Dermatofibroma

Context.-Common fibrous histiocytoma (cFH) or dermatofibroma and dermatofibrosarcoma protuberans (DFSP) are 2 spindle cell mesenchymal tumors that are distinguished in part by their microscopic growth patterns and clinically by the greater propensity for DFSP to recur. Matrix metalloproteinases (MMPs) potentially play a role in modulating the growth patterns of cFH and DFSP by remodeling the extracellular matrix.

Objective.-To evaluate the immunohistochemical (IHC) expression of MMP-1, MMP-2, MMP-9, and MMP-14 in DFSP and cFH, because (1) MMP-1, MMP-2, MMP-9, and MMP-14 are synthesized by dermal fibroblasts, the major constituent of DFSP and cFH; and (2) platelet-derived growth factor B, which is overexpressed in most examples of DFSP because of t(17;22), activates ets-1, a transcription factor that regulates molecules associated with tumor invasion and metastasis, including MMP-1, MMP-3, and MMP-9.

Design.-Immunohistochemical studies were performed on archived, formalin-fixed, paraffin-embedded tissue of DFSP (n = 48) and cFH (n = 47).

Results.-Significant IHC expression (>10% of tumor cells) in cFH included MMP-14 (27 [59%] of 46 tumors positive), MMP-2 (21 [47%] of 45 tumors positive), MMP-9 (9 [20%] of 45 tumors positive), and MMP-1 (6 [13%] of 46 tumors positive). No DFSPs showed significant IHC expression of any of the MMPs evaluated. However, anti-MMP-2 highlighted a rich microvascular element within deep tumor tissue present in 81% of DFSPs with a prominent subcutaneous component.

Conclusion.-Our IHC results indicate that MMP-1 and MMP-9 are not up-regulated in DFSP. Convincing expression of MMP-14 in cFH suggests that this MMP may affect the growth pattern of the lesion, perhaps by activating MMP-2 expression in tumor cells. In DFSP, MMP-2 may play a role in tumor angiogenesis.

(Arch Pathol Lab Med. 2004;128:1136-1141)

The common fibrous histiocytoma (cFH), commonly referred to as dermatofibroma, and dermatofibrosarcoma protuberans (DFSP) are well-characterized, spindle cell mesenchymal tumors. Histopathologic and immunohistochemical differences between these 2 tumors have been emphasized in the literature in view of the fact that the cFH (dermatofibroma) is a wholly benign tumor with a reported recurrence rate of approximately 2% after simple excision,1 whereas DFSP has a recurrence rate ranging from 20% to 50%, depending on the completeness of initial excision and distance from tumor to the surgical margin.2

Although the microscopic features of a typical cFH generally present the pathologist little challenge, large, predominantly spindled variants of cFH with involvement of subcutaneous fat may cause diagnostic confusion with the more ominous DFSP. Useful light microscopic features that assist in distinguishing the 2 processes are the quality of the dermal collagen associated with the infiltrating tumor cells3,4 and the pattern of cell growth in subcutaneous fat.4 The cells of cFH surround and entrap thick bundles of polarizable collagen within the reticular dermis, which is composed primarily of type I collagen.3 In contrast, the tumor cells of DFSP proliferate in an altered dermal stroma consisting of thin, nonpolarizable collagen fibers. This stroma reportedly has a relatively higher proportion of collagen type III fibers than normal dermis.6

With respect to tumor interface with subcutaneous tissue, DFSP characteristically shows a more infiltrative growth pattern in subcutaneous fat than cFH. Cells of DFSP proliferate along interlobular septa, in multilayered bands paralleling the dermis, and in between individual fat cells, resulting in a honeycomb appearance of the process at low-power magnification. When cFH involves subcutaneous tissue, it grows along the superficial aspect of the interlobular septa as short, jagged extensions of tumor.

The chief constituent of both DFSP and cFH is a mesenchymal cell resembling a fibroblast. The fibroblast has the capability to produce a number of stroma-altering proteases.7,8 One such family of proteases, the matrix metalloproteinases (MMPs), are zinc-dependent endopeptidases that are divided into subgroups according to their connective tissue substrate specificity and that have the ability to remodel stromal matrix, facilitate cell motility, and promote angiogenesis by degrading various connective tissue elements.9 To date, only a few studies examining the presence of MMPs in examples of cFH"MS and DFSP12"14 have been reported in the English-language literature. We undertook this study to explore IHC expression of MMPs 1, 2, 9, and 14, which are synthesized by dermal fibroblasts, in examples of DFSP and cFH in an effort to further elucidate the pathogenesis of these 2 tumors.

