Role for Activation of Matrix Metalloproteinases in the ... - CiteSeerX

4 downloads 0 Views 1MB Size Report
Among the members of the MMP family, MMP-2 is unique in that it is activated at the cell surface by membrane-type (MT) MMPs, particularly by. MT1-MMP.10.
Role for Activation of Matrix Metalloproteinases in the Pathogenesis of Pulmonary Lymphangioleiomyomatosis Kazuhiro Matsui, MD, PhD; Kazuyo Takeda, MD, PhD; Zu-Xi Yu, MD, PhD; William D. Travis, MD; Joel Moss, MD, PhD; Victor J. Ferrans, MD, PhD

● Background.—Matrix metalloproteinases (MMPs) have been shown to be involved in the pathogenesis of the destructive pulmonary lesions in patients with lymphangioleiomyomatosis (LAM); in the present report, the activation of these enzymes is examined. Objective.—To evaluate the role of MMPs and their activating enzymes, immunohistochemical and confocal microscopic techniques were used to localize a-smooth muscle actin (a-SMA), HMB-45, proliferating cell nuclear antigen (PCNA), MMP-2, membrane-type 1 MMP (MT1MMP), MT2-MMP, and MT3-MMP in lung tissues from 10 women with LAM. Tissue samples were obtained from 5 patients before treatment and in 5 patients after hormone treatment (progesterone and/or tamoxifen citrate). Results.—Staining for a-SMA and MMP-2 was present in all the abnormal smooth muscle cells (LAM cells) in both

groups. The percentages of PCNA-, MMP-2–, or MT1MMP–positive LAM cells were much higher in the untreated group than in the treated group, whereas the percentages of HMB-45–reactive LAM cells were similar in both groups. The reactions for MT1-MMP and PCNA were preferentially localized in small spindle-shaped LAM cells; the reaction for HMB-45 was found in large epithelioid LAM cells. Many of the PCNA-positive cells were also positive for MT1-MMP. Staining for MT2-MMP and MT3-MMP was negative. Conclusions.—This study demonstrates an association between cellular proliferation and the presence of MT1MMP in LAM cells. The activation of MMP-2 by MT1-MMP may play an important role in the destruction of lung tissue in this disorder. (Arch Pathol Lab Med. 2000;124:267–275)

P

connective tissue, with formation of cysts throughout the parenchyma.8 Three types of LAM cells are recognizable: (1) small round or oval cells, (2) small to medium-sized spindle-shaped cells, and (3) large epithelioid cells.8,9 However, we consider these types to be part of a continuous morphologic spectrum rather than indicative of 3 distinct, independent subpopulations of LAM cells. Furthermore, the reactivity for HMB-45 antibody, which is empirically used as a diagnostic marker for LAM, is most frequently observed in the large epithelioid cells.9 For this reason, we currently use reactivity for a-smooth muscle actin (a-SMA) to identify LAM cells in tissue sections, being careful to exclude reactive cells that are components of vascular, bronchial or bronchiolar structures. Early in the disease, the LAM cells tend to form nodular masses in which the small cells are centrally located and the epithelioid cells are peripherally distributed. In the late stages, the LAM cells tend to be much more irregularly arranged around the walls of the pulmonary cyst, in which they often appear intermingled with variable amounts of connective tissue. In this stage, most of the HMB-45–reactive LAM cells are medium to large and spindle shaped rather than epithelioid. The pathogenesis of these changes is poorly understood. However, the matrix metalloproteinases (MMPs) and their tissue inhibitors (TIMPs) deserve critical consideration as contributors to tissue destruction in LAM,8 since they play an important role in the remodeling of the extracellular matrix and basement membranes in normal and pathologic conditions. Recent work in our lab-

ulmonary lymphangioleiomyomatosis (LAM), an uncommon disorder that occurs either sporadically or in association with tuberous sclerosis, mainly affects women of reproductive age.1 The clinical signs and symptoms are distinctive and include recurrent spontaneous pneumothorax, chylothorax, hemoptysis, and slowly progressive dyspnea. Patients with LAM usually die of respiratory failure within 5 to 10 years after the onset of symptoms,1,2 although long-term survival has been documented.2 In view of the predominant occurrence of LAM in premenopausal women, hormone-based therapy has been used for this disorder. In some patients, stabilization or improvement occurs with hormonal therapy (progesterone, tamoxifen citrate, or oophorectomy), which is believed to be mediated through its effects on estrogen and progesterone receptors.3–5 Histologically, LAM is characterized by 2 major findings: (1) proliferation of abnormal smooth muscle cells (LAM cells), which are distinguished from other types of smooth muscle cells by their reactivity with HMB-45 antibody,6,7 and (2) progressive destruction of pulmonary Accepted for publication July 8, 1999. From the Pathology Section (Drs Matsui, Takeda, Yu, and Ferrans) and Pulmonary-Critical Care Medicine Branch (Dr Moss), National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Md; and Armed Forces Institute of Pathology, Washington, DC (Dr Travis). Reprints: Victor J. Ferrans, MD, PhD, National Institutes of Health, National Heart, Lung, and Blood Institute, Bldg 10, Room 2N240, 10 Center Dr, MSC-1518, Bethesda, MD 20892-1518 (e-mail: Ferransv@ gwgate.nhlbi.nih.gov). Arch Pathol Lab Med—Vol 124, February 2000

MMPs and MT-MMPs in Pulmonary LAM—Matsui et al 267

Clinical and Immunohistochemical Data on 10 Patients With Lymphangioleiomyomatosis Patient No.

