SOX10 mediates glioblastoma cell-state plasticity.

Phenotypic plasticity is a cause of glioblastoma therapy failure. We previously showed that suppressing the oligodendrocyte-lineage regulator SOX10 promotes glioblastoma progression. Here, we analyze SOX10-mediated phenotypic plasticity and exploit it for glioblastoma therapy design. We show that low SOX10 expression is linked to neural stem-cell (NSC)-like glioblastoma cell states and is a consequence of temozolomide treatment in animal and cell line models. Single-cell transcriptome profiling of Sox10-KD tumors indicates that Sox10 suppression is sufficient to induce tumor progression to an aggressive NSC/developmental-like phenotype, including a quiescent NSC-like cell population. The quiescent NSC state is induced by temozolomide and Sox10-KD and reduced by Notch pathway inhibition in cell line models. Combination treatment using Notch and HDAC/PI3K inhibitors extends the survival of mice carrying Sox10-KD tumors, validating our experimental therapy approach. In summary, SOX10 suppression mediates glioblastoma progression through NSC/developmental cell-state transition, including the induction of a targetable quiescent NSC state. This work provides a rationale for the design of tumor therapies based on single-cell phenotypic plasticity analysis.

Glioblastoma epigenome profiling identifies SOX10 as a master regulator of molecular tumour subtype.

Glioblastoma frequently exhibits therapy-associated subtype transitions to mesenchymal phenotypes with adverse prognosis. Here, we perform multi-omic profiling of 60 glioblastoma primary tumours and use orthogonal analysis of chromatin and RNA-derived gene regulatory networks to identify 38 subtype master regulators, whose cell population-specific activities we further map in published single-cell RNA sequencing data. These analyses identify the oligodendrocyte precursor marker and chromatin modifier SOX10 as a master regulator in RTK I-subtype tumours. In vitro functional studies demonstrate that SOX10 loss causes a subtype switch analogous to the proneural-mesenchymal transition observed in patients at the transcriptomic, epigenetic and phenotypic levels. SOX10 repression in an in vivo syngeneic graft glioblastoma mouse model results in increased tumour invasion, immune cell infiltration and significantly reduced survival, reminiscent of progressive human glioblastoma. These results identify SOX10 as a bona fide master regulator of the RTK I subtype, with both tumour cell-intrinsic and microenvironmental effects.

A high-content RNAi screen reveals multiple roles for long noncoding RNAs in cell division.

Genome stability relies on proper coordination of mitosis and cytokinesis, where dynamic microtubules capture and faithfully segregate chromosomes into daughter cells. With a high-content RNAi imaging screen targeting more than 2,000 human lncRNAs, we identify numerous lncRNAs involved in key steps of cell division such as chromosome segregation, mitotic duration and cytokinesis. Here, we provide evidence that the chromatin-associated lncRNA, linc00899, leads to robust mitotic delay upon its depletion in multiple cell types. We perform transcriptome analysis of linc00899-depleted cells and identify the neuronal microtubule-binding protein, TPPP/p25, as a target of linc00899. We further show that linc00899 binds TPPP/p25 and suppresses its transcription. In cells depleted of linc00899, upregulation of TPPP/p25 alters microtubule dynamics and delays mitosis. Overall, our comprehensive screen uncovers several lncRNAs involved in genome stability and reveals a lncRNA that controls microtubule behaviour with functional implications beyond cell division.

Specificity of RNAi, LNA and CRISPRi as loss-of-function methods in transcriptional analysis.

Loss-of-function (LOF) methods such as RNA interference (RNAi), antisense oligonucleotides or CRISPR-based genome editing provide unparalleled power for studying the biological function of genes of interest. However, a major concern is non-specific targeting, which involves depletion of transcripts other than those intended. Little work has been performed to characterize the off-target effects of these common LOF methods at the whole-transcriptome level. Here, we experimentally compared the non-specific activity of RNAi, antisense oligonucleotides and CRISPR interference (CRISPRi). All three methods yielded non-negligible off-target effects in gene expression, with CRISPRi also exhibiting strong clonal effects. As an illustrative example, we evaluated the performance of each method for determining the role of an uncharacterized long noncoding RNA (lncRNA). Several LOF methods successfully depleted the candidate lncRNA but yielded different sets of differentially expressed genes as well as a different cellular phenotype upon depletion. Similar discrepancies between methods were observed with a protein-coding gene (Ch-TOG/CKAP5) and another lncRNA (MALAT1). We suggest that the differences between methods arise due to method-specific off-target effects and provide guidelines for mitigating such effects in functional studies. Our recommendations provide a framework with which off-target effects can be managed to improve functional characterization of genes of interest.