MATERIALS AND METHODS

Archived examples of DFSP (n = 48) and cFH (n = 47) with unstained slides or paraffin-embedded blocks prepared from formalin-fixed material were collected for the study. cases of cFH and DFSP accessioned to the surgical pathology department between 1991 and 2003 and cases of DFSP retrieved from the personal files of one of the authors (E.L.W.) formed the basis of this study. All hematoxylin-eosin-stained slides were examined.

Dermatofibrosarcoma protuberans is a dermal-based neoplasm characterized by a moderate to highly cellular population of mildly atypical spindle cells growing primarily in a distinct storiform pattern (Figure 1, A). Intralesional inflammation or histiocytes (except in areas of prior biopsy) were not identified. In subcutaneous tissue, tumor cells grew along interlobular septa and between fat cells. Two examples of the pigmented variant of DFSP (so-called Bednar tumor) and 1 tumor with features of giant cell fibroblastoma were included in the study. No DFSP showed definitive microscopic evidence of fibrosarcomatous transformation. CD34 expression was identified in all 30 exampies of DFSP tested. All cases of benign fibrous histiocytoma included in the study arose in the dermis and demonstrated histologie features of the cFH or dermatofibroma.1 Forty of the cFHs were moderate to highly cellular and consisted of plump spindle cells arranged in ill-defined fascicles, whorled arrays, and loose storiform configurations that coursed around bundles of dermal collagen (Figure 1, B). The remaining 7 rumors were relatively smaller and less cellular. They were composed almost entirely of slender spindle cells haphazardly proliferating around bundles of dermal collagen. A few scattered foamy or hemosiderin-laden histiocytes were identified in many of the tumors, but only in 1 case were they a substantial component of the lesion. Superficial growth along the interlobular septa of subcutaneous fat was noted in several of the larger, more cellular lesions. Additional microscopic features noted only in cFH included overlying squamous or basaloid hyperplasia and a perivascular lymphoplasmacytic infiltrate typically located at the periphery of the lesion.

Immunohistochemical analysis was performed using the standard avidin-biotin complex immunoperoxidase technique1" with 3,3-diaminobenzidene tetrachloride as the chromogen and hematoxylin as the counterstain. Antibodies used in the study, their sources, dilutions, and pretreatments are listed in Table 1.

Primary antibodies to MMP-I, MMP-2, MMP-9, and MMP-14 were incubated with the histologie sections on an automated immunostainer (Ventana BioTech Systems, Tucson, Ariz). Positive controls were reviewed. For negative control sections, primary antibody was replaced with normal serum. Immunostaining was scored semiquantitatively based on the number of spindled tumor cells stained. Cells with conspicuous histiocytic differentiation manifested by the presence of abundant foamy cytoplasm or intracytoplasmic deposits of hemosiderin were not counted. Positive ( + ) reactivity was defined as cytoplasmic and was graded semiquantitatively in the following manner: O, no reactivity; +1, 1% to 10% of tumor cells reactive; +2, 11% to 25% of tumor cells reactive; +3, 26% to 50% of cells reactive; and +4, >50% of cells reactive. For the purpose of this study, only tumors with greater than 10% of cells reactive ( + 2 score or higher) were considered positive.

RESULTS

Immunohistochemical results for cFH and DFSP are summarized in Tables 2 and 3, respectively. Immunohistochemical expression (>10% of tumor cells) of MMP-14 (Figure 2, A and B) and MMP-2 (Figure 2, C) was observed in 27 (59%) of 46 cases and 21 (47%) of 45 cases of cFH, respectively. In 10 cases (22%), more than 50% of the tumor cells demonstrated strong expression of MMP-14. In 80% of cases, the frequency of MMP-14 expression was equal to or greater than the frequency of MMP-2 expression, and correlation between MMP-14 and MMP-2 expression within 1 grade was found in 31 (67%) of 46 cases. No quantitative differences in IHC expression of MMP-14 and MMP-2 were observed when tumors composed exclusively of plump spindle cells (46% of cases positive; Figure 2, A) were compared with cFHs consisting primarily of slender spindle cells (45% of cases positive; Figure 2, B). However, tumor cells with conspicuous cytoplasm tended to exhibit a greater intensity of reaction compared to the delicate spindle cells. Variable expression of MMP-9 and MMP-I was noted in 9 (20%) of 45 tumors and in 6 (13%) of 46 tumors, respectively.