Age, y

Untreated group 1 2 3 4 5 Mean 6 SD

35 46 55 32 37 41.0 6 9.4

Presenting Complaints*

D PNT, C W, PNT, C R-PNT, C W

Duration of Illness, mo

2 19 23 3 10 11.4 6 9.4

Therapy

HMB-45†, %

MT-1-MMP†, %

PCNA†, %

None None None None None

20 15 19 3 30 17.2 6 9.7

80 45 20 75 75 59.0 6 25.8‡

85 25 22 62 25 43.8 6 28.3‡

Treated group 6 26 D, PNT 67 T, M 22 10 NR 7 31 D, PNT 28 M 3 10 10 8 42 D, PNT 43 M 10 8 5 9 41 R-CHT 63 M 3 3 NR 10 36 D, PNT 75 M, L 3 3 5 Mean 6 SD 35.0 6 6.7 55.2 6 19.2 8.6 6 8.0 6.8 6 3.6‡ 6.7 6 2.9‡ * C indicates cough; CHT, chylothorax; D, dyspnea; L, leuprolide acetate; M, medroxyprogesterone acetate; NR, not reported; PNT, pneumothorax; R, recurrent; T, tamoxifen; and W, wheezing. † Percentage of lymphangioleiomyomatosis cells giving a positive reaction for this component. ‡ Values in the 2 groups of patients are significantly different (P , 0.05, Student’s t test).

oratory has demonstrated an association of MMPs with the destructive lesions of LAM. The LAM cells show strong immunoreactivity for MMP-2 (the 72-kd gelatinase A) and less intense reactivity for MMP-1 and MMP9 (the 92-kd gelatinase B).8 These MMPs are secreted as inactive proenzymes (pro-MMPs) and become functional only after activation. Among the members of the MMP family, MMP-2 is unique in that it is activated at the cell surface by membrane-type (MT) MMPs, particularly by MT1-MMP.10 The activation of pro-MMPs by MT-MMPs can be carried out by 4 members of this subgroup: MT1-MMP, MT2MMP, MT3-MMP, and MT4-MMP.10–13 Activation of MMP2, in which MT1-MMP appears to play the most important role, has attracted attention, because a close association between this MMP and remodeling of the extracellular matrix has been found in various types of tumors,14–16 ovulation,17 embryonic growth and differentiation,18 and human placentation.19 To our knowledge, no biochemical studies have been reported of the activation of MMP-2 in LAM. However, we observed in preliminary experiments that all LAM cells were immunoreactive for MMP-2, whereas only a subpopulation of these cells gave a positive immunohistochemical reaction for MT1-MMP. In view of these findings, the present study was designed to evaluate, by means of immunohistochemical and confocal microscopy techniques, the following aspects of LAM: (1) the expression and localization of MT1-MMP in LAM cells; (2) the relationship between the immunoreactivity for MT1-MMP and that for HMB-45; (3) the association of the immunoreactivity for MT1-MMP with proliferating cells, as reflected by their immunoreactivity for proliferating cell nuclear antigen (PCNA); and (4) the correlation of these immunohistochemical findings with the clinical stage of the disease, as documented in 2 groups of patients: an untreated group, from whom lung tissue was obtained at the time of initial evaluation, and a group from whom tissue was obtained after prolonged courses of hormonal treatment (progesterone, tamoxifen, or oophorectomy) and who were in more advanced stages of their disease. 268 Arch Pathol Lab Med—Vol 124, February 2000

MATERIALS AND METHODS Patients Studied The study group consisted of 10 women ranging from 26 to 55 years of age (mean 6 SD, 39.1 6 8.5 years) in whom the diagnosis of LAM was made on the basis of clinical, radiologic, pulmonary function, and histologic studies. These patients were divided into 2 groups (n 5 5 for each group), which differed in 3 important respects: (1) the patients in the first group had not undergone hormonal therapy at the time the tissues were obtained for study, whereas patients in the second group had received various forms of hormonal treatment (Table) before tissues were obtained at pulmonary transplantation; (2) the duration of the illness, from the onset of symptoms to the time when tissue was taken for study, was longer in the second group of patients; and (3) the latter patients had more advanced disease, which was evident both clinically and pathologically (their LAM lesions were more cystic and less proliferative than those in the other patients). In this communication, the patients in the first group will be referred to as ‘‘untreated’’ and those in the second group as ‘‘treated.’’ One of the patients included in the untreated group underwent pulmonary transplantation without first receiving hormonal therapy, after a short illness (10 months), due to rapid clinical deterioration. Tissues were fixed in buffered 10% formalin and embedded in paraffin. For histologic study, 5-mm-thick sections were stained with hematoxylin-eosin, the Masson trichrome and the Movat pentachrome methods, and the periodic acid–Schiff reaction with and without prior digestion by amylase.

