Portal de Periódicos da CAPES

Low‐grade diffusely infiltrative tumour (LGDIT), SMARCB1‐mutant: A clinical and histopathological distinct entity showing epigenetic similarity with ATRT‐MYC

2022; Wiley; Volume: 48; Issue: 4 Linguagem: Inglês

10.1111/nan.12797

1365-2990

Martin Hasselblatt, Christian Thomas, Aniello Federico, Susanne Bens, Mats Hellström, Olivera Casar‐Borota, Uwe Kordes, Julia E. Neumann, Matthias Dottermusch, Fausto J. Rodríguez, Andrea Lo, Sylvia Cheng, Glenda Hendson, Juliette Hukin, Christian Hartmann, Arend Koch, David Capper, Reiner Siebert, Werner Paulus, Karolina Nemes, Pascal D. Johann, Michael C. Frühwald, Marcel Kool,

Tumors and Oncological Cases

Low-grade diffusely infiltrative tumour (LGDIT), SMARCB1-mutant, is a histopathological distinct low-grade lesion encountered in older children and young adults that shows epigenetic similarity with ATRT-MYC and has the potential for malignant progression. Atypical teratoid/rhabdoid tumour (AT/RT) is a highly malignant central nervous system tumour characterised by loss of SMARCB1/INI1 protein expression [1]. Most AT/RTs occur in infants, but older children and adults may also be affected. AT/RT comprises three molecular subgroups, ATRT-SHH, ATRT-TYR and ATRT-MYC [2]. The median age of ATRT-MYC patients is older [2], and many AT/RTs diagnosed in adolescents and adults represent ATRT-MYC [3]. In older patients, secondary AT/RTs developing from tumours with retained SMARCB1/INI1 staining such as pleomorphic xanthoastrocytoma [4] and ganglioglioma [5] are also on record. Recently, central nervous system low-grade diffusely infiltrative tumour with INI1 deficiency (CNS LGDIT-INI1) has been described as a low-grade lesion with diffuse growth of small inconspicuous SMARCB1/INI1-deficient tumour cells and a background of neurons and reactive astrocytes, which has the potential to progress to AT/RT [6]. DNA methylation profiling performed in one case yielded similarity with ATRT-MYC [6], but the relationship of CNS LGDIT-INI1 and ATRT-MYC has not been further examined. Here, we confirm that CNS LGDIT-INI1 represents a clinical and histopathological distinct entity and provide evidence for epigenetic similarity of CNS LGDIT-INI1 with ATRT-MYC. In line with recent developments of WHO nomenclature, we propose to change the name from CNS LGDIT-INI1 to LGDIT, SMARCB1-mutant. Six consult cases of LGDIT, SMARCB1-mutant (including one recently published case), were retrieved from the archives of the Institute of Neuropathology Münster (Table 1). The median age of the four males and two females was 16 years (range: 10–28 years). All tumours were of supratentorial location. Histopathologically, cases represented tumours of low to moderate cellularity and diffuse growth of inconspicuous small SMARCB1-deficient tumour cells and reactive pleomorphic neuronal cells and glial cells with retained SMARCB1-staining in the background as originally described by Nobusawa et al. [6] (Figure 1A,C). Tumour cells stained positive for vimentin (6/6 cases) and CD34 (3/6 cases) but were negative for GFAP, S100, actin, desmin, epithelial membrane antigen, cytokeratins (MNF116), neuronal antigens (synaptophysin, NeuN) and Olig2 (0/4), CD45 (0/3) and CD68 (0/3). Pleomorphic glial cells in the background stained positive for GFAP and S100. Mitoses were absent, and proliferative activity was low to moderate in most cases (median Ki67/MIB1 index 8%, Figure 1E). In addition, two cases also showed a high-grade component composed of densely packed rhabdoid tumour cells showing SMARCB1/INI1 loss as well as increased mitotic and proliferative activity (Figure 1B,D,F). Low grade cases had initially been diagnosed as diffuse astrocytoma, desmoplastic myxoid tumour, SMARCB1-mutant (DMT) or SMARCB1-deficient tumour, not elsewhere classified. In the two cases featuring a high-grade component, the possibility of AT/RT with a ganglioglioma component had been considered. Virtual slides of all cases of the series are available online (https://omero-imaging.uni-muenster.de/webclient/img_detail/42361/). Fluorescence in situ hybridization (FISH) performed in five cases revealed a homozygous deletion of the SMARCB1 locus. None of the six cases showed a BRAF-V600E mutation using Sanger sequencing. After DNA isolation from formalin-fixed paraffin-embedded tumour samples, purification and bisulfite conversion using standard protocols, all six cases were subjected to DNA methylation profiling using MethylationEPIC BeadChip array (Illumina Inc., San Diego, CA). Copy-number variation analysis from methylation array data confirmed the presence of a homozygous SMARCB1 deletion in all six cases. Apart from a second more distal 22q deletion present in three cases, various chromosomal losses, but no recurrent chromosomal alterations, were noted (Figure S1). Using DNA methylation-based classification and the Heidelberg Brain Tumour Classifier (version v11b4) [8], all tumours were classified as ATRT-MYC (median calibrated score: 0.97). Next, t-distributed stochastic neighbour embedding (t-SNE) analysis was performed in comparison with DNA methylation profiles of 112 AT/RT samples (29 ATRT-MYC, 37 ATRT-TYR and 46 ATRT-SHH), 58 malignant rhabdoid tumours of the kidney, 8 poorly differentiated chordomas, SMARCB1-mutant, and 6 desmoplastic myxoid tumours, SMARCB1-mutant [9]. DNA methylation profiles of the six LGDIT, SMARCB1-mutant, grouped together in proximity to ATRT-MYC (Figure 1g). Furthermore, unsupervised hierarchical clustering results confirmed epigenetic similarity of LGDIT, SMARCB1-mutant, and ATRT-MYC (Figure S2). Follow-up information was available for four cases (including the two cases with a high-grade component). Patients responded well to therapy and experienced stable disease or complete remission after an observation time of three