Homozygous but not heterozygous loss of Pten in a sporadic mouse model of SmoM2 SHH-MB increases penetrance and reduces time of disease onset

(A) An Oncoprint of 27 patient samples considered to be SHH-MB3 using the cBioPortal20, 21 showing the frequency of mutations in PTEN and genes that activate the SHH pathway showing co-occurrences. (B) Schematic showing Mosaic mutant Analysis with Spatial and Temporal Regulation (MASTR) technique used to model SHH-MB by expressing SmoM2 in scattered GCPs starting at P0 and deleting one or two copies of Pten (Atoh1-FlpoER/+; R26MASTR/LSL-SmoM2; Ptenfl/fl or SmoM2-Ptenfl/fl mice). (C) Kaplan-Meier curve and statistics (below) for survival of SmoM2 mice compared to SmoM2-Ptenfl/fl and SmoM2-Ptenfl/+ mice. Statistics were calculated using the log-rank test. (D-F) Representative images of hematoxylin and eosin (H&E) stained sagittal sections of end stage tumors from SmoM2, SmoM2-Ptenfl/+, and SmoM2-Ptenfl/fl mice with higher maginification images shown below (n>8 per genotype). Scale bars indicates 1mm (D-F) and 100 μm (D’-F’). (G-I) GFP stained saggital sections of end stage tumors from the the indicated mutants (n=5-8 per genotype). Scale bar indicates 1mm. (J-L) Saggital sections of end stage tumors from the indicated mutants stained for PTEN (J-L) or pAKT (M-O) with higher magnifications below. Scale bars indicates 1mm (J-O) and 20 μm (J’-O’). (D-O) Dashed lines outline the tumors, solid lines outline IGL, dotted rectangles indicate location of high magnification images.

Pten loss in SmoM2 sporadic SHH-MB induces highly differentiated tumors

(A, B) Sagittal sections of the cerebellum showing end stage tumors from SmoM2 and SmoM2-Ptenfl/fl mice stained for P27 (post mitotic cells) and Ki67 (proliferating cells) (SmoM2, n=5; SmoM2-Ptenfl/fl n=7). Dashed lines outline each tumor, solid lines outline IGL, dotted squared indicate location of high magnification images in (C-H). (C-H) High power images of sections stained for Ki67 and DAPI (C, D), P27 (E, F) or NeuN (G, H)(SmoM2, n=5; SmoM2-Ptenfl/fl n=7). Dashed line separates the internal granule layer (IGL) or molecular layer (ML) from the tumor. (I,J) Sagittal sections of the cerebellum showing lateral tumors from SmoM2 and SmoM2-Ptenfl/flmice stained for Ki67 and DAPI (SmoM2, n=7; SmoM2-Ptenfl/fln=8)). (K-N) RNA in situ hybridization staining of sections for Gli1 (K, L) and Syp mRNA (M, N) (n=3 per genotype). (O, P) Immunohistochemical staining of sections for SYP protein (n=6 per genotype). (I-P) Dashed lines outline the tumors, solid lines outline IGL. (A, B, I-P) Scale bars indicate 1mm. (C-H) Scale bars indicates 100μm.

Pten loss in SmoM2 SHH-MB protects differentiated tumor cells from cell death

(A, B) Sagittal sections of end stage tumors in SmoM2 and SmoM2-Ptenfl/flmice stained as indicated (SmoM2, n=3; SmoM2-Ptenfl/fl n=4). Dashed lines outline the tumors, solid lines outline IGL. Dotted line square indicates location of high magnification images show in (C-H). Scale bars indicate 1mm. (C-H) High magnification images of tumors stained as indicated showing reduced cell death (TUNEL particles) in differentiated regions (NeuN+ and SYP+) of tumors in SmoM2-Ptenfl/flmice compared to SmoM2 (SmoM2, n=3; SmoM2-Ptenfl/fln=4). Scale bars indicates 50μm. (I-K) Quantification of density of TUNEL particles in tumor regions with high SYP (SYP HI) compared to low (SYP Low) staining in SmoM2-Ptenfl/fl mice (I), in SYP Low regions in SmoM2 (M2), SmoM2-Ptenfl/+ (M2; Ptenfl/+) and SmoM2-Ptenfl/fl(M2; Ptenfl/fl) mice (J), and in SYP High and Low regions of SmoM2-Ptenfl/+ mice (K). n=3 mice/genotype. All statistics were determined using an unpaired t-test comparing SYP High to Low (I, K) or two pairs of genotypes (J).

