Bile ductules and stromal cells express hedgehog ligands and/or hedgehog target genes in primary biliary cirrhosis


  • Potential conflict of interest: Nothing to report.


Indian Hedgehog (Ihh) regulates tissue morphogenesis. Hedgehog (Hh) activity has been demonstrated in human cholangiocarcinoma and hepatocellular carcinoma lines, and in myofibroblasts and progenitors from adult rodent livers. We evaluated Hh pathway involvement in the response to biliary injury in primary biliary cirrhosis (PBC). Liver sections from 3 PBC patients and 3 controls without liver disease were studied. Immunohistochemistry was used to determine if cells that accumulate in PBC livers express Ihh or Hh-target genes including the Hh-receptor, Patched (Ptc), and the Hh-transcriptional activator glioblastoma (Gli) 2. Positive cells were further identified by costaining for cytokeratin (CK) 19, a biliary marker, or OV6, a hepatic progenitor marker. In all subjects, Gli2 and Ptc expression localized in portal areas. The numbers of Gli2- or Ptc-expressing cells/portal triad were each 10-fold greater in patients with PBC than in controls (P < 0.05). In PBC livers, some CK19(+) cells coexpressed Gli2 or Ptc. Many stromal fibroblastic cells were also Gli2(+). Strong Ihh expression was detected in most bile ductular cells. Scattered stromal cells also expressed Ihh. The number of Ihh(+) cells/portal triad was 6-fold greater in PBC livers than controls (P < 0.05). OV6(+) progenitors increased significantly in PBC livers, and some of these cells coexpressed Ihh, Ptc, and/or Gli2. Conclusion: This is the first direct evidence that noncancerous, adult human livers harbor several types of cells that produce and/or respond to Hh ligands. Such Hh-responsive cells accumulate during the fibroproliferative response to chronic cholestatic liver injury, suggesting a role for Hh signaling in this process. (HEPATOLOGY 2007;45:1091–1096.)

Primary biliary cirrhosis (PBC) is a chronic cholestatic liver disease characterized by progressive loss of interlobular bile ducts. Immune-mediated destruction of bile duct epithelial cells is thought to drive the pathogenesis of PBC. Recurrent injury to these small bile ducts evokes a repair response that involves expansion of portal tracts with proliferating bile ductules. The latter are typically associated with a myofibroblastic response that results in net deposition of collagen matrix. With time, the fibroproliferative response bridges adjacent portal areas, leading to the typical biliary pattern of fibrosis and, eventually, biliary cirrhosis.1, 2

The mechanisms that mediate the fibroproliferative response to bile duct injury have been the focus of considerable research because understanding this process may suggest novel therapies to promote normal re-epithelialization of the injured biliary tree in PBC and other cholestatic liver diseases.2 It has been shown that immature cholangiocytes produce autocrine growth factors, such as epidermal growth factor and hepatocyte growth factor, as well as cytokines, such as platelet-derived growth factor-β, that induce the proliferation of myofibroblastic cells.1 These findings suggest that cross-talk between biliary epithelial cells and stromal cells may modulate the fibroproliferative response to biliary injury. Recent data from our laboratory provides further support for this concept. We found that cultured cholangiocytes and myofibroblastic cells derived from adult rodents release Hedgehog family ligands that promote the viability and growth of both cell types.3

Hedgehog (Hh) signaling is extensively involved in embryogenesis.4 Many types of progenitor cells express Patched (Ptc), the cell surface receptor for Hh ligands. Ligand-receptor interaction liberates the Ptc coreceptor, Smoothened (Smo), from the inhibitory influence of Ptc, permitting propagation of intracellular signals that culminate in the nuclear localization of Glioblastoma (Gli) DNA-binding proteins that regulate the transcription of Hh-responsive genes.4 Gli1 and Gli2 activate transcription, whereas Gli3 can function as a transcriptional activator or repressor depending on its posttranslational modification.5 Several components of the Hh signaling pathway, including Ptc, Gli1, and Gli2, are themselves Hh-regulated genes. Hence, expression of one or more of these gene products marks cells as being Hh-responsive.6

