Drak/STK17A Drives Neoplastic Glial Proliferation Through Modulation of MRLC Signaling Alexander S

Published OnlineFirst December 10, 2018; DOI: 10.1158/0008-5472.CAN-18-0482

Cancer Molecular Cell Biology Research

Drak/STK17A Drives Neoplastic Glial Proliferation through Modulation of MRLC Signaling Alexander S. Chen1, Joanna Wardwell-Ozgo1, Nilang N. Shah1, Deidre Wright1, Christina L. Appin4,5, Krishanthan Vigneswaran6, Daniel J. Brat3,4,5, Harley I. Kornblum7,8, and Renee D. Read1,2,3

Abstract

Glioblastoma (GBM) and lower grade gliomas (LGG) are chain (MRLC), which was necessary for transformation. More- the most common primary malignant brain tumors and are over, Anillin, which is a binding partner of phosphorylated resistant to current therapies. Genomic analyses reveal that Sqh, was upregulated in a Drak-dependent manner in mitotic signature genetic lesions in GBM and LGG include copy cells and colocalized with phosphorylated Sqh in neoplastic gain and amplification of chromosome 7, amplification, cells undergoing mitosis and cytokinesis, consistent with their mutation, and overexpression of receptor tyrosine kinases known roles in nonmuscle myosin-dependent cytokinesis. (RTK) such as EGFR, and activating mutations in components These functional relationships were conserved in human of the PI3K pathway. In Drosophila melanogaster, constitutive GBM. Our results indicate that Drak/STK17A, its substrate co-activation of RTK and PI3K signaling in glial progenitor Sqh/MRLC, and the effector Anillin/ANLN regulate mitosis cells recapitulates key features of human gliomas. Here we use and cytokinesis in gliomas. This pathway may provide a new this Drosophila glioma model to identify death-associated therapeutic target for gliomas. protein kinase (Drak), a cytoplasmic serine/threonine kinase orthologous to the human kinase STK17A, as a downstream Significance: These findings reveal new insights into differ- effector of EGFR and PI3K signaling pathways. Drak was ential regulation of cell proliferation in malignant brain necessary for glial neoplasia, but not for normal glial prolif- tumors, which will have a broader impact on research regard- eration and development, and Drak cooperated with EGFR to ing mechanisms of oncogene cooperation and dependencies promote glial cell transformation. Drak phosphorylated Sqh, in cancer. the Drosophila ortholog of nonmuscle myosin regulatory light See related commentary by Lathia, p. 1036

Introduction more variable responses to therapeutics (2). To understand their genesis, these tumors have been subject to extensive Glioblastomas (GBM), the most common primary malig- genomic analyses, which show that signature genetic lesions nant brain tumors, infiltrate the brain, grow rapidly, and are in LGG and GBM include copy gain and amplification of resistant to current therapies (1). Low-grade gliomas (LGG), chromosome 7, amplification, mutation, and/or overexpres- which are related infiltrative malignant neural neoplasms, have sion of receptor tyrosine kinases (RTK), such as EGFR, slower tumor growth rates, longer patient survival, and display and activating mutations in components of the PI3K pathway (1, 3, 4). Nearly 60% of GBMs show focal EGFR copy gain or amplification, which are often accompanied by gain-of-func- 1Department of Pharmacology, Emory University School of Medicine, Atlanta, 2 tion EGFR mutations (4). The most prevalent EGFR mutant Georgia. Department of Hematology and Medical Oncology, Emory University VIII School of Medicine, Atlanta, Georgia. 3Winship Cancer Institute, Emory Univer- variant in GBM is EGFR (5), in which exon 2 to 7 deletion sity School of Medicine, Atlanta, Georgia. 4Department of Pathology, Emory confers constitutive kinase activity, which potently drives University School of Medicine, Atlanta, Georgia. 5Department of Pathology, tumorigenesis (1, 6). The most frequent PI3K pathway muta- Northwestern University Feinberg School of Medicine, Chicago, Illinois. 6Depart- tion in gliomas is loss of PTEN lipid phosphatase (4), which ment of Neurosurgery, Emory University School of Medicine, Atlanta, Georgia. results in unopposed PI3K signaling. 7 Department of Molecular and Medical Pharmacology, University of California Recent mouse models demonstrate that co-activation of Los Angeles, Los Angeles, California. 8Department of Psychiatry and Behavioral Sciences, and Semel Institute for Neuroscience and Human Behavior, University EGFR and PI3K pathways in glial cells or neuro-glial stems of California – Los Angeles, Los Angeles, California. cells induces GBM-like tumors, although these tumors do not show the full range of GBM phenotypes (7–9). Furthermore, to Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). date, pharmacologic inhibitors of EGFR and PI3K pathway components are ineffective in improving LGG and GBM out- A.S. Chen and J. Wardwell-Ozgo are co-first authors. comes (10). Genomic studies indicate that LGGs and GBMs Corresponding Author: Renee D. Read, Emory University, 1510 Clifton Road, have other genomic alterations (3–5); however, it is unknown Atlanta, GA 30322. Phone: 404-727-5985; E-mail: [email protected] how these changes contribute to gliomagenesis. Taken collec- doi: 10.1158/0008-5472.CAN-18-0482 tively, these data suggest that there are still undiscovered 2018 American Association for Cancer Research. biological factors that drive tumorigenesis. Given the

www.aacrjournals.org 1085

Chen et al.

