Liver receptor homolog 1 contributes to intestinal tumor formation through effects on cell cycle and inflammation

Abstract

Liver receptor homolog 1 (LRH-1) is an orphan nuclear receptor that synergizes with beta-catenin/T cell factor 4 signaling to stimulate intestinal crypt cell renewal. We evaluated here the impact of haploinsufficiency of LRH-1 on intestinal tumorigenesis by using two independent mouse models of human colon tumorigenesis. Haploinsufficiency of LRH-1 blunts intestinal tumorigenesis in the ApcMin/+ mice, a genetic model of intestinal cancer. Likewise, Lrh-1+/- mice are protected against the formation of aberrant crypt foci in the colon of mice exposed to the carcinogen azoxymethane. LRH-1 gene expression is reduced in tumors that express elevated levels of the proinflammatory cytokine TNF-alpha. Reciprocally, decreased LRH-1 expression in Lrh-1+/- mice attenuates TNF-alpha expression. Compared with normal human colon, expression and subcellular localization of LRH-1 is significantly altered in neoplastic colon. In combination, these data suggest a role of LRH-1 in the initiation of intestinal tumorigenesis both by affecting cell cycle control as well as through its impact on inflammatory pathways.

Fig. 1.

Haploinsufficiency of LRH-1 reduces tumor multiplicity in Apc^Min/+ mice. (A and B) Total tumor number in the gastrointestinal tract of 18-week-old male (A) and female (B) single C57BL/6J-Lrh-1^+/+/Apc^Min/+ and double C57BL/6J-Lrh-1^+/–/Apc^Min/+ heterozygous mutant littermates (n = 10 per genotype and per gender). Processing and analysis of the gastrointestinal tract is described in Experimental Procedures. Animals were studied on a pure C57BL/6J background and on a standard chow diet. Single and double heterozygous mutant littermates are represented by lines. The mean tumor number is indicated by a bar. Mann–Whitney U test was used to compare data from different genotypes. Differences were considered statistically significant at P < 0.05. (C and D) Tumor distribution by number and diameter in proximal small intestine (SI), distal small intestine, and colon of male (C) and female (D) single C57BL/6J-Lrh-1^+/+/Apc^Min/+ and double heterozygous C57BL/6J-Lrh-1^+/–/Apc^Min/+ mutant mice. x axis represents tumor diameter (in mm). *, Statistically significant at P < 0.05 (ANOVA). (E and F) Representative hematoxylin/eosin staining of normal (E) and adenomatous (F) ileum of single heterozygous C57BL/6J-Lrh-1^+/+/Apc^Min/+ and double heterozygous C57BL/6J-Lrh-1^+/–/Apc^Min/+ mice.

Fig. 2.

Haploinsufficiency of LRH-1 reduces azoxymethane-induced neoplasia. (A) Incidence of mice with ACF in male and female Lrh-1^+/+ and Lrh-1^+/– mice after short-term treatment with AOM. Six-week-old animals (n = 5 per gender and per genotype) were injected i.p. twice for 2 weeks with 7.5 mg/kg AOM and killed 5 weeks after the first injection. Analysis of the colon is described in Experimental Procedures. Statistical significance (P < 0.05; χ² test) is indicated. (B) Total number of ACF in male and female Lrh-1^+/+ and Lrh-1^+/– mice (n = 5 per genotype and gender). Comparisons were only performed between littermates. (C) BrdUrd labeling index expressed as the percentage of BrdUrd-positive cells in the colon of male and female BrdUrd-injected Lrh-1^+/+ and Lrh-1^+/– littermates. BrdUrd immunostaining was performed as described in Experimental Procedures. Percentage of BrdUrd-positive cells represents the number of BrdUrd-positive epithelial cells compared with the total number of epithelial cells in three random fields (×200). (D) BrdUrd immunofluorescence staining of colon sections from BrdUrd-injected Lrh-1^+/+ and Lrh-1^+/– mice that were challenged with AOM.

Fig. 3.

Proinflammatory factors decrease LRH-1 expression in normal and tumoral intestine. (A and B) LRH-1 and TNF-α expression are inversely correlated in the intestine. Relative mRNA levels of LRH-1 and TNF-α were analyzed in normal (N) versus tumoral (T) ileum of 22-week-old female C57BL/6J-Lrh-1^+/+/Apc^Min/+ mice (A) or in normal versus tumoral colon of C57BL/6J mice exposed to long-term AOM treatment (B) as described in Experimental Procedures.(C) Effect of LPS treatment on colon TNF-α and LRH-1 mRNA. Relative mRNA levels of TNF-α and LRH-1 were examined on colon biopsies from C57BL/6 mice killed before or 6 and 24 h after vehicle (CTRL) or LPS i.p. injection. (D) Effect of LRH-1 haploinsufficiency on the expression level of LRH-1 in normal and tumoral ileum of 18-week-old male Apc^Min/+ mice.

Fig. 4.

LRH-1 is differentially expressed in normal versus neoplastic colon. Shown is hematoxylin/eosin staining [A and D (magnification, ×80)] and LRH-1 immunostaining [B and E (magnification, ×80) and C and F (magnification, ×320)] of human normal colon mucosa (A–C) and neoplastic polyp (D-F). (A–C) In normal colon, LRH-1 is found predominantly in the nucleus of the crypt epithelial cells (filled arrowheads). LRH-1 immunostaining is detected in the epithelial cells lining the crypt compartment but absent in the surface epithelial cells (open arrowheads). (D–F) Neoplastic polyp characterized with high-grade dysplasia. A marked elongation and lack of differentiation and maturation of crypt and surface epithelial cells is observed. Nuclei are elongated and pseudostratified. Increased cytoplasmic (open circle) and nuclear (filled arrowheads) LRH-1 immunostaining is observed in all epithelial cells, including those lining the surface, which are normally nonproliferative and unstained.