Potential conflict of interest: Nothing to report.
This study was supported by a grant from the Dutch Digestive Diseases Foundation (MLDS WO 08-69; to P.L.J. and F.G.S.).
Fibroblast growth factor 19 (FGF19) plays a crucial role in the negative feedback regulation of bile salt synthesis. In the postprandial state, activation of ileal farnesoid X receptor (FXR) by bile salts results in transcriptional induction of FGF19 and elevation of circulating FGF19 levels. An intestinal-liver axis of FGF19 signaling results in down-regulation of bile salt synthesis. The aim of this study was to explore a broader signaling activity of FGF19 in organs engaged in the enterohepatic circulation of bile salts. For this aim, FGF19 expression and aspects of FGF19 signaling were studied in surgical specimens and in cell lines of hepatobiliary and intestinal origin. FGF19 messenger RNA was found to be abundantly expressed in the human gallbladder and in the common bile duct, with only minor expression observed in the ileum. Interestingly, human gallbladder bile contains high levels of FGF19 (21.9 ± 13.3 versus 0.22 ± 0.14 ng/mL in the systemic circulation). Gallbladder explants secrete 500 times more FGF19 than FXR agonist-stimulated ileal explants. Factors required for FGF19 signaling (i.e., FGFR4 and βKlotho) are expressed in mucosal epithelial cells of the gallbladder and small intestine. FGF19 was found to activate signaling pathways in cell lines of cholangiocytic, enteroendocrine, and enterocytic origin. Conclusion: The combined findings raise the intriguing possibility that biliary FGF19 has a signaling function in the biliary tract that differs from its established signaling function in the portal circulation. Delineation of the target cells in bile-exposed tissues and the affected cellular pathways, as well as a possible involvement in biliary tract disorders, require further studies. (HEPATOLOGY 2012)
The family of fibroblast growth factors (FGFs) comprises subfamilies of canonical, intracellular, and hormone-like FGFs.1 The latter group contains three secreted FGFs (i.e., FGF19, FGF21, and FGF23) that are devoid of mitogenic activity, but have a metabolic function.2-4 Unlike their canonical counterparts, the hormone-like FGFs lack a heparin-binding site and, therefore, have little affinity for the extracellular matrix. This enables their release into the circulation and allows an endocrine mode of action. Other distinguishing features of these endocrine FGFs are the requirement of Klotho family members for signaling via FGF receptors (FGFRs) and the involvement of nuclear receptors in regulating their expression.
FGF19 mediates the negative feedback regulation of bile salt synthesis by bile salts.5-7 After gallbladder contraction, bile salts enter the small intestinal lumen, where they aid in the breakdown and uptake of dietary lipids. Bile salts are efficiently reclaimed by enterocytes in the terminal ileum, resulting in activation of ileal FXR and concomitant up-regulation of FGF19 gene expression and release of FGF19 into the portal circulation. Binding of FGF19 to the hepatocytic cell-surface receptor, FGFR4, activates intracellular pathways that target the expression of the cytochrome P450 7A1 gene that encodes the rate-limiting enzyme in bile salt synthesis.5, 8 Postprandial elevation of circulating FGF19 signals the return of bile salts to the liver and serves to reduce bile salt synthesis.9, 10 In addition, FGF19 has been implicated in terminating the enterohepatic cycling of bile salts in mice by inducing gallbladder relaxation and refilling.11
FGF19 and its murine counterpart, Fgf15, have a restricted pattern of expression, with expression in the adult mouse confined to the small intestine (ileum>jejunum).12-14 FGF19/Fgf15 is not expressed in the healthy adult liver.14, 15 However, compensatory hepatic expression is observed in patients with an interrupted enterohepatic circulation, but this does not occur in bile duct–ligated mice.15 Treatment of a small group of healthy volunteers with the bile salt sequestrant cholestyramine greatly reduced serum FGF19 levels,10 suggesting that the ileum is the main source of circulating FGF19 in health. However, FGF19 messenger RNA (mRNA) expression was also observed in a number of fetal human tissues as well as in adult human gallbladder.13 Quantitative data on FGF19 expression levels in the respective tissues have not been reported. Nonetheless, the presence of FGF19 mRNA in human gallbladder is surprising in view of the alleged role of FGF19 in gallbladder relaxation. In the current study, we sought to determine the relative expression of FGF19 in the human biliary tract and small intestine. Having noted that human bile contains high levels of FGF19, we further explored a signaling role of biliary FGF19.
