P., Bernardini, M. L., Biagini, G. and Offidani, A. (2002), Immunohistochemical expression of vascular endothelial growth factor, matrix metalloproteinase 2, and matrix metalloproteinase 9 in cutaneous melanocytic lesions. Cancer, 95: 1963–1970. doi: 10.1002/cncr.10888
- Issue published online: 25 OCT 2002
- Article first published online: 25 OCT 2002
- Manuscript Accepted: 3 JUN 2002
- Manuscript Revised: 17 MAY 2002
- Manuscript Received: 3 SEP 2001
- vascular endothelial growth factor;
- matrix metalloproteinase 2;
- matrix metalloproteinase 9
Vascular endothelial growth factor (VEGF), an endothelial cell mitogen, plays a hierarchical role in regulating physiologic and pathologic angiogenesis. Moreover, the transformation from noninvasive to invasive carcinomas is accompanied by focal disruption and discontinuity of the basement membrane. Several groups of proteases have been implicated in tumor cell invasion, including the 72-kDa gelatinase A/Type IV collagenase (matrix metalloproteinase 2 [MMP-2]) and the 92-kDa gelatinase B/Type IV collagenase (MMP-9).
The authors assessed the immunohistochemical expression of VEGF and metalloproteinases MMP-2 and MMP-9 in paraffin embedded biopsy specimens of malignant melanomas (18 invasive melanomas and 10 in situ melanomas); dysplastic nevi with architectural disorder and cytologic atypia of melanocytes; Spitz nevi; and compound or predominantly intradermal, ordinary, benign melanocytic nevi.
Strong cytoplasmic staining for VEGF was observed in melanoma cells in as many as 77% of primary invasive melanomas, whereas only 25% of the in situ melanomas exhibited a detectable immunoreactivity for VEGF. It is interesting to note that no immunoreactivity was shown by any nevi; Spitz nevi, in particular, showed negative immunoreactivity to VEGF. Invasive melanomas and in situ melanomas displayed coexpression of MMP-2 and MMP-9, although to a variable extent. In particular, high MMP-2 staining was observed in 14 of 18 invasive melanomas; moreover, strong MMP-2 expression also was observed in 60% of in situ melanomas, whereas the residual 40% of those melanomas showed a moderate level of positivity.
On the basis of the current data showing that malignant melanocytic tumors displayed strong VEGF expression, whereas benign melanocytic proliferations showed no immunoreactivity for VEGF, VEGF also may be used as a discriminating factor to distinguish malignant melanoma from lesions of uncertain histology. Cancer 2002;95:1963–70. © 2002 American Cancer Society.
The series of events by which a putatively normal cell develops into a metastatic tumor cell have been designated as tumor progression, the mechanisms of which have not been well defined in malignant melanoma. Crucial events include the ability of the tumor to grow, to cross mechanical and molecular barriers of invasion, to escape immune surveillance, and to disseminate.1 Early steps in this progression require tumor proliferation and the formation of the new vascular channels to provide nutrients and an escape route from the primary site.2, 3 Often, the tumor-associated angiogenesis occurs during the radial growth phase of melanoma, without competence for metastasis, and it may result in the development of the vertical growth phase with metastatic competence and a worse clinical prognosis.4, 5
Many angiogenic factors are expressed by melanoma cells in vivo.6, 7 Among the numerous chemical mediators, it is believed that vascular endothelial growth factor (VEGF) or vascular permeability factor is one of the most potent stimulators of endothelial activation and proliferation. VEGF is a heparin-binding dimeric polypeptide that, because of differential splicing, can be found in five different isoforms encoding proteins of 121 amino acids (VEGF121), 145 amino acids (VEGF145), 165 amino acids (VEGF165), 189 amino acids (VEGF189), and 206 amino acids (VEGF206).8 VEGF exerts a number of important biologic actions on endothelial cells. It increases the permeability of microvessels to circulating macromolecules,9 and it also acts as a selective endothelial cell mitogen.10, 11 In addition, VEGF alters endothelial cell gene expression, inducing increased production of tissue factor and several proteases, including interstitial collagenase and both urokinase-like plasminogen activators and tissue plasminogen activators.12, 13
Moreover, the transformation from noninvasive carcinoma to invasive carcinoma is accompanied by focal disruption and discontinuity of the basement membrane. Tissue invasion requires the expression of proteinases that are specific for interstitial extracellular matrix. Several groups of proteases have been implicated in tumor cell invasion, including matrix metalloproteinases (MMPs),14–16 serine proteases,17 and cysteine protease.18 It is postulated that the 72-kDa gelatinase A/Type IV collagenase (MMP-2) and the 92-kDa gelatinase B/Type IV collagenase (MMP-9)15 are two members of the MMP family that play a critical role in tumor invasion and angiogenesis. In fact, degradation of interstitial and basement membrane extracellular matrix represents a key element in the multistage process of tumor invasion and metastasis.16, 19, 20 The objective of this study was to evaluate the immunohistochemical expression of VEGF, MMP-2, and MMP-9 in cutaneous melanocytic lesions representing different stages of progression to malignant melanoma.