No significant frequency of MMP expression was identified in cases of DFSP, but scattered positive tumor cells (+1 score) were noted in 37% of tumors tested for MMP14 expression, 22% of tumors evaluated for MMP-2 expression, 19% of tumors tested for MMP-I expression, and 8% of cases tested for MMP-9 expression.

Matrix metalloproteinase 2 was expressed in endothelial cells and in smooth muscle cells lining vascular spaces. In cFH and in the superficial dermal aspect of DFSP (Figure 2, D), anti-MMP-2 staining highlighted vessels with rounded, straight, and arcuate profiles that were irregularly distributed throughout the lesions. However, in examples of DFSP, anti-MMP-2 decorated numerous small, rounded vessels that were typically concentrated in the deep and oftentimes more cellular component of the neoplasm (Figure 2, D and E). This finding was observed in 17 of 21 examples of DFSP with a prominent subcutaneous component. Anti-MMP-14 weakly decorated endothelial cells in cFH, but was virtually inconspicuous in endothelial cells associated with vessels of DFSP. Matrix metalloproteinases 1 and 9 were not convincingly expressed in the vascular component of any of the tumors evaluated.

COMMENT

The pathobiologic mechanisms responsible for the divergent growth patterns observed in conventional examples of cFH and DFSP have not been fully elucidated in the literature. Connective tissue endopeptidases, including cathepsins, MMPs, and plasminogen activators, may play a role in the growth patterns of these 2 tumors by their ability to remodel the extracellular matrix. To date, only a few studies investigating cathepsin17 and MMP-2, MMP-9, MMP-Il, and MMP-14 expression KM5 in a variable number of cFHs and DFSPs have appeared in the English-language literature. In this study, we hoped that a more comprehensive MMP evaluation would assist in clarifying the pathogenesis of these 2 processes.

Our selection of the gelatinase/type IV collagenase, MMP-2, and its membrane-bound activator MMP-14; the gelatinase/type IV collagenase, MMP-9; and the interstitial collagenase, MMP-I, was based on the following data. First, since the key mesenchymal element of DFSP and cFH is a spindle cell resembling the fibroblast, we wanted to test MMPs that have been demonstrated in dermal fibroblasts. MMP-2 is constitutively expressed in skin fibroblasts18 as an inactive propeptidase and is activated by membrane-type MMPs,19 including MMP-14 (membrane type 1 MMP),18,20 whereas MMP-I and MMP-9 synthesis has been demonstrated in fibroblasts after exposure to exogenous stimuli, including cytokines21 and ultraviolet irradiation.-22,23 Second, the protooncogene, ets-1, has a binding site that acts as a promoter for MMPs 1, 3, and 9, and the urokinase-like plasminogen activator,24 which activates MMPs by proteolytic cleavage. Finally, Ets-1 expression is regulated in part by certain cytokines and growth factors, including platelet-derived growth factor B,25 which is overexpressed in most cases of DFSP on account of the characteristic translocation involving the collagen type IAl and platelet-derived growth factor B genes on chromosomes 17 and 22, respectively.26,27

In this study, we found significant IHC expression (>10% of tumor cells) of MMP-14 (59%), MMP-2 (47%), MMP-9 (20%), and MMP-I (13%) in examples of cFH, but no significant expression of these MMPs in any of the cases of DFSP tested.

The absence of MMP-I and MMP-9 expression in DFSP indicates that the wild-type platelet-derived growth factor B, which is commonly overexpressed in this neoplasm, probably does not activate ets-1 in tumor cells as anticipated. Alternatively, cfs-2-targeted MMP expression in endothelial cells could facilitate growth of the neoplasm by promoting angiogenesis.2!i However, we failed to find convincing MMP-I and MMP-9 expression in vessels associated with either DFSP or cFH.

A small number of cFHs in our study exhibited variable, but generally low levels of MMP-I and MMP-9 expression. One recent study reported MMP-9 messenger RNA (mRNA) synthesis by in situ hybridization and strong, diffuse immunohistochemical expression of the molecule in a small number of cFHs.1" Since cells with a monocyte/ macrophage phenotype are a source of MMP-9,2" it is not surprising that cFH, a tumor demonstrating the potential for histiocytic differentiation, would show evidence of MMP-9 synthesis.

