If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to Guido Carpino, M.D., Ph.D., Division of Health Sciences, Department of Movement, Human and Health Sciences, University of Rome “Foro Italico,” Piazza Lauro De Bosis 6, 00135 Rome, Italy.
Norwegian PSC Research Center, Division of Cancer, Surgery and Transplantation, Department of Transplantation Medicine, Oslo University Hospital Rikshospitalet, Oslo, Norway
Norwegian PSC Research Center, Division of Cancer, Surgery and Transplantation, Department of Transplantation Medicine, Oslo University Hospital Rikshospitalet, Oslo, NorwayResearch Institute of Internal Medicine, Division of Surgery, Inflammatory Diseases and Transplantation, Oslo University Hospital Rikshospitalet and Institute of Clinical Medicine, University of Oslo, Oslo, Norway
Primary sclerosing cholangitis (PSC) and primary biliary cholangitis (PBC) are human primary cholangiopathies characterized by the damage of mature cholangiocytes and by the appearance of ductular reaction (DR) as the results of hepatic progenitor cell activation. This study evaluated the differences in progenitor cell niche activation between these two cholangiopathies. Liver tissue was obtained from healthy liver donors (n = 5) and from patients with PSC (n = 20) or PBC (n = 20). DR, progenitor cell phenotype, and signaling pathways were investigated by IHC analysis and immunofluorescence. Our results indicated that DR was more extended, appeared earlier, and had a higher proliferation index in PBC compared with PSC. In PBC, DR was strongly correlated with clinical prognostic scores. A higher percentage of sex determining region Y–box (SOX)9+ and cytokeratin 19+ cells but fewer features of hepatocyte fate characterized progenitor cell activation in PBC versus PSC. Lower levels of laminin and neurogenic locus notch homolog protein 1 but higher expression of wingless-related integration site (WNT) family pathway components characterize progenitor cell niche in PSC compared with PBC. In conclusion, progenitor cell activation differs between PSC and PBC and is characterized by a divergent fate commitment and different signaling pathway predominance. In PBC, DR represents a relevant histologic prognostic marker.
Primary sclerosing cholangitis (PSC) is a chronic cholangiopathy characterized by diffuse inflammation and fibrosis of the bile ducts.
The progressive injury primarily involves the extrahepatic and large intrahepatic bile ducts, leading to concentric periductal fibrosis and obliterating strictures.
Biliary tree stem/progenitor cells in glands of extrahepatic and intrahepatic bile ducts: an anatomical in situ study yielding evidence of maturational lineages.
Massive BTSC proliferation and peribiliary gland hyperplasia were revealed in PSC but not in primary biliary cholangitis (PBC). Peribiliary gland activation resulted in obliterative concentric fibrosis through the accumulation of myofibroblasts in the walls of the extrahepatic and large intrahepatic bile ducts in PSC.
Moreover, PSC and PBC share the appearance of ductular reaction (DR) within the liver parenchyma. DR is thought to represent the activation of the hepatic stem/progenitor cell (HpSC) niche that resides within the bile ductules.
In the HpSC niche, the neurogenic locus notch homolog protein (NOTCH) and wingless-related integration site (WNT) family pathways can be differently modulated based on disease etiology and stage.
However, differences in HpSC response and DR phenotype between PSC and PBC have not been yet investigated.
Therefore, the aims of the present study were to compare PSC and PBC in terms of i) DR extent and HpSC activation in progressive disease stages, ii) DR phenotype and commitment, and iii) expression of signaling pathways by HpSCs.
Materials and Methods
Samples
Formalin-fixed and paraffin-embedded liver tissue from patients with PSC or PBC or from healthy livers was obtained from the Norwegian PSC Research Center, Oslo University Hospital Rikshospitalet (Oslo, Norway), from the Sapienza University of Rome (Rome, Italy), and from the University of Padua (Padua, Italy).
Specimens included i) livers with normal findings on histologic examination (controls) from liver donors (n = 5); ii) PSC (n = 20), obtained from liver biopsy samples and from explanted livers; and iii) PBC (n = 20), from liver biopsy samples and from explanted livers.
Informed consent was obtained from each patient, and the study protocol was conformed to the ethics guideline of the 1975 Declaration of Helsinki. The research protocol was reviewed and approved by the ethics committee at Umberto I Policlinico (Rome, Italy), and the regional committees for medical and health research ethics in southeastern Norway. No donor organs were obtained from executed prisoners or other institutionalized individuals.
Clinical Parameters
The diagnostic criteria for PSC and PBC followed the Clinical Practice Guidelines from the European Association for the Study of the Liver.
Clinical and biochemical parameters were collected during biopsy or during transplantation, and, if appropriate, 1 year after ursodeoxycholic acid administration. Clinical parameters were collected to determine prognostic classes according to the Mayo Clinic's Revised Natural History Model for PSC and the Updated Natural History Model for PBC, while the Global PBC
Global PBC Study Group Development and validation of a scoring system to predict outcomes of patients with primary biliary cirrhosis receiving ursodeoxycholic acid therapy.
