BRAFV600E drives dedifferentiation in small intestinal and colonic organoids and cooperates with mutant p53 and Apc loss in transformation

Abstract

BRAF^V600E confers poor prognosis and is associated with a distinct subtype of colorectal cancer (CRC). Little is known, however, about the genetic events driving the initiation and progression of BRAF^V600E mutant CRCs. Recent genetic analyses of CRCs indicate that BRAF^V600E often coexists with alterations in the WNT- and p53 pathways, but their cooperation remains ill-defined. Therefore, we systematically compared small and large intestinal organoids from mice harboring conditional Braf^floxV600E, Trp53^LSL-R172H, and/or Apc^flox/flox alleles. Using these isogenic models, we observe tissue-specific differences toward sudden BRAF^V600E expression, which can be attributed to different ERK-pathway ground states in small and large intestinal crypts. BRAF^V600E alone causes transient proliferation and suppresses epithelial organization, followed by organoid disintegration. Moreover, BRAF^V600E induces a fetal-like dedifferentiation transcriptional program in colonic organoids, which resembles human BRAF^V600E-driven CRC. Co-expression of p53^R172H delays organoid disintegration, confers anchorage-independent growth, and induces invasive properties. Interestingly, p53^R172H cooperates with BRAF^V600E to modulate the abundance of transcripts linked to carcinogenesis, in particular within colonic organoids. Remarkably, WNT-pathway activation by Apc deletion fully protects organoids against BRAF^V600E-induced disintegration and confers growth/niche factor independence. Still, Apc-deficient BRAF^V600E-mutant organoids remain sensitive toward the MEK inhibitor trametinib, albeit p53^R172H confers partial resistance against this clinically relevant compound. In summary, our systematic comparison of the response of small and large intestinal organoids to oncogenic alterations suggests colonic organoids to be better suited to model the human situation. In addition, our work on BRAF-, p53-, and WNT-pathway mutations provides new insights into their cooperation and for the design of targeted therapies.

Fig. 1. RNA-Seq of Braf^V600E/+ (VE) and Trp53^R172H/+ (p53) single and Braf^V600E/+,Trp53^R172H/+ (VE,p53) double-mutant SI and COL organoids (GSE132551).

a Heatmap of the log2 fold changes (induced vs. non-induced, color coded) of the genes listed in the HALLMARK_P53_PATHWAY MSigDB gene set. b Heatmap of selected differentially regulated transcripts. Color code represents the log2 fold change (induced vs. non-induced). Asterisks marked genes have already been found to be among the top 50 of differentially regulated genes in human CRC cell line spheroids upon BRAF^V600E inhibition ([10]; blue = downregulated; red = upregulated). c Western blots (WB) of SI and COL organoids using the indicated antibodies. GAPDH and HSP90 serve as loading controls. Each subpanel identified by its pERK detection reflects a distinct biological experiment. d MUC2 IF staining of formalin-fixed paraffin-embedded (FFPE) SI and COL organoid sections shows enhanced mucin production within mutant organoids, with highest levels in double-mutant ones. Scale bars: 50 µm.

Fig. 2. Primary small and large intestine exhibit differences in basal ERK-signaling activity (GSE132546).

a PCA of RNA-Seq of freshly isolated small intestinal (SI, circles) and colonic (COL, triangles) crypts from two female (no. 1, 4) and two male (no. 2, 3) donor mice. SI and COL show a clear separation according to PC1. b Heatmap of transcripts encoding ERK-pathway components shows differential expression between SI and COL crypts freshly isolated from the four donor mice described in a. Color code represents the row-wise scaled (Z score) RNA intensity. c Simplified cartoon visualizing ERK-pathway components of particular interest. d WB of SI and COL crypts with the indicated antibodies. E-cadherin serves as loading a control. Of note, RNA-Seq revealed that other loading controls, such as GAPDH or beta-actin were differentially expressed between SI and COL crypts. Only E-cadherin displayed a negligible log2 fold change of 0.004. e Quantification of WB analyses of pERK and DUSP6. ERK phosphorylation was normalized to total ERK expression, and DUSP6 was normalized to internal loading control. Green/brown colors indicate whole-tissue lysates, while blue colors indicate isolated crypt lysates. Each color depicts one donor mouse. Data are presented as mean ± SD. **P ≤ 0.01 (paired t test, n ≥ 3 donor mice).

Fig. 3. BRAF^V600E induces a fetal signature in COL organoids.

a Gene set enrichment analysis (GSEA) of our transcriptomic data with genes that are up- (left) and downregulated (right), respectively, in Cdx1/Cdx2 double-knockout (DKO) mice (GSE24633). b GSEA of our transcriptomic data against genes that are up- (left) and downregulated (right), respectively, in mouse fetal intestinal spheroids. c GSEAs showing the comparison of the LGR5-independent fetal signature to human data sets. The left panel shows the enrichment of the fetal spheroids signature in TCGA BRAF^V600E mutant vs. healthy colon. The right panel shows the enrichment of the Popovici signature, i.e., 314 differentially expressed probe sets between WT and BRAFm samples, in fetal spheroids. Note, a positive fold change of the “Popovici genes” indicates higher expression in WT vs BRAFm, which results in an inverse correlation to the fetal signature. NES normalized enrichment score, PV P-value.

