Immunohistological features related to functional impairment in lymphangioleiomyomatosis

Lymphangioleiomyomatosis (LAM) is a rare low-grade neoplasm that affects women of reproductive age, characterized by the proliferation of immature smooth muscle-like cells (LAM cells), mainly in the lungs [1‐3]. LAM often arises spontaneously in patients with no evidence of genetic disease and is present in approximately one-third of women with tuberous sclerosis complex (TSC) [4]. The main clinical features in LAM patients include progressive dyspnea and recurrent pneumothorax, whereas the main findings in pulmonary function tests (PFTs) are an obstructive pattern and reduced diffusion capacity for carbon monoxide (DL_CO). The clinical course is variable; some patients have slow progression over many years, while others develop progressive respiratory failure a few years after disease onset [5, 6].

The pathologic changes in pulmonary LAM (LAM lesions) are characterized by two distinct components: the cystic and the cellular components [7]. The former consists of multiple cysts distributed diffusely throughout the lungs and the latter is composed by two cellular subpopulations (LAM cells): spindle myofibroblast-like cells and polygonal cells with epithelioid morphology. LAM cells predominantly form nodules, mainly localized at the cysts walls, but small cell clusters can also be found scattered in the lung parenchyma [8]. The spindle cell population expresses smooth muscle proteins and forms the core of the nodules, surrounded by epithelioid cells that exhibit immunoreactivity for theHMB-45 antibody, a typical melanocytic cell marker [9]. The cell proliferation has been proposed to play a central role in the destruction of lung parenchyma due to high proliferative activity, growth factor release and metalloproteinase (MMP) production [10]. Moreover, the dysregulation and hyperactivation of the mTOR kinase pathway leads to increased protein translation and ultimately to inappropriate cellular proliferation, migration and invasion [11, 12].

Previous studies have shown the correlation of morphological parameters with prognosis determined by overall survival. Matsui et al. (2001) showed that certain histological criteria, such as the percentage of lung tissue cysts, the degree of cell infiltration into the tissue and the accumulation of hemosiderin-laden macrophages, correlate with worse prognosis [7]. Kitaichi et al. (1995) also suggested that higher grades of predominantly cystic lesions correlated inversely with survival from 2 to 5 years after biopsy. However, the previous studies that evaluated the histological parameters related to prognosis mainly focused on survival; the morphological determinants of impairment of PFTs have yet to be identified [13].

To investigate the association of immunohistological features with impairment in PFTs, we evaluated morphological parameters related to tissue destruction, LAM cell infiltration, and the expression of proteins related to the mTOR pathway, angiogenesis and enzymatic activity in LAM patients with different functional severities.

The study was approved by the review board for the human ethics committee of Sao Paulo University (CAPPesq-FMUSP n° 0623/09). The study is retrospective and used medical records and archived material from the Department of Pathology of Sao Paulo University Medical School.

Thirty-three women with LAM were included in the study, 7 (21%) of whom underwent lung transplantation while 26 (79%) underwent lung biopsy for diagnostic purposes. The clinical features are described in Table 1. The mean age at pulmonary biopsy or transplantation was 42 ± 8 years, 3 (9%) had underlying TSC, and 15 (46%) had renal angiomyolipoma. The most common symptom was dyspnea (70%), and 22 (67%) patients had a history of pneumothorax.

Pulmonary function tests are described in Table 2. The mean FEV₁ and DL_CO were 2.05 ± 0.93 L (72 ± 31% of predicted) and 17.4 ± 7.6 ml/min/mmHg (67 ± 29% of predicted), respectively.

The histological features were consistent with those previously reported [7, 12, 13]. There was significant variability in the number of LAM lesions among the cases, with extensive lung involvement in the more severe group. The immunohistochemical evaluation showed cytoplasmatic localization for each antibody. All analyzed proteins were expressed in the LAM cells. SMA was also expressed in the normal smooth muscle fibers of the airway and vessel walls; D240 in the lymphatic vessels; MMP9 in the macrophages, and VEGF-D in the areas of capillary proliferations, mainly within the pleural surface.

