A comprehensive analysis of GATA3 expression in carcinomas of various origins with emphasis on lung carcinomas

GATA3 is a transcription factor involved in the embryogenesis of multiple human tissues and organs and in maintaining cell differentiation and tissue homeostasis in the adult organism. GATA3 is also involved in carcinogenesis and is regarded as a sensitive marker for urothelial and breast carcinomas, although its expression in carcinomas of non-breast/urothelial origin has been frequently reported. In this study, we sought to examine the extent and intensity of GATA3 expression in various carcinomas, mainly lung, urothelial, breast, and various other primary sites. Patients with breast carcinoma (n=40), carcinoma of the urinary bladder/renal pelvis (n=40), lung carcinoma (n=110), and various other origins (n=45) were included in the study. 165 patients had a primary tumor diagnosis, and 70 cases had a metastatic tumor diagnosis. Our results showed that GATA3 expression was significantly more common in carcinomas of the breast, urinary bladder, and renal pelvis compared to all other origins. All primary and 93% of metastatic urinary bladder carcinomas and 94% of primary and 80% of metastatic breast carcinomas expressed GATA3. Expression was lower in the non-urothelial histology of urinary primaries and in triple-negative breast carcinomas (TNBC). Focal staining, mostly faint, was seen in 5.6% of the primary lung adenocarcinomas and 35% of the primary lung squamous cell carcinomas. More extensive and intense staining was seen in 3.7% of the primary lung adenocarcinomas and 12% of the primary lung squamous cell carcinomas. Expression, mostly focal, was also seen in 30% of the metastatic lung carcinomas. Finally, high expression was seen in 12.5% of the other tumors (one metastatic pancreatic carcinoma, one metastatic salivary gland adenocarcinoma not otherwise specified, one metastatic squamous cell carcinoma of the skin, one primary uterine cervix serous carcinoma, and one squamous cell carcinoma of the head and neck), and focal expression was present in another 22% of them. No ideal cut-off for positivity for GATA3 staining could be identified, as increasing the cut-off in either the extent or the intensity of staining increased specificity but decreased sensitivity. In conclusion, our study shows that although GATA3 staining is very helpful in everyday practice in determining the breast/urothelial origin of carcinomas, there are two caveats to its use: the first is that nonclassical histologies of urothelial carcinomas and TNBC may be negative for the marker, and secondly, carcinomas of various origins may show (although rarely) intense positivity.