Bcl-2–dependent oxidation of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during apoptosis

The close association between autoantibodies against pyruvate dehydrogenase-E2 (PDC-E2), a ubiquitous mitochondrial protein, and primary biliary cirrhosis (PBC) is unexplained. Many autoantigens are selectively modified during apoptosis, which has focused attention on apoptotic cells as a potential source of “neo-antigens” responsible for activating autoreactive lymphocytes. Since increased apoptosis of bile duct epithelial cells (cholangiocytes) is evident in patients with PBC, we evaluated the effect of apoptosis on PDC-E2. Autoantibody recognition of PDC-E2 by immunofluorescence persisted in apoptotic cholangiocytes and appeared unchanged by immunoblot analysis. PDC-E2 was neither cleaved by caspases nor concentrated into surface blebs in apoptotic cells. In other cell types, autoantibody recognition of PDC-E2, as assessed by immunofluorescence, was abrogated after apoptosis, although expression levels of PDC-E2 appeared unchanged when examined by immunoblot analysis. Both overexpression of Bcl-2 and depletion of glutathione before inducing apoptosis prevented this loss of autoantibody recognition, suggesting that glutathiolation, rather than degradation or loss, of PDC-E2 was responsible for the loss of immunofluorescence signal. We postulate that apoptotic cholangiocytes, unlike other apoptotic cell types, are a potential source of immunogenic PDC-E2 in patients with PBC.

Primary biliary cirrhosis (PBC) is a slowly progressive liver disease characterized by the chronic nonsuppurative destruction of intrahepatic bile duct epithelial cells (cholangiocytes) and high titers of IgG anti-mitochondrial Ab’s (1). Although it is an uncommon disease, PBC is a leading indication for liver transplantation among women. Approximately 70% of patients also have salivary gland involvement (2). The only widely used medical treatment, ursodeoxycholate (UDCA), is only moderately effective in preventing progression to cirrhosis (3–5). Interestingly, Gershwin and others (6–8) have determined that over 90% of patients with PBC produce autoantibodies specific for a conformation-dependent epitope of the E2 subunit of the pyruvate dehydrogenase complex (PDC-E2), a ubiquitous mitochondrial matrix protein associated with the inner mitochondrial membrane. Autoreactive T cells specific for PDC-E2 self-peptides have also been isolated from patients with PBC (9, 10). High-titer anti–PDC-E2 autoantibodies with the same specificity are rarely seen in other autoimmune diseases, nor in unaffected relatives of patients with PBC (11). Understanding why an immune response against this particular autoantigen is so closely associated with PBC may provide insight into the pathogenesis of PBC.

As a group, autoantigens have no common cellular distribution or function that distinguishes them from nonautoantigens. However, a high percentage of autoantigens are specifically cleaved by caspases, apoptosis-specific cysteine proteases (12, 13), and become concentrated in cytoplasmic surface blebs or apoptotic bodies during apoptosis (14). Other autoantigens are phosphorylated or otherwise modified during apoptosis (15). Recent studies suggest that under normal conditions, apoptotic cells engulfed by dendritic cells serve as a source of self-antigens for the induction of peripheral self-tolerance (16, 17). Conceivably, under aberrant conditions, apoptosis may generate unique “neo-antigens” for which peripheral self-tolerance has not been induced. For example, granzyme B, released during cytotoxic T lymphocyte–mediated (CTL-mediated) apoptosis of target cells during inflammatory responses, cleaves many systemic autoimmune disease-associated autoantigens at sites distinct from those of caspases (18). The possibility that variation in the apoptotic signaling pathway between cell types might also lead to the generation of “neo-antigens” in select cell types has not been closely studied.

