Druggable genome CRISPRi screen in hydrogels reveals regulators of cortactin-driven actin remodeling promoting glioblastoma invasion
Mufeng Hu, Anna Weldy, Isabella M. Lovalvo, Erin A. Akins, Saket Jain, Alexander Chang, Ankita Sati, Meeki Lad, Austin Lui, Akhil Rajidi, Ameya Kothekar, Erika A. Ding, Juan A. Oses Prieto, Pablo Estevez, Alma L. Burlingame, Sanjay Kumar, Manish K. Aghi
Mufeng Hu, Anna Weldy, Isabella M. Lovalvo, Erin A. Akins, Saket Jain, Alexander Chang, Ankita Sati, Meeki Lad, Austin Lui, Akhil Rajidi, Ameya Kothekar, Erika A. Ding, Juan A. Oses Prieto, Pablo Estevez, Alma L. Burlingame, Sanjay Kumar, Manish K. Aghi
View: Text | PDF
Research Article Cell biology Oncology

Druggable genome CRISPRi screen in hydrogels reveals regulators of cortactin-driven actin remodeling promoting glioblastoma invasion

Abstract

To identify therapeutic targets limiting glioblastoma invasion, we applied druggable genome CRISPRi screens and multiomic analysis to patient-derived glioblastoma cells in micro-dissectible biomimetic 3D hydrogels that permitted separation and analysis of core versus invasive fractions. Of 2,550 genes screened, 12 encoded druggable targets whose suppression limited invasion, of which AURKB (encoding aurora kinase B) and ACP1 (encoding low molecular weight protein tyrosine phosphatase, LMW-PTP) were validated in neurosphere assays and in vivo. Proximity labeling identified cortactin as a link between LMW-PTP and aurora B, and we observed that cortactin underwent serine phosphorylation by aurora B and tyrosine dephosphorylation by LMW-PTP. Targeting ACP1 or AURKB via CRISPRi or inhibitors in culture and in vivo shifted the cortactin phosphorylation balance in glioblastoma, reducing levels of cortactin and the actin-related protein 2/3 (Arp2/3) complex that mediates cortactin-induced actin stabilization, thereby reducing actin-cortactin-Arp2/3 colocalization and subsequent actin polymerization. AURKB or ACP1 targeting shifted actin from cytoplasm to the nucleus, reducing mesenchymal gene expression. Biophysical analysis implicated AURKB in glioblastoma cell adhesion and stiffness needed for initial migration and ACP1 in mechanical stress resistance required for later migration. These findings revealed a targetable axis balancing kinase and phosphatase activities to regulate actin polymerization during glioblastoma invasion.

Authors

Mufeng Hu, Anna Weldy, Isabella M. Lovalvo, Erin A. Akins, Saket Jain, Alexander Chang, Ankita Sati, Meeki Lad, Austin Lui, Akhil Rajidi, Ameya Kothekar, Erika A. Ding, Juan A. Oses Prieto, Pablo Estevez, Alma L. Burlingame, Sanjay Kumar, Manish K. Aghi

×

Figure 2

CRISPRi screen identifies druggable genes mediating GBM invasion in 3D hydrogels.

Options: View larger image (or click on image) Download as PowerPoint
CRISPRi screen identifies druggable genes mediating GBM invasion in 3D h...
(A) Volcano plot displaying statistical significance (y axis) versus gene score, quantifying the average of the top 3 enrichments among 5 sgRNAs targeting the gene in the invasive front versus core (x axis) of GBM43 cells expressing an sgRNA library targeting the druggable genome before invading 3D hydrogels. Dashed lines represent cutoff for hits (FDR = 0.01); genes left of the dashed lines had sgRNAs enriched in the invasive front versus core. (B) Quantification of spheroid invasion assays of 8 single-gene CRISPRi knockdown GBM43 cell lines from screen hits compared with control cells expressing dCas9 and sgRNA targeting GAL4 (NT, nontargeting control). Only targeting AURKB or ACP1 slowed spheroid invasion (P < 0.0001; n = 30 spheres/group from 3 independent experiments). (C) Repeat quantification with larger sample size and representative images of GBM43 spheroid invasion assays with CRISPRi knockdown of AURKB and ACP1. (P < 0.0001; n = 49–56 spheres/group from 3 independent experiments). Original magnification, ×10; scale bar: 200 μm. (D) Quantification and representative images of spheroid invasion assays of GBM43 cells treated with aurora B inhibitor AZD1152-HQPA (barasertib) or LMW-PTP inhibitor I (MLS-0322825 Cmpd23); both slowed spheroid invasion (P = 0.01 AZD1152-HQPA, P = 0.006 LMW-PTP inhibitor; n = 7–13 spheres/group from 3 independent experiments). Original magnification, ×10; scale bar: 200 μm. (E) GBM43 spheroids treated with AZD1152-HQPA or LMW-PTP inhibitor I revealed a comparable timeline of slowed invasion beginning on day 3 (P < 0.05) and becoming robust from day 4 until day 6 (P < 0.0001); data assessment revealed that AZD1152-HQPA or LMW-PTP inhibitor I dramatically slowed invasive velocity over 6 days (P < 0.0001). n = 25–37 spheroids/group; ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. For box and whisker plots, the horizontal line is the median; the box extends from 25th to 75th percentile and the whiskers from minimum to maximum value. One-way (B–D and right side of E) and 2-way (left side of E) ANOVA with Tukey’s pairwise post hoc comparisons.