Discussions of cancer biology at the molecular level often become alphabet soup – the complex signaling pathways are myriads of acronyms acting on one another. However, keeping track of these complex pathways is essential as many tumors are characterized by mutation or amplification of genes that initiate signals for biosynthesis and proliferation. Research on signaling networks has identified key oncogenic proteins and yielded targets for potent therapeutics. For example, tumors containing an activating mutation in the signaling protein, BRAF, often respond well to BRAF inhibitors. This approach targets the mutated protein directly so it can no longer drive the cancer phenotype, but this is not always effective. Many oncogenic proteins cannot be targeted using traditional chemotherapy approaches and in some instances tumors develop drug resistance. Knowing these challenges, Dr. Patrick Paddison’s Laboratory (Human Biology Division) uses a less biased approach to identify new drug targets. Genetic screening with CRISPR/Cas9 or RNAi technologies allows these researchers to sift through thousands of genes and identify the ones essential for survival of cancer cells but dispensable for growth of healthy cells. In a recent Oncotarget publication, Dr. Yu Ding found that tumor-initiating glioblastoma cells were sensitive to the inhibition of a zinc finger protein, ZNF131, while non-cancerous neural progenitor cells tolerated inhibition. These researchers further demonstrated that ZNF131 likely functioned as a transcriptional regulator of proteins required for microtubule organization and thus cell division.
Glioblastoma is an aggressive brain cancer, most recently in the news because of Senator John McCain’s diagnosis. Glioblastoma tumors contain a mix of cell types; within this mix is commonly a population that expresses similar genes to neural progenitor cells. These glioblastoma stem-like cells are very potent at forming tumors in mice – dozens of these cells are sufficient for tumor engraftment, whereas tens of thousands of more differentiated glioblastoma cells are required. Thus glioblastoma treatments need to target this population. For this screen Dr. Ding targeted a pool of approximately 19,000 small hairpin RNAs (shRNA) into both glioblastoma stem-like cells and neural progenitor cells such that individual cells contained one shRNA that inhibited a single gene. This revealed that inhibiting any of 162 genes was extremely lethal to glioblastoma cells, while neural progenitor cells were largely unaffected. The rest of the study followed up on ZNF131, one of the most effective targets in the screen. The screen was performed in glioblastoma cells isolated from three different tumors, and ZNF131 showed the same behavior in four more samples, suggesting it was a widespread target for glioblastoma.
Image provided by Dr. Jacob Herman
Creating a therapy targeting ZNF131 would be easier if the function of ZNF131 were understood, thus Dr. Ding looked at RNA profiles of control glioblastoma stem-like cells and those depleted of ZNF131. Zinc finger proteins commonly regulate gene transcription, which was observed from this experiment, but interestingly ZNF131 appeared to function very specifically - only affecting the RNA levels of dozens of genes rather than hundreds. Researchers found that ZNF131 depleted glioblastoma cells could survive if they expressed one of its targets, HAUS5. This was consistent with microscopic analysis of ZNF131 depleted cells. HAUS5 is a component of the Augmin complex that binds laterally to microtubules and nucleates new polymers helping create a robust mitotic spindle. When ZNF131 was depleted in glioblastoma cells they arrested in mitosis with diminished spindles (less tubulin staining) or multipolar spindles (more than a single spindle axis).
Further experiments suggested that the Augmin complex is essential in glioblastoma and non-cancerous cells but that HAUS5 is a limiting component of the complex only in the glioblastoma cells. When either ZNF131 or HAUS5 were depleted, a different Augmin complex member, HAUS6, was delocalized from mitotic spindles of glioblastoma cells. This suggests that the function of the entire complex is dependent of HAUS5 in the glioblastoma cells. On the other hand neural progenitor cells could survive HAUS5 depletion but simultaneous depletion of other complex members, HAUS2 or HAUS4, was lethal. Collaborative experiments with the Olson Lab (Clinical Research Division) also demonstrated that glioblastoma cells expressing shRNAs targeting ZNF131 or HAUS5 failed to form tumors in mice brains, indicating these are useful targets translate to an in vivo context.
The last question to ask was if a certain pathway caused the ZNF131/HAUS5 sensitivity. If a common oncogene like EGFR mutation or BRAF activation induced the behavior it could be used in personalized medicine to identify the right patients for a therapy. “This is a question we tried hard to answer. We have previously found that over-activation of oncogenic signaling pathways in tumor cells can weaken chromosomal attachments to the mitotic spindle causing unique therapeutic opportunities. However, for ZNF131 requirement our best guess is that aneuploidy (abnormal chromosome number) is driving the requirement. One possibility is that to accommodate capture of additional chromosomes the spindle needs to bulk up to counter balance the forces at play. Similar to an overloaded bridge, the spindle in cancer cells may simply need more supports to prevent a collapse. Thus they require the spindle nucleated microtubules that ZNF131 helps maintain in the cancer cells” said Dr. Paddison.
This research was funded by the National Cancer Institute, the Department of Defense, and Pew Biomedical Scholars Program.
Ding Y, Herman JA, Toledo CM, Lang JM, Corrin P, Girard EJ, Basom R, Delrow JJ, Olson JM, Paddison PJ. 2017. ZNF131 suppresses centrosome fragmentation in glioblastoma stem-like cells through regulation of HAUS5. Oncotarget. Jul 25;8(30):48545-48562