A syngeneic spontaneous zebrafish model of tp53-deficient, EGFR^vIII, and PI3KCA^H1047R-driven glioblastoma reveals inhibitory roles for inflammation during tumor initiation and relapse in vivo

Alex Weiss
Cassandra D’Amata
Bret J. Pearson author has email address
Madeline N. Hayes author has email address

Developmental and Stem Cell Biology Program, The Hospital for Sick Children, Toronto, Canada
Department of Molecular Genetics, University of Toronto, Toronto, Canada
Knight Cancer Institute, Oregon Health & Science University, Portland, Oregon, 97239, USA
Department of Pediatrics, Papé Research Institute, Oregon Health & Science University, Portland, Oregon, 97239, USA

https://doi.org/10.7554/eLife.93077.2

Open access
Copyright information

Figures and data

her4.1-driven over-expression of patient-relevant oncogenes drives glial-derived brain tumor formation in syngeneic tp53 loss-of-function mutant zebrafish.
(A) Schematic of modeling strategy where linearized transgene vectors with the zebrafish her4.1 promoter driving human EGFR^vIII, human PI3KCA^H1047R, and mScarlet fluorescent protein are co-injected at the one-cell stage into syngeneic (CG1 strain) tp53^−/− mutant zebrafish embryos. Starting at 15 days post fertilization (dpf), mosaic-injected zebrafish were screened for CNS tumor formation, indicated by mScarlet expression in the brain region of live zebrafish. Co-injection of gfap:GFP linearized transgene is used to assess glial-specific cell fate specification in vivo. (B) her4.1:mScarlet and gfap:GFP expression in the anterior CNS of mosaic-injected syngeneic (CG1 strain) tp53^−/−zebrafish at 30dpf. (C) Whole brain dissected from a p53EPS mosaic-injected zebrafish at 30dpf. (D) Cumulative frequencies of mScarlet+ CNS lesions in syngeneic tp53^−/− mutant (CG1tp53^−/−) and wild-type (CG1) zebrafish injected at the one cell stage with her4.1:EGFR^vIII (E), her4.1:PI3KCA^H1047R (P), and/or her4.1:mScarlet (S). (E) Syngeneic (CG1 strain) zebrafish at 30dpf engrafted with her4.1:mScarlet+/gfap:GFP+ brain tumor cells, following primary transplantation (1T) at 2dpf into the embryonic brain ventricle. (F) Whole brain dissected from engrafted syngeneic host (CG1) zebrafish at 30dpf. (G) FACS plot of bulk syngeneic host brain following primary transplant (1T) of her4.1:EGFR^vIII + her4.1:PI3KCA^H1047R + her4.1:mScarlet + gfap:GFP brain tumor cells. (H-O) Histological staining of uninjected control (H-K) and p53EPS tumor burdened brains (L-O). (H,L) Hematoxylin and eosin (H&E) staining of coronal sections highlighting telecephalon and diencephalon regions of representative control (H) and p53EPS (L) brains. (I,M) Proliferating Cell Nuclear Antigen (PCNA) staining of control (I) and p53EPS (M) brain sections. (J,N) Phosphorylated-ERK (p-ERK) staining and quantifications reveal increased MAPK signaling pathway activation in p53EPS tumors (p<0.001, n=3 independent tumor sections). (K,O) Phosphorylated-Akt (p-Akt) staining and quantifications reveal increased Akt signaling pathway activity in p53EPS tumors (p=0.007, n=3 independent tumor sections). Scare bars represent 50μm.

RNA expression analysis establishes enrichment of mesenchymal glioblastoma and inflammation signatures in p53EPS model.
(A) Principal component analysis (PCA) of mRNA sequencing from whole control injected brains (CTRL), p53EPS and p53KPS tumor-burdened brains. (B) Heatmap of normalized counts for genes upregulated in p53EPS tumor-burdened brains (log2FoldChange > 2, padj < 0.05), compared to whole control injected brains (CTRL). A selected list of upregulated transcripts is indicated. (C) Gene set enrichment analysis (GSEA) plots of published gene signatures for mesenchymal subtype glioblastoma for genes differentially regulated in p53EPS compared to control injected brains (McLendon et al., 2008; Wang et al., 2017). Normalized Enrichment Scores (NES) and nominal p-values are indicated. (D) Bar plot of Normalized Enrichment Scores from GSEA of Hallmark gene sets (Liberzon et al., 2011). (E) Volcano plots of differentially expressed genes between sorted mScarlet+ p53EPS tumor cells and control injected whole brain tissue (CTRL WB), as well as between sorted mScarlet-negative cells from p53EPS tumor-burdened brains and control injected whole brains (CTRL WB). (F) her4.1:mScarlet and rag2:EGFP expression in live zebrafish with a p53EPS tumor at 30 dpf. (G) FACS plot of p53EPS brain with rag2:EGFP co-expression from (F). (H, I) her4.1:mScarlet and mpeg1.1:EGFP expression in live zebrafish with a p53EPS tumor at 30 dpf.