UK-PBC Consortium The UK-PBC risk scores: derivation and validation of a scoring system for long-term prediction of end-stage liver disease in primary biliary cholangitis.
risk scores were used to stratify the risk for progression in PBC patients. Patients who presented with overlapping syndromes or concomitant liver diseases, comprising alcoholic steatohepatitis or nonalcoholic fatty liver disease, were excluded.
Light Microscopy, Histopathology, and IHC Analysis
Specimens were fixed in 10% buffered formalin for 2 to 4 hours and embedded in low-temperature–fusion paraffin (55°C to 57°C), and 3- to 4-μm sections were stained with hematoxylin and eosin and Sirius Red/Fast Green, according to standard protocols. PSC and PBC samples were staged in accordance to the common histologic scoring systems
and further categorized as early (not cirrhosis) or cirrhosis. The extent of fibrosis was evaluated in Sirius Red/Fast Green stains, as detailed previously.
Docosahexanoic acid plus vitamin D treatment improves features of NAFLD in children with serum vitamin D deficiency: results from a single centre trial.
Briefly, Sirius Red/Fast Green–stained slides were scanned by a digital scanner (ScanScope CS System, Aperio Technologies, Inc., Oxford, UK) and processed by ImageScope software version 11.2 (Leica Biosystems, Milan, Italy). An image analysis algorithm was used to quantify the proportion of Sirius Red–stained area. The algorithm was applied on the entire section. The extent of collagen deposition was expressed as the proportion (%) of Sirius Red–stained area with respect to the total biopsy area, providing a quantitative value on a continuous scale.
For immunohistochemistry analysis, endogenous peroxidase activity was blocked by a 30-minute incubation in methanolic hydrogen peroxide (2.5%). Antigens were retrieved, as indicated by the vendor, by applying proteinase K (catalog number S3020; Dako Cytomation, Glostrup, Denmark) for 10 minutes at room temperature. Sections were then incubated overnight at 4°C with primary antibodies (Table 1). Samples were rinsed twice with phosphate-buffered saline for 5 minutes and incubated for 20 minutes at room temperature with secondary biotinylated antibody and then with streptavidin–horseradish peroxidase (LSAB + System-HRP; catalog number K0690; Dako Cytomation). Diaminobenzidine (Dako Cytomation) was used as a substrate, and sections were counterstained with hematoxylin.
For immunofluorescence, nonspecific protein binding was blocked by 5% normal goat serum. Specimens were incubated with primary antibodies. Specimens were washed and incubated for 1 hour with labeled isotype-specific secondary antibodies (anti-mouse Alexa Fluor 546, anti-mouse Alexa Fluor 488, anti-rabbit Alexa Fluor 488, and anti-goat Alexa Fluor 546; Invitrogen, Life Technologies Ltd, Paisley, UK) and counterstained with DAPI for visualization of cell nuclei.
For all immunoreactions, negative controls (the primary antibody was replaced with preimmune serum) were also included. Sections were examined in a coded fashion by DM 4500 B Light and Fluorescence Microscopy (Leica Microsystems, Weltzlar, Germany), equipped with a Prog Res C10 Plus Videocam (Jenoptik, Jena, Germany). Immunofluorescence stains were also analyzed by confocal microscopy (model TCS-SP2; Leica Microsystems). Slides were further processed with Image Analysis System software version 7 (Delta Sistemi, Rome, Italy) and were independently evaluated by two researchers (G.C. and P.O.) in a blinded fashion.
Table 1List of Antibodies Used in This Study
Name
Host
Supplier
Catalog no.
Dilution
CD326/EpCAM
Mouse
Dako Cytomation
M3525
1:100
Cytokeratin 7
Mouse
Dako Cytomation
M7018
1:100
Cytokeratin 19
Mouse
Dako Cytomation
M0888
1:100
Hep-Par1
Mouse
Dako Cytomation
M7158
1:100
Laminin
Rabbit
Dako Cytomation
Z0097
1:25
Notch1
Goat
Santa Cruz Biotechnology, Heidelberg, Germany
Sc-6014
1:50
pβ-Catenin
Mouse
Cell Signaling Technology, Leiden, the Netherlands
#4176S
1:100
PCNA
Mouse
Dako Cytomation
M0879
1:100
SOX9
Rabbit
EMD Millipore, Darmstadt, Germany
AB5809
1:200
WNT1
Goat
Santa Cruz
Sc.6280
1:50
WNT3a
Mouse
Santa Cruz
Sc-136163
1:50
EpCAM, epithelial cell adhesion molecule; Hep-Par1, hepatocyte paraffin 1; Notch1, neurogenic locus notch homolog protein 1; p-, phosphorylated; PCNA, proliferating cell nuclear antigen; SOX9, sex determining region Y-box (SOX)9; WNT, wingless-related integration site family.
The extent of DR was evaluated by cytokeratin (CK)-7 immunoreactivity. CK7-stained slides were scanned by a digital scanner (ScanScope CS System; Aperio Digital Pathology, Leica Biosystems) and processed by ImageScope software.
