3.1. Modulation of FXR
FXR is implicated in bile acid, lipid, and glucose homeostasis, hepatic inflammation, regeneration, and fibrosis and is widely distributed in organs such as the liver, kidney, intestinal tract, and adrenal gland [
63,
64,
65,
66]. FXR is encoded by two genes, namely, FXRα and FXRβ, albeit the latter is a pseudogene in primates. Bile acids, the natural ligands of FXR, have been considered potential intestinal tumor promoters [
67]. FXRα is an attractive drug target for the treatment of diverse metabolic diseases, including NASH, primary biliary cirrhosis (PBC), diabetes [
68], and atherosclerosis [
5,
68,
69,
70,
71] in addition to cancer [
72]. Interaction with FXRα recruits coactivators such as SRC1 or corepressors such as nuclear receptor corepressor (NCoR1) [
5]. Extensive research effort ended up with the discovery of some clinical FXR ligands, including obeticholic acid, the only approved FXR modulator to date for PBC therapy, and under clinical phase III for the treatment of NASH [
73].
FXR expression is found to be decreased in human intestinal tumors due to the promotion of Wnt signaling, while the reactivation of FXR in a xenograft model via adenoviral infection induced cytotoxicity through the induction of apoptosis and inhibition of proinflammatory and antiapoptotic genes [
60]. On the contrary, FXR activation promotes transforming growth factor β (TGF-β) induced epithelial-mesenchymal transition in hepatocellular carcinoma cells [
74]. By virtue of its complicated role in cholesterol and bile acids homeostasis, both FXR agonists and antagonists may be therapeutically useful for the treatment of metabolic diseases [
5].
The first report relating
OA with FXR was disclosed by Liu and Wong in 2010 [
75]. They supposed that
OA health benefits are partially attributed to FXR modulation. Luciferase assay using hepatocellular carcinoma cells showed that
OA competitively suppressed the activity of FXR-LBD induced by its endogenous activator chenodoxycholic acid (
CDCA) without affecting the latter’s metabolism.
OA not only bound to FXR-LBD and suppressed its activity in a dose-dependent manner but also partially blocked the interaction with the coactivator SRC-3, as shown in a cell-free model. At 25 μM concentration of
OA in HepG2, quantitative RT-PCR (RT-qPCR) showed that
OA partially blocked
CDCA induction of bile salt export pump (BSEP) and cytochrome P450 7A1 (CYP7A1) target genes but did not affect the expression level of another target gene, organic solute transporter (OST-β) and slightly enhanced SHP suggesting that
OA works as a gene selective modulator of FXR.
OA did not significantly reduce LXRα and LXRβ activity induced by their known synthetic ligand TO901317 [
75,
76].
Another interesting work showcased the effect of
OA on mice models with obstructive cholestasis by bile duct ligation (BDL) [
77]. Basically, the histological examination of hepatocytes indicated that 20 mg/kg
OA administration ameliorates BDL-induced liver injury. Furthermore, pretreatment with
OA or ursodeoxycholic acid (UDCA), the only approved drug for treatment of PBC by Food and Drug Administration (FDA), in BDL mice lowered the level of alanine aminotransferase (ALT) by 59% and 41%, aspartate transaminase (AST) by 33% and 28%, alkaline phosphatase (ALP) by 44% and 39%, respectively. Furthermore,
OA attenuated BDL-induced extrahepatic cholestasis in association with association with enhancement of urine bile acid output. Mechanistically, gene expression analysis proved that
OA resulted in increased mRNA expression of bile acid efflux transporters MRP2, MRP3, and MRP4 [
45], which is ascribed to
OA-mediated accumulation of nuclear factor-erythroid 2-related factor (NRF2). A significant decrease in Bsep expression when
OA was administrated to Sham mice was also further confirmed by RT-qPCR. The latter effect was confirmed to be due to
OA antagonism of FXR through a dual luciferase reporter assay in HepG2 cells. In the latter assay, CDCA was used as a positive control and significantly enhanced the luciferase activity of the FXR reporter gene, which was opposed by co-treatment with
OA in a dose-dependent manner. Taken together, Chen et al. concluded that
OA’s protective role against BDL-induced extrahepatic cholestasis is ascribed to increasing basolateral bile acid export, probably via NRF2-mediated upregulation of MRP2, MRP3, and MRP4, meanwhile, decreased canalicular Bsep expression by
OA which is mediated by FXR antagonism may also have a paramount role in attenuating bile duct injury [
44,
45,
77].
Fang research group succeeded in affording the first semi-synthetic
OA derivatives as FXR antagonists in 2017 [
78]. They designed and synthesized four
OA derivatives through the transformation of its 3β-OH, considering the docking score of the candidates versus fexaramine as a control. They used the crystal structure of FXR as a model (PDB code: 1OSH), utilized Autodock 4.2 software in computational work, and synthesized the four top-scored compounds
1-
4. These esters were afforded through reacting free
OA or protected benzyl
OA ester with the appropriate activated acid, followed by catalytic hydrogenolysis for deprotection. In human embryonic kidney 293T cells, compounds
1—
4 suppressed FXR transactivation in a concentration-dependent manner with respective IC
50 19.41, 7.03, 13.74, and 9.03 μM as settled by a dual-luciferase reporter assay. This is in quite agreement with their predicted docking energy values pKi (
Figure 3). Seemingly,
2 and
4 have similar placement in FXR-LBD, forming two T-shaped π-π stacking with Trp458 in helix 11. This crucial interaction is missing in the case of
OA,
1, and
3 [
78].
