Colonic epithelial adaptation to EGFR-independent growth induces chromosomal instability and is accelerated by prior injury

Abstract

Although much is known about the gene mutations required to drive colorectal cancer (CRC) initiation, the tissue-specific selective microenvironments in which neoplasia arises remains less characterized. Here, we determined whether modulation of intestinal stem cell niche morphogens alone can exert a neoplasia-relevant selective pressure on normal colonic epithelium. Using adult stem cell-derived murine colonic epithelial organoids (colonoids), we employed a strategy of sustained withdrawal of epidermal growth factor (EGF) and epidermal growth factor receptor (EGFR) inhibition to select for and expand survivors. EGFR-signaling-independent (iEGFR) colonoids emerged over rounds of selection and expansion. Colonoids derived from a mouse model of chronic mucosal injury showed an enhanced ability to adapt to EGFR inhibition. Whole-exome and transcriptomic analyses of iEGFR colonoids demonstrated acquisition of deleterious mutations and altered expression of genes implicated in EGF signaling, pyroptosis, and CRC. iEGFR colonoids acquired dysplasia-associated cytomorphologic changes, an increased proliferative rate, and the ability to survive independently of other required niche factors. These changes were accompanied by emergence of aneuploidy and chromosomal instability; further, the observed mitotic segregation errors were significantly associated with loss of interkinetic nuclear migration, a fundamental and dynamic process underlying intestinal epithelial homeostasis. This study provides key evidence that chromosomal instability and other phenotypes associated with neoplasia can be induced ex vivo via adaptation to EGF withdrawal in normal and stably euploid colonic epithelium, without introducing cancer-associated driver mutations. In addition, prior mucosal injury accelerates this evolutionary process.

Keywords: Chromosomal instability; Colitis; Colorectal cancer; Intestinal organoids; Transformation.

Fig. 1

Colonoids adapt to culture conditions devoid of critical niche factors. (A) Survival rate depicting viability of colonoids at the end of each selection cycle (7 d) as a percentage of colonoid growth in control media (n= 6 biological replicates). (B) Survival rate of colonoid lines in other selective media after 7 d (n=3 biological replicates), **** P < 0.0001, 2-tailed non-paired student t-test.

Fig. 2

Long-term adaptation of colonoids to EGF-deficient conditions results in transcriptional changes. (A) Samples analyzed by RNA-seq were plotted by principal component 1 (PC1) and principal component 2 (PC2) using raw count data following regularized logarithm transformation. Samples from the same experimental condition were grouped with the same colors. (B) Volcano plots displaying log₂-transformed fold change and -log₁₀-transformed p value of genes assessed by RNA-seq in iEGFR vs. control colonoids and DSS iEGFR vs. DSS control colonoids. Selected differentially expressed genes are highlighted. Genes highlighted in green are differentially expressed in both iEGFR and DSS iEGFR compared to the respective control. Genes highlighted in red and blue are differentially expressed only in iEGFR or DSS iEGFR, respectively. (C) Gene set enrichment analysis of overlapping upregulated and downregulated genes in both iEGFR and DSS iEGFR compared to the respective control. All enriched gene sets (P value < 0.05) are shown. (D) Quantitative RT-PCR validation of select upregulated (left) and downregulated (right) genes detected by RNA-seq. Results are expressed as log₂ fold change to control and DSS control (n=3).

Fig. 3

iEGFR colonoids show morphologic features of dysplasia and increased proliferation. (A) Representative H&E stained human tissues (upper panels, 20X) and cultured colonoids (lower panels, 40X). Squares denote nuclear hyperchromasia and loss of nuclear polarity, arrowheads denote architectural complexity, asterisks denote squamoid features, and the triangle denotes overall normal epithelial morphology. (B) Representative confocal maximal Z-stacks images for colonoids stained with the thymidine analogue EdU (green) and the counterstain DAPI (blue). No-EdU and Nutlin served as negative and positive controls, respectively. Scale bars = 50 µm. n = 3 independent experiments. (C) Box and whiskers plot for the percentage of EdU-positive nuclei per EdU-positive colonoid. Transverse lines represent the median, boxes show 25th−75th percentile and the whiskers represent the lowest and highest values within 1.5 times the interquartile range. **** P < 0.0001, *P ≤ 0.05; 2-tailed, non-paired student t-test. (D) Bar plot for the percentage of colonoids with at least one EdU-positive cell (n= 3 biological replicates). Error bars represent standard deviation. ****P < 0.0001, *P ≤ 0.05; 2-tailed, non-paired student t-test.

Fig. 4

iEGFR colonoids are aneuploid and demonstrate chromosomal instability. (A) Dot plot of the number of chromosomes in metaphase spreads. The number of counted spreads and the percentage of metaphase spreads with euploid chromosomes are shown at the top. The red line represents tetraploidy. iEGFR(H) and iEGFR(L) correspond to high passage number (‘H’ high, passage 66) and lower passage number (‘L’ low, passage 40), respectively. (B) Representative images of metaphase spreads from control (euploid) and DSS control (tetraploid) colonoids. 60x. (C) Representative color depth coded images of chromosome segregation errors as revealed by H2B-mNeon labeling of colonoids. Insets highlight mitoses in white boxes. White arrows indicate mitotic errors, corresponding to Supplementary Videos 1-6. n= 4 or 5 independent experiments. (D) Box and whiskers plot of the percentage of segregation errors. Transverse lines represent the median, boxes show 25th−75th percentile and the whiskers represent the lowest and highest values within 1.5 times the interquartile range. The number of divisions and colonoids analyzed are shown at the top. ****P < 0.0001, *P ≤ 0.05; 2-tailed, non-paired student t-test. (E) Bar plot of the percentage of different segregation errors in analyzed mitotic figures. Other types of errors include multipolar mitoses, mitotic failure, and fusion of nuclei. (F-H) Illustrative cartoons and violin plots for time distribution of duration from nuclear envelope breakdown (NEB) to chromosome alignment (F), chromosome alignment to completion of mitosis (G), and total mitotic time (H). Transverse solid lines represent the median, and the dotted lines border the 25th−75th percentiles. ****P < 0.0001, *P ≤ 0.05; 2-tailed, non-paired student t-test.

Fig. 5

INM loss is frequent in iEGFR colonoids and significantly associated with mitotic errors. (A) Sequential still images captured from representative individual mitoses (highlighted by white arrowheads) as revealed by H2B-mNeon labeling of control, iEGFR, and Apc^mut colonoids. (B) Bar graph stratifying the presence of mitotic errors with the presence (n=79) or loss (n=98) of INM in all analyzed mitoses across colonoid lines. ****P < 0.0001; 2-tailed Fisher's exact test. (C) Bar graph detailing the percentage of mitoses with INM loss in each colonoid line. The number of mitoses and colonoids evaluated per group are shown at top. ***P < 0.001, *P ≤ 0.05; 2-tailed Fisher's exact test.