Bisceglie AM, Houghton M, Reyes GR, Kim JP, Choo. Q-L, Alter MJ: Hepatitis C virus antigen in hepatocytes: immunomorphologic detection and identification.
American journal of Pathology, Vol. 150, No. 3, March 1997 Copyright (0 American Society for Investigative Pathology
Lymphocyte and Macrophage Phenotypes in Chronic Hepatitis C Infection Correlation with Disease Activity
Salim 1. Khakoo,* Paresh N. Soni,* Kay Savage,t David Brown,* Amar P. Dhillon,t Leonard W. Poulter,f and Geoffrey M. Dusheiko* From the Academic Department of Medicine* and University Departments of Histopathologyt and Immunology,5 The Royal Free Hospital and School of
Medicine, London, United Kingdom
The pathogenesis of chronic hepatitis C and the mechanisms underlying progressive liver disease in patients with chronic hepatitis C infection are poorly understood. To demonstrate which inflammatory ceUs might be responsible for the necroinflammatory damage in chronic hepatitis C infection, we have correlated the phenotype of the intrahepatic lymphocytes and macrophages with histological activity in liver biopsy and explant specimens from 19 patients with chronic hepatitis C infection. In aU stages of disease, more CD8+ than CD4 + lymphocytes were found. However, histologicaly active versus histologicaly mild hepatitis was associated with a trend toward greater parenchymal concentrations ofCD4+ lymphocytes (0. 71 ± 0.27per
104 pmn2 versus 0.35 + 0.15; not significant), significantly less parenchymal CD8+ lymphocytes (0.90 ± 0.1 versus 1.70 ±0.3,- t = 2.32, P = 0.03) and a greaterparenchymal CD4/CD8 ratio (4.1 ± 2.8 versus 0.91 ± 0.3; t = 1.65, P = 0.07). No difference was found in the number of cells containing cytotoxic granules between the two groups. Greater numbers of CD4+ lymphocytes were found in liver biopsy specimens with little or no staining for hepatitis C virus antigen (1.47 ± 0.88 versus 0.27 + 0.277; t = 2.28, P < 0.05). No significant differences werefound in the macrophage subsets between the three stages of disease. Our data suggest that active histological disease in chronic hepatitis C infection may be
associated with an increase in CD4 + lymphocytes and suggest that CD4+ T ceUs may play an important role in the hepatic injury in these patients. (Am JPathol 1997, 150:963-970)
Hepatitis C virus (HCV) infection is a major cause of chronic liver disease worldwide. Acute infection becomes chronic in a majority of cases,1'2 and chronic infection progresses to cirrhosis in approximately 10 to 20% of patients after 10 years of infection.3'4 There are no pathognomonic histological features, but characteristic findings include mild portal tract inflammation, periportal piecemeal necrosis, parenchymal steatosis, apoptosis, and lobular inflammation, with lymphoid aggregates found in up to 78% of cases.5 The mechanisms of hepatocellular injury and disease progression are poorly understood.6 Despite its relationship to the flaviviruses and pestiviruses, 8 which are known to exert directly cytopathic effects on target organs, evidence for a direct cytopathic effect by HCV is limited. Serum levels of HCV RNA may correlate with lobular inflammation,9 but immunohistochemical studies do not show a direct correlation between hepatic HCV antigen staining and degree of inflammation.10 Immune-mediated mechanisms may therefore be important in the hepatocyte injury of chronic hepatitis C infection. Induction of adhesion molecules and HLA class and 11 antigens on hepatocytes and bile ducts is seen in chronic infection and has been correlated with disease activity, as has the presence of Fas antigen.11'12 Lymphoid follicles composed of B lymphocytes, CD4+ T lymphocytes, and CD8+ T lymphocytes13 are commonly seen in the portal tracts of liver biopsy specimens. Accepted for publication November 4, 1996. Address reprint requests to Dr. G. M. Dusheiko, University Department of Medicine, The Royal Free Hospital and School of Medicine, Pond Street, London NW3 2QG, UK.