aggressive nature of these tumors, there is a pressing need to Materials and Methods better understand their biology and to identify additional Drosophila Strains and culture conditions factors that could serve as new drug therapy targets. Drosophila stocks were obtained from the Bloomington stock To investigate the biology of malignant gliomas, we developed center and VDRC and the specific genotypes and stocks used are models in Drosophila melanogaster (11). Drosophila offers unique listed in the Supplementary Tables S1 and S2. Draknull(Drakdel) advantages for modeling gliomas: flies have orthologs for 70% of was a gift of David Hipfner (18). UAS-sqhD20D21 was a gift of human genes, including most known gliomagenic genes (12); l Guang-Chao Chen (25). UAS- dEGFR was a gift of Trudi Schup- Drosophila neural and glial cell types are homologous to their bach. All stocks were cultured on standard corn meal molasses human counterparts (13); and versatile genetic tools are available food at 25 C. Prior to publication, the UAS-DrakdsRNA stocks were for in vivo cell-type specific gene manipulation including RNAi validated by PCR amplification of the dsRNA element followed by (14). Although Drosophila models cannot address all aspects of sequence validation against the published sequence available at glioma biology, our model demonstrates that constitutive acti- the VDRC. To create UAS-Drak and UAS-DrakKD (kinase dead) vation of EGFR and PI3K signaling in glial progenitor cells gives constructs, the RE12147 Drak cDNA was cloned into pUAS-T, site rise to malignant glial tumors that recapitulate key biological directed mutagenesis was used to convert Lys-66 to Ala, and the features of human gliomas (11). resulting DNA was injected into embryos and stocks for each To discover new pathways that contribute to EGFR–PI3K– construct were established and sequence verified. All genotypes mediated glioma, we performed a kinome-wide RNAi-based were established by standard genetic crossing. To normalize for genetic screen in our EGFR–PI3K Drosophila GBM model (15). l UAS transgene dosage in repo>dEGFR ;dp110CAAX animals, a Kinases were screened because they are highly conserved in terms UAS-lacZ transgene was included in control genotypes. of protein function between Drosophila and mammalian systems. One of the top candidates from this screen was death-associated protein kinase related (Drak; ref. 15). Drak and its human ortholog, Immunohistochemistry fi STK17A, are cytoplasmic serine/threonine kinases in the Drak Larval brains were dissected with forceps, xed in 4% parafor- subfamily of cytoplasmic death-associated protein (DAP) kinases, maldehyde, processed, stained, and imaged as previously which regulate cytoskeletal dynamics, cytokinesis, and cell adhe- described (11). The following antibodies were used: 8D12 sion and mobility (16–18). In Drosophila development, Drak mouse anti-Repo (1:10, Developmental Studies Hybridoma promotes epithelial morphogenesis, acting downstream of Bank), anti-phospho-S21-Sqh (1:500, gift of Robert Ward; ref. 26), Rho-GTPase signaling to control the actin cytoskeleton through anti-Anillin (1:500, gift of Maria Giansanti; ref. 27). Secondary phosphorylation of Spaghetti Squash (Sqh; refs. 18, 19). MRLC, antibodies were conjugated to Cy3 (1:150), Alexa Fluor 488, or the human Sqh ortholog, is a regulator of nonmuscle myosin type Alexa Fluor 647 (1:100; Jackson Laboratories). Brains were II (NMII) motor proteins, and phosphorylated MRLC binds to mounted on glass slides ventral side down in vectashield and NMII and stimulates NMII-dependent contractile activity to pro- whole mount imaged on a Zeiss LSM 700 confocal system. For mote cytoskeletal re-organization, morphogenesis, and cytokine- experiments where protein levels were compared between geno- sis (20, 21). MRLC phosphorylation is dynamic and tightly types, all samples were prepared, subjected to IHC, imaged, and regulated, which allows for precise temporal and spatial changes image processed in a parallel manner side by side. Six or more to the cytoskeleton (20). Like other Drak family kinases, STK17A, brains were stained with each Ab combination, and representative which is expressed in GBM (22), can directly phosphorylate images are shown for each result. All brain phenotypes shown purified MRLC protein in vitro (18, 23). However, little is known were highly penetrant, with approximately 75% to 100% of about the mechanisms by which Drak/STK17A promotes animals showing the growth phenotypes described. Images were tumorigenesis. analyzed in Zeiss Zen Software and processed in Photoshop. Here we characterize genetic and functional requirements for Larval Drosophila brain hemisphere volumes were analyzed using Drak in EGFR- and PI3K-driven neoplastic glia. We report that Imaris software. Larval glial cells were counted manually in Drak operates downstream and in concert with EGFR signaling to representative optical sections of age-matched brain hemispheres, phosphorylate and activate Sqh to drive proliferation of neoplas- matched for section plane. Statistical analyses were done using tic glia. Moreover, our data show that, in mitotic neoplastic glial Prism. cells undergoing cytokinesis, phosphorylated Sqh colocalizes with Anillin, a cytoskeletal protein and known Sqh binding Mammalian tissue culture partner (24), and that Anillin is required for proliferation of GBM39, shared by C. David James, was created from human neoplastic glia. We show that these interactions are conserved in GBMs serially xenografted. GBM157, GBM281, GBM301, and human gliomas. STK17A is overexpressed in primary human GBM309 gliomasphere cultures were derived at UCLA and main- GBM and LGG tumors and patient-derived GBM cell cultures, tained in culture as described (28). HNPCs were obtained from and elevated STK17A expression correlates with elevated EGFR, Lonza. U87 and U87-EGFRvIII cells were gifts of Frank Furnari. Cell MRLC phosphorylation, and ANLN (human Anillin ortholog) culture was performed as previously described (15). Lentiviral levels. STK17A activity is required for tumor cell proliferation and shRNAs were prepared and used as previously described on serum for elevated levels of MRLC phosphorylation and ANLN. More- cultured cells and adherent serum-free cultured gliomasphere over, we found that STK17A, phosphorylated MRLC, and ANLN lines (15). WST-1 assays on shRNA-treated cells were performed are all colocalized to the cleavage furrow in tumor cells under- as previously described (15). For immunofluorescence, U87- going cytokinesis. Taken together, our data suggest that Drak/ EGFRvIII cells were plated on glass coverslips, fixed with 4% STK17A potentiates EGFR signaling to drive activation of Sqh/ paraformaldehyde, and stained on the coverslips with anti- MRLC, which in turn regulates mitosis in gliomas through effects b-tubulin (1:25, Developmental Studies Hybridoma Bank, on Anillin/ANLN. AA4.3), anti-STK17A (1:1,000, Sigma, HPA037979), anti-ANLN

1086 Cancer Res; 79(6) March 15, 2019 Cancer Research

Drak/STK17A Drive Neoplastic Glial Proliferation

(1:100, Sigma, HPA005680), anti-phospho-S19-MRLC (1:200, tional driver to co-overexpress constitutively active versions of l Abcam, ab2480; note this antibody detects S19 according to our dEGFR (dEGFR ) and dp110 (dp110CAAX), the catalytic subunit sequence alignment and not S20), and/or DRAQ7 (Cell Signaling of PI3K, that together drive malignant transformation of post- Technology, 1:200) to stain nuclear DNA and chromosomes. embryonic larval glia (11). The resulting glial tumors exhibit phenotypic and molecular characteristics similar to human GBM Immunoblot analysis (11). To identify novel modifiers of EGFR–PI3K–driven glial Cultured cells were collected and washed with 1 PBS and neoplasia, we used our Drosophila GBM model in a genetic screen, lysed in RIPA buffer containing protease and phosphatase inhi- and identified Drak, which encodes the sole Drosophila ortholog of bitors. Whole third instar larval brains were dissected were washed STK17A and STK17B serine/threonine kinases (15). with 1 PBS and lysed in RIPA buffer containing protease and Glial-specific Drak RNAi reduced neoplastic glial prolifera- phosphatase inhibitors. The following antibodies were used tion and altered glial morphogenesis, with DrakdsRNA#1 yielding for immunoblotting following the manufacturer's recommenda- significantly reduced brain sizes and glial cell numbers com- l tions: anti-STK17A (1:1,000, Sigma, HPA037979), anti-phospho- pared with dEGFR -dp110CAAX (Fig. 1A–I). The transforming S19-MRLC (1:500, Cell Signaling Technology, 3671), anti-phos- effects of EGFR–PI3K signaling were also reduced in Drak null pho-S19-MRLC (1:200, Abcam, ab2480), anti-MRLC (1:500, Cell mutants (Draknull; ref. 18) or by co-overexpression of kinase- Signaling, 3672), anti-EGFR (1:5000, BD), anti-ANLN (1:500, dead Drak (DrakKD): such mutants showed near wild-type Sigma, HPA005680), and anti-actin (1:200, Developmental Stud- brain sizes and reduced glial cell numbers compared with l ies Hybridoma Bank, JLA20). Bands were quantified using ImageJ. dEGFR -dp110CAAX controls (Fig. 1A–I), indicating that Drak l catalytic activity is essential for proliferation of dEGFR - In silico analysis dp110CAAX-mutant glia. STK17A mRNA expression data were obtained from cBioportal Growth inhibition of neoplastic glia induced by Drak knock- (www.cbioportal.org) and exacted and analyzed using RStudio. down or loss-of-function was not due to nonspecific glial lethal- All graphs and statistics were generated from this data using Prism ity: Drak is a nonessential gene and homozygous null mutants are (29, 30). The in silico analysis results shown here are based solely viable and show normal brain morphology and glial develop- upon data generated by the The Cancer Genome Atlas (TCGA) ment, and Drak RNAi in wild-type larval glia caused no obvious – Research Network (http://cancergenome.nih.gov; refs. 2 4). defects (Supplementary Fig. S1A and S1B; refs. 15, 18). Thus, although Drak is not required for normal glial proliferation and Tissue microarray processing development, Drak kinase activity is essential for neoplastic glial All human tumor specimens were collected from surgical proliferation. specimens donated for research with written informed consent Because reduced Drak function had a dramatic effect on of patients and were collected and used according to recognized l dEGFR -dp110CAAX mutant neoplastic glia, and because Drak ethical guidelines (Declaration of Helsinki, CIOMS, Belmont functions downstream of EGFR in epithelia (19), we predicted Report, GCP, Nuremberg Code) in a protocol (IRB00045732) that Drak may function downstream of EGFR in neoplastic glia. approved by the Institutional Review Board at Emory University. l Overexpression of constitutively active dEGFR alone elicits Paraffin-embedded human brain tumor specimens and tumor hyperplasia in glia (11), which we found was suppressed by loss tissue microarrays (TMA) with matched control tissue were pre- of Drak function (Supplementary Fig. S1C and S1D), consistent pared and sectioned using the Winship Core Pathology Labora- with our prediction. tory. Antigen retrieval and IHC staining was performed as spec- Previous studies show that, in developing epithelia, Drak acts ified by manufacturer's guidelines for each specific antibody (15). downstream of EGFR and RhoGTPase (RhoA) signaling in parallel The following antibodies were used: anti-STK17A (1:1,000, Sigma with Rho kinase (Rok; refs. 18, 19). To investigate whether HPA037979), anti-ANLN (1:50, Sigma, HPA005680), anti-phos- RhoGTPases are also required in neoplastic glia, we tested RhoA pho-S19-MRLC (1:200, Abcam, ab2480), anti-EGFR (1:50, Cell RNAi and dominant negative constructs in our GBM Drosophila Signaling Technology, 4267). Results were scored by neuro- model. We found that RhoA loss-of-function caused a significant pathologists according to standard clinical criteria on a scale of l reduction in dEGFR -dp110CAAX mutant glial proliferation, but 1 and 2 (low staining), 3 or 4 (high staining, with 4 being more did not obviously affect wild-type glia proliferation (Fig. 1A–I; uniform), and images of immunoreactivity were taken on an Supplementary Table S1). Glial-specific inhibition of Rho family Olympus DP72 CCD camera. GTPases Cdc42 and RhoL and glial-specific Rok RNAi did not as Statistical analyses strongly suppress neoplastic glial proliferation (Supplementary Comparisons between two groups were performed by Mann– Table S1; ref. 15). Thus, Drak may phosphorylate substrates to Whitney U test (nonparametric) for TCGA RNA-seq data analyses. drive neoplastic glial proliferation downstream of RhoA GTPase Comparisons between of three or more groups were performed by signaling, independent of Rok. Together, our data demonstrate – – one-way ANOVA with multiple comparisons to experimental that Drak pathways are necessary for EGFR PI3K dependent glial controls. Comparisons between two groups were done using neoplasia. either paired or unpaired parametric t tests. Fisher exact test was used to analyze correlations in IHC data. Drak cooperates with EGFR to promote glial transformation Because Drak reduction suppressed glial neoplasia in the context of constitutive EGFR–PI3K and EGFR signaling, we tested Results whether Drak overexpression cooperates with constitutive EGFR Drak is required for glial neoplasia in Drosophila signaling. We found that glial-specific Drak overexpression To understand GBM pathology, we developed a Drosophila had no obvious effect on normal larval glia morphology or GBM model. The model uses the glial-specific repo-Gal4 transcrip- number compared with wild-type (Fig. 2A–C). Glial-specific

