PAX2, PAX5, and PAX8 Expression in Pulmonary Neuroendocrine Tumors
Sean Kirby, Wendy Frankel, William Marsh, Fukuoka Junya, Jin Jen, Teri Franks, William Travis, Konstantin Shilo. The Ohio State University Medical Center, Columbus, OH; Toyama University Hospital, Toyoma, Japan; Mayo Clinic, Rochester, MN; JPC, Silver Spring, MD; Memorial Sloan-Kettering Cancer Center, New York, NY
Background: PAX genes 1-9 are transcription factors with DNA-binding capabilities, whcih play key roles in fetal development, tissue maintenance, and repair. Recent studies have documented PAX gene expression in a wide variety of cancers with preferential expression of PAX2 in mullerian and renal carcinomas; PAX5 in pulmonary small cell lung carcinoma (SCLC) cell lines, mullerian tumors, and B-cell lymphomas; and PAX8 in mullerian, renal, and thyroid carcinomas, as well as in pancreatic/duodenal well differentiated neuroendocrine tumors (NETs). Based on patterns of staining, these markers have been suggested as useful site specific tumor markers. Since the biological and diagnostic significance of PAX genes in NETs is not well elucidated, we investigated their expression and correlation with clinicopathological findings in a wide spectrum of pulmonary NETs.
Design: Tissue microarray based samples from 178 patients with pulmonary NET were studied for PAX8 (rabbit polyclonal, Cell Marque), PAX5 (clone 1EW, Novocastra), and PAX2 (rabbit polyclonal, Invitrogen) expression by immunohistochemistry. Nuclear expression was recorded as negative (0, absent), or positive with 1+ (weak), 2+ (moderate), and 3+ (strong) intensity. PAX expression was correlated with clinical-pathological variables including patients' age, gender, overall survival and tumor type, grade and stage.
Results: PAX8 expression was identified in 16.0% (23/144) of pulmonary NETs, and included 2.4% (1/42) of typical carcinoids, 11.1% (3/27) of atypical carcinoids, 33.3% (8/24) of large cell neuroendocrine carcinomas (LCNECs), and 21.6% (17/51) of SCLCs. PAX8 expression correlated with tumor grade (p<.05), but not with other clinical pathological variables. PAX2 expression was detected in 1 SCLC representing 0.5% (1/181) of pulmonary NETs. PAX5 expression was detected in 1 SCLC representing 0.7% (1/145) of pulmonary NETs.
Conclusions: PAX8 expression correlates with NET differentiation/grade and is observed in a significant percentage of LCNEC and SCLC. Since PAX8 expression is also seen in rare pulmonary carcinoids, its suggested site specificity for pancreatic/duodenal NETs should be used with caution. The vast majority of pulmonary NETs is negative for PAX2 and PAX5 (clone 1EW). The rarity of PAX5 expression in SCLCs in this study may be related to the choice of antibody clone.
Tuesday, March 20, 2012 1:00 PM
Poster Session IV # 295, Tuesday Afternoon