Loss of Pten leads to cell autonomous changes in SmoM2 GCPs at P12

(A-C) Midline cerebellar sagittal sections of P12 SmoM2, SmoM2-Ptenfl/+and SmoM2-Ptenfl/fl mice stained for GFP and EdU to show the EGL with proliferating mutant GCPs (SmoM2, n=11; SmoM2-Ptenfl/+n=5; SmoM2-Ptenfl/fl n=6). Scale bars indicate 0.5 mm. Dotted line boxes indicate locations of images in (A for G, J, M; B for H, K, N; C for I, L, O). (D) Quantification of the area of the cerebellum on midline sections of SmoM2 (M2), SmoM2-Ptenfl/+ (M2; Ptenfl/+) and SmoM2-Ptenfl/fl (M2; Ptenfl/fl) mice (n=3 mice/genotype). (E) Quantification of the area of the EGL on midline sections in the three genotypes (n=3 mice/genotype). (F) Quantification of TUNEL particle density (cell death) in the EGL of midline sections in the three genotypes (n=3 mice/genotype). (G-O) Staining of sections of the three genotypes for the proteins indicated showing reduced cell death (TUNEL density) and increased pAKT and NeuN in the EGL of SmoM2-Ptenfl/fl mice (n>3 per genotype). Scale bars indicates 50μm. EGL, external granule layer; ML, molecular layer (region between IGL and EGL); IGL, internal granule layer. (P) Schematic of experimental design. (Q) Quantification of the percentage of proliferating GCPs that became post mitotic (Ki67-) between BrdU labeling at P10 and analysis at P12 in a region of lobule 5 (L5) (n=3 mice/genotype). (R) Quantification of the percentage of GCPs that are in S phase (EdU+ Ki67+ of all Ki67+ GCPs) at P12 (n=3 mice/genotype). All statistics were determined using an unpaired t-test comparing two pairs of genotypes.

Loss of Pten leads to cell autonomous changes SmoM2 GCP behaviors at P8

(A-C) Midline cerebellar sagittal sections of P8 SmoM2, SmoM2-Ptenfl/+and SmoM2-Ptenfl/fl mice stained for GFP and EdU to show the EGL with proliferating mutant GCPs. Dotted line boxes indicate locations of where quantifications were performed. Scale bars indicate 0.5 mm. (A’-C’) High magnification images of areas indicated in (A-C). Scale bars indicate 0.5μm. (D) Quantification of the area of the cerebellum on midline sections of SmoM2 (M2), SmoM2-Ptenfl/+ (M2; Ptenfl/+) and SmoM2-Ptenfl/fl(M2; Ptenfl/fl) mice (n=3 mice/genotype). (E) Quantification of the area of the external granule layer (EGL) on midline sections in the three genotypes (n=3 mice/genotype). (F) Quantification of the proliferation index in lobule 5 (L5) at P8: percentage of mutant GCPs that are in S phase (GFP+ EdU+ of all GFP+ GCPs), (n=4 mice/genotype). (G) Quantification of the percentage of mutant GCPs that remained in the proliferative outer EGL (oEGL) at P8 and thus were progenitors. Analysis performed in wall of lobule 5 (lower rectangles in (A-C) (n=4 mice/genotype). All statistics were determined using an unpaired t-test comparing two pairs of genotypes.

