Huntington Beach, CA 92649, USA. (Received 9 June 1998; accepted in revised form 10 August 1998). We have sought to determine the production and activity ...
Clin. Exp. Metastasis, 1998, 16, 713–719
11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111
Expression and localization of urokinase-type plasminogen activator in human spinal column tumors Ziya L. Gokaslan*, Shravan K. Chintala*, Julie E. York*, Venkaiah Boyapati*, Sushma Jasti*, Raymond Sawaya*, Gregory Fuller+, David M. Wildrick*, Garth L. Nicolson‡, and Jasti S. Rao* *Department of Neurosurgery and +Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030 and ‡The Institute for Molecular Medicine, Huntington Beach, CA 92649, USA (Received 9 June 1998; accepted in revised form 10 August 1998)
We have sought to determine the production and activity of serine proteases in primary and metastatic spinal tumors and the association of these enzymes with the invasive and metastatic properties of spinal column tumors. Using immunohistochemical techniques, the cellular localization and expression of urokinase-type plasminogen activator (uPA) was assessed, whereas its activity was determined by fibrin zymography, and the amounts of enzyme were measured by an enzyme-linked immunosorbent assay (ELISA) in primary spinal column tumors (chordoma, chondrosarcoma, and giant cell tumor) and metastatic tumors of the spine arising from various malignancies (breast, lung, thyroid, and renal cell carcinomas, and melanomas). Metastatic tumors displayed higher levels of uPA activity than did primary spinal tumors (P < 0.001). Immunohistochemical analysis revealed that uPA expression was highest in metastases from lung and breast carcinomas and melanomas, followed by metastatic tumors from thyroid and renal cell carcinomas. Similar results were obtained for uPA activity and enzyme level as determined by fibrin zymography and ELISA, respectively. We conclude that metastatic spinal tumors possess higher levels of uPA expression and activity than the primary spinal tumors, which tend to be less aggressive and only locally invasive malignancies. The results suggest that the plasminogen system may participate in the metastasis of tumors to the spinal column. Keywords:
immunohistochemistry, proteases, spinal metastases, uPA
Introduction The vast majority of spinal neoplasms arise as a result of hematogenous spreading from distant malignancies, most commonly from breast, lung, or prostate carcinomas. Among adults, breast, lung, and prostate carcinomas account for approximately 65% of spinal metastases (The University of Texas M. D. Anderson Cancer Center, Department of Medical Informatics, 1994). Although renal cell carcinomas Address correspondence to: Ziya L. Gokaslan, Department of Neurosurgery, Box 64, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030 USA. Tel: (713) 792–2400; Fax: (713) 794–4950.
© 1998 Kluwer Academic Publishers
show a lower incidence of spinal column metastases, these tumors represent a large proportion of metastatic spinal neoplasms in surgical series because of their relatively poor response to medical therapy [1]. Other tumors, such as thyroid carcinomas, have a lower propensity toward formation of spinal metastases and more commonly spread to lung, liver, and bone before involving the spinal column [2]. Autopsy studies demonstrate metastatic deposits in the spinal columns of up to 90% of patients who die of cancer, with approximately onethird of these patients having symptoms referable to spinal tumor involvement during the course of their
Clinical & Experimental Metastasis Vol 16 No 8 713
Z. L. Gokaslan et al. 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111
illness [3]. Also, spinal column neoplasms may arise from the direct invasion of adjacent tumors, such as lung or chest wall malignancies [4]. Primary malignant tumors of the spinal column exist but are rare [3]. Chondrosarcomas, chordomas, and giant-cell tumors comprise the majority of primary malignant spinal column tumors in adults [3]. After their surgical removal, chondrosarcomas and chordomas tend to recur locally for prolonged periods of time before spreading hematogenously to distant sites [5–7]. Giant-cell tumors are associated with a poor prognosis because of their aggressive nature, high likelihood of recurrence, and ability to metastasize [8,9]. For tumors to metastasize to distant sites, they must produce proteases that proteolytically degrade the surrounding extracellular matrix and allow tumor cells to penetrate into the blood circulation. Numerous studies have shown that plasminogen activators, and in particular, urokinase-type plasminogen activator (uPA), play an essential role in tumor metastasis. When bound to cell surface receptors, uPA triggers a directed and localized proteolysis at the leading edge of the tumor [10]. uPA accomplishes this by converting plasminogen to plasmin, which in turn degrades extracellular matrix proteins and activates certain collagenases [11–14]. Recent studies have demonstrated higher levels of uPA in various carcinomas, such as those of the colon [15], lung [16], breast [16–18], uterus [19], gastrointestinal tract [20], thyroid [21,22], and prostate [23–25] than in surrounding tissues. However, there are no published reports addressing the role of uPA, or proteases in general, in the primary and metastatic tumors of the spinal column. To assess the role of uPA in this process, we determined the amounts and enzymatic activites of uPA in both primary and metastatic spinal tumors using immunohistochemical analysis and fibrin zymography.