Table 3 compares the morphological, laboratorial, functional and survival/lung transplantation data between the two groups (less severe and more severe). Two patients presented borderline predicted values of FEV₁ (67 and 68%) and were not included in this comparative analysis. There was no difference in the expression of D2-40 between the more severe and the less severe groups. All other immunohistological parameters showed significantly higher values in the more severe group (p ≤ 0.002). Serum levels of VEGF-D were available for 23 patients; there was no difference between the less severe and the more severe groups. Figure 4 shows the variability in the number of LAM lesions, density of cysts, and density of nodules among cases in the two groups. Representative photomicrographs of immunohistological changes in both groups are shown in Fig. 5.The correlations among immunohistological parameters are shown in Table 4. The expression levels of VEGF-D, MMP-9 and mTOR in LAM cells were associated with both the density of cysts and the density of nodules. Further, the density of cysts showed a positive correlation with the density of nodules (r = 0.62; p < 0.001).

Table 5 presents the correlations between immunohistological findings and functional parameters. The density of cysts, the density of nodules and the expression of MMP-9 and VEGF-D were associated with impairment in PFTs (FEV₁ and DL_CO).

Figure 6 shows the Kaplan-Meier survival curves for tissue expression levels of VEGF-D, mTOR, and MMP-9. For each marker, two subgroups of patients were defined according to the marker median values. In the subgroup with lower median values, no patient died or underwent lung transplantation, whereas in the subgroup with higher median values 47% (VEGF-D), 53% (MMP-9) or 56% (mTOR) of the patients died or underwent lung transplantation (p ≤ 0.007). The median time to death or lung transplantation was 72 months.

In this study examining open lung biopsy specimens and lung explants from patients with LAM, we analyzed a range of histological and immunohistochemical markers that represent parameters of tissue destruction, lymphangiogenesis, LAM-cells nodules and mTOR-pathway alterations. Our results showed that 1) the more severe, advanced phase of disease presents with increased mTOR, VEGF-D and MMP-9 expression, as well as greater extent of cystic areas and LAM cell nodules; and 2) the main immunohistochemical and histological parameters associated with loss of pulmonary function were MMP-9, VEGF-D, the density of cysts and the density of nodules. These results provide new insights into the natural history of the disease and show for the first time that the tissue expression of VEGF-D in LAM cells is associated with cystic tissue destruction, LAM cell infiltration and loss of pulmonary function.

Lymphangioleiomyomatosis (LAM) is a rare disease of unknown etiology characterized by an abnormal infiltration of cells expressing smooth muscle protein in the lungs, lymph nodes or other locations, and often associated with renal angiomyolipoma [19, 20]. The disorder was previously classified as an interstitial lung disease and tumor-like lesion but is now considered a low-grade malignancy, with destructive metastasizing potential and evidence of clonal origin, being part of the spectrum of PEComatous tumors [1, 2, 21]. It has a variable course but may lead to progressive lung cystic destruction, respiratory failure and terminal lung disease [21].

