MAD1 upregulation sensitizes to inflammation-mediated tumor formation

Abstract

Mitotic Arrest Deficient 1 (gene name MAD1L1), an essential component of the mitotic spindle assembly checkpoint, is frequently overexpressed in colon cancer, which correlates with poor disease-free survival. MAD1 upregulation induces two phenotypes associated with tumor promotion in tissue culture cells-low rates of chromosomal instability (CIN) and destabilization of the tumor suppressor p53. Using CRISPR/Cas9 gene editing, we generated a novel mouse model by inserting a doxycycline (dox)-inducible promoter and HA tag into the endogenous mouse Mad1l1 gene, enabling inducible expression of HA-MAD1 following exposure to dox in the presence of the reverse tet transactivator (rtTA). A modest 2-fold overexpression of MAD1 in murine colon resulted in decreased p53 expression and increased mitotic defects consistent with CIN. After exposure to the colon-specific inflammatory agent dextran sulfate sodium (DSS), 31% of mice developed colon lesions, including a mucinous adenocarcinoma, while none formed in control animals. Lesion incidence was particularly high in male mice, 57% of which developed at least one hyperplastic polyp, adenoma or adenocarcinoma in the colon. Notably, mice expressing HA-MAD1 also developed lesions in tissues in which DSS is not expected to induce inflammation. These findings demonstrate that MAD1 upregulation is sufficient to promote colon tumorigenesis in the context of inflammation in immune-competent mice.

Fig 1. Overexpression of MAD1L1 is common in colon cancer, where it correlates with worse prognosis.

(A) Alterations in the gene encoding MAD1, MAD1L1, in 524 samples of normal versus colon adenocarcinoma from TCGA PanCancer Atlas (cancer study identifier: coadread_tcga_pan_can_atlas_2018). z threshold = 2.0. (B-C) Kaplan-Meier survival analysis (KMPlot [25]) showing high expression of MAD1L1 mRNA is associated with worse relapse free survival in a cohort of 1336 patients (B) and post-progression survival in a cohort of 311 colon cancer patients (C). Q1 (lower quartile; “low”) vs Q4 (upper quartile; “high”) analysis is shown. HR = hazard ratio.

Fig 2. Generation of dox-inducible HA-MAD1 mouse model.

(A) Schematic showing strategy for editing the endogenous mouse Mad1l1 locus (top) and the donor DNA template used for homology directed repair (bottom). (B) Breeding scheme used to obtain mice with both the inducible Mad1l1 allele and the rtTA-M2 reverse tetracycline transactivator protein. KI = Knock In. (C) Schematic shows dox-inducible Tet-ON expression of endogenous mouse Mad1l1. (D) Immunoblots show inducible expression of HA-MAD1 after 1 week of 6 mg/kg dox feed. Red arrows indicate HA-MAD1. (E) Quantification of MAD1 protein levels (mean +/- SE). Each dot represents tissue from a different animal. (F) qRT-PCR quantification of Mad1l1 mRNA expression after 1 week of 6 mg/kg dox feed. * = p<0.05. Biorender was used to generate illustrations in 2B.

Fig 3. Dox-inducible expression of HA-MAD1 in colon is sufficient to decrease p53 expression and cause chromosome missegregation.

(A-B) Immunofluorescence of colonic epithelium showing expression of HA (A) and increased expression of MAD1 (B) in HA-MAD1 animals after 1 week of 6 mg/kg dox feed. (C) Quantification of MAD1 expression as in B. Colors indicate samples from different animals. Large circles show mean of cells from 5 fields of view from a single animal. (D) Immunoblot showing p53 protein expression is decreased in colon tissue from 5-month old HA-MAD1 mice after 1 week of 6 mg/kg dox feed as compared to control. Lanes 2 and 3 are from different mice of the same genotype. (E) Quantification (mean +/- SEM) of p53 protein levels, based on immunoblotting with the 1C12 antibody. (F) Immunofluorescence showing p53 is decreased in the colonic epithelium after HA-MAD1 expression. (G) Quantification of p53 expression as in F. Colors indicate individual animals. Large circles show mean from 5 fields of view from a single animal. (H-I) 1 week of 6 mg/kg dox feed is sufficient to induce mitotic defects indicative of CIN in colons of 10–27 week old HA-MAD1 mice. (H) Representative H&E images of normal anaphase and anaphase with evidence of chromosome missegregation in mouse colon. (I). Quantitation (mean +/- SEM) of anaphase defects as in H. n≥20 anaphase cells per mouse in each of 3 mice/genotype. * = p<0.05. ** = p<0.001.

