Methylation-based reclassification and risk stratification of skull-base chordomas

Abstract

BackgroundSkull-base chordomas are rare malignant bone cancers originating from the remnant of the notochord. Survival is variable, and clinical or molecular factors cannot reliably predict their outcomes. This study therefore identified epigenetic subtypes that defined new chordoma epigenetic profiles and their corresponding characteristics.MethodsMethylation profiles of 46 chordoma-resected neoplasms between 2008 and 2014, along with clinical information, were collected. K-means consensus clustering and principal component analysis were used to identify and validate the clusters. Single-sample gene set enrichment analysis, methylCIBERSORT algorithm, and copy number analysis were used to identify the characteristics of the clusters.ResultsUnsupervised clustering analysis confirmed two clusters with a progression-free survival difference. Gene set enrichment analysis indicated that the early and late estrogen response pathways and the hypoxia pathway were activated whereas the inflammatory and interferon gamma responses were suppressed. Forty-six potential therapeutic targets corresponding to differentially methylated sites were identified from chordoma patients. Subgroups with a worse outcome were characterized by low immune cell infiltration, higher tumor purity, and higher stemness indices. Moreover, copy number amplifications mostly occurred in cluster 1 tumors and the high-risk group. Additionally, the presence of a CCNE1 deletion was exclusively found in the group of chordoma patients with better outcome, whereas RB1 and CDKN2A/2B deletions were mainly found in the group of chordoma patients with worse outcome.ConclusionsChordoma prognostic epigenetic subtypes were identified, and their corresponding characteristics were found to be variable.

Unravelling the role of immune cells and FN1 in the recurrence and therapeutic process of skull base chordoma

“…Conversely, recurrent chordomas showed an enrichment of Treg cells (Foxp3) and natural killer (NK) cells (CD56) (Figure S6). 30 Despite the elevated proportion of Tregs observed in recurrent chordomas (Figure S5), the cells exhibited significant expression of TIGHT, CTLA4, PDCD1, ENTPD1 and HAVCR2 (Figure 2B), indicating the immunosuppressive nature of Tregs. One of the primary ways in which chordoma lesions exhibit immunosuppression is through the identification of CTLA4 and PDCD1 expression on Tregs 31,32 .…”

“…We found that FAP (a fibroblast marker) has higher expression in recurrent chordoma, and this result was confirmed by IHC experiments (Figures S16 and S17). 30 Using specific markers, a total of 1796 fibroblasts were identified, which were classified into four distinct subclusters by U‐map analysis (Figures 4A and S18 and Table S6). Confirming their identity as fibroblasts, all four subclusters demonstrated elevated expression levels of well‐known fibroblast markers including DCN, LUM and COL1A2 8,29 …”

“…Immune cells, which are a crucial part of the TME, frequently demonstrate suppressive effects on the immune system 30 . The TME consists of a wide range of cellular components that enable interactions among cancerous cells, immune cells and stromal cells 59 .…”

Multimodal profiling of chordoma immunity reveals distinct immune contextures

“…Studies focused on methylation and RNA sequencing have identified groups of patients with either an inflamed or non-inflamed phenotype and associated these groups to clinical outcome, all with varying implications. 34–36 For instance, Zuccato et al found through methylation profiling that a group of patients presenting an inflamed phenotype had a worse prognosis than patients without these inflammatory features, 35 whereas Huo et al observed that patients from the non-inflamed group had a worse clinical outcome. 34 Furthermore, Baluszcek et al described two similar groups based on methylation and transcriptomic profiling, without finding an association between these groups and survival.…”

Multimodal profiling of chordoma immunity reveals distinct immune contextures

“…Studies focussed on methylation and RNA sequencing have identified groups of patients with either an inflamed or non-inflamed phenotype and associated these groups to clinical outcome, all with varying implications 29–31 . For instance, Zuccato et al found through methylation profiling that a group of patients presenting an inflamed phenotype had a worse prognosis than patients without these inflammatory features 30 , whereas Huo et al observed that patients from the non-inflamed group had a worse clinical outcome 29 . Furthermore, Baluszcek et al described two similar groups based on methylation and transcriptomic profiling, without finding an association between these groups and survival 31 .…”