The first hepatocyte-specific autoantibody reported in autoimmune hepatitis (AIH) was an antibody to liver-specific membrane lipoprotein (LSP).1–3 Studies of LSP revealed that antibody to the asialoglycoprotein receptor was a specific marker of AIH.4, 5 In addition, recent evidence indicates that there may be several autoantibodies that recognize hepatocyte-associated molecules.6–10 We have previously reported an autoantibody to hepatocyte membrane antigen (HMA) that was detected using HepG2 cells and flow cytometry; this autoantibody was frequently found in patients with AIH but not in those with other liver diseases.11 The serum titer of anti-HMA correlated with the serum alanine aminotransferase (ALT) levels in patients with AIH.12 However, the precise role of these various autoantibodies in the immunopathogenesis of AIH is still an enigma. Several previous studies13, 14 have indicated that some hepatocyte-specific autoantibodies are able to kill hepatocytes by either a complement-dependent or a K cell-dependent mechanism; these findings suggest that some autoantibodies may play a role in the immunopathogenesis of AIH. In this study, we generated a human hepatocyte-specific monoclonal antibody (MoAb) to study its properties and its possible role in the immunopathogenesis of AIH.
It has been reported that autoantibodies to hepatocytes are frequently found in patients with autoimmune hepatitis (AIH). To elucidate the nature of these hepatocyte-specific autoantibodies, we attempted to generate a hepatocyte-specific monoclonal antibody (MoAb) from Epstein-Barr virus-transformed peripheral blood mononuclear cells obtained from a patient with AIH. We established a single clone, 2E3, that continued to produce an immunoglobulin M (IgM) antibody (λ-type). This MoAb had the following properties: it reacted mainly with hepatocyte-derived cell lines, rather than with other cell lines, and it reacted with liver tissue but not with other tissues. By immunoblot analysis, we found that this MoAb recognized a 190 kDa molecule on hepatocytes. The MoAb was able to kill hepatocyte-derived cell lines in the presence of fresh human serum. This cytotoxic effect was completely abrogated by heat inactivation of human serum prior to its addition to cell lines. In addition, an IgM autoantibody that recognized a 190 kDa molecule was also found in patients with AIH but not in those with chronic hepatitis C; its titer correlated significantly with serum alanine aminotransferase (ALT) levels in patients with AIH. In conclusion, we generated a human MoAb that recognizes a 190 kDa molecule on hepatocytes. Because of its ability to mediate complement-dependent cytotoxicity and the presence of similar IgM autoantibody in patients with AIH, we hypothesize this autoantibody may play a role in the immunopathogenesis of AIH. (HEPATOLOGY 2004;40:687–692.)
Materials and Methods
We used 9 human cell lines (Table 1). They included 3 hepatocyte-derived cell lines, HuH2, HuH7, and HepG2, that were maintained in Dulbecco's modified eagle's medium (Invitrogen Co., Grand Island, NY) containing 10% fetal calf serum (FCS). Other cell lines were 1 gastric cancer cell line (SS90), 2 colon cancer cell lines (Cw2 and MT), 1 gallbladder cancer cell line (AlaGB), and 2 lymphoma cell lines (Daudi and Ragi). Both Cw2 and AlaGB were established in our laboratory tumor tissue obtained at operation. These 6 lines were maintained in RPMI 1640 (Invitrogen Co.) containing 10% FCS.
|Cell Lines||FI ratio|
|HuH2 (Hepatocellular carcinoma)||10.7|
|HuH7 (Hepatocellular carcinoma)||16.7|
|Cw2 (Colon cancer)||3.6|
|MT (Colon cancer)||0.8|
|SS90 (Gastric cancer)||2.9|
|AlaGB (Gallbladder cancer)||2.0|
Preparation of Human MoAb Using PBMC From a Patient With AIH.
To prepare human MoAb, we used peripheral blood mononuclear cells (PBMC) infected with Epstein-Barr virus (EBV) in vitro.15, 16 In brief, we cultured PBMC from a patient with the classical features of type 1 AIH17; this EBV-producing cell line was designated SB91. Supernatants from these EBV-transformed cells were selected by FACS (see Flow Cytometry Analysis). Using a limiting dilution, we then cloned positive cell lines. The positive clones were selected using the same method as that used in the first selection method. The selected clones were maintained in long-term cultures using FCS-free, GIT medium (Wako Pure Chemical Industries, Ltd., Osaka, Japan). We purified this MoAb in the culture supernatant of the established clone using HiTrap immunoglobulin M (IgM) purification (Amersham Pharmacia Biotech AB, Uppsala, Sweden), according to the manufacturer's protocol. A control IgM myeloma (MM) protein was obtained from the serum of a patient with IgM myeloma using the same purification method.
