INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary lymphangioleiomyomatosis (LAM) is an uncommon disorder that occurs exclusively in women (1). Its clinical manifestations include recurrent spontaneous pneumothorax, chylothorax, hemoptysis, slowly progressive dyspnea, and a high incidence of single or multiple angiomyolipomas, which are most frequently located in the kidneys. The pathological changes in LAM are characterized by nodules of abnormal, proliferating smooth muscle cells (LAM cells) and by the development of multiple, diffusely distributed pulmonary cysts (2, 5). The LAM cells grow in a haphazard arrangement, unlike the orderly patterns of organization of normal vascular and bronchiolar smooth muscle cells.

The occurrence of LAM in women primarily of reproductive age and the exacerbation of the disease by pregnancy or by the administration of estrogen, has led to speculation that the proliferation of the LAM cells depends on female sex hormones. Estrogen receptors (ER) and/or progesterone receptors (PR) have been demonstrated by biochemical techniques in lung tissue of patients with LAM (6, 7) or, more specifically, by immunohistochemical techniques in LAM cells (8). However, such receptors do not appear to be detectable in all patients with this disease. Furthermore, a clear correlation has not been found between the hormone-receptor status and the clinical response to hormonal therapy.

The LAM cells are morphologically heterogeneous, with a spectrum of shapes and sizes ranging from small to medium-sized, spindle-shaped cells to large epithelioid cells with a clear cytoplasm (13). In this article, we follow the classification of Bonetti and coworkers (14), in which the LAM cells are designated as either spindle-shaped or epithelioid. The LAM cells differ immunohistochemically from other types of smooth muscle cells in that they show reactivity for HMB-45, a mouse monoclonal antibody that reacts with a glycoprotein (gp-100) present in premelanosomes of cells of melanocytic lineage (19). We have recently reported studies describing the immunohistochemical heterogeneity of LAM cells (12, 16). We have confirmed the observations of Bonetti and coworkers (15) showing that the histochemical reactivity of LAM cells with HMB-45 is localized more frequently in the large, epithelioid cells. We have also demonstrated that LAM cells contain increased amounts of matrix metalloproteinases (MMPs), especially MMP-2 and, to a lesser extent, MMP-1 and MMP-9. These enzymes are considered to be important in the pathogenesis of the connective tissue destruction that results in the formation of the pulmonary cystic lesions in LAM (16).

In view of these considerations, we have undertaken the present study to evaluate the presence of ER and PR in the lesions of LAM and to correlate these findings with those concerning HMB-45, proliferating cells nuclear antigen (PCNA), and membrane-type 1-MMP (MT-1-MMP), the enzyme responsible for the activation of MMP-2. For this purpose, we studied two groups of patients that were classified according to whether their tissues were obtained before or after hormonal treatment. We used conventional immunohistochemical methods as well as multiple antibody labeling techniques in conjunction with laser scanning confocal fluorescence microscopy, with emphasis on the morphology of LAM cells positive for ER and PR, and on the colocalization of PR and ER with HMB-45, MT-1-MMP, and PCNA, since the latter two components appear to be markers for proliferating LAM cells.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients Studied

The study group consisted of 10 women (age range, 26 to 55 yr; mean ± SD, 39.8 ± 8.7 yr) in whom the diagnosis of LAM was made on the basis of clinical, radiologic, and histopathologic findings. Data on the therapy administered to these patients are presented in Table 1. Lung tissues were obtained either by biopsy in four patients at the time of their initial diagnostic evaluation or at pulmonary transplantation (six patients). The patients were classified into two groups, each consisting of five patients according to whether or not they had received hormonal therapy before the tissues were obtained for study. The duration of the disease, i.e., from the time of onset of symptoms to the time when tissue was obtained for study, ranged from 2 to 23 mo (mean, 11.4 mo) in the untreated group and from 28 to 75 mo (mean, 55.2 mo) in the treated group. One of the patients (number 5) in the untreated group underwent pulmonary transplantation after a relatively short period (10 mo) of illness, owing to rapid clinical deterioration.

Immediately after removal, all tissues were cut into thin slices and fixed by immersion with buffered 10% formalin. Particular care was taken to avoid any delay in the fixation process. The tissues were dehydrated, embedded in paraffin, and sectioned at a thickness of 5 µm. Morphologically normal lung tissues from lobectomy specimens resected for the evaluation of solitary nodules were used as negative controls. For histologic study, the sections were stained by the following methods: hematoxylin-eosin, periodic acid-Schiff with and without predigestion with diastase, Masson trichrome, and Movat pentachrome.