Intriguingly, Pan
et al. revealed that
OA is likely to work as an FXR agonist. They studied the
OA effect in human umbilical vein endothelial cells HUVECs atherosclerosis model by treating them with oxidized low-density lipoprotein (ox-LDL, 100 μg/mL) for 24 h [
79]. Basically,
OA interrupted the ox-LDL-induced cell apoptosis in the HUVECs. Luciferase reporter assay revealed that the FXR relative luciferase activity was significantly higher in the case of
OA treatment in a dose-dependent manner, leading to angiotensin (Ang)-(1–7) upregulation, which, in turn, perturbs the development of atherosclerosis. The results were further validated in New Zealand white (NZW) rabbits as an atherosclerosis animal model. The atherosclerosis assessment at the abdominal aorta and thoracic aorta of the animal models was performed by histopathological analysis in the presence of
OA and simvastatin as a positive control. Both
OA and simvastatin inhibited the development of atherosclerosis by minimizing the aortic lesion area size and enhancing the collagen content. Although this study is in discrepancy with previous studies that reported FXR antagonism by
OA, it introduced an insight into the therapeutic potential of the latter for the treatment of atherosclerosis [
79]. It is worth repeating that the NRs modulation is tissue-specific; therefore, it is quite normal to find different pharmacodynamics in different tissues.
Another study explored the underlying mechanism of
OA in alleviating alpha-naphthol isothiocyanate (ANIT)-induced cholestatic liver damage in rats instead of BDL [
80,
81]. As anticipated,
OA decreased hepatocyte necrosis and reduced inflammatory cell infiltration in a similar way to UDCA. In rat hepatocytes,
OA significantly restored glutathione levels of rat primary hepatocytes reduced by ANIT. This is by reversing the high serum levels of ALP, ALT, AST, total bile acid and (TBA), total bilirubin (TBiL), and gamma-glutamyl transferase (γ-GT) levels in the ANIT-induced model as shown by RT-qPCR. This is attributed to the restoration of FXR and Nrf2 mRNA and protein levels, which were reduced in ANIT. Consequently, treatment with
OA decreased the expression of Cyp7a1 mRNA and protein in rats and restored Bsep levels [
82].
The same research group conducted an extensive mechanistic study [
83]. They compared
OA protective effect on ANIT-induced cholestatic liver injury in wild-type and Nrf2 gene knockdown rats and demonstrated that the effect was much weaker in the latter case. This highlights the important role of
OA in stimulating the Nrf2 pathway. Likewise, the protective effect of
OA against ANIT-induced cholestatic liver injury was relatively weaker in FXR knockdown than in type rats. This points out that the protective effects of
OA on ANIT-induced injury and its regulatory role of the bile acids homeostasis gene are mainly ascribed to the simultaneous activation of NRF2 and FXR dual signaling pathways. The authors found a correlation between NRF2 and FXR signaling [
83].
By virtue of the cellular context effect of NRs modulators, Fallon
et al. studied the effect of hederagenin and
OA on FXR in human colonic epithelial cells T84 in comparison to the GW4064, a synthetic FXR agonist. Surprisingly, they found that both hederagenin and
OA compounds don’t have direct agonistic FXR activity in this model. Having said that, they induced the overexpression of FXR mRNA and protein and upregulated GW4064-induced FXR signaling. This opens the way for the potential application of
OA in colon cancer and secretory diarrheas [
59].
The same group of Fang reported a class of 12β-oxygenated oleanolic alkyl esters with FXR modulation properties [
84]. In brief,
OA 28-COOH was protected by benzylation, and 3-OH was protected by etherification with
t-butylmethylsilyl (TBS). Olefin of the double-protected derivative was oxidized by
meta-chloroperoxybenzoic acid (
m-CPBA), affording a 12-oxo derivative. The latter was reduced by sodium borohydride (NaBH
4) to a mixture of 12β- and 12α-hydroxyl derivatives in ca. 2:1 ratio. The 12β-OH compound was reacted with the corresponding carboxylic acids in the presence of N,N’-diisopropylcarbodiimide (DIC) or 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDCI) and 4-dimethylaminopyridine (DMAP), followed by deprotection of TBS and Bn groups with BF
3.Et
2O and Pd/C hydrogenolysis, respectively, to furnish a series of 12-
OA alkyl esters. Various aliphatic and aromatic substituents were used with different polarities.