MATERIALS AND METHODS
All tissues used in this study were obtained from the archives of the Pathology Department. The formalin fixed, paraffin embedded melanocytic lesions represented 18 primary invasive (Clark level IV) malignant melanomas (7 lesions with a Breslow thickness > 1 and 11 lesions with a Breslow thickness > 3), none of which presented regression areas; 10 malignant melanoma in situ; 9 atypical (dysplastic) melanocytic nevi with architectural disorder and cytologic atypia of melanocytes; 6 Spindle and epithelioid cell (Spitz) nevi; and 10 ordinary, benign, melanocytic nevi (either compound or predominantly intradermal).
Specimens were fixed routinely in 10% buffered formalin, embedded in paraffin, and sections were stained with hematoxylin and eosin. Conventional 3-μm-thick histologic sections were obtained with a microtome from the selected blocks, mounted on slides that were pretreated with 3-aminopropyltriethoxylane, deparaffinized, and rehydrated. To better unmask the antigenic sites, sections were incubated with a solution of TVF (Alexis Biochemical) for 10 minutes at 90 °C. After washing in H2O and then in Tris buffer solution, the sections were incubated overnight at 4 °C with the monoclonal antibody anti-VEGF (dilution, 1:200; Santa Cruz Biochemicals, Santa Cruz, CA) and the polyclonal antibodies anti MMP-2 (dilution, 1:400) and anti MMP-9 (dilution, 1:500; kindly provided by Stetler Stevenson, National Institutes of Health, Bethesda, MD). The antibodies were linked to an alkaline phosphatase-antialkaline phosphatase (APAAP) catalyzed reaction (Kit; Dako, SpA, Italy). The APAAP technique was used according to Cordell et al.21 For visualization of the antibodies, sections were incubated with the substrate Fas red (Dako). Sections finally were counterstained with Mayer hematoxylin and mounted in glycerol gelatin. Each series included appropriate negative control sections that omitted the primary antibody.
The immunohistochemical expression of the antibodies studied was evaluated by light microscopy by observing 10 fields for each sample at × 250 magnification and identifying zone representatives. We evaluated particular immunohistochemical expression of the positive cells as follows: negative, no cells stained; low, 1–10% of cells stained; moderate, 15–30% of cells stained; and high, > 40% cells stained.
No VEGF immunoreactivity was detected in the cells of ordinary benign melanocytic nevi, atypical melanocytic nevi, or Spitz nevi. However, VEGF immunoreactivity was observed in endothelial cells of microvessels in all samples.
Moderate percentages (15–30%) of melanoma cells were detected in 5 of 10 melanomas in situ (Fig. 1), whereas, in the other 5 melanomas in situ, faint immunopositivity was detected. Few host response cells were positive for VEGF.
High percentages cytoplasmic tumor cell staining were observed in 14 of 18 invasive melanomas (Fig. 2). The other melanomas exhibited faint immunopositivity. Focal cytoplasmic VEGF staining was observed in tumor-infiltrating inflammatory cells in all samples examined.
In normal skin, no MMP-2 or MMP-9 immunoreactivity was observed; whereas moderate or faint MMP-2 expression was observed in ordinary, benign, melanocytic nevi and in typical, junctional, melanocytic nevi. In nevic cells, MMP-2 was observed more commonly in the papillary dermis than in the reticular dermis; whereas, in junctional nevi, MMP-2 staining was a prominent feature of the dermal:epidermal junction.
