MYC Promotes Aggressive Growth and Metastasis of a WNT-Medulloblastoma Mouse Model

Abstract Medulloblastoma (MB), the most common malignant pediatric brain tumor, comprises four molecularly and clinically distinct subgroups (termed WNT, SHH, group 3, and group 4). Prognosis varies based on genetic and pathological features associated with each molecular subgroup. WNT-MB, considered low-risk, is rarely metastatic and contains activating mutations in CTNNB1; group 3-MB (GRP3-MB), commonly classified as high-risk, is frequently metastatic and can contain genomic alterations, resulting in elevated MYC expression. Here, we compare model systems of low-risk WNT-MB and high-risk GRP3-MB to identify tumor and microenvironment interactions that could contribute to features associated with prognosis. Compared to GRP3-MB, we find that WNT-MB is enriched in gene sets related to extracellular matrix (ECM) regulation and cellular adhesion. Exogenous expression of MycT58A in a murine WNT-MB model significantly accelerates growth and results in metastatic disease. In addition to decreased ECM regulation and cell adhesion pathways, we also identified immune system interactions among the top downregulated signaling pathways following MycT58A expression. Taken together, our data provide evidence that increased Myc signaling can promote the growth and metastasis in a murine model of WNT-MB.


Introduction
Medulloblastoma (MB), the most common malignant pediatric brain tumor, comprises four molecularly and clinically distinct subgroups (termed WNT, SHH, group 3, and group 4) [1].Across all subgroups, metastatic dissemination is a feature of high-risk disease and poor outcome [2].Depending on subgroup identity, recurrent MB tends to occur at primary or metastatic sites [3], but all MB subgroups commonly display MYC or MYCN amplifications and p53 loss at relapse [4,5].While these and other genomic alterations found in MB metastases provide biological distinctions from the primary tumor, prior work suggests that it does not result in a change in MB subgroup identity [6,7].
WNT-MB commonly harbor activating mutations in CTNNB1, resulting in hyperactivation of the Wntsignaling pathway [8][9][10].WNT-MB is known to have the best prognosis of all MB subgroups, with greater than 95% patients surviving, following standard of care therapy that includes surgery, radiation, and chemotherapy [11,12].
Prior studies identified features that are unique to WNT subgroup tumors.First, compared to the other 3 MB subgroups (SHH, group 3, and group 4), which arise from cells in the developing cerebellum, WNT-MB originate from progenitor cells in the lower rhombic lip of the brainstem [13].The unique cellular origin of WNT-MB relative to the other MB subgroups has been established through the creation of mouse models [10,13] and transcriptional comparisons of tumors against the developing murine brain [14].Second, tumor-associated vasculature within WNT-MB harbors a distinct loss of blood-brain barrier function [15].Wnt signaling is essential for the establishment and maintenance of the endothelial bloodbrain barrier [16][17][18].The secretion of Wnt antagonists such as DKK1 and WIF1 from WNT-MB tumor cells inhibits Wnt signaling in neighboring endothelial cells, consequently blocking the transcriptional program required to induce endothelial blood-brain barrier properties.This results in increased levels of drug penetration into WNT-MB, which may be one of the reasons why this MB subgroup responds exceptionally well to standard of care treatment [15].
There is a critical need to understand mechanisms that drive differences in prognosis among MB subgroups and to develop new biomarkers and therapies to improve overall outcomes.Group 3-MB (GRP3-MB), commonly classified as high-risk, is frequently metastatic and contains genomic alterations, resulting in increased MYC expression [8][9][10]19].The presence of metastasis is one of the most important prognostic markers for MB as data show near total and subtotal resection does not faithfully predict outcome [20].Metastasis also plays an important role in relapse as recurrent GRP3-and group 4-MB frequently occur at metastatic sites with the tumor bed void of disease [3].Yet relatively little is known about the mechanisms that promote MB metastatic dissemination or why differences exist between molecular subgroups.Prior studies have identified genetic alterations that promote MB metastasis, including increased phosphatidylinositol-3-OH kinase pathway activation [21,22], IGFR1 activity [23], WIP1 amplification [24], ATOH1 expression [25], and p53 loss of function [22][23][24]26].Moreover, in addition to circulating through the cerebral spinal fluid, MB cells also utilize the CCL2-CCR2 signaling axis to seed leptomeningeal metastasis after circulation through the blood stream [27].
In the present study, we leverage a combination of patient data and mouse model systems of low-risk WNT-MB to high-risk GRP3-MB to further delineate MB subgroup-specific cellular programs that may participate in regulating biological features associated with outcome.
We demonstrate that WNT-MB display enrichment in extracellular matrix (ECM) and cellular adhesion pathways.Overexpression of a stable Myc mutant in a murine WNT-MB model results in aggressive metastatic disease and disruption of ECM and cellular adhesion pathways.Our data further advocate the importance of MYC signaling in MB biology and identifies both tumor cell intrinsic and microenvironment interactions that it alters to promote MB growth and metastasis.