p53EPS recruits activated microglia/macrophages at early stages of tumor initiation.
(A-C) Neutral red staining of p53EPS mScarlet tumor-negative (A) and mScarlet tumor-positive brains (B,C) at 10 dpf. Neutral red foci in early-stage lesions are highlighted with arrows and are indicative of phagocytic cells. (D) Whole brain with p53EPS-induced tumor in a transgenic Tg(tnfa:EGFP) zebrafish at 20 dpf. (E) Z-stack projection of live confocal imaging of p53EPS tumor in transgenic Tg(tnfa:EGFP) background. (F-K) Z-stack projections of control uninjected (F-H) and p53EPS brains (I-K) at 5 dpf (F,I), 7 dpf (G,J), and 9 dpf (H,K) in transgenic Tg(mpeg1.1:EGFP) background. White arrows highlight an early-stage p53EPS lesion and associated mpeg:EGFP+ cells. (L) Z-stack projections of two independent p53EPS brains at 12 dpf in transgenic Tg(mpeg1.1:EGFP) background. (M) Quantification of tumor-associated mpeg1.1:EGFP+ cells with overlapping and/or internalized her4.1:mScarlet+ punctae (n=3 independent tumors).

Inflammation-associated irf7 and irf8 inhibit p53EPS formation in vivo.
(A-C) Primary (1°) control (A), irf7 CRISPR/Cas9 (B), and irf8 CRISPR/Cas9 (C) injected p53EPS at 30 dpf. (D) p53EPS incidence at 30 dpf in control (n=3 independent experiments, 108 zebrafish), irf7 CRISPR/Cas9 (***p<0.0001, Fisher’s exact test, n=2 independent experiments, 31 zebrafish), and irf8 CRISPR/Cas9 (*p=0.0155, Fisher’s exact test, n= 2 independent experiments, 36 total injected zebrafish). (E) Representative fluorescent in situ hybridization (FISH) images of whole mount control p53EPG (EGFR^vIII + PI3KCA^H1047 + GFP) and p53EPG + irf7 CRISPR/Cas9-injected zebrafish at 8 dpf. p53EPG (GFP, magenta) and irf7 (green) images represent Z-stack projections through tumor lesions (11 optical sections each). Merged images represent single optical sections at two spatially separated levels within control and irf7 knock-down tumors. DAPI staining (blue) is used to label nuclei. White arrowheads highlight irf7 expression specific to the TME. Scale bars represent 10μm. (F) Tg(her4.1:Cas9-2A-EGFP) expression at 30 dpf. (G) mScarlet+ p53EPS at 30 dpf in Tg(her4.1:Cas9-2A-EGFP) injected with irf7 gRNAs at the one-cell stage. (H) p53EPS incidence at 30 dpf in Tg(her4.1:Cas9-2A-EGFP)-negative gRNA-injected control siblings, and Tg(her4.1:Cas9-2A-EGFP) zebrafish injected at the one-cell stage with irf7 or irf8 gRNAs. n.s. not significant, Fisher’s exact test.

Inflammation-associated phagocytes inhibit p53EPS tumor engraftment.
(A,B) CG1 syngeneic host zebrafish at 20 dpf engrafted with p53EPS tumor cells transplanted with vehicle control (A) or Clodronate liposomes (B) at 2 dpf. (C-E) Quantification of p53EPS control engrafted and p53EPS tumors engrafted into CG1 host embryos with (C) Clodronate liposomes (p=0.0048, Fisher’s exact test, n=2 independent experiments, total 56 transplanted vehicle control and 50 transplanted Clodrosome-injected hosts), (D) engrafted into irf8 CRISPR/Cas9-injected into CG1 syngeneic host embryos (p=0.0002, Fisher’s exact test, n=2 independent experiments, total 100 transplanted control and 74 transplanted irf8 CRISPR/Cas9-injected hosts).