Each of the Spitz nevi showed immunostaining for MMP-2 that varied from moderate to high intensity (Fig. 3). In atypical, melanocytic nevi, MMP-2 staining showed slightly higher expression compared with the expression in ordinary, benign, melanocytic nevi, especially in atypical, compound nevi that were characterized by severe cellular atypia. No MMP-9 immunostaining was observed in tissue sections from benign, Spitz, or atypical nevi.
Invasive melanomas and melanomas in situ both showed variable immunostaining for MMP-2 and MMP-9. Strong MMP-2 expression was observed in each of the melanomas, with different grades of tumor invasion. In particular, high MMP-2 staining was observed in 14 of 18 invasive melanomas (Fig. 4). Moreover, strong MMP-2 expression was found in 60% of the melanomas in situ, whereas the residual 40% of melanomas in situ showed moderate levels of positivity (Fig. 5). MMP-2 was present in nests of melanoma cells within the epidermis as well as the dermis but also was found in peripheral melanoma cells and in adjacent dermal stroma.
Thirteen of 18 invasive melanomas with Clark Level IV invasion were positive for MMP-9 immunostaining. In the other five invasive melanomas, MMP-9 immunoreactivity occurred in scattered single and nested neoplastic cells in the deeper regions, with staining intensity ranging from low to moderate (Fig. 6).
Six of 10 melanomas in situ exhibited positive immunoreactivity for MMP-9, and its distribution within specimens of melanoma in situ showed extensive variations in terms of staining intensity and percentage of tumor cells (Fig. 7). All results are summarized in Table 1.
|Lesion||No. of patients||VEGFa||MMP-2a||MMP-9a|
|None||< 10%||15–30%||> 40%||None||< 10%||15–30%||> 40%||None||< 10%||15–30%||> 40%|
|In situ melanoma||10||-||5||5||-||-||-||4||6||4||1||2||3|
|Atypical (dysplastic) melanocytic nevi||9||9||-||-||-||2||2||5||-||-||-||-||-|
|Ordinary benign melanocytic nevi||10||10||-||-||-||-||8||2||-||-||-||-||-|
Angiogenesis, which is the formation of new capillaries, is a crucial component in the growth, invasive progression, and metastatic spread of solid tumors.2, 3 Several potential regulators of angiogenesis have been identified.22–24 In particular, it was found that VEGF is specific for endothelial cells, suggesting that this molecule may play a hierarchical role in the regulation of physiologic and pathologic angiogenesis.22, 25 We assessed the immunohistochemical expression of VEGF in biopsy specimens of malignant melanomas; dysplastic nevi with architectural disorder and cytologic atypia of melanocytes; Spitz nevi; and compound or predominantly intradermal, ordinary, benign, melanocytic nevi. In agreement with previous studies,26, 27 our results indicated that VEGF expression was greater in invasive malignant melanomas compared with melanomas in situ, suggesting a correlation between VEGF-induced angiogenesis and aggressiveness of invasive melanomas. In fact, we observed strong cytoplasmic staining for VEGF in melanoma cells in as many as 77% of primary invasive melanomas, whereas only 25% of the melanomas in situ exhibited a detectable immunoreactivity for VEGF. It is interesting to note that no immunoreactivity was shown by any nevi. In particular, Spitz nevi, which are difficult to differentiate diagnostically from melanoma, showed negative immunoreactivity for VEGF.
Different factors seem to play a critical role in regulating VEGF expression. They include cyclooxygenase 2,28 heregulin β 1,29 insulin,30 estrogens and progestins,31, 32 epidermal growth factor,33, 34 basic fibroblast growth factor and fibroblast growth factor 2,35 and N-acetylcysteine.36 Oncogenes also are involved directly in controlling VEGF expression, with a prominent inductive role of the ras oncogene and a suppressive function of the p53 tumor suppressor gene.37, 38 Moreover, it has been suggested that the overexpression of VEGF in melanoma cells may be a consequence of hypoxia.39 It is well recognized that hypoxia up-regulates VEGF expression in a variety of cells in tissues, both in vitro and in vivo.40, 41 In addition, it has been observed that transient hypoxia may promote the development of metastases in melanomas.40 Conversely, there also are some in vitro data showing the overexpression of VEGF in melanoma cells that were cultured under normal oxygen conditions.42 Therefore, the regulation of the mechanism that leads to tumor angiogenesis is complex and depends on multiple and interactive factors. For this reason, the inhibition of angiogenesis for therapeutic purposes preferably should not concentrate on a single factor.
