Search

OBJECTIVE Chordomas are rare bone neoplasms characterized by a high recurrence rate and no benefit from any approved medical treatment to date. However, the investigation of molecular alterations in chordomas could be essential to prognosticate, guide clinical decision-making, and identify theranostic biomarkers. The aim of this study was to provide a detailed genomic landscape of a homogeneous series of 64 chordoma samples, revealing driver events, theranostic markers, and outcome-related genomic features. METHODS The authors conducted whole-exome sequencing (WES), targeted next-generation sequencing, and RNA sequencing of 64 skull base and spinal chordoma samples collected between December 2006 and September 2020. Clinical, histological, and radiological data were retrospectively analyzed and correlated to genetic findings. RESULTS The authors identified homozygous deletions of CDKN2A/2B, PIK3CA mutations, and alterations affecting genes of SWI/SNF chromatin remodeling complexes (PBRM1 and ARID1A) as potential theranostic biomarkers. Using matched germline WES, they observed a higher frequency of a common genetic variant (rs2305089; p.(Gly177Asp)) in TBXT (97.8%, p < 0.001) compared to its distribution in the general population. PIK3CA mutation was identified as an independent biomarker of short progression-free survival (HR 10.68, p = 0.0008). Loss of CDKN2A/2B was more frequently observed in spinal tumors and recurrent tumors. CONCLUSIONS In the current study, the authors identified driver events such as PBRM1 and PIK3CA mutations, TBXT alterations, or homozygous deletions of CDKN2A/2B, which could, for some, be considered potential theranostic markers and could allow for identifying novel therapeutic approaches. With the aim of a future biomolecular prognostication classification, alterations affecting PIK3CA and CDKN2A/2B could be considered as poor prognostic biomarkers.

Purpose Long-term use of cyproterone acetate (CPA) is associated with an increased risk of developing an intracranial meningioma. Discontinuation of CPA most often induce stabilization or regression of the tumor. The exact mechanism of regression is unknown as well as the reason why some meningiomas are still growing after CPA discontinuation.We are reporting four patients with multiple meningiomas, showing opposite tumor evolutions after stopping the CPA highlighting the underlying histologic and genetic features.MethodsPatients presenting several meningiomaswith opposite evolutions following the discontinuation of CPA were identified. The clinical and radiological data’s were reviewed. A retrospective volumetric analysis of the meningiomas was performed. All the growing meningiomas were operated. Each tumor was characterized histopathologically and by molecular and genetic analyses.ResultsFour female withmultiple meningiomas and opposite tumor volume evolution after CPA discontinuation were identified. The histopathological results found fibroblastic meningiomas for tumorsgrowinglocated in the convexity and a morefibrousmeningioma in the skull-base tumor which decreased. Meningothelial and transitional meningiomaswere found in two skull-base growing meningiomas.The molecular characterization found twoNF2-mutations among the 4 growing meningiomas. In one patient who presented both patterns, the shrinking skull-basetumor harbored a PIK3CA-mutation whereas the growing tumor, a NF2-mutation.ConclusionTo our knowledge, this is the first report of such an atypical tumor evolution of CPA-associated meningiomatosis after CPA discontinuation in the same patient. Underlying biological mechanisms explaining this observationespecially, the close relationship between mutational landscapes and the meningeal embryologyin CPA-related meningiomasrequire further research.

Medulloblastoma (MB), the most frequent embryonic tumor of the cerebellum is classified into four molecular subgroups (WNT group, SHH group, group 3 and group 4). Although the vast majority of MB are sporadic, predisposing genetic diseases have been described in rare WNT MB and more frequently in the SHH group. In a recent pediatric series of SHH-MB, germline alterations of the ELP1 gene have been described in 14% of cases, making this gene the most frequent genetic predisposition in MB. We have investigated the potential interest of ELP1 immunostaining on a large cohort of 132 MB. A complete loss of ELP1 staining was observed in 12 SHH MB (among 57 total SHH MB: 21%). The loss of ELP1 immunostaining was well correlated with the presence of a bi-allelic alteration of the gene except for one case for which the MB had a loss of ELP1 protein expression demonstrated by immunohistochemistry (IHC) and confirmed by whole proteome analysis, although no obvious genetic alteration in the coding sequence of ELP1 could be found. Molecular analysis of a large “molecular” cohort of 266 MB from French centers for which somatic ELP1 was sequenced allows to identify 12 additional MB with bi-allelic ELP1 genetic alterations. Our results demonstrate the benefit of the ELP1 IHC as an accurate and reliable tool to screen ELP1-deficient MB. This new immunohistochemical tool will now be advantageously used to screen SHH MB upfront for genetic alteration in ELP1, and will subsequently help orientating these patients towards genetic counseling.