Loss of Pten in normal GCPs causes increased proliferation and decreased differentiation

(A) Schematic showing mosaic mutant analysis approach to study scattered GFP+ Pten homozygous mutant GCPs in Atoh1-FlpoER/+; R26FSF-GFPcre/+; Ptenfl/fl (Atoh1-M-Ptenfl/fl) mice compared to control GFP+ cells in Atoh1-FlpoER/+; R26FSF-GFPcre/+; Pten+/+ mice (Atoh1-M-Pten+/+) at P8 following tamoxifen at P0. (B, C) Midline sagital sections of lobule 5 stained with GFP marking mutant cells and EdU (S phase) at P8 in the two genotypes. Scale bars indicates 100μm. (D) Quantification of the proliferation index in lobule 5 (L5) at P8: percentage of mutant GCPs that are in S phase (GFP+ EdU+ of all GFP+ GCPs; n=3 mice per genotype). (E) Quantification of the percentage of mutant GFP+ GCPs that remained in the proliferative outer EGL (oEGL) at P8 and thus were progenitors (n=3 mice per genotype). (F-M) Midline sagittal H&E stained sections of the cerebellum of conditional mutant mice lacking Pten in the GCP lineage (Atoh1-Cre; Ptenfl/fl) compared to controls (Atoh1-Cre; Ptenfl/+). Scale bars indicates 500μm. (N) Quantification of the area of midline sagittal sections of the cerebellum across development and aging (n=3 mice per age). All statistics were determined using an unpaired t-test comparing two pairs of genotypes (D, E) or two genotypes at each age (N).

Loss of Pten in SmoM2 GCPs results in a cell non-autonomous decrease in macrophages in SHH-MB tumors

(A, B) Sagittal sections of end stage tumors in SmoM2 and SmoM2-Ptenfl/fl mice stained with IBA1 (macrophage marker) and DAPI (n=3 mice/genotype). Dashed lines outline the tumors, solid lines outline IGL. Dotted line squares indicate locations of high magnification images show in (C-H). Scale bars indicate 1mm. (C-H) Higher magnification images of tumors stained with IAB1 and TUNEL (C, D), IAB1, NeuN and Ki67 showing enrichment of macrophages in undifferentiated regions of SmoM2-Ptenfl/fl mice (E, F) and with CD31 to label blood vessels (G, H) (n=5-9 mice/genotype). IGL, internal granule layer. Scale bars indicates 100μm. (I, J) Lateral sagittal cerebellar sections of P12 SmoM2 and SmoM2-Ptenfl/fl mice stained for GFP and DAPI to show the EGL with mutant GCPs (n=3 mice/genotype). Dotted line boxes indicate locations of higher magnifications images in (K-P). Scale bars indicate 500μm. (K-P) Sections of P12 mice stained with IAB1, Ki67 and NeuN showing macrophages in large tumor lesions are enriched in proliferative regions of SmoM2-Ptenfl/fl mice (n=3 mice/genotype). Dotted line outlines the tumor lesion. Scale bars indicate 200μm.

scRNA-seq analysis of tumors from SmoM2 and SmoM2-Ptenfl/flmice reveals altered differentiation in SHH-MB tumor cells lacking Pten

(A) UMAP of integrated cells from SmoM2 (n=2) and SmoM2-Ptenfl/fl (n=2) tumors showing scRNA-seq GCP-lineage clusters 0-9 (Pax6-expressing) (left) and the assignment of cell types to each cluster (right) based on marker gene expression and cell cycle phase. (B) Pax6 expression (GCP-lineage marker) shown across all clusters in UMAP of all tumor cells. (C) UMAP of GCP-lineage cells shown by genotye. (D) UMAP of GCP-lineage cells (two genotypes) showing cell cycle phases. (E) Dot plot graph showing expression levels of GCP-lineage marker genes and desgnation of the clusters as GCP-like or maturing GC-like cell types. (F) Graph of GCP-lineage cells showing percentage of cells from each genotype that are GCP-like, transitioning GC-like, early GC-like and mature GC-like showing a higher percentage of mature GCs in tumors from SmoM2 mice. (G) Graph showing proportions of cells of each genotype in each of the GCP-lineage cell types excuding mature GCs that are mostly from SmoM2 mice. Dotted line represents the expected ratio between the two genotypes. (H) Pseudotime analysis reflecting the differentiation pathway of cells within the GCP-lineage with an origin in the GCP-like clusters, indicating cluster 7 is stalled or altered during differentiation.