Fig. 4. Trp53^R172H supports proliferative and invasive characteristics of Braf^V600E mutant organoids.

a, b GSEA of the delta log2 fold changes of double-mutant vs. BRAF^V600E-only organoids was performed. Shown are enrichment heatmaps for p53-related gene sets (a) and for the top ten significantly (P < 0.05) regulated chemical and genetic perturbations (CGP) (b). On both heatmaps, color code and circle size represent NES. c Representative bright-field (BF) images of three independent experiments show colony growth capacity of COL organoids. Quantification is shown in (d), d Quantification of colony growth capacity, normalized to the corresponding non-induced control. The longest straight lines of the crypts were measured. e Control or 4-HT-induced COL crypts were disaggregated, and grown on PolyHEMA-coated culture dishes for 6 days before BF images were taken and the diameters of the formed cell clusters were measured. Representative BF pictures of ≥3 independent experiments that are quantified in (f), are shown. f Quantification of anchorage-independent growth. The longest straight lines of the cell clusters were measured. Note that neither non-induced controls nor p53^R172H-mutant organoids were able to form cell clusters on PolyHEMA. g Control or 4-HT-induced COL organoids were grown in diluted (50%) Matrigel. BF images were taken at day 7, and organoids attached to the plastic surface were counted (highlighted by dashed lines). Representative pictures of ≥3 independent experiments are shown, which are quantified in (h). Higher-magnification BF images of “invaded” organoids are shown in Supplementary Fig. S7. h Quantification of “invaded” organoids. In (c, e, g), scale bars: 50 µm. In (d, f, h), symbol colors refer to donor mice, symbol shapes refer to independent experiments. Data are presented as mean ± SD, and statistical significance was determined by one-way ANOVA (corrected for multiple comparison by Bonferroni). *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001.

Fig. 5. Loss of APC rescues the BRAF^V600E-driven disintegrative phenotype and confers growth factor independence.

a COL organoids with the indicted genotypes were treated with 3 µM 4-HT for 24 h. Representative microscopy pictures at days 2 and 6 (see also Supplementary Video 3) and MTT staining at day 7 are shown. b COL organoids with the indicated genotypes were induced with 3 µM 4-HT and cultured without the growth factors (GFs) EGF, R-Spondin, Noggin, and Wnt3a. BF images and MTT staining at day 9 are shown. c COL organoids with the indicated genotypes were treated with DMSO or indicated trametinib concentrations 1 day after induction with 3 µM 4-HT. Representative MTT staining at day 9 after 4-HT induction is shown. Note that the dark-blue colonies indicate metabolic activity. d Quantification of trametinib treatment shown in (c) of two (for Apc^Δ/Δ) and three (for Braf^V600E/+,Apc^Δ/Δ and Braf^V600E/+,Apc^Δ/Δ,Trp53^R172H/+) independent experiments. Colony count was normalized to corresponding DMSO control, and statistical significance was determined by two-way ANOVA (corrected for multiple comparison by Bonferroni). *P ≤ 0.05; **P ≤ 0.01. In (a, b), scale bars: 50 µm.

Fig. 6. Trp53^R172H supports proliferative and invasive characteristics of Braf^V600E/+,Apc^Δ/Δ mutant organoids.

a Representative BF images of COL organoid colony growth capacity of ≥3 independent experiments are shown, which are quantified in (b). b Quantification of colony growth assay, normalized to corresponding non-induced control. The longest straight lines of the crypts were measured. c COL organoids with the indicated genotypes were grown in diluted (50%) Matrigel. BF images were taken at day 7, and organoids attached to the plastic surface were counted (highlighted by dashed lines). Representative pictures of three independent experiments are shown, and are quantified in (d). d Quantification of “invaded” organoids. In (a, c), scale bars: 50 µm. In (b, d), symbol colors refer to donor mice, symbol shapes refer to independent experiments. Data are presented as mean ± SD, and statistical significance was determined by one-way ANOVA (corrected for multiple comparison by Bonferroni). *P ≤ 0.05.

Fig. 7. Graphical illustration summarizing the model of the multistep colorectal carcinogenesis investigated in this study.

From left to right: Expression of oncogenic BRAF^V600E induces a fetal gene signature in adult wildtype colonic organoids, but also leads to rapid disintegration and subsequent cell death. Co-expression of p53^R172H extends organoid survival and conveys proliferative and invasive properties. Additional loss of APC prevents the collapse of the intestinal stem cell (ISC) niche, thereby promoting the survival of the mutant organoids. Importantly, additional APC loss confers growth factor independence and modulates the sensitivity to MEK inhibitor (MEKi) treatment. AJ adherens junctions, TJ tight junctions.