Human FXR regulates SHP expression through direct binding to an LRH-1 binding site, independent of an IR-1 and LRH-1

Abstract

Background: Farnesoid X receptor/retinoid X receptor-alpha (FXR/RXRα) is the master transcriptional regulator of bile salt synthesis and transport in liver and intestine. FXR is activated by bile acids, RXRα by the vitamin A-derivative 9-cis retinoic acid (9cRA). Remarkably, 9cRA inhibits binding of FXR/RXRα to its response element, an inverted repeat-1 (IR-1). Still, most FXR/RXRα target genes are maximally expressed in the presence of both ligands, including the small heterodimer partner (SHP). Here, we revisited the FXR/RXRα-mediated regulation of human SHP.

Methods: A 579-bp hSHP promoter element was analyzed to locate FXR/chenodeoxycholic acid (CDCA)- and RXRα/9cRA-responsive elements. hSHP promoter constructs were analyzed in FXR/RXRα-transfected DLD-1, HEK293 and HepG2 cells exposed to CDCA, GW4064 (synthetic FXR ligand) and/or 9cRA. FXR-DNA interactions were analyzed by in vitro pull down assays.

Results: hSHP promoter elements lacking the previously identified IR-1 (-291/-279) largely maintained their activation by FXR/CDCA, but were unresponsive to 9cRA. FXR-mediated activation of the hSHP promoter was primarily dependent on the -122/-69 region. Pull down assays revealed a direct binding of FXR to the -122/-69 sequence, which was abrogated by site-specific mutations in a binding site for the liver receptor homolog-1 (LRH-1) at -78/-70. These mutations strongly impaired the FXR/CDCA-mediated activation, even in the context of a hSHP promoter containing the IR-1. LRH-1 did not increase FXR/RXRα-mediated activation of hSHP promoter activity.

Conclusion: FXR/CDCA-activated expression of SHP is primarily mediated through direct binding to an LRH-1 binding site, which is not modulated by LRH-1 and unresponsive to 9cRA. 9cRA-induced expression of SHP requires the IR-1 that overlaps with a direct repeat-2 (DR-2) and DR-4. This establishes for the first time a co-stimulatory, but independent, action of FXR and RXRα agonists.

Figure 1. The IR-1 at −291/−279 is required for 9cRA-, but not for CDCA-mediated induction of the human SHP promoter.

DLD-1 cells were transfected with hFXR and hRXRα expression plasmids and various hSHP promoter constructs as indicated. Cells were treated with or without 100 µmol/L CDCA and/or 1 µmol/L 9cRA. The synergistic effect of the FXR ligand (CDCA) and RXRα ligand (9cRA) on SHP promoter activity depends on the previously identified IR-1 located at −291/−279. Mutation or deletion of this IR-1 sequence did not abolish SHP promoter activation by FXR/CDCA. Luciferase activity was measured to determine the SHP promoter activity. Data are presented as means ± SD; n≥3. Vehicle-treated conditions are set to 1. p≤0.05 for *) in a pairwise comparison by Mann-Whitney U test.

Figure 2. FXR is required for CDCA-induced activation of the −278/+10 SHP promoter.

DLD-1 cells were transfected with hFXR and hRXRα expression plasmids and various hSHP promoter constructs as indicated. Cells were treated with or without 100 µmol/L CDCA. Luciferase activity was measured to determine the SHP promoter activity. Data are presented as means ± SD; n≥3. P≤0.05 for *) in a pairwise comparison by Mann-Whitney U test.

Figure 3. A FXR/RXRα/CDCA-responsive element is located in the −122/−69 region of the SHP promoter.

A) shows an overview of the different constructs used to localize the FXR-responsive element in the −278/−69 region of the SHP promoter. Relevant binding sites for other NRs are included. (B, C) DLD-1 cells were transfected with the indicated hSHP promoter constructs and expression plasmids for hFXR and hRXRα. Cells were treated with or without 100 µmol/L CDCA (B) or 1 µmol/L GW4064 (C). Luciferase activity was measured to determine SHP promoter activity. Data are presented as means of ± SD; n≥3. Significant differences are indicated when compared to CDCA/GW4064-treated −569/+10 (a); CDCA/GW4064-treated −303/+10 (b); CDCA/GW4064-treated −278/+10(c). P≤0.05 in a pairwise comparison by Mann-Whitney U test.