The area occupied by CK7+ cells was quantified by an image analysis algorithm. The extent of DR was expressed as the percentage of the parenchymal area occupied by reactive ductules.
Cholangiocytes lining the interlobular bile ducts were excluded from the counts. CK7-stained slides were further counterstained with Sirius Red for better visualization of the spatial relationship between DR and fibrosis.
With regard to DR phenotype, sex determining region Y–box (SOX)9 and epithelial cell adhesion molecule (EpCAM) were used as stem/progenitor cell markers; the coexpression of CK7 and CK19 was considered indicative of the maintenance of a biliary phenotype, and the expression of the hepatocyte marker hepatocyte paraffin 1 (Hep-Par1) and the loss of the biliary CK19 expression were considered as signs of hepatocyte commitment.
The DR proliferation index (PI) was automatically calculated as the percentage of DR cells expressing the proliferating cell nuclear antigen by specific algorithms.
For all other immunoreactions, the number of positive cells was automatically calculated by an algorithm on the entire section and, then, a semiquantitative scoring system was applied (0 = <1%; 1 = 1% to 5%; 2 = 6% to 30%; 3 = 31% to 50%; and 4 = >50%).
Immunofluorescence stains were scanned with a digital scanner (ScanScope FL System; Aperio Technologies) and processed with ImageScope software. An algorithm was used to quantify the proportion of positive pixel area for a single fluorophore or the area with colocalization of two fluorophores.
Intermediate EpCAM+ Hepatocytes and Bud Definition
Buds were defined as hepatocytes, single or in clusters, immediately adjacent to cholangiocytes and located in fibrous septa (ie, in regions with parenchymal extinction). Buds were classified as levels 0 to 4. Buds of each maturation level (1 to 4) were counted on the CK19 slides.
To quantitate the presence of apparent bud-derived tissue, we defined the immaturity index (IMI)-EpCAM as the presence of hepatocytes with EpCAM positivity.
The presence of EpCAM+ hepatocytes was further confirmed by double immunofluorescence for EpCAM and Hep-Par1 and scored as: 0 = no positive cells; 1 (level 1) = single occasional; or 2 (level 2) = clusters of EpCAM+ hepatocytes.
Analyses were independently performed by two different researchers (G.C. and P.O.) in a blinded fashion.
Statistical Analysis
Data are expressed as means ± SD. The t-test or Mann–Whitney U-test was used to determine the significance of differences between groups for normally or non-normally distributed data, respectively. The Pearson correlation coefficient or the Spearman nonparametric correlation was used. A P value of <0.05 was considered statistically significant. Analyses were performed using SPSS software version 23 (IBM, Milan, Italy).
Results
Clinical and Serologic Parameters
In PSC patients, ALP was similar between early-PSC and cirrhotic-PSC patients (Table 2). In contrast, serum bilirubin, serum albumin, and international normalized ratio were significantly altered in cirrhotic- compared with noncirrhotic-PSC patients. The cirrhotic-PSC patients presented a high Mayo risk score, whereas noncirrhotic-PSC patients showed a low or intermediate score.
Table 2Biochemical and Prognostic Parameters in Patients with Primary Sclerosing Cholangitis
Parameter
Early PSC (n = 10)
Cirrhotic PSC (n = 10)
P value
Age (y)
32.9 ± 5.5
57.13 ± 4.13
<0.001
M/F, n
6/4
6/4
Bilirubin (mg/dL)
0.95 ± 0.58
6.43 ± 5.57
<0.05
Albumin (g/L)
44.5 ± 3.5
25.3 ± 4.1
<0.001
AST (× ULN)
2.14 ± 0.63
3.79 ± 2.05
0.07
ALP (× ULN)
4.23 ± 2.11
1.99 ± 1.61
0.14
GGT (× ULN)
3.55 ± 1.99
2.73 ± 3.08
0.61
INR (0.8–1.2)
0.92 ± 0.04
1.9 ± 0.5
<0.001
Mayo PSC risk
Low to intermediate
High
Unless otherwise indicated, data are expressed as means ± SD. P values <0.05 are in bold.
F, female; M, male; ALP, alkaline phosphatase; AST, aspartate aminotransferase; GGT, γ-glutamyl transferase; INR, international normalized ratio; PSC, primary sclerosing cholangitis; ULN, upper limit of normal.
PBC patients were young adult and middle-aged women (Table 3). ALP was significantly increased in cirrhotic-PBC compared with early-PBC patients (P < 0.05), whereas none of the differences were significant in ALT, bilirubin, or albumin level. PBC patients, according to the Mayo Clinic's Updated Natural History Model for PBC, had an estimated probability of survival at 24 months of 99% to 100%. Liver biochemistry was evaluated 1 year after the start of ursodeoxycholic acid treatment for the calculation of the Global PBC and UK-PBC risk scores. Both the UK-PBC and Global PBC risk scores discriminated between early-PBC and cirrhotic-PBC patients (both, P < 0.05).
