Genomic characterization of DICER1-associated neoplasms uncovers molecular classes

Article https://doi.org/10.1038/s41467-023-37092-w Genomic characterization of DICER1associated neoplasms uncovers molecular classes Received: 6 October 2022 1234567890():,; 1234567890():,; Accepted: 28 February 2023 Check for updates Felix K. F. Kommoss 1, Anne-Sophie Chong 2,3,4, Anne-Laure Chong2,3,5, Elke Pfaff6,7,8, David T. W. Jones 6,7, Laura S. Hiemcke-Jiwa9,10, Lennart A. Kester 10, Uta Flucke10,11, Manfred Gessler 12, Daniel Schrimpf13,14, Felix Sahm 13,14, Blaise A. Clarke15, Colin J. R. Stewart16,17, Yemin Wang 18,19, C. Blake Gilks18, Friedrich Kommoss20, David G. Huntsman18,19, Ulrich Schüller 21,22,23, Christian Koelsche 1, W. Glenn McCluggage24, Andreas von Deimling 13,14,25 & William D. Foulkes 2,3,5,25 DICER1 syndrome is a tumor predisposition syndrome that is associated with up to 30 different neoplastic lesions, usually affecting children and adolescents. Here we identify a group of mesenchymal tumors which is highly associated with DICER1 syndrome, and molecularly distinct from other DICER1-associated tumors. This group of DICER1-associated mesenchymal tumors encompasses multiple well-established clinicopathological tumor entities and can be further divided into three clinically meaningful classes designated “low-grade mesenchymal tumor with DICER1 alteration” (LGMT DICER1), “sarcoma with DICER1 alteration” (SARC DICER1), and primary intracranial sarcoma with DICER1 alteration (PIS DICER1). Our study not only provides a combined approach to classify DICER1-associated neoplasms for improved clinical management but also suggests a role for global hypomethylation and other recurrent molecular events in sarcomatous differentiation in mesenchymal tumors with DICER1 alteration. Our results will facilitate future investigations into prognostication and therapeutic approaches for affected patients. DICER1 is a cytoplasmic endoribonuclease that is critical for the correct processing (cleavage) of precursor micro-RNA (pre-miRNA) double-stranded hairpins with 3’ and 5’ ends to their mature singlestranded forms1,2. DICER1 utilizes its RNAse IIIa and IIIb domains to process pre-miRNAs, yielding a duplex containing either a mature 5p or 3p miRNA as well as a complementary ‘passenger strand’ that is ultimately degraded. The mature miRNA is then loaded onto an AGO protein to form an RNA-induced silencing complex, eventually resulting in down-regulation or silencing of mRNA targets. Specific metal-ion binding amino acids within the IIIa and IIIb domains are crucial for pre-miRNA cleavage. Failure of dicing of the pre-miRNA by A full list of affiliations appears at the end of the paper. Nature Communications | (2023)14:1677 RNase IIIb is a critical event in most DICER1-associated tumors. This failure arises because of the occurrence of tumor-confined missense variants that result in amino acid substitutions at these critical premiRNA-interacting residues within RNase IIIb3,4. DICER1 syndrome is a tumor susceptibility syndrome, characterized in 2009, determined by the occurrence of a germline pathogenic variant (PV) in DICER15. Typically, this is a loss of function (LOF) variant that is predicted to result in inactivation of the affected DICER1 allele. For the syndrome to occur, a second hit affecting exons encoding the RNase IIIb domain of DICER1 (a “hotspot” PV), as discussed above, is usually required3,4. The phenotypic spectrum of DICER1 syndrome is e-mail: 1 Article wide but some of the manifestations almost exclusively occur in persons with DICER1 syndrome. Both benign and malignant neoplasms are part of the syndrome. Pleuropulmonary blastoma (PPB), the most frequent primary lung malignancy in children, is highly characteristic of the syndrome. Other manifestations include ovarian Sertoli-Leydig cell tumor (SLCT), pediatric cystic nephroma, thyroid adenoma and carcinoma, embryonal rhabdomyosarcoma (ERMS) (particularly of the gynecological tract) and other rare entities6–9. Sarcomas are amongst the most common neoplasms in this syndrome and DICER1-associated sarcomas exhibit several characteristic morphological features, which can aid the pathologist in suspecting an association with DICER1 PVs, irrespective of the site of origin. These features comprise a subepithelial layer of malignant mesenchymal cells (cambium layer), areas of rhabdomyoblastic differentiation with positive staining for myogenin and myoD1, cellular/immature and occasionally malignant cartilage, foci of bone/osteoid and areas of anaplasia10–13. Furthermore, we have recently shown that both ERMS with DICER1 PVs and a tumor entity termed “primary intracranial sarcoma, DICER1-mutant” (PIS DICER1) are associated with DNA methylation signatures that are distinct from their morphological counterparts that are not DICER1-associated14–16. This finding, together with our recent speculations regarding the histomorphological similarities between DICER1-associated sarcomas11–13 arising at different sites led us to question whether in general, DICER1-associated tumors share common features such that they represent a distinct tumor entity, arising at various anatomical locations. Here, we address this hypothesis by analyzing 534 tumors, including a large number with DICER1 PVs, by DNA methylation profiling and identify three classes of mesenchymal tumors with DICER1 alteration, comprising tumors from various anatomical locations. Results DNA methylation profiling of DICER1-associated neoplasms We analyzed whole genome DNA methylation data of 534 tumors including various histotypes associated with the DICER1 syndrome, as well as reference entities representing morphological counterparts of the tumors studied. The study set included a total of 431 tumors with known DICER1 mutational status (431/534, 81%) of which 176 were reported to harbor DICER1 alterations (176/431, 41% of tumors analyzed for DICER1 variants). Detailed information on the tumor histotypes studied and DICER1 PVs of the study cohort is provided in Supplementary Table 1 and Supplementary Data 1. Unsupervised hierarchical clustering and t-SNE dimensionality reduction of DNA methylation data segregated tumors into distinct and stable clusters (Fig. 1a–c and Supplementary Fig. S1). Wilms tumor (WILMS), MYOD1-mutant spindle cell and sclerosing rhabdomyosarcoma (SRMS), ERMS, alveolar rhabdomyosarcoma (ARMS), lowgrade endometrial stromal sarcoma (LGESS), high-grade endometrial stromal sarcoma (HGESS), Müllerian adenosarcoma (MAS), embryonal tumor with multilayered rosettes (ETMR), lung adenocarcinoma (LAC), clear cell renal cell carcinoma (RCC), ciliary body medulloepithelioma (MEPL), multinodular goiter (MG) and papillary thyroid carcinoma (PCA) each formed a distinct molecular cluster defined by diagnoses based on histology and established molecular testing, irrespective of DICER1 alteration status. For SLCT, we identified a subcluster that correlated with DICER1 PV status, which we termed “SLCT with DICER1 alteration” (SLCT DICER1). F (...truncated)