Fig. 1

 CLSTN1 expression is reduced in primary MB tumor samples. (A) Differential protein expression in the plasma membrane-associated proteome between sgCTRL and sgMAP4K4 cells assessed by spatial mass-spectrometry analysis identified CLSTN1 as one of several specifically upregulated proteins9. The mean Z-score and SD of up-regulated proteins in sgMAP4K4 cells are shown. N = 3–4 biological replicas. P values of multiple unpaired T-tests are shown when P = < 0.05. (B) mRNA expression analysis in three controls (diseased brain, normal brain regions (white), and cerebellum (blue)), five primary MB (orange), one ATRT, and one ependymoma dataset (grey). Significant differences between CLSTN expression in tumor samples compared to cerebellum controls were calculated using ANOVA multiple comparisons and Kruskal-Wallis test and are indicated by asterisks: * p = < 0.05, ** p = < 0.01, **** p = < 0.0001. (C) CLSTN mRNA expression analysis across the 12 MB subtypes. ANOVA multiple comparisons and Kruskal-Wallis test were used. * p = < 0.05, ** p = < 0.01, *** p = < 0.001, **** p = < 0.0001. (D) XY plots showing the correlation between CLSTN and MAP4K4 expression across the four MB subgroups WNT, SHH, Grp3 and Grp4. Interrupted line indicates best fit of linear regression analysis. The slope and p-value of the slope are indicated. (E) Analysis of CLSTN1 expression across human organ development. (F, G) snRNAseq analysis of CLSTN1 (F) and MAP4K4 (G) expression across cerebellar cell types. Upper: Manifold Approximation and Projection (UMAP) of 180,956 human cells colored by cell type (from29). Lower: Expression levels of CLSTN1 across projected cell types. CPM, Counts per million; GABA, Gamma-aminobutyric acid, DN, Dentate nucleus; GC, Granule cells; UBC, Unipolar brush cells; Glut_DN, Glutamatergic deep nuclei neurons; Isth_N, Isthmic nuclei neurons; NTZ, Nuclear transitory zone; VZ, Ventricular zone.