www.aacrjournals.org Cancer Res; 79(6) March 15, 2019 1087

Chen et al.

Figure 1. Drak is required for glial neoplasia in Drosophila. A, Optical projections of whole brain–nerve cord complexes from third instar larvae approximately 130 hours old. Dorsal view; anterior up. CD8-GFP (green) labels glial cell bodies. Knockdown of Drak (repo>dEGFRl;dp110CAAX; DrakdsRNA) and Rho1 (repo>dEGFRl; dp110CAAX;Rho1dsRNA), the Drak null genetic background (Draknull; repo>dEGFRl;dp110CAAX), or overexpression of catalytically inactive Drak (repo>dEGFRl; dp110CAAX;DrakKD) decreased brain overgrowth relative to repo>dEGFRl; dp110CAAX. B, Total volumes (in mm3) of third instar larval brains, measured using Imaris. repo>dEGFRl;dp110CAAX(n ¼ 3), repo>dEGFRl;dp110CAAX; DrakdsRNA#1(n ¼ 5), repo>dEGFRl; dp110CAAX;DrakdsRNA#2(n ¼ 5), Draknull;repo>dEGFRl; dp110CAAX(n ¼ 3), repo>dEGFRl; dp110CAAX;DrakKD(n ¼ 3), repo>dEGFRl;dp110CAAX; Rho1dsRNA#1(n ¼ 4), repo>dEGFRl; dp110CAAX;Rho1dsRNA#2(n ¼ 5), wild- type (n ¼ 3). C, Glial cell numbers in representative 3 mm optical projections of third instar brain hemispheres (n ¼ 3 per genotype). Statistics generated using one-way ANOVA; , P < 0.0001. D–I, The 3 mm optical projections of brain hemispheres, age-matched third instar larvae. Frontal sections, midway through brains. Anterior up; midline to left. Repo (magenta) labels glial cell nuclei; CD8-GFP (green) labels glial cell bodies; anti- horseradish peroxidase (blue) counterstains for neurons and neuropil. D, repo>dEGFRl; dp110CAAX showed increased glial cell numbers (magenta nuclei, green cell bodies) relative to wild- type (I). E–G, Drak knockdown (DrakdsRNA), genetic reduction of Drak using a null allele (Draknull), or overexpression of catalytically inactive Drak (DrakKD) signiﬁcantly reduced glial cells compared with repo>dEGFRl;dp110CAAX.

l l dEGFR overexpression induced a significant increase in glial cell (15). Similar to dEGFR , hEGFRvIII overexpression caused glial numbers compared with wild-type controls (Fig. 2A–D; ref. 11). hyperplasia (15), and Drak and hEGFRvIII co-overexpression l Co-overexpression of Drak and dEGFR increased glial cell num- cooperated to drive increased proliferation and alterations in bers, and these glia lost their normal stellate shape and formed larval glial morphology consistent with neoplastic transformation abnormal cellular aggregates that disrupted normal larval brain (Supplementary Fig. S2A–S2D). Thus, although Drak overexpres- architecture, indicative of neoplastic transformation (Fig. 2A–E). sion alone has little effect in glia, Drak co-overexpression aug- We previously showed that overexpression of human EGFRvIII ments the ability of constitutively active EGFR to promote tumor- (hEGFRvIII), an oncogenic constitutively active mutant variant of igenesis. Thus, Drak acts as a bona fide genetic modifier in that EGFR found in GBM, cooperates with dp110CAAX to drive neo- Drak loss or gain exerts the observed effects only in the context of plastic glial transformation, which was suppressed by Drak RNAi other oncogenic mutations.