Flow Cytometry Analysis.
Screening for MoAb was done using a method similar to that applied in our previous studies.11, 12 In brief, either HepG2 or HuH7 cells were obtained by trypsinization, followed by preincubation for 1 hour at room temperature and then washing with phosphate-buffered saline containing 0.2% bovine serum albumin. Cells (1 × 105) were incubated at 37°C for 60 minutes with 100 μL of supernatant obtained from cultures of EBV-transformed cells. After washing, they were incubated further with 5 μL of a 1:10 dilution of fluorescence isothiocyanate (FITC)-conjugated goat anti-human immunoglobulin antibody (Biosource Co., Camarillo, CA) for 30 minutes at 4°C. Positively stained cells were then analyzed by FACScan (BD Pharmingen, San Diego, CA). Other human tumor cell lines were also analyzed using the same method. To identify the isotype of MoAb, we subsequently undertook the same study using both FITC-anti-immunoglobulin G (IgG) and FITC-anti-IgM (Biosource Co.). The results were expressed as relative fluorescence intensity (FI), i.e., the ratio of intensity of each sample to that of the second antibody when added alone.
Immunohistochemical Staining for 2E3.
Human liver tissue was obtained from a liver biopsy obtained from a patient with chronic hepatitis C (CH-C). Other biopsy specimens were obtained from the stomach and colon of patients with chronic gastritis and nonspecific colitis, respectively. Liver, kidney, and thyroid tissue were also obtained from an autopsy of a 37-year-old man). The samples were embedded in OCT compound (Tissue Tek II; Miles Laboratories, Naperville, IL) and frozen in liquid nitrogen. Four μm-thick sections, obtained using a cryostat microtome, were reacted with either the MoAb or myeloma protein for 120 minutes at 37°C; they were then exposed to FITC-labeled anti-human IgM for 1 hour at 37°C. The slides were then viewed under a fluorescence microscope.
The cell membrane fraction was obtained using a previously described method.18 In brief, 50 to 100 μg of membrane protein was adsorbed into sodium dodecyl sulfate-10% polyacrylamide gels and transferred to nitrocellulose filters using an electroblotting transfer system. Filters were incubated for 2 hours in Tris-buffered saline (TBS) containing 5% nonfat dry milk at room temperature, and then overnight with primary antibody to specific proteins in TBS containing 3% nonfat dry milk containing sodium azide. Filters were washed for 5 minutes in TBS-0.2% Tween 20 4 times, incubated with secondary antibodies in TBS for 1 hour, and then washed for 10 minutes in TBS-0.2% Tween 20 5 times. Immune complexes were visualized using an enhanced chemiluminescence system.
Various tumor cell lines were labeled with 51Cr and used as targets for an in vitro test of cytotoxicity. The targets were washed and incubated for 4 hours at 37°C with 50 μL of phosphate-buffered saline containing either purified MoAb (20 μg/mL) or myeloma protein together with 50 μL of fresh human serum or human serum that had been inactivated by heating to 56°C for 30 minutes. For this in vitro assay, the human serum was obtained from 5 carefully selected healthy subjects. The radioactivity in the supernatant was counted using a gamma counter. Specific lysis was calculated using the following equation:
Differences in cytotoxicity between groups were assessed using Student t test. A P value of less than .05 was considered significant.
Serum Autoantibody to the Molecule Isolated Using the MoAb.
We used an affinity column to purify HMA from a crude cell lysate of HuH7 cells. In brief, the MoAb was coupled with CNBr-activated Sepharose 4B (Sigma, St. Louis, MO) according to the manufacturer's protocol. The cell lysate obtained from HuH7 cells was passed though this column, and bound molecules were eluated using a low-pH solution. This purified material was diluted in coating buffer (pH 8.1; total protein concentration approximately 500 ng/mL). Fifty μL of this solution was placed in wells of a 96-well micro plate (Immulon 2, Dynex Technologies, Inc., Chantilly, VA) and incubated overnight at 37°C. Each well was blocked and washed; 50 μL of a 1:200 dilution of serum was then added to each well followed by further incubation for 1 hour at room temperature. After washing, 50 μL of a 1:5,000 dilution of 2nd antibody, goat anti-human IgG or IgM conjugated with horse radish peroxidase (Biosource Co.), was added to each well; they were then incubated again for 1 hour at room temperature. The reaction of each fraction with the corresponding antibody was measured spectophotometrically at 490 nm after incubation of each plate with peroxidase substrate, 2,2′-azino-di-3-ethyl-benzothiazoline sulfonate, for 10 to 15 minutes at room temperature. The serological data were expressed as cutoff values, which were calculated as follows:
The standard serum was obtained from 5 randomly selected healthy female subjects. Serum samples examined in this study were obtained from 16 patients who fully met the criteria for diagnosis of AIH17; all of these patients were females in the age range of 21 to 72 years. As controls, serum samples obtained from age-matched healthy female subjects (n = 22), and patients with CH-C (n = 16) and elevated serum ALT levels (range, 100 to 829 U/L; median 237 U/L) were also examined.