In PBC, as well as other inflammatory cholangiopathies, increased cholangiocyte apoptosis in the presence of activated CTLs is evident in biopsy specimens (19–21). We addressed whether PDC-E2, similar to autoantigens in several systemic autoimmune diseases, is structurally altered or becomes concentrated at the cell surface during apoptotic cell death. Immunoblot analysis of PDC-E2 using PBC patient autoantibodies indicated that PDC-E2 was not a substrate for caspase- or granzyme B–mediated cleavage and remained localized to mitochondria following apoptosis. However, there was loss of immunofluorescent staining of PDC-E2 in several noncholangiocyte cell lines (HeLa, Caco-2, Jurkat T cells, 3T3 fibroblasts, and human skin–derived fibroblasts) following apoptosis, although not in a cholangiocyte cell line, a salivary gland cell line, nor in freshly isolated intrahepatic biliary epithelial cells. Loss of PDC-E2 staining among the different cell types correlated with the expression level of Bcl-2, which has antioxidant properties and inhibits protein oxidation during cell death (22–24). Overexpression of Bcl-2 by transfection inhibited loss of PDC-E2 staining in apoptotic HeLa cells. Cholangiocytes in vivo express significantly higher levels of Bcl-2 compared with most cell types (25, 26). These results suggest that in patients with PBC, apoptotic cholangiocytes are a source of immunogenic PDC-E2 responsible for the chronic activation of autoreactive lymphocytes.

PDC-E2 is not cleaved during apoptosis. Apoptosis was induced in NRCs using UV-B irradiation. PARP, an autoantigen in SLE, is known to be cleaved by caspases during apoptosis. In reduced lysates of control and apoptotic NRCs, autoantigens were immunoblotted using PBC and SLE patient sera. The levels of intact SLE autoantigens were reduced in the apoptotic versus control lysate (Figure 2, lanes 8 and 7, respectively) concomitant with the appearance of the expected caspase cleavage fragment of each SLE autoantigen, confirming induction of apoptosis. In contrast, no cleavage of PDC-E2 or any other PBC autoantigen was observed in the apoptotic NRC lysate (Figure 2, lanes 2, 4, and 6). Densitometric analysis of the PBC autoantigen bands showed no significant difference in intensity. Similarly, no loss of intact PDC-E2 was seen by using serum withdrawal, staurosporine, or CTL granule contents to induce NRC apoptosis (data not shown). Apoptosis was also induced in HSG, HeLa, and Jurkat T cells by UV-B irradiation as well as other apoptotic stimuli as noted in Methods. In all cases, no loss of intact PDC-E2 was observed in apoptotic lysates by immunoblotting (data not shown).

Figure 2

PDC-E2 is not cleaved during apoptosis. Reduced NRC lysates from control cells (C) and apoptotic (A) cells were immunoblotted with three different PBC patient sera (lanes 1–6) and SLE patient sera monospecific for NuMA (lanes 7 and 8, top panel), PARP (lanes 7 and 8, middle panel), and U1 70K (lanes 7 and 8, bottom panel). Apoptosis was induced by UV-B irradiation. Equal amounts of protein were electrophoresed in each gel lane. Neither loss of intact PDC-E2 (66 kDa) nor any PDC-E2 cleavage fragments were seen in the apoptotic lysates (lanes 2, 4, and 6) compared with control lysates (lanes 1, 3, and 5, respectively). Likewise, no cleavage of any other PBC autoantigen was detected. Blotting of NuMA, PARP, and U1 70K in the apoptotic lysate showed generation of their expected caspase cleavage fragments (lane 8), confirming induction of apoptosis. Molecular size markers (M_r × 10^–3) are shown on the left. Blotting by each serum was examined at least three times with identical results. A representative blot is shown.

Recognition of PDC-E2 by PBC patient autoantibodies after apoptosis varies in a cell type–specific manner. The location of PDC-E2 in control and apoptotic cells was compared by indirect immunofluorescent staining. NRCs were costained with PBC patient sera monospecific for PDC-E2 and a mAb specific for COX -1, also a mitochondrial inner membrane protein. PDC-E2 stained in a punctate, perinuclear pattern (Figure 3a) typical of mitochondrial staining, which was similar to that of COX-1 (Figure 3e). PDC-E2 staining also colocalized with staining by mitotracker red (data not shown). Using the same secondary Ab, serum from a healthy control and sera from two patients with AIH and two patients with PSC did not stain mitochondria (data not shown). No staining was seen with secondary Ab alone (data not shown).

