TRIM28 inactivation in epithelial nephroblastoma is frequent and often associated with predisposing TRIM28 germline variants

Abstract

Wilms tumors (WTs) are histologically diverse childhood cancers with variable contributions of blastema, stroma, and epithelia. A variety of cancer genes operate in WTs, including the tripartite-motif-containing-28 gene (TRIM28). Case reports and small case series suggest that TRIM28 mutations are associated with epithelial morphology and WT predisposition. Here, we systematically investigated the prevalence of TRIM28 inactivation and predisposing mutations in a cohort of 126 WTs with >2/3 epithelial cells, spanning 20 years of biobanking in the German SIOP93-01/GPOH and SIOP2001/GPOH studies. Overall, 44.4% (56/126) cases exhibited loss of TRIM28 by immunohistochemical staining. Of these, 48 could be further analyzed molecularly, revealing TRIM28 sequence variants in each case – either homozygous (~2/3) or heterozygous with epigenetic silencing of the second allele (~1/3). The majority (80%) of the mutations resulted in premature stops and frameshifts. In addition, we detected missense mutations and small deletions predicted to destabilize the protein through interference with folding of key structural elements such as the zinc-binding clusters of the RING, B-box-2, and PHD domains or the central coiled-coil region. TRIM28-mutant tumors otherwise lacked WT-typical IGF2 alterations or driver events, except for rare TP53 progression events that occurred with expected frequency. Expression profiling identified TRIM28-mutant tumors as a homogeneous subset of epithelial WTs that mostly present with stage I disease. There was a high prevalence of perilobar nephrogenic rests, putative precursor lesions, that carried the same biallelic TRIM28 alterations in 7/7 cases tested. Importantly, 46% of the TRIM28 mutations were present in blood cells or normal kidney tissue, suggesting germline events or somatic mosaicism, partly supported by family history. Given the high prevalence of predisposing variants in TRIM28-driven WT, we suggest that immunohistochemical testing of TRIM28 be integrated into diagnostic practice as the management of WT in predisposed children differs from that with sporadic tumors. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.