An FXR Agonist Reduces Bile Acid Synthesis Independently of Increases in FGF19 in Healthy Volunteers

Abstract

Bile acid (BA) synthesis is regulated through suppression of hepatic cholesterol 7α-hydroxylase via farnesoid X receptor (FXR) activation in hepatocytes and/or enterocytes; in enterocytes, this process requires FGF19 signaling. To study these pathways, we quantified markers of BA synthesis (7α-hydroxy-4-cholesten-3-one [C4]) and cholesterol production (lathosterol), fibroblast growth factor (FGF)19, and BAs in serum from healthy male volunteers given 1 oral dose of the nonsteroidal FXR agonist Px-102 (0.15 mg/kg, 0.3 mg/kg, 0.6 mg/kg, 1.12 mg/kg, 2.25 mg/kg, 3.38 mg/kg, or 4.5 mg/kg). After 8 hours, serum levels of C4 decreased by 80% in volunteers given 0.15 mg/kg, whereas serum levels of FGF19 were unchanged. Serum levels of FGF19 increased significantly, in a dose-dependent manner, in volunteers given >0.3 mg/kg Px-102, up to as much as 1600%, whereas C4 levels remained significantly reduced (by >80%). For all doses, FGF19 levels returned to normal 24 hours after administration of Px-102. Serum levels of C4 decreased before levels of FGF19 levels increased, and were still reduced by 95% 24 hours after the highest dose (4.5 mg/kg) of Px-102, even though levels of FGF19 had returned to baseline. Our findings indicate that activation of hepatic FXR is able to suppress BA synthesis, independent of FGF19. Bile acid (BA) synthesis is regulated through suppression of hepatic cholesterol 7α-hydroxylase via farnesoid X receptor (FXR) activation in hepatocytes and/or enterocytes; in enterocytes, this process requires FGF19 signaling. To study these pathways, we quantified markers of BA synthesis (7α-hydroxy-4-cholesten-3-one [C4]) and cholesterol production (lathosterol), fibroblast growth factor (FGF)19, and BAs in serum from healthy male volunteers given 1 oral dose of the nonsteroidal FXR agonist Px-102 (0.15 mg/kg, 0.3 mg/kg, 0.6 mg/kg, 1.12 mg/kg, 2.25 mg/kg, 3.38 mg/kg, or 4.5 mg/kg). After 8 hours, serum levels of C4 decreased by 80% in volunteers given 0.15 mg/kg, whereas serum levels of FGF19 were unchanged. Serum levels of FGF19 increased significantly, in a dose-dependent manner, in volunteers given >0.3 mg/kg Px-102, up to as much as 1600%, whereas C4 levels remained significantly reduced (by >80%). For all doses, FGF19 levels returned to normal 24 hours after administration of Px-102. Serum levels of C4 decreased before levels of FGF19 levels increased, and were still reduced by 95% 24 hours after the highest dose (4.5 mg/kg) of Px-102, even though levels of FGF19 had returned to baseline. Our findings indicate that activation of hepatic FXR is able to suppress BA synthesis, independent of FGF19. What You Need to KnowBackground and ContextThe regulation of bile acid synthesis from cholesterol in the liver is not completely understood. Marked species differences make human studies essential.New FindingsHepatic rather than intestinal FXR signaling rapidly downregulates bile acid synthesis following administration of the FXR agonist Px-102. This suppression does not require an increase in circulating FGF19 levels in humans.LimitationsThe study cohort only contained males and was limited in size, making details of inter-individual variation and possible gender differences difficult to evaluate.ImpactA better understanding of how bile acid metabolism is regulated in humans makes possible the development of new concepts for therapeutic interventions in conditions such as dyslipidemia, fatty liver disease, and diabetes. The regulation of bile acid synthesis from cholesterol in the liver is not completely understood. Marked species differences make human studies essential. Hepatic rather than intestinal FXR signaling rapidly downregulates bile acid synthesis following administration of the FXR agonist Px-102. This suppression does not require an increase in circulating FGF19 levels in humans. The study cohort only contained males and was limited in size, making details of inter-individual variation and possible gender differences difficult to evaluate. A better understanding of how bile acid metabolism is regulated in humans makes possible