Figure 3

PDC-E2 staining by PBC patient sera in apoptotic cells does not localize to cell membrane blebs or apoptotic bodies. Both control cells and cells treated with UV-B to induce apoptosis were examined by confocal, immunofluorescence microscopy. Cells were stained with DAPI (blue) (a–h) to distinguish apoptotic cells (cells labeled with stars) with characteristic condensed, fragmented chromatin from nonapoptotic cells (unlabeled cells). Control NRC were costained with PBC patient serum monospecific for PDC-E2 (green) (a) and a mAb specific for COX-1 (red) (e). Immunoreactivity against PDC-E2 and COX-1 colocalized in mitochondria. In apoptotic NRCs, immunoreactivity against PDC-E2 (green) (b) remained perinuclear and colocalized with COX-1 (red) (f). Preincubation of PBC patient sera with purified PDC blocked staining of PDC-E2 (green) in both nonapoptotic and apoptotic NRCs (g). Cells were stained with PI (red) to distinguish cell membrane blebs or apoptotic bodies in apoptotic cells (c and d). Staining of PDC-E2 (green) in apoptotic NRCs did not localize to cell membrane blebs or apoptotic bodies (c). In contrast, immunoreactivity against PDC-E2 (green) was not detected in apoptotic HeLa cells (d), although immunoreactivity against COX-1 (red) was detected (h). Each experiment was repeated twice with each of the monospecific PBC patient sera with identical results. Bar, 20 μm. Representative images are shown.

After UV-B irradiation to induce apoptosis, staining of PDC-E2 in apoptotic NRCs (Figure 3b; apoptotic cells are labeled with a star) again colocalized with staining of COX-1 (Figure 3f) and did not colocalize with propidium iodide (PI) staining of surface membrane blebs or apoptotic bodies (Figure 3c). Preincubation of the PBC patient sera with purified porcine PDC blocked staining of PDC-E2 in both apoptotic and nonapoptotic NRCs (Figure 3g; apoptotic cells are labeled with a star). Surprisingly, PDC-E2 in apoptotic HeLa cells was undetectable by immunostaining after UV-B irradiation (Figure 3d; apoptotic cells are labeled with a star). In contrast, UV-B–irradiated, nonapoptotic HeLa cells (cells lacking chromatin condensation and membrane blebbing) stained strongly for PDC-E2 (Figure 3d; cells not marked with a star). In contrast, COX-1 staining in apoptotic HeLa (Figure 3h) was readily detectable and similar to that in apoptotic NRCs (Figure 3f).

To better understand this difference in PDC-E2 staining in apoptotic HeLa cells compared with apoptotic NRCs, we examined both additional cell types and other apoptotic stimuli. After UV-B treatment, there was strong staining of PDC-E2 in apoptotic HSGs (Figure 4j) as in NRCs, while there was loss of PDC-E2 staining in apoptotic Jurkat T cells (Figure 4l) as in HeLa cells. HSGs were studied since salivary gland damage in PBC is quite common. Loss of PDC-E2 staining in apoptotic cells was also observed after staurosporine treatment of HeLa cells (Figure 4n; apoptotic cells are labeled with stars), Caco-2 cells (data not shown), and Jurkat T cells (data not shown). In contrast, staurosporine treatment of neither NRCs (Figure 4p) nor HSGs (data not shown) led to a loss of PDC-E2 staining in apoptotic NRCs or HSGs. Likewise, CTL granule treatment caused a loss of PDC-E2 staining in apoptotic HeLa cells, but not in apoptotic NRCs (data not shown). For each stimulus, persistence of PDC-E2 staining in apoptotic cells was dependent on the cell type undergoing apoptosis.

