Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Hepatic stellate cells (HSCs) in the normal liver are quiescent, long-lived cells situated in the space of Disse and bearing cytoplasmic processes that embrace the sinusoids, thereby facilitating their interaction with neighboring cell types.1 HSCs are the predominant source of extracellular matrix collagen in the liver, and their activation is the earliest step in hepatic fibrogenesis, collagen deposition, and progressive fibrosis leading to cirrhosis.2 Following liver injury, HSCs undergo activation and transdifferentiation to myofibroblast-like cells.2 The expression of alpha-smooth muscle actin (α-SMA) is a reliable and widely used marker of activation of HSCs in patients with chronic hepatitis C.3–7 Although the correlation between HSC activation and necroinflammatory activity and/or fibrosis stage is controversial,3, 5, 6, 8, 9 an improvement in the histological activity index in responders to interferon-based therapy coincides with decreased activation of HSCs.3, 5, 6 Recent evidence concerning recurrence of hepatitis C virus (HCV) infection following liver transplantation shows that early activation of HSCs predicts fibrosis progression.10, 11 Finding precocious markers of activation of HSCs will be helpful in identifying early stages of hepatic fibrosis when the antiviral therapy may reduce fibrosis progression.12–14
In the last decades, many markers have been tested in HSCs.15–20 Glial fibrillary acidic protein (GFAP) immunoreactivity was detected in rat and human HSCs, which exhibit neural/neuroendocrine features. In the rat liver, GFAP increases in the acute response to injury and decreases in the chronic response.17 Reports concerning GFAP expression in human liver are conflicting.18, 20, 21 The difference between GFAP expression in rat and human liver HSCs may reflect species diversity for the molecular phenotype of human and rat HSCs.
In the central nervous system, GFAP expression is required for blood-brain barrier repair following brain injury.22 The analogy between HSCs and astrocytes suggests that GFAP expression by HSCs could also be involved in vascular remodeling in the hepatic damaged tissue, in which neovascularization significantly increases during the development of liver fibrosis, as described in both human and animal experimental studies.23–26
In this study, we investigated the utility of GFAP in relation to α-SMA as an immunohistochemical marker of early activated HSCs in chronic hepatitis C. For the first time in the model of HCV infection following liver transplantation, we also tested the correlation between GFAP and fibrosis progression, postulating a predicting value of GFAP immunoreactivity. To further characterize the biological significance of hepatic GFAP, we finally investigated the relationship between GFAP expression and vascular remodeling.
α-SMA, alpha-smooth muscle actin; DL, donor liver; GFAP, glial fibrillary acidic protein; HCV, hepatitis C virus; HCV-C, hepatitis C virus cirrhosis; HCV-CH, hepatitis C virus chronic hepatitis; HCV-PTR, posttransplant hepatitis C virus recurrent hepatitis; HSC, hepatic stellate cell; MVD, microvessel density; ns, not significant.
PATIENTS AND METHODS
This study involved 68 hepatic tissue samples divided into donor liver (DL; n = 21), hepatitis C virus cirrhosis (HCV-C; n = 16), hepatitis C virus chronic hepatitis (HCV-CH; n = 12), and posttransplant hepatitis C virus recurrent hepatitis (HCV-PTR; n = 19) groups. Liver specimens were obtained from heart-beating cadaveric liver donors at laparotomy for the DL group, from explanted cirrhotic livers at liver transplantation for the HCV-C group, and from percutaneous needle biopsies with histological aspects of HCV hepatitis and fibrosis for the HCV-CH and HCV-PTR groups, which were taken for diagnostic purposes with the informed consent of the patients. In the HCV-PTR group, the duration of the infection period corresponds to the time that elapsed from the transplantation procedure, whereas in the HCV-CH group for patients who could recall their date of exposure and underwent liver biopsy, we estimated the duration of the infection period from the date of exposure until the first liver biopsy.
The study followed the ethical procedures approved for the Policlinico Umberto I of the University of Rome “La Sapienza.”