Single Antibody Labeling For single immunolabeling, the sections were stained by the peroxidase method (EnVision System; Dako Corporation, Carpinteria, Calif) to investigate the localization and degree of reactivity for a-SMA, HMB-45, PCNA, MT1-MMP, MT2-MMP, MT3-MMP, and MMP-2. The antibodies against a-SMA (Dako; dilution, 1: 200), HMB-45 (Dako; dilution, 1:200), and PCNA (Dako; dilution, 1:100) were mouse monoclonal antibodies. The antibodies against MT-MMPs (Chemicon International Inc, Temecula, Calif; MT1MMP, dilution, 1:400; MT2-MMP, dilution, 1:200; and MT3-MMP, dilution, 1:200) were rabbit polyclonal antibodies. Two antibodies against MMP-2 were used. One of these was a rabbit polyclonal antibody (dilution, 1:1000), prepared at the National Institutes of Health (William Stetler-Stevenson). The other was a mouse monoclonal antibody (Chemicon; dilution, 1:200). The tissue sections MMPs and MT-MMPs in Pulmonary LAM—Matsui et al

were rehydrated and treated with 0.04% pepsin in 0.01-mol/L HCl for 10 to 15 minutes at 378C before the immunohistochemical procedures. When staining for PCNA, the sections were placed in heat-resistant plastic staining jars that contained an antigen retrieval solution (Citra; BioGenex, San Ramon, Calif), heated in a microwave oven for 10 minutes, and then allowed to cool to room temperature. The reactions for MT-MMPs also were performed on sections not treated with pepsin, because such treatment often resulted in loss of immunoreactivity for these components. Endogenous peroxidase was inactivated by incubation with 0.3% H2O2 in methanol for 15 minutes at room temperature. After washing in phosphate-buffered saline and blocking nonspecific binding of secondary antibody with 5% normal horse or goat serum for 30 minutes, sections were incubated with suitable primary antibodies for 30 minutes at room temperature. After 3 washes with phosphate-buffered saline, appropriate biotinylated secondary antibodies were applied for 30 minutes at room temperature. Finally, after washing with phosphate-buffered saline, the color was developed at the sites of reactivity with the Vector VIP substrate kit (Vector Laboratories) and counterstained with hematoxylin. After immunostaining for HMB-45, PCNA, MTMMPs, and MMP-2, the percentage of positive cells was determined in each patient by counting a total of 500 cells in at least 5 high-power fields.

Dual Staining for MT1-MMP Plus a-SMA, HMB-45, PCNA, or MMP-2 For the simultaneous visualization of the reactivity for MT1MMP and a-SMA, HMB-45, PCNA, or MMP-2 by laser scanning confocal microscopy, the sections were stained by the double indirect immunofluorescence method, using combinations of a rabbit polyclonal antibody against MT1-MMP (dilution, 1:50) and mouse monoclonal antibodies against a-SMA (dilution, 1:50), HMB-45 (dilution, 1:25), PCNA (dilution, 1:20), or MMP-2 (dilution, 1:25). Sections were incubated simultaneously with the two primary antibodies overnight at 48C. After staining with the two primary antibodies, the sections were reacted for 1 hour at room temperature with a mixture of the two secondary antibodies, which consisted of goat antirabbit IgG (Vector; dilution, 1:100) labeled with fluorescein isothiocyanate and horse antimouse IgG (Vector; dilution, 1:100) labeled with Texas red. This was followed by nuclear counterstaining with 49,6-diamidino-2-phenylindole (DAPI; Sigma Chemical Co, St Louis, Mo). Yellow fluorescence was considered indicative of colocalization of the red and green fluorescence signals. However, yellow autofluorescence was found in elastic fibers and lipofuscin granules. This autofluorescence was distinguished from that due to colocalization by its presence in unstained sections and in negative control preparations. A purple nuclear fluorescence resulted from the superimposition of the red staining for PCNA and the blue staining with DAPI. All preparations were examined with a confocal microscope (model TCS-4D/DMIR-BE; Leica, Heidelberg, Germany) equipped with argon and argon-krypton lasers.

Immunohistochemical Control Procedures Normal bronchial and vascular smooth muscle cells served as internal positive controls for a-SMA. Human carcinoma of the breast tissue was used as a positive control for MT1-MMP. Negative control immunohistochemical procedures included omission of the primary antibody and replacement of the primary antibody by normal rabbit or mouse IgG in appropriate concentrations.

Statistical Analysis All data are expressed as mean 6 SD. Differences were considered significant when P was .05 or less (Student t test).

RESULTS Histopathologic Findings The lung biopsy specimens obtained from 4 untreated patients were characterized by a combination of nodules Arch Pathol Lab Med—Vol 124, February 2000

