Multi-level interactions between the nuclear receptor TRα1 and the WNT effectors β-catenin/Tcf4 in the intestinal epithelium

Abstract

Intestinal homeostasis results from complex cross-regulation of signaling pathways; their alteration induces intestinal tumorigenesis. Previously, we found that the thyroid hormone nuclear receptor TRα1 activates and synergizes with the WNT pathway, inducing crypt cell proliferation and promoting tumorigenesis. Here, we investigated the mechanisms and implications of the cross-regulation between these two pathways in gut tumorigenesis in vivo and in vitro. We analyzed TRα1 and WNT target gene expression in healthy mucosae and tumors from mice overexpressing TRα1 in the intestinal epithelium in a WNT-activated genetic background (vil-TRα1/Apc mice). Interestingly, increased levels of β-catenin/Tcf4 complex in tumors from vil-TRα1/Apc mice blocked TRα1 transcriptional activity. This observation was confirmed in Caco2 cells, in which TRα1 functionality on a luciferase reporter-assay was reduced by the overexpression of β-catenin/Tcf4. Moreover, TRα1 physically interacted with β-catenin/Tcf4 in the nuclei of these cells. Using molecular approaches, we demonstrated that the binding of TRα1 to its DNA target sequences within the tumors was impaired, while it was newly recruited to WNT target genes. In conclusion, our observations strongly suggest that increased β-catenin/Tcf4 levels i) correlated with reduced TRα1 transcriptional activity on its target genes and, ii) were likely responsible for the shift of TRα1 binding on WNT targets. Together, these data suggest a novel mechanism for the tumor-promoting activity of the TRα1 nuclear receptor.

Figure 1. Analysis of TRα1 target genes in mice of different genotypes.

RT-qPCR experiments were performed in the intestine of 6-month-old mice of the indicated genotype to analyze the mRNA levels of Ctnnb1 (A), Sfrp2 (B), Ccnb1 (C) and Cdc2a (D). Values represent fold change ± sd after normalization to the wild-type [WT] animals. *: P<0.05, **: P<0.001, compared with the WT; #: P<0.05, ##: P<0.01 compared with the vil-TRα1 animals, by two-tailed Student's t-test (n = 4). N, normal mucosa; T, tumor.

Figure 2. The Wnt3a ligand is not sufficient to impair TRα1 transcriptional activity ex vivo.

The primary cultures of intestinal epithelial cells were treated with 10 ng/ml of Wnt3a and/or 10⁻⁷ M of T3 for 24 hours. (A) The number of proliferating cells in the different experimental conditions was analyzed by Ki67 immunolabeling; all of the nuclei were labeled by Hoechst. The percentage of Ki67-positive nuclei was determined by counting under a fluorescence microscope (Zeiss Axioplan). The histograms represent the summary (mean ± sd) of the scoring of specific immunolabeling in two independent experiments each conducted in triplicate (n = 50). (B, C) Analysis of β-catenin in intestinal epithelial primary cultures by immunolabelling (B) and WB (C). Cells were in control, T3, Wnt3a and T3+Wnt3a conditions as indicated. Pictures in B show the fluorescent staining of the nuclei (blue), β-catenin (red) and the merging of each simple labeling. Bar: 15 µm. For the WB (C), we used a specific antibody allowing the detection of activated non-phosphorylated β-catenin , . Actin was used as loading control. The image is representative of two independent experiments. Each lane represents whole protein extracts (50 µg/lane). (D–F) RT-qPCR analysis to evaluate mRNA levels of Ccnd1, Ctnnb1 and Sfrp2. Results are from three independent experiments, each conducted in duplicate. Values represent fold change ± sd after normalization to the control condition (Ctrl). *: P<0.05, **: P<0.01, ***: P<0.001 by two-tailed Student's t-test (n = 6).

Figure 3. Analysis of Tcf7l2 (Tcf4) and Wnt target genes in mice of different genotypes.