Inter-tumoral heterogeneity in p53EPS induced tumors. Sample screens of p53EPS injected fish showing distinct tumor initiation sites and varying degrees of mScarlet-fluorescent intensity. Tumors predominantly arise in the optic tectum/mesencephalon region with a subset appearing in the telencephalon/diencephalon region. Distinct tumors did not arise in the cerebellum/rhombencephalon regions over the course of our experiments.

Hematoxylin and eosin (H&E) staining of three independent p53EPS tumors. Scale bars represent 50μm.

(A) her4.1:mScarlet and gfap:GFP expression in the anterior CNS of 30dpf syngeneic (CG1) tp53^−/− zebrafish injected at the one-cell stage with linearized her4.1:KRAS^G12D + her4.1:PI3KCA^H1047R + her4.1:GFP + gfap:mScarlet transgenes. (B) Whole brain dissected from mosaic-injected her4.1:KRAS^G12D + her4.1:PI3KCA^H1047R + her4.1:GFP zebrafish at 30dpf. (C) Cumulative frequencies of GFP+ brain lesions in syngeneic tp53^−/− mutant (CG1tp53^−/−) and wild-type (CG1) zebrafish injected at the one cell stage with her4.1:KRAS^G12D (K), her4.1:PI3KCA^H1047R (P), and/or her4.1:GFP (G). (D) Immunohistochemical staining of proliferating cell nuclear antigen (PCNA), phosphorylated ERK (p-ERK), and phosphorylated AKT (p-AKT) on her4.1:KRAS^G12D + her4.1:PI3KCA^H1047R + her4.1:GFP tumor section. Quantification of p-ERK and p-AKT positive cells within total field. Scale bars represent 50μm.

Gene set enrichment analysis (GSEA) plots of published gene signatures for alternative molecular subtypes of human glioblastoma and medulloblastoma for genes differentially regulated in p53EPS compared to control injected brains (Cavalli et al., 2017; McLendon et al., 2008; Wang et al., 2017). Normalized Enrichment Scores (NES) and nominal p-values are indicated.

Quantitative real-time PCR analysis of neural stem cell (NSC) genes and genes associated with inflammatory gene expression signatures identified using bulk RNA sequencing. Gene expression in pooled her4.1:mScarlet+/gfap:GFP+ FACS-sorted cells relative to non-tumor control brain tissue. *p<0.01, **p<0.001, Student’s t-test.

Gene set enrichment analysis (GSEA) plots of established gene signatures for classical M1 polarized macrophages (Classical_M1_VS_Alternative_M2_Macrophage_UP), compared to alternative M2 macrophages (Classical_M1_VS_Alternative_M2_Macrophage_DN). GSEA plots of macrophage-specific gene expression at early (DAY3 UP) and late (DAY3 Down) time points of stimulation with macrophage colony stimulating factor (MCSF). Normalized Enrichment Scores (NES) and nominal p-values are indicated.

Quantitative real-time PCR analysis of irf7 CRISPR/Cas9-injected (A) and irf8 CRISPR/Cas9-injected p53EPS (B). Gene expression is represented relative to control p53EPS tumor-burdened whole brains. At least 5 tumor-burdened brains were pooled for each cohort and expression was normalized to mScarlet mRNA expression to account for differences in tumor size. *p<0.01, Student’s t-test.

(A,B) Images of neutral red staining in control (A) and irf8 CRISPR/Cas9-injected (B) zebrafish larvae at 8 dpf, A’ and B’ highlight boxed regions. Black arrows highlight examples of neutral red foci. (C) Quantification of neutral red foci. *p<0.01, n=4 fish for each cohort, Student’s t-test.

(A) Quantitative real-time PCR analysis of neural stem cell (NSC) genes and inflammation genes in non-tumor control brains, pooled p53EPS-burdened brains and pooled brains engrafted with p53EPS tumor cells at 20 dpf. *p<0.01 compared to non-tumor control whole brain tissue, Student’s t-test. (B) Quantitative real-time PCR analysis of neural stem cell (NSC) genes and inflammation genes at 20 dpf in primary transplanted (1T) p53EPS-burdened brains injected with vehicle control liposomes (1T p53EPS) and Clodrosomes at 2 dpf. At least 5 tumor-burdened brains were pooled for each cohort and expression was normalized to mScarlet mRNA expression to account for differences in tumor size. *p<0.01 compared to 1T p53EPS, Student’s t-test.

Sign up for email alerts