Hh pathway signaling is initiated when extracellular ligands such as Indian hedgehog (Ihh) interact with Ptc on Hh target cells.4 Although a study of early embryo explants demonstrated that induction of Hh ligands in ventral endoderm suppressed pancreatic differentiation and permitted hepatic specification of endodermal progenitor cells,7 little else is known about the role of Hh signaling in liver development. In contrast, there is growing evidence that this pathway regulates key aspects of morphogenesis in many other endoderm-derived tissues, including lung and the proximal gastrointestinal tract.8 In these tissues, Hh signaling between mesenchymal and epithelial cells also modulates remodeling responses after injury during adult life.4, 9 This information stimulated us to search for Hh-responsive cells in adult livers. We found that in adult rats and mice, hepatic stellate cells, myofibroblasts, and certain hepatic epithelial cells, including immature cholangiocytes, are capable both of producing and responding to Hh ligands. Moreover, hepatic mRNA levels of Ihh, Gli1, and Gli2 increase significantly as populations of Hh-responsive myofibroblasts and bile ductular cells expand after bile duct ligation.3, 9, 10 The purpose of this study was to determine if adult human livers also accumulate Hh-responsive cells when bile ducts are chronically injured. Given our findings in rodents, evidence published by another group demonstrating Hh transcriptional activity in some human cholangiocarcinoma cell lines,11 and the fact that myofibroblasts and cholangiocytes are the key cell types involved in the fibroproliferative response to chronic nonsuppurative cholangitis, the present analysis focuses on human PBC.


CK, cytokeratin; DAPI, 4′6′-diamindino-2-phenylindole; Gli, Glioblastoma; Hh, hedgehog; Ihh, Indian hedgehog; α-SMA, α-smooth muscle actin; PBC, primary biliary cirrhosis; Ptc, patched.

Patients and Methods

Human Subjects.

Anonymized liver sections were examined from 3 patients with PBC and 3 control patients without chronic liver disease who had liver resections for colorectal metastases. Tissues were obtained from the Duke University School of Medicine Tissue Bank Shared Resource and studied in accordance with NIH and institutional guidelines for human subjects research.


Specimens fixed in formalin and embedded in paraffin were cut into 4-μm sections, dewaxed, hydrated, subsequently incubated for 10 minutes in 3% hydrogen peroxide to block endogenous peroxidase. Antigen retrieval was performed by heating in 10 mM sodium citrate buffer (pH 6.0) for 10 minutes or incubation with 0.25% pepsin for 15 minutes. Sections were blocked in Dako protein block (X9090; Dako) for 30 minutes and incubated with primary antibodies, Gli2 (ab 26056, 1:500; Abcam Ltd.), Ihh (sc-13088, 1:250; Santa Cruz Biotechnology), and Ptc (sc-6149, 1:80; Santa Cruz Biotechnology) at 4°C overnight. Other sections were also incubated overnight at 4°C in nonimmune sera. Rabbit anti-goat IgG-HRP (sc-2922, 1:500; Santa Cruz Biotechnology) for Ptc and polymer/horseradish peroxidase–conjugated anti-rabbit (K4003; Dako) for Ihh and Gli2 were used as secondary antibody. DAB reagent was employed in the detection procedure. Omitting primary antibodies from the reactions eliminated staining, demonstrating staining specificity. For double immunofluorescent staining, frozen liver sections were used. Samples were fixed and permeabilized, saturated, and processed for immunostaining with primary antibody Ihh, Gli2, OV6 (MAB2020, 1:100; R&D Systems), CK19 (Dako) and α-smooth muscle actin (α-SMA) (M0851;Dako). Alexa Fluor 568 and Alexa Fluor 488 (Molecular Probes) were used as secondary antibodies. Counterstaining with 4′,6-diamidino-2-phenylindole (DAPI) was employed to demonstrate nuclei.

Cells with cytosol/membranes that strongly stained for Ihh or Ptc were considered to be positive for Ihh or Ptc. Because Gli2 is a transcription factor, only cells expressing nuclear Gli2 were considered to be Gli2(+). Cells expressing the respective Hh pathway antigens were counted in all portal tracts on liver sections from patients with PBC and individuals with normal livers (n = 3 per group). In each sample, the average number of positive cells was obtained by dividing the total number of positive cells by the total number of portal tracts. To further characterize the Hh-responsive population in PBC liver, the number of Gli2(+) cells with cytokeratin (CK) 19, OV6, or α-SMA were counted, and the relative ratio of Gli2(+)/α-SMA or OV6(+) cells to Gli2(+)/CK19(+) cells was derived.

Statistical Analysis.

Results are expressed as the mean ± SD. Statistical differences were determined by Student t test. P values < 0.05 were considered statistically significant.


We performed immunohistochemical analysis on liver sections from 3 adults with PBC to identify Hh-responsive cells (i.e., cells that were expressing Gli2 or Ptc). In all patients, the expression of Gli2 and Ptc localized in areas of proliferating bile ducts. Many more bile ductular cells in portal areas expressed Gli2 and/or Ptc in patients with PBC than in healthy controls, whose livers harbored only rare Hh-responsive cells (Figs. 1 and 2). Double immunofluorescent staining of human PBC liver samples revealed that CK19-positive bile ductular cells express Ptc (Fig. 3A) and Gli2 (Fig. 3B). Gli2 expression was also demonstrated in fibroblastic-appearing cells in the adjacent stroma (Fig. 2).