Materials and methods Surgical tissue specimens Tissue specimens were obtained from patients undergoing spinal surgery at The University of Texas M. D. Anderson Cancer Center (M. D. Anderson). Five specimens from each of the metastatic tumor types (breast, lung, thyroid and renal cell carcinoma, and melanoma) and the primary spinal tumors (chordoma, giant-cell tumor, and chondrosarcoma) were obtained intraoperatively. The histological diagnosis of the samples was confirmed by a neuropathologist. 714 Clinical & Experimental Metastasis Vol 16 No 8
The tissue samples were flash-frozen in liquid nitrogen immediately after surgical removal and stored at –80°C prior to the determination of uPA activity by fibrin zymography. Samples for immunohistochemical analysis were fixed in 10% formalin and embedded in paraffin. Immunohistochemical analysis Immunolocalization of uPA was analyzed in formalin-fixed, paraffin-embedded sections using specific antihuman polyclonal antibodies raised against human uPA (American Diagnostica, Greenwich, CT). The appropriate concentrations of primary antibody were determined by titrating the antibody on positive and negative control tissue sections. Paraffin sections (4 m thick) were prepared and mounted onto Silane-coated glass slides. uPA was detected by incubating the dewaxed, blocked sections with rabbit antihuman uPA antibody (1:200 dilution) in 1% bovine serum albumin in phosphate-buffered saline (PBS) for 1 h at room temperature in a humidified chamber. After the tissue samples were briefly washed in buffer, they were incubated in biotinylated, goat anti-rabbit secondary antibody and streptavidin-alkaline phosphatase (Biogenese, San Ramon, CA). Alkaline phosphatase activity was visualized by adding a substrate solution containing naphthol AS-BI phosphate, levamisole, and fast-red TR, which makes the sample intensely pink in regions containing alkaline phosphatase. The sections were later counterstained with hematoxylin. Positive immunoreactivity of uPA was identified by pink staining. A nonspecific immunoglobulin G (IgG) was used as the primary antibody instead of the uPA antibody for the negative control studies. Fibrin zymography The enzymatic activities and molecular weights of electrophoretically separated forms of uPA in tissue extracts were determined by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) using purified plasminogen and fibrinogen added as substrates before gel polymerization [26]. After gel polymerization, tumor extract samples containing equal amounts of protein (20 g/ml) were loaded into the gel slots and electrophoresis was performed. Various types of plasminogen activators present were identified based on a comparison of the molecular weight observed for each of them with the standard molecular weight of uPA. Each gel was then given two 30 min washes with 2.5% Triton X-100 and incubated at 37°C overnight in 0.1 M glycine buffer (pH 7.5). The gels were stained with Coomassie blue
Expression and localization of urokinase-type plasminogen activator 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111
and destained with isopropanol. The final gel had a uniform blue background except for the transparent bands in regions in which the plasminogen activator was present and caused cleavage of plasminogen to plasmin. Quantitative analysis Densitometric analysis was performed to quantitate the relative proteolytic activities of uPA in the fibrin gels. After electrophoresis, the samples from different patients, grouped according to tumor histology, were washed with 2.5% Triton X-100 and incubated at 37 °C for a shorter time than that used for zymography to ensure improved quantitative estimation of uPA bands. The relative band intensities were determined densitometrically. Enzyme-linked immunosorbent assay The uPA protein content in spinal tumor extracts (25 g) was measured by enzyme-linked immunosorbent assay (ELISA) using an antibody specific for uPA. In this assay, tissue extracts and buffer containing uPA were mixed with phosphate buffer in the wells of 96-well microtiter plates and incubated overnight at 4°C. The wells were then washed with PBS and incubated with anti-uPA antibody at 25°C for 3 h. The plates were subsequently washed with PBS and incubated with a secondary antibody (an alkaline-phosphatase antibody conjugate), and the color was developed with p-nitrophenol phosphate. The concentration of uPA in the tissue extracts was determined using a standard curve for uPA.