Hepatic Histology and Immunohistochemistry
Liver fragments were fixed in buffered formalin for 2-4 hours and embedded in paraffin with a melting point of 55-57°C. Three- to four-micrometer sections were cut and stained with hematoxylin-eosin and Masson's trichrome stains. For immunohistochemical studies, the sections were mounted on glass slides coated with 0.1% poly(L-lysine). After deparaffination and subsequent blockage of the endogenous peroxidase activity by incubation in 2.5% methanolic hydrogen peroxide (30 minutes), the endogenous biotin was blocked by the Biotin Blocking System (Dako, Milan, Italy) according to the instructions received from the vendor. The sections were then washed 3 times in phosphate-buffered saline. For immunohistochemical studies, the sections were processed as described elsewhere.7 Mouse monoclonal anti–α-SMA antibody (Dako 1A4) diluted 1:40, mouse monoclonal anti-GFAP antibody (Neomarker Ab-1) diluted 1:100, and mouse monoclonal anti-CD34 antibody (Dako QBEnd 10) diluted 1:50 were used as primary antibodies. The sections were incubated for 1 hour at room temperature with anti–α-SMA and anti-CD34 and overnight at 4°C with anti-GFAP. After 3 washings in phosphate-buffered saline, the sections were incubated for 30 minutes with the appropriate secondary biotinylated antibody labeled with the avidin-biotin complex (labeled streptavidin-biotin; Dako code K0675). Negative controls were performed with normal mouse antiserum instead of the primary antibody, which uniformly demonstrated no reaction. The sections were developed with 3,3′-diaminobenzidine and finally counterstained with hematoxylin. The hepatic cell population examined in the present study was represented by parenchymal HSCs that were distinguished from the other myofibroblasts of the liver (such as portal, interface, and septal myofibroblasts) by their specific position. Parenchymal HSCs were individuated by morphological criteria (perisinusoidally located, stellate-shaped cells residing in the parenchymal lobules or nodules).19 Estimation of the number of anti–α-SMA and anti-GFAP immunoreactive HSCs was performed independently by 2 researchers. Intraobserver agreement was higher than 90%. The number of positive HSCs and negative HSCs was separately counted for α-SMA and GFAP under a light microscope at 200× magnification: only the cells that displayed nuclei on the section were considered. For each slide, at least 7-10 microscopic fields were randomly chosen. The percentage of positive HSCs was calculated on the total number of HSCs counted in each slide. The average percentage of activated HSCs was then calculated for each group. In the HCV-PTR and HCV-CH groups, histopathological lesions of chronic HCV hepatitis were evaluated according to the histology activity index of Knodell et al.27; however, necroinflammatory and fibrosis scores were given separately to distinguish ongoing hepatitis from parenchymal remodeling with fibrosis. In the HCV-PTR group, the grading of acute allograft rejection was assessed according to the rejection activity index score.28
Microvessel Density (MVD) Evaluation
MVD was evaluated according to the method described by Weidner et al.29 simultaneously by 2 independent observers, without knowledge of the patient's data, using a double-headed light microscope. Both observers agreed that any brown-stained (CD34-positive) endothelial cell or endothelial cell cluster, which was clearly separated from adjacent microvessels and other connective tissue elements, was considered a single, countable blood vessel. Screening of the hepatic tissue was first performed at a low power (40×) to identify areas of the highest MVD. Counting was performed in the 3 highest MVD areas at a high power (100×). There was no significant interobserver difference, and cases with wide differences were re-evaluated by a third observer. The data were expressed as MVD [countable blood vessels per observed area (mm2)].
Quantitative analysis was performed on specimens stained with Masson's trichrome as previously described.30 In brief, light microscopy micropictures were captured by a videocam (SPOT Insight; Diagnostic Instrument, Inc., Sterling Heights, MI) connected to an Olympus BX-51 light microscope (Olympus, Tokyo, Japan) and processed with an image analysis system (Delta Sistemi, Roma, Italy). The green-stained collagen fibers were automatically measured as the volume fraction of the entire liver tissue specimen, including the collagen fibers that normally existed in the portal tract or central vein. Fibrosis progression was calculated as the ratio of the volume fraction of fibrosis (assessed by morphometry) to the months that elapsed between liver biopsy and transplantation.
Results are presented as the mean ± standard deviation. Statistical analysis was conducted with the Mann-Whitney U test. To assess significant correlations, the Pearson or Spearman correlation coefficients were calculated when the data had a normal or not normal distribution, respectively. P values < 0.05 were regarded as statistically significant.
The diagnosis of cirrhosis was confirmed by histological examination of the explanted liver in all cases of the HCV-C group. The biopsies collected in the HCV-PTR and HCV-CH groups were performed at a mean distance from infection of 31.4 ± 44.0 and 102.0 ± 54.5 months, respectively. The biopsies of the HCV-PTR group showed chronic hepatitis in all cases, with grading and staging Knodell scores not significantly different from those in HCV-CH: 6.9 ± 3.6 and 1.2 ± 1 versus 6.8 ± 2.4 and 1.7 ± 1 in HCV-PTR and HCV-CH, respectively (P = not significant). In the HCV-PTR group, the rejection activity index score was 2.2 ± 1.9 (range 0-5).