Figure 1.

Ptc expression in PBC and normal human livers. Representative sections from (A,B) patients with PBC and (C) a healthy control liver. Original magnification: (A,C) ×40; (B) ×100. (D) Ptc(+) cells (brown) were counted in all portal triads and expressed as number of Ptc(+) cells per portal triad (PT). Mean ± SD results from 3 samples per group are graphed. (*P < 0.05)

Figure 2.

Gli2 expression in PBC and normal human livers. Representative sections from (A,B) patients with PBC and (C) a healthy control liver. Original magnification: (A,C) ×40; (B) ×100. (D) Gli2(+) cells (brown) were counted in all portal triads and expressed as number of Gli2(+) cells per portal triad (PT). Mean ± SD results from 3 samples per group are graphed. (*P < 0.05)

Figure 3.

Ptc and Gli2 expression by CK19-positive cells. (A) Double immunofluorescent staining for Ptc (Alexa Fluor 568, red) and CK19 (Alexa Fluor 488, green). (B) Double immunohistochemical staining for Gli2 (Alexa Fluor 568, red) and CK19 (Alexa Fluor 488, green). Ptc and Gli2 are shown in cell membranes and nuclei, respectively, whereas CK19 is cytoplasmic. DAPI counterstaining (blue) demonstrates cell nuclei. Original magnification: (A) ×100; (B) ×100.

Because earlier findings in rodents with chronic biliary obstruction suggested that Ihh may act in an autocrine and/or paracrine fashion to initiate Hh signaling during cholestatic liver injury, we used immunohistochemical analysis to localize Ihh expression in liver sections from patients with PBC. In all samples examined, strong Ihh expression was detected in many bile ductular cells (Fig. 4). Scattered stromal cells also expressed this ligand (Fig. 4). Because the same cell populations contain Hh-responsive cells (demonstrated by expression of Gli2 and/or Ptc), Hh signaling may be involved in the autoregulation of bile ductular cells and myofibroblastic cells during cholestatic liver injury.

Figure 4.

Ihh expression in livers with PBC and normal human livers. Representative sections from (A,B) patients with PBC and (C) a healthy control liver. Original magnification: (A,C) ×40; (B) ×100. (D) Ihh(+) cells (brown) were counted in all portal triads and expressed as number of Ihh(+) cells per portal triad (PT). Mean ± SD results from 3 samples per group are graphed. (*P < 0.05)

It has been suggested that progenitor populations, including bipotent hepatoblasts (i.e., oval cells) that reside along canals of Hering, participate in the repair of injured bile ducts.12, 13 Many types of progenitor cells are Hh-responsive.4, 6, 7, 9 To determine if Hh-responsive cells reside among hepatic progenitor populations, sections were incubated with primary antibodies to both OV6, an oval cell cytokeratin marker,13 and Ptc or Gli2 to identify double-positive cells. Some Gli2-positive and Ptc-positive populations coexpress OV6 (Fig. 5), indicating that some of the hepatic progenitors that accumulate during chronic cholestatic liver injury are Hh-responsive. Next, we investigated whether these hepatic progenitor cells express Ihh ligand. Double immunofluorescent staining for Ihh and OV6 demonstrated that Ihh was expressed in OV6-positive hepatic progenitor cells (Fig. 6).

Figure 5.

Gli2 and Ptc expression by some OV6-positive cells in human PBC samples. Double staining for OV6 (Alexa Fluor 488, green) and Gli2 (A) or Ptc (B) (Alexa Fluor 568, red). Gli2 is in cell nuclei, whereas Ptc localizes in cell membranes. DAPI counterstaining (blue) demonstrates cell nuclei. Original magnification: (A) ×100; (B) ×100.

Figure 6.

Ihh expression by a portion of OV6-positive cells. OV6 (Alexa Fluor 488, green) and Ihh (Alexa Fluor 568, red) are stained. Because Ihh and OV6 are cytoplasmic, in merged images of Ihh and OV6 staining, cells that coexpress Ihh and OV6 are yellow. Original magnification: ×63.

To better characterize the population of Hh-responsive cells in PBC livers, cells coexpressing Gli2 and a marker for bile ductular cells (CK19), progenitors (OV6), or myofibroblasts (α-SMA) were counted (Fig. 7A). Only 5% of CK19(+) cells coexpressed Gli2. On the other hand, although PBC livers contained many fewer OV6(+) cells than did CK19(+) cells, the proportion of OV6(+) cells coexpressing nuclear Gli2 was 5-fold higher than the proportion of CK19(+) cells coexpressing Gli2. Nuclear Gli2 was also noted in cells that expressed neither CK19 nor OV6, some of which were α-SMA(+) (Fig. 7B). The proportion of α-SMA(+) cells coexpressing Gli2 was 6-fold greater than those coexpressing CK19 and Gli2. These results suggest that fibroblastic and progenitor populations in PBC livers are enriched with Hh-responsive cells, whereas more mature-appearing ductular cells mostly produce Hh ligand (Fig. 4).