Table 3Biochemical and Prognostic Parameters in Patients with Primary Biliary Cholangitis
Parameter
Early PBC (n = 10)
Cirrhotic PBC (n = 10)
P value
Age (y)
46.33 ± 11.5
46.0 ± 8.54
0.96
M/F, n
0/10
0/10
Bilirubin (mg/dL)
0.63 ± 0.29
1.43 ± 1.22
0.14
Albumin (g/L)
40.3 ± 4.3
42.4 ± 4.3
0.5
AST (× ULN)
0.89 ± 0.5
1.72 ± 2.04
0.39
ALP (× ULN)
2.11 ± 1.94
5.62 ± 2.48
<0.05
GGT (× ULN)
3.52 ± 3.54
14.83 ± 5.98
<0.01
INR (0.8–1.2)
0.92 ± 0.04
1.9 ± 0.5
<0.001
UK risk score, 5 years (%)
1.73 ± 1.25
9.69 ± 9.12
<0.05
UK risk score, 15 years (%)
7.05 ± 4.3
29.47 ± 19.32
<0.05
Global PBC score, 3 years (%)
95.93 ± 1.2
90.9 ± 3.24
<0.01
Global PBC score, 5 years (%)
92.85 ± 2.13
84.3 ± 5.4
<0.01
PBC score, 10 years (%)
82.47 ± 7.72
66.87 ± 8.61
<0.01
Mayo PBC risk, 24 months (%)
99–100
99–100
Unless otherwise indicated, data are expressed as means ± SD. P values <0.05 are in bold.
F, female; M, male; ALP, alkaline phosphatase; AST, aspartate aminotransferase; GGT, γ-glutamyl transferase; INR, international normalized ratio; PBC, primary biliary cholangitis; ULN, upper limit of normal.
Ductular Reaction Extent and HpSC Proliferation Indexes
In livers with normal findings on histologic examination (controls), resident HpSCs can be recognized as CK7 (or EpCAM) positive cells located in the smallest branches (canals of Hering and bile ductules) of the biliary tree at the interface with portal tracts (Figure 1A ). In the noncirrhotic stages of PBC (stages I to III; early PBC) (Figure 1B), the presence of reactive ductules was revealed, and DR extent was increased (5.05 ± 0.61) compared with that in healthy livers (2.41 ± 1.67; P < 0.01). Cirrhotic-PBC samples (stage IV) (Figure 1, C and D) were characterized by a greater DR extent (11.90 ± 1.40) compared with those in early PBC (P < 0.01) and controls (P < 0.001). DR extent was strongly correlated with PBC stage (r = 0.83, P < 0.001).
Figure 1Ductular reaction (DR) in human cholangiopathies. A–C: Immunohistochemistry(IHC) analysis for cytokeratin (CK)-7 in controls, primary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC). A: In control livers, CK7 immunoreaction individuates canals of Hering and bile ductules. The boxed area in the left panel corresponds with the magnified view of a canal of Hering in the right panel. B: In the early stages of PBC and PSC, the presence of DR is revealed. C: In cirrhotic samples, DR extent is greater compared with that in early stages. D: Differences in DR extent among groups. E: IHC analysis for CK7 was counterstained with Sirius Red for the visualization of the association between DR and fibrosis extent. F: Scatterplot showing correlation between DR and fibrosis in PBC and PSC. Data are expressed as means ± SD (D). ∗P < 0.05 versus control; †P < 0.05 versus early PSC; ‡P < 0.05 versus early stages (PBC and PSC); §P < 0.05 versus cirrhotic PSC. Scale bars: 100 μm (A–C); 200 μm (E). Original magnification, ×40 (A, right panel)
In PSC (Figure 1, B and D), DR extent was slightly, but not significantly, increased in early-PSC biopsy samples (3.16 ± 1.15) compared with livers with normal findings on histologic examination (P = 0.28). Cirrhotic-PSC samples showed a greater DR extent (7.01 ± 1.96) compared with those in early-PSC stages (P < 0.05) and controls (P < 0.05). DR extent was significantly less in PSC compared with PBC samples (P < 0.01).
In both PBC and PSC samples, DR extent was significantly correlated with fibrosis extent (PBC: r = 0.63, P < 0.01; PSC: r = 0.62, P < 0.01) (Figure 1, E and F).
When the PI of cells within reactive ductules (DR-PI) was calculated (Figure 2), the DR-PI in cirrhotic PBC (38.4 ± 6.2) was higher in comparison with controls (5.8 ± 2.9; P < 0.01) and early-PBC samples (18.0 ± 2.1; P < 0.01). The DR-PI in early PBC was significantly higher compared with that in control specimens (P < 0.05). The DR-PI in cirrhotic PSC (24.4 ± 3.8) was significantly higher in comparison with those in control (P < 0.001) and early-PSC samples (10.0 ± 3.7; P < 0.001), and in early-PSC samples compared with normal specimens (P < 0.05). The DR-PI was lower in PSC compared with PBC samples (P < 0.001).