α-SMA HSC Immunohistochemistry
In the DL group, very few α-SMA–positive HSCs were found only along the sinusoids, mostly in the peripheral zones of the hepatic lobule close to the portal spaces. In the HCV-C groups, α-SMA–positive HSCs were strongly and diffusely immunostained. In cirrhosis, many HSCs were present near the expanding septa and in the perisinusoidal spaces of residual hepatic parenchyma. Stronger α-SMA expression was detected in the HCV-PTR and HCV-CH groups (Figs. 1 and 2). Figure 3 shows box plots of the percentages of α-SMA–positive HSCs that were found in the 4 groups. The percentage of α-SMA–positive HSCs was significantly higher in the HCV-PTR, HCV-CH, and HCV-C groups (41.5% ± 17.2%, 30.4% ± 21.3%, and 30.1% ± 18.9%, respectively) compared to the DL group (4.4% ± 3.1%). No other intergroup differences were found in the percentage of α-SMA–positive HSCs.
GFAP HSC Immunohistochemistry
With respect to the distribution of GFAP-positive HSCs, in the DL group, GFAP expression is mostly limited to a small population of stellate cells located in the lobule near the portal tracts. In the HCV-PTR and HCV-CH group, the GFAP-positive HSCs are more evenly distributed throughout the lobule (Figs. 1 and 2). In the C group, GFAP-positive HSCs are not strongly and diffusely immunostained, and GFAP-positive HSCs are mostly confined to the periphery of the hepatic lobule (Fig. 1). Figure 4 shows box plots of the percentages of GFAP-positive HSCs that were found in the 4 groups. The percentage of GFAP-positive HSCs was significantly higher in the HCV-PTR group (35.0% ± 18.9%) compared to the DL, HCV-CH, and HCV-C groups (3.8% ± 8.3%, 18.4% ± 17.4%, and 8.5% ± 11.3% respectively). A significant difference was also found in the percentage of GFAP-positive HSCs between the HCV-CH and DL groups (18.4% ± 17.4% and 3.8% ± 8.3%, respectively).
Microvessels were mostly restricted to enlarged portal spaces and expanding septa in the HCV-PTR and HCV-CH groups. In the HCV-C group, CD34-positive microvessels were also found within the lobules (Fig. 5). Figure 6 shows box plots of the MVD values that were found in the 4 groups. MVD is significantly higher in the HCV-PTR, HCV-CH, and HCV-C groups (216.3 ± 66.2, 231.8 ± 24.8, and 313.1 ± 72.9, respectively) compared to the DL group (124.6 ± 55.7) and in the HCV-C group compared to the HCV-PTR and HCV-CH groups.
The volume fraction of fibrosis was represented only by the connective tissue of the portal tracts in the DL group. The extent of fibrosis was significantly higher in the HCV-PTR, HCV-CH, and HCV-C groups (6.9% ± 4.1%, 8.3% ± 4.8%, and 28.5% ± 6.9%, respectively) compared to the DL group (4.0% ± 0.6%) and in the HCV-C group compared to the HCV-PTR and HCV-CH groups (Fig. 7).
Relationships of Immunohistochemical, Morphometrical, and Histological Parameters
An inverse correlation was found between the percentage of GFAP-positive HSCs and volume fraction of fibrosis (r = −0.447; P < 0.01) and between the percentage of GFAP-positive HSCs and MVD (r = −0.417; P < 0.01) in liver tissue for all patients (HCV-PTR, HCV-CH, and HCV-C groups), whereas no correlation was individuated for α-SMA–positive HSCs. A direct correlation between MVD and the volume fraction of fibrosis (r = 0.616; P < 0.01) was found for all patients (HCV-PTR, HCV-CH, and HCV-C groups; Table 1). In liver tissue of patients with HCV recurrence, an inverse correlation was found between the time that elapsed from transplantation to liver biopsy and the percentage of GFAP-positive HSCs (r = −0.590, P < 0.01; Fig. 8), and a direct correlation was detected between the rating of fibrosis progression and the percentage of GFAP-positive HSCs (r = 0.577, P < 0.01; Fig. 9). In liver tissue of the same patients, no correlation was observed for the percentage of α-SMA–positive HSCs (r = 0.180; not significant).