SmoM2 tumors lacking Pten have altered neurogenesis and immune signaling

(A) Volcano plot showing differentially expressed genes in GCP-like tumor cell clusters (3, 4, 8) (red: adjusted p-value <0.05, ab[log2fold-change]>0.5), in SmoM2-Ptenfl/fl tumors compared to SmoM2. (B) Gene set enrichment analysis (GSEA) plots (Hallmark terms) showing that mTORC1 signaling pathway is significantly unregulated in SmoM2-Ptenfl/fl tumors compared to SmoM2. (C) Violin plots showing expression levels of mTORC1 signaling pathway genes in the two genotypes. (D) GSEA plots (Hallmark terms) showing that Interferon alpha response is significantly downregulated in SmoM2-Ptenfl/fl tumors compared to SmoM2. (E) Violin plots showing expression levels of Interferon alpha response genes in the two genotypes. (F) GSEA plots of Gene ontology (GO) terms (biological processes) showing that regulation of neural differentiation is significantly upregulated in SmoM2-Ptenfl/fl tumors compared to SmoM2. (G) Violin plots showing expression levels of neural differentiation genes in the two genotypes, indicating an early stage of neurogenesis is enhanced. (H) Volcano plot showing differentially expressed genes in early GC-like cluster 1 compared to GCP-like cell clusters (3, 4, 8) in SmoM2-Ptenfl/fl tumors (red: adjusted p-value <0.05, ab[log2fold-change]>0.5). (I) GSEA plots of Gene ontology (GO) terms (biological processes) confirmed upregulation of neural differentiation in cluster 1 compared to the GCP-like cell clusters. (J) Violin plots showing expression levels of neural differentiation genes in the two cell types. (K) GSEA plots of Gene ontology (GO) terms (biological processes) showed downregulation of ‘positive regulation of programmed cell death’ in cluster 1 compared to the GCP-like cell clusters. (L) Violin plots showing expression levels of programmed cell death genes in the two cell types. (M-R) Lateral sagittal sections of tumors from SmoM2 mice (P22, M, O, Q) or SmoM2-Ptenfl/fl mice (P16, N, P, R) stained for RNA in situ analysis of Cdkn1a, Vldr, Asns, indicating upregulation of mTORC1 signaling pathway in Pten mutant tumors.

Loss of Pten does not increase metastasis to the spinal cord in SmoM2 SHH-MB

(A-C) Coronal sections of spinal cords stained with H&E from mice of the indicated three genotypes showing metastatic lesions at end stage. Dashed rectangles indicate locations of images below stained for the proteins and DAPI indicated on the left. Scale bars indicates 200μm. (D) Quantification of percentage of sections with a tumor for each animal analyzed. (E) Quantification of the total area of the tumor on all the sections. (F) Quantification of the area of each tumor as a percentage of the spinal cord area in each section. (D-F) Number of animals per genotype (n=) are indicated below graphs. All statistics were determined using an unpaired t-test comparing SmoM2 tumors to each of the Pten mutants or the Pten mutants to each other.