Figure 2Ductular reaction (DR) proliferation index (PI) in human cholangiopathies. A: Immunohistochemistry (IHC) analysis for proliferating cell nuclear antigen (PCNA) in control livers, primary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC). PCNA+ cells within DR are more numerous (arrows) in PBC and PSC compared with controls. PCNA expression is higher in PBC compared with PSC samples and in cirrhotic compared with early stages. B: Immunofluorescence for cytokeratin (CK)-7 and PCNA in late stages of PBC and PSC samples. Nuclei are displayed in blue. Immunofluorescence confirms the higher PCNA expression in CK7+ DR in PBC versus PSC samples (arrows). C: Differences in DR-PI among groups. Data are expressed as means ± SD (C). ∗P < 0.05 versus control; †P < 0.05 versus early PSC; ‡P < 0.05 versus early stages (PBC and PSC); §P < 0.05 versus cirrhotic PSC. Scale bars: 100 μm (A); 25 μm (B).
With regard to clinical parameters, DR in PSC samples was correlated with the Mayo PSC risk score (r = 0.827, P < 0.05) (Supplemental Figure S1A) and with bilirubin (r = 0.841, P < 0.05) and was inversely correlated with serum albumin (r = −0.838, P < 0.05). No correlation with ALP serum levels was evident in PSC patients. In PBC samples, DR was correlated with total bilirubin (r = 0.937, P = 0.001); γ-glutamyl transferase (r = 0.736, P = 0.037); and UK-PBC risk scores up to 5 years (r = 0.785, P < 0.001), 10 years (r = 0.820, P < 0.001), and 15 years (r = 0.847, P = 0.001) (Supplemental Figure S1B) and Global PBC score (r = 0.791, P = 0.019) and Global PBC score orthotopic liver transplantation–free survival at 3 years (r = −0.852, P = 0.007), 5 years (r = −0.850, P = 0.008), and 10 years (r = −0.833, P = 0.010) (Supplemental Figure S1B). No correlation was found with the History Model for PBC from the Mayo Clinic.
Phenotype of Reactive Ductules
In control specimens with normal findings on histologic examination (Figure 3), double immunofluorescence showed that the majority of CK7+ cells lining the canals of Hering and the bile ductules coexpressed SOX9 (score = 3.6 ± 0.5) (Figure 3A) and CK19 (score = 4 ± 0) (Figure 3B). In cirrhotic PBC, double immunofluorescence showed that reactive ductules were mainly located within fibrous septa and that the majority of CK7+ cells were SOX9+ (score = 3.8 ± 0.4) (Figure 3A) and CK19+ (score = 3.7 ± 0.5) (Figure 3B). There were no or very few cells within the bile and reactive ductules that were Hep-Par1+ (score = 0; ie, <5%) in control and PBC samples (Figure 3C). Conversely, in cirrhotic PSC, reactive ductules were located at the interface with the fibrotic septa, and numerous cells penetrated deep into the cirrhotic nodules; variable percentages of CK7+ DR cells were positive for SOX9 (score = 1.2 ± 0.4) (Figure 3A) and CK19 (score = 1.65 ± 0.5) (Figure 3B), with lower values compared with controls and cirrhotic-PBC samples (both, P < 0.01). Interestingly, reactive ductules contained Hep-Par1+ cells in the cirrhotic-PSC samples (score = 1 ± 0.8) (Figure 3C) but not in the early-PSC samples (data not shown).
Figure 3Phenotype of reactive ductules in human cholangiopathies. A: Immunofluorescence for cytokeratin (CK)-7 and transcription factor sex determining region Y-box (SOX)9 in control livers, primary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC). In cirrhotic PSC, ductular reaction (DR) is composed of a higher percentage of CK7+/SOX9− (blue arrows) compared with controls and cirrhotic PBC. Red arrows indicate CK7+/SOX9+ cells. The graph on the right summarizes differences in SOX9 positivity among groups. B: Immunofluorescence for CK7 and CK19. In cirrhotic PSC, a higher percentage of DR cells are CK19– compared with controls and cirrhotic PBC. Green arrows indicate CK7+/CK19− DR, and yellow arrows indicate CK7+/CK19+ DR. The graph on the right summarizes differences in CK19 positivity among groups. C: IHC analysis for hepatocyte paraffin 1 (Hep-Par1). Reactive ductules contain Hep-Par1+ cells (arrows) in cirrhotic PSC, but not in controls or cirrhotic PBC. Arrowheads indicate negative DR cells. The graph on the right summarizes differences in Hep-Par1 positivity among groups. Data are expressed as means ± SD. ∗P < 0.05 versus other groups. Scale bars = 50 μm.