Cell Culture
Primary cultures of murine mWnt-and mGRP3-MB cell lines were previously generated in other studies [13,15,28].Cultures were maintained in ultra-low adherence plates (SBio) in defined Neurobasal medium (Gibco) containing N2 and B27 supplements (Gibco), penicillin/streptomycin (Gibco), and recombinant human bFGF and EGF growth factors (Shenandoah Biotechnology) [13,28].The proliferation of mWnt-MB cells transduced with EV or Myc T58A lentiviral constructs was measured by counting the number of cells present in individual wells of a low-adherent 6-well plate.Three independent experiments were performed, with each experiment including a minimum of triplicate wells per condition.Data were analyzed in GraphPad Prism and presented as mean ± SEM.Statistical tests were performed using unpaired t test with Mann-Whitney post hoc comparison.

Orthotopic MB Mouse Models
All mouse work was done according to institutional and IA-CUC Review Board (University of Cincinnati).6-8 week old female CD1-nude mice were obtained from Charles River (Crl:CD1-Foxn1nu, code 086).Orthotopic implantation of tumor cells was performed as described previously [15,30].Briefly, 1 × 10 6 tumor cells resuspended in 5 μL of Matrigel was injected via a borehole into the cerebellum of anesthetized CD-1 nude mice.For immunecompetent studies, 6-8 week old female CD1-IGS mice were obtained from Charles River (Crl:CD1[ICR], code 022).Orthotopic implantation studies were performed as described above.

RNA Preparation, Whole-Transcriptome Sequencing, and Analysis
Total RNA was isolated from freshly isolated samples using the NucleoSpin Plus RNA kit (Macherey-Nagel) as previously described [31,32].Samples included mWnt-MB + EV and mWnt-MB + Myc T58A tumors orthotopically implanted into CD1-nude mice.The control normal cerebellum was derived from 6 to 8 week old CD1nude mice.RNA-sequencing was performed as previously described [31].RNA quality control was performed on a bioanalyzer (BioRad) to ensure the quality of each sample submitted.For isolation of polyA RNA, a NEBNext Poly(A) mRNA Magnetic Isolation Module (New England BioLabs) was used for polyA RNA purification with a total of 1 μg good quality total RNA as input.The SMARTer Apollo NGS library prep system (Takara) was used for automated polyA RNA isolation.For RNA-sequencing library preparation, the library for RNA-seq was prepared by using the NEBNext Ultra II Directional RNA Library Prep Kit (New England BioLabs).After indexing via PCR enrichment (8 cycles), the amplified libraries together with the negative control were cleaned up for quality control analysis.To study differential gene expression, individually indexed and compatible libraries were proportionally pooled (~25 million reads per sample in general) for clustering in the cBot system (Illumina).Libraries at the final concentration of 15 pM were clustered onto a single-read flow cell using the IlluminaTruSeq SR Cluster Kit v3 and sequenced to 51 bp using the TruSeq SBS Kit v3 on the Illumina HiSeq system.Sequence reads were aligned to the reference genome using the TopHat aligner, and reads aligning to each known transcript were counted using Bioconductor packages for next-generation sequencing data analysis.The differential expression analysis between different sample types was performed using the negative binomial statistical model of read counts as implemented in the edgeR Bioconductor package.Transcriptional profiles were interrogated with Interactive Gene Expression Analysis Kit (for microarray and RNA-seq data), an R (v3.3.2), and JavaScript-based open-source desktop application [33] and integrated Differential Expression and Pathway analysis [34].Differential gene expression heatmaps and gene set enrichment analysis (GSEA) files were generated and downloaded in Interactive Gene Expression Analysis Kit and then analyzed using GSEA 4.0.Functional enrichment analysis of differentially expressed gene lists between tumor groups was performed in integrated Differential Expression and Pathway analysis using the ShinyGO app.All RNAseq files are deposited in Gene Expression Omnibus as GSE227631.All MB patient-derived transcriptome analysis was analyzed in the R2 genomics analysis and visualization platform [35] using publicly available datasets [1,36].