1088 Cancer Res; 79(6) March 15, 2019 Cancer Research

Drak/STK17A Drive Neoplastic Glial Proliferation

studies have shown that DAP family kinases, including STK17A, the human ortholog of Drak, phosphorylate MRLC at serine and threonine residues (Thr-18 and Ser-19; refs. 18, 21, 23). Similarly, Drak phosphorylates Sqh at the conserved serine residue (Ser-21 in Sqh is equivalent to Ser-19 in MRLC; refs. 18, 21, 23). Phos- phorylated Sqh binds to and stimulates the ATPase-dependent motor activity of Zipper, the sole NMII ortholog, which functions in cellular processes that require cytoskeletal contractility, such as cell migration, cytokinesis, and morphogenesis; all processes that play pivotal roles in tumorigenesis (21, 31, 32). We found that, similar to Drak knockdown, glial-specific sqh knockdown significantly reduced glial cell numbers and rescued l brain size in dEGFR -dp110CAAX mutants (Fig. 3A–F), which suggests that Drak activates Sqh by phosphorylation in trans- formed glia. To test this hypothesis, we used a validated phospho- specific antibody to examine levels of Ser-21-phosphorylated Sqh l (Sqh-S21-P) protein in dEGFR -dp110CAAX mutant glia in the l presence or absence of Drak (26). dEGFR -dp110CAAX mutant glia showed increased Sqh-S21-P compared with wild-type glia, with mitotic cells showing cortical enrichment of Sqh-S21-P (Fig. 3G), consistent with published observations regarding Sqh phosphorylation during cytokinesis (27). Drak RNAi reduced Sqh-S21-P levels (Fig. 3G), indicative of loss of Sqh activation. We next asked whether concurrent activation of EGFR and Drak signaling influences levels of activated Sqh. Consistent with Drak activity to phosphorylate Sqh, we observed increased levels of l Sqh-S21-P in dEGFR ;DrakOE glia compared with Drak-over expressing glia (Fig. 3G). By Western blot analysis, we observed increased levels of Sqh-S21-P with DrakOE;hEGFRvIII glia compared with hEGFRvIII-overexpressing glia (Supplementary Fig. S2E). These data are consistent with a model in which Drak increases the amount of activated, phosphorylated Sqh, which in turn supports EGFR-dependent tumorigenesis. If Sqh phosphorylation promotes EGFR-dependent neoplasia, then overexpression of a phospho-mimetic version of Sqh should Figure 2. enhance oncogenic EGFR. A prior study engineered a Sqh trans- Drak cooperates with EGFR to promote glial transformation. A, Glial cell gene with serine-21 and threonine-20 converted to aspartic acid numbers in representative 3 mm optical projections of third instar brain D20D21 P < (Sqh ) to mimic the phosphorylated state (33). As pre- hemispheres. Statistical analysis generated using one-way ANOVA; , D20D21 l 0.0001. Comparisons of wild-type to DrakOE and dEGFRl to dEGFRl;DrakOE dicted, Sqh and dEGFR co-overexpression increased num- were performed using paired parametric t tests. n.s., nonsignificant, P > 0.05; bers of small proliferative glia that disturbed normal larval brain , P < 0.05. wild-type (n ¼ 5), repo>DrakOE (n ¼ 7), repo>dEGFRl (n ¼ 4), architecture; these changes are indicative of enhanced hyperplasia l OE repo>dEGFR ;Drak (n ¼ 3). B–E, The 3 mm optical projections of brain and/or neoplastic transformation (Fig. 4A–E). Thus, Sqh is a hemispheres from third instar larvae, approximately 130 hours old. Frontal functionally relevant Drak substrate in EGFR–PI3K–mediated sections, midway through brains. Repo (magenta) labels glial cell nuclei; glial neoplasia. Taken together, our data provide strong support CD8-GFP (green) labels glial cell bodies; anti-horseradish peroxidase (blue) counter-stains for neurons and neuropil. A and B, Drak overexpression for a model in which Drak cooperates with oncogenic EGFR (DrakOE) using repo-Gal4 had no obvious effect on glial cell development signaling to create and sustain a pool of activated, phosphorylated compared with wild-type. A and C, dEGFRl overexpression alone increased Sqh necessary for glial cell transformation. glial cells compared with wild-type. D, repo>dEGFRl;DrakOE showed increased numbers of abnormal glia (magenta, nuclei; green, cell bodies), compared with DrakOE or dEGFRl alone, which lost their normal stellate A Sqh binding partner, Anillin, is required for neoplastic shape and formed aggregates that disrupt normal brain architecture and stunt brain development, evidenced by decreased neural tissue (HRP-stain). growth Following determination that glial neoplasia in our GBM model requires Sqh, we next sought to determine which processes Drak acts downstream of oncogenic EGFR to phosphorylate downstream of Sqh are essential in neoplastic glia. We used our Sqh Drosophila GBM model to test RNAi constructs against eight We next sought to understand the mechanism whereby Drak published Sqh binding partners (Supplementary Table S2). We contributes to glial transformation. Given that Drak catalytic found that glial-specific RNAi of anillin, which encodes a well- activity was essential for EGFR–PI3K glial neoplasia, we predicted established Sqh binding partner and multifunctional actin-bind- that Sqh, which is the Drosophila ortholog of human NMII ing scaffolding protein important for cytoskeletal reorganization l regulatory light chain (MRLC) and only known Drak substrate during cytokinesis (24), significantly reduced dEGFR -dp110CAAX (18), would be essential for EGFR–PI3K glial neoplasia. Prior mutant glia proliferation (Fig. 5A–F). Thus, anillin RNAi in glial

www.aacrjournals.org Cancer Res; 79(6) March 15, 2019 1089

Chen et al.

Figure 3. Drak acts downstream of oncogenic EGFR to phosphorylate Sqh. A, Optical projections of whole brain–nerve cord complexes from third instar larvae approximately 130 hours old. Dorsal view; anterior up. CD8-GFP (green) labels glia. Sqh knockdown (repo>dEGFRl;dp110CAAX;sqhdsRNA) decreased neoplastic overgrowth relative to repo>dEGFRl;dp110CAAX. B–D, The 3 mm optical projections of brain hemispheres from age-matched third instar larvae. Frontal sections, midway through brains. Repo (magenta) labels glial cell nuclei; CD8-GFP (green) labels glial cell bodies; anti-horseradish peroxidase (blue) counterstains for neurons and neuropil. B, repo>dEGFRl;dp110CAAX showed increased glial cell numbers (magenta, nuclei; green, cell bodies) relative to wild-type (D). C, Drak knockdown dramatically reduced neoplastic glial proliferation compared with controls. E, Total volumes (in mm3) of third instar larval brains, measured using Imaris. sqh RNAi in the context of dEGFRl;dp110CAAX reduced brain volume compared with dEGFRl;dp110CAAX controls. Statistical analysis generated using one- way ANOVA; , P < 0.0001. repo>dEGFRl;dp110CAAX (n ¼ 3), repo>dEGFRl;dp110CAAX;sqhdsRNA#1 (n ¼ 5), repo>dEGFRl;dp110CAAX;sqhdsRNA#2 (n ¼ 4), wild- type (n ¼ 3). F, Glial cell numbers in representative 3 mm optical projections of third instar brain hemispheres. Statistical analysis generated using one-way ANOVA; , P < 0.01 (n ¼ 3 per genotype). G, The 3 mm optical projections of brain hemispheres from third instar larvae. Repo (blue) labels glial cell nuclei; CD8- GFP (green) labels glial cell bodies; and Sqh-S21-P (magenta; Sqh-S21-P alone top, merge bottom). repo>dEGFRl;dp110CAAX brains (second from left) showed increased Sqh-P21-P compared with wild-type (leftmost). Drak knockdown in dEGFRl;dp110CAAX glia (middle) reduced Sqh-P21-P compared with repo>dEGFRl; dp110CAAX glia (second from left). Co-overexpression of Drak and dEGFRl increased Sqh-P21-P levels (rightmost) compared with Drak overexpression alone (second from right).

1090 Cancer Res; 79(6) March 15, 2019 Cancer Research

Drak/STK17A Drive Neoplastic Glial Proliferation

Figure 4. Sqh is a functionally relevant Drak substrate in glial neoplasia. A–D, The 3 mm optical projections of brain hemispheres from third instar larvae approximately 130 hours old. Frontal sections, midway through brains. Repo (magenta) labels glial cell nuclei; CD8- GFP (green) labels glial cell bodies; anti-horseradish peroxidase (blue) counterstains for neurons and neuropil. B, sqhD20D21 overexpression in glia caused no observable phenotype compared with wild-type (A). D, sqhD20D21 and dEGFRl co-overexpression yielded small proliferative glia that disturbed normal brain architecture, as compared with wild-type (A)and dEGFRl alone (C) brains. E, Glial cell numbers in representative 3 mm optical projections of third instar brain hemispheres. Statistical analysis generated using one-way ANOVA; , P < 0.0001; cell numbers of wild-type flies to sqhD20D21 flies compared using unpaired parametric t tests. n.s., nonsignificant, P > 0.05; dEGFRl flies to dEGFRl;sqhD20D21 compared using paired parametric t tests; , P < 0.05 (n ¼ 4per genotype).