The fast advances in molecular biology have led to the discovery of new genetic and epigenetic driver events that play a key role in the pathogenesis of central nervous system (CNS) tumors. These data showed the importance of molecular biology, and in particular methylation profiling and fusion detection, to complete the histological examination. Exploring all these complex genomic and epigenomic alterations with the technologies implemented in the routine diagnosis requires massive laboratory equipment. In addition, it could take weeks before getting data that could be important to adapt the diagnosis, prognosis, and to ultimately guide the treatment’s choice. We describe the innovative potential of a nanopore sequencing device to obtain fast and multiple information to characterize CNS tumors. Adaptive-sampling is a software-controlled enrichment unique method that allows simultaneous methylation profiling, copy number landscape assessment, and fusion gene detection. Thereby, this technology brings comprehensive molecular information that could be very important for CNS tumor classification in a very short delay (below one week) and with a relative low cost compared to the combined other molecular analysis needed to get the same information. We report ten cases of diverse CNS tumors with already known oncogenic structural variants such as YAP1-MAMLD1 fused ependymoma, CNS neuroblastoma with FOXR2 activation, CNS high-grade neuroepithelial tumor with BCOR and MEN1 alteration, and atypical teratoid rhabdoid tumors (ATRT). For these tumors, using adaptive sampling with the nanopore sequencing, we were able to detect fusion genes as driver events, while obtaining at the same time the methylation profile to corroborate the tumor classification. Our results with this combined approach, never described before, demonstrate its feasibility as a new extremely powerful tool and show remarkable perspectives for this technology to bring a fast and comprehensive tumor molecular characterization.

The long-term use of cyproterone acetate (CPA) is associated with an increased risk of developing intracranial meningiomas. CPA discontinuation most often induces a stabilization or regression of the tumor. The underlying biological mechanisms as well as the reasons why some meningiomas still grow after CPA discontinuation remain unknown. We reported a series of patients presenting CPA-induced meningiomatosis with opposed tumor evolutions following CPA discontinuation, highlighting the underlying histological and genetic features. Patients presenting several meningiomas with opposite tumor evolution (coexistence of growing and shrinking tumors) following CPA discontinuation were identified. Clinical and radiological data were reviewed. A retrospective volumetric analysis of the meningiomas was performed. All the growing meningiomas were operated. Each operated tumor was characterized by histological and genetic analyses. Four women with multiple meningiomas and opposite tumor volume evolutions after CPA discontinuation were identified. Histopathological analysis characterized the convexity and tentorial tumors which continued to grow after CPA discontinuation as fibroblastic meningiomas. The decreasing skull base tumor was characterized as a fibroblastic meningioma with increased fibrosis and a widespread collagen formation. The two growing skull base meningiomas were identified as meningothelial and transitional meningiomas. The molecular characterization found two NF2 mutations among the growing meningiomas and a PIK3CA mutation in the skull base tumor which decreased. To our knowledge, this is the first report describing an atypical tumor evolution of CPA-associated meningiomas after CPA discontinuation. The underlying biological mechanisms explaining this observation and especially the close relationship between mutational landscapes and embryologic origins of the meninges in CPA-related meningiomas as well as their clonal origin require further research.

International audience

Medulloblastoma (MB) can be classified into four molecular subgroups (WNT group, SHH group, group 3, and group 4). The gold standard of assignment of molecular subgroup through DNA methylation profiling uses Illumina EPIC array. However, this tool has some limitation in terms of cost and timing, in order to get the results soon enough for clinical use. We present an alternative DNA methylation assay based on nanopore sequencing efficient for rapid, cheaper, and reliable subgrouping of clinical MB samples. Low-depth whole genome with long-read single-molecule nanopore sequencing was used to simultaneously assess copy number profile and MB subgrouping based on DNA methylation. The DNA methylation data generated by Nanopore sequencing were compared to a publicly available reference cohort comprising over 2,800 brain tumors including the four subgroups of MB (Capper et al. Nature; 2018) to generate a score that estimates a confidence with a tumor group assignment. Among the 24 MB analyzed with nanopore sequencing (six WNT, nine SHH, five group 3, and four group 4), all of them were classified in the appropriate subgroup established by expression-based Nanostring subgrouping. In addition to the subgrouping, we also examine the genomic profile. Furthermore, all previously identified clinically relevant genomic rearrangements (mostly MYC and MYCN amplifications) were also detected with our assay. In conclusion, we are confirming the full reliability of nanopore sequencing as a novel rapid and cheap assay for methylation-based MB subgrouping. We now plan to implement this technology to other embryonal tumors of the central nervous system.