Tissue Collection and Immunostaining
Brains were collected and processed as previously described [15,31].Briefly, brains were rapidly dissected in ice-cold phosphate-buffered saline (PBS) and then fixed in 4% paraformaldehyde overnight.Fixed brains were washed in PBS and then incubated at 4°C in 30% sucrose/PBS solution overnight before embedding in tissue-freezing media (Ted Pella).50 microns thick free-floating sections were made using a cryostat (Leica).Freefloating sections were transferred to blocking solution (PBS +0.5% Triton X-100 + 10% normal donkey serum) at room temperature for 30-60 min before adding specific combinations of antibodies.Staining with Hoechst dye was done for 10 min (1:1000 in PBS) and washed in PBS again before being mounted onto slides (Fisher, Superfrost) and cover slipped (ProLong Gold Antifade, Ther-moFisher).Whole brain images were acquired on a stereomicroscope equipped with a digital camera (Motic) and fluorescent light and filter adapters (NightSea).Tissue section images were acquired on a confocal microscope (Nikon A1).All image analysis was performed in Image J (National Institutes of Health).

Cellular Adhesion and Extracellular Matrix Programs Are Enriched in WNT-MB
We surveyed available transcriptome datasets from MB patient samples [1,36] to determine differences in transcriptional programs between WNT-and GRP3-MB.Analysis of gene ontology categories overrepresented in the differentially (FDR <0.01) upregulated genes in WNT-MB identified ECM and cellular adhesion pathways among the most significantly enriched (shown in Fig. 1a).Common genes enriched within these pathways included different ECM components (laminins, collagens, vitronectin, and tenascin) and integrin receptors (shown in Fig. 1b).Within GRP3-MB, overrepresented gene ontology categories included GABA-ergic and neuronal signaling pathways, which have been previously identified to be enriched in GRP3-MB [1,37].
We next extended our analyses to established orthotopic implant mouse models of WNT-and GRP3-MB (denoted from herein as mWnt-MB and mGrp3-MB) [10,13,15,28].We noted differences between the invasiveness of mWnt-and mGrp3-MB tumors, where mWnt-MBs grew as well-defined tumors with a demarcated border at the normal brain interface (shown in Fig. 1c).This contrasted with mGrp3-MBs, where tumor cells invaded into neighboring normal hindbrain tissue and to more distant sites including the leptomeninges and lateral ventricle wall (shown in Fig. 1c).These differences are also present and noted in prior native forming genetic mouse models [10,13,38], suggesting invasive or metastatic differences are not an artifact induced by cell culture or orthotopic implantation.Further examination of transcriptomic datasets from mWnt and mGrp3-MB mouse models [13,28] identified gene sets related to ECM and cellular adhesion pathways as the most significantly enriched in mWnt-MB mouse models, while mGrp3-MB displayed enrichment for neuronal and MYC-related pathways (shown in Fig. 1d).This paralleled results in MB patient samples that identified increased ECM and cellular adhesion pathways in WNT-MB and provides Hartley/Phoenix further support that these murine models accurately recapitulate the molecular and pathological features of their human disease counterparts.