Single Antibody Labeling

For single labeling, the sections were immunostained by the peroxidase method (EnVision System; DAKO, Carpinteria, CA) to evaluate the localization and degree of activity of ER, PR, HMB-45, MMP-2, MT-1-MMP, and PCNA. The antibodies against ER (DAKO; dilution, 1:25), HMB-45 (DAKO; dilution, 1:200), and PCNA (DAKO; dilution, 1:100) were mouse monoclonal antibodies (MA). The antibodies against PR (DAKO; dilution, 1:50) and MT-1-MMP (Chemicon, Temecula, CA; dilution, 1:400) were rabbit polyclonal antibodies (PA). When staining for ER, PR, and PCNA, the sections were placed in an antigen-retrieval solution (Citra; BioGenex, San Ramon, CA), heated in a microwave oven for 10 min, and allowed to cool to room temperature (RT). The sections were then treated with 0.04% pepsin in 0.01 M HCl for 10 to 15 min at 37° C before immunohistochemical staining.

The reaction for MT-1-MMP was performed on sections not treated with pepsin, because such treatment often resulted in loss of the specific immunoreactivity for this component. Endogenous peroxidase activity was inactivated by incubation with 0.3% H₂O₂ in methanol for 15 min at RT. After washing in phosphate-buffered saline (PBS) and blocking nonspecific binding of the secondary antibody with 5% normal horse or goat serum for 30 min, the sections were incubated with the primary antibodies for 30 min at RT. After three washes with PBS, the peroxidase-labeled polymer from the EnVision System kit was applied for 30 min at RT. After washing with PBS, the color was developed with the Vector VIP substrate kit (Vector Laboratories, Burlingame, CA) and the sections were counterstained with Meyer's hematoxylin.

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. These control procedures consistently gave negative results. In addition, comparisons were made with the results obtained on sections of histologically normal lung. Sections of human carcinoma of the breast were used as a positive control for the immunoreactivity for MT-1-MMP. The percentages of positive cells were determined by counting 500 cells, identified morphologically as LAM cells, in a minimum of five high-power fields in each case.

Dual Labeling Procedures for PR/ER, PR/HMB-45, and MT-1-MMP/ER

Dual labeling using immunofluorescence methods and laser scanning confocal fluorescence microscopy was employed to evaluate the colocalization of the immunoreactivity for the following pairs of components: PR (PA) and ER (MA); HMB-45 (MA) and PR (PA); ER (MA) and MT-1-MMP (PA). The sections were stained by the double indirect immunofluorescence method, as previously described in detail (16). The following antibody dilutions were used: PR, 1:10; ER, 1:10; HMB-45, 1:25; and MT-1-MMP, 1:50. The sections were incubated overnight at 4° C with a mixture of the two primary antibodies. The primary PAs were reacted with a secondary antibody (goat anti-rabbit IgG; Vector; dilution, 1:100) labeled with fluorescein isothiocyanate (FITC). The primary MAs were reacted with a secondary antibody (horse anti-mouse IgG; Vector; dilution, 1:100) labeled with Texas red, followed by nuclear counterstaining with 4',6-diamidino-2-phenylindole (DAPI; Sigma Chemical, St. Louis, MO) for 15 min. In these preparations, a total of 500 immunoreactive cells were counted and classified into three categories, according to whether they showed green, red, or yellow fluorescence. The latter was indicative of colocalization of the red and the green fluorescent signals. However, a yellow autofluorescence was observed in elastic fibers and lipofuscin granules, both of which also showed this fluorescence in unstained sections and in negative control preparations. Identification of the morphological features was facilitated by nuclear staining with DAPI. Nuclei that were reactive for ER or PR and were counterstained with DAPI showed a purple (instead of red) fluorescence. Immunohistochemical control procedures similar to those described for single labeling were employed in conjunction with the dual staining methods. All preparations were examined with a confocal microscope (model TCS4D/DMIRBE; Leica, Heidelberg, Germany), equipped with argon and argon-krypton laser sources.