The cell(s) of origin of AT/RTs remain(s) unknown. We previously developed a mouse model consisting of tamoxifen inducible system in a Smarcb1Flox/Flox;Rosa26-CreERT2 background. We obtained two molecular subgroups of intracranial tumors, one with neuronal and the other with non-neuronal features, consistent with the diversity observed in human AT/RTs. To investigate the potential cell(s) of origin of these various AT/RTs, we first explored whether different time points of Smarcb1 inactivation correlated with anatomic location and/or molecular subgroups. We observed that the neuronal group, primarily developing from the subventricular zone and the spinal cord, was almost exclusively obtained with the earliest inactivation time points (E6-E7). In contrast, the non-neuronal group emerged upon Smarcb1 inactivation at any time point (E6-E10) and showed intracranial but extra-parenchymal/meningeal origins. High-resolution analysis of anatomic distribution of 55 human AT/RT and molecular subgroups is in progress. In order to more specifically identify the cell(s) of origin for the neuronal group, we next generated developmental stage-specific conditional knockout mice carrying Smarcb1 inactivation by restricting Cre expression with promoters characteristic for various neural stem cells/progenitors. While Smarcb1Flox/Flox;Atoh1CreERT2 showed ataxia but failed to give rise to any tumor at any embryonal time point, Smarcb1Flox/Flox;Ascl1CreERT2 did not show any phenotype. Targeting Nestin-expressing cells led to tumors with morphologic rhabdoid features; these tumors again showed molecular diversity as observed in human AT/RTs. In conclusion, we show that deletion of Smarcb1 determines mouse AT/RT subtypes depending on spatial and temporal factors. Our new mouse models not only give insight into the cell(s) of origin, but also provide interesting preclinical models of AT/RTs. Citation Format: Zhi-Yan Han, Mamy Andrianteranagna, Maria Jesus-Lobon-Iglesias, Arnault Tauziède-Espariat, Kevin Beccaria, Paul Fréneaux, Joshua J. Waterfall, Julien Masliah-Planchon, Christine Bourneix, Gaelle Pierron, Amaury Leruste, Céline Chauvin, Didier Surdez, Pascale Varlet, Christelle Dufour, Olivier Delattre, Volodia Dangouloff-Ros, Franck Bourdeaut. Spatial and temporal conditions for Smarcb1 deletion determines mouse AT/RT (atypical teratoid/rhabdoid tumor) subtype [abstract]. In: Proceedings of the AACR Special Conference on the Advances in Pediatric Cancer Research; 2019 Sep 17-20; Montreal, QC, Canada. Philadelphia (PA): AACR; Cancer Res 2020;80(14 Suppl):Abstract nr B68.

The cell(s) of origin of AT/RTs remains unclear. We previously developed a mouse model consisting in tamoxifen inducible system in a Smarcb1(Flox/Flox);Rosa26-Cre(ERT2) background. We obtained two subgroups of intracranial tumors, one with neuronal and the other with non-neuronal features, in keeping with the diversity observed in human AT/RTs. To investigate the potential cell(s) of origin of those various AT/RTs, we first endeavored to specify whether different time points of Smarcb1 inactivation correlated with specific anatomic location and/or molecular subgroups. Thereby, we identified that the neuronal group, mostly developing from the sub-ventricular zone and the spinal cord, was almost exclusively obtained with the earliest inactivation time point(E6-E7). In contrary, the non-neuronal group emerged after Smarcb1 inactivation at any time point(E6-E10), and showed intracranial but extra-parenchyma/meningeal origins. In order to more specifically identify the cell(s) of origin for the neuronal group, we next generated developmental stage-specific conditional knockout mice carrying Smarcb1 inactivation by restricting Cre expression upon promoters characteristic for various neural stem cells/progenitors. While Smarcb1(Flox/Flox);Atoh1Cre(ERT2) showed ataxia but failed to give rise to any tumor whatever the embryonal time point, targeting Nestin-expressing cells led to tumors with morphological rhabdoid features; these again showed some molecular diversity as observed in human AT/RTs. The study of the impact of Smarcb1 inactivation in other promoter-restricted backgrounds is ongoing. In conclusion, we show that spatio-temporal deletion of Smarcb1 determines mouse AT/RT subtypes. Our new mouse models not only give insight on the cell(s) of origin, but also provide interesting preclinical models.