Drivers underpinning the malignant transformation of giant cell tumour of bone

Abstract

The rare benign giant cell tumour of bone (GCTB) is defined by an almost unique mutation in the H3.3 family of histone genes H3-3A or H3-3B; however, the same mutation is occasionally found in primary malignant bone tumours which share many features with the benign variant. Moreover, lung metastases can occur despite the absence of malignant histological features in either the primary or metastatic lesions. Herein we investigated the genetic events of 17 GCTBs including benign and malignant variants and the methylation profiles of 122 bone tumour samples including GCTBs. Benign GCTBs possessed few somatic alterations and no other known drivers besides the H3.3 mutation, whereas all malignant tumours harboured at least one additional driver mutation and exhibited genomic features resembling osteosarcomas, including high mutational burden, additional driver event(s), and a high degree of aneuploidy. The H3.3 mutation was found to predate the development of aneuploidy. In contrast to osteosarcomas, malignant H3.3-mutated tumours were enriched for a variety of alterations involving TERT, other than amplification, suggesting telomere dysfunction in the transformation of benign to malignant GCTB. DNA sequencing of the benign metastasising GCTB revealed no additional driver alterations; polyclonal seeding in the lung was identified, implying that the metastatic lesions represent an embolic event. Unsupervised clustering of DNA methylation profiles revealed that malignant H3.3-mutated tumours are distinct from their benign counterpart, and other bone tumours. Differential methylation analysis identified CCND1, encoding cyclin D1, as a plausible cancer driver gene in these tumours because hypermethylation of the CCND1 promoter was specific for GCTBs. We report here the genomic and methylation patterns underlying the rare clinical phenomena of benign metastasising and malignant transformation of GCTB and show how the combination of genomic and epigenomic findings could potentially distinguish benign from malignant GCTBs, thereby predicting aggressive behaviour in challenging diagnostic cases.

DNA methylation‐based profiling of bone and soft tissue tumours: a validation study of the ‘DKFZ Sarcoma Classifier’

“…Our work and that of others on DNA methylation profiling of sarcoma has demonstrated key insights into specific tumour types [6][7][8][9][10][11], and has shown that DNA methylation can add value to whole-genome and RNA sequencing data. To enhance the benefit of the genomes delivered from 1,200 patients with sarcoma as part of the UK's 100,000 Genomes Project [12], we have undertaken methylation profiling of a significant proportion of these patients' samples, with the aim of providing greater insight into the pathogenesis of sarcoma and its mimics.…”

“…DNA methylation profiling data were available from 986 samples; many of the samples were included in other studies over the last 5 years including the 100,000 Genomes Project [6][7][8][9]13]. For details of generation of these data, see Supplementary materials and methods.…”

“…By asking rigorous biological and clinically relevant questions, the classification of disease has the potential to provide novel prognostic and predictive biomarkers as well as identify therapeutic targets. Studies of the methylation profiles of GCTBs, dedifferentiated liposarcoma, MPNST, and CHA serve as examples [6][7][8]11]. In the DKFZ Classifier, benign and malignant GCTBs cluster together; however, directed analysis of methylation data of these subtypes not only allowed them to be separated, but also implicated CCND1 as a likely cancer driver gene in the malignant transformation of GCTB [7].…”

“…Studies of the methylation profiles of GCTBs, dedifferentiated liposarcoma, MPNST, and CHA serve as examples [6][7][8]11]. In the DKFZ Classifier, benign and malignant GCTBs cluster together; however, directed analysis of methylation data of these subtypes not only allowed them to be separated, but also implicated CCND1 as a likely cancer driver gene in the malignant transformation of GCTB [7]. Second, The Cancer Genome Atlas analysis of soft tissue sarcomas demonstrated that dedifferentiated liposarcoma, conventionally classified by the presence of MDM2 amplification, formed two methylation groups which corresponded to significantly different clinical outcomes [11].…”

Cell Biology of Giant Cell Tumour of Bone: Crosstalk between m/wt Nucleosome H3.3, Telomeres and Osteoclastogenesis

“…Next to histological examination, interpretation of the radiological features may be of major help in distinguishing both entities. Molecularly, a recent publication reported that maligant bone tumours with a H3.3-G34W mutation possess features resembling osteosarcoma [105]. More specifically, they carry an increased burden of somatic mutation variants.…”