Statistical Analysis

All values are expressed as mean ± SD. Differences were analyzed using the Student's t test and were considered significant when a p value was  0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Histopathologic Observations

The histopathologic appearance of the lung tissues was clearly different in the untreated group (Figure 1A) and in the treated group (Figure 1B). In the former group, the LAM cells formed nodules of various sizes, and the cysts were smaller and thicker-walled than in the treated group. In the latter group, the proliferation of LAM cells was less pronounced; the nodules were organized in a haphazard manner, and the tissue had a honeycomb-like appearance. The LAM cells in both groups of patients were variable in shape and size. We consider that these variations formed part of a continuous spectrum of morphological diversity, and that they do not represent separate, distinct cell types. Nevertheless, for reasons of convenience (particularly with respect to the description of the immunohistochemical observations), we classified these cells into small to medium-sized, spindle-shaped cells and large, polygonal or epithelioid cells with a clear cytoplasm (13). This classification corresponds to that outlined by Bonetti and coworkers (15). In the untreated group, the nodules showed a preferential distribution of the spindle-shaped cells in their central regions. In contrast, the large epithelioid cells were most frequently situated in the periphery of the nodules. This pattern of organization was much less clearly evident in the treated group. Furthermore, in the latter group the epithelioid cells were found less frequently and tended to be smaller and less well developed than in the untreated group. These observations are similar to those made by Fukuda (17) on lung tissues from patients in the early and late stages of LAM.

View larger version (66K): [in this window] [in a new window]   Figure 1.   Histopathologic findings in pulmonary LAM. (A) Biopsy from an untreated patient. Note the morphological heterogeneity of the LAM cells: small to medium-sized spindle-shaped cells (S) and large epithelioid cells (E) with relatively large nucleoli. (Hematoxylin-eosin stain; original magnification, ×1,000.) (B) View of a cyst wall in lung of a treated patient. Larger LAM cells (arrowhead ) are spindle-shaped rather than epithelioid and are associated with fibrous stroma. (Hematoxylin-eosin stain; original magnification, ×1,000.)

Single Antibody Labeling

ER and PR. Immunoreactivity for ER was observed in four of the five patients in the untreated group, but not in any of the patients in the treated group (Table 1). The specific staining for ER was distributed mainly in the nuclei of the LAM cells (Figure 2A), but some cells showed weak cytoplasmic staining. In patients from the treated group, a few LAM cells (about 5%) showed weak cytoplasmic immunostaining for ER, but did not show a distinct reaction in the nuclei (Figure 2B). LAM cells were positive for PR in all five patients in the untreated group. The immunoreactivity for PR was confined to the nuclei of the LAM cells (Figures 2C and 2D). The intensity of the reactivity was higher for PR than for ER in each of the untreated patients. However, the frequencies of the two reactivities were not significantly different (29.6 ± 28.2 versus 10.0 ± 5.0%, respectively, p = 0.197). Although the intensity of the staining was variable, the reaction for ER was positive in large, epithelioid LAM cells and in a few of the larger spindle-shaped cells; however, it was negative in other LAM cells. In parallel with this, most of the large, epithelioid LAM cells and some of the relatively large, spindle-shaped LAM cells also gave a positive reaction for PR. Bronchiolar and vascular smooth muscle cells were negative for both ER and PR, as were all other types of cells examined in both groups of patients.

View larger version (119K): [in this window] [in a new window]   Figure 2.   Immunohistochemical reactivity of LAM cells. Single antibody labeling, peroxidase method and hematoxylin counterstain. (A and B) ER. Many of the large epithelioid LAM cells show a weak to moderate reaction in the nuclei (arrows) in a biopsy specimen from an untreated patient (A). No reaction is evident in tissue from a treated patient (B). (Each, original magnification ×1,000.) (C and D) PR. The nuclei of large epithelioid cells and some medium-sized spindle-shaped LAM cells give a positive reaction in a biopsy specimen (C ) and in a transplant specimen (D), both from untreated patients. (Each, original magnification ×1,000.) (E and F ) HMB-45. LAM cells positive for HMB-45 are observed mainly in the periphery of proliferating nodules (E; original magnification, ×250). Large epithelioid LAM cells from an untreated patient show a strong granular reaction in their cytoplasm (E, inset; original magnification, ×1,000), whereas spindle-shaped cells in the wall of a cyst in a treated patient also are positive (F; original magnification, ×1,000). (G) MT-1-MMP. Most LAM cells in the wall of a cyst in a treated patient show a strong positive reaction. (Original magnification, ×1,000.) (H ) PCNA. Most of the spindle-shaped cells from an untreated patient are positive for PCNA. (Original magnification, ×1,000.)