Figure 4. The −122/−69 region is required for optimal FXR-ligand-mediated induction of the SHP promoter in DLD-1, HEK293 and HepG2 cells.

The colon carcinoma (DLD-1), human embryonic kidney (HEK293) and hepatoma (HepG2.rNtcp) cell lines were transfected with the indicated hSHP promoter constructs and expression plasmids for hFXR and hRXRα. Cells were treated with or without 100 µmol/L CDCA (A) or 1 µmol/L GW4064 (B). Luciferase activity was measured to determine SHP promoter activity. Data are presented as means of ± SD; n≥3. Promoter activity in CDCA/GW4064-treated condition is significantly different from the −278/+10 construct in DLD-1 (a), HEK293 (b) or HepG2.rNtcp (c) cells. P≤0.05 in a pairwise comparison by Mann-Whitney U test.

Figure 5. The LRH-1 site is required for FXR-induced expression of SHP.

(A) shows the location of IR-1-like half sites and an LRH-1 site in the −122/−69 region of the hSHP promoter. The latter was previously identified in the murine Shp promoter and conserved in rat and human (see Figure S6). All 4 IR-1-like sites and the LRH-1 site were mutated and analyzed for the effect on FXR/CDCA-mediated induction of the −278/+10 hSHP promoter (B) mutating one of the IR-1 half-sites did not or only partially reduce FXR/CDCA-dependent activation of the −122/−69 hSHP promoter fragment, whereas mutations in the LRH-1 site strongly reduced the response of the −122/−69 hSHP promoter fragment to FXR/CDCA-stimulation. *) significantly different from CDCA treated 278/+10 WT (C) in the context of the −569/−10 hSHP promoter fragment the LRH-1 site is the dominant FXR/CDCA response element (over the previously identified IR-1). DLD-1 cells were transfected with hFXR and hRXRα expression plasmids and various hSHP promoter constructs as indicated. Cells were treated with or without 100 µmol/L CDCA. Luciferase activity was measured to determine the SHP promoter activity. a) significantly different from CDCA treated 569/+10 WT. b) significantly different from CDCA treated −569/+10 IR-1 KO. Data presented as means ± SD; n≥3. P≤0.05 for *), a) and b) in a pairwise comparison by Mann-Whitney U test.

Figure 6. No synergy between FXR and LRH-1 in human SHP regulation.

LRH-1 dose-dependently induced activation of the −303/+10 hSHP promoter fragment, confirming the presence of a functional LRH-1. In the presence of FXR a similar dose response curve is observed. However, in the presence of CDCA, FXR and LRH-1 do not synergistically activate the SHP promoter. LRH-1 rather limits the FXR/CDCA-dependent activation at a higher dose. DLD-1 cells were transfected with hFXR and hRXRα expression plasmids, the −303/+10 hSHP promoter construct and/or increasing amounts of an mLrh-1 expression plasmid as indicated. Cells were treated with or without 100 µmol/L CDCA. Luciferase activity was measured to determine the SHP promoter activity. a) significantly different from 0 ng mLrh-1 vehicle, between vehicle treated conditions. b) significantly different from 0 ng mLrh-1 FXR/RXRα, between FXR/RXRα treated conditions. Data presented as means ± SD; n≥3. P≤0.05 for a) and b) in a pairwise comparison by Mann-Whitney U test.

Figure 7. FXR binds to the LRH-1 responsive element in the hSHP promoter.

FXR was precipitated from nuclear extracts of hFXR-overexpressing DLD-1 cells using a DNA probe containing the SHP −122/−69 region (“wild type”(WT) or “mutated”(mut)), the IR-1 from the hBSEP promoter (positive control), the LacI binding site (negative control) or empty beads (EB, negative control). A) DNA probes of SHP −122/−69 and BSEP-IR-1 bind FXR. Competition experiments were performed with 3-fold excess hBSEP-IR-1 or LacI lacking a biotin label. B) the SHP −122/−69 region with a mutated LRH-1 site failed to precipitate or compete for FXR binding. Competition experiments were performed with 3-fold excess wild type SHP −122/−69, mutated SHP −122/−69 or LacI lacking a biotin label.