963
964
Khakoo et al
AJP March 1997, Vol. 150, No. 3
Table 1. Histological Activity Scores of the Individual Cases
Ishak et al21 classification Category Mild Mild Mild Mild Mild Mild Mild Active Active Active Active Active Active Cirrhosis Cirrhosis Cirrhosis Cirrhosis Cirrhosis Cirrhosis
Interface hepatitis
Confluent necrosis
Focal necrosis
Portal inflammation
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1 2 1 1 2 2 2 2 3 2 2 1
1 1
1 1 3
0 0 0
3
0
31 1 2
0
0
2
2 0 0 0 0
HLA class-l-restricted CD8+ cytotoxic T cells have been isolated from the peripheral blood and livers of patients with chronic HCV infection, and in vitro cytolytic responses to several HLA-A2-restricted peptides derived from the HCV core and nonstructural regions have been reported.14'15 The degree of CD8 response has not been correlated with disease activity or viral eradication. High levels of natural killer cell activity, in comparison with other liver diseases, have also been demonstrated in non-A, non-B hepatitis. 16 CD4+ T lymphocytes, which recognize antigen in the context of HLA class 11 molecules, have also been isolated from liver tissue and peripheral blood of patients with acute and chronic HCV infection.17'18 Their role in the mediation of hepatic injury is at present unclear, but it has been suggested that in vitro proliferative responses of peripheral blood CD4+ lymphocytes to several HCV antigens is greater in patients with a benign course of infection.19 However, it is known that CD4+ T lymphocytes can be localized to the liver, and so peripheral blood responses may not be indicative of current intrahepatic responses.20 The aim of this study was to phenotype intrahepatic lymphocytes and macrophages in biopsy and liver explant specimens in patients with different stages of chronic HCV infection to clarify the role of these cells in hepatic injury.
1
2 2 2
0
0 0
0 0
Fibrosis
Histological activity grading
3 2 2 2 3 3 3 2 2 3 1
3 2 2 3 3 3 6 6 6 6 6 6
1
2 2 2 3 8 4 4 7 7 8 7 7
5 6 3
Materials and Methods Patients All available frozen liver biopsy material (1992 to 1995) from patients with HCV infection without previous interferon therapy was reviewed. This material was taken either at the time of liver transplantation or diagnostic biopsy and had been frozen when the biopsy had yielded sufficient material for both routine histopathology and experimental phenotyping, according to a best practice protocol approved by an ethical committee. Suitable material was available from nineteen patients, including six who were transplanted for endstage hepatitis C infection. All were anti-HCV positive by second-generation enzyme-linked immunosorbent assay. Seventeen of eighteen were also HCV RNA positive by nested polymerase chain reaction. Histological activity was assessed by a single observer on the routine paraffin processed sample according to the classification of Ishak et al.21 Patients were categorized into 1) those with mild hepatitis (seven), determined as an absence of interface erosion on liver biopsy, 2) those with active hepatitis (six) as determined by the presence of interface erosion, and 3) those with advanced cirrhosis (six) necessitating orthotopic liver transplantation (Table 1). The demographic data, including serum viral load as measured by the HCV bDNA assay (Quantiplex, Chiron Corp., Emeryville, CA) are given in Table 2.
Intrahepatic Lymphocytes and HCV Infection 965 AJP March 1997, Vol. 150, No. 3
Table 2. Demographic Data of the Three Patient Groups
Age (mean Mild hepatitis Active hepatitis Advanced cirrhosis
+
SD)
31.6 (+8.6) 45 (+11.7) 50.5 (+7.9)*
Sex (M:F) 3:4 4:2 6:0
Serum alanine transaminase (mean + SD)
Serum viral load (mean + SD); (genome equivalents/ml x 105)
60.2 ± 68.4 58.1 + 34.6 100.5 + 57.3
60.2 ± 68.4 69.0 + 72.7 23.7 + 30.3
M, male; F, female. *P < 0.01.