Hepatocyte buds (Figure 4A ) were not present in controls and were more numerous in cirrhotic PSC compared with cirrhotic PBC (11.4 ± 3.7 versus 5.4 ± 3.4; P < 0.05). The immaturity index, calculated as the presence of EpCAM+/Hep-Par1+ newly derived hepatocytes (Figure 4, B and C), was higher in cirrhotic-PSC samples compared with cirrhotic-PBC samples (1.8 ± 0.5 versus 0.6 ± 0.6; P < 0.01). No EpCAM+ hepatocytes were present in control livers (Figure 4, B and C). Few EpCAM+ hepatocytes were present in early stages of PSC and PBC (data not shown).
Figure 4Features of hepatocyte commitment in hepatic stem/progenitor cell (HpSC) compartment in human cholangiopathies. A: Immunohistochemistry (IHC) analysis for cytokeratin (CK)-19 in control (CTR) livers, primary biliary cholangitis (PBC) and primary sclerosing cholangitis (PSC). The evaluation of hepatocyte buds indicated a higher prevalence in cirrhotic PSC (arrows) compared with controls and cirrhotic PBC. The graph on the right summarizes differences in hepatocyte bud number among groups. B: IHC analysis for epithelial cell adhesion molecule (EpCAM). EpCAM+ hepatocytes are rarely present in normal and cirrhotic-PBC samples, whereas clusters of EpCAM+ hepatocytes are frequently observed in cirrhotic-PSC samples (arrows). The graph on the right summarizes differences in EpCAM+ hepatocytes (Heps) among groups. C: Immunofluorescence for EpCAM and hepatocyte paraffin 1 (Hep-Par1). Nuclei are displayed in blue. Immunofluorescence confirms the presence of clusters of EpCAM+/Hep-Par1+ hepatocytes in PSC (yellow asterisks) but not in PBC (white asterisks). Dashed lines individuate the interface between the portal tract (PT) and the liver parenchyma. Data are expressed as means ± SD (A and B). ∗P < 0.05 versus other groups. Scale bars: 50 μm (A and B); 100 μm (C).
Signaling Pathways in the HpSC Niche: Expression of Laminin, NOTCH, and WNT Pathways
The presence of laminin around the reactive ductules and the expression of the NOTCH and WNT pathways by DR cells were evaluated by immunohistochemistry analysis on serial sections and confirmed by double immunofluorescence (Figures 5 and 6).
Figure 5Laminin and neurogenic locus notch homolog protein 1 (NOTCH1) expression in human cholangiopathies. A: Immunohistochemistry (IHC) analysis for laminin in controls, primary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC). In cirrhotic PBC but not in controls and cirrhotic PSC, ductular reaction (DR) is surrounded by a laminin-rich matrix (arrows). Vessels represent positive controls (arrowheads). B: Immunofluorescence for cytokeratin (CK)-7 and laminin shows laminin (arrowheads) around DR (arrows) in PBC. Nuclei are displayed in blue. C: IHC analysis for NOTCH1. NOTCH1+ DR cells (arrows) are more numerous in cirrhotic PBC compared with controls and cirrhotic PSC. D: Immunofluorescence for NOTCH1 and sex determining region Y-box (SOX)9 confirms high NOTCH1 (green arrows) expression in SOX9+ (red arrows) cells in cirrhotic PBC. Nuclei are displayed in blue. Scale bars: 100 μm (A and C); 50 μm (B and D).
Figure 6Wingless-related integration site family (WNT)/β-catenin expression in human cholangiopathies. A: Immunohistochemistry (IHC) analysis for WNT1 in primary biliary cholangitis (PBC) and primary sclerosing cholangitis (PSC). WNT1+ (arrows) cells in ductular reaction (DR) are more numerous in cirrhotic PSC compared with cirrhotic PBC. B: Immunofluorescence for WNT1 and cytokeratin (CK)-7 confirms WNT1 (red arrows) expression by CK7+ DR (yellow arrows) in cirrhotic PSC. Nuclei (NU) are displayed in blue. C and D: Immunofluorescence for WNT3a and sex determining region Y-box (SOX)9 shows fewer WNT3a+/SOX9+ DR cells (arrows) in cirrhotic PBC compared with cirrhotic PSC. Nuclei are displayed in blue. E: IHC analysis for phosphorylated (p)β-catenin shows higher expression (arrows) by DR in cirrhotic PSC compared with cirrhotic PBC. F: Immunofluorescence for pβ-catenin and CK7 confirms the cytoplasmic/nuclear pβ-catenin expression by CK7+ DR (arrows) in cirrhotic PSC. Nuclei are displayed in blue. G: The heat map shows differences in IHC analysis expression of laminin, NOTCH1, WNT1, WNT3a, and pβ-catenin by DR. Color range refers to the used semiquantitative scoring system, as follows: 0 = <1%; 1 = 1% to 5%; 2 = 6% to 30%; 3 = 31% to 50%; and 4 = >50%. Data are expressed as means ± SD in each cell (G). ∗P < 0.05 versus control; †P < 0.05 versus PSC; ‡P < 0.05 versus PBC. Scale bars: 100 μm (A, C–E); 50 μm (B and F).