the development of new concepts for therapeutic interventions in conditions such as dyslipidemia, fatty liver disease, and diabetes. Synthesized from cholesterol in liver, bile acids (BAs) undergo efficient enterohepatic circulation. Current thinking regarding regulation of their metabolism is largely based on animal experiments.1Hofmann A.F. et al.J Lipid Res. 2014; 55: 1553-1595Crossref PubMed Scopus (246) Google Scholar However, in addition to major species differences, including BA spectrum, synthesis and signaling pathways, gallbladder presence, and intestinal microflora, dissimilarities also exist regarding diurnal rhythms of BA and cholesterol metabolism.2Li T. et al.Pharmacol Rev. 2014; 66: 948-983Crossref PubMed Scopus (634) Google Scholar, 3Zhou M. et al.J Hepatol. 2017; 66: 1182-1192Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar, 4Galman C. et al.Gastroenterology. 2005; 129: 1445-1453Abstract Full Text Full Text PDF PubMed Scopus (166) Google Scholar, 5Li J. et al.Biochim Biophys Acta. 2018; (pii: S0925–4439(18)30181–30189) ([Epub ahead of print])https://doi.org/10.1016/j.bbadis.2018.05.011Google Scholar Understanding regulation of BA metabolism in humans is also important in dyslipidemia, type 2 diabetes, liver steatosis, and cholestasis.6Chavez-Talavera O. et al.Gastroenterology. 2017; 152: 1679-1694.e3Abstract Full Text Full Text PDF PubMed Scopus (594) Google Scholar BA production is controlled by pathways targeting hepatic cholesterol 7α-hydroxylase (CYP7A1), strongly influenced by the nuclear BA receptor, farnesoid X receptor (FXR) by at least 2 parallel pathways (Supplementary Figure 1). Their relative importance is still unclear, and the role of circulating fibroblast growth factor (FGF)19 in humans remains controversial.7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar Efficient hepatic clearance generates a high and robust portosystemic gradient for BAs, but not for FGF19.8Johansson H. et al.J Clin Exp Hepatol. 2017; 473: 1-7Google Scholar Both have diurnal rhythms, with postprandial BA peaks followed by FGF19.9Lundasen T. et al.Journal of internal medicine. 2006; 260: 530-536Crossref PubMed Scopus (337) Google Scholar Treatment with cholestyramine decreases serum FGF19 by approximately 90% and increases hepatic CYP7A1 activity and its serum marker 7α-hydroxy-4-cholesten-3-one (C4) many-fold.7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar Conversely, treatment with chenodeoxycholic acid (CDCA) lowers CYP7A1 activity and C4 by approximately 90%, whereas FGF19 only increases 2.5-fold.7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar, 10Ghosh Laskar M. et al.J Intern Med. 2017; 281: 575-585Crossref PubMed Scopus (46) Google Scholar This suggests that intestinal (FGF19) and hepatic FXR pathways may contribute differently in suppression and derepression of BA synthesis. Separating their contributions is difficult, however, since BA administration changes size and composition of the BA pool. Administration of FGF19 is not feasible due to hepatoproliferative effects,5Li J. et al.Biochim Biophys Acta. 2018; (pii: S0925–4439(18)30181–30189) ([Epub ahead of print])https://doi.org/10.1016/j.bbadis.2018.05.011Google Scholar but infusion of high doses of an FGF19 analog reduces BA synthesis in humans,11Luo J. et al.Science translational medicine. 2014; 6: 247ra100Crossref PubMed Scopus (135) Google Scholar similar to findings in mice.7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar Potent nonsteroidal FXR agonists, such as Px-102,12Deuschle U. et al.Int J Cancer. 2015; 136: 2693-2704Crossref PubMed Scopus (12) Google Scholar may be used to explore the respective contribution of intestinal and hepatic pathways suppressing CYP7A1 in humans (Supplementary Figure 1). We here characterize responses to administration of increasing single doses of Px-102 to healthy male volunteers (Supplementary Figure 2), monitoring temporal patterns of serum BAs, FGF19, and C4. This allows establishment of immediate effects without compensatory changes expected with prolonged treatment. We tested the hypothesis that circulating FGF19 predominates feedback inhibition of BA synthesis, predicting that Px-102 treatment reduces C4 subsequent to elevated FGF19 levels. To avoid postprandial interference,9Lundasen T. et al.Journal of internal medicine. 