Related Experiment Video
Updated: Sep 22, 2026

Genetic Profiling and Genome-Scale Dropout Screening to Identify Therapeutic Targets in Mouse Models of Malignant Peripheral Nerve Sheath Tumor
Published on: August 25, 2023
A Reproducible Transcriptomic Program of Neurofibroma-to-Malignant Peripheral Nerve Sheath Tumor Transformation
Quanyu Jin1, Zhenxuan Gao1, Ze Zhang2
1Department of Neurosurgery, China-Japan Friendship Hospital (Institute of Clinical Medical Sciences), Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, China.
Abstract:
Malignant peripheral nerve sheath tumor (MPNST) carries the highest excess mortality of the cancers associated with neurofibromatosis Type 1, and usually arises from a neurofibroma. We asked whether the transcriptomic differences that distinguish MPNST from neurofibroma reproduce across independent cohorts, and whether a compact score derived from them transports between platforms. We meta-analyzed five public discovery cohorts of primary nerve-sheath tissue (113 MPNST, 79 neurofibromas; 3 platforms) and held out a sixth, independently generated cohort. Neurofibroma and MPNST tissue differed at 1489 genes in two opposing themes: up-regulated mitotic machinery (TOP2A, EZH2, AURKA) and loss of Schwann-cell identity (the transcription factors SOX10 and EGR2, and the myelin genes PMP2 and MPZ). These differences were reproducible. Among the significant genes assayed in all five cohorts, almost all changed in the same direction in everyone, and 93.3% of those detectable in the held-out cohort agreed in direction. Leave-one-cohort-out re-derivation of the pipeline gave a mean AUC of 0.90 across five folds (per-fold 0.84 to 0.97). CD276 (B7-H3) was selectively up-regulated, whereas CD274 and PDCD1LG2 were not. The myelin marker PMP2, a readout chosen retrospectively after the held-out results were known, separated MPNST from neurofibroma in that cohort but left atypical and conventional neurofibromas unseparated (AUC 0.67, p = 0.10). A program score derived without the discovery cohort carrying survival data was associated with shorter survival there, an exploratory single-cohort association. The composition of a compact score varied with platform and preprocessing, so no fixed single-sample classifier is warranted.