to 56 months (Table 1). Our findings confirm that LGDIT, SMARCB1-mutant, represents a clinical and histological distinct entity. Of note, LGDIT, SMARCB1-mutant, lacks the polyphenotypic expression profile encountered in AT/RT. Inconspicuous tumour cells and lack of specific diagnostic markers pose a diagnostic challenge and highlight the value of DNA methylation profiling in this setting. From our experience, the possibility of LGDIT, SMARCB1-mutant, should be considered in any low-grade lesion with epigenetic similarity with AT/RT. This holds especially true for supratentorial tumours in older children and young adults, an age group where AT/RTs are only rarely encountered. The exact relationship between LGDIT, SMARCB1-mutant, and AT/RT remains to be determined. The view that LGDIT, SMARCB1-mutant, may represent an early stage of ATRT-MYC could be supported by the epigenetic similarity with ATRT-MYC, but also by the fact that two LGDIT, SMARCB1-mutant, featured a high-grade AT/RT component. Furthermore, malignant progression to AT/RT has been described in two of three cases of the original series [6] and in another tumour with histological features of LGDIT, SMARCB1-mutant, as long as 10 years after the initial diagnosis [10]. Taken into account the relatively short follow-up of the present series, it cannot be excluded that eventually all LGDIT, SMARCB1-mutant, may progress to AT/RT. Therefore, careful follow-up examinations are warranted, and the possibility of early therapeutic intervention needs to be considered. On the other hand, because DNA methylation profiles of LGDIT, SMARCB1-mutant, grouped closely together and the number of cases of the present series was relatively small, it cannot be excluded that epigenetic distinction of LGDIT, SMARCB1-mutant, and ATRT-MYC may be possible once a larger number of cases is available for examination. The possibility that LGDIT, SMARCB1-mutant, and ATRT-MYC may originate from different cells of origin needs also to be considered, because epigenetic similarity of LGDIT, SMARCB1-mutant, and ATRT-MYC could rather represent a consequence of the well documented effects [11] of SMARCB1/INI1 loss on the epigenome. The older median age of patients harbouring LGDIT, SMARCB1-mutant, as compared to AT/RT could point toward a different vulnerable time period for tumourigenesis. Other potential mechanisms involved in the biology of LGDIT, SMARCB1-mutant, may include decreased susceptibility of tumour cells to SMARCB1/INI1 loss and possibly even interactions with pleomorphic neighbouring cells showing retained SMARCB1/INI1 staining. Pleomorphic glial cells in the background stained positive for GFAP and S100 and showed retained (normal) SMARCB1/INI1 staining. Most likely, they represent reactive non-neoplastic cells. We have also considered the possibility of a pre-existing low-grade glioma, but from our experience in AT/RT with molecular features of pleomorphic xanthoastrocytoma [4], such tumours retain the epigenetic and genetic signature of the primary lesion, which seems not to be the case in LGDIT, SMARCB1-mutant. The cell of origin of LGDIT, SMARCB1-mutant (and ultimately ATRT-MYC), remains to be determined. Interestingly, in malignant rhabdoid tumours of the kidney [12], which show epigenetic similarity with ATRT-MYC [13], neighbouring Schwann cells have been shown to partly share genetic alterations of the malignant rhabdoid tumour, including homozygous deletions of the SMARCB1 locus [12], suggesting that SMARCB1/INI1 loss may not be sufficient to drive the development of malignant rhabdoid tumours [12]. In conclusion, this series confirms that LGDIT, SMARCB1-mutant, is a clinically and histologically distinct entity. LGDIT, SMARCB1-mutant, shows epigenetic similarity with ATRT-MYC and has the potential for malignant progression. The study was further supported by IZKF Münster (Ha3/017/20) and "The Everest Centre for Research into Paediatric Low Grade Brain Tumours" (The Brain Tumour Charity, UK). The authors have no competing interest to declare. MH and CT designed the study. All authors were involved in the acquisition, analysis or interpretation of data for the work. MH drafted the manuscript, and all other co-authors revised it and approved the final version. Cases were retrieved from the archives of the Institute of Neuropathology Münster. The tumour bank of the Institute of Neuropathology Münster has received continuous local ethics committee approval (Ethics committee of the University Hospital Münster). Patients or the guardians gave informed consent for the scientific use of archival materials. Patient level data included are consistent with governance approvals/consents. The peer review history for this article is available at https://publons.com/publon/10.1111/nan.12797. The data that support the findings of this study are available from the corresponding author upon reasonable request. Supplemental Figure 1: Copy-number variation analysis from methylation array data. A homozygous deletion of the SMARCB1 locus on 22q is present in all cases. Apart from a second more distal 22q deletion encountered in three cases (arrows), further recurrent chromosomal alterations are absent. Note that low grade and high-grade component of case #5 show comparable copy number alterations. Supplemental Figure 2: Clustering analysis. Unsupervised hierarchical clustering analysis confirms the epigenetic similarity of LGDIT, SMARCB1-mutant, within ATRT-MYC. Unsupervised hierarchical clustering was performed on the pairwise Pearson correlation coefficient matrix using the 927 most variable CpGs (standard deviation > 0.3) across the whole dataset. Samples were clustered using the Euclidean distance as the distance measure and single linkage as the clustering method. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.