Peroperative bile and tissue specimens were collected from patients undergoing elective surgery in the abdominal region. Gallbladder bile was collected from patients undergoing laparoscopic cholecystectomy for symptomatic gallstone disease (n = 5) or pancreaticoduodenectomy for periampullar malignancies (n = 6). The majority of patients undergoing pancreaticoduodenectomy were noncholestatic at the time of operation. In a number of these patients (n = 4), hepatic bile was also sampled from the common hepatic bile duct after removal of the surgical clamp at the time of hepaticojejunostomy. Gallbladder tissue specimens (n = 15) were procured from aforementioned groups of patients. In addition, common bile duct (n = 5) and duodenal tissue (n = 3) was obtained from the patients scheduled for pancreaticoduodenectomy. Terminal ileum specimens (n = 4) were acquired from patients undergoing right hemicolectomy for nonobstructive colon carcinoma. For reference purposes, previously collected wedge liver biopsies (n = 5) taken from patients undergoing liver resection for benign processes were included in this study.15 All tissue specimens were judged to be free of gross inflammatory or malignant processes. Bile specimens and RNA later-stabilized tissues were stored at −80°C until further processing. All patients gave informed consent for the use of bile and surgical specimens.
Detailed descriptions of the experimental procedures are provided in the Supporting Materials.
Nonparametric t testing (Mann-Whitney) and analysis of variance were used to test for differences between groups. Statistical significance was accepted at P < 0.05. Data are expressed as means ± standard deviation.
The Gallbladder Is the Main Site of FGF19 Expression in the Enterohepatobiliary System.
To extend the initial observation that FGF19 is expressed in the adult human gallbladder,13 we determined the relative expression of FGF19 in enterohepatobiliary tissues (Fig. 1A). Using our methodology, quantification of FGF19 mRNA level was found to be reliable in complementary DNA (cDNA) samples with threshold cycle (Ct) values <35 (data not shown). FGF19 mRNA was considered absent in cDNA samples with Ct values >40. FGF19 was found to be expressed in the small intestine in both duodenum (Ct range: 36-39) and ileum (Ct range: 36-37), with quantification being hampered by high Ct values. These low levels of expression are likely a consequence of procuring specimens in a fasted state during surgery. Expression in the common bile duct and gallbladder was notably higher, with respective median values 13- and 400-fold higher than in the ileum. In line with earlier findings,15 FGF19 was not expressed in healthy liver (data not shown). Immunohistochemistry (IHC) revealed that FGF19 is expressed in gallbladder epithelial cells (Fig. 1B). A cytoplasmic, dot-like staining pattern re-miniscent of secretory vesicles is apparent.
Human Gallbladder Bile Contains High Levels of FGF19.
To explore the possibility that gallbladder epithelial cells release FGF19 into bile, we determined FGF19 levels in gallbladder bile. FGF19 levels in gallbladder bile (21.9 ± 13.3 ng/mL; range, 4.2-47.4) were considerably higher than either fasted (0.22 ± 0.14 ng/mL; range, 0.043-0.48) or postprandial (0.84 ± 0.43 ng/mL; range, 0.28-1.60; taken from reference9) levels found in the systemic circulation (Fig. 2A). Interestingly, analysis of four paired samples of hepatic (7.7 ± 2.1 ng/mL) and gallbladder bile (28.0 ± 14.9 ng/mL) revealed that hepatic bile also contains high levels of FGF19 in comparison with circulating levels. FGF19 levels in gallbladder bile were comparable in patients with symptomatic gallstone disease and periampullar malignancies (22.1 ± 17.3 versus 21.8 ± 11.8 ng/mL, respectively; P = 0.92). The presence of FGF19 in hepatic bile is in line with strong IHC staining of bile plugs in cholestatic liver sections (data not shown).
To ascertain that the above-described observations were not the result of aberrant performance of the enzyme-linked immunosorbent assay (ELISA), we tested for interference by biliary components and checked the molecular mass of the biliary protein re-cognized by the ELISA antibodies. The employed ELISA has a high sensitivity (linear detection range: 0.02-0.60 ng/mL), requiring at least 20-fold dilution of bile samples. Spiking of diluted bile samples resulted in full recovery of added recombinant FGF19, indicating that biliary components did not interfere with the FGF19 ELISA (data not shown). Furthermore, proteins with similar molecular mass (22 kDa) were immunoprecipitated from gallbladder bile and recombinant FGF19 solutions by the employed ELISA antibodies (Fig. 2B). The above-described findings demonstrate that the observation of high levels of FGF19 in gallbladder bile is authentic.
The Ex Vivo Production of FGF19 Is Highest in Gallbladder Mucosal Explants.
