An unexpected tumor-resistant phenotype from floxing PAK1 in a mouse model of colitis associated cancer

Abstract

Inflammatory bowel disease (IBD) and colitis-associated cancer are associated with activation of PAK1 (p-21 activated kinase 1). We previously found that total knockout of PAK1 (PAK1KO) reduced tumorigenesis upon AOM/DSS but enhanced tumorigenesis in another model of IBD with total knockout of IL10 (IL10KO). To better understand the specific role of epithelial PAK1, we crossed Pak1 floxed (PAK1fl) with VillinCre mice for a conditional knockout of PAK1 in intestinal epithelia (PAK1CKO). PAK1fl were included as additional controls. Unexpectedly, inflammation and tumorigenesis were greatly reduced in PAK1fl compared to WT or PAK1KO after AOM/DSS treatment. PAK1CKO had higher tumor incidence and counts compared to PAK1fl, but was still lower in comparison to PAK1KO or WT. When crossed with IL10KO mice, PAK1CKO exacerbated the expected hyperproliferative phenotype, resulting in early mouse morbidity. Despite normal Pak1 mRNA expression in PAK1fl colonic lysates, PAK1 protein expression on immunohistochemistry was higher that WT. Both PAK1fl and PAK1CKO mice were more resistant to shifts in microbiome, and remained clustered together compared to WT or PAK1KO. Altogether, our results suggest that floxing itself may have altered Pak1 expression, which conferred protection from AOM/DSS carcinogenesis.

Keywords: AOM/DSS; Colitis-associated cancer; Cre-LoxP; IL10; Inflammatory bowel disease; P-21 activated kinase 1; PAK1.

Fig. 1

Inflammation is reduced upon floxing of PAK1 A. Treatment Regimen. WT (n=7), PAK1KO (n=9), PAK1CKO (n=11) and PAK1fl (n=11) mice were injected with 10mg/kg BW Azoxymethane and underwent four cycles of 1.7% DSS in the drinking water for four days with 2-week recovery periods. B. Weight curve and calculated area under the curve (AUC) across the experiment divided by the number of days in the experiment. C. Disease activity index (DAI) across the experiment and calculated AUC divided by the number of days in the experiment. D. Native colon lengths at sacrifice. E. Representative histological images of hematoxylin and eosin stains of intestinal epithelia and evaluated histological inflammation score. *p ≤ 0.05, ** ≤ 0.01, *** ≤ 0.001.

Fig. 2

PAK1 protein expression is increased in floxed mice A. Cropped representative PCR blots from tail genotyping for floxed PAK1 and VillinCre, edited to increase contrast (Original images in Supplementary Figure 3B) B. Relative expression of PAK1 in whole colon lysates from untreated animals WT (n=3), PAK1CKO (n=3), PAK1fl (n=3). Normalized to Actin. C. Representative images of PAK1 staining in colons from AOM/DSS-treated mice D. Mean percentage of PAK1 positive cells in epithelial cells and lamina propria (LP) WT (n=4), PAK1CKO (n=5), PAK1fl (n=4). PAK1KO is only included on the graph as a reference and is not included in statistics ** ≤ 0.01, *** ≤ 0.001.

Fig. 3

Floxing of PAK1 protects against AOM/DSS tumorigenesis A. Representative colonoscopy images from colonoscopy performed after the 3rd cycle of DSS. B. Representative histological images of swiss-rolled colons, with tumors outlined in blue C. Tumor incidence, multiplicity, total tumor burden, and mean tumor size. ** p ≤ 0.01, *** ≤ 0.001 WT (n=6), PAK1KO (n=9), PAK1flVilcre (n=11), PAK1fl (n=10).

Fig. 4

Genotype co-housing and AOM/DSS treatment do not shift fecal microbial profiles in PAK1fl mice A. NMDS plot of fecal microbial profiles after co-housing genotypes, but prior to DSS treatment (pre-) B. NMDS plots comparing fecal microbial profiles in pre- versus post-AOM/DSS treatment. C. Shannon index of pre- samples. D. Differentially expressed genera in pre- samples of floxed PAK1 versus WT or PAK1KO mice. WT (n=7), PAK1KO (n=6), PAK1CKO (n=7) and PAK1fl (n=6) * ≤ 0.05, ** ≤ 0.01.