Figure 4

Persistence of PDC-E2 staining by PBC patient sera in apoptotic cells is cell-type dependent and independent of the stimulus used to induce apoptosis. Both control cells and cells treated with UV-B or staurosporine (STS) to induce apoptosis were examined by confocal, immunofluorescence microscopy. Cells were stained with DAPI (a–h) to distinguish apoptotic cells (cells labeled with stars) from nonapoptotic cells (unlabeled cells). After UV-B treatment, PBC patient sera staining of PDC-E2 was detected in apoptotic HSGs (j), but not in apoptotic Jurkat T cells (l). When inducing apoptosis with STS, PDC-E2 immunoreactivity again was undetectable in apoptotic HeLa (n), while persisting in apoptotic NRCs (p), and was undetectable in apoptotic HeLa (n). Bars, 20 μm. All cell types were stained independently with two different PDC-E2 monospecific sera at least three times with similar results each time. Representative images are shown.

Loss of autoantibody recognition of PDC-E2 after apoptosis is due to glutathiolation of its sulfhydryl group(s). Possible explanations for the loss of PDC-E2 staining in certain cell types after apoptosis include the following: (a) leakage of PDC-E2 out of the mitochondria, (b) degradation of PDC-E2, and (c) loss only of the PDC-E2 epitope(s) recognized by PBC patient autoantibodies. To detect leakage of PDC-E2 into the cytosol after UV-B–induced apoptosis, subcellular fractions were prepared from control and apoptotic HeLa cells. The high degree of PARP cleavage detected in the nuclear fraction of UV-B–treated cells (Figure 5, lane 2) confirmed that apoptosis was induced in a high percentage of the cells. Little PDC-E2 was detected in the lysate of the apoptotic cytosol fraction (Figure 5, lane 4). Additionally, there was equivalent immunoblotting of PDC-E2 in the apoptotic mitochondrial lysate as compared with the control mitochondrial lysate (Figure 5, lanes 6 and 5, respectively). Since autoantibody recognition of PDC-E2 by immunoblotting (Figure 2) was equivalent in lysates of control and apoptotic cells, degradation of PDC-E2 is an unlikely explanation. Loss of only the PDC-E2 epitope recognized by PBC patient autoantibodies cannot be similarly ruled out by the immunoblotting results since loss of the epitope may be transient and dependent on the local mitochondrial environment.

Figure 5

PDC-E2 does not leak out of mitochondria during apoptosis. Reduced lysates of subcellular fractions of control (C) and UV-B–irradiated, apoptotic (A) HeLa cells were immunoblotted with SLE and PBC patient sera monospecific for PARP (lanes 1 and 2) and PDC-E2 (lanes 3–6), respectively. Intact PARP predominated in the control nuclear (Nucl) fraction (lane 1), while its caspase cleavage fragment was the predominant form detected in the apoptotic nuclear fraction (lane 2), confirming induction of a high degree of caspase activity in the UV-B–treated cells. PDC-E2 was strongly detected in both control and apoptotic mitochondrial (Mito) fractions (lanes 5 and 6), but not in either cytosolic (Cyto) fraction (lanes 3 and 4). Molecular size markers (M_r × 10^–3) are shown on the left. The experiment was repeated twice with identical results. A representative blot is shown.

Previous studies of PDC purified from tissue specimens have shown that the PDC-E2 epitope recognized by PBC patient autoantibodies is sensitive to the sulfhydryl redox potential of the preparation (36). The intensity of PDC-E2 immunoblotting by PBC patient antisera correlated with the concentration of the sulfhydryl reducing agent in the sample buffer added to purified PDC before SDS-PAGE, suggesting that PBC patient autoantibodies preferentially recognize PDC-E2 with reduced sulfhydryl groups. Coincidentally, the cellular sulfhydryl redox potential during apoptosis has been shown to decrease during apoptosis in some cell types (37) due to increased reactive oxygen species (ROS) production, particularly within the mitochondria.