In cirrhotic PBC, the percentage of DR elements surrounded by laminin was higher (score = 2.8 ± 0.5) compared with those in the control (score = 0.5 ± 0.1) and cirrhotic-PSC samples (score = 0.6 ± 0.4) (Figure 5, A and B). NOTCH1 expression by DR was increased in cirrhotic-PBC samples (score = 3.6 ± 0.5) and in cirrhotic-PSC samples (score = 2.8 ± 0.4) compared with controls (0.5 ± 0.1) (P < 0.05) (Figure 5, C and D); cirrhotic-PBC samples showed significantly higher values with respect to those in cirrhotic PSC (P < 0.05). The study of the WNT pathway revealed that the expression of WNT1 (Figure 6, A and B), WNT3a (Figure 6, C and D), and phosphorylated β-catenin (Figure 6, E and F) were higher in cirrhotic-PSC samples (2.8 ± 0.9, 2.0 ± 1.0, and 3.3 ± 1.0, respectively; all, P < 0.05) and in cirrhotic-PBC samples [0.6 ± 0.5 (P < 0.01); 0.2 ± 0.5 (P < 0.01); and 1.5 ± 1.0 (P < 0.05)] compared with controls (0, 0, and 0.5 ± 0.1); cirrhotic-PSC samples showed significantly higher values compared with cirrhotic PBC (all, P < 0.05). Results are visually summarized by the heat map in Figure 6G.
Discussion
The main results of the present study indicate that i) DR in PBC is more expansive, appears earlier, and has a higher PI compared with PSC; ii) DR extent correlates with fibrosis and clinical prognostic risk scores both in PSC and PBC patients; iii) the phenotype of reactive ductules differs between PSC and PBC, with a predominant biliary phenotype in PBC and prominent signs of hepatocyte commitment in PSC; and iv) the observed differences in fate commitment were associated with signaling pathways in HpSC niche, as characterized by lower levels of laminin and NOTCH1 and higher WNT/β-catenin pathway expression in PSC compared with PBC.
DR is composed of anastomosing strands of tortuous ductules, and its presence is thought to be the result of HpSC activation and proliferation.
Under normal conditions, they are mostly quiescent cells and do not take part in the renewal of liver parenchyma because mature hepatocytes and cholangiocytes have high proliferative capabilities.
In chronic biliary diseases, a variety of insults induce the apoptosis of cholangiocytes and determine the impairment of cholangiocyte replicative capabilities
However, our results indicate that the presence, extent, and phenotype of DR significantly differs between PBC and PSC. In PBC, a prominent DR appeared at an early stage, and its extent was correlated with the disease stage; contrarily, in PSC, DR did not correlate with the histologic stage and was less prominent in comparison with PBC. Accordingly, the HpSC PI was higher in samples obtained from PBC compared with PSC patients matched for the same histologic stage.
Interestingly, DR extent was correlated with fibrosis extent in both PBC and PSC samples; DR is considered a main driver of fibrogenesis,
and a strict correlation with fibrosis severity has been described in several chronic hepatic diseases, thus individuating patients with a worse prognosis and an increased risk for mortality.
In this context, our results suggest that DR extent is associated with fibrosis extent in both PBC and PSC; thus, the degree of HpSC activation, more than its phenotype, may influence the progression to cirrhosis and to a more aggressive course of the disease.
In PBC, histologic staging according to Ludwig predicts clinical evolution.
Interestingly, our data indicate that DR extent strongly correlates with prognostic clinical scores predicting PBC outcome and ursodeoxycholic acid response (ie, Global PBC and UK-PBC risk scores); thus, the evaluation of DR extent in liver biopsy may represent an additional important histologic prognostic marker.
In PSC, the Ludwig and other histologic staging systems are considered clinically relevant, given their association with transplant-free survival and time to liver transplantation.
However, although considered a progressive cholangiopathy, spontaneous stage regression evaluated by applying the Ludwig scoring system on paired liver biopsies has been observed in a proportion (nearly 15%) of PSC patients.
Although DR extent was correlated with the Mayo risk score, our observations exclude a possible clinical utility in patients with early histologic lesions and in patients with low Mayo risk scores.
Besides DR extent, HpSC phenotype differed between PSC and PBC; a stem/biliary (CK19+/SOX9+/Hep-Par1−) phenotype was prominent in cirrhotic PBC, with few signs of hepatocyte commitment. Conversely, the loss of CK19/SOX9 expression within DR, the Hep-Par1 positivity in reactive ductules, the presence of numerous hepatocyte buds, and EpCAM+ hepatocytes indicated that signs of hepatocyte commitment were frequently observable in cirrhotic PSC. Interestingly, the proliferation of reactive ductules preceded the appearance of Hep-Par1+ DR cells and the accumulation of EpCAM+ hepatocytes. Taken together, our results indicate a different pattern underlying HpSC activation in PBC and PSC. In PBC, the HpSC compartment is activated early, highly proliferative, and shows signs of an undifferentiated phenotype or biliary fate commitment.