Table 1. Relationships Between the Percentage of Activated Hepatic Stellate Cells and Morphometrical and Microvascular Parameters
HSCs have a crucial role in determining the pathogenesis and clinical course of liver fibrosis and cirrhosis. Chronic hepatitis C is one of the most important health problems,31 and end-stage liver disease from hepatitis C is the leading indication for liver transplantation.32 Moreover, recurrent hepatitis C is one of the most serious problems presented by liver recipients because 25% of patients who undergo transplantation for hepatitis C progress to advanced fibrosis within 5 years.33 Because treatment of recurrent hepatitis is characterized by poor tolerability and efficacy, finding early markers of activation of HSCs that could identify patients at higher risk for severe recurrence might be very useful in the selection of candidates for antiviral therapy.
To investigate the expression of different HSC markers in different phases of chronic hepatitis C, we planned to compare the immunohistochemical staining and volume fraction of fibrosis in posttransplant HCV recurrence, chronic HCV hepatitis, and cirrhosis. Posttransplantation HCV recurrence represents a human model of the earlier stage of fibrosis.4, 5 Moreover, in the posttransplantation HCV reinfection model, we exactly know the moment of the infection. This is very useful for correlating the immunohistochemical expression of different HSC markers and the time that has elapsed from reinfection (that is, transplantation procedure) and the rate of fibrosis progression.
In the present study, significant numbers of α-SMA–positive HSCs were detected throughout the liver in patients with chronic HCV infection, whereas in normal DLs, α-SMA–positive HSCs were poorly detected. In contrast, their prevalence was not associated with fibrosis severity.
Because α-SMA is a specific marker for smooth muscle cell differentiation,34 it has been used to identify activated HSCs that show a myofibroblastic phenotype.35, 36 However, the association between the α-SMA–positive HSCs and the extent of fibrosis is controversial. A lack of correlation between the prevalence of α-SMA–positive HSCs and fibrosis severity in chronic liver disease was reported by Levy et al.,9 whereas a positive correlation has been found in other studies.5, 7, 21 The precise subpopulation of α-SMA–positive HSCs/myofibroblasts related to fibrosis is still debated.10, 11 The characterization of a more specific marker of HSCs for activation of HSCs could be helpful in improving the predictive role of HSC detection.
Many markers of human HSCs have been described that may identify phenotypic subsets more closely associated with the presence of fibrosis.9, 19
GFAP is an intermediate filament first identified and characterized in astroglial cells.37 HSCs share several features with astrocytes of the central nervous system, such as juxtaposition to the capillaries, a stellate-shaped appearance,1 and response to tissue injuries.38
The role of GFAP expression in HSCs is currently unknown. Previous studies in rodents showed that HSCs contained an unusually broad spectrum of intermediate filament proteins.17, 39, 40 In the rat, GFAP identifies most HSCs in normal liver and HSCs at the margins of fibrotic septa following CCl4 damage.16, 17 When rodent HSCs were activated, the expression of GFAP decreased.17 The decreased GFAP expression in an advanced stage of fibrosis suggested that GFAP could be a marker for quiescent cells in rodents. However, the accumulation of GFAP/desmin-positive HSCs in an early stage of fibrosis,17 the proliferation of cells,41 and the expression of extracellular matrix genes and proteins,42, 43 which are hallmarks of activated HSCs, suggested that an increased expression level of GFAP by HSCs could be related to their initial activation changes.
Reports concerning GFAP expression in human liver are conflicting. A rim of GFAP-positive cells around portal tracts in normal liver and increased staining in the cirrhotic nodule without staining in the fibrous septa have been described.18 In another study,20 GFAP was not detected in normal liver HSCs but was detected in focal periseptal areas in cirrhotic liver. Few studies have been performed in order to quantify the hepatic expression of GFAP at different stages of human chronic hepatitis,9, 19, 21 and none has investigated GFAP expression in a representative group of HCV recurrent hepatitis.
