Ouabagenin is a naturally occurring LXR ligand without causing hepatic steatosis as a side effect

Ouabagenin (OBG) is an aglycone of the cardiotonic steroid ouabain and until now was considered a biologically inactive biosynthetic precursor. Herein, we revealed that OBG functions as a novel class of ligand for the liver X receptor (LXR). Luciferase reporter assays and in silico docking studies suggested that OBG has LXR-selective agonistic activity. In addition, OBG repressed the expression of epithelial sodium channel (ENaC), a LXR target gene, without causing hepatic steatosis, a typical side effect of conventional LXR ligands. This remarkable biological activity can be attributed to a unique mode of action; the LXR agonist activity mainly proceeds through the LXRβ subtype without affecting LXRα, unlike conventional LXR ligands. Thus, OBG is a novel class of LXR ligand that does not cause severe side effects, with potential for use as an antihypertensive diuretic or a tool compound for exploring LXR subtype-specific biological functions.


Supplementary Figures and
. S1. OBG did not remarkably affect the transcriptional activities of many NRs.
OBG was subjected to luciferase reporter assay with many NRs and its cognate reporter plasmids in 293T cell lines. The transcriptional activity of steroid receptors (AR, ER, PR, GR, MR), VDR and FXR were not affected by OBG. 10 -8 M or 10 -9 M OBG and 10 -8 M corresponding NR ligands were treated 48h before the assay. Each luciferase activity of OBG was normalized by renilla luciferase activity (as a control), comparing with that of vehicle. Data are represented as mean ± SD and the values followed by different letters are statistically different according to analysis of variance followed by SNK test (P < 0.01).
V: vehicle, D3: vitamin D3, T09: T0901317, DHT: dehydrotestosterone, Aldo: aldosterone, DEX: dexamethasone, E2: estradiol, P4: progesterone. Using the previous reported data sets of the crystal structure for VDR/1a,25-dihydroxyvitamin D3 complex and FXR/fexaramine complex, OBG was subjected to in silico docking simulation and compared with corresponding known ligands. In left panel, gray ribbon; VDR, green; OBG, magenta; 1a,25-dihydroxyvitamine D3. In right panel, gray ribbon; FXR, green; OBG, magenta; fexaramine. The data are represented as mean ± SD and the values followed by different letters are statistically different according to analysis of variance followed by SNK test (P < 0.01).  The amounts of mRNA were normalized by that of gapdh, then compared with that of vehicle. The data are represented as mean ± SD and the values followed by different letters are statistically different according to analysis of variance followed by SNK test (P < 0.01). 22-OH cholesterol means 22-hydroxycholesterol.

Fig. S7. The mRNA level of lxra in M-1 cells was less than 100 times that of lxrb.
The mRNA level between LXRa and b in M-1 cells was compared by qRT-PCR. The data were normalized by that of gapdh and represented as mean ± SD.

Fig. S8. The confirmation of the efficiency for the over-expression efficiency of LXRa in LXRb knocked down M-1 cells.
The M-1 cells were treated with LXRa plasmid for 6 h after treatment with siRNA for LXRb for 24 h.
The expression level of lxrs was evaluated by real time-qPCR analysis in the same manner as Fig. 5A. The data were normalized by that of gapdh then compared with that of intact. The graphs are represented as mean ± SD and the values followed by different letters are statistically different according to analysis of variance followed by SNK test (P < 0.01).

Fig. S9. LXR subtype-selective knockdown in 293T cells
LXRa or b subtype-selective knockdown in M-1 cells was performed by transfection of corresponding siRNAs. The efficacy and selectivity were confirmed by qRT-PCR. The amount of mRNA was normalized against GAPDH, then compared with siControl.

Fig. S10. mRNA expression level of both subtype of LXR in 293T cells were almost same.
The mRNA level between LXRa and b in 293T cells was compared by qRT-PCR. The data were normalized by that of gapdh and represented as mean ± SD.

Fig. S11. mRNA expression level of LXRb was a little more than that of LXRa in mouse kidney.
The mRNA level between LXRa and b in mouse kidney was compared by qRT-PCR. The data were normalized by that of gapdh and represented as mean ± SD.

Fig. S12. OBG but not T0901317 agonistic activity towards C-terminal tagged-mouse LXRs could not be observed in luciferase reporter assay,
The expression vector for mouse LXRa or LXRb C-terminal tagged with Myc-DDK was subjected to luciferase reporter assay in 293T cells as described in Fig. S1. The fold change in transcriptional activity is expressed as the mean ± SD of triplicate experiments. Values followed by different letters are statistically different according to analysis of variance (ANOVA) followed by SNK tests (P < 0.01).