Many of the factors that induce angiogenesis also are known as regulators of MMP gene expression (VEGF, tumor necrosis factor α, and basic fibroblast growth factor)43, 44 both in endothelial cells and in other cell types. In particular, it has been shown that VEGF specifically induces expression of interstitial collagenase (MMP-1).12 It also has been demonstrated that invasion of human melanoma cells through Type IV collagen and a reconstituted basement membrane seem to be dependent of MMP-1 expression.45
Metalloproteinases play important roles in neoplastic processes, such as tumor invasion and metastases,16 although conflicting results have been reported regarding the relation of MMP expression and the invasiveness of melanoma cells both in vitro and in an in vivo murine model.46, 47 Therefore, we analyzed the patterns of expression of MMP-2 and MMP-9 in benign lesions and in different stages of invasive growth of melanoma.
Ordinary, benign, melanocytic nevi as well as typical, junctional, melanocytic nevi showed faint MMP-2 expression, suggesting a limited degree of normal connective tissue remodeling. Moderate expression of MMP-2 was detected in atypical, melanocytic nevi. These findings are in agreement with other studies that presented a weak distribution of MMP-2 in the junctional nest of most benign nevi.48 Moreover, Spitz nevi showed moderate-to-strong labeling for MMP-2. According to other studies,49 negative MMP-9 immunostaining was observed in all tissue sections from benign, Spitz, and atypical nevi.
Invasive melanomas and in situ melanomas displayed coexpression of MMP-2 and MMP-9, although to a variable extent. In particular, high MMP-2 staining was observed in 14 of 18 invasive melanomas and in 6 of 10 melanomas in situ. Our results have confirmed previous observations that demonstrated the variable production of this enzyme both in melanomas in situ and in invasive melanomas.50 Moreover, in agreement with other reports,51, 52 our results showed that MMP-2 was present in melanoma cells as well as in tumor-surrounding stromal or host cells.
In the current study, positive immunoreactivity for MMP-9 was present in 6 of 10 melanomas in situ, and the staining patterns were similar to those observed in invasive melanoma (13 of 18 lesions). Some authors found that MMP-9 was expressed variably in the radial growth phase of primary melanoma, indicating that MMP-9 expression is correlated with early invasion of melanoma.49 Conversely, other immunohistochemical studies revealed that MMP-9 expression was present in advanced-stage melanoma, but not in early-stage melanoma.53 It also has been observed that a number of aggressive tumors, when they were examined for expression in situ, showed tumor cell-associated expression of MMP-9,54–56 suggesting that this molecule may be an important determinant of aggressive behavior in some tumors. It follows that the role of MMP-9 in melanoma progression is puzzling and has not been defined well, and it is reasonable to speculate that variable MMP-9 expression may be related to the presence of different clones in melanoma.
In conclusion, our data showed that malignant melanocytic tumors displayed strong VEGF expression, whereas benign melanocytic proliferations showed no immunoreactivity for VEGF. In particular, Spitz nevi, which are difficult to differentiate diagnostically from melanoma, were negative for VEGF. MMP-2 was expressed variably in Spitz nevi as well as in melanoma, characterizing all infiltrating lesions. On the basis of our data, VEGF showed a high specificity; therefore, it also may be used as a discriminating factor to distinguish malignant melanoma from lesions of uncertain histology.
- 1Molecular biology of cutaneous malignant melanoma. In: De VitaVTJr., HellmanS, RosembergSA, editors. Cancer. Principles and practice of oncology. New York: Lippincott-Raven, 1997: 1935., , .
- 29Up-regulation of vascular endothelial growth factor in breast cancer by the heregulin-beta 1 activated p38 signalling pathway enhance endothelial cell migration. Cancer Res. 2001; 15: 1727–1732., , , et al.
- 32Regulation of vascular endothelial growth factor expression by estrogens and progestins. Environ Health Perspect. 2000; 108 S5: 785–790., , , , , .