As recently demonstrated, the primary damage to these portions of the biliary tree triggers the proliferation of BTSCs located within peribiliary glands.
PSC lesions progress to concentric periductal fibrosis with subsequent duct stenosis (strictures), which may determine the development of secondary intrahepatic cholestasis. In this context, the late activation of the HpSC compartment seems to indicate a secondary response of the hepatic parenchyma to cholestasis more than a primary reaction to the biliary damage; this corresponds with the signs of HpSC commitment toward hepatocytic lineage, tentatively supporting hepatocyte regeneration. Accordingly, with regard to biochemical serum markers of liver damage, DR was correlated with markers of cholestasis both in PSC and PBC, but it was correlated with a reduction in albumin levels only in PSC patients.
The expression of the components of the NOTCH- and WNT-signaling pathways within the HpSC compartment has been studied. The activation of NOTCH and WNT pathways sustains HpSC proliferation.
In particular, NOTCH signaling is implicated in the commitment toward cholangiocyte fate, whereas WNT can trigger differentiation toward the hepatocyte lineage.
Furthermore, a crucial element for the HpSC response is represented by extracellular matrix composition, in which the degradation of the collagen matrix, coupled with production of a laminin-rich niche, leads to HpSC expansion and inhibits hepatocyte differentiation.
Our results indicate that DR cells in PSC were surrounded by a niche with low laminin content and with higher expression of WNT pathway elements compared with PBC samples, corresponding to the presence of hepatocyte commitment signs in PSC.
The different activation of HpSC and BTSC niches in PBC and PSC may also have implications in the risk for primary liver cancers. PSC, but not PBC, is an established risk factor for cholangiocarcinoma.
On the other hand, PBC patients have an increased risk for hepatocellular carcinoma occurrence, and this risk has been associated with advanced histologic stage and nonresponse to ursodeoxycholic acid treatment.
This association seems to be in accordance with the prominent DR observed in cirrhotic-PBC patients, the correlation with prognostic clinical scores, and the assumed role of HpSCs in hepatocellular carcinogenesis.
Moreover, the activation of signaling pathways within the niche is in agreement with the well-established role of the NOTCH and WNT deregulation in liver tumor development.
In conclusion, our results indicate a different pattern of HpSC activation in the context of human primary cholangiopathies, which reflects the involvement of different portions of the biliary tree as the primary target of damage. These aspects may have implications in the pathogenesis of different cholangiopathies. Interestingly, the evaluation of DR in biopsy samples may individuate patients with progressive fibrosis and could have prognostic value, especially in PBC. Future studies may reveal the role of the HpSC-activation patterns in furnishing valuable information in the clinical management of cholestatic diseases.
Acknowledgments
We thank Melissa Kerr for proofreading the manuscript.
G.C. conceived, designed, and performed experiments, analyzed data, and wrote the manuscript; V.C. and D.O. analyzed data, performed experiments, and wrote the manuscript; T.H.K., T.F., A.Fr., N.C., P.O., P.B.B., and A.Fl. acquired data and wrote the manuscript; and D.A. and E.G. conceived and designed experiments, interpreted data, provided financial support, and edited the manuscript. G.C. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Supplemental Data
Supplemental Figure S1Correlation between ductular reaction (DR) and prognostic clinical risk score. A: In patients affected by primary sclerosing cholangitis (PSC), the scatterplot shows correlation between DR and Mayo PSC risk score. B: In patients affected by primary biliary cholangitis (PBC), the scatterplot shows correlation between DR and UK-PBC risk (15 years) and Global PBC score (10 years). ∗P < 0.05.
Biliary tree stem/progenitor cells in glands of extrahepatic and intrahepatic bile ducts: an anatomical in situ study yielding evidence of maturational lineages.
The UK-PBC risk scores: derivation and validation of a scoring system for long-term prediction of end-stage liver disease in primary biliary cholangitis.
Docosahexanoic acid plus vitamin D treatment improves features of NAFLD in children with serum vitamin D deficiency: results from a single centre trial.
Supported by a research project grant from the Sapienza University of Rome , Fondo per gli Investimenti della Ricerca di Base (FIRB) grant RBAP10Z7FS_001 , and Progetti di ricerca di Rilevante Interesse Nazionale (PRIN) grant 2009X84L84_001 (E.G.); FIRB grant RBAP10Z7FS_004 , PRIN grant 2009X84L84_002 , and a grant from Consorzio Interuniversitario Trapianti d'Organo (Rome, Italy; D.A.); a sponsored research agreement from Vesta Therapeutics (Bethesda, MD; E.G.); and a grant from the Norwegian Primary Sclerosing Cholangitis Research Center (Oslo, Norway; T.H.K.).
30740-X&id=gr1.jpg){kind=link}
30740-X&id=gr2.jpg){kind=link}
30740-X&id=gr3.jpg){kind=link}
30740-X&id=gr4.jpg){kind=link}
30740-X&id=gr5.jpg){kind=link}
30740-X&id=gr6.jpg){kind=link}
30740-X&id=figs1.jpg){kind=link}