2006; 260: 530-536Crossref PubMed Scopus (337) Google Scholar no food was given until 12 AM. Serum BAs and FGF19 remained unchanged this period, thereafter showing expected postprandial increases (Figure 1). C4 levels increased before the meal, consonant with a meal-independent diurnal rhythm of BA synthesis.4Galman C. et al.Gastroenterology. 2005; 129: 1445-1453Abstract Full Text Full Text PDF PubMed Scopus (166) Google Scholar Following intake, serum Px-102 levels peaked after 60 minutes, displaying rapid clearance with estimated half-life of approximately 3 hours (Supplementary Figure 2); peak concentrations varied between 44 and approximately 2250 ng/mL Px-102 was detected after 24 hours at doses greater than 0.3 mg/kg, Serum FGF19 increased dose-dependently following Px-102 (Figure 1E). There was no response at the lowest dose, and increases were only seen postprandially at doses of 0.3 to 1.2 mg/kg (Figure 2). At higher doses, increases were seen both fasting and postprandially, reaching elevations of 800% to 1600%. The higher the Px-102 dose, the earlier the response in FGF19. Px-102 levels of >100 ng/mL were required to elicit a clear FGF19 response. For all doses, FGF19 levels were normalized after 24 hours. Circulating C4 was rapidly and markedly diminished following Px-102 (Figure 1F), even when no FGF19 increase was observed. Already at the lowest dose, C4 was reduced by 80% (Figure 2), and prolonged suppression lasting overnight was seen at the 2 highest doses. Interestingly, the degree of C4 inhibition at 0.30 mg/kg and at 4.45 mg/kg was not different (P > .05), except at 24 hours (P < .001) (Supplementary Figure 3). The C4 reduction was not preceded by increased FGF19, except at the highest doses. At Px-102 doses of 1.2 mg/kg or higher, cholesterol synthesis (lathosterol) was reduced, indicating lower precursor demand (Figure 1G). Total and individual BAs increased during the fasting part of the study (8 AM to 12 AM), whereas there was no clear change in the postprandial state (Figure 1D; Supplementary Figure 4). Whereas serum cholesterol was unchanged, triglycerides tended to decline during the night at higher doses (Figure 1H and I). Our data thus demonstrate that human BA synthesis is almost maximally suppressed at very low doses of Px-102 (not influencing FGF19), and that the C4 reduction is generally not preceded by increased FGF19 levels. Although the findings do not support our primary hypothesis, they do not exclude a role of circulating FGF19 in regulation of BA synthesis. First, although Px-102 is a well-documented nonsteroid FXR agonist, it also may exert other effects on BA synthesis, so similar experiments should be performed using alternative agonists. Second, FGF19 may have additive (synergistic) effects at very high levels. As mentioned, parenteral administration of an equipotent FGF19 analog elicits a robust reduction of C4 in humans,11Luo J. et al.Science translational medicine. 2014; 6: 247ra100Crossref PubMed Scopus (135) Google Scholar but the dose used (3 mg/d) corresponds to approximately 5000 times the estimated plasma compartment content of FGF19.7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar Finally, upholding an adequate (permissive) level of FGF19 may be crucial for maintaining intact feedback even if immediate regulation is executed through hepatic FXR-agonism.7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar Presence of such mechanisms would be compatible with the extensive data on marked stimulation of BA synthesis in mice where genes such as FXR, FGF15, FGF receptor-4, and β-Klotho have been deleted.5Li J. et al.Biochim Biophys Acta. 2018; (pii: S0925–4439(18)30181–30189) ([Epub ahead of print])https://doi.org/10.1016/j.bbadis.2018.05.011Google Scholar, 7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar Administration of a potent nonsteroidal FXR agonist suppressing BA synthesis should eventually diminish the BA pool. In mice, FGF19 suppresses activity of the apical sodium-dependent BA transporter, facilitating ileal reabsorption.5Li J. et al.Biochim Biophys Acta. 2018; (pii: S0925–4439(18)30181–30189) ([Epub ahead of print])https://doi.org/10.1016/j.bbadis.2018.05.011Google Scholar FGF15/19 is also proposed to induce gallbladder relaxation and refilling.13Choi M. et al.Nature medicine. 