cells may interfere with cytokinesis, thereby inhibiting tumor cell NMII/Zipper proteins in the contractile ring (24). During embry- proliferation. onic cellularization, Drak phosphorylation of Sqh is responsible To determine if Anillin is an effector of dEGFR–Drak–Sqh for proper organization of contractile rings (34), most likely signaling, we tested glial-speciﬁc anillin knockdown in either because Sqh phosphorylation was necessary for binding of Anillin l l dEGFR ;DrakOE or dEGFR ;sqhD20D21 mutant glia, and observed to Zipper. To determine if Drak-dependent phosphorylation of a signiﬁcant reduction in glial proliferation (Supplementary Fig. Sqh is necessary for Anillin binding and/or localization in mitosis, S3A–S3G), indicating that Anillin operates downstream of we examined Anillin localization in relation to Sqh-S21-P in l dEGFR–Drak–Sqh signaling. In contrast, anillin knockdown in neoplastic dEGFR -dp110CAAX glia. Consistent with their binding wild-type glia had no impact on glial cell proliferation compared in mitotic glia, Sqh-S21-P and Anillin were colocalized and with wild-type controls (Supplementary Fig. S3A–S3G), showing enriched at the cortex and cleavage furrow (observed in mitotic l a differential requirement for the Drak–Sqh–Anillin pathway in cells in all dEGFR ;dp110CAAX brains imaged, n ¼ 8), and this l neoplastic glia, but not wild-type glia. enrichment was lost upon Drak depletion (in all dEGFR ; MRLC phosphorylation mediates cellular processes that dp110CAAX;DrakdsRNA#1 brains imaged, n ¼ 6; Fig. 5G). Moreover, require NMII-dependent cytoskeletal contractility, including overall Anillin levels were reduced by Drak depletion (Fig. 5G). mitosis and cytokinesis (21, 31, 32). Previous studies demon- Thus, our data support a model wherein Drak-dependent phos- strate that phosphorylated Sqh recruits Anillin to the cortex during phorylation of Sqh promotes Anillin binding to coordinately mitosis, where it coordinates cytokinesis by linking actin and drive cytokinesis and proliferation in neoplastic glia.

www.aacrjournals.org Cancer Res; 79(6) March 15, 2019 1091

Chen et al.

Figure 5. A Sqh binding partner, Anillin, is required for neoplastic growth. A, Optical projections of whole brain–nerve cord complexes from third instar larvae approximately 130 hours old. Dorsal view; anterior up. CD8-GFP labels glia (green). Knockdown of anillin (repo>dEGFRl;dp110CAAX;anillindsRNA) shows decreased neoplastic brain overgrowth relative to repo>dEGFRl;dp110CAAX. B–D, The 3 mm optical projections of brain hemispheres from age-matched third instar larvae. Frontal sections, midway through brains. Repo (magenta) labels glial cell nuclei; CD8-GFP (green) labels glial cell bodies; anti-horseradish peroxidase (blue) counterstains for neurons and neuropil. C, Anillin knockdown dramatically reduced neoplastic glial proliferation in the context of dEGFRl;dp110CAAX compared with repo>dEGFRl;dp110CAAX controls (B). E, The total volume (in mm3) of third instar larval brains, measured using Imaris. Anillin knockdown drastically reduced brain volume in the context of dEGFRl;dp110CAAX. Statistical analysis generated using one-way ANOVA; , P < 0.001. repo>dEGFRl;dp110CAAX (n ¼ 3), repo>dEGFRl;dp110CAAX;anillindsRNA#1(n ¼ 4), repo>dEGFRl;dp110CAAX;anillindsRNA#2(n ¼ 4), wild-type (n ¼ 3). F, Glial cell numbers in representative 3 mm optical projections of brain hemispheres from third instar larvae. Statistical analysis generated using one-way ANOVA with multiple comparisons; , P < 0.0001 (n ¼ 3 per genotype). G, The 3 mm optical projections, third instar larval brains. CD8-GFP (green) labels glial cell membranes. F, Immunostaining showed low levels of Anillin protein (magenta; top) and Sqh-S21-P (blue; middle) in glia (white arrows; green; bottom) and neighboring neurons (GFP-negative) in wild-type. G, High levels of Anillin protein (magenta; top) and Sqh-S21-P (blue; middle) were detected in dEGFRl;dp110CAAX neoplastic glia (bottom), with Anillin showing nuclear localization and Sqh-S21-P showing cytoplasmic localization in nonmitotic cells. In mitotic cells (white arrow in large panel), Anillin and Sqh-S21-P showed membrane and cytoplasmic colocalization. In cytokinesis, Anillin localized to the cleavage furrow (white arrows in insets), with top inset showing cells in cytokinesis costained with Anillin and Sqh-S21-P with the bottom inset showing cells in cytokinesis costained Anillin and DRAQ7 DNA dye (cyan) to showcell nuclei. H, Immunostaining revealed reduced levels of Anillin protein (magenta; top) and Sqh-S21-P (blue; middle) upon Drak RNAi in dEGFRl;dp110CAAX glia.

1092 Cancer Res; 79(6) March 15, 2019 Cancer Research

Drak/STK17A Drive Neoplastic Glial Proliferation

STK17A expression correlates with EGFR status, alterations and well-characterized genetic lesions in gliomas (2–4, phosphorylated MRLC levels, and ANLN expression in human 29, 30). Well-characterized lesions include full or partial ampli- tumors fication of chromosome 7, which includes regions encoding both To determine whether the Drak–Sqh–Anillin pathway operates EGFR and STK17A (39), and focal EGFR amplification and muta- in human GBM and/or LGGs, we examined expression and tion. In TCGA cohorts of LGGs and GBMs, STK17A mRNA function of the human orthologs of Drak, STK17A, and STK17B, expression was significantly correlated with copy number gain and found that STK17A is overexpressed in GBMs (15). We next in 13p on chromosome 7 (7p13; Fig. 7C and D). To understand examined STK17A levels in a panel of patient-derived human whether STK17A mRNA overexpression is specific or is passively GBM stem-cell containing gliomasphere (GSC) cultures. Com- driven by copy gain, we examined mRNA expression of neigh- pared with cultured normal human neural progenitor cells boring genes on chromosome 7. We found that genes in close (HNPC), EGFRvIII-positive and EGFR-mutant GSC cultures proximity to STK17A (i.e., NACAD) showed no statistically sig- express higher levels of STK17A and ANLN (Fig. 6A). Moreover, nificant difference in mRNA expression between LGG specimens in GSC cultures with high STK17A levels, we also saw a modest with chromosome 7 copy gain compared with LGG specimens increase in Ser-19-phosphorylated MRLC (MRLC-S19-P) relative with no observable copy number alterations in the same region of to HNPCs (Fig. 6A). Thus, at the protein level, we observed that chromosome 7 (Fig. 7E), indicating that STK17A mRNA expres- the relationships between EGFR, STK17A, ANLN (Anillin), and sion may be selectively upregulated. Thus, increased STK17A MRLC phosphorylation in GBM cells recapitulate our observa- expression may have a specific role in glioma pathology. tions from Drosophila. To further assess whether STK17A is a driver of gliomagenesis, To further explore pathway conservation, we examined STK17A we used cBioportal to examine STK17A mRNA expression function and localization in serum-cultured GBM cell lines and relative to IDH1 status in patients with LGG. IDH1 mutation is GSC lines using RNAi and immunofluorescence. Consistent with a common genetic alteration in LGG, and patients who harbor prior reports (22), we found that, in serum cultured lines, STK17A IDH1 mutations have a better prognosis than those with wild-type is required for proliferation, with STK17A knockdown inducing IDH1 (40–42). LGG patients with wild-type IDH1 have more slower proliferation, apoptosis, reduced MRLC-S19-P levels, and aggressive tumors that behave much like primary GBMs (2, 40). altered cell shape and adhesion (Fig. 6B; Supplementary Fig. S4A Furthermore, wild-type IDH1, not mutant IDH1, is typically and S4B), which is consistent with alterations in MRLC regulation found in gliomas with chromosome 7 alterations (2, 40). In LGG, (35–37). We examined the effects of STK17A loss in GSCs treated we found elevated STK17A mRNA expression in tumors with with or without ZVAD to control for the effects of apoptosis, and wild-type IDH1 compared with tumors with mutant IDH1 we found that STK17A knockdown caused GSC adhesion defects, (Fig. 7F). In IDH1 mutant LGGs, there was not a statistically slower proliferation, apoptosis, and reduced levels of MRLC-S19- significant difference between STK17A mRNA expression between P and ANLN levels relative to control GSCs (Fig. 6C and D; tumors with or without chromosome 7 gain (Fig. 7G), suggesting Supplementary Fig. S4C and S4D). We also observed that total that STK17A levels are not as relevant to progression in IDH1 MRLC levels were reduced upon STK17A knockdown, suggesting mutant tumors. Our observations are consistent with a previous that phosphorylation may regulate total MRLC protein in GBM study showing STK17A mRNA overexpression in LGGs is corre- cells. This is consistent with published studies showing MRLC lated with disease severity and worse prognosis (22). levels are regulated by proteosomal turn-over (38). Furthermore, To investigate associations between STK17A expression and we examined STK17A localization in EGFRvIII-positive GBM cells patient survival, we analyzed cBioportal TCGA data to find that and observed that STK17A protein, in conjunction with MRLC- LGG and GBM patients with at least two-fold STK17A copy gain S19-P and ANLN, was upregulated in mitotic cells and localized to showed worse overall survival compared with patients with no the cleavage furrow in tumor cells undergoing cytokinesis (Fig. 6E STK17A copy gain (Supplementary Fig. S5A and S5B). Thus, LGG and F). Thus, STK17A expression is required to promote MRLC and GBM patients with STK17A copy gain have a worse overall phosphorylation and ANLN upregulation in GBM cells to coor- prognosis. dinately regulate cytokinesis and proliferation. Together, our results suggest that elevated STK17A expression We used IHC to examine protein expression of STK17A, drives MRLC- and ANLN-dependent cytoskeletal changes during EGFR, MRLC-S19-P, and ANLN in a collection of graded human mitosis and cytokinesis to facilitate disease progression (Fig. 7H), tumor specimens. In our LGG TMA, specimens with high validating our identification of the STK17A ortholog Drak as a STK17A expression showed a statistically significant correlation driver of tumorigenesis in our Drosophila GBM model. with high EGFR, MRLC-S19-P, and ANLN expression (Fig. 7A). In our GBM TMA, we observed high expression of EGFR, STK17A, MRLC-S19-P, and ANLN in the majority of specimens Discussion (Fig. 7B). However, we observed some GBM specimens with Although EGFR and PI3K pathways play important roles in high STK17A but low MRLC-S19-P expression (Fig. 7B): in glioma progression and maintenance, effective therapies targeting these surgical GBM specimens, it is possible that the MRLC- these pathways remain elusive (10). We developed a Drosophila S19-P phospho-epitope was not properly fixed and preserved. melanogaster GBM model based on co-activation of EGFR and Thus, in human tumors, elevated STK17A levels co-occur with PI3K in glia in order to gain insight into genetic and cellular elevated EGFR, MRLC-S19-P, and ANLN levels, which recapi- mechanisms underlying gliomas (11, 15). Using our system, we tulates the relationship we observed between EGFR, Drak, Sqh, identified a pathway through which the cytoplasmic serine/thre- and Anillin in Drosophila. onine kinase Drak specifically drives neoplastic proliferation. We used cBioportal to process genomic data catalogued by Consistent with published results (22), we show that the ortho- TCGA to assess prevalence of STK17A mRNA expression and copy logous kinase STK17A drives proliferation human GBM cells, gain alterations and to examine relationships between STK17A through a conserved pathway that regulates cytokinesis.