of proliferating LAM cells and cystic spaces (Figure 1, A and B). The LAM cells were distributed mainly around bronchioles and blood vessels and along the walls of the cysts. In contrast, the lungs explanted at pulmonary transplantation from the other 6 patients contained prominent cysts with thin walls and a honeycomb-like arrangement (Figure 1, C and D). The cystic lesions were less cellular than the nodular lesions. Immunoperoxidase Staining A summary of the results of immunoperoxidase staining is presented in the Table. a-Smooth Muscle Actin. A positive reaction for aSMA was observed in normal vascular and bronchial smooth muscle cells. Immunoreactivity for a-SMA was diffuse and strong in the cytoplasm of LAM cells situated in areas of proliferation (Figure 2, A and B). HMB-45. Only a subpopulation of the a-SMA–positive LAM cells showed immunoreactivity for HMB-45. Most of the LAM cells that showed a strong reaction for HMB-45 were of the epithelioid type and tended to be localized in the periphery of the nodular LAM lesions just subjacent to the epithelial cells that covered the surfaces of the nodules (Figure 2, C and D). The percentages of LAM cells that were positive for HMB-45 were similar in the untreated group and the treated group (17.2% 6 9.7% vs 8.6% 6 8.0%, respectively; P 5 .26). Proliferating Cell Nuclear Antigen. Cells positive for PCNA were found in each of the 5 untreated patients but in only 3 of the 5 treated patients. In the other 2 patients, an adequate immunoreaction was not obtained, even after use of the antigen retrieval method. The PCNA reaction was more frequently positive in the small round or spindle-shaped LAM cells than in the large epithelioid cells (Figure 3, A). The mean percentage of PCNA-positive cells was significantly higher in the untreated group than in the treated group (43.8% 6 28.3% vs 6.7% 6 2.9%, respectively; P , .05). Some (, 10%) of the cuboidal epithelial cells covering the proliferating lesions also showed a positive reaction for PCNA. Other epithelial cells, including normal bronchial and alveolar epithelial cells, were also positive for PCNA but much less frequently. A few (, 5%) smooth muscle cells of normal blood vessels, bronchi, and bronchioles showed a positive reaction for PCNA. Matrix Metalloproteinase 2. As reported previously,8 most of the LAM cells gave a positive reaction with the rabbit polyclonal antibody against MMP-2. The reactivity for MMP-2 in LAM cells was less frequent when the corresponding mouse monoclonal antibody was used. With the latter antibody, the mean percentage of MMP-2–positive cells was higher in the untreated group than in the treated group (52.5% 6 24.0% vs 10.0% 6 6.3%, respectively; P , .05). The reason for this difference is not clear, but it appears to be related to the decreased reactivity of the monoclonal antibody. Membrane Type 1 MMP. Reactivity for MT1-MMP was found in the cytoplasm and sometimes also on the surfaces of LAM cells but not in the extracellular space. Most of the small or spindle-shaped LAM cells showed a moderate to strong reaction, whereas most of the large epithelioid cells tended to be negative or only weakly reactive (Figure 3, B through D). The percentage of MT1MMP–positive LAM cells was much higher in the untreated group than in the treated group (59.0% 6 25.8% vs MMPs and MT-MMPs in Pulmonary LAM—Matsui et al 269

Figure 1. Histopathologic findings in pulmonary lymphangioleiomyomatosis (LAM). A, Low-magnification view of lung from an untreated patient showing nodules of LAM cells and relatively thick-walled cysts (hematoxylin-eosin, original magnification 350). B, Higher-magnification view of tissue shown in A. The heterogeneity of LAM cells is evident. Small round or oval cells, small to medium-sized spindle-shaped cells, and large epithelioid cells are present (hematoxylin-eosin, original magnification 31000). C, Lung tissue obtained at the time of pulmonary transplantation, after hormonal treatment, shows prominent cysts with thin walls and a honeycomb-like arrangement (hematoxylin-eosin, original magnification 325). D, The wall of a cyst in tissue shown in C is less cellular and contains variable amounts of fibrous stroma (hematoxylin-eosin, original magnification 3400).

6.8% 6 3.6%, respectively; P 5 .01). Approximately 70% of the normal vascular and bronchiolar smooth muscle cells were weakly reactive. The endothelial cells in the LAM lesions showed a weak-to-moderate reaction, whereas those in other areas were negative. Some of the cuboidal epithelial cells covering the nodules of LAM cells showed a weak reaction; other types of epithelial cells were negative. About 80% of the alveolar macrophages also were weakly reactive. There was strong reactivity in the cytoplasm of human breast cancer cells in metastatic lesions in the liver, lung, and lymph nodes (Figure 3, F). Membrane Types 2 and 3 MMPs. No reactivity with the antibodies against MT2-MMP or MT3-MMP was observed in LAM cells or in other tissue components in any of the samples examined. Dual Immunofluorescent Labeling Membrane Type 1 MMP and a-SMA. In preparations stained by dual labeling for MT1-MMP and a-SMA, almost all the LAM cells showed a distinct cytoplasmic colocalization of these two components (Figure 4, A). In general, LAM cells in the central areas of the nodules were more intensely reactive for both components than were those in the periphery. A similar but less intense colocalization of MT1-MMP and a-SMA was observed in endo270 Arch Pathol Lab Med—Vol 124, February 2000

thelial cells in proliferating LAM lesions and in normal smooth muscle cells of the vascular and bronchiolar wall. In some areas, the reaction for MT1-MMP appeared to be stronger in the cellular membrane than in the cytoplasm of the LAM cells. Membrane Type 1 MMP and HMB-45. Colocalization of MT1-MMP and HMB-45 was not observed. Most HMB45–positive LAM cells were negative for MT1-MMP and vice versa (Figure 4, B and C). Membrane Type 1 MMP and MMP-2. The reaction obtained with the monoclonal antibody against MMP-2 was positive in many of the MT1-MMP–positive cells. A distinct colocalization of MT1-MMP and MMP-2 was observed in the untreated group but not in the treated group (Figure 4, D and E). Colocalization was most frequently evident in the smaller LAM cells located in central regions of the nodules. Membrane Type 1 MMP and PCNA. The reaction for MT1-MMP was positive in many of the PCNA-positive LAM cells, but the colocalization of these two reactivities was incomplete in all patients (Figure 4, F). In the periphery of the nodular lesions, MT1-MMP–negative large epithelioid LAM cells were not always negative for PCNA. MMPs and MT-MMPs in Pulmonary LAM—Matsui et al

