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


Mod Pathol. 2015;28(8):1043-1057. 

In This Article

Abstract and Introduction


The distinction between malignant mesothelioma and reactive mesothelial proliferation can be challenging both on histology and cytology. Recently, variants of the BRCA1-associated protein 1 (BAP1) gene resulting in nuclear protein loss were reported in hereditary and sporadic mesothelioma. Using immunohistochemistry, we evaluated the utility of BAP1 expression in the differential diagnosis between mesothelioma and other mesothelial proliferations on a large series of biopsies that included 212 mesotheliomas, 12 benign mesothelial tumors, and 42 reactive mesothelial proliferations. BAP1 stain was also performed in 70 cytological samples (45 mesotheliomas and 25 reactive mesothelial proliferations). BAP1 was expressed in all benign mesothelial tumors, whereas 139/212 (66%) mesotheliomas were BAP1 negative, especially in epithelioid/biphasic compared with sarcomatoid/desmoplastic subtypes (69% vs 15%). BAP1 loss was homogeneous in neoplastic cells except for two epithelioid mesotheliomas showing tumor heterogeneity. By fluorescence in situ hybridization, BAP1 protein loss was paralleled by homozygous deletion of the BAP1 locus in the vast majority of BAP1-negative tumors (31/41, 76%), whereas 9/10 BAP1-positive mesotheliomas were normal. In biopsies interpreted as reactive mesothelial proliferation BAP1 loss was 100% predictive of malignancy, as all 6 cases subsequently developed BAP1-negative mesothelioma, whereas only 3/36 (8%) BAP1-positive cases progressed to mesothelioma. On cytology/cell blocks, benign mesothelial cells were invariably positive for BAP1, whereas 64% of mesotheliomas showed loss of protein; all 6 cases showing BAP1 negativity were associated with histological diagnosis of BAP1-negative mesothelioma. BAP1 stain also showed utility in the differential of mesothelioma from most common pleural and peritoneal mimickers, such as lung and ovary carcinomas, with specificity and sensitivity of 99/70% and 100/70%, respectively. Our results show that BAP1 protein is frequently lost in mesothelioma, especially of epithelioid/biphasic subtype and is commonly associated with homozygous BAP1 deletion. BAP1 immunostain represents an excellent biomarker with an unprecedented specificity (100%) in the distinction between benign and malignant mesothelial proliferations. Finding BAP1 loss in mesothelial cells should prompt to immediately reevaluate the patient; moreover, it might be useful in mapping tumor extent and planning surgical resection.


The diagnosis of malignant mesothelioma and its distinction from serosal involvement by other malignant processes and from benign mesothelial proliferations is of primary importance not only to patient treatment and prognosis, but also for its forensic implications, because of the occupational nature of mesothelioma. In the United States, after a peak registered in 2000–2004, malignant mesothelioma incidence decreased,[1,2] with a rate between 2007 and 2011 of 1.9 and 0.4 per 100 000 inhabitants, respectively, in men and women.[3] Higher incidence has been observed in Italy where between 1993 and 2008, the National Mesothelioma Registry recorded 3.55 (male) and 1.35 (female) cases of malignant mesothelioma per 100 000 inhabitants.[4,5] Moreover, either the late asbestos ban or the long latency period for full transformation to mesothelioma validate the expected increasing incidence trend and peak within 10–15 years in Italy, as well as in other European countries.[4,6,7]

Immunohistochemistry is of definite support to the differential diagnosis between mesothelioma and serosal involvement by extraserosal neoplasms,[8,9] and although the distinction of mesothelioma from reactive mesothelial proliferations remains challenging, it is fundamentally based on the demonstration of stromal invasion,[10–12] with limited support by immunohistochemistry.[12] In fact, a variety of markers, such as desmin, epithelial membrane antigen, p53, IMP3, GLUT-1, CD146, and CD147, have been evaluated on both tissue and cytological samples, but none of them appeared to achieve sufficient diagnostic adequacy in the separation between malignant and benign mesothelial lesions.[13–36]

Using fluorescence in situ hybridization (FISH), the homozygous deletion of CDKN2A gene is found in 52–88% of mesotheliomas, but not in reactive mesothelial proliferations;[23,37–39] using a cut-off value of 10% positive mesothelial cells, p16 protein expression resulted to be closely related to CDKN2A status in some,[23,38] but not all, studies,[37] thus hampering its use as a reliable marker to distinguish mesothelioma from reactive mesothelial proliferations.

Recently, Carbone et al[40] identified a novel cancer syndrome related to germline BRCA1-associated protein 1 (BAP1) loss-of-function mutations and inherited with dominant autosomal transmission. Family members bearing BAP1 germline mutations show increased susceptibility to develop a variety of neoplasms, including uveal melanoma, cutaneous melanoma, atypical Spitz tumor, clear cell renal cancer, basal cell carcinoma, and mesothelioma, the latter occurring independently from occupational or environmental asbestos exposure.[40]

Initially identified as a BRCA1-binding protein, BAP1 is a deubiquitinating enzyme with C-terminal active hydrolase domain (UCH) and N-terminal nuclear localization signals (NLS1, NLS2).[41,42] Its tumor suppressor functions have been recently described and it has been found that BAP1 plays a role in cycle-cell progression, DNA ionizing radiation breaks repair, gene expression regulation through hystone H2A deubiquitinase activity, and subsequent chromatin remodeling.[42–49] As further evidence of BAP1 as a tumor suppressor gene, somatic gene inactivating deletions or point mutations have been detected in the same types of neoplasms associated with BAP1-cancer syndrome.[50–53] Bott et al[54] first reported BAP1 somatic variants in malignant mesothelioma, identifying gene losses and/or mutations in 22 of 53 pleural mesotheliomas (42%) that resulted in absent BAP1 immunoreactivity in tumor cells; interestingly, immunohistochemical analysis revealed a negative BAP1 staining also in eight cases in which mutations were not detected.[54] Yoshikawa et al[55] observed frequent deletion of the 3p21.1 region in mesothelioma primary cultures and cell lines; the same authors subsequently detected somatic biallelic BAP1 alterations in 14 of 23 cases of mesothelioma (61%);[56] similar percentages of cases showing loss of BAP1 protein expression were found in other more recent studies.[57,58]

In a recent study performed on tissue microarray containing 49 benign and 26 malignant mesothelial proliferations, Sheffield et al[39] showed that BAP1 immunostain separates benign from malignant processes, with a 100% specificity and 27% sensitivity; sensitivity increased to 58% when BAP1 stain was combined with 9p21 FISH analysis. In their study, a roughly equivalent number of epitheloid/biphasic and sarcomatoid mesothelioma cases were included, but whether BAP1 reactivity differed between subtypes was not specified.

In this study, we evaluated the diagnostic utility of BAP1 expression using a large series of histological samples, including 12 benign mesothelial tumors, 42 simple or atypical reactive mesothelial proliferations, and 212 mesotheliomas. The results fully supported the diagnostic role of anti-BAP1 as biomarker to distinguish benign from malignant mesothelial proliferations; specificity and sensitivity reached 100% and 66%, respectively, the latter raising to 69% when only epithelioid/biphasic subtypes were included. FISH analysis of BAP1 gene demonstrated that homozygous deletion of the BAP1 locus was commonly associated with BAP1 protein loss. Finally, the utility of BAP1 immunohistochemistry was also proven on cytological and cell-block samples.