Results Immunohistochemical analysis Immunohistochemical staining was performed to determine the distribution of uPA protein in tissue sections from various samples of primary and metastatic spinal tumors. Higher amounts of uPA were found in metastases from lung and breast carcinomas and melanomas than in metastases from thyroid and renal carcinomas (Figure 1). Fibrin zymography revealed similar differences in terms of uPA enzymatic activity. Immunohistochemical localization in tissue sections showed predominantly cytoplasmic uPA in metastatic tumors originating from breast and lung carcinomas and melanomas. In addition, sections of melanoma in spinal metastases membranous uPA staining demonstrated. Sections from renal and thyroid spinal metastases displayed only weak uPA cytoplasmic staining. Among the
primary spinal tumors, giant-cell tumors of the bone showed intense cytoplasmic and membranous staining, whereas staining was less intense in chondrosarcoma and chordoma tissue samples. In general, metastatic tumors showed higher amounts of uPA than did the primary tumors of the spine. Quantitation of tumor-associated uPA We performed fibrin zymography to determine the amounts of endogenous uPA in human spinal primary and metastatic tumor tissues and measured the resulting enzyme activity in fibrin gels using densitometry. Figure 2A shows that the intensity of the Mr 55 000 uPA band was significantly higher in extracts of metastatic tumors from lung and breast carcinomas and melanomas than in those from renal and thyroid spinal metastases. When the amounts of Mr 55 000 uPA activity were quantitated by densitometry, the uPA activity was found to be four to five times higher in extracts of metastatic tumors from lung and breast carcinomas and melanomas than in the renal cell carcinomas (P < 0.001), whereas thyroid tumor samples contained twice the amount of uPA activity found in renal cell extracts (Figure 2B). The uPA activity was slightly higher for chondrosarcoma and giant-cell tumor extracts than for chordoma samples. ELISA The levels of uPA protein in tumor tissue extracts was quantitated by ELISA using an antibody specific for uPA. We found higher (five- to eight-fold) levels of uPA protein in extracts of metastatic lung and breast carcinomas and melanomas than in metastatic thyroid and renal cell tumors (P < 0.001) (Figure 3). The uPA protein content was similar in chordoma, chondrosarcoma and giant-cell tumor extracts and of approximately the same level as that found in extracts of thyroid carcinoma metastases to the spine.