Previous studies have claimed strong immunohistochemical expression of MMP-2 protein10,15 and MMP-2 mRNA synthesis by in situ hybridization10 in a small number of cFHs. In addition, Ohnishi et al11 found strong expression of membrane type 2 MMP and weak expression of membrane type 1 MMP (MMP-14) in cFH. In comparison, we observed a higher frequency and a greater intensity of reactivity with anti-MMP-14 than with anti-MMP-2 in cFH, although MMP-14 and MMP-2 immunohistochemical expression correlated within 1 grade in 67% of cases. Our results and those presented by Ohnishi et al15 indirectly support the observation that MMP-2 expression is not regulated at the transcriptional level, but instead is governed by enzyme activators (membrane-type MMPs) and inhibitors (tissue inhibitors of metalloproteinases [TIMPs]).18 In fact, membrane-type MMPs, including MMP-14, are probably the chief mediators of MMP-2 function in cFH, as TIMP-2, the main inhibitor of MMP-2 activity, is synthesized and expressed at extremely low levels in this tumor.11

Our MMP-2, MMP-9, and MMP-14 results differ somewhat from results from the aforementioned studies10,15 in that we report weaker reactivity for anti-MMP-2 and antiMMP-9 and stronger reactivity for anti-MMP-14 in cases of cFH. These discrepancies could be attributed to a number of factors, including the larger number of cases analyzed in our study, differences in immunohistochemical scoring methods, or technical factors such as differences in antibodies, immunohistochemical processing techniques, or antigen retrieval methods used.

Our data suggest that MMP-2 and MMP-14 may affect the pathogenesis of cFH and DFSP in different ways. Although significant expression of MMP-2 and MMP-14 was not observed in tumor cells of DFSP, MMP-2, which is also synthesized by endothelial cells,30 highlighted numerous capillary-sized vessels typically concentrated in the deep, and oftentimes more cellular, invasive front of the neoplasm. Consequently, MMP-2 might indirectly promote DFSP growth and invasive potential through its recognized role in tumor angiogenesis.v The absence of the activator molecule MMP-14 in the microvasculature of DFSP raises the possibility that one of the other 4 members of the membrane-type MMP family is responsible for activating MMP-2.15 Conversely, tumor cell expression of MMP-14 and MMP-2 in cFH might directly impact tumor growth characteristics. Several lines of experimental evidence support our contention. First, fibroblast culture experiments have shown that cellular up-regulation of MMP14 requires the presence of fibrillar (type I) collagen.18 The abundance of type I collagen in the reticular dermis provides the putative fibroblast-like tumor cell of cFH, the needed substrate to up-regulate MMP-14, which in turn governs expression of MMP-2. second, these 2 molecules facilitate cell migration through the interstitial matrix by their ability to cleave extracellular proteins, including type I collagen,11 laminin,32,33 and the extracellular domain of the hyaluronan receptor (CD44).34 Finally, they act in concert to activate and release transforming growth factor β,35,36 a molecule that modulates fibroblast motility and collagen production.

Our data coupled with results of previous studies indicate that the tumor cells of cFH have the capacity to produce a variety of MMPs.10,12-15 In contrast, DFSP cells have been shown to primarily produce cathepsins B, proL, and pro-D.17 The wide range of extracellular proteins normally cleaved by these 3 cathepsin molecules, including collagens, fibronectin, and proteoglycans, may provide an explanation for DFSP's notorious potential for invasive growth.17

In summary, the relatively high frequency of immunohistochemical expression of MMP-14 and MMP-2 in examples of cFH tested in our study strongly indicates that these MMPs may play a direct role in the pathogenesis of this tumor. Although MMP immunoexpression in DFSP was minimal, strong expression of MMP-2 in tumor microvasculature concentrated in the deep invasive front of the tumor suggests that endothelium-derived MMP-2 production might facilitate tumoral angiogenesis in this neoplasm. Future endeavors evaluating a broader range of MMPs and other connective tissue proteinases with more sensitive molecular techniques might further elucidate the pathogenesis of these tumors.

