Expression of Matrix Metalloproteinase (MMP)-1, MMP-2, MMP-9, Cathepsin B, and Urokinase Plasminogen Activator in Non Skull Base Chordoma

Transcription

1 Anatomic Pathology / EXPRESSION OF PROTEASES IN CHORDOMA Expression of Matrix Metalloproteinase (MMP)-1, MMP-2, MMP-9, Cathepsin B, and Urokinase Plasminogen Activator in Non Skull Base Chordoma Takahiko Naka, MD, 1 Carsten Boltze, MD, 1 Doerthe Kuester, MD, 1 Torss-Oliver Schulz, MD, 1 Amir Samii, MD, 2 Christian Herold, MD, 2 Helmut Ostertag, MD, 3 and Albert Roessner, MD 1 Key Words: Chordoma; Matrix metalloproteinase; Cathepsin B; Urokinase plasminogen activator; Prognosis DOI: /C8T7APJDAUPR8TLL Abstract We analyzed the expression of proteases and the clinicopathologic significance in non skull base chordoma (NSBC). By using immunohistochemical techniques, we studied the expression of matrix metalloproteinase (MMP)-1, MMP-2, MMP-9, cathepsin B (CatB), and urokinase plasminogen activator (upa) in 29 NSBCs and compared these data with clinicopathologic parameters and the expression of cell differentiation markers. Expression of MMP-1 (P =.092), MMP-2 (P =.041), and CatB (P =.058) was associated with nuclear pleomorphism, a previously described adverse prognostic indicator. Expression of cytokeratin 8 correlated with that of MMP-1 (P =.005), MMP-2 (P =.002), and upa (P =.032). Patients with higher MMP-2 expression had a poorer prognosis than those with lower MMP-2 expression (P =.013). We believe that NSBCs with nuclear pleomorphism or stronger epithelial character have a higher invasive ability than those without. In addition, high MMP-2 expression was an indicator of an unfavorable clinical outcome in NSBC. Chordoma is a rare malignant bone tumor that occurs in the notochordal remnants. It is located most often in the sacrococcygeal region, followed by the spheno-occipital region and other sites. Although it is a slow-growing tumor and metastasis generally is restricted to later clinical stages, it frequently recurs after excision, which suggests its locally strong invasiveness. In particular, patients with a nonskull lesion frequently have a giant mass exhibiting an infiltrating growth pattern to soft tissues after the destruction of the cortical part of vertebral bones. During the last few years, tumor-associated proteases have been shown to have an important role in invasive growth in a variety of human carcinomas. These proteases include matrix metalloproteinases (MMPs), cysteine proteases, serine proteases, and others. Several studies have revealed that higher expression of cathepsin B (CatB), 1 urokinase plasminogen activator (upa), 1,2 MMP-1, 1,3 MMP-2, 1,4 and MMP-9 1,3,5 in giant cell tumors or chondrosarcomas also is indicative of invasiveness and progression in human bone tumors. Haeckel et al 6 reported that cathepsin K contributes to invasive growth in skull base chordoma. However, to the best of our knowledge, data on the expression of proteases in non skull base chordoma (NSBC) are not available. As seen in the physiologic bone absorption process, various proteolytic enzymes can be involved in the bone invasion process in chordoma. In the present study, we evaluated the expression of proteases in NSBC and determined its clinicopathologic and prognostic implications. Materials and Methods Tumors A total of 29 conventional chordomas (23 primary and 6 recurrent lesions) were used: 19 lesions (13 patients) were 926 Am J Clin Pathol 2004;122: DOI: /C8T7APJDAUPR8TLL

2 Anatomic Pathology / ORIGINAL ARTICLE obtained from the Department of Pathology, Nordstadt Medical Center, Hanover Clinic, Hanover, Germany, and 10 (10 patients) from the Department of Pathology, Otto-von- Guericke University Magdeburg, Magdeburg, Germany. Patient age ranged from 31 to 80 years (median, 58 years). Twenty-two lesions (18 patients) were located in the sacrococcygeal region and 7 (5 patients) in the mobile spines. All patients gave informed consent, and the study was approved by the local ethics committee. Histologic sections were stained routinely with H&E for diagnostic purposes. The following studies were conducted using sections from 10% formalin-fixed, paraffinembedded tissue samples, highlighting the representative areas of the tumor. Light Microscopic Study We subclassified tumors into solid or nonsolid types and recorded nuclear pleomorphism. Mitosis was evaluated as positive or negative because it was scarce, as previously reported Immunohistochemical Study Immunohistochemical study was done using the following antibodies: anti MMP-1 (41-1E5, dilution 