We have shown previously that treatment of cholangiocytes with BSO, a glutathione synthetase inhibitor, depletes cholangiocytes of glutathione (38), increasing their sensitivity to the effects of ROS. To determine if the cell type–specific difference in PDC-E2 staining after apoptosis was due to differences in the cellular sulfhydryl redox potential and subsequent ROS production during apoptosis, NRCs were pretreated with BSO for 24 hours before inducing apoptosis. HSG and HeLa cells were similarly treated. Pretreatment of each cell type with BSO reduced total glutathione levels greater than 90% compared with untreated cells (Table 1) and did not by itself induce apoptosis. There was no difference in PDC-E2 staining of apoptotic NRCs and HSG cells with or without BSO treatment before UV-B irradiation (data not shown). Unexpectedly, staining of PDC-E2 persisted in apoptotic HeLa cells after UV-B irradiation of cells pretreated with BSO (Figure 6d). As before, in the absence of pretreatment with BSO, PDC-E2 staining was undetectable in apoptotic HeLa cells (Figure 6c; apoptotic cells are labeled with a star). Preincubation of the PBC patient sera with purified PDC inhibited staining of PDC-E2 in apoptotic HeLa cells pretreated with BSO (data not shown). Depletion of total cellular glutathione with BSO prevented the loss of PDC-E2 staining after HeLa cell apoptosis.

Figure 6

Depletion of cellular glutathione by BSO treatment before UV-B irradiation prevents the loss of the PDC-E2 staining in apoptotic HeLa cells. UV-B–irradiated cells were stained with DAPI (a and b) to differentiate nonapoptotic from apoptotic cells and with monospecific PBC patient serum against PDC-E2 (c and d). Apoptotic cells are labeled with stars. Staining was analyzed by confocal, immunofluorescence microscopy. In the absence of pretreatment with BSO, staining of PDC-E2 was again not detected in apoptotic HeLa cells after UV-B irradiation (c). However, staining of PDC-E2 persisted in apoptotic HeLa cells pretreated with BSO before UV-B irradiation (d). Bars, 20 μm. Each experiment was repeated at least three times with identical results. Representative images are shown.

Table 1

Cellular glutathione levels in the presence and absence of BSO

This result suggested a role for glutathione (GSH) in the oxidation of PDC-E2 sulfhydryl groups leading to the loss of PDC-E2 staining in apoptotic HeLa cells. Increased concentrations of protein-S-S-G conjugates in apoptotic cells have been reported (24). To determine if oxidized glutathione (GSSG) directly oxidizes PDC-E2 sulfhydryl groups to cause loss of the PDC-E2 epitope recognized by PBC patient autoantibodies, detergent lysates of apoptotic NRCs were treated with DTT (5mM) and SDS (2%) and then boiled to inactivate lysate proteins before the addition of GSSG (10mM) and subjection to PAGE. After transfer of proteins to nitrocellulose, immunoblotting of PDC-E2 by PBC patient autoantibodies was absent in lysate to which GSSG was added (Figure 7a, lane 2). In the absence of GSSG (lane 1) or in the presence of excess DTT (25 mM) relative to GSSG (10 mM) (lane 3), PDC-E2 was strongly detected. Glutathiolation of PDC-E2 in apoptotic lysate reversibly inhibited its recognition by PBC patient autoantibodies.