References

1. Calonje E, Fletcher CDM. Cutaneous fibrohistiocytic tumors: an update. Adv Anat Pathoi. 1994;1:2-1 5.

2. Kempson RL, Fletcher CDM, Evans HL, Hendrickson MR, Sibley RK. Fibrous histiocytomas. In: Kempson RL, Fletcher CDM, Evans HL, Hendrickson MR, Sibley RK, eds. Washington, DC: Armed Forces Institute of Pathology: 2001:113186. Atlas of Tumor Pathology, 3rd series, fascicle 30.

3. Barr R), Young EM, King DF. Non-polarizable collagen in dermatofibrosarcoma protuberans: a useful diagnostic aid. I Cutan Pathoi 1986;13:339-346.

4. Kamino H, lacobson M. Dermatofibroma extending into the subcutaneous tissue: differential diagnosis from dermatofibrosarcoma protuberans. Am I Surg Pathoi. 1990:14:1156-1164.

5. Meigel WN, Gay S, Weber L. Dermal architecture and collagen type distribution. Arch Dermatol Res. 1977:259:1-10.

6. Weber L, Meigel WN. Nature of collagen in dermatofibrosarcoma protuberans. Arch Dermatol Res. 1979:265:55-62.

7. Himelstein BP, Canete-Soler R, Bernhard E), Dilks DW, Muschel RJ. Metalloproteinases in tumor progression: the contribution of MMP-9. Invasion Metastasis. 1994:14:246-258.

8. Wolf C, Lefebvre O, Rouyer N, et al. Stromal proteinases and tumor progression. MedSci. 1994;10:507-515.

9. Stamenkovic I. Extracellular matrix remodeling: the role of matrix metalloproteinases. I Pathoi. 2003:200:448-464.

10. Soini Y, SaIoT, Oikarinen A, Autio-Harmainen H. Expression of 72 and 92 kDa type IV collagenase in malignant fibrous histiocytomas and dermatofibromas. Lab Invest. 1993:69:305-311.

11. Hurskainen T, Soini Y, Tuuttila A, Hoyhtya M, Oikarinen A, Autio-Harmainen H. Expression of the tissue metalloproteinase inhibitors TIMP-I andTIMP-2 in malignant fibrous histiocytomas and dermatofibromas as studied by in situ hybridization and immunohistochemistry. Hum Pathoi. 1996:27:42-49.

12. Unden AB, Sandstedt B, Bruce K, Hedblad MA, Stahle-Backdahl M. Stromelysin-3 mRNA associated with myofibroblasts is overexpressed in aggressive basal cell carcinoma and in dermatofibroma but not in dermatofibrosarcoma. / Invest Dermatol. 1996;107:147-1 53.

13. Thewes M, Worret Wl, Engst R, Ring J. Stromelysin-3 (ST-3): immunohistochemical characterization of the matrix metalloproteinase (MMP)-11 in benign and malignant skin tumours and other skin disorders. Clin Exp Dermatol. 1999; 24:122-126.

14. Cribier B, Noacco G, Peltre B, Grosshans E. Stromelysin 3 expression: a useful marker for the differential diagnosis dermatofibroma versus dermatofibrosarcoma protuberans. J Am Acad Dermatol. 2002:46:408-413.

15. Ohnishi Y, lto Y, Tajima S, lshibashi A, Arai K. lmmunohistochemical study of membrane type-matrix metalloproteinases (MT-MMPs) and matrix metalloproteinase-2 IMMP-2) in dermatofibroma and malignant fibrous histiocytoma. / Derma toi Sd. 2002:28:119-125.

16. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. I Histochem Cytochem. 1981:29:137-145.

17. Thewes M, Engst R, Boeck K, Ring ). Expression of cathepsins in dermal fibrous tumors: an immunohistochemical study. Eur j Dermatol. 1998;8:86-89.

18. Ruangpanit N, Chan D, Holmbeck K, et al. Gelatinase A (MMP-2) activation by skin fibroblasts: dependence on MTl -MMP expression and fibrillar collagen form. Matrix Biol. 2001:20:193-203.

19. TakinoT, Sato H, Shinagawa A, Seiki M. Identification of the second membrane-type matrix metalloproteinase (MT-MMP-2) gene from a human placenta cDNA library. I Biol Chem. 1995;270:2301 3-23020.

20. Zigrino P, Drescher C, Mauch C. Collagen-induced proMMP-2 activation by MT1-MMP in human dermal fibroblasts and the possible role of α2β1 integrins. Eur I Cell Biol. 2001 ;80:68-77.