1:100; Oncogene Research Products, Cambridge, MA), anti MMP-2 (42-5D11, dilution 1:40; Oncogene Research Products), anti MMP-9 (56-2A4, dilution 1:60; Oncogene Research Products), anti-catb (polyclonal, dilution 1:50; Oncogene Research Products), anti-upa (polyclonal, dilution 1:100; DAKO, Hamburg, Germany), antipancytokeratin (AE1/AE3, dilution 1:200; BioGenex, San Ramon, CA), anti cytokeratin 8 (CAM5.2, dilution 1:5; Becton Dickinson, San Jose, CA), and antivimentin (Vim3B4, dilution 1:400; DAKO). The sections were treated with primary antibody, followed by staining with avidin-biotin-peroxidase complex (Immunotech, Marseille, France) or an alkaline phosphatase detection kit (Vector, Burlingame, CA). The antigen was retrieved with enzyme digestion for both pancytokeratin and cytokeratin 8 (CK8) and with microwave for vimentin. Evaluation of Immunostaining Proteolytic activity in tumor cells was estimated by the expression of MMP-1, MMP-2, MMP-9, CatB, and upa. Immunoreactivity for the proteases was scored as described previously. 11 The distribution of positive cells was graded semiquantitatively as follows: 0, none; 1, less than 10%; 2, 10% to 50%; and 3, more than 50%. Staining intensity was graded as follows: 0, negative; 1, faint; 2, moderate; and 3, strong. We determined the sum of these 2 parameters to evaluate the expression of proteases, from 0 to 6. Reactivity for cell differentiation markers (pancytokeratin, CK8, vimentin) was graded semiquantitatively according to the number of positive cells: 2+, 80% or more; 1+, less than 80%; and 0, none. Statistical Analysis By using the median value, we subdivided patients into older and younger groups. Analyses regarding age and sex and comparison between recurrent and nonrecurrent tumors were done using primary lesions; other analyses included both primary and recurrent lesions. Correlations between one protease score and another were described by using the Spearman rank correlation test. Correlations between protease scores and other clinicopathologic parameters were estimated by using the Mann-Whitney U test. The prognostic implication of protease expression was estimated by using the log-rank test. Results Immunohistochemical Study Expression of MMP-1, MMP-2 Image 1, CatB, and upa showed a cytoplasmic staining pattern with diverse intensity. Mean scores ± SD were 4.9 ± 1.4 in MMP-1, 4.8 ± 1.4 in MMP- 2, 1.5 ± 2.0 in CatB, and 3.0 ± 2.1 in upa. Immunoreactivity for these proteases occasionally was enhanced in invasion fronts to bone compared with areas away from such fronts. MMP-9 was expressed in only a few cells near bone invasion fronts in a small number of lesions. MMP-1 expression correlated with that of MMP-2; there was a trend toward association in upa, MMP-1, MMP-2, and CatB expression Table 1. Image 1 Expression of matrix metalloproteinase (MMP)-2 in a non skull base chordoma (score, 6) (MMP-2 immunostaining, original magnification 200). Am J Clin Pathol 2004;122: DOI: /C8T7APJDAUPR8TLL 927

3 Naka et al / EXPRESSION OF PROTEASES IN CHORDOMA Table 1 P Values for Correlations in Protease Expression MMP-1 MMP-2 CatB MMP CatB upa CatB, cathepsin B; MMP, matrix metalloproteinase; upa, urokinase plasminogen activator. Table 2 Correlation Between Protease Scores and Clinicopathologic Parameters in Non Skull Base Chordomas MMP-1 MMP-2 CatB upa Primary lesions (n = 23) Age <57 (n = 11) (n = 12) P Sex Male (n = 16) Female (n = 7) P Recurrent or nonrecurrent Recurrent tumor (n = 8) Nonrecurrent tumor (n = 4) P Primary and recurrent lesions (n = 29) Clinical status Primary (n = 23) Recurrent (n = 6) P Subtype Solid (n = 4) Nonsolid (n = 25) P Nuclear pleomorphism Negative (n = 15) Positive (n = 14) P Mitosis Negative (n = 15) Positive (n = 14) P Cytokeratin 8 * 1+ (n = 14) (n = 15) P Pancytokeratin * 1+ (n = 6) (n = 23) P Vimentin * 1+ (n = 2) (n = 27) P CatB, cathepsin B; MMP, matrix metalloproteinase; upa, urokinase plasminogen activator. * Reactivity was graded semiquantitatively according to the number of positive cells: 2+, 80% or more; 1+, less than 80%; and 0, none. Relationships between the expression of protease and other parameters are summarized in Table 2. upa expression was significantly higher in the older than in the younger patient group. Men showed higher expression of MMP-2, CatB, and upa than women; the difference for upa was statistically significant. Nuclear pleomorphism correlated with MMP-2 expression and showed a trend toward an association with the expression of MMP-1 and CatB. Mitosis was associated with MMP-1 expression. Protease