CD10 (neprilysin) expression: a potential adjunct in the distinction of hibernoma from morphologic mimics

Abstract

Although the morphologic diagnosis of hibernoma is usually straightforward, some hibernomas have atypical morphologic features, mimicking atypical lipomatous tumors/well-differentiated liposarcomas (ALT/WDLs). In addition, the multivacuolated brown fat cells may be mistaken for lipoblasts by pathologists, especially those without significant soft tissue tumor exposure. Thus, we continue to receive in consultation cases of hibernoma sent for MDM2 fluorescence in situ hybridizationtestingto exclude ALT/WDL. Testing hibernomas for MDM2 amplification, however, adds cost and delays the final diagnosis. Recently, we have noted expression of neprilysin (CD10, CALLA), a zinc-dependent metalloproteinase involved in the inactivation of various peptide hormones, in brown fat cells, and wished to explore the potential utility of this widely available, inexpensive ancillary test in the differential diagnosis of hibernoma. Formalin-fixed, paraffin-embedded tissue sections from well-characterized cases of hibernoma (n=48), brown fat (n=21), ALTs/WDLs (n=17), pleomorphic liposarcomas (PLPSs) (n=6), lipomas (n=5), and fat necrosis (n=5) were immunostained for CD10, using a commercially available antibody and routine laboratory protocols. CD10 expression was evaluated in both adipocytes and in surrounding stromal cells. The hibernomas occurred in 28 men and 20 women, ranging from 11 to 76 years of age and involved the extremities (n=25), pelvis (n=7), abdomen/pelvis/retroperitoneum (n=7), head and neck region (n=6), back (n=2), and chest (n=1). All showed diffuse, strong CD10 expression in multivacuolated brown fat cells and in the majority of adjacent univacuolated fat cells. Brown adipose tissue from various anatomic structures showed an identical pattern of immunoreactivity. In contrast, CD10 expression was present in the adipocytes of only 3 of 17 (18%) ALTs/WDLs and was absent in lipomas and fat necrosis. Lipoblasts expressed CD10 in 3 PLPSs. Expression of CD10 by surrounding fibroblastic stromal cells was more widespread, present in 13 hibernomas, 10 ALTs/WDLs, 1 instance of fat necrosis, 6 PLPSs, and 4 examples of brown fat. We conclude that immunohistochemistry for CD10 may represent a useful, rapid and inexpensive ancillary test in the differential diagnosis of hibernoma from potential morphologic mimics, especially when morphologic features favor hibernoma. CD10 expression in adipocytes, however, should be rigorously distinguished from fibroblastic stromal cell CD10 expression, a nonspecific finding.