Differences in mean anti-HMA titers in each group were assessed using Student t test. The correlation between anti-HMA titer and serum ALT levels was determined as the Spearman-correlation coefficient. A P value of less than 0.05 was considered significant.
Reactivity of Isolated MoAb to Various Human Cell Lines.
After the cloning procedure, we established a single clone, 2E3. This clone continued to produce λ type-IgM antibody. The specificity of the MoAb in the supernatant of 2E3 was examined by determining its reactivity to 4 cell lines, HuH7, HepG2, SS90, (gastric cancer) and MT (colon cancer) by flow cytometry. It reacted with HuH7 and HepG2, but not with the other 2 cell lines (Fig. 1). We also determined its reactivity with other cell lines. It reacted strongly with 3 hepatocyte-derived cell lines, HuH7, HuH2, and HepG2, with the FI ratio ranging from 7.2 to 16.7; however, it reacted weakly with others, the FI ratio ranging from 0.8 to 3.6. Thus, this MoAb reacted mainly with hepatocyte-derived cells.
Immunohistochemical Staining of Various Human Tissues With the MoAb.
To confirm the specificity of this MoAb, we undertook immunohistochemical staining of several tissue samples obtained by biopsy or at autopsy. Only liver tissue was positively stained by this MoAb (Figs. 2A and B). Specific fluorescence was observed only on the hepatocyte surface membrane. Thus, the MoAb produced by 2E3 specifically reacts with the suface membrane of hepatocytes (Fig. 1, Table 1).
Characterization of the Molecule in Hepatocyte Membranes Recognized by the MoAb.
Characteristics of this MoAb were determined using by immunoblotting. A 190 kDa molecule was precipitated by the MoAb in both HuH7 and HuH2 lysates, but not in control cell (MT) lysates (Fig. 3A). Using MM instead of the MoAb, no such precipitation occurred in these lysates (Fig. 3B).
Complement-Dependent Cytotoxicity Mediated by the MoAb.
To elucidate the biological properties of this MoAb, we examined whether it was able to kill target cells in the presence of fresh human serum. When 51Cr-labeled HuH2 cells were incubated with the MoAb in the presence of fresh human serum, an appreciable release of 51Cr occurred (Table 2). The cytolysis induced by the MoAb was also induced in HuH2 cells but not in SS90 (gastric cancer) or AlaGB (gallbladder cancer) cells. When control IgM was used instead of the MoAb, no significant cytolysis was observed. 2E3-dependent cytotoxic activity mediated by the MoAb was completely abrogated by heat-inactivation of human serum prior to its addition to cultures, indicating that cytotoxicity induced by this MoAb is complement-dependent.
|Target Cell Lines||Treated with|
|HuH2‡ (Hepatocellular carcinoma)||2.9 ± 1.9||33.6 ± 11.5||0.7 ± 0.9|
|SS90 (Gastric cancer)||0.0 ± 0.0||1.0 ± 1.4||0.6 ± 1.0|
|AlaGB (Gallbladder cancer)||0.0 ± 0.0||2.7 ± 0.3||0.8 ± 1.4|
Serum Titers of Autoantibody to HMA190.
We assessed whether an autoantibody like the one derived from 2E3 exists in the serum of patients with AIH. We found that the mean titer of IgM autoantibody was 3.44 ± 1.5 in patients with AIH, 1.33 ± 0.7 in healthy subjects, and 0.66± 0.6 in patients with CH-C; comparisons of data in patients with AIH with those in other cohorts yielded P values for that were less than .05. Furthermore, we did not find antibody to HMA190 of the IgG class in the same samples (Fig. 4b). The association between anti-HMA titers and serum ALT levels was examined in sera from patients with AIH. There was a significant correlation between anti-HMA titers and serum ALT levels (P = .016).
In this study, we generated a human MoAb using EBV-infected PBMC from a patient with type 1 AIH. A single clone, 2E3, produced IgM antibody (λ-type light chain) for more than a year. This MoAb reacted mainly with hepatocyte-derived cell lines (Fig. 1, Table 1) and liver tissue (Fig. 2). Some nonhepatocyte-derived cell lines also reacted with this MoAb, but this reactivity was much lower than that of hepatocyte-derived cells (Table 1). Furthermore, using this MoAb we did not find any specific fluorescence in non-liver tissues (Fig. 2). These findings suggest that even though a molecule recognized by this MoAb may be expressed by nonhepatocytes, this expression is much weaker than that expressed by hepatocytes.