In the present study, GFAP-expressing HSCs seem to be related to early phases of fibrotic tissue deposition. We found that the percentage of GFAP-positive HSCs, but not that of α-SMA–positive HSCs, is higher in posttransplant HCV recurrence than in chronic HCV hepatitis and cirrhosis. Furthermore, the extent of GFAP expression correlates negatively with the fraction volume of fibrosis for all patients (HCV-PTR, HCV-CH, and HCV-C groups). To strengthen the precocity of GFAP expression, the mean time that elapsed from infection was 31 and 102 months in posttransplant HCV recurrence and in chronic HCV hepatitis, respectively. Because this finding could be caused by the more aggressive fibrotic process characterizing HCV recurrence compared to chronic HCV hepatitis, we analyzed what occurs in the HCV-PTR group in term of fibrosis progression and distance from transplantation. We observed that in the HCV-PTR group, the percentage of GFAP-positive HSCs, but not that of α-SMA–positive HSCs, correlates negatively with the time of HCV infection and positively with fibrosis progression.
In recent years, studies were performed in order to identify the liver allograft recipients early after transplantation who are destined to develop aggressive recurrent disease on the basis on HSC activation as determined by α-SMA immunostaining. Russo et al.10 demonstrated that the activity of HSCs determined from a liver biopsy obtained 4 months after liver transplantation is associated with the development of advanced fibrosis within 2 years of liver transplantation. A similar study conducted in HCV-infected patients with liver transplantation found that early activation of mesenchymal but not parenchymal HSCs was a marker for progressive fibrosis.11 The discrepancies in the HSC subpopulations indicated as fibrosis predictors in these studies could be due to the relative aspecificity of α-SMA.
The immunoreactivity of GFAP was correlated with the fibrosis score in a previous study performed in a limited number of HCV-related hepatitis patients without any information about the time of infection and the rate of fibrosis progression.9
In the present study, GFAP-positive HSCs in posttransplant HCV recurrence positively correlated with fibrosis progression. Moreover, GFAP immunostaining seemed to identify a subset of HSCs detectable in liver fibrosis mainly in the early stages compared to α-SMA, and this strengthens the role of GFAP as an early marker of activation of HSCs in posttransplant HCV recurrence. In fact, the closer to transplantation recurrence was observed, the more diffusely GFAP expression was found.
If we consider GFAP as an early HSC marker, it is not surprising that it is more useful in predicting fibrosis progression in recurrent hepatitis than in simply reflecting the fibrosis stage in a mixed population of chronic hepatitis.
GFAP immunoreactivity mainly involved an HSC/myofibroblast subpopulation residing in the parenchyma and at the septal-parenchymal interface, whereas α-SMA–positive cells were more extensively distributed in human fibrotic livers.19 In our study, α-SMA and GFAP immunoreactivities partially overlapped, some HSCs being α-SMA–positive and GFAP-positive; others stained only for α-SMA or GFAP. Interestingly, α-SMA–positive HSCs were widely distributed in the lobule or nodules in all diseased groups, whereas GFAP-positive HSCs, diffuse in recurrent hepatitis, were mainly confined to periseptal parenchyma in cirrhosis. The GFAP-positive cells could be the precursors of fully activated HSCs identified by α-SMA immunoreactivity, or they could represent a subpopulation of different origin. Further studies are needed to investigate the origin of GFAP HSCs in human liver.
New insights into the role of GFAP in activating HSCs are suggested by the relationship between GFAP expression and vascular remodeling. In the present study, the percentage of GFAP-positive HSCs negatively correlates with MVD as determined by CD34 immunostaining for all the HCV-PTR, HCV-CH, and HCV-C groups. Neovascularization and vascular remodeling significantly increased during the development of liver fibrosis in both human and animal experimental studies.23–26 In experimental chronic liver disease, sinusoidal endothelial cells were chronologically the first cell type to undergo pathological changes.44 It is conceivable that HSCs in close contact with sinusoidal endothelial cells could respond to early endothelial modifications, finely tuning their intermediate filaments, as in the central nervous system, where GFAP expression by astrocytes is required for blood-brain barrier repair after injury.22, 45
In conclusion, GFAP could represent a useful marker of early activation of HSCs in the HCV chronic hepatitis setting. The percentage of GFAP-positive HSCs correlates with fibrosis progression in the model of posttransplant recurrent HCV chronic hepatitis, suggesting that the level of GFAP expression in early activating HSCs could be related to the disease severity. GFAP-dependent activation of HSCs precedes fibrotic tissue deposition, inversely correlating with fibrosis and vascular remodeling. Even if further studies are needed, GFAP expression of HSCs could be one of the first steps in the hepatic angiogenic switch in the course of HCV infection and may be the hallmark of incoming fibrogenesis process.
Thanks are due to Mrs. Liliana Domizi for her skillful technical assistance.