21. Wong WR, Kossodo S, Kochevar IE. Influence of cytokines on matrix metalloproteinases produced by fibroblasts cultured in monolayer and collagen gels. J Formes Med Assoc. 2001:100:377-382.

22. Brenneisen P, Oh ), Wlaschek M, et al. Ultraviolet B wavelength dependence for the regulation of two major matrix-metalloproteinases and their inhibitor TIMP-I in human dermal tibroblasts. Photochem Photobiol. 1996:64:877-885.

23. Kut C, Hornebeck W, Groult N, Redziniack G, Godeau G, Pellat B. Influence of successive and combined ultraviolet A and B irradiations on matrix metalloelastases produced by human dermal fibroblasts in culture. Cell Biol Int. 1997:21:347-352.

24. Behrens P, Rothe M, Wellmann A, Krischler J, Wernert N. The Ets-1 transcription factor is up-regulated together with MMP1 and MMP9 in the stroma of pre-invasive breast cancer. J Pathoi. 2001:194:43-50.

25. Naito S, Shimizu S, Maeda S, Wang ), Paul R, Fagin JA. Ets-1 in an early response gene activated by ET-1 and PDGF-BB in vascular smooth muscle cells. Am I Physiol. 1998:43:472-480.

26. Simon MP, Pedeutour F, Sirvent N, et al. Deregulation of the platelet-derived growth factor B-chain gene via fusion with collagen gene COLlAI in dermatofibrosarcoma protuberans and giant-cell fibroblastoma. NatCcnet. 1997; 15: 95-98.

27. O'Brien KP, Seroussi E, DaI Cin P, et al. Various regions within the alpha-helical domain of the COLI Al gene are fused to the second exon of the PDCFB gene in dermatofibrosarcomas and giant-cell fibroblastomas. Genes Chromosomes Cancer. 1998:23:187-193.

28. Wernert N, Okuducu A-F, Pepper MS. EtS 1 is expressed in capillary blood vessels but not in lymphatics. J Pathol. 2003:200:561-567.

29. Hibbs MS. Expression of 92 kDa phagocyte gelatinase by inflammatory and connective tissue cells. Matrix Suppl. 1992;1:51-57.

30. Hanemaaijer R, Koolwijk P, Ie Clercq L, de Vree WJ, van Hinsbergh VW. Regulation of matrix metalloproteinase expression in human vein and microvascular endothelial cells: effects of tumour necrosis factor alpha, interleukin 1 and phorbol ester. Biochcm I. 1993;296:803-809.

31. Aimes RT, Quigley |P. Matrix metalloproteinse-2 is an interstitial collagenase: inhibitor-free enzyme catalyzes the cleavage of collagen fibrils and soluble native type I collagen generating the specific 3/4- and 1/4-length fragments. J Biol Chem. 1995:270:5872-5876.

32. Gianelli G, Falk-Marzillier J, Schiraldi O, Stetler-Stevenson WG, Quaranta V. Induction of cell migration by matrix metalloprotease-2 cleavage of laminin5. Science. 1997;277:225-228.

33. Koshikawa N, Giannelli G, Cirulli V, Miyazaki K, Quaranta V. Role of cell surface metalloprotease MTI-MMP in epithelial cell migration over laminin-5. J Cell Biol. 2000:148:615-624.

34. Kajita M, ltoh Y, Chiba T, et al. Membrane-type 1 matrix metalloproteinase cleaves CD44 and promotes cell migration. J Cell Biol. 2001:153:893-904.

35. Yu Q, Stamenkovic I. Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-B and promotes tumor invasion and angiogenesis. Genes Dev. 2000:14:163-176.

36. Mu D, Gambier S, Fjellbirkeland L, et al. The integrin α(v)β8 mediates epithelial homeostasis through MT1 -MMP-dependent activation of TGF-β1.1 Cell Biol. 2002:157:493-507.

David M. Weinrach, MD; Kirn L. Wang, MD; Elizabeth L. Wiley, MD; William B. Laskin, MD

Accepted for publication May 25, 2004.

From the Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, III.

The authors have no relevant financial interest in the products or companies described in this article.

Reprints: William B. Laskin, MD, Department of Pathology, Northwestern Memorial Hospital, Feinberg Pavilion 7-325, 251 E Huron St, Chicago, IL 60611-3053 (e-mail: wbl769@northwestern.edu).

Copyright College of American Pathologists Oct 2004
Provided by ProQuest Information and Learning Company. All rights Reserved