Figure 2. Immunohistochemical reactivity of lymphangioleiomyomatosis (LAM) cells. Immunoperoxidase method and nuclear counterstaining with hematoxylin. A and B, a-Smooth muscle actin. Most LAM cells in a biopsy specimen from untreated patients show strong reactivity (A). Vascular smooth muscle cells are also strongly positive (arrow). Most of the LAM cells in the thin walls of the cysts from a treated patient also are strongly reactive (B) (original magnification 3250 for each). C, HMB-45. Positive cells are chiefly distributed in the periphery of LAM nodules in a biopsy specimen from an untreated patient (original magnification 3250). D, HMB-45. Higher-magnification view of tissue shown in C. A strong granular reaction is present in large epithelioid LAM cells adjacent to epithelial cells covering LAM nodules (D) (original magnification 31000).

COMMENT The MMPs are classified into classic MMPs and novel MMPs. The classic MMPs include 3 subgroups: (1) the interstitial collagenases (MMP-1, MMP-8, and MMP-13), (2) the gelatinases (MMP-2 and MMP-9, ie, the 72-kd gelatinase A and the 92-kd gelatinase B), and (3) the stromelysins (MMP-3, MMP-7, and MMP-10). Matrix metalloproteinase 12, a metalloelastase, is also included among the classic MMPs. The novel MMPs are secreted in active form or are associated with the cell surface. They include the membrane type (MT1-MMP, MT2-MMP, MT3-MMP, and MT4-MMP), which plays an important role in the activation of some of the classic MMPs.10–13 The classic MMPs are synthesized and secreted into the extracellular space as proenzymes, which require activation either by MT-MMPs or alternative mechanisms.20,21 Four mechanisms for the activation of the MMPs have been identified: (1) extracellular activation by non-MMP proteins (plasmin and thrombin)20,21; (2) extracellular activation by other MMPs (ie, MMP-7 can activate MMP-3 and MMP-9)22,23; (3) intracellular activation of MT-MMPs by furin, a transGolgi network serine protease24; and (4) MT-MMP membrane-associated activation of MMP-2.10 We have presented evidence showing that the immunoreactivity for MMPArch Pathol Lab Med—Vol 124, February 2000

1, MMP-2, and MMP-9 is greater in LAM cells than in vascular and bronchiolar smooth muscle cells.8 However, the antibodies used in these studies do not distinguish between the pro-MMPs and the activated forms of these enzymes. Other studies have demonstrated that the activities of MMPs in tissues are controlled at 3 levels: gene expression,25,26 activation of the pro-MMPs,27 and inhibition by complexation of the MMPs with specific TIMPs.28 The expression of MMPs is also controlled at the transcriptional level by cytokines, hormones, and growth factors.29,30 The importance of these modes of regulation of MMPs in LAM remains to be evaluated. The biochemical mechanisms involved in the activation of MMP-1 and MMP-9 are not well understood. The activation of MMP-2 occurs at the cell surface and appears to be mediated by binding of pro–MMP-2 or its complex with TIMP-2 to the membrane followed by proteolysis by MT1-MMP.27 The observations reported in the present study demonstrate the expression of MT1-MMP in a distinct subpopulation of LAM cells and the topographic distribution of this protein relative to that of a-SMA, HMB-45, MMP2, and PCNA. In contrast to this, reactivity for MT2-MMP and MT3-MMP could not be demonstrated in LAM cells. MMPs and MT-MMPs in Pulmonary LAM—Matsui et al 271

Figure 3. Immunohistochemical reactivity of lymphangioleiomyomatosis (LAM) cells. Immunoperoxidase method and nuclear counterstaining with hematoxylin. A, Proliferating cell nuclear antigen. Most of the small round or oval cells and small to medium-sized spindle-shaped LAM cells in a biopsy specimen from an untreated patient are positive (original magnification 31000). B through D, Membrane-type 1 matrix metalloproteinase (MT1-MMP). Most large epithelioid LAM cells in the periphery of a LAM nodule are weakly positive in a biopsy specimen from an untreated patient (B) (original magnification 31000). In contrast, small to medium-sized spindle-shaped LAM cells (C) in the wall of a cyst are strongly positive in a transplant specimen from an untreated patient (C) (original magnification 3400). Most of the small and medium-sized LAM cells in the wall of a cyst from a treated patient show a strong reaction (D) (original magnification 31000). E, Negative control (omission of primary antibody for MT1-MMP) showing same area as in C. No specific staining is evident (original magnification 3400). F, Positive control for MT1MMP. Most breast cancer cells give strongly positive reaction in their cytoplasm (original magnification 3800).