www.aacrjournals.org Cancer Res; 79(6) March 15, 2019 1093

Chen et al.

Figure 6. STK17A function is required for GBM cell proliferation, and regulates MRLC phosphorylation and ANLN expression. A, Panel of GBM gliomasphere cultures. GBM39 and GBM301 harbor amplified EGFRVIII; GBM281 harbors an activated variant mutant EGFR; GBM157 is PDGFR-overexpressing; GBM309 has neither (15, 28). Band intensities were normalized to actin (numbers above bands). B, WST-1 assay on U87 cells. Following selection for shRNA expression, U87 cell proliferation was measured by WST-1 reagent and quantified as fold-increase in absorbance between day 0 and day 3 after plating, normalized to controls treated with nontargeting shRNA. Three shRNAs tested for STK17A. P values from one-way ANOVA with Dunnett posttest. C, STK17A knockdown using three separate shRNAs in GBM301 decreased expression of ANLN and MRLC-S19- P compared with control cells treated with GFP shRNAs; all samples were treated with 20 mM ZVAD (Supplementary Fig. S4C and S4D shows non-ZVAD–treated cells). Cells were infected with lentiviral vectors 3 days prior to harvest. Band intensities were normalized to actin (numbers above bands), reported as fold changes relative to control cells. D, Adherently cultured GBM301 cells 3 days after infection with control lentivirus or shSTK17A#1, which reduced proliferation and altered cell shape and reduced adhesion; cells were treated with 20 mM ZVAD for 24 hours prior to imaging. E and F, U87-EGFRVIII cells were fixed and probed for either STK17A, phosphorylated MRLC (MRLC-S19- P), or ANLN (magenta) and a-tubulin (green) and DRAQ7 to label DNA (blue). E, STK17A, MRLC- S19-P, and ANLN were upregulated in tumor cells undergoing mitosis (cells marked with white arrows). F, STK17A, MRLC-S19-P, and ANLN were localized in the cleavage furrow of tumor cells undergoing cytokinesis (cells marked with white arrows).

During Drosophila development, Drak acts downstream of and GBM, STK17A is frequently subject to copy gain and over- EGFR and Rho-GTPase signaling to regulate epithelial tissue expression in association with EGFR and chromosome 7 altera- morphogenesis through Sqh phosphorylation (18, 19). Activated tions. We show that, in Drosophila and human tumor cells, Sqh/ Sqh, like human MRLC, modulates cytoskeletal reorganization in MRLC is a key mediator of Drak/STK17A: Sqh/MRLC is phos- cellular processes such as cytokinesis (32, 33, 43), which is phorylated at equivalent conserved sites in EGFR–PI3K mutant fundamental to cancer progression (44). Drak harbors latent tumor cells in a Drak/STK17A-dependent manner, and is neces- oncogenic activity, as it can cooperate with constitutively active sary for neoplastic growth. Thus, Sqh/MRLC is a functionally EGFR to stimulate glial transformation. Similarly, in human LGG relevant and evolutionarily conserved substrate of Drak/STK17A

1094 Cancer Res; 79(6) March 15, 2019 Cancer Research

Drak/STK17A Drive Neoplastic Glial Proliferation

Figure 7. STK17A expression correlates with EGFR status, MRLC-S19-P levels, and ANLN expression in human tumors. A and B, IHC on TMAs of human LGG tissue (A) or GBM tissue for STK17A, EGFR, MRLC-S19-P, or ANLN (reddish brown; B) showing cytoplasmic enrichment in tumor cells. Hematoxilin counterstain. Statistical analysis of low-grade glioma of LGG tumor specimens expressing STK17A (A) show a statistically signiﬁcant (Fisher exact test) correlation between high STK17A expression and high EGFR expression, high MRLC-S19-P expression, and high ANLN expression. B, Bottom, table of the number of GBM tumor specimens with either high or low expression of STK17A and high or low expression of either EGFR, MRLC-S19-P, or ANLN. C–G, Analysis of STK17A mRNA expression in relation to chromosome 7 gain, NACAD expression, and IDH1 status in LGGs and GBMs using TCGA datasets. Statistical analysis generated using Mann–Whitney U test; , P < 0.0001; , P < 0.001 (number of TCGA cases used: C, 282; D, 160; E, 281; F, 250; G, 219). H, Diagram depicting the role of Drak/STK17A in promoting gliomas.

in the context of EGFR–PI3K–driven glial tumorigenesis. This MRLC phosphorylation regulates many cellular processes that corroborates studies that have shown MRLC hyper-phosphoryla- require NMII-dependent contractility (21, 31, 32). To distinguish tion occurs in GBM tumors and that targeted inhibition of MRLC which of these processes are most relevant to Drak/STK17A activity inhibits GBM growth and invasion (35–37, 45, 46). function, we used a genetic approach and found that, among Together, our results establish STK17A as a disease-relevant MRLC known Sqh binding partners, Anillin is essential in neoplastic glia. kinase in gliomas. Anillin is a scaffolding protein that acts downstream of

www.aacrjournals.org Cancer Res; 79(6) March 15, 2019 1095

Chen et al.