Immunoblot analysis revealed that this MoAb recognizes a 190 kDa molecule. This molecule was not precipitated by MM (Fig. 3), indicating that the 190 kDa molecule was a specific target of the MoAb. Of potential relevance to this finding was the report by Komatsu et al. of an autoantibody that recognized a 210 kDa molecule in patients with AIH.19 However, that molecule was associated with microtubules rather than cell membranes, ruling out the possibility that our 190 kDa molecule was identical to their 210 kDa molecule. In an earlier study of LSP, Swanson et al. reported the presence of several hepatocyte-related polypeptides, between 15 kDa and 220 kDa, in the serum of patients with AIH.20 We assume that the 190 kDa molecule recognized by our MoAb may be one of them.
The precise mechanism of the immunopathogenesis of AIH is still uncertain. Because some altered functions of T cells have been frequently observed in this disease, T cell-mediated cytotoxicity may be a mechanism of hepatocellular injury in AIH.21–23 However, the involvement of an autoantibody against hepatocytes in hepatocellular injury in AIH has also been suggested by several previous studies. For example, acute hepatocellular injury has been generated in mice by immunization with LPS, and complement-dependent hepatocellular toxicity has been induced by an LPS-specific antibody in vitro.2, 13 In addition, several previous studies, including our own, have shown that the serum titer of hepatocyte-reactive autoantibody correlated closely with serum ALT levels.1, 12, 24
Prompted by the findings that some hepatocyte-specific serum autoantibodies mediate cytotoxic effects on human hepatocyte-derived cell lines, we assessed whether our MoAb was toxic to hepatocyte-derived cell lines, and, if this cytotoxicity occurred, whether it was dependent on the presence of human complement. Our results demonstrated that this human MoAb, in the presence of fresh human serum, killed hepatocyte-derived cell lines but not other cell lines. This cytotoxic effect was completely abrogated by heat-inactivation of human serum prior to its addition to cultured cells, suggesting that the cytotoxic effect was complement-dependent.
Recently, it has been reported that several organ-specific autoimmune diseases are mediated by the direct effect of organ-specific autoantibodies.25, 26 Our results suggest that a similar mechanism may also apply to the hepatocellular injury in AIH.
To investigate this possibility, we isolated the HMA recognized by our MoAb using an affinity column, and used the HMA in an enzyme-linked immunosorbent assay system to screen for a serum autoantibody to this molecule in patients with AIH. This enzyme-linked immunosorbent assay reacted with the MoAb in a dose-dependent manner, but it did not react with MM (data not shown). The serum titer of this autoantibody was determined in patients with AIH and control subjects. In patients with AIH, the mean anti-HMA titer was 3.44, which was significantly higher than corresponding means for control groups (Fig. 4). The antibody that recognized the 190 kDa molecule belonged to the IgM—but not the IgG—class of immunoglobulins. We also found a significant correlation between anti-HMA titer and serum ALT levels in serum samples from patients with AIH (Fig. 5). These findings clearly demonstrate that an IgM-class autoanitbody, which is similar to our MoAb, exists in patients with AIH. Our results suggest that measurement of this autoantibody may be useful in the diagnosis of AIH. Because the number of patients in this study was limited, a definitive conclusion concerning the possible clinical use of measurements of this IgM autoantibody will have to be based on studies of a larger population of patients.
The prediction that an autoantibody that recognizes a 190 kDa molecule on hepatocellular surface membranes may play an important role in the immunopathogenesis of AIH is supported by other recent observations (K. Y., unpublished data). Specifically, following the observation that this MoAb cross-reacted with murine hepatocytes, we have successfully induced acute hepatocellular injury, including elevated serum ALT levels and massive hepatocellular necrosis, in mice by administration of this MoAb. These findings indicate that a serum autoantibody that recognizes a 190 kd molecule on the surface membrane of hepatocytes may play a role in the immunopathogenesis of AIH. Characterization of this 190 kDa molecule, including its amino acid sequence, is currently being investigated.
The authors thank Dr. Ezaki, Department of Anatomy and Developmental Biology, Tokyo Women's Medical University, for helpful advice, and Drs. Oda and Noguchi, Department of Pathology, Tokyo Women's Medical University, and Dr. Sawada, Dept. of Path., Tokyo Metropolitan Geriatric Medical Center for providing autopsy samples.