Fig 4. Modest MAD1 upregulation sensitizes to colon tumorigenesis.

(A) Experimental schematic for DSS treatment. All mice were given dox feed for the duration of the experiment. Control animals were on water for 304 days while DSS treated animals were given 3 cycles of DSS for 4 days followed by 17 days of recovery on normal water. n^finished indicates the number of animals that survived to 10 months after the first round of DSS in their cohort. (B) Representative H&E images of mouse colon 7 days after the second treatment with water (left) or 2% DSS (right) showing chronic architectural changes caused by DSS treatment. (C) Representative stills from serial colonoscopies following two HA-MAD1 tumors over time. White dotted lines indicate tumor boundaries. Solid and dashed green lines indicate tumors in D. (D) Percent of colon lumen occluded by tumor over time for tumors shown in C. Solid line corresponds to the tumor in the top row of C. Dashed line corresponds to the tumor in the bottom row of C. (E) Representative stills from colonoscopy movies showing normal colon (water), scarring of the colon lining 6 months after initiation of DSS treatment (DSS), or tumors that formed 7 months after DSS treatment in HA-MAD1 and p53^+/- animals. White dotted lines outline tumors. (F) Incidence of tumor formation detected by colonoscopy 10 months after start of DSS and confirmed histologically. Significance shown is compared to water treated control animals (Mad1l1^+/+;rtTA-M2^KI/KI). (G) Size of tumors at 10 months following the start of DSS observed via colonoscopy and measured as percent of colon occluded by the tumor. n = 4 tumors in HA-MAD1 animals and 1 tumor in p53^+/- mice. (H) Colon tumor incidence after DSS treatment categorized by sex. Numbers indicate number of mice with tumor/number of DSS-treated mice of the indicated genotype. * = p<0.05. ns = not significant.

Fig 5. Pathology of colon tumors in DSS-treated HA-MAD1 mice.

Lesions in HA-MAD1 and p53+/- colons were excised, embedded in paraffin, sectioned, and stained with H&E, Ki-67, and β-catenin. Example images of normal colon, hyperplastic polyp, tubular adenoma, tubular adenoma with high grade dysplasia, and a mucinous adenocarcinoma are shown. Black insets in the 1^st column are enlarged in the 2^nd column. Red insets in the 2^nd column are enlarged in the Ki-67 and β-catenin columns. Serial sections were used for staining with H&E, Ki67 and β-catenin. Ki67 staining is confined to the base of the crypts in the no lesion control and the hyperplastic polyp. Ki67 positive cells are identified higher up in the crypts in the tubular adenoma. The tubular adenoma with high grade dysplasia loses crypt structure, but has positive Ki67 staining throughout. The adenocarcinoma had pockets of epithelial cells with positive Ki67 lining the large mucin pools. Arrows indicate cells with positive Ki-67 staininig. β-catenin staining was primarily found to be membranous and excluded from the nucleus in all samples with the exception of the tubular adenoma with high grade dysplasia, which contained β-catenin positive nuclei (arrowhead).

Fig 6. HA-MAD1 tumors have increased CIN and decreased p53.

(A) Immunofluorescence showing MAD1 expression in colonic epithelium. (B) Quantification of MAD1 expression as in A. Each color represents 1 animal. Shapes indicate type of lesion. Each dot represents the average of 5 fields of view. (C) Representative H&E images of normal anaphase and defective anaphase consistent with CIN in mouse colon isolated from 12-month-old mice 10 months after the start of DSS treatment. (D). Quantitation (mean +/- SEM) of anaphase defects as in C. n≥20 anaphase cells per mouse in each of 3 (no lesion) or 2 (hyperplastic polyp, tubular adenoma with high grade dysplasia) mice per genotype. Statistics shown are in comparison to water treated control mice. (E) Immunofluorescence showing p53 expression in colon epithelium. TA w/ HGD = tubular adenoma with high grade dysplasia. (F) Quantification of p53 expression as in E. Each color represents 1 animal. Shapes indicate type of lesion. Each dot represents the average of 5 fields of view. “Normal” indicates normal tissue adjacent to the lesion. * = p<0.05. ** = p<0.01.* ns = not significant.