Patients with cirrhosis were significantly older than those in the other two groups (t = 2.85, P < 0.01). All liver specimens were snap-frozen in isopentane in a liquid nitrogen bath. The 5-,um sections were cut on a cryostsat, fixed for 10 minutes in a 1:1 solution of chloroform and acetone, and stored at -200C before staining. The following murine monoclonal antibodies (MAbs) were used: for pan-T-cell staining, a cocktail of CD2, CD7, and CD8; pan-Bcell staining, CD22 and CD37R; T-cell subsets CD4 and CD8 MAbs; cytotoxic granule-containing cells, anti-TIA-1 (Coulter Corp., Hialeah, FL); natural killer cells CD11b and anti-human bcl-2 (Dako Corp., Glostrup, Denmark). Macrophages were stained using the MAbs RFD1 and RFD7, which distinguish the following functionally distinct macrophage subsets: antigen-presenting cells (RFD1+, RFD7-), mature phagocytes (RFD1 -, RFD7+), and suppressive macrophages (RFD1+, RFD7+).22 Sections were stained using immunoperoxidase, immunofluorescence, and alkaline phosphatase anti-alkaline phosphatase (APAAP) techniques as described below. Positive and negative controls were included for all cases. Negative controls consisted of specimens in which the primary antibody was omitted, and sections of palatine tonsil were used as positive controls and also to determine optimal titers of MAbs and isotype specificity.
Immunoperoxidase Method Briefly, sections were incubated in a moist chamber for 10 minutes in normal rabbit serum diluted 1:100 in phosphate-buffered saline (PBS). The first-layer antibody was applied at the appropriate dilution and incubated for 45 minutes. The sections were rinsed and the rabbit anti-mouse antibody/immunoperoxidase conjugate, code P161 (Dakopatts) diluted 1:100 in PBS containing normal horse serum (1:25), was added. Sections were incubated for 30 minutes, rinsed, and developed with 3,3'-diaminobenzidine. They were then rinsed, counterstained in Harris' hematoxylin, and mounted in DPX.
Dual Immunoperoxidase/APAAP Method The following method was used to evaluate dual staining in which both first-layer MAbs were of the same class. The method for immunoperoxidase staining was followed until just before the counterstaining step. The sections were then rinsed in Trisbuffered saline (TBS), and the first-layer MAb diluted in TBS was applied for 45 minutes. Sections were rinsed and the second-layer, rabbit anti-mouse immunoglobulin code Z 259 (Dakopatts) was applied for 30 minutes. After rinsing again, the mouse APAAP conjugate code D651 (Dakopatts) was applied for 30 minutes. After additional rinsing, color was developed with a solution containing levamisole, naphthol, and Fast red. Sections were then counterstained in Harris' hematoxylin and mounted in PBS/glycerol 1:9.
Double-Immunofluorescence Method A double-immunofluorescence method was employed to evaluate dual staining and co-localization of antigens, in which the first-layer MAbs were of different classes. Sections were incubated with both MAbs simultaneously at appropriate dilutions in a moist chamber for 45 minutes. After rinsing in PBS, the second-layer mixture of goat anti-lgG tetraethyl rhodamine isothiocyanate (TRITC) and goat anti-lgM fluorescein isothiocyanate (FITC) in PBS was added and incubated for 30 minutes. The slides were then rinsed and mounted in PBS/glycerol 1:9.
Evaluation of Staining Sections stained by the immunoperoxidase and immunoperoxidase/APAAP techniques were quantified by means of an image analysis system (Seescan, Cambridge, UK). Areas to be measured were framed and recorded in square microns.23 Using this system of framing, hepatic parenchyma, portal tracts, and lymphoid aggregates could thus be distinguished and counted separately. Cells exhibiting identifiable
966
Khakoo et al
AJPMarch 1997, Vol. 150, No. 3
reactions on their cell membranes were recorded as positive. Five such defined areas were counted in the parenchyma of each section, and up to five portal tracts were counted where available. Results are expressed as cell counts per unit area. Sections stained by the double-immunofluorescence method were also quantified using the image analysis system (Seescan). Again, parenchyma, portal tract areas, and lymphoid aggregates could be distinguished, and quantification of the relative proportions of IgG-TRITC+, IgM-FITC+, and dually labeled (both IgG-TRITC+ and IgM-FITC+) cells in each area was performed.