Figure 7

Direct oxidation of PDC-E2 by oxidized glutathione abrogates recognition of PDC-E2 by PBC patient sera. (a) Apoptotic NRC lysate was treated with SDS, boiled, and then sequentially treated with 5 mM DTT, 10 mM GSSG, and 25 mM DTT. Strong blotting of PDC-E2 was detected in lysate treated with 5 mM DTT (lane 1). Blotting of PDC-E2 was absent after addition of GSSG to the DTT-treated lysate (lane 2), but subsequent treatment with additional DTT restored detection of PDC-E2 (lane 3). (b) The difference between recognition of oxidized versus reduced PDC-E2 by PBC patient autoantibodies was quantified by immunoblotting lysate treated with 10 mM GSSG vs. 5 mM DTT. A representative blot is shown. (c) The oxidative state of PDC-E2 does not significantly affect its transfer to nitrocellulose. PDC-E2 was immunoprecipitated using PBC patient sera from lysate of HeLa cells incubated with [³⁵S]-methionine to label proteins. Half the immunoprecipitate was treated with DTT and the other half with GSSG before SDS-PAGE was performed. The autoradiogram performed after transfer to nitrocellulose showed that similar amounts of both reduced and oxidized PDC-E2 were transferred. Results using two different PBC patient sera are shown (lanes 1 and 2). (d) The expression level of Bcl-2 correlates with persistence of PDC-E2 staining following apoptosis. Eighty micrograms of lysate protein was loaded per lane and immunoblotted with an mAb specific for human Bcl-2. High levels of Bcl-2 were only detected in the Bcl-2–transfected HeLa (lane 1) and HSG cell lysates (lane 2).

In titration experiments using serum from five different patients with PBC, the difference in autoantibody recognition of PDC-E2 in GSSG-treated versus DTT-treated lysates of apoptotic NRCs ranged from 100- to 1,000-fold (Figure 7b and data not shown) based on the calculated autoantibody dilution at which the OD of the PDC-E2 band would equal 1.5 (the midpoint of the OD scale). That poor autoantibody recognition of PDC-E2 in GSSG-treated lysates was not due to inefficient transfer of glutathiolated PDC-E2 to nitrocellulose was demonstrated as follows. PDC-E2 was oxidized (GSSG treated) or reduced (DTT treated) after its immunoprecipitation from lysates of [³⁵S]-methionine–labeled HeLa cells with antisera from patients with PBC. After SDS-PAGE, both oxidized and reduced [³⁵S]-methionine labeled PDC-E2 transferred to nitrocellulose in equal amounts as assessed by autoradiography (Figure 7c).

High-level Bcl-2 expression preserves the PDC-E2 epitope recognized by PBC patient autoantibodies in apoptotic cells. The formation of protein–S-S-G conjugates during apoptosis is regulated by the expression level of Bcl-2, an antiapoptotic protein with antioxidant properties (24). Human cholangiocytes (25, 26) and NRCs (39) are unusual in that they constitutively express high levels of Bcl-2. High levels of Bcl-2 were detected in HSG cells as well (Figure 7d, lane 2), but not in HeLa or Jurkat T cells (Figure 7d, lanes 3 and 4, respectively). The expression level of Bcl-2 in each cell type correlated with persistence of PDC-E2 staining after apoptosis. To directly address whether the level of Bcl-2 expression affects glutathiolation of PDC-E2 in apoptotic cells, HeLa cells expressing high levels of Bcl-2 (Figure 7d, lane 1) after transfection with a Bcl-2 cDNA were studied.