mWnt-MB Growth and Metastasis Is Enhanced by Expression of Stabilized Myc T58A
Increased MYC expression is a hallmark across many types of human cancer [39].This includes MB, where MYC family amplifications are associated with high-risk disease and metastasis [1,36,37].To investigate whether activation of Myc signaling alters the phenotype of WNT-MB, we transduced established mWnt-MB cell lines [10,13,15] to express a stabilized form of Myc (+Myc T58A ) or an empty vector control (+EV).Relative to +EV, expression of +Myc T58A significantly increased mWnt-MB proliferation in vitro (shown in Fig. 2a).We next utilized these transduced cells to test their ability to form tumors in vivo following orthotopic implantation into the hindbrain of immune-compromised mice (CD1-nude).Orthotopic allografts of mWnt-MB + Myc T58A cells developed tumors with a significantly shorter latency (median survival = 14 days post implantation) compared to mWnt-MB + EV control samples (median survival = 52 days post implantation) (shown in Fig. 2b).Samples collected from mWnt-MB + Myc T58A tumors showed local and distant metastasis that included GFP+ tumor cells along the ventral leptomeningeal axis, olfactory bulbs, and ventricular space (shown in Fig. 2c).In agreement with our prior observations in native and orthotopic implant models, mWnt-MB + EV tumors did not show evidence of invasion or metastasis (shown in Fig. 2c).GFP+ tumor cells could be found in the spinal cord of most mWnt-MB + Myc T58A samples (8 out of 10) but were not detected in any of the mWnt-MB + EV samples (0 out of 5) (shown in Fig. 2d, e).Thus, increased Myc signaling promotes increased growth, invasion, and metastasis of mWnt-MB.

Myc T58A Expression Downregulates Extracellular Matrix and Cellular Adhesion Transcriptional Signatures in mWnt-MB
To delineate the transcriptional programs altered by Myc T58A expression in mWnt-MB cells, we performed whole-transcriptome sequencing (RNA-seq) on mWnt-MB + EV, mWnt-MB + Myc T58A , and normal mouse cerebellum samples.Hierarchical clustering of all samples based on the top 1,000 most variably expressed genes showed each condition segregated independently (shown in Fig. 3a), which was also confirmed by k-means clustering (shown in Fig. 3b).Genes in k-means cluster C, which was enriched under both mWnt-MB +EV and +Myc T58A conditions relative to normal cerebellum, were strongly enriched in cell-cycle and proliferation pathways (shown in Fig. 3b).Genes in cluster D were elevated in only mWnt-MB + EV samples and displayed enrichment in pathways that included ECM organization (shown in Fig. 3b).
Prior studies have established that the molecular signature of MB subgroup identity is stable at recurrence/ relapse [3,5,7].We examined the expression of genes that comprise the core WNT-MB transcriptional signature [13,40] and did not find any significant changes in their expression between mWnt-MB + EV and mWnt-MB + Myc T58A samples (shown in Fig. 3c), suggesting the expression of Myc T58A did not result in a change of MB subgroup identity.We then utilized GSEA to compare mWnt-MB + Myc T58A versus mWnt-MB + EV samples.ECM and cellular adhesion pathways were among the top enriched gene sets in mWnt-MB + EV tumors (shown in Fig. 3d).Gene sets enriched in mWnt-MB + Myc T58A tumors were dominated by DNA replication and proliferation-related pathways (shown in Fig. 3d).