HMB-45. Some of the LAM cells gave a positive reaction with HMB-45 antibody in each of the 10 patients. The frequency of this reactivity varied from one patient to another (Figures 2E and 2F), but the percentages of positive cells were not significantly different in the two groups of patients (17.2 ± 9.7 and 8.6 ± 8.0%, respectively; p = 0.167). The majority of HMB-45-positive LAM cells were of the epithelioid type and were present in the peripheral areas of the nodules.

MT-1-MMP. In all 10 patients, many of the spindle-shaped cells gave a moderate to strong reaction for MT-1-MMP (Figure 2G), but the epithelioid LAM cells were negative or weakly positive. The percentage of positive cells was much higher in the untreated group than in the treated group (59.0 ± 25.8 versus 6.8 ± 3.6%, respectively; p = 0.01). The breast cancer cells used as positive controls showed a strong reaction for MT-1-MMP in the cytoplasm.

PCNA. A positive reaction for PCNA was observed in the nuclei of LAM cells in eight of the 10 patients. For reasons that are not clear, a technically adequate staining reaction for PCNA could not be obtained in two of the patients in the treated group. A positive reaction was more frequent in the spindle-shaped LAM cells than in the epithelioid cells (Figure 2H). The percentage of PCNA-positive cells was higher in the untreated group than in the treated group (43.8 ± 28.3 versus 6.7 ± 2.9%, respectively; p < 0.05).

Dual Antibody Labeling

PR and ER. PR and ER colocalized in about 30 to 50% of all hormone receptor-positive LAM cells. However, the frequency of this colocalization varied among different patients. The LAM cells that showed colocalization of PR and ER in the nuclei were observed in the periphery of the nodules and along the cyst walls (Figures 3A and 3B).

View larger version (110K): [in this window] [in a new window]   Figure 3.   Confocal microscopy images of dual labeling. (A and B) ER and PR in an untreated patient. Two images of the same field, showing staining for ER ( green fluorescence), PR (red fluorescence), and nuclear counterstaining with DAPI (blue fluorescence). A shows the combination of the red and blue signals and demonstrates their colocalization in almost half of the nuclei. B shows the combination of the red and green signals and demonstrates yellow fluorescence that results from the colocalization of these two signals in approximately half of the cells. Some nuclei stain only for PR (red fluorescence, arrowheads). (Original magnification, each, ×1,000.) (C ) Three-signal image showing staining for PR ( green), HMB-45 (red), and DAPI (blue). The red cytoplasmic staining for HMB-45 contrasts sharply with the staining of some nuclei for PR (arrows). The latter nuclei appear light blue as a result of the superimposition (colocalization) of the blue and green signals. (Original magnification, ×400.) (D) MT-1-MMP and ER. The reaction for MT-1-MMP is localized in the cytoplasm ( green fluorescence); the reaction for ER is present in the nuclei (purple color that results from the superimposition of the red and blue signals). There is no cellular colocalization of the red and green signals. Note areas in which LAM cells are positive only for MT-1-MMP (left half  ). In contrast, other areas show a positive reaction only for ER (right half  ). (Original magnification, ×400.)

PR and HMB-45. As mentioned previously, most of the LAM cells that gave a positive reaction for HMB-45 were of the large, epithelioid type. Such cells were present at the periphery of the LAM nodules and showed colocalization of HMB-45 and PR. In preparations stained by the dual labeling method, the nuclear staining for PR contrasted sharply with the cytoplasmic staining for HMB-45 (Figure 3C).