RhoGTPase to organize the cytoskeleton and contractile ring in nism by which Drak-Sqh and EGFR cooperate to drive tumori- cytokinesis (24, 47–50). As part of this functionality, Anillin binds genesis, and the lack of NMII phosphorylation underlies the to Zipper (NMII) when Sqh is phosphorylated and activated (24). inability of Drak to promote glial proliferation when overex- Drak-dependent activation of Sqh promotes Anillin binding to pressed alone. Further studies are required to explore mechanisms Zipper and organization into contractile rings during cytokinesis by which Drak/STK17A cooperates with EGFR activation to pro- (32, 34). We observed that phosphorylated Sqh and Anillin mote tumorigenesis. colocalize in a Drak-dependent manner at the cleavage furrow In summary, our data validate use of invertebrate model in EGFR–PI3K mutant glia undergoing cytokinesis. Similarly, we organisms as a means to elucidate new aspects of glioma biology. observed that elevated levels of phosphorylated MRLC, ANLN, Our research reveals that Drak/STK17A dependency may provide and STK17A co-occur in human gliomas, and that these proteins a molecular vulnerability and therapeutically relevant target for colocalize at the cleavage furrow in EGFR-mutant GBM cells GBM and LGG. undergoing cytokinesis. Our results suggest that the primary role of Drak/STK17A in neoplastic glia is to promote cytokinesis to Disclosure of Potential Conflicts of Interest drive proliferation. No potential conflicts of interest were disclosed. Two other MRLC kinases, MLCK and ROCK, and two different NMII isoforms, including NMIIA and NMIIB, regulate GBM cell Authors' Contributions migration (35–37, 45, 46). Previous studies also show that Conception and design: A.S. Chen, J. Wardwell-Ozgo, R.D. Read STK17A promotes GBM cell migration in vitro (22). Therefore, Development of methodology: A.S. Chen, J. Wardwell-Ozgo, H.I. Kornblum, we did not study STK17A function in tumor cell invasion; instead, R.D. Read given the effects of Drak loss, we focused on Drak/STK17A Acquisition of data (provided animals, acquired and managed patients, function in GBM proliferation and cytokinesis. While regulation provided facilities, etc.): A.S. Chen, J. Wardwell-Ozgo, N.N. Shah, D. Wright, of cytokinesis in glioma cells is not well understood, our results K. Vigneswaran, D.J. Brat, R.D. Read Analysis and interpretation of data (e.g., statistical analysis, biostatistics, imply that cytokinesis in glioma cells is differentially regulated computational analysis): A.S. Chen, J. Wardwell-Ozgo, C.L. Appin, D.J. Brat, relative to normal developing glia or neural stem cells. In other R.D. Read tumor cell types, cytokinesis is preferentially controlled by Writing, review, and/or revision of the manuscript: A.S. Chen, J. Wardwell- NMIIC, which is also expressed in GBM cells (45), and may Ozgo, N.N. Shah, K. Vigneswaran, D.J. Brat, H.I. Kornblum, R.D. Read therefore mediate STK17A function. Further work is needed to Administrative, technical, or material support (i.e., reporting or organizing determine mechanisms of Drak/STK17A-dependent regulation of data, constructing databases): A.S. Chen, J. Wardwell-Ozgo, R.D. Read Study supervision: J. Wardwell-Ozgo, R.D. Read cytokinesis, and whether defective cytokinesis actively provokes growth arrest and apoptosis in GBM cells. Acknowledgments Although Drak overexpression alone causes no phenotype, We thank Tim Fenton, Clay Coston Rowe, Colleen Mosley, and Hye Rim Kim fi Drak overexpression intensi ed the proliferative output of for technical assistance, and Ken Moberg for critical reading of the manuscript. EGFR and PI3K signaling pathways through Sqh. Given that This work was supported by grants from the NIH/NINDS (NS065974), the Drak modifies glial EGFR–PI3K–driven neoplasia but does not Southeastern Brain Tumor Foundation, and Emory University Research Com- affect normal glial development, Drak and STK17A may require mittee to R.D. Read, and a K12-IRACDA career development award from the other signaling outputs downstream of EGFR or PI3K to drive NIH/NIGMS (GM000680) to J. Wardwell-Ozgo. proliferation. This is consistent with known requirements for RTK The costs of publication of this article were defrayed in part by the payment of and PI3K activity in cytokinesis and other NMII-dependent pro- page charges. This article must therefore be hereby marked advertisement in cesses (32). For example, previous reports indicate that increased accordance with 18 U.S.C. Section 1734 solely to indicate this fact. phosphorylation of NMII occurs in GBM cells in response to EGFR signaling (37). Thus, perhaps EGFR-dependent differential Received February 13, 2018; revised September 21, 2018; accepted December phosphorylation and activation of NMII underlies the mecha- 5, 2018; published first December 10, 2018.

References 1. Dunn GP, Rinne ML, Wykosky J, Genovese G, Quayle SN, Dunn IF, et al. confers enhanced tumorigenicity. Proc Natl Acad Sci U S A 1994;91: Emerging insights into the molecular and cellular basis of glioblastoma. 7727–31. Genes Dev 2012;26:756–84. 7. Bachoo RM, Maher EA, Ligon KL, Sharpless NE, Chan SS, You MJ, et al. 2. Cancer Genome Atlas Research N, Brat DJ, Verhaak RG, Aldape KD, Yung Epidermal growth factor receptor and Ink4a/Arf: convergent mechanisms WK, Salama SR, et al. Comprehensive, integrative genomic analysis of governing terminal differentiation and transformation along the neural diffuse lower-grade gliomas. N Engl J Med 2015;372:2481–98. stem cell to astrocyte axis. Cancer Cell 2002;1:269–77. 3. Comprehensive genomic characterization deﬁnes human glioblastoma 8. Zhu H, Acquaviva J, Ramachandran P, Boskovitz A, Woolfenden S, Pfannl genes and core pathways. Nature 2008;455:1061–8. R, et al. Oncogenic EGFR signaling cooperates with loss of tumor suppres- 4. Brennan CW, Verhaak RGW, McKenna A, Campos B, Noushmehr H, sor gene functions in gliomagenesis. Proc Natl Acad Sci U S A 2009; Salama SR, et al. The somatic genomic landscape of glioblastoma. Cell 106:2712–6. 2013;155:462–77. 9. Holland EC, Hively WP, DePinho RA, Varmus HE. A constitutively active 5. Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, et al. epidermal growth factor receptor cooperates with disruption of G1 cell- Integrated genomic analysis identiﬁes clinically relevant subtypes of glio- cycle arrest pathways to induce glioma-like lesions in mice. Genes Dev blastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and 1998;12:3675–85. NF1. Cancer Cell 2010;17:98–110. 10. Cloughesy TF, Cavenee WK, Mischel PS. Glioblastoma: from 6. Nishikawa R, Ji XD, Harmon RC, Lazar CS, Gill GN, Cavenee WK, et al. A molecular pathology to targeted treatment. Ann Rev Pathol 2014;9: mutant epidermal growth factor receptor common in human glioma 1–25.

