Transcriptomics-Based Phenotypic Screening Supports Drug Discovery in Human Glioblastoma Cells

Simple Summary Glioblastoma (GBM) remains a particularly challenging cancer, with an aggressive phenotype and few promising treatment options. Future therapy will rely heavily on diagnosing and targeting aggressive GBM cellular phenotypes, both before and after drug treatment, as part of personalized therapy programs. Here, we use a genome-wide drug-induced gene expression (DIGEX) approach to define the cellular drug response phenotypes associated with two clinical drug candidates, the phosphodiesterase 10A inhibitor Mardepodect and the multi-kinase inhibitor Regorafenib. We identify genes encoding specific drug targets, some of which we validate as effective antiproliferative agents and combination therapies in human GBM cell models, including HMGCoA reductase (HMGCR), salt-inducible kinase 1 (SIK1), bradykinin receptor subtype B2 (BDKRB2), and Janus kinase isoform 2 (JAK2). Individual, personalized treatments will be essential if we are to address and overcome the pharmacological plasticity that GBM exhibits, and DIGEX will play a central role in validating future drugs, diagnostics, and possibly vaccine candidates for this challenging cancer. We have used three established human glioblastoma (GBM) cell lines-U87MG, A172, and T98G-as cellular systems to examine the plasticity of the drug-induced GBM cell phenotype, focusing on two clinical drugs, the phosphodiesterase PDE10A inhibitor Mardepodect and the multi-kinase inhibitor Regorafenib, using genome-wide drug-induced gene expression (DIGEX) to examine the drug response. Both drugs upregulate genes encoding specific growth factors, transcription factors, cellular signaling molecules, and cell surface proteins, while downregulating a broad range of targetable cell cycle and apoptosis-associated genes. A few upregulated genes encode therapeutic targets already addressed by FDA approved drugs, but the majority encode targets for which there are no approved drugs. Amongst the latter, we identify many novel druggable targets that could qualify for chemistry-led drug discovery campaigns. We also observe several highly upregulated transmembrane proteins suitable for combined drug, immunotherapy, and RNA vaccine approaches. DIGEX is a powerful way of visualizing the complex drug response networks emerging during GBM drug treatment, defining a phenotypic landscape which offers many new diagnostic and therapeutic opportunities. Nevertheless, the extreme heterogeneity we observe within drug-treated cells using this technique suggests that effective pan-GBM drug treatment will remain a significant challenge for many years to come.

Automated summary

Glioblastoma remains a challenging cancer with limited treatment options, and future therapies will require personalized approaches targeting aggressive cellular phenotypes. The use of DIGEX has identified genes encoding drug targets and validated effective antiproliferative agents in human GBM cell models. Further research using this method may uncover new therapeutic opportunities and diagnostic strategies for the complex drug response networks in GBM.