Journal: bioRxiv
Article Title: Quantitative dissection of the metastatic cascade at single colony resolution
doi: 10.64898/2026.02.20.706841
Figure Lengend Snippet: CREBBP suppresses SCLC metastasis through regulating Cdx2 expression. a. UMAP of single-nucleus RNA-seq from Crebbp- WT and Crebbp- KO liver metastases showing four Ascl1⁺ SCLC clusters. Crebbp- KO metastases are enriched in cluster 4 SCLC. Left, UMAP for combined samples with all cell type identified. A total of 55,558 nuclei were analyzed. Right-Top, UMAP for SCLC clusters identified in Crebbp- WT sample with Asc1 UMI >1 per cell. Right-Bottom, UMAP for SCLC clusters identified in Crebbp -KO sample with Asc1 UMI >1 per cell. b. Alluvial plot showing cluster composition of Ascl1 ⁺ SCLC cells from Crebbp- WT and Crebbp- KO liver metastases. Crebbp- KO samples are dominated by Cluster 4 SCLC cells, comprising over 70% of the population. c. Cluster 4 SCLC displays MYC signaling activation. Gene set enrichment analysis (GSEA) of Ascl1 ⁺ SCLC clusters, with circle color indicating enrichment z-score and size reflecting the percentage of genes enriched in each pathway. d. Crebbp- KO SCLC cells exhibit increased proliferation and loss of neuroendocrine identity. Violin plots showing pairwise comparisons of S- and G2/M-phase scores and neuronal scores across the four SCLC clusters in Crebbp- WT and Crebbp- KO metastases. e. Schematic of co-transplantation experiment for bulk RNAseq. GFP-labeled Crebbp- WT and mCherry-labeled Crebbp- KO RP48 cells were mixed at a 1:1 ratio and transplanted via tail vein injection. After 3 weeks, liver metastases were dissociated, and GFP⁺ and mCherry⁺ SCLC cells were isolated by FACS for bulk RNA-seq analysis. PCA plot shows a well separation between WT and KO samples. f. Volcano plot showing differentially expressed genes between Crebbp- WT and Crebbp -KO metastatic SCLC cells. A total of 459 genes were upregulated (>2-fold) in Crebbp- KO cells, while 474 genes were upregulated (>2-fold) in Crebbp- WT cells, indicating balanced transcriptional activation and repression upon Crebbp loss. g. GSEA analysis of bulk RNA-seq from sorted Crebbp- WT and Crebbp- KO SCLC cells showing loss of neuroendocrine programs and activation of type I interferon signaling in Crebbp- KO cells. Circle color represents enrichment FDR, and circle size reflects the number of genes contributing to each pathway. h. Integration of H3K27Ac ChIP-seq and RNA-seq identifies Cdx2 as a direct CREBBP target. Differential H3K27Ac peak signals between Crebbp- WT and Crebbp -KO SCLC cells (two biological replicates) were cross-compared with differential gene expression from RNA-seq. Circle color represents –log₁₀FDR, and circle size reflects the total number of differential peaks per gene. Cdx2 is among the most significantly enriched and shared targets in WT cells comparing to KO cells. i. Crebbp loss specifically reduces H3K27Ac modification at the Cdx2 locus. H3K27Ac ChIP-seq tracks from Crebbp- WT and Crebbp- KO SCLC cells are shown for the indicated chromosome 5 region. Crebbp- KO cells display a selective loss of H3K27Ac at the Cdx2 locus, while neighboring gene loci remain unaffected. j. Crebbp l oss diminished CDX2 protein expression in SCLC cells. Western blot analysis of CREBBP, CDX2, and H3K27Ac in Crebbp- WT and Crebbp- KO RP48 cells. Total H3 and GAPDH are blotted as control. k. Doxycycline-induced CDX2 re-expression in Crebbp- KO RP48 cells suppresses proliferation and growth advantage in co-culture with Crebbp- WT cells. Crebbp- WT-TetON-CDX2 and Crebbp- KO-TetON-CDX2 RP48 cells were mixed at a 1:1 ratio and treated with 0, 100, or 500 ng/mL doxycycline in co-culture for 6 days. GFP/mCherry cell ratios were quantified by flow cytometry. n= 7 wells l. Schematic of the MOBA-seq screen using control or CDX2-overexpressing RP48 cells. Lentiviral sgRNA libraries were constructed using control (PuroR) or CDX2-overexpression (PuroR-2A-CDX2) backbones. 4N-barcoded control and CDX2-overexpressing RP48 cell libraries were mixed at a 1:1 ratio and transplanted via tail vein into five NGS mice. Three weeks post-transplantation, metastasis-bearing livers and lungs were collected, and over 20,000 barcoded metastases were identified from both tissues for MOBA-seq analysis. m. Crebbp-mediated metastatic suppression in the liver is CDX2-dependent. MOBA-seq analysis deconvoluted control and CDX2-overexpressing RP48 cells using their 4N-static barcodes. Genotype-specific metastatic burden was calculated as fold change relative to the mean of all control colonies ( sgCtrl ). The blue regression line with confidence interval represents genotypes unaffected by CDX2 expression, while those preferentially suppressed under CDX2-overexpressing conditions are highlighted in yellow. Crebbp inactivation increased liver metastatic burden only in control RP48 cells, but not in CDX2-overexpressing RP48 cells. n. Crebbp -mediated metastatic suppression in the lung is CDX2-dependent. MOBA-seq analysis of genotype-specific metastatic burden in control and CDX2-overexpressing RP48 cells that metastasized to the lung as described in m. Crebbp inactivation markedly increased lung metastatic burden in control RP48 cells but had a reduced effect in CDX2-overexpressing RP48 cells.
Article Snippet: Antibodies used in this study: GAPDH (Cell Signaling, 5174S, 1:2000), CREBBP (Cell Signaling, 7389S, 1:200), H3 (Cell Signaling, 4499S, 1:1000), CDX2 (Invitrogen, MA5-35215, 1:2000), H3K27Ac (Cell Signaling, 8173S, 1:1000), Goat-anti-Rabbit IgG Antibody, HRP-conjugate (Sigma-Aldrich, 12-348), Goat-anti-Mouse IgG Antibody, HRP-conjugate (Thermo Fisher Scientific, 62-6520).
Techniques: Expressing, RNA Sequencing, Activation Assay, Transplantation Assay, RNA sequencing, Labeling, Injection, Isolation, ChIP-sequencing, Gene Expression, Modification, Western Blot, Control, Co-Culture Assay, Flow Cytometry, Construct, Over Expression
Cell Signaling Technology Inc is a verified supplier 
Cell Signaling Technology Inc manufactures this product 
