BAP1 (BRCA1-Associated Protein 1) Is a Highly Specific Marker for Differentiating Mesothelioma From Reactive Mesothelial Proliferations

Marta Cigognetti; Silvia Lonardi; Simona Fisogni; Piera Balzarini; Vilma Pellegrini; Andrea Tironi; Luisa Bercich; Mattia Bugatti; Giulio Rossi; Bruno Murer; Mattia Barbareschi; Silvia Giuliani; Alberto Cavazza; Gianpietro Marchetti; William Vermi; Fabio Facchetti

Disclosures

Mod Pathol. 2015;28(8):1043-1057. 

In This Article

Results

Normal Mesothelium and Benign Mesothelial Tumors Regularly Express BAP1 Protein

In all control cases the flat monolayer surface mesothelium showed moderate to intense nuclear BAP1 expression. Similarly, BAP1 signal was easily detectable in all benign mesothelial tumors, independently from subtype and location (Figure 1).

BAP1 Loss in Reactive Mesothelial Proliferations Is Predictive of Malignancy

In all, 36 cases of reactive mesothelial proliferations (25/27 simple and 11/15 atypical) showed BAP1 nuclear expression, whereas 2/27 simple reactive mesothelial proliferations and 4/15 cases of atypical reactive mesothelial proliferation were BAP1 negative. All 6 (100%) BAP1-negative reactive mesothelial proliferation cases were diagnosed having BAP1-negative mesothelioma within a time period between 2 and 104 weeks, whereas only 3/36 (8%) cases of BAP1-positive reactive mesothelial proliferation (1 simple and 2 atypical) had mesothelioma diagnosed within a time period between 3 and 56 weeks (Figure 2).

Figure 2.

BAP1 immunostain in cases of noninvasive mesothelial proliferations. (a and b) Case 28 in Table 2. Hematoxylin and eosin shows the irregular mesothelial proliferation with papillae (a) and cytological atypical features (inset); BAP1 is strongly reactive in mesothelial and stromal-inflammatory cells. (c–f) Three different cases of mesothelial proliferation showing BAP1 negativity that were associated with mesothelioma (c and d: case 19, e: case 18; and f: case 40, all from Table 2).

The incidence of mesothelioma in BAP1+ and BAP1− reactive mesothelial proliferation cases was significantly different (8% vs 100%, respectively; Fisher's exact test: P<0.0001); considering all reactive mesothelial proliferation excluding the 5 cases associated with pleural metastases (cases 22–26; Table 2 ) the lack of BAP1 expression had 100% (95% confidence interval: 54–100%) positive predictive value and 90% (95% confidence interval: 74–98%) negative predictive value for mesothelioma development, respectively.

BAP1 Expression Is Frequently Lost in Malignant Mesothelioma, Especially in the Epithelioid Subtype

Loss of BAP1 was observed in 139 of the 212 mesothelioma cases (66%), particularly in the epithelioid (128/184; 70%) and the biphasic subtypes (9/15; 60%) (Figure 3a–d), whereas it was infrequent in the sarcomatoid and desmoplastic variants (2/13; 15%). BAP1 protein loss was consistently observed in tumor cells in all cases except for two epithelioid mesotheliomas that showed tumor heterogeneity for the marker (Figure 3e and f). Interestingly, some invasive mesothelioma cases were associated with a surface mesothelium layer showing expression of BAP1 similar to that of the underlying invasive component, independently from its morphological growth pattern (single flat layer versus papillary) and the degree of cell atypia. On occasion, in BAP1-negative cases, focal areas of surface mesothelium with preserved BAP1 expression were detected.

Figure 3.

(a–d) Examples of three BAP1-negative malignant epithelioid and biphasic mesotheliomas. (a) The BAP1-positive lung parenchyma (upper left) infiltrated by the mesothelioma (lower right); (b) BAP1 is expressed by inflammatory cells and vascular endothelium within the tumor. (c and d) A case of biphasic mesothelioma double stained for BAP1 (brown) and epithelial membrane antigen (blue), the latter being helpful to identify the BAP1-negative spindle cells in the sarcomatoid areas. (e and f) The only two cases of epithelioid mesothelioma containing distinct tumor populations regarding BAP1 expression.

In five additional cases, extensive sampling from wide surgical resection of deeply invasive mesothelioma included tissue blocks defined as atypical reactive mesothelial proliferation as no invasion was detectable in the underlying stroma. In four of them, both the atypical reactive mesothelial proliferation and the invasive mesothelioma lacked BAP1, whereas the fifth case represented one of the mesotheliomas containing both BAP1+ and BAP1− tumoral cells and such heterogeneous neoplastic population was also clearly noticeable in the atypical reactive mesothelial proliferation areas.