Staining for MT4-MMP could not be performed because of the lack of a suitable antibody. Membrane type 1 MMP was also expressed, but to a much lesser extent, in type II alveolar epithelial cells and smooth muscle cells of bronchi, bronchioles, and blood vessels and in alveolar macrophages. The staining patterns for MT1-MMP in these 272 Arch Pathol Lab Med—Vol 124, February 2000

types of normal cells were similar to those that we have observed in normal lung tissue from control subjects. Immunoreactivity for MMP-2 and for a-SMA was detected in the cytoplasm of practically all LAM cells. However, the reactivity for MT1-MMP was essentially limited to LAM cells that were small, centrally located in the nodMMPs and MT-MMPs in Pulmonary LAM—Matsui et al

Figure 4. Confocal images of sections of lung after dual labeling for membrane-type 1 matrix metalloproteinase (MT1-MMP), shown by a green fluorescence in all panels, plus either a-smooth muscle actin (a-SMA) (A), HMB-45 (B and C), MMP-2 (D and E), and proliferating cell nuclear antigen (PCNA) (F), all of which are indicated by a red fluorescence. Colocalization of these two colors is shown in yellow, and all nuclei show blue fluorescence due to counterstaining with 49,6-diamidino-2-phenylindole (DAPI). A, MT1-MMP and a-SMA. Lymphangioleiomyomatosis (LAM) cells in the center of proliferating nodules show a strong colocalization of these two components (yellow fluorescence) (original magnification 3400). B and C, MT1-MMP and HMB-45. Positive reactions for MT1-MMP and HMB-45 are independently seen both in a biopsy specimen from an untreated patient (B) and in lung explanted at transplantation in a treated patient (C). In the former, a cluster of HMB-45–positive cells is located in the periphery of a LAM nodule, whereas most MT1-MMP–positive cells are seen in the central area (original magnification 3400 for each). D and E, MT1MMP and MMP-2. A distinct colocalization of these two components (orange to yellow color) is observed in about half of the LAM cells in a biopsy specimen from an untreated patient (D) (original magnification 3600). In tissue from a treated patient (E), most of the LAM cells show a positive reaction for MMP2; however, a reaction for MT1-MMP is barely detectable (original magnification 3400). F, MT1-MMP and PCNA. Most small to medium-sized spindle-shaped LAM cells in proliferating nodules show a positive reaction for both components. The PCNA reaction appears as a purple color due to the superimposition of its red signal with the blue signal of the DAPI nuclear counterstain. The presence of the cytoplasmic reaction for MT1MMP and the nuclear reaction for PCNA in the same cells is clearly evident in many but not all cells (original magnification 3400).

ules, and unreactive for HMB-45. Most of the LAM cells with these characteristics were also positive for PCNA, a finding considered indicative of a high degree of proliferative activity. Thus, we interpret these observations as showing that, although all LAM cells are capable of producing MMP-2, only those cells that are actively proliferating have the capacity to convert the proenzyme to the catalytically active enzyme form of this MMP. These findings show that both MMP-2 and its activator, MT1-MMP, are present in the lesions of LAM and, therefore, can contribute to the lysis of connective tissue components in this disorder. This lysis also may facilitate the migration and spreading of LAM cells from developing nodules to previously uninvolved areas of the lungs. We have not obArch Pathol Lab Med—Vol 124, February 2000

served infiltration of cells such as neutrophils, eosinophils, or macrophages, which could serve as other sources of lytic enzyme activity directed against connective tissue components in LAM lesions. The reason for the reciprocal relationship between the expression of MT1-MMP and the reactivity for HMB-45 antibody in LAM cells is not currently understood. Although the ontogenetic relationships among the 3 histologic morphologically distinct subtypes of LAM cells remain unclear, Bonetti et al have suggested that these cells represent different functional stages of the same type of cell rather than 3 different types of cells.9 We also found that the percentage of HMB-45–positive cells in proliferating lesions of LAM varied from one patient to another MMPs and MT-MMPs in Pulmonary LAM—Matsui et al 273

