To the Editor: We greatly appreciate the interest of Chou et al1 in our recent work reporting the coexistence of TFE3 gene rearrangement and alterations of SDHB in renal cell carcinoma.2 Indeed, as the authors note, it seems counterintuitive to find these two alterations in the same neoplasm, which might be hypothesized to be mutually exclusive, since both are presumed to be key driver events in tumorigenesis. However, with increasing application of molecular techniques,3, 4 it is now also apparent that some neoplasms exhibit overlapping and complex alterations.5 Coexistence of translocation with key driver mutations in the same tumor has been reported previously.6, 7, 8 The same alterations may produce different phenotypes in different tumor types or body sites, and there may be considerable clonal evolution and clonal heterogeneity within a single tumor or metastases.9, 10

It is not unprecedented to hypothesize that some alterations may occur as secondary events in tumors driven by other genetic alterations, including renal cell carcinoma. As examples relevant to the study discussed here, Papathomas et al11 encountered one renal cell carcinoma from a cohort of 348 unselected tumors (including 130 renal tumors) that exhibited abnormal negative SDHB immunohistochemical staining in the high-grade component of a clear cell renal cell carcinoma. Genetic studies demonstrated large intragenic SDHD and SDHAF2 deletions in only this high-grade and sarcomatoid component of this tumor, with absence of germline mutation, suggesting that in this case SDH subunit alterations occurred as a secondary event in a tumor likely representing a usual clear cell renal cell carcinoma.11 Conversely, another recent study suggested that some gene fusions with TFEB may be secondary, occurring in the setting of amplification, rather than as a main driver event,12 a phenomenon that has also been reported in other contexts.13 As another example, some of us have found that from the Cancer Genome Atlas database of clear cell renal cell carcinoma,14 two tumors were recently noted to be TFE3-SFPQ rearrangement-associated tumors8 also had VHL gene mutation, chromosome 3p deletion, and morphology indistinguishable from clear cell renal cell carcinoma,15 suggesting that there may be unexpected overlap in molecular alterations between tumor histologies, especially as various testing techniques with varying sensitivities are employed.

As noted by Chou et al1 the fraction of cells showing a TFE3 gene rearrangement pattern in our study was low,2 which is also somewhat counterintuitive, as one might predict an overwhelming majority of tumor cells to harbor a molecular alteration if it were the main driver event of tumorigenesis. However, a number of studies on TFE3 rearrangement-associated renal cell carcinoma have used similar cutoff thresholds,16, 17 with possible explanations for the low fraction of rearrangement signals including nuclear truncation due to histologic sectioning, inability to distinguish normal from neoplastic cells during evaluation, and tendency to underestimate the composition of normal cells in tumor tissues.3, 4, 16, 18, 19 Strong immunohistochemical labeling for TFE3 has been found to be a reliable biomarker of TFE3 translocation. In our study, all of the cases were also strongly positive by TFE3 immunohistochemical staining, which may support abnormal protein production, although of course fluorescence in situ hybridization (FISH) is generally considered superior for detecting true translocation in this context.2, 16, 20, 21

The particular cases reported in our study were identified because of the unusual morphologic features that, despite areas morphologically suggestive of SDH-deficient renal cell carcinoma, also raised the possibility of translocation-associated carcinoma and prompted immunohistochemical evaluation for TFE3 protein. This included a papillary architecture in three tumors with psammomatous calcifications in two. These features have been rarely described in series of SDH-deficient renal cell carcinoma in the literature.22, 23 One case included in the series from Gill et al22 had a predominant papillary architecture, whereas only ‘very focal abortive papillary architecture’ was noted in a few cases. The study by Williamson et al23 did not identify any neoplasms with papillary architecture, although most of the tumors in both studies were identified based on morphologic suspicion rather than unselected screening, which might introduce a bias toward detection of those with prototypical morphology. Although limited pathology data are provided, the study by Ricketts et al24 also noted some tumors in patients with known SDH subunit gene mutations to be usual clear cell renal cell carcinomas, suggesting that the morphology of SDH-deficient renal cell carcinomas may be more heterogeneous when patients are identified based on known gene mutations rather than tumor morphology.

In the study by Green et al,16 seven of the 31 tumors interpreted as TFE3 rearranged had a split signal FISH pattern making up less than 40% of cells (16–37%). This raises an intriguing question for future research, as to whether all of these tumors should be considered biologically equivalent, or whether there are differences between tumors with high and low percentages of rearranged cells. This question becomes increasingly relevant, as the spectrum of TFE3 rearranged renal cell carcinoma continues to expand to include a highly heterogeneous group of renal cell carcinomas with morphology beyond that which was initially described.16, 20, 25 Cutoff values for TFE3 translocation have also varied among studies from 7% in one study26 to 20% in another study.27 Chou et al indicate that in their laboratory a cutoff of ≥10% is utilized.

Overall, we agree with the interpretation of Chou et al that in the reported tumors, SDHB alterations are likely to represent the primary driver alteration; nonetheless, our finding may have relevance in the diagnostic setting and for understanding of intratumoral heterogeneity and clonal evolution. As such, when encountering a mutation, rearrangement, or other molecular alteration in an unusual or unexpected context, it may be warranted for pathologists and scientists to keep open consideration for other alterations, such as the scenario of diagnosing renal cell carcinoma in young patients posed by Chou et al. The clinical and biological significance of TFE3 translocation also remain to be further explored; an increasing number of non-renal tumors also harbor TFE3 translocation, including perivascular epithelioid cell neoplasms (PEComas) and rare ovarian tumors, in addition to the prototypical entity, alveolar soft part sarcoma.27, 28, 29, 30, 31, 32, 33, 34