Aberrant nuclear immunoreactivity for TFE3 in neoplasms with TFE3 gene fusions: a sensitive and specific immunohistochemical assay

We report the aberrantly strong nuclear immunoreactivity for the C-terminal portion of TFE3 protein in tumors characterized by chromosome translocations involving the TFE3 gene at Xp11.2. This group of tumors includes alveolar soft part sarcoma and a specific subset of renal carcinomas that tend to affect young patients. They contain fusion genes that encode chimeric proteins consisting of the N-terminal portion of different translocation partners fused to the C-terminal portion of TFE3. We postulated that expression of these fusion proteins may be dysregulated in these specific tumors and detectable by immunohistochemistry. We performed immunohistochemistry using a polyclonal antibody to the C-terminal portion of TFE3 in 40 formalin-fixed, paraffin-embedded tumors characterized by TFE3 gene fusions, including 19 alveolar soft part sarcoma (of which nine were molecularly confirmed) and 21 renal carcinomas with cytogenetically confirmed characteristic Xp11.2 translocations and/or fusion transcripts involving TFE3 (11 PRCC-TFE3, 7 ASPL-TFE3, 3 PSF-TFE3). We also screened 1476 other tumors of 64 histologic types from 16 sites for TFE3 immunoreactivity using tissue microarrays and evaluated a broad range of normal tissues. Thirty-nine of 40 neoplasms characterized by TFE3 gene fusions (19 of 19 alveolar soft part sarcoma, 20 of 21 renal carcinomas) demonstrated moderate or strong nuclear TFE3 immunoreactivity. In contrast, only 6 of 1476 other neoplasms labeled for TFE3 (sensitivity 97.5%, specificity 99.6%). Nuclear immunoreactivity in normal tissues was extremely rare. We then applied this assay to a set of 11 pediatric renal carcinomas for which only paraffin-embedded tissue was available, to assess if morphologic features could predict TFE3 immunoreactivity. Of the eight cases in which we suspected that a TFE3 gene rearrangement might be present based on morphology, seven scored positive for nuclear TFE3 labeling. Of the three tumors whose morphology did not suggest the presence of a TFE3 gene fusion, none showed nuclear TFE3 labeling. In summary, we find that nuclear immunoreactivity for TFE3 protein by routine immunohistochemistry is a highly sensitive and specific assay for neoplasms bearing TFE3 gene fusions. Furthermore, the finding in our set of test cases (i.e., that morphologic features can be used to predict TFE3 immunoreactivity) further supports the notion that renal carcinomas with TFE3 gene fusions have a distinctive morphology that corresponds to their genetic distinctiveness. Carcinomas associated with TFE3 gene fusions may account for a significant proportion of pediatric renal carcinomas, and this immunohistochemistry assay may help to clarify their true prevalence.