Figure 7.

Coexpression fo GLi2 with markers for ductular cells (CK19), progenitors (OV6), or myofibroblastic cells (α-SMA). (A) Relative ratio of Gli2(+)/α-SMA or OV6(+) cells to Gli2(+)/CK19(+) cells. (*P < 0.05, **P < 0.005) (B) Double staining for α-SMA (Alexa Fluor 488, green) and Gli2 (Alexa Fluor 568, red). Original magnification: ×63.


In PBC, biliary epithelial cells are targeted for immune-mediated destruction. The resultant bile duct injury triggers reparative responses to reconstruct the damaged epithelia. This reconstruction often includes an exuberant “ductular response” that is characterized by periportal accumulation of proliferating bile ductular cells and associated stromal elements, including myofibroblastic cells and fibrous matrix. Various factors that are released from the damaged ductular epithelia appear to fuel the ductular response by encouraging the proliferation of cholangiocytes and/or myofibroblasts.1

This study identifies an additional trophic substance, namely Ihh, which is likely to be involved in the ductular response to cholestatic damage in PBC. This discovery is particularly exciting because Hh ligands are known to function as viability factors for many types of progenitors, and have been shown to play pivotal roles in tissue morphogenesis during fetal development.4, 7 Although there is growing evidence that Hh signaling modulates repair in many adult tissues,4 the possibility that this might occur in adult livers was ignored until very recently because Hh activity had not been demonstrated in mature hepatocytes or cholangiocytes. However, within the last year or so, several groups demonstrated Hh transcriptional activity in some human HCC and cholangiocarcinoma cell lines,6, 11, 14 and our laboratory reported that hepatic stellate cells, myofibroblasts, and certain immature but nonmalignant hepatic epithelial cells can produce and respond to Hh ligands.9, 10 These findings raised the intriguing possibility that like many other tissues, adult livers activate Hh signaling in response to certain types of injury. Indeed, while reviewing earlier literature on PBC, we learned that Shackel et al.15 reported that Ptc was the fourth most up-regulated transcript in their microarray analysis of PBC livers.

Using immunohistochemistry, we demonstrated that proliferating bile ductular cells are rich sources of Ihh ligand. Ihh may be involved in bile ductular cell autoregulation, because such cells sometimes express the Hh receptor Ptc, as well as other Hh target genes such as Gli2. Gli2(+) stromal cells are other potential targets for Ihh generated by hepatic progenitor cells, ductular cells, and/or the stromal cells themselves. Our earlier work showed that Hh ligands inhibited caspase activation and promoted the viability of rat myofibroblastic cells in culture.10 Therefore, it is conceivable that Hh ligands participate in epithelial-mesenchymal cross-talk that promotes the fibroproliferative response to bile duct injury. Indeed, we reported supranormal induction of α-SMA, collagen alpha-1, and fibronectin in bile duct–ligated mice with a genetic defect that limits appropriate down-regulation of Hh pathway activity.3

Like myofibroblastic cells, hepatic progenitor cells accumulate in close proximity to proliferating bile ductules in PBC.13 In PBC liver sections, we demonstrated expression of Ihh in oval cells (a type of hepatic epithelial progenitor). Also, we showed that some of these cells express Gli2 and Ptc. The latter finding suggests that Hh signaling is involved in autocrine and/or paracrine regulation of progenitor populations in adult livers. To our knowledge, this is the first direct evidence that Hh signaling regulates putative hepatic progenitors in adult humans.

Because we did not examine ductular reactions in other types of liver damage, such as severe parenchymal injury, it is uncertain if our findings are specific for PBC. This seems unlikely, however, given that similar responses were noted in rodents with another type of cholestatic liver damage (i.e., mechanical bile duct obstruction).3 Nevertheless, further research on the role of Hedgehog signaling in regulating the fate of biliary progenitors is certainly justified. Such knowledge will be particularly pertinent to PBC, because this disease is characterized by progressive loss of intralobular bile ducts. Why ductopenia develops in PBC, despite a seemingly robust ductular response to biliary epithelial injury, is unknown. Closer scrutiny of the Hedgehog pathway and other systems that modulate progenitor cell proliferation, apoptosis, and differentiation might identify novel therapeutic and diagnostic targets in patients with PBC.