2006; 12: 1253-1255Crossref PubMed Scopus (248) Google Scholar Based on these observations, we expected reduced postprandial serum BAs following Px-102. This was not observed, however, and instead both fasting and postprandial serum BAs increased (Figure 1; Supplementary Figure 3). A likely explanation is that hepatic BA uptake by transporters like sodium/taurocholate cotransporting polypeptide NTCP is rapidly reduced by hepatic FXR activation.5Li J. et al.Biochim Biophys Acta. 2018; (pii: S0925–4439(18)30181–30189) ([Epub ahead of print])https://doi.org/10.1016/j.bbadis.2018.05.011Google Scholar, 6Chavez-Talavera O. et al.Gastroenterology. 2017; 152: 1679-1694.e3Abstract Full Text Full Text PDF PubMed Scopus (594) Google Scholar In contrast, when 0.5 mg/kg Px-102 was given to humans for 7 days, serum levels of BAs decreased,12Deuschle U. et al.Int J Cancer. 2015; 136: 2693-2704Crossref PubMed Scopus (12) Google Scholar and similar findings were reported following 7 days of treatment with modified FGF19,11Luo J. et al.Science translational medicine. 2014; 6: 247ra100Crossref PubMed Scopus (135) Google Scholar underscoring the different response patterns following acute and chronic treatment. Suppression of CYP7A1 may result in elevated low-density lipoprotein (LDL) cholesterol, as seen following treatment with steroidal FXR agonists.14Albers J.J. et al.Gastroenterology. 1982; 82: 638-646PubMed Scopus (54) Google Scholar, 15Neuschwander-Tetri B.A. et al.Lancet. 2015; 385: 956-965Abstract Full Text Full Text PDF PubMed Scopus (1724) Google Scholar For CDCA, this is linked to lowered hepatic LDL receptors and reduced LDL clearance.10Ghosh Laskar M. et al.J Intern Med. 2017; 281: 575-585Crossref PubMed Scopus (46) Google Scholar CDCA also lowers very low-density lipoprotein triglyceride secretion, presumably through hepatic FXR activation; the possible involvement of FGF19 is still unclear, however.7Angelin B. et al.Cell Metab. 2012; 16: 693-705Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar Although the present study was not designed to study the effects of Px-102 on lipoprotein metabolism, it is clear that although total cholesterol was unchanged, high-dose treatment lowered postprandial triglycerides. The possible role of FXR-mediated effects on PCSK9 or apoC-III as seen for CDCA,10Ghosh Laskar M. et al.J Intern Med. 2017; 281: 575-585Crossref PubMed Scopus (46) Google Scholar or effects on fat absorption or transintestinal cholesterol excretion can only be speculated on. Further studies exploring lipoprotein metabolism following treatment with Px-102 and other FXR agonists will be important. To conclude, Px-102 rapidly reduces BA synthesis, and increases serum FGF19 dose-dependently. However, the response pattern challenges the concept that circulating FGF19 is a major mediator of FXR-activated suppression of BA synthesis in humans. Drs Claus Kremoser and Ulrich Deuschle from Phenex Pharmaceuticals AG, Heidelberg, Germany, kindly provided serum samples for our analyses. We thank Ingela Arvidsson and Lisbet Benthin for expert technical assistance. Author Contributions: Study concept and design: Bo Angelin, Mats Rudling. Acquisition of data: Amani Al-Khaifi. Analysis and interpretation of data: Bo Angelin, Mats Rudling, Amani Al-Khaifi. Drafting of the manuscript: Bo Angelin, Mats Rudling, Amani Al-Khaifi. Critical revision of the manuscript for important intellectual content: Bo Angelin and Mats Rudling. Statistical analysis: Amani Al-Khaifi. Obtained funding: Bo Angelin and Mats Rudling. Administrative, technical, or material support, study supervision: Bo Angelin and Mats Rudling. Parts of this work were presented at the Cell Symposia: Metabolic Disease Therapies in San Diego, California, October 2017 (Al-Khaifi A, Straniero S, Ghosh M, Rudling M, Angelin B. Rapid regulation of hepatic lipid metabolism by modulation of bile acid fluxes in humans). Download .pdf (1.16 MB) Help with pdf files Supplementary Material