Discussion Previous studies have shown that the plasminogen system plays an important role in local tumor invasion of normal tissues and formation of distant metastases [10]. During tumor invasion, the basement membrane must be degraded. Basement membranes contain type IV collagen and other proteins, such as fibronectin and laminin. Proteinases such as uPA have been implicated in tumor invasion and metastasis [11,12,14,27], but uPA also plays a vital role in tissue remodeling that occurs in various physiological conditions like wound healing. uPA is a highly specific Clinical & Experimental Metastasis Vol 16 No 8 715
Z. L. Gokaslan et al. 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111
Figure 1. Immunohistochemical localization of uPA in spinal metastatic tumor samples from lung, breast, renal, and thyroid carcinoma, and melanoma, as well as in primary spinal tumors, including chordoma, chondrosarcoma, and giantcell tumor of bone was determined using antibody against human uPA as described in ‘Materials and methods’.
serine protease that is secreted as a single-chain protein (pro-uPA) that is then proteolytically converted into the active two-chain uPA. Active uPA molecules bind to a cell surface uPA receptor, thereby accelerating the conversion of plasminogen to plasmin at cell surfaces [28,29]. Plasmin, in turn, is a member of the serine proteinase family with a broad substrate range [10]. In addition to directly degrading basement membrane components and activating plasmin, uPA can activate other degradative enzymes such as type IV collagenases [14]. During the metastatic process, uPA is highly upregulated in a number of different types of tumor 716 Clinical & Experimental Metastasis Vol 16 No 8
cells [11,12,16]. For example, the majority of patients with prostatic carcinoma develop skeletal metastasis, and uPA has been implicated in this process [23,25]. In an interesting metastatic model system, it was shown that intracardiac inoculation of Dunning R-3227 Mat LyLu rat prostate carcinoma cells leads to preferential implantation, invasion and growth of tumor cells that over express uPA, predominantly at skeletal sites [30]. As noted by Achbarou et al. [25], this is remarkable as the skeleton receives only a small fraction of the total blood supply from the left ventricle [31]. Moreover, secreted uPA is entrapped by bone cells during the skeletal invasion of tumor
Expression and localization of urokinase-type plasminogen activator 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111
Figure 2. A: Results of fibrin zymography of tissue extracts obtained from various primary and metastatic spinal tumor samples. Samples containing equal amounts of protein (20 g) were electrophoresed on sodium dodecylsulfate-polyacrylamide gels containing plasminogen and fibrinogen. For details, see the ‘Materials and methods’. Gel sample lanes 1 and 2 contain 20 g each of urokinase-type plasminogen activator (uPA) and tissue-type plasminogen activator (tPA) as 55 and 70 (kDa) kilodalton reference standards, respectively. Lanes 3–7 include extracts from breast carcinoma, melanoma, lung carcinoma, thyroid carcinoma, and renal carcinoma, respectively. Lanes 8–10 include extracts from chordoma, chondrosarcoma, and giant-cell tumor of bone, respectively. B: Quantitative estimation of uPA activity obtained on fibrin gels. Bands on gels as shown in Figure 2A were scanned on a densitometer, and their relative intensities were plotted as a histogram. The data are the mean ± S.D. of five different samples in each tumor category. The units of peak area are arbitrary. The 8 tumor samples are arranged and labeled as for lanes 3–10 in Figure 2A.
Figure 3. uPA content in primary and metastatic spinal tumor tissue extracts. uPA content was determined by ELISA as described in ‘Materials and methods’. The 8 samples are arranged and labeled as for gel lanes 3–10 in Figure 2A. The data are the mean ± S.D. of five different samples in each tumor category.