FGF19 is abundantly expressed in the gallbladder, and gallbladder bile contains high levels of FGF19 protein. In contrast, basal expression of FGF19 in ileum, which is considered to be the source of FGF19 in the circulation, is quite low. Short-term culture experiments were conducted to address FGF19 protein production in ileal and gallbladder tissue explants. FGF19 production was defined as the amount of FGF19 secreted in the medium during the incubation period. FGF19 production by ileal explants was only measurable when FXR agonist (100 μM of chenodeoxycholate or 10 μM of GW4064) was applied to the incubation medium and amounted to 0.14-0.16 pg FGF19/mg tissue (Fig. 3A). This appears in line with the low level of FGF19 expression in (unstimulated) ileum (Fig. 1A). FGF19 production by gallbladder explants was markedly higher (75.0 ± 42.0 pg FGF19/mg tissue). Only a single specimen of common bile duct could be procured for these measurements, showing intermediate FGF19 production (9.4 pg FGF19/mg tissue). The observed differences in tissue FGF19 production are roughly proportional to the differences in basal FGF19 mRNA level.
To address whether FXR regulates FGF19 expression in the gallbladder, gallbladder explants were treated with the synthetic FXR agonist GW4064. This resulted in the induction of both the prototypical FXR target, small heterodimer partner (SHP) (2.4-fold; P = 0.002) and FGF19 (2.2-fold; P = 0.001) (Fig. 3B). However, there was no effect of GW4064 on FGF19 protein secretion (28.9 ± 16.1 versus 31.7 ± 23.9 pg FGF19/mg tissue; P = 0.62) (Fig. 3C). Pilot experiments indicated that GW4064 stimulation of FGF19 protein production required extended incubation times (data not shown). Note that the presence of a FGF19-deficient serosal layer in the gallbladder explants used in this particular experiment accounted for the somewhat lower values for FGF19 protein secretion in comparison with the values reported in Fig. 3A.
FGF19 Signaling Components Are Expressed Throughout the Enterohepatobiliary System.
The high level of FGF19 in hepatic and gallbladder bile suggests that FGF19 can activate signaling cascades in tissues in contact with bile. Enterohepatobiliary tissues were analyzed for the presence of mRNAs encoding components required for FGF19 signaling (i.e., FGFRs and Klotho family members).16-18 Relative tissue expression levels of these signaling components and the FGFR profile of the examined tissues are depicted in Fig. 4. FGF19 signaling proceeds via FGFR4 and requires βKlotho,19 both of which are most abundantly expressed in the liver, an established FGF19 target organ. Expression levels of these factors were considerably lower in the duodenum, ileum, gallbladder, and common bile duct. FGF19 signaling may also proceed via FGFRs other than the classical FGF19 receptor (FGFR4).19-21 Although tissue expression levels varied, alternative receptors (e.g., FGFR1, FGFR2, and FGFR3), as well as αKlotho, could be detected in all examined tissues.
FGFR4 and βKlotho Are Expressed in Gallbladder Mucosal Cells and Intestinal Goblet Cells.
IHC was employed to further delineate the site of expression of FGFR4 and βKlotho in gallbladder and small intestine. Intense staining for both FGFR4 and βKlotho was observed in gallbladder epithelium (Fig. 5, upper panels), with little or no staining of submucosal layers. In addition to a strong staining of the apical plasma membrane, a granular staining pattern resembling endoplasmic reticulum/Golgi structures was observed for FGFR4, whereas βKlotho showed strong cytoplasmic staining. In the small intestine, mucosa cells showed strong apical plasma membrane staining, whereas goblet cells showed additional staining of intracellular components (Fig. 5, lower panels). βKlotho showed cytoplasmic reactivity throughout the mucosal epithelial layer. The specificity of the antibodies employed in Figs. 1B and 5 was confirmed in parallel immunostainings, where the respective first antibodies were omitted (Supporting Fig. 1).
FGF19 Activates Signaling Cascades in Cholangiocyte and Intestinal Cell Lines.
The occurrence of FGFR4 and βKlotho in mucosal cells of bile-exposed tissues implies that these cells may be responsive to biliary FGF19. To test this idea, cultured cells were exposed to recombinant FGF19 and readouts for FGF19 responsiveness were assessed (i.e., activation of the extracellular signal-related kinase [ERK]1/2 pathway and transcriptional induction of early growth response protein 1; EGR1). Treatment with FGF19 resulted in the phosphorylation of ERK1/2 in human H69 cholangiocytes, rat AR42j pancreatic acinar cells, mouse STC-1 enteroendocrine cells, and human HT29 enterocytes (Fig. 6A). The FGFR and Klotho mRNA expression profile of these cell lines is depicted in Supporting Fig. 2. Induction of EGR1 mRNA was observed in AR42j and STC-1 cells stimulated with FGF19 (Fig. 6B). Furthermore, treatment of STC-1 cells with FGF19 resulted in a modest (1.3-fold; P < 0.001) elevation of cholecystokinin (CCK) levels in conditioned medium (Fig. 6C).