Previous studies have shown that quantitative computed tomography (CT) can be used to evaluate the severity of lung disease in LAM patients. The quantitative CT scan grading has a good correlation with lung function tests, exercise performance and gas exchange [22]. Several authors have demonstrated a correlation between the extent of cysts on CT and the degree of airflow obstruction and reduced DLCO levels [23‐25]. In consonance with CT studies, we used a quantitative histological method to assess cystic tissue destruction and also demonstrated that the results are associated with loss of pulmonary function, reductions in both FEV₁ and DL_CO, which are considered important prognostic markers in LAM. Semiquantitative histologic studies already demonstrated that the percentage of predominantly cystic lesions is associated with prognosis. Kitaichi et al. (1995) used an individual score for each component of LAM lesions, that is, cysts and muscular-type cell infiltration, to demonstrate that predominantly cystic lesions have a tendency toward poor prognosis compared with predominantly muscular-type lesions [13]. The LAM histologic score established by Matsui et al. (2001), based on the total percentage of lung tissue involvement by both the smooth muscle-type cell and the cystic components, showed a strong correlation with the overall survival of LAM patients [7]. Both studies have an important aspect in common: the percentage of cystic component alone was a predictor of survival, independently of the extent of LAM cell infiltration. In this study, we used quantitative image-analysis methods to measure LAM lesion components individually (cysts and nodules), adding new elements to previous histological studies on LAM. Our results showed that both tissue destruction and LAM cell infiltration, assessed respectively by the density of cysts and the extent of LAM cell nodules, are associated with loss of pulmonary function. The stronger association of cystic areas with lower values of FEV₁ indicates that cystic tissue destruction is likely to be more relevant than cell infiltration in disease progression, as previously suggested. However, patients with more severe disease presented an increased extent of LAM cell nodules with higher expression of MMP9, VEGF-D and mTOR, suggesting that rather than a quiescent predominantly cystic phase, LAM cells present high metabolic activity in the advanced phases of the disease, likely participating in the process of tissue destruction.

Metalloproteinases are involved in both cystic destruction and the metastatic potential associated with pulmonary LAM cells [26]. The pathogenesis of cystic destruction of lung parenchyma seems to be associated with an imbalance between MMPs and tissue inhibitor of MMPs, with increased levels of MMPs [10]. Our results are consistent with this concept, showing a strong association of MMP expression with the extent of cystic areas and also with functional impairment.

Vascular endothelial growth factor (VEGF) is a major angiogenic growth factor produced by malignant cells. VEGF-D, a ligand for the lymphatic growth-factor receptor VEGFR-3/Flt-4, induces the formation of lymphatics and promotes the spread of tumor cells to the lymph nodes [27, 28]. Seyama et al. reported that levels of VEGF-D, but not VEGF-A or VEGF-C, are elevated in patients with sporadic lymphangioleiomyomatosis compared with healthy controls [29]. It has been suggested that the serum level of VEGF-D is a valuable marker of LAM diagnosis and severity and to evaluate the response to treatment [30‐32]. VEGF-D is upregulated through the mTOR signaling pathway, is expressed by LAM cells, and is associated with LAM cell proliferation, dissemination and lymphangiogenesis [28‐30, 33]. Previous studies have assessed the tissue expression of VEGF-C by LAM cells as a marker of lymphangiogenesis, which correlated with the LAM histologic score [19]. Xu et al. (2013) and Baldi et al. (2014) reported that levels of serum VEGF-D are higher in patients with a greater extent of cysts as evaluated by CT scan [23, 34]. To the best of our knowledge, this study is the first to show an association of the tissue expression of VEGF-D with pulmonary cystic destruction and loss of pulmonary function in LAM patients.

The 10-year survival of LAM patients varies from 40 to 91% [35, 36] and DL_CO, FEV₁, histology score and tomographic changes are considered the main prognostic markers [37]. We observed that, except for D2-40, all immunohistological parameters evaluated were significantly different between the two severity groups, which reinforces their association with functional impairment. Furthermore, the observation that patients with higher tissue expression of mTOR, VEGF-D and MMP-9 presented significantly lower survival or a greater risk to undergo lung transplantation, emphasizes a potential association of the extent of those immunohistological changes with prognosis.

We conclude that severe LAM patients present with high expression of VEGF-D, mTOR, and MMP-9, as well as cystic destruction and LAM cell infiltration. The tissue expression levels of VEGF-D and MMP-9 are important parameters associated with the loss of pulmonary function and could be considered as potential severity markers in open lung biopsies of LAM patients. Future studies are necessary to establish whether the expression of such tissue markers affects the response to treatment modalities.