cells, and this could result in an additional proteolytic activation cascade. In another rat prostate carcinoma model in male Copenhagen rats [25,30,32], the amino-terminal and carboxyl-terminal domains of uPA were found to potentially contribute differentially to the characteristics of skeletal metastasis. To determine if uPA also plays a role in the process of metastasis formation or local bone invasion/destruction in spinal column tumors, we examined the production of uPA in several primary and malignant spinal neoplasms. Our analyses demonstrated that most metastatic spinal tumors contained uPA. In general, metastatic tumors displayed higher levels of uPA production and enzymatic activity than primary spinal tumors. Immunohistochemical analysis revealed that uPA production was greatest in spinal metastases from lung carcinoma, followed by those from breast carcinoma and melanoma. In addition, in all tissue specimens, the amount of uPA protein correlated with the amount of enzymatic activity determined by fibrin zymography. The uPA protein content determined by ELISA confirmed the increased expression of uPA in the metastatic lesions. Clinical & Experimental Metastasis Vol 16 No 8 717
Z. L. Gokaslan et al. 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111
We found that the level of uPA production and the amount of uPA enzymatic activity were roughly proportional to the degree of tumor aggressiveness. In general, metastatic tumors from lung carcinoma and melanoma are considered to be more aggressive and are known to be associated with a poorer prognosis than carcinomas of renal or thyroid origin [33]. In the present study, uPA production was significantly higher in the more aggressive tumors (lung carcinomas and melanomas). Breast carcinomas are not considered as aggressive as lung carcinomas or melanomas; thus, the significance of the high uPA levels in breast carcinomas is not clear. It could be that other enzymes, including matrix metalloproteinases and endoglycosidases, play important roles as well. Studies are underway at M. D. Anderson to investigate the role of these other degradative enzymes in spinal column tumors. Although the relevance of elevated uPA levels to tumor invasion and metastasis formation is not exactly clear, studies that examine the mechanisms by which tumors spread to distant sites may yield new information that can be used to develop novel treatments for metastatic spinal tumors. Further studies are in progress to elucidate the mechanism of uPA activation in both primary and metastatic tumors of the spine.
Acknowledgments We thank Lena Jones for help with preparation of the manuscript, Alan Rayford for technical assistance, and Beth Notzon for manuscript review.
References: 1. Black P, 1979, Spinal metastasis: current status and recommended guidelines for management. Neurosurgery, 5, 726–46. 2. Harrington KD, 1986, Metastatic disease of the spine. J Bone Joint Surg Am, 68, 1110–15. 3. Sundaresan N, Galicich JH, Chu FC and Huvos AG, 1979, Spinal chordomas. J Neurosurg, 50, 312–19. 4. Kostuik JP and Weinstein JN, 1991, Differential diagnosis and surgical treatment of metastastic spine tumors. In: Frymoyer JW, ed: The Adult Spine: Principles and Practice. New York: Raven Press, 861–88. 5. Cummings BJ, Esses S and Harwood AR, 1982, The treatment of chordomas. Cancer Treat Rev, 9, 299–311. 6. Shives TC, Wold LE, Dahlin DC and Beabout JW, 1983, Chondrosarcoma and its Variants. Diagnosis and Treatment of Bone Tumors: A Team Approach. Sim FH, ed. Thorofare, NJ: Slack, Inc., 211–12. 7. Sanerkin NG, 1980, The diagnosis and grading of chondrosarcoma of bone: a combined cytologic and histologic approach. Cancer, 45, 582–94. 718 Clinical & Experimental Metastasis Vol 16 No 8