In HEK293T cells, screening of the synthesized series at ten μM showed that compound
5 with an acetopropionyl group and
6 with a 4-acetobutanoyl group of the 12-
O-alkanoic acid esters are pronounced FXR antagonists opposing the CDCA effect without observable cytotoxicity at ten µM (
Figure 4). Compounds with carboxy-, amino-, or phenyl terminal groups demonstrated less activity. Owing to its prominent antagonistic activity at 10 and 1 µM, the IC
50 of compound
6 was calculated to be 0.10 µM. Its binding mode and pharmacophore placement into FXR-LBD is comparable to the FXR antagonist, ivermectin, as calculated by docking simulation [
52]. To assess its selectivity, the authors tested compound
6 for the action on an array of NRs, including RXRα, RXRβ, RXRγ, PPARα, PPARβ, PPARγ, LXRα, LXRβ, PXR, and another bile acid receptor GPBAR compared to guggulsterone. While the latter antagonizes almost all tested NRs, compound
6 demonstrated an outstanding antagonism (>90 %) against FXR and inconsiderable antagonism (10–20 %) against LXRα and PPARα.
Transcription of FXR downstream genes was then evaluated in HepG2 cells. Ten μM of
6 efficiently reversed the induction of small heterodimeric partner (SHP) and BSEP by 50 μM of the endogenous agonist CDCA without significant effect on sterol regulatory element-binding protein-1c (SREBP-1c and CYP7A1). This is controversial due to the discrepancy with the effect of the parent
OA, which suppresses CYP7A1 expression [
75]. Intriguingly, in the absence of CDCA, compound
6 clearly hindered SHP, BSEP, and CYP7A1 at ten μM but did not affect the expression of SREBP-1c, revealing a unique FXR downstream regulation.
To explore the possible effect of
6 on FXR-controlled genes that regulate blood glucose level and gluconeogenesis [
86], it was found to suppress mRNA levels of phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase (G6Pase) in the presence or absence of CDCA. In addition, in KKay mice, treatment with
6 significantly decreased fasting and non-fasting blood glucose levels. Concomitantly,
6 improved glucose tolerance and insulin sensitivity and lowered HbA1c levels [
84].
In continuation of their work, Fang group designed and synthesized more
OA hybrids with 12β-
O-β-aspartyl or 12β-
O-γ-glutamyl moiety trying to enhance the affinity of compound
6 towards FXR-LBD as shown for compounds
7–
10 [
84,
85]. The design considered the presence of unoccupied hydrophobic space around the compound
6 terminal chain, thus aimed at using more branched chains. The authors adopted similar synthetic procedures for
OA 12β-OH esterification that were used for the synthesis of compound
6, as mentioned above. Among the new series, compound
10 with
S-γ-glutamyl moiety possesses the highest antagonism to FXR with IC
50 0.44 µM in HEK293T cells. Using the
R-γ-glutamyl chain lowered the activity (
9, IC
50 1.95 µM).
OA-bearing
S-aspartyl chain, compound
7, is slightly less active than
10 with IC
50 0.95 µM. Protection of the terminal amino- group by
tert-butyloxycarbonyl (Boc) and the terminal carboxyl with
tert-butyl (
t-Bu) group is crucial for activity; for example, compound
8 with both free terminal groups dramatically lost activity. In general,
OA derivatives bearing (
S) configuration side chains outperform those bearing (
R) configuration ones, and glutamyl derivatives outperform aspartyl derivatives. In a similar fashion to compound
6,
10 showed favorable selectivity against FXR with much less effect on LXRα and PPARα and no effect on other tested metabolic NRs. Consequently,
10 significantly inhibited the expression of PEPCK and G6Pase at 1 and 10 µM concentrations. In hepatic stellate cell line LX-2, compound
10 lowered mRNA expressions of liver fibrosis marker genes, including collagen type I α-1, actin-α-2, transforming growth factor β-1, connective tissue growth factor, and integrin α-V, whereas guggulsterone lowered expression of only collagen type I α-1, actin-α-2.
In a bile duct ligation (BDL) rat model, the oral administration of compound
10 for two weeks effectively decreased the levels of AST, ALT, TBA, and TBiL, which means less liver fibrosis. Liver histopathology using hematoxylin and eosin (H&E) staining showed that compound
10 intake alleviated bile duct hyperplasia and parenchymal necrosis that accompany BDL. The collagen-specific Sirius red staining showed less collagen accumulation in the compound
10 treated group. These
in vitro data of reducing hepatic fibrosis markers were further validated
in vivo. Interestingly, in a NASH mice model, the titled compound reduced intrahepatic steatosis and hepatic lobular inflammation, indicating less liver fat accumulation. This is accompanied by a reduction of mRNA expression of fibrosis marker genes. Collectively, compound
10 is an
OA derivative and a promising FXR modulator with NASH and diabetes therapeutic potential [
85].
Collectively,
OA reprograms the liver to protect against hepatotoxic chemicals, but its intake should be with care since its high doses are reported to be hepatotoxic and can develop cholestatic liver injury [
42,
87,
88]. It is worth noting that the knockdown of FXR ameliorated
OA-induced cholestatic liver injury [
87]. Such paradoxical hepatoprotection and hepatotoxicity are common for natural herbs. In conclusion,
OA dose is the one to differentiate between its role as a remedy and a poison [
89].