Spontaneous loss of the second allele of Pten in rare cells of some SmoM2-Ptenfl/+ tumors

(A-D) Sagittal sections of end stage tumors (GFP+) showing upregulation of pS6 in SmoM2-Ptenfl/fl mice compared to SmoM2 (n=3 per genotype). Dashed lines outline the tumors, solid lines outline IGL. Scale bars indicates 1mm. (E) Section of end stage tumor from SmoM2-Ptenfl/+ mice showing coherent patches of tumor cells (GFP+) devoid of PTEN (n=3). Dashed line square indicates location of high magnification images shown in (F-L). Scale bar indicates 1mm. (F-L) Sections of lateral tumors stained for the indicated protein and DAPI (n=3). Dashed line outlines tumor area where PTEN is absent (G) or pAKT (H), SYP (J) and NeuN (K) are upregulated and proliferation (Ki67) and cell death (TUNEL) are reduced (I, L). Scale bars indicates 200μm.

Method of quantifying TUNEL particles in tumors

(A-C) Examples of stained sagittal cerebellar sections showing where quantifications of density of TUNEL particles were performed in the three genotypes in end stage lateral tumor regions distinguishing high SYP (H) compared to low SYP (L) regions. Dashed lines outline each tumor, solid lines outline IGL, dotted boxes indicate high SYP (H) and low SYP (L) regions. Scale bars indicate 1mm.

Method of quantifying TUNEL particles in P12 EGL

(A, B) Examples of stained midline sagittal cerebellar sections showing where quantification of density of TUNEL particles was performed in the EGL in medial regions of the cerebellum. Scale bars indicate 1mm.

The cellular phenotypes of large tumors from P16 SmoM2-Ptenfl/fl mice and P22 SmoM2 mice are similar to end stage

(A-C) Sagittal sections of cerebellar sections from P16 and P22 SmoM2 mice and P16 SmoM2-Ptenfl/flmice stained with GFP and DAPI to highlight tumor cells (n=3 mice/genotype). Dashed line squares indicate location of high magnification images shown in (D-O). Scale bars indicate 1mm. (D-O) Higher magnification images of tumors stained for indicated proteins to confirm that the cellular phenotypes of large tumors in P16 SmoM2-Ptenfl/fl mice and P22 SmoM2 mice are like tumors at end stage. Dashed lines outline the internal granule layer (IGL). Scale bars indicate 200μm.

ScRNA-seq comparison of tumors from P16 SmoM2-Ptenfl/fl mice to P22 SmoM2 mice

(A) Number of cells from each replicate and genotype used for downstream analyses after filtering. (B) Violin plots showing the number of features, and RNA counts and percent mitochondrial RNA counts across the biological replicates of the scRNA-seq data set after filtering out poor quality cells (where number of detected genes was ≤ 500, the number of detected transcripts was ≥ 18,000 and mitochondrial gene percentage ≥ 20%). (C) PCA plot after filtering out poor quality cells and integrating the replicates and genotypes showing the two genotypes separate well. (D) UMAP showing projections of all cells showing cluster annotations. (E) Dot plot graph showing expression levels of cell type marker genes across all clusters and their cell type desgnations. (F) UMAPs of all cells shown by genotype and sample. (G) Graph showing proportions of cells of each genotype when all cells are included. Dotted line represents the expected ratio between the two genotypes.

Differential expression analyses confirms activation of mTOR

A. Bar plots showing top 10 upregulated and downregulated Hallmark terms in SmoM2-Ptenfl/fl GCP-like cells compared to SmoM2. B. Bar plots showing top 10 upregulated and downregulated Gene ontology (GO) terms (biological processes) in GCP-like cells from SmoM2-Ptenfl/fl mice compared to SmoM2. C. Bar plots showing top 10 upregulated and downregulated Gene ontology (GO) terms (biological processes) in early GC-like cluster 1 compared to GCP-like cell clusters (3, 4, 8) in SmoM2-Ptenfl/fl tumors. D. Bar plots showing the top 10 upregulated and downregulated Gene ontology (GO) terms (biological processes) in macrophages in SmoM2-Ptenfl/fltumors compared to SmoM2. E. Violin plots showing expression levels of Cd36 and Irf2 in macrophages in the two tumor genotypes.