1096 Cancer Res; 79(6) March 15, 2019 Cancer Research

Drak/STK17A Drive Neoplastic Glial Proliferation

11. Read RD, Cavenee WK, Furnari FB, Thomas JB. A drosophila model for 31. Kasza KE, Zallen JA. Dynamics and regulation of contractile actin-myosin EGFR-Ras and PI3K-dependent human glioma. PLoS Genet 2009;5: networks in morphogenesis. Curr Opin Cell Biol 2011;23:30–8. e1000374. 32. Cabernard C. Cytokinesis in Drosophila melanogaster. Cytoskeleton (Hobo- 12. Reiter LT, Bier E. Using Drosophila melanogaster to uncover human disease ken) 2012;69:791–809. gene function and potential drug target proteins. Expert Opin Ther Targets 33. Mitonaka T, Muramatsu Y, Sugiyama S, Mizuno T, Nishida Y. Essential 2002;6:387–99. roles of myosin phosphatase in the maintenance of epithelial cell integrity 13. Freeman MR, Doherty J. Glial cell biology in Drosophila and vertebrates. of Drosophila imaginal disc cells. Dev Biol 2007;309:78–86. Trends Neurosci 2006;29:82–90. 34. Chougule AB, Hastert MC, Thomas JH. Drak is required for actomyosin 14. Gonzalez C. Drosophila melanogaster: a model and a tool to investigate organization during Drosophila cellularization. G3 (Bethesda) 2016;6:819– malignancy and identify new therapeutics. Nat Rev Cancer 2013;13: 28. 172–83. 35. Gillespie GY, Soroceanu L, Manning TJ Jr., Gladson CL, Rosenfeld SS. 15. Read RD, Fenton TR, Gomez GG, Wykosky J, Vandenberg SR, Babic I, et al. Glioma migration can be blocked by nontoxic inhibitors of myosin II. A kinome-wide RNAi screen in Drosophila Glia reveals that the RIO kinases Cancer Res 1999;59:2076–82. mediate cell proliferation and survival through TORC2-Akt signaling in 36. Salhia B, Hwang JH, Smith CA, Nakada M, Rutka F, Symons M, et al. Role of glioblastoma. PLoS Genet 2013;9:e1003253. myosin II activity and the regulation of myosin light chain phosphoryla- 16. Lin Y, Hupp TR, Stevens C. Death-associated protein kinase (DAPK) and tion in astrocytomas. Cell Motil Cytoskeleton 2008;65:12–24. signal transduction: additional roles beyond cell death. FEBS J 2010; 37. Ivkovic S, Beadle C, Noticewala S, Massey SC, Swanson KR, Toro LN, et al. 277:48–57. Direct inhibition of myosin II effectively blocks glioma invasion in the 17. Bialik S, Kimchi A. The death-associated protein kinases: structure, func- presence of multiple motogens. Mol Biol Cell 2012;23:533–42. tion, and beyond. Annu Rev Biochem 2006;75:189–210. 38. Bornhauser BC, Olsson PA, Lindholm D. MSAP is a novel MIR-interacting 18. Neubueser D, Hipfner DR. Overlapping roles of Drosophila Drak and Rok protein that enhances neurite outgrowth and increases myosin regulatory kinases in epithelial tissue morphogenesis. Mol Biol Cell 2010;21: light chain. J Biol Chem 2003;278:35412–20. 2869–79. 39. Beroukhim R, Getz G, Nghiemphu L, Barretina J, Hsueh T, Linhart D, et al. 19. Robertson F, Pinal N, Fichelson P, Pichaud F. Atonal and EGFR signalling Assessing the significance of chromosomal aberrations in cancer: meth- orchestrate rok- and Drak-dependent adherens junction remodelling dur- odology and application to glioma. Proc Natl Acad Sci U S A ing ommatidia morphogenesis. Development 2012;139:3432–41. 2007;104:20007–12. 20. Vicente-Manzanares M, Ma X, Adelstein RS, Horwitz AR. Non-muscle 40. Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA, Yuan W, et al. IDH1 myosin II takes centre stage in cell adhesion and migration. Nat Rev Mol and IDH2 mutations in gliomas. N Engl J Med 2009;360:765–73. Cell Biol 2009;10:778–90. 41. Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, et al. An 21. Heissler SM, Manstein DJ. Nonmuscle myosin-2: mix and match. Cell Mol integrated genomic analysis of human glioblastoma multiforme. Science Life Sci 2013;70:1–21. 2008;321:1807–12. 22. Mao P, Hever-Jardine MP, Rahme GJ, Yang E, Tam J, Kodali A, et al. Serine/ 42. Guo C, Pirozzi CJ, Lopez GY, Yan H. Isocitrate dehydrogenase mutations in threonine kinase 17A is a novel candidate for therapeutic targeting in gliomas: mechanisms, biomarkers and therapeutic target. Curr Opin glioblastoma. PLoS One 2013;8:e81803. Neurol 2011;24:648–52. 23. Sanjo H, Kawai T, Akira S. DRAKs, novel serine/threonine kinases related to 43. Jordan P, Karess R. Myosin light chain-activating phosphorylation death-associated protein kinase that trigger apoptosis. J Biol Chem sites are required for oogenesis in Drosophila. J Cell Biol 1997;139: 1998;273:29066–71. 1805–19. 24. Straight AF, Field CM, Mitchison TJ. Anillin binds nonmuscle myosin II and 44. Newell-Litwa KA, Horwitz R, Lamers ML. Non-muscle myosin II in disease: regulates the contractile ring. Mol Biol Cell 2005;16:193–201. mechanisms and therapeutic opportunities. Dis Mod Mech 2015;8:1495– 25. Tang HW, Wang YB, Wang SL, Wu MH, Lin SY, Chen GC. Atg1-mediated 515. myosin II activation regulates autophagosome formation during starva- 45. Beadle C, Assanah MC, Monzo P, Vallee R, Rosenfeld SS, Canoll P. The tion-induced autophagy. EMBO J 2011;30:636–51. role of myosin II in glioma invasion of the brain. Mol Biol Cell 26. Zhang L, Ward REt. Distinct tissue distributions and subcellular localiza- 2008;19:3357–68. tions of differently phosphorylated forms of the myosin regulatory light 46. Wong SY, Ulrich TA, Deleyrolle LP, MacKay JL, Lin JM, Martuscello RT, et al. chain in Drosophila. Gene Expr Patterns 2011;11:93–104. Constitutive activation of myosin-dependent contractility sensitizes glio- 27. Giansanti MG, Vanderleest TE, Jewett CE, Sechi S, Frappaolo A, Fabian L, ma tumor-initiating cells to mechanical inputs and reduces tissue invasion. et al. Exocyst-dependent membrane addition is required for anaphase cell Cancer Res 2015;75:1113–22. elongation and cytokinesis in Drosophila. PLoS Genet 2015;11:e1005632. 47. Field CM, Coughlin M, Doberstein S, Marty T, Sullivan W. Characterization 28. Laks DR, Masterman-Smith M, Visnyei K, Angenieux B, Orozco NM, Foran of anillin mutants reveals essential roles in septin localization and plasma I, et al. Neurosphere formation is an independent predictor of clinical membrane integrity. Development 2005;132:2849–60. outcome in malignant glioma. Stem Cells 2009;27:980–7. 48. Field CM, Alberts BM. Anillin, a contractile ring protein that cycles from the 29. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. nucleus to the cell cortex. J Cell Biol 1995;131:165–78. Integrative analysis of complex cancer genomics and clinical profiles using 49. Thomas JH, Wieschaus E. src64 and tec29 are required for microfilament the cBioPortal. Sci Signal 2013;6:pl1. contraction during Drosophila cellularization. Development 2004;131: 30. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio 863–71. cancer genomics portal: an open platform for exploring multidimensional 50. Hickson GR, O'Farrell PH. Rho-dependent control of anillin behavior cancer genomics data. Cancer Discov 2012;2:401–4. during cytokinesis. J Cell Biol 2008;180:285–94.

www.aacrjournals.org Cancer Res; 79(6) March 15, 2019 1097

Drak/STK17A Drives Neoplastic Glial Proliferation through Modulation of MRLC Signaling

Alexander S. Chen, Joanna Wardwell-Ozgo, Nilang N. Shah, et al.

Cancer Res 2019;79:1085-1097. Published OnlineFirst December 10, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-18-0482

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2018/12/08/0008-5472.CAN-18-0482.DC1

Cited articles This article cites 50 articles, 23 of which you can access for free at: http://cancerres.aacrjournals.org/content/79/6/1085.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/79/6/1085.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/79/6/1085. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.