MT-1-MMP and ER. In the preparations stained for MT-1-MMP by the immunofluorescence method, the LAM cells showed a positive cytoplasmic reaction that was similar to that observed with the immunoperoxidase method. However, the immunofluorescent staining reaction appeared to be stronger in the cell membranes of some LAM cells. Colocalization of the immunoreactivity for ER (in the nuclei) and MT-1-MMP (in the cytoplasm) was observed only very rarely in LAM cells (Figure 3D). Thus, the reactions for these two components were distributed in different types of LAM cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study provides new insights into the expression of hormone receptors in the abnormal smooth muscle cells in the lungs of patients with LAM. Our results demonstrate a new aspect of the morphological and immunohistochemical heterogeneity of LAM cells, in which ER and PR were localized mainly in large, epithelioid LAM cells, rather than in the less differentiated, highly proliferative cells. Furthermore, the frequency of occurrence of receptor-positive cells and their intensity of staining differed in the two groups of patients in the study, being higher in the untreated group than in the treated group.

Colocalization of the Immunoreactivity for Hormonal Receptors and HMB-45 Antibody in LAM Cells

In agreement with previous morphological reports (14, 18), we recognized the heterogeneity of LAM cells according to their size and shapes. The large, epithelioid cells showed a greater frequency and intensity of staining for HMB-45 and ER and PR, with a high degree of colocalization, than did the other LAM cells. In contrast to this, the reactivities for MT-1-MMP and PCNA were localized mainly in the spindle-shaped LAM cells, which were situated in the central regions of the nodules and were negative for HMB-45, ER, and PR. We emphasize the fact that microwave treatment of the sections was required to obtain optimal immunohistochemical reactivities for ER and PR. Using this technique, we were able to demonstrate that ER and PR colocalized in 30 to 50% of the receptor-positive epithelioid LAM cells in the four untreated patients who showed this reactivity. Our observations on the nuclear localization of these receptors are similar to those previously described in other studies of LAM cells (8).

The preferential distribution of the receptor-positive cells in peripheral regions in the nodules of LAM cells suggests that such cells originate in the central regions of the nodules as small cells that undergo a process of growth and differentiation as they become peripherally displaced by newly forming LAM cells.

Relationship Between the Presence of Hormonal Receptors and Hormonal Therapy of LAM

The finding that ER and PR were easily detectable in all patients of the untreated group, but not in those studied after receiving hormonal treatment, is of particular interest. The most likely cause of this difference is downregulation of the receptors by intensive therapy with hormones before pulmonary transplantation. This downregulation may have reduced the immunoreactivity for these receptors to the point that no staining was observed under the conditions employed in the present study. The percentage of receptor-positive LAM cells decreased markedly in the treated group, whereas that of HMB-45-positive cells did not, suggesting that these two reactivities are modulated independently. Similarly, the frequency of the reactivity for PCNA also was considerably decreased in the treated group, indicating that hormonal treatment induces a reduction in the proliferation of LAM cells. Nevertheless, we cannot exclude the possibility that the reduction in the reactivity for PCNA and hormone receptors in the treated group of patients may have been related, at least in part, to burning out of the disease in its late stage. We have given careful consideration to the possibility that variations in the quality of the tissue fixation may have contributed to the decrease in immunoreactivity for hormone receptors that we observed in the treated group of patients. However, we have found the quality of tissue fixation to be comparable in both groups of patients. Therefore, we do not believe that the differences in immunoreactivity in these two groups of patients could be explained on the basis of inadequate fixation of the tissues removed at transplantation.

Our finding that ER and PR are distributed selectively in LAM cells that are large, epithelioid and are negative for MT-1-MMP and PCNA may be interpreted as suggesting that the hormonal therapy in this disorder is targeted against a subtype of LAM cell that is not the most important in contributing to the progression of the disease. Oophorectomy and the administration of lyprolide and progesterone have become the most widely used modalities of treatment for LAM. Therapy with tamoxifen is uncommonly employed for LAM because of its partial agonist activity with ER and the association of its use with possible exacerbation of the disease (2, 20). Oophorectomy has been of clinical benefit in some patients. The progression of the disease has been arrested by hormonal therapy only in relatively few patients. Progesterone therapy may have slowed the progression of the disease and led to clinical improvement in some patients (10, 20). Treatment with hormones may contribute to the decrease in the reactivity of LAM cells for PCNA and MT-1-MMP, and this would be expected to reduce the severity of the disease. Furthermore, recent studies have demonstrated that the administration of estrogen and progesterone exerts significant effects on the production of MMPs by hormone-sensitive tissues (21).

