Gene fusion characterisation of rare aggressive prostate cancer variants-adenosquamous carcinoma, pleomorphic giant-cell carcinoma, and sarcomatoid carcinoma: an analysis of 19 cases.

AIMS: To evaluate the molecular underpinnings of the rare aggressive prostate cancer variants adenosquamous carcinoma, pleomorphic giant-cell carcinoma, and sarcomatoid carcinoma. METHODS AND RESULTS: We retrieved 19 tumours with one or more variant(s), and performed ERG immunohistochemistry, a next-generation sequencing assay targeting recurrent gene fusions, and fluorescence in-situ hybridisation (FISH) for ERG and BRAF. Divergent differentiation included: sarcomatoid carcinoma (n = 10), adenosquamous carcinoma (n = 7), and pleomorphic giant-cell carcinoma (n = 7). Five patients had more than one variant. Four had variants only in metastases. ERG rearrangement was detected in nine (47%, seven via sequencing, showing TMPRSS2-ERG fusions and one GRHL2-ERG fusion, and two via FISH, showing rearrangement via deletion). ERG was immunohistochemically positive in the adenocarcinoma in eight of nine (89%) patients, but was immunohistochemically positive in the variant in only five of nine patients (56%, typically decreased). One patient had a false-positive ERG immunohistochemical result in the sarcomatoid component despite a negative FISH result. Two (11%) harboured BRAF fusions (FAM131A-BRAF and SND1-BRAF). CONCLUSIONS: ERG fusions are present in these rare prostate cancer variants with a frequency close to that in conventional prostate cancer (9/19, 47%). ERG immunohistochemistry usually detects rearrangement in the adenocarcinoma, but is less sensitive for the variant histology, with weak to negative staining. Adenosquamous and sarcomatoid variants can, particularly, occur together. Molecular assessment may be an additional tool in selected cases to confirm the prostatic origin of unusual tumours. The presence of two BRAF rearrangements suggests that this gene fusion may be enriched in this setting, as RAF kinase fusions have been previously reported in 1-2% of prostate cancers.

Sclerosing TSC1 mutated renal cell carcinoma: An unusual pattern mimicking MITF family translocation renal cell carcinoma.

The tuberous sclerosis genes and MTOR are increasingly being found to have important roles in novel subtypes of renal cancer, particularly emerging entities eosinophilic solid and cystic renal cell carcinoma (RCC) and high-grade oncocytic renal tumor (HOT)/RCC with eosinophilic and vacuolated cytoplasm. We report a unique renal neoplasm in a 66-year-old woman that initially mimicked MITF family translocation RCC due to mixed clear and eosinophilic cells, extensive stromal hyalinization, and psammoma bodies, yet which was negative for TFE3 and TFEB fluorescence in situ hybridization and a next generation sequencing (NGS) gene fusion assay. Cytoplasmic stippling triggered consideration of TSC-associated neoplasms, and a targeted NGS assay revealed a variant in exon 21 of TSC1 resulting in c.2626G>T p.(Glu876*) truncating mutation. This report adds to the morphologic spectrum of TSC-related renal neoplasms, including prominent stromal hyalinization as a potentially deceptive pattern. Due to the overlap in cytoplasmic stippling between eosinophilic solid and cystic RCC and HOT/RCC with eosinophilic and vacuolated cytoplasm, it is debatable which category this example would best fit. Further understanding of these entities and other renal neoplasms with alterations in the TSC genes will elucidate whether they should be considered a family of tumors.

Aurora A is differentially expressed in gliomas, is associated with patient survival in glioblastoma and is a potential chemotherapeutic target in gliomas.

Aurora A is critical for mitosis and is overexpressed in several neoplasms. Its overexpression transforms cultured cells, and both its overexpression and knockdown cause genomic instability. In transgenic mice, Aurora A haploinsufficiency, not overexpression, leads to increased malignant tumor formation. Aurora A thus appears to have both tumor-promoting and tumor-suppressor functions. Here, we report that Aurora A protein, measured by quantitative protein gel blotting, is differentially expressed in major glioma types in lineage-specific patterns. Aurora A protein levels in WHO grade II oligodendrogliomas (n=16) and grade III anaplastic oligodendrogliomas (n=16) are generally low, similar to control epilepsy cerebral tissue (n=11). In contrast, pilocytic astrocytomas (n=6) and ependymomas (n=12) express high Aurora A levels. Among grade II to grade III astrocytomas (n=7, n=14, respectively) and grade IV glioblastomas (n=31), Aurora A protein increases with increasing tumor grade. We also found that Aurora A expression is induced by hypoxia in cultured glioblastoma cells and is overexpressed in hypoxic regions of glioblastoma tumors. Retrospective Kaplan-Meier analysis revealed that both lower Aurora A protein measured by quantitative protein gel blot (n=31) and Aurora A mRNA levels measured by real-time quantitative RT-PCR (n=58) are significantly associated with poorer patient survival in glioblastoma. Furthermore, we report that the selective Aurora A inhibitor MLN8237 is potently cytotoxic to glioblastoma cells, and that MLN8237 cytotoxicty is potentiated by ionizing radiation. MLN8237 also appeared to induce senescence and differentiation of glioblastoma cells. Thus, in addition to being significantly associated with survival in glioblastoma, Aurora A is a potential new drug target for the treatment of glioblastoma and possibly other glial neoplasms.