and that such cells tended to be distributed more frequently beneath the epithelial cells covering the LAM nodules rather than in their central regions. Nevertheless, it was surprising to find that the percentage of HMB-45–positive LAM cells was similar in the treated and untreated groups of patients. Immunoprecipitation and immunoblotting studies have established that the antigen recognized by HMB-45 antibody is a 100-kd glycoprotein termed gp100.31 Other glycoproteins with a variety of smaller molecular sizes also reacted with HMB-45; at least some of these smaller molecules were considered to represent products of the degradation of gp100. This protein may play an important role in the differentiation of LAM cells. We have recently reported that in LAM cells there is an inverse relationship between reactivity for HMB-45 and PCNA.32 We consider that the nodular lesions of LAM are proliferative, in contrast to the cystic lesions that predominate in later stages of the disease. This hypothesis is consistent with the finding that the percentages of LAM cells positive for PCNA or MT1-MMP were much higher in the untreated than in the treated group of patients. In previous studies, we found the reactivity of LAM cells for TIMPs to be weak.8 This would suggest that the interaction of TIMPs with MMP-2 and MT1-MMP is of relatively limited importance in the pathogenesis of LAM lesions. HMB-45 has been used as a diagnostic marker for LAM. It remains to be determined, on the basis of analysis of a large group of patients, whether the frequency and distribution of LAM cells positive for MT1-MMP are useful in predicting the progression of this disorder. Currently, it is not known whether the lower frequency of MT1-MMP–positive LAM cells in tissue obtained at the time of pulmonary transplantation is due to a ‘‘burned-out’’ stage of the disease, hormonal therapy, or a combination of these two factors. Hormonal therapy would be expected to result in a decrease in cellular proliferation, as has been shown to occur in hormone-sensitive carcinomas of the breast.33 In this context, hormonal therapy has been reported to induce a decrease in MMP activity in these tumors.34 Similarly, progesterone treatment decreased MMPs in uterine stromal cells, cervical fibroblast, and trophoblasts.35–37 However, no data have been reported on the effects of this drug on MT1-MMP. In conclusion, the present study demonstrates that MMP-2 is present in all LAM cells, whereas MT1-MMP, its activating enzyme, is detectable only in a subpopulation of LAM cells. The latter cells are characterized by small size, a high degree of proliferative activity, and a negative reaction with HMB-45 antibody. The percentage of LAM cells that were positive for MT1-MMP was much lower in tissues from patients who had received hormonal therapy than in those of untreated patients. The prognostic significance of these findings needs to be evaluated in a larger group of patients. References 1. Taylor JR, Ryu J, Colby TV, Raffin TA. Lymphangioleiomyomatosis: clinical course in 32 patients. N Engl J Med. 1990;323:1254–1260. 2. Kitaichi M, Nishimura K, Itoh H, Izumi T. Pulmonary lymphangioleiomyomatosis: a report of 46 patients including a clinicopathologic study of prognostic factors. Am J Respir Crit Care Med. 1995;151:527–533. 3. Colley MH, Geppert E, Franklin WA. Immunohistochemical detection of steroid receptors in a case of pulmonary lymphangioleiomyomatosis. Am J Surg Pathol. 1989;13:803–807. 4. Brock ET, Votto JJ. Lymphangioleiomyomatosis: treatment with hormonal manipulation. N Y State J Med. 1986;86:533–536.