Comparing the results of BAP1 expression in benign mesothelial lesions and mesothelioma, the sensitivity and specificity of BAP1 loss for diagnosing mesothelioma was 66% (95% confidence interval: 59–72%) and 100%, respectively; sensitivity increased to 69% (95% confidence interval: 62–75%) if only cases of epithelioid and biphasic mesotheliomas were considered.

BAP1 Expression on Cytological and Cell-block Samples

Immunostain on cytological and cell-block samples was easily evaluable similar to tissue sections (Figure 4). In 15 cytological samples from obvious inflammatory pleural effusions and in 2 cases containing lung adenocarcinoma tumor cells, mesothelial cells regularly expressed BAP1. In 29 out of 45 cases of mesothelioma (64%), BAP1 stain resulted negative in atypical mesothelial nuclei; one of these cases contained a mixed BAP1+ and BAP1− tumor cell population according to the finding of BAP1 heterogeneity on the corresponding biopsy. In a further case, a BAP1− cytological sample belonged to a patient on whom a histological diagnosis of BAP1+ mesothelioma was done 2 years before. Six out of 8 samples containing atypical mesothelial cells of unknown significance were BAP1 negative and all of them had a previous, concomitant, or subsequent biopsy showing BAP1− mesothelioma. In the remaining two cases, mesothelial cells were BAP1 positive: one patient had BAP1+ mesothelioma and the other a non-Hodgkin's lymphoma involving the pleura.

Figure 4.

BAP1 immunostain on cytological (a) and cell-block (b) samples from two cases of epithelioid mesothelioma showing BAP1-negative malignant cells surrounded by numerous BAP1-positive inflammatory cells. Double stain for BAP1 (brown) and CD11c (blue) reveals the histiocytic nature of the BAP1-positive cells embedding the tumor papillae (b). (c–h) Shown is a patient (no. 42 in Table 2) who showed pleural effusion containing rare atypical cells (c) that resulted negative for BAP1 (d, single immunostain for BAP1; e, double immunostain for BAP1 and cytokeratin 5/6, respectively brown and blue); pleural biopsy was unremarkable (f); 11 months later, frank invasive BAP1-negative mesothelioma was diagnosed (g and h).

BAP1 Is Expressed in the Vast Majority of Lung and Ovarian Carcinomas

In order to evaluate whether BAP1 might also be useful in the distinction between malignant mesothelioma and potential mimickers involving the pleura and peritoneum, a large cohort of tumor samples from lung and ovary carcinomas were analyzed. With the exception of one acinar and one solid lung adenocarcinomas that were totally BAP1 negative, all other cases strongly expressed BAP1 ( Table 4 and Supplementary Figure 1 http://www.nature.com/modpathol/journal/v28/n8/suppinfo/modpathol201565s1.html?url=/modpathol/journal/v28/n8/full/modpathol201565a.html). Interestingly, a minority of cases contained a BAP1-negative component associated with the predominant BAP1-positive one (Supplementary Figure 1 http://www.nature.com/modpathol/journal/v28/n8/suppinfo/modpathol201565s1.html?url=/modpathol/journal/v28/n8/full/modpathol201565a.html). All eight samples of epithelioid hemangioendothelioma of the lung resulted intensely positive for BAP1 (Supplementary Figure 1 http://www.nature.com/modpathol/journal/v28/n8/suppinfo/modpathol201565s1.html?url=/modpathol/journal/v28/n8/full/modpathol201565a.html).

BAP1 Loss in Mesothelioma Is Frequently Sustained by Homozygous Deletion of BAP1 Gene by FISH

All six control samples containing normal mesothelium showed unaltered BAP1 gene pattern on FISH (2G–2R). In 36 of 41 (88%) BAP1-negative mesothelioma cases, at least one abnormality at the BAP1 locus was observed. As defined by the established cutoff values, homozygous deletion was detected in 31 cases (76%; 1/2G–0R or >2G–0R; Figure 5b); 6 of them also contained cells with heterozygous deletion. Heterozygous-only deletion was found in one case (2G–1R or G>R), whereas three cases showed chromosome 3 monosomy (1G–1R). Finally, one case showed unexpected chromosome 3 polysomy.

Figure 5.

Fluorescence in situ hybridization using the BAP1/CEN3q probe in a case of BAP1-positive mesothelioma that shows normal copy numbers (a) and in a case of BAP1-negative mesothelioma with homozygous deletion of BAP1 (b).

Remarkably, in one mesothelioma case, tumor heterogeneity at the protein level was supported by heterogeneous deletion of the BAP1 locus in mesothelioma cells. Eight of 10 BAP1-positive mesotheliomas showed no BAP1 anomalies (Figure 5a). One homozygous deletion and one gene amplification pattern were detected in the remaining two cases, respectively.

processing....