It seems likely that therapy with tamoxifen or progesterone may have induced a marked reduction in the corresponding hormone receptors in LAM cells. Such a response has been documented in breast cancer cells of patients receiving these types of long-term hormonal therapy (24). Antiestrogens are considered to be competitive inhibitors of estrogen at the level of ER protein, with the subsequent prevention, switched-on by the drug-ER complex, of many of the endogenous hormone-induced responses within tumor cells. The suppression of tumor growth by tamoxifen correlates with its potency to bind to ER. The tamoxifen-ER complex prevents estrogen- ER-mediated gene transcription, DNA synthesis, and cellular proliferation, and modulates the secretion of polypeptides, involved in the control of cell growth. It has been suggested that cellular proliferation also may be inhibited by tamoxifen through mechanisms not mediated by ER. Moreover, mutations in ER are considered to be related to the development of resistance to the therapeutic effects of tamoxifen. Nevertheless, it has not yet been determined whether the considerations outlined above are of functional importance in the hormonal treatment of LAM.

Nuclear Hormone Receptors and MMPs

Nuclear receptors control the production of MMPs through a variety of mechanisms (25). They act on the promoters of MMP genes to enhance or suppress transactivation. Few consensus hormone-responsive elements have been demonstrated in MMP promoters. Inhibition of MMPs occurs primarily at activator protein-1 (AP-1) sites, in which nuclear receptors form complexes on the DNA through interactions with AP-1 proteins and sequester Fos/Jun and/or decrease the messenger RNAs (mRNAs) for these transcription factors (26). Nuclear receptors and their ligands also can indirectly inhibit MMPs. For instance, both retinoids and glucocorticoids induce the transcription of tissue inhibitor of metalloproteinases (TIMPs) (27), and progesterone stimulates the production of transforming growth factor- (TGF-) (which suppresses MMP-7) (28). Furthermore, nuclear receptors bind to coactivators, corepressors, and components of the general transcriptional apparatus (29), but the potential role of these interactions remains to be determined.

Marbaix and colleagues showed that the release of MMP-1, MMP-2, and MMP-9 from cultured human endometrial explants was largely inhibited by physiological concentrations of progesterone (30). These effects, which were antagonized by mifepristone (RU38486), suggest that progesterone restrains endometrial tissue breakdown by blocking the secretion and activation of MMPs. Lockwood and coworkers evaluated the expression of MMPs in confluent cultures of human endometrial stromal cells (21). In this study, the basal output of proMMP-1 was unaffected by estradiol, but was inhibited by estradiol plus medroxyprogesterone. In contrast, mRNA levels for MMP-2 and TIMP-1 were not altered by either estradiol or medroxyprogesterone.

Imada and colleagues (31) showed that uterine cervical fibroblasts prepared from rabbits at 23 d of gestation did not produce proMMP-9; in contrast, interleukin-1, interleukin-1 or phorbol myristate acetate induced proMMP-9 production along with an increase of proMMP-9 mRNA. These effects were inhibited by progesterone.

As reviewed by Shimonovitz and associates (22), progesterone inhibits the secretion of MMP-9 by endometrial cells, myometrium, and cervical fibroblasts. It is of interest to note that the MMP-9 promoter contains a progesterone-responsive element that may explain the transcriptional activation of this MMP by progesterone. Shimonovitz and associates found that primary cell cultures of trophoblasts from the first trimester of pregnancy constitutively elaborated MMP-2 and MMP-9. Treatment with progesterone decreased the accumulation of MMP-9 in a dose-dependent fashion. Administration of ZK-98.299, a progesterone receptor antagonist, abolished this inhibition, thus providing support to the concept that this change is receptor-mediated.

Taken together, the studies summarized in this report clearly indicate that progesterone can exert a substantial inhibitory effect on the production of MMPs in hormone-sensitive tissues. This is in accord with the results of the present study, in which the reactivity for MT-1-MMP appeared to be considerably decreased after hormonal treatment of LAM. It will be of considerable interest to determine whether or not changes in MMPs and LAM are influenced at the transcriptional level by hormone-sensitive regulatory elements.

In conclusion, the results of the present study show that: (1) immunoreactivity for ER and PR can be demonstrated immunohistochemically, by means of the microwave method, in the LAM cells of patients who have not received hormonal treatment; (2) the immunoreactive LAM cells have an epithelioid morphology and are also reactive for HMB-45 antibody, and (3) the reactivity for ER and PR is markedly diminished or abolished by hormonal treatment, most likely because of downregulation of the receptors.