scores were higher in CK8 lesions scored as 2+ than in those scored as 1+, and differences were statistically significant for MMP-1, MMP-2, and upa. Pancytokeratin expression also was associated with MMP-1 expression. There were no differences in protease scores between primary recurrent tumors and primary nonrecurrent tumors, as well as between primary and recurrent tumors. Similarly, no correlations were seen between recurrence and age, sex, histologic subtype, nuclear pleomorphism, or mitosis. There were no correlations between expression of protease and expression of other parameters. Prognosis Detailed follow-up data, available for 12 patients, ranged from 42 to 158 months after the initial surgery. The Kaplan- Meier survival curve was used for calculating survival rates. Patients with higher MMP-2 expression (score 5 or 6) had a poorer prognosis than those with lower MMP-2 expression (score 0-4) (P =.013) Figure 1. Nuclear pleomorphism also correlated with survival (P =.029). Neither mitosis nor expression of other proteases correlated with patient survival. Discussion Data on the relationship between the expression of MMPs, cysteine proteases, or serine proteases and tumor malignancy have accumulated during the past few years. Correlation was reported between increased expression of these proteases and poor prognosis in carcinomas of the stomach, 12 colon, 13,14 lung, 15 ovary, 16 and other sites. 17,18 Regarding bone tumors, Haeckel et al 1 reported correlations between the expression of CatB and upa and recurrence in chondrosarcoma. In addition, we have shown increased expression of cathepsin K in invasion fronts of skull base chordomas, suggesting its role in the bone invasion process. 6 However, little is known about protease expression in NSBCs. In the present study, we observed divergent expression of MMP-1, MMP-2, CatB, and upa in NSBCs and found that they often are accentuated in invasion fronts to bone. In addition, we detected coexpression of some proteases, which led us to speculate that these proteases might cooperate in the invasion process of NSBCs. Moreover, expression of some proteases significantly correlated with or showed a trend toward an association with nuclear pleomorphism and mitosis; the former was shown to correlate with worse prognosis in a study by Naka et al 19 and in the present study. These findings indicate that proteases have an important role in the invasive process and are expressed more strongly in lesions at histologically 928 Am J Clin Pathol 2004;122: DOI: /C8T7APJDAUPR8TLL

4 Anatomic Pathology / ORIGINAL ARTICLE Survival Months Figure 1 Kaplan-Meier survival curve according to matrix metalloproteinase (MMP)-2 expression in non skull base chordomas. Survival times for patients who were still alive at the study s end are indicated by circles or triangles, based on the most recent census. Patients with higher MMP-2 expression (n = 9; circles) had a poorer prognosis than those with lower MMP-2 expression (n = 3; triangles) (P =.013). advanced stages of NSBC. We found no differences in the expression of MMP-1, MMP-2, CatB, and upa between primary and recurrent lesions or between recurrent and nonrecurrent tumors, suggesting no correlation between protease expression and recurrence. However, patients with higher MMP-2 expression had a worse prognosis than those with lower MMP-2 expression. Although the results should be confirmed by investigating larger series, these findings have far-reaching implications for the improvement of the outcome of this locally aggressive tumor. The expression of upa was significantly higher in patients in the older age group and in men, findings for which the causes and significance remain to be clarified. Chordoma is a unique neoplasm with epithelial and mesenchymal differentiation. However, Naka et al 20 have shown divergent expression of cytokeratin subsets in chordoma. In the present cytokeratin analysis in which the difference from a completely immunopositive status was emphasized, there were some lesions with lower expression of CK8 and pancytokeratin. Interestingly, CK8 expression correlated with MMP-1, MMP-2, and upa expression; pancytokeratin expression was associated with MMP-1 expression. These findings strongly suggest that the degree of the epithelial character is related to invasive ability in NSBC. Divergent cytokeratin expression in NSBC warrants further study. NSBCs with nuclear pleomorphism or stronger epithelial character had a higher invasive ability than those without. In addition, high MMP-2 expression was an indicator