R-Spondin chromosome rearrangements drive Wnt-dependent tumour initiation and maintenance in the intestine

Abstract

Defining the genetic drivers of cancer progression is a key in understanding disease biology and developing effective targeted therapies. Chromosome rearrangements are a common feature of human malignancies, but whether they represent bona fide cancer drivers and therapeutically actionable targets, requires functional testing. Here, we describe the generation of transgenic, inducible CRISPR-based mouse systems to engineer and study recurrent colon cancer-associated EIF3E-RSPO2 and PTPRK-RSPO3 chromosome rearrangements in vivo. We show that both Rspo2 and Rspo3 fusion events are sufficient to initiate hyperplasia and tumour development in vivo, without additional cooperating genetic events. Rspo-fusion tumours are entirely Wnt-dependent, as treatment with an inhibitor of Wnt secretion, LGK974, drives rapid tumour clearance from the intestinal mucosa without effects on normal intestinal crypts. Altogether, our study provides direct evidence that endogenous Rspo2 and Rspo3 chromosome rearrangements can initiate and maintain tumour development, and indicate a viable therapeutic window for LGK974 treatment of RSPO-fusion cancers.

Figure 1. Induction of EIF3E–RSPO2 and PTPRK–RSPO3 fusions using inducible CRISPR.

(a) Schematic representation of chromosomal rearrangements involving Rspo2 and Eif3e (left), and Ptprk and Rspo3 (right). (b) Dox-inducible lentiviral vector (upper). Paired U6-sgRNAs are cloned into the vector upstream of TRE^3G promoter. Detection of EIF3E–RSPO2 (E-Rspo2) and PTPRK–RSPO3 (P-Rspo3) genomic rearrangements by fusion-specific PCR, on genomic DNA extracted from puromycin-selected 3T3 cells at multiple time points (lower). (c) Detection of the E-Rspo2 and P-Rspo3 fusion transcripts using fusion-specific PCR primers on cDNA from day 4 dox-treated 3T3 cells in b.

Figure 2. Induction of E-Rspo2 and P-Rspo3 fusions in transgenic iCRISPR ESCs.

(a) Schematic of iCRISPR col1a1-targeting vector. (b) Detection of the E-Rspo2 and P-Rspo3 fusion transcripts using fusion-specific PCR primers on cDNA from targeted ESC clones, 4 days post doxycycline treatment (upper). Sanger sequencing of the PCR product shows the expected splice junctions in expressed RNA transcripts (lower). (c) Details of the DNA FISH strategies to detect the E-Rspo2 deletion, and P-Rspo3 inversion (upper). Multi-colour DNA FISH on WT and isogenic E-Rspo2 or P-Rspo3 ESC clones. Rearranged alleles are highlighted with white arrows (lower).

Figure 3. Intestinal organoids carrying the Ptprk–Rspo3 fusion are RSPO1-independent.

(a) Detection of E-Rspo2 and P-Rspo3 rearrangements using fusion-specific PCR primers on genomic DNA extracted from c3GIC9-E-Rspo2 and c3GIC9-P-Rspo3 mouse small intestinal organoids, 7 days post doxycycline treatment. (b) Bright-field images of dox-naive and dox-treated P-Rspo3 organoids cultured in ENR or EN medium, as indicated. Scale bar, 50 μm. (c) Confocal immunofluorescent images of dox-naive, P-Rspo3 and Apc-deleted organoids cultured in ENR and EN medium, as indicated, showing markers of proliferation (EdU), differentiation (K20 and alkaline phosphatase activity) and Paneth cells (lysozyme). Scale bar, 50 μm. (d) Graphs represent qRT-PCR results of Rspo3, Ptprk, K20 and Wnt target genes (Axin2, Lgr5 and Ascl2) on WT, P-Rspo3 and Apc-deleted organoids (n≥4, bars represent mean values +/−s.d., *P<0.05, **P<0.01, ***P<0.001, two-sided t-test with Welch correction). (e) Schematic of culture conditions of WT, P-Rspo3 and Apc-deleted organoids for RNAseq (upper). Heatmap indicates up (yellow) and downregulated (blue) transcripts (log2FC≥1) in P-Rspo3 organoids (grey) compared to WT/naive cultures (green) (lower).

Figure 4. Rspo rearrangements initiate tumour growth in vivo.

(a) Immunohistochemical (H&E, alkaline phosphatase and lysozyme), and immunofluorescent (Ki67 and keratin 20) stains of intestinal sections from control sgRNA (CR8), Apc sgRNA and P-Rspo3 animals, treated with dox (200 mg kg⁻¹) for 10 days, and collected at indicated times. Scale bars, 100 μm. (b) DNA FISH staining on intestinal sections from R26-rtTA/c3GIC9-P-Rspo3 animals. Immune cells within Peyers patches are shown as a normal control. P-Rspo3 adenomas are enriched for the P-Rspo3 rearrangement, highlighted with white arrows. (c) Graphs represent fraction of genomes containing P-Rspo3 inversions (upper) and mRNA expression level of Rspo3 (lower) in mouse small intestine following dox treatments for indicated times (n≥4, bars represent mean values+/−s.d., ***P<0.001, ****P<0.0001, two-sided t-test with Welch correction).

Figure 5. P-Rspo3 tumours are molecularly distinct from Apc-mutant tumours.

(a) Graphs represent expression (RPKM value) of Sox17, Sox9 and Axin2 on WT, P-Rspo3 and Apc-deleted organoids (n≥4, bars represent mean values+/−s.d., **P<0.01, ***P<0.001, ****P<0.0001, two-sided t-test with Welch correction). (b) Immunohistochemical staining of Sox17, Sox9 and Axin2 on small intestinal sections from dox-treated CR8, Apc sgRNA and P-Rspo3 animals. Scale bars are labelled in the figure.

Figure 6. Rspo rearranged tumours are sensitive to Porcn inhibition.

(a) Bright-field images of WT, P-Rspo3, and Apc-deleted organoids treated with DMSO, C59 (500 nM) or LGK974 (500 nM) for 4 days. Scale bars, 50 μm. (b) Immunofluorescent images of P-Rspo3 and Apc-deleted organoids treated with DMSO, C59 (500 nM), and LGK974 (500 nM) for 4 days. EdU (red) was stained for proliferation. Scale bars, 50 μm. (c) Schematic of the in vivo LGK974 treatment experiment. (c) Immunohistochemical images of intestinal sections from E-Rspo2 and R26-rtTA/P-Rspo3 mice treated with either DMSO or LGK974. BrdU was stained for proliferation. Scale bars, 100 μm. (d) Representative immunohistochemical images of proximal small intestine from P-Rspo3 mice treated with either DMSO or LGK974. (e) Quantification of hyperproliferative and histologically abnormal regions in the LGK974 treatment experiment (n≥4, bars represent mean values+/−s.d., **P<0.01, ***P<0.001, two-sided t-test with Welch correction). (f) Schematic representation of sequential gene editing in intestinal organoids to create BR3 and BRPS cultures. Red indicates proliferative cells. (g) Bright-field and immunofluorescent images of DMSO or LGK974 treated, BR3 and BRPS organoids, as indicated. In both genotypes, 4 days of LGK treatment induce cell cycle arrest and drives differentiation. Scale bars, 50 μm.