HCVAntigen Staining Staining for HCV antigens was performed according to the method of Krawczynski et al.10 Five-micron frozen sections were fixed in chloroform, air dried, and then incubated for 40 minutes with a FITC-conjugated anti-HCV polyclonal IgG isolated from a patient with a high titer of anti-HCV antibody by ELISA 11 (Abbott Laboratories, North Chicago, IL). Sections were washed three times in PBS and mounted in citifluor (UKC Chemical Laboratories, North Chicago, IL). Analysis was performed using a direct view confocal laser imaging system (DVC 250; Biorad, Poole, UK). Sections in which >50% hepatocytes stained positive for HCV antigens were given a score of 3+, those in which 5 to 50% stained positive were scored as 2+, and those with fewer than 5% positive hepatocytes were assigned a score of 1+. Intensity of staining was graded on an arbitrary scale from 1+ to 3+.
Statistics Mean and standard errors were calculated from the cell counts taken over the representative areas in each specimen for each stain. The mean and standard error for the cell counts of each stain in the separate disease categories was calculated. Where appropriate, significance was determined using Student's t-test for unpaired data.
Results Quantification of immunostaining in the parenchyma of all specimens was performed. Areas across the entire hepatic lobule of each specimen were included, and the portal tracts excluded, from this analysis. However, in only those with advanced cirrhosis (ie, the explant specimens) was enough tissue
obtained to reliably count an adequate number of areas to accurately quantify staining of the connective tissue, and therefore these data are considered separately. T lymphocytes were the predominant lymphocyte population in the parenchyma, in all stages of disease, and also in the connective tissue in all specimens studied. Mean parenchymal T lymphocyte counts were 3.13 ± 0.52 cells per 104 p.m2 in mild disease, 4.78 ± 1.39 in active disease, and 2.28 ± 0.43 in advanced cirrhosis. Very few parenchymal B lymphocytes were found in all stages of disease (0.024 + 0.018, 0.014 ± 0.009, and 0.083 ± 0.04, respectively). In all stages of disease, more CD8+ than CD4+ lymphocytes were found, but there was a trend toward higher parenchymal CD4 cell counts in active disease versus mild disease, which was not statistically significant (0.71 ± 0.27 per 104 tLm2 versus 0.35 ± 0.15; t = 1.2, P = 0.12). The CD8 count was, however, significantly lower (0.90 + 0.1 versus 1.70 ± 0.3; t = 2.32, P = 0.03), and the CD4/CD8 ratio, as assessed by the dual-immunofluorescence technique, was significantly higher (4.1 ± 2.8 versus. 0.91 ± 0.3; t = 1.65, P = 0.07) in active versus mild disease (Figure 1). In patients with advanced cirrhosis, the CD4/CD8 pattern was more similar to that seen in mild disease. Lymphocytes were predominantly concentrated within the septa and at the interface between the portal tract and hepatic lobule rather than across the hepatic lobule itself. The highest concentration of CD4+ cells was noted at the parenchymal-portal tract interface (Figure 2). A large proportion of these CD4+ cells (up to 43%) stained for the marker bcl-2. Proportions of CD4 cells that were also bcl-2 positive were similar in all stages of disease. Intralobular lymphocytes were also observed in all specimens. These intralobular lymphocytes were of both CD4 and CD8 phenotypes, although the latter predominated. In general a greater proportion of the intralobular lymphocytes were CD8+ compared with CD4+ in mild versus active disease. Care was taken to distinguish the CD4+ lymphocytes from CD4+ Kupffer cells, which were also noted, especially in sections from cases of active disease. Again, these cells were concentrated at the interface between the portal tract and the hepatic lobule. We found no difference in the number of cytotoxic granule-containing (TIA-1 +) cells in active versus mild disease, but the number of these cells was lower in cases with advanced cirrhosis (Figure 1). CD4+ cell counts were lower in sections with >5% hepatocytes positive for HCV antigens (10 specimens) than in those with