of an unfavorable clinical outcome in this locally aggressive tumor. From the 1 Department of Pathology, Faculty of Medicine, Magdeburg University, Magdeburg and 2 Departments of Neurosurgery and 3 Pathology, Nordstadt Medical Center, Klinikum Hanover, Hanover, Germany. Supported in part by the Canon Foundation Europe, Leiden, the Netherlands. Address reprint requests to Dr Naka: Dept of Orthopaedic Surgery, Kyushu Rosai Hospital, Kuzuharatakamatsu, Kokuraminami-ku, Kitakyushu, Fukuoka, , Japan. Acknowledgments: We thank Brigitte Roettger, Carola Kuegler, Claudia Miethke, and Nadine Wiest, Department of Pathology, Faculty of Medicine, Magdeburg University, for technical assistance, and Bernd Wueshoff for editing the manuscript. References 1. Haeckel CG, Krueger S, Grote H-J, et al. Overexpression of cathepsin B and urokinase plasminogen activator is associated with increased risk of recurrence and metastasis in patients with chondrosarcoma. Cancer. 2000;89: Haeckel C, Czerniak B, Ayala AG, et al. Expression of plasminogen activators and plasminogen activator inhibitor 1 in dedifferentiated chondrosarcoma. Cancer. 1997;79: Kawashima A, Okada Y, Nakanishi I, et al. Immunolocalization of matrix metalloproteinases and tissue inhibitors of metalloproteinases in human chondrosarcomas. Gen Diagn Pathol. 1996;142: Sakamoto A, Oda Y, Iwamoto Y, et al. Expression of membrane type 1 matrix metalloproteinase, matrix metalloproteinase 2 and tissue inhibitor of metalloproteinase 2 in human cartilaginous tumors with special emphasis on mesenchymal and dedifferentiated chondrosarcoma. J Cancer Res Clin Oncol. 1999;125: Ueda Y, Imai K, Tsuchiya H, et al. Matrix metalloproteinase 9 (gelatinase B) is expressed in multinucleated giant cells of human giant cell tumor of bone and is associated with vascular invasion. Am J Pathol. 1996;148: Haeckel C, Krueger S, Kuester D, et al. Expression of cathepsin K in chordoma. Hum Pathol. 2000;31: O Connell JX, Renard LG, Liebsch NJ, et al. Base of skull chordoma: a correlative study of histologic and clinical features of 62 cases. Cancer. 1994;74: Volpe R, Mazabraud A. A clinicopathologic review of 25 cases of chordoma: a pleomorphic and metastasizing neoplasm. Am J Surg Pathol. 1983;7: Bergh P, Kindblom L, Gunterberg B, et al. Prognostic factors in chordoma of the sacrum and mobile spine: a study of 39 patients. Cancer. 2000;88: Holton J, Steel T, Luxsuwong M, et al. Skull base chordomas: correlation of tumour doubling time with age, mitosis and Ki67 proliferation index. Neuropathol Appl Neurobiol. 2000;26: Kumaki F, Matsui K, Kawai T, et al. Expression of matrix metalloproteinases in invasive pulmonary adenocarcinoma with bronchioloalveolar component and atypical adenomatous hyperplasia. Am J Pathol. 2001;159: Sier CFM, Kubben FJGM, Ganesh S, et al. Tissue levels of matrix metalloproteinase MMP-2 and MMP-9 are related to the overall survival of patients with gastric carcinoma. Br J Cancer. 1996;74: Am J Clin Pathol 2004;122: DOI: /C8T7APJDAUPR8TLL 929

5 Naka et al / EXPRESSION OF PROTEASES IN CHORDOMA 13. Herszenyi L, Plebani M, Carraro P, et al. The role of cysteine and serine proteases in colorectal carcinoma. Cancer. 1999;86: Murray GI, Duncan ME, O Neil P, et al. Matrix metalloproteinase-1 is associated with poor prognosis in colorectal cancer. Nat Med. 1996;2: Kawano N, Osawa H, Ito T, et al. Expression of gelatinase A, tissue inhibitor of metalloproteinases-2, matrilysin, and trypsin(ogen) in lung neoplasms: an immunohistochemical study. Hum Pathol. 1997;28: Young TN, Rodriguez GC, Rinehart AR, et al. Characterization of gelatinases linked to extracellular matrix invasion in ovarian adenocarcinoma: purification of matrix metalloproteinase 2. Gynecol Oncol. 1996;62: Kallakury BVS, Karikehalli S, Haholu A, et al. Increased expression of matrix metalloproteinases 2 and 9 and tissue inhibitors of metalloproteinases 1 and 2 correlate with poor prognostic variables in renal cell carcinoma. Clin Cancer Res. 2001;7: Murray GI, Duncan ME, O Neil P, et al. Matrix metalloproteinase-1 is associated with poor prognosis in esophageal cancer. J Pathol. 1998;185: Naka T, Boltze C, Samii S, et al. Skull base and non skull base chordomas: clinicopathologic and immunohistochemical study with special reference to nuclear pleomorphism and proliferative ability. Cancer. 2003;98: Naka T, Iwamoto Y, Shinohara N, et al. Cytokeratin subtyping in chordomas and fetal notochord: an immunohistochemical analysis of aberrant expression. Mod Pathol. 1997;10: Am J Clin Pathol 2004;122: DOI: /C8T7APJDAUPR8TLL