8. Caballes RL, 1981, The mechanism of metastasis in the so-called ‘benign giant cell tumor of bone’. Human Pathol, 12, 762–7. 9. Rock MG, Prichard DJ and Unni KK, 1984, Metastasis from histologically benign giant-cell tumor of bone. J Bone Joint Surg Am, 66A, 269–74. 10. Danø K, Andreasen PA, Grondahl-Hansen J, et al. 1985, Plasminogen activators, tissue degradation and cancer. Adv Cancer Res, 44, 139–266. 11. Pollanen J, Stephens RW and Vaheri A, 1991, Directed plasminogen activation at the surface of normal malignant cells. Adv Cancer Res, 57, 273–328. 12. Ossowski L, 1988, Plasminogen activator dependent pathways in the dissemination of human tumor cells in the chick embryo. Cell, 52, 321–8. 13. Liotta LA, Steeg PS and Stetler-Stevenson WG, 1991, Cancer metastasis and angiogenesis: an imbalance of positive and negative regulation. Cell, 64, 327–36. 14. Reich R, Thompson EW, Iwamoto Y, et al. 1988, Effects of inhibitors of plasminogen activator, serine proteinases, and collagenase IV on the invasion of basement membranes by metastatic cells. Cancer Res, 48, 3307–12. 15. de Bruin PA, Griffioen G, Verspaget HW, et al. 1988, Plasminogen activator profiles in neoplastic tissues of the human colon. Cancer Res, 48, 4520–4. 16. Sappino AP, Busso N, Belin D and Vassalli JD, 1987, Increase of urokinase-type plasminogen activator gene expression in human lung and breast carcinomas. Cancer Res, 47, 4043–6. 17. Evers JL, Patel J, Madeja JM, et al. 1982, Plasminogen activator activity and composition in human breast cancer. Cancer Res, 42, 219–26. 18. Janicke F, Schmitt M, Ulm K, et al. 1989, Urokinasetype plasminogen activator antigen and early relapse in breast cancer. Lancet, 2, 1049. 19. Saito K, Nagashima M, Iwata M, et al. 1990, The concentration of tissue plasminogen activator and urokinase in plasma and tissues of patients with ovarian and uterine tumors. Thromb Res, 58, 355–66. 20. Nishino N, Aoki K, Tokura Y, et al. 1988, The urokinase type of plasminogen activator in cancer of digestive tracts. Thromb Res, 50, 527–35. 21. Ito Y, Takeda T, Kobayashi T, et al. 1996, Plasminogen activation system inactive even in thyroid tumors; an immunohistochemical study. Anticancer Res, 16, 81–89. 22. Ito Y, Takeda T and Wakasugi E, 1995, Expression of urokinase-type plasminogen activator in various human thyroid tissues. Oncol Rep, 2, 525–28. 23. Chiarodo A, 1991, National cancer institute roundtable on prostate cancer: Future research directions. Cancer Res, 51, 2498–2505. 24. Cook GB and Watson FR, 1968, Events in the natural history of prostate cancer: using salvage curves, mean age distributions and contingency coefficients. J Urol, 99, 87–96. 25. Achbarou A, Kaiser S, Tremblay G, et al. 1994, Urokinase overproduction results in increased skeletal metastasis by prostate cancer cells in vivo. Cancer Res, 54, 2372–77. 26. Yamamoto M, Sawaya R, Mohanam S, et al. 1994, Expression and localization of urokinase-type plasminogen activator in human astrocytomas in vivo. Cancer Res, 54, 3656–61.
Expression and localization of urokinase-type plasminogen activator 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111
27. Blasi F, 1993, Urokinase and urokinase receptor: A paracrine/autocrine system regulating cell migration and invasiveness. Bioessays, 15, 105–11. 28. Pollanen J, Hedman K, Nielsen LS, et al. 1988, Ultrastructural localization of plasma membrane-associated urokinase-type plasminogen activator at focal contacts. J Cell Biol, 106, 87–95. 29. Stephens RW, Pollanen J, Tapiovaara H, et al. 1989, Activation of pro-urokinase and plasminogen on human sarcoma cells: a proteolytic system with surface-bound reactants. J Cell Biol, 108, 1987–95. 30. Isaacs JT, Isaacs WB, Feitz WF and Scheres J, 1986, Establishment and characterization of seven Dunning
rat prostatic cancer cell lines and their use in developing methods for predicting metastatic abilities of prostatic cancers. Prostate, 9, 261–81. 31. Ray RD, 1976, Circulation and bone. In: Bourne GH, ed. The Biochemistry and Physiology of Bone. New York: Academic Press, Vol. 4, 385–402. 32. Haq M, Goltzman D, Tremblay G and Brodt P, 1992, Rat prostate adenocarcinoma cells disseminate to bone and adhere preferentially to bone marrow-derived endothelial cells. Cancer Res, 52, 4613–19. 33. Lang FF and Sawaya R, 1996, Surgical management of cerebral metastases. Neurosurg Clin N Am, 7, 459–84.
Clinical & Experimental Metastasis Vol 16 No 8 719