274 Arch Pathol Lab Med—Vol 124, February 2000

5. Zanella A, Toppan P, Nitti D, Lise M. Pulmonary lymphangioleiomyomatosis: a case report in postmenopausal woman treated with pleurodesis and progesterone (medroxyprogesterone acetate). Tumori. 1996;82:96–98. 6. Guinee DGJ, Feuerstein I, Koss MN, Travis WD. Pulmonary lymphangioleiomyomatosis: diagnosis based on results of transbronchial biopsy and immunohistochemical studies and correlation with high-resolution computed tomography findings. Arch Pathol Lab Med. 1994;118:846–849. 7. Hoon V, Thung SN, Kaneko M, Unger PD. HMB-45 reactivity in renal angiomyolipoma and lymphangioleiomyomatosis. Arch Pathol Lab Med. 1994;118: 732–734. 8. Hayashi T, Fleming MV, Stetler-Stevenson WG, et al. Immunohistochemical study of matrix metalloproteinases (MMPs) and their tissue inhibitors (TIMPs) in pulmonary lymphangioleiomyomatosis (LAM). Hum Pathol. 1997;28:1071–1078. 9. Bonetti F, Pea M, Martignoni G, Zamboni G, Iuzzolino P. Cellular heterogeneity in lymphangiomyomatosis of the lung. Hum Pathol. 1991;22:727–728. 10. Sato H, Takino T, Okada Y, et al. A matrix metalloproteinase expressed on the surface of invasive tumour cells. Nature. 1994;370:61–65. 11. Will H, Hinzmann B. cDNA sequence and mRNA tissue distribution of a novel human matrix metalloproteinase with a potential transmembrane segment. Eur J Biochem. 1995;231:602–608. 12. Takino T, Sato H, Shinagawa A, Seiki M. Identification of the second membrane-type matrix metalloproteinase (MT-MMP-2) gene from a human placenta cDNA library: MT-MMPs form a unique membrane-type subclass in the MMP family. J Biol Chem. 1995;270:23013–23020. 13. Puente XS, Pendas AM, Llano E, Velasco G, Lopez-Otin C. Molecular cloning of a novel membrane-type matrix metalloproteinase from a human breast carcinoma. Cancer Res. 1996;56:944–949. 14. Nomura H, Sato H, Seiki M, Mai M, Okada Y. Expression of membranetype matrix metalloproteinase in human gastric carcinomas. Cancer Res. 1995; 55:3263–3266. 15. Li H, Bauzon DE, Xu X, Tschesche H, Cao J, Sang QA. Immunological characterization of cell-surface and soluble forms of membrane type 1 matrix metalloproteinase in human breast cancer cells and in fibroblasts. Mol Carcinog. 1998;22:84–94. 16. Yoshizaki T, Sato H, Maruyama Y, et al. Increased expression of membrane type 1-matrix metalloproteinase in head and neck carcinoma. Cancer. 1997;79: 139–144. 17. Liu K, Wahlberg P, Ny T. Coordinated and cell-specific regulation of membrane type matrix metalloproteinase 1 (MT1-MMP) and its substrate matrix metalloproteinase 2 (MMP-2) by physiological signals during follicular development and ovulation. Endocrinology. 1998;139:4735–4738. 18. Fortunato SJ, Menon R, Lombardi SJ. Expression of a progelatinase activator (MT1-MMP) in human fetal membranes. Am J Reprod Immunol. 1998;39:316– 322. 19. Nawrocki B, Polette M, Marchand V, et al. Membrane-type matrix metalloproteinase-1 expression at the site of human placentation. Placenta. 1996;17: 565–572. 20. Corcoran ML, Hewitt RE, Kleiner DEJ, Stetler-Stevenson WG. MMP-2: expression, activation and inhibition. Enzyme Protein. 1996;49:7–19. 21. Nagase H. Activation mechanisms of matrix metalloproteinases. Biol Chem. 1997;378:151–160. 22. Imai K, Yokohama Y, Nakanishi I, et al. Matrix metalloproteinase 7 (matrilysin) from human rectal carcinoma cells: activation of the precursor, interaction with other matrix metalloproteinases and enzymic properties. J Biol Chem. 1995; 270:6691–6697. 23. Fridman R, Toth M, Pena D, Mobashery S. Activation of progelatinase B (MMP-9) by gelatinase A (MMP-2). Cancer Res. 1995;55:2548–2555. 24. Pei D, Weiss SJ. Furin-dependent intracellular activation of the human stromelysin-3 zymogen. Nature. 1995;375:244–247. 25. Brenner DA, O’Hara M, Angel P, Chojkier M, Karin M. Prolonged activation of jun and collagenase genes by tumour necrosis factor-alpha. Nature. 1989; 337:661–663. 26. White LA, Brinckerhoff CE. Two activator protein-1 elements in the matrix metalloproteinase-1 promoter have different effects on transcription and bind Jun D, c-Fos, and Fra-2. Matrix Biol. 1995;14:715–725. 27. Cao J, Sato H, Takino T, Seiki M. The C-terminal region of membrane type matrix metalloproteinase is a functional transmembrane domain required for progelatinase A activation. J Biol Chem. 1995;270:801–805. 28. Emmert-Buck MR, Emonard HP, Corcoran ML, Krutzsch HC, Foidart JM, Stetler-Stevenson WG. Cell surface binding of TIMP-2 and pro-MMP-2/TIMP-2 complex. FEBS Lett. 1995;364:28–32. 29. Bigg HF, Cawston TE. Effect of retinoic acid in combination with plateletderived growth factor-BB or transforming growth factor-beta on tissue inhibitor of metalloproteinases and collagenase secretion from human skin and synovial fibroblasts. J Cell Physiol. 1996;166:84–93. 30. Sato T, Ito A, Mori Y, Yamashita K, Hayakawa T, Nagase H. Hormonal regulation of collagenolysis in uterine cervical fibroblasts: modulation of synthesis of procollagenase, prostromelysin and tissue inhibitor of metalloproteinases (TIMP) by progesterone and oestradiol-17 beta. Biochem J. 1991;275:645–650. 31. Adema GJ, de Boer AJ, Vogel AM, Loenen WA, Figdor CG. Molecular characterization of the melanocyte lineage-specific antigen gp100. J Biol Chem. 1994;269:20126–20133. 32. Matsumoto Y, Horiba K, Usuki J, Chu SC, Ferrans VJ, Moss J. Characterization of the melanosome-containing smooth muscle cells in lymphangioleiomyomatosis (LAM): inverse correlation between markers of cell proliferation and

MMPs and MT-MMPs in Pulmonary LAM—Matsui et al

expression of melanosomal antigens. Am J Respir Cell Mol Biol. 1999;21:327– 336. 33. Raabe NK, Hagen S, Haug E, Fossaa SD. Hormone receptor measurements and survival in 1335 consecutive patients with primary invasive breast carcinoma. Int J Oncol. 1998;12:1091–1096. 34. Nutt JE, Lunec J. Induction of metalloproteinase (MMP1) expression by epidermal growth factor (EGF) receptor stimulation and serum deprivation in human breast tumour cells. Eur J Cancer. 1996;32A:2127–2135. 35. Shimonovitz S, Hurwitz A, Hochner-Celnikier D, Dushnik M, Anteby E,

Arch Pathol Lab Med—Vol 124, February 2000

Yagel S. Expression of gelatinase B by trophoblast cells: down-regulation by progesterone. Am J Obstet Gynecol. 1998;178:457–461. 36. Lockwood CJ, Krikun G, Hausknecht VA, Papp C, Schatz F. Matrix metalloproteinase and matrix metalloproteinase inhibitor expression in endometrial stromal cells during progestin-initiated decidualization and menstruation-related progestin withdrawal. Endocrinology. 1998;139:4607–4613. 37. Imada K, Ito A, Sato T, Namiki M, Nagase H, Mori Y. Hormonal regulation of matrix metalloproteinase 9/gelatinase B gene expression in rabbit uterine cervical fibroblasts. Biol Reprod. 1997;56:575–580.

MMPs and MT-MMPs in Pulmonary LAM—Matsui et al 275