Myc T58A Expression Enhances Immune Evasion of mWnt-MB
Beyond changes in ECM and cell adhesion, we also noted that immune-related pathways were consistently among the most significantly downregulated gene sets in mWnt-MB + Myc T58A tumors.Over-representation analysis of genes significantly (FC ≥2, FDR ≤0.05) downregulated in mWnt-MB + Myc T58A versus mWnt-MB + EV found multiple gene sets related to regulation of immune system interactions (shown in Fig. 4a, b).Moreover, these and other immune-related pathways, such as antigen-processing cross presentation and interferon gamma signaling, were present among the top gene sets enriched in mWnt-MB + EV tumors in our GSEA, confirming their downregulation following expression of Myc T58A (shown in Fig. 4c).Increased MYC expression is known to promote tumor-immune evasion by regulating the expression of key immunomodulatory molecules [41].To test whether Myc T58A expression allowed mWnt-MB tumors to evade the immune system, we orthotopically implanted mWnt-MB + EV or mWnt-MB + Myc T58A cells into immunecompetent outbred mice.While we did not observe growth of mWnt-MB + EV control tumors in this setting, mWnt-MB + Myc T58A cells rapidly formed tumors (shown in Fig. 4d), and displayed regional invasion within the brain (shown in Fig. 4e).GFP+ tumor cells were also found in the spinal cord of 50% (5 out of 10) of the mWnt-MB + Myc T58A samples, suggesting these cells maintained the ability to spread to distant areas of the CNS (shown in Fig. 4e, f).

Discussion
Our findings identify increased expression of ECM and cellular adhesion components in WNT-MB relative to GRP3-MB.We show that expression of stabilized Myc T58A altered the growth characteristics of mWnt-MB cells, generating aggressive and metastatic tumors that displayed downregulation of ECM and cellular adhesion pathways.Taken together, this suggested a potential connection between the expression of ECM and cellular adhesion pathways and established outcome and pathological differences of WNT-MB.Beyond this, we also find that Myc T58A expression alters multiple immunerelated pathways in mWnt-MB and promotes tumor growth even in the setting of an intact immune system.
MB subgroups arise from distinct cellular origins, driven by unique mutations that block differentiation and promote tumor formation and growth [13,[42][43][44][45][46].This heterogeneity is further manifested in their clinical behavior, where SHH-, group 3-, and group 4-MB are fatal in 40-70% of cases [12] while WNT-MB are essentially curable (95%) [1,11,12].Prior work comparing MB subgroups has established links between tumor phenotypes and clinical prognosis, such as the lack of a blood-brain barrier in WNT-MB due to the inhibition of vascular Wnt-signaling by tumor secreted Wnt antagonists [15].
While we find that Myc T58A expression drives metastasis, it did not significantly alter the expression of genes that make up the core WNT-MB subgroup signature.This agrees with clinical data, which show that MB subgroups are stable at recurrence and/or relapse [6,7].The stability of MB subgroup signatures could be regulated by multiple factors, including the transcriptional programs of the (1) cell-of-origin and (2) initial driver mutations.Our data agree with the idea that acquisition of additional mutations or genetic alterations does not result in changes in the core MB subgroup identity transcriptional program but can induce additional transcriptional changes.These changes will be dependent on the specific mutations or genetic alterations acquired and how they interact with the transcriptional program of the primary tumor.Our data would support the idea that the selection of specific mutations or genetic alterations in recurrent and relapsed MB, such as MYC amplification or high expression [4,5,47] provide advantages through multiple mechanisms including cell proliferation, metastasis, and immune evasion.
Our data demonstrate that the expression of a stable Myc T58A mutant promotes the aggressive growth of mWnt-MB.Prior reports have highlighted hyperactivation of the Wnt-signaling pathway to inhibit MB tumorigenesis [48].Within SHH-MB mouse models, activation of Wnt-signaling significantly inhibited granule neuron precursor proliferation and MB formation [49].In GRP3-MB, Wnt activation is associated with decreased tumorigenic potential.A subset of cells within patient-derived GRP3-MB cell lines show high levels of Wnt activation and display decreased tumor initiating capacity.Moreover, activation of Wnt-signaling in GRP3-MB cells by treatment with Wnt-agonists significantly impairs tumorigenesis [48].We postulate the enhanced stability of Myc T58A helps overcome any restrictions imposed by Wnt-hyperactivation. Expression of wildtype Myc did not significantly enhance the proliferation of mWnt-MB cells in vitro (data not shown), suggesting that simply increasing Myc mRNA expression does not have the same effect as enhanced protein stability.Prior data from patient samples show that WNT-MB and GRP3-MB express similar mRNA levels of MYC  [50], further supporting regulation beyond transcription as an important participant in downstream effects.Further work is needed to delineate differences driven by Myc expression versus stability and its relationship to MB subgroup pathology.
Prior work has shown that MYC signaling can influence different aspects of tumor-host immune interactions that facilitate tumor growth and maintenance [51,52].This includes regulation of MHC class 1 antigens ( [53][54][55][56]), immune check point gene such as PD-L1 and CD47 [57], and inflammatory cytokines [58].We find that increased Myc signaling in our mWnt-MB model drove alterations in several immune system-related pathways and allowed robust growth in immune-competent hosts.The ability of Myc to promote tumor cell survival through multiple mechanisms, including immune escape, will be important to further delineate as immune-based therapies are being rapidly translated into the clinic.Heterogeneity in MYC expression, at both inter-and intra-tumoral levels, could provide routes to develop resistance to new therapies such as CAR T cells that utilize expression of specific antigens, check point inhibitors, or other immune-mediated treatment strategies.
Our data show that increased Myc signaling, in the context of mWnt-MB, promotes tumor growth and progression.These findings implicate multiple mechanisms downstream of Myc signaling, including changes in ECM and cellular adhesion interactions and components important for tumor-immune system interactions.This should help guide future studies examining the mechanisms that contribute to MB subgroup heterogeneity and difference at MB relapse and/or recurrence.