In this study, we demonstrated that human bile contains high levels of the endocrine factor FGF19. To the best of our knowledge, the abundant presence of a signaling factor in bile is unprecedented. Factors indispensable for FGF19 signaling (i.e., FGFR4 and βKlotho) were found to be expressed in mucosal epithelial cells of the gallbladder and the small intestine. Moreover, FGF19 signaling could be demonstrated in cell lines representing distinct cell types encountered in bile-exposed tissues. The above-described findings implicate a role for biliary FGF19 in the regulation of yet unknown processes in the enterohepatobiliary tract.
Despite an early report on FGF19 expression in the adult human gallbladder,13 this finding has thus far not been followed up. Earlier work on FGF19 (or its rodent counterpart, Fgf15) established that FGF19 is an enteric factor that affects hepatic bile salt and lipid synthesis and —at least in mice— induces gallbladder relaxation.5, 11, 22 We noted that FGF19 expression in the ileum was dwarfed by expression in the gallbladder and the common bile duct, a heretofore unknown site of FGF19 expression (Fig. 1A). In contrast with the situation in humans, Fgf15 mRNA was virtually undetectable in murine gallbladder (Ct>35; data not shown). It is currently unknown what drives FGF19 expression in gallbladder mucosa. Analogous to the regulation observed in the ileum, abundant expression of FXR in gallbladder mucosa (data not shown) suggests that this bile salt–activated transcription factor is involved in regulating FGF19 expression in the gallbladder. This notion is supported by the induction of FGF19 mRNA after treatment of gallbladder explants with an FXR agonist (Fig. 3B). Abundant expression of the apical sodium-dependent bile salt transporter in gallbladder mucosa (data not shown) is likely to result in efficient uptake of bile salts, the main FXR-activating ligands. FXR involvement may explain part of the large difference in FGF19 expression in the ileum and gallbladder. In the fasted state, the bile salt pool is largely stored in the gallbladder, resulting in maximal FXR target gene expression in the gallbladder. In line with such notion, FGF19 mRNA expression and protein secretion is barely detectable in unstimulated ileal explants, but is highly inducible in FXR agonist-stimulated ileal explants (Fig. 3A).
The circulating levels of FGF19 in control subjects (0.22 ng/mL) are two orders of magnitude lower than the levels of FGF19 that we have found in human gallbladder bile (21.9 ng/mL). The likely sources of biliary FGF19 are the gallbladder and the extrahepatic bile duct, both exhibiting high levels of FGF19 mRNA expression. Taking into account tissue mass, the gallbladder seems to be the main contributor to the pool of FGF19 in bile. In the gallbladder, FGF19 protein was present in cytoplasmic granules reminiscent of secretory vesicles (Fig. 1B). It is currently unclear whether the gallbladder contributes to the circulating pool of FGF19. The large effect of bile salt sequestrants on serum FGF19 levels,10 however, argues against a major effect of tissues other than the ileum on circulating FGF19 levels. Analysis of plasma FGF19 levels in patients before and after cholecystectomy could shed further light on the contribution of the gallbladder to the pool of FGF19 in the systemic circulation.
Interestingly, FGF19 was also found at considerable levels in hepatic bile (Fig. 2A). It should be noted that the collected hepatic bile was not flowing unimpeded; rather, hepatic bile was obtained from the common hepatic bile duct after removal of a clamp that had been in place for an estimated 1-2 hours. Although the healthy liver does not appear to express FGF19 mRNA, the above-described method of hepatic bile sampling may have resulted in some degree of local cholestasis with concomitant induction of hepatic FGF19, a phenomenon that we observed earlier in patients with extrahepatic cholestasis.15 Nonetheless, it cannot be excluded that FGF19 is expressed in the intrahepatic biliary tree under conditions of unobstructed bile flow. Such expression may be limited to larger biliary ducts not typically present in the needle or wedge liver biopsies that were used to demonstrate the absence of FGF19 mRNA in the healthy liver.15 A less likely explanation for the presence of FGF19 in hepatic bile would be the transhepatocellular transport and subsequent concentration of FGF19 derived from the circulation. Irrespective of the exact origin of FGF19 in hepatic and gallbladder bile, the presence of high levels of this endocrine factor strongly suggests a functional role of biliary FGF19.