RT-qPCR experiments were performed in the intestine of 6-month-old mice of the indicated genotype. (A) Tcf4, (B) Lef1, (C) c-Myc and (D) Ccnd1 mRNA levels were analyzed. Values represent fold change ± sd after normalization to wild-type (WT) animals. *: P<0.05, **: P<0.01 compared with the WT; $: P<0.05, $$: P<0.01 compared with the healthy mucosa of the same genotype; @: P<0.05, @@: P<0.01 compared with vil-TRα1/Apc N and T; #: P<0.05 compared with vil-TRα1, by two-tailed Student's t-test (n = 4). N, normal mucosa; T, tumor.

Figure 4. β-catenin/Tcf4 complex interferes with TRα1 functionality in luciferase assay in vitro.

(A) The DR4-luc luciferase reporter was transfected into Caco2 cells maintained in T3-depleted serum with or without supplementation with T3 as indicated, together with TRα1, Tcf4 or β-catenin expression vectors in different combinations. (B, C) The DR4-luc luciferase reporter (B) or TopFlash luciferase reporter (C) was transfected into Caco2 cells maintained in culture medium containing physiological concentrations of T3, together with the β-catenin/Tcf4 complex in the presence or absence of the TRα1 expression vector. Histograms represent mean ± sd from three independent experiments, each conducted in triplicate (n = 9). *: P<0.05, **: P<0.01 compared with the control condition (Ctrl); #: P<0.05, ##: P<0.01, compared with the TRα1 condition; $: P<0.05 compared with the TRα1+β-catenin condition; ££: P<0.01 compared with the β-catenin or β-catenin+Tcf4 condition, by two-tailed Student's t-test.

Figure 5. Physical interaction between TRα1 and the β-catenin/Tcf4 complex.

Nuclear extracts from Caco2 cells, maintained in T3-depleted (−) or T3-supplemented (+) serum, were immunoprecipitated with antibodies directed against endogenous Tcf4, β-catenin or TRα1 and analyzed by WB by using the antibodies as indicated. Rabbit IgG was used as negative control. Ponceau red was used as whole-protein (50 µg/lane) loading control. Histone H1 was used to check the enrichment and was the loading control for the nuclear proteins in the inputs. The pictures are representative of at least three independent experiments.

Figure 6. Chromatin occupancy of TRα1, β-catenin and Tcf4 on genomic regions of target genes.

ChIP analysis was performed with chromatin isolated from the intestine of WT or vil-TRα1/Apc mice (healthy mucosa and tumor) and immunoprecipitated with anti-TRα1, anti-β-catenin, anti-Tcf4 antibodies or rabbit IgG (negative control). qPCR was performed using specific primers covering the TRE of Sfrp2 (A) and Ctnnb1 (B), the WRE of Axin2 (C) and c-Myc (D) or the promoters of Villin (E) and 36B4 (F); the Ppia gene was used as internal control. Data are representative of one of two experiments. Histograms represent the specific-DNA enrichment in each sample immunoprecipitated with the indicated antibody. The black bar in A–D delineates the threshold of binding specificity determined by the IgG non-specific binding. N, normal mucosa; T, tumor.

Figure 7. Proposed molecular model for the action of TRα1 in controlling the WNT signaling pathway in the intestine.

In physiological conditions (WT), TRα1 binding on TREs positively regulates the expression and stabilization of β-catenin, and then contributes to maintaining tissue homeostasis. In vil-TRα1 mice, there is an increased level of β-catenin expression and stabilization that leads to WNT activation and hyper-proliferation. In vil-TRα1/Apc mice, the stronger β-catenin stabilization and Tcf4 overexpression might be responsible of the shift in TRα1 binding from TREs to WREs. The physical interaction between TRα1 and β-catenin/Tcf4 we showed can explain its presence on the WREs. We speculate that this is a novel mechanism of WNT induction promoting the activation and/or the acceleration of the tumorigenic process. TREs, Thyroid hormone response elements; WREs, WNT response elements. Solid lines indicate genomic actions; dotted lines represent the speculative model of a negative regulatory loop involving non-genomic actions and eventually other(s) factor(s). Double black arrows indicate crypt width in WT and vil-TRα1 intestinal sections.