Statement of Ethics
This study protocol was reviewed and approved by the IACUC Review Board at the University of Cincinnati (16-07-06-01).

Fig. 1 .Fig. 2 . 3 (
Fig. 1.WNT-MB display enrichment in cellular adhesion and ECM signatures compared to GRP3-MB. a The top ten gene ontology categories overrepresented in differentially expressed genes between human WNT-and GRP3-MB.Sorted by -log10 (p value).b Heatmap depicting expression of genes in the collagencontaining extracellular matrix GO gene set.c Representative images of mWnt-and mGrp3-MB samples immunostained for Hoechst.Dashed lines demarcate tumor-brain boundaries.d The top five Gene Ontology (GO) Cellular Component and Reactome categories overrepresented in differentially expressed genes between mWnt-and mGrp3-MB models.Internal granule layer (IGL), lateral ventricle (LV).Scale bars = 100 µm.

Fig. 3 .Fig. 4 .
Fig.3.Myc T58A expression downregulates ECM and cellular adhesion transcriptional signatures in mWnt-MB.a Heatmap of unsupervised hierarchical clustering of cerebellum (n = 3), mWnt-MB + EV (n = 3), and mWnt-MB + Myc T58A (n = 3) samples using the top 1,000 most variably expressed genes.b Heatmap of K-means clustering of cerebellum, mWnt-MB + EV, and mWnt-MB + Myc T58A samples using the top 2,000 most variably expressed genes.Enrichment of gene ontology gene sets from cluster C and cluster D genes.c Heatmap depicting the expression of 34 WNT-MB signatures genes in mWnt-MB + EV and mWnt-MB + Myc T58A samples.d Gene set enrichment analysis of mWnt-MB + Myc T58A versus mWnt-MB + EV samples displaying enrichment of specific gene ontology gene sets under each condition.