FGF19 signaling proceeds via FGFR4, a ubiquitously expressed protein, and is dependent on ßKlotho, a protein with a more restricted expression pattern.14 Although our IHC studies focused on identifying cells expressing the high-affinity FGF19 receptor, FGFR4, the high levels of FGF19 in bile may also allow signaling via FGFRs with lower affinity for FGF19. At the mRNA level, alternative FGFRs were found to be expressed throughout the enterohepatobiliary system (Fig. 4). The liver appears to be the main target for FGF19 action in mice.16 However, in the present study, we observed strong expression of both FGFR4 and ßKlotho in mucosal epithelial cells of the gallbladder and small intestine (Fig. 5), suggestive for an extrahepatic action of FGF19. Intriguingly, a direct effect of FGF19/Fgf15 on gallbladder relaxation and filling has been previously demonstrated in mice, implying an effect of FGF19 on the smooth muscle layer.11 It is currently unknown whether FGF19 has a similar action on the human gallbladder. Though we could not detect significant expression of Fgf15 mRNA in the murine gallbladder (i.e., Ct>35; data not shown), the abundant expression of FGF19 in the human gallbladder raises questions about a possible action of FGF19 in the dilatation of the gallbladder in humans. Strict separation of bile and portal blood compartments may allow a relatively low level of FGF19 in the circulation to act on gallbladder smooth muscle in the presence of high levels of FGF19 in bile. Further studies are required to shed light on this issue. Notwithstanding the above, the expression of FGFR4 and ßKlotho in epithelial cells suggests that biliary FGF19 does have an effect on human gallbladder mucosa. FGF19 responsiveness could be demonstrated in a number of cell lines representing cell types encountered in enterohepatobiliary tissues, but has thus far not been explored in relevant tissues.
The following concept is emerging from the present findings. Hepatic bile and gallbladder bile contain high levels of the signaling factor FGF19. After gallbladder contraction and relaxation of the sphincter of Oddi, bile rich in FGF19 flows through the common bile duct towards the duodenum and further downstream. Biliary FGF19 interacts with FGFRs to initiate signaling pathways in ßKlotho-expressing cells lining this trajectory. Dilution with pancreatic juices and stomach contents and/or intraluminal proteolytic degradation is expected to limit the action of FGF19 towards the proximal parts of the small intestine.
What could be the function of FGF19 in bile? Although the cellular processes affected by FGF19 signaling in the studied cell lines have not been characterized as of yet, it is tempting to speculate that biliary FGF19 protects against detrimental effects of biliary bile salts. This would broaden the protective role of FGF19 observed earlier in the cholestatic liver.15 This postulated function may be especially relevant in structures exposed to concentrated bile (e.g., the gallbladder) and may involve FGF19-mediated regulation of mucin expression. The ERK1/2 pathway activated by FGF19 signaling has been implicated in the induction of mucin expression in the lung and intestines.7, 23 The actual involvement of FGFs in regulating mucin expression in the gallbladder —or other tissues— has not been described thus far. Through its recently described anti-inflammatory properties, FGF19/FGFR4 signaling may counteract the proinflammatory effects of certain bile salts, thus conferring protection of bile-exposed structures.24 The postulated protective function of biliary FGF19 may be less relevant for mice, a species that has a less toxic bile salt pool composition and is apparently devoid of biliary Fgf15 protein. Apart from a hypothesized protective function, biliary FGF19 may modulate the secretion of entero-endocrine factors. This line of thought is supported by the observation of FGF19-stimulated CCK release by enteroendocrine STC-1 cells (Fig. 6C) and would require an intraluminal action of FGF19. Such modus operandi is not unprecedented. Earlier work established that gastric juice–derived leptin stimulates secretion of CCK by duodenal cells and may even act way further downstream by regulating mucin expression in colonic goblet cells.25, 26 Stimulation of I-cell CCK release by biliary FGF19 may be a feed-forward signal to maximize gallbladder emptying and, hence, minimize bile stasis, a risk factor for the development of gallstones. The latter is a frequent complication of obesity-related disorders that appear to be accompanied by —at least in the liver— impaired FGF19 responsiveness.9 It is currently unexplored whether genetic variation in the FGF19 gene is a risk factor for cholelithiasis, cholecystitis, or biliary pancreatitis.
In conclusion, human bile contains high levels of the endocrine factor FGF19. Components required for FGF19 signaling are expressed in cell types in tissues in contact with bile. Experiments with cell lines identified as FGF19 responsive could shed light on the postulated protective roles of biliary FGF19.
The authors are indebted to Drs. Shanna Tol and the surgeons at the Academic Medical Center (Amsterdam, The Netherlands) for their assistance in collecting surgical specimens.