UV-B–irradiated, Bcl-2–transfected HeLa cells were stained with PBC patient sera monospecific for PDC-E2. To achieve an equivalent extent of apoptosis, the incubation time following UV-B irradiation was doubled in the Bcl-2–transfected HeLa cells compared with wild-type (WT) HeLa cells. Intense punctate, perinuclear staining was detected in apoptotic, Bcl-2–transfected HeLa cells (Figure 8h; apoptotic cells are labeled with a star), unlike apoptotic, WT HeLa cells (Figure 3d). The PDC-E2 epitope(s) recognized by PBC patient autoantibodies persisted after apoptosis of Bcl-2–transfected HeLa cells, but not WT HeLa cells. Consistent with this finding, glutathione levels are 50% higher in nonirradiated Bcl-2–transfected HeLa cells than WT HeLa cells and only decrease dramatically in the WT HeLa cells, not in the Bcl-2–transfected HeLa cells, after UV-B irradiation (Table 1). To strengthen the correlation between cellular Bcl-2 levels and preservation of the PDC-E2 epitope recognized by PBC patient autoantibodies after apoptosis, primary human skin fibroblasts, known to express low levels of Bcl-2 (40), were also examined. In apoptotic primary fibroblasts, PBC patient autoantibody staining of PDC-E2 was absent after UV-B irradiation (Figure 8j), as observed in cell lines expressing relatively low levels of Bcl-2. In both cell lines and primary cells, persistence of the PDC-E2 epitope recognized by PBC patient autoantibodies in apoptotic cells correlated with the expression level of Bcl-2 in that cell type. Additionally, staining of PDC-E2 persisted in freshly isolated, rat IBDECs after UV-B induced apoptosis (Figure 8l) as observed with NRCs, suggesting that the studies of NRC apoptosis pertain to apoptosis of native cholangiocytes lining the intrahepatic biliary tree.

Figure 8

Oxidation of PDC-E2 during apoptosis is Bcl-2 dependent. Control and UV-B–irradiated cells were stained with DAPI (a–f) to differentiate nonapoptotic from apoptotic cells and with monospecific PBC patient serum against PDC-E2 (g–l). Apoptotic cells are labeled with stars. Staining was analyzed by confocal, immunofluorescence microscopy. Staining of PDC-E2 was present in apoptotic, Bcl-2–transfected HeLa cells (h) and apoptotic, freshly isolated IBDECs (j), though not in apoptotic, primary fibroblasts (l). Bars, 20 μm. Representative images are shown.

The selective modification of autoantigens during apoptosis has led to the suggestion that apoptotic cells may be a source of immunogen in patients with autoimmune disease. A high percentage of autoantigens are cleaved by caspases or otherwise altered during apoptosis and become concentrated in surface membrane blebs or apoptotic bodies. These studies show that PDC-E2 is not cleaved by caspases during apoptosis, nor does it become concentrated at the cell surface. However, in most cell types, PDC-E2 did undergo a structural alteration during apoptosis, as detected by the loss of the PDC-E2 epitope(s) recognized by PBC patient autoantibodies. A number of different apoptosis-inducing stimuli had the same effect on recognition of PDC-E2. Loss of recognition of this antigenic epitope(s) was apparently due to the covalent modification of a PDC-E2 sulfhydryl group by glutathione to form a mixed disulfide (PDC-E2-S-SG). These results indicate that the reduced form of PDC-E2, not glutathiolated PDC-E2, is responsible for the activation of autoreactive B cells specific for PDC-E2 in patients with PBC. Significantly, cholangiocytes and salivary gland epithelial cells, the cell types most frequently affected in patients with PBC, retain the PDC-E2 epitope(s) recognized by PBC patient autoantibodies after apoptosis. Thus, a lack of PDC-E2 modification after apoptosis appears to be unusual and preserves its antigenic epitope.

In each cell type, autoantibody recognition of PDC-E2 after apoptosis correlated with the expression level of Bcl-2. Bcl-2 expression has been shown to inhibit oxidative protein damage during cell death (22, 23). Under the conditions used to induce apoptosis, expression of Bcl-2 retarded, but did not prevent, apoptosis. Additionally, overexpression of Bcl-2 by transfection in HeLa cells preserved recognition of PDC-E2 after apoptosis, in contrast to WT HeLa cells. Thus, Bcl-2 appears to inhibit protein glutathiolation during apoptosis. The apparent mechanism by which Bcl-2 inhibits oxidative damage during apoptosis varies among cell types and remains unclear in some cases. In Bcl-2–transfected HeLa, Bcl-2 may act by increasing basal glutathione levels (41). Intrahepatic cholangiocytes express high levels of Bcl-2, unlike most other cell types in vivo (25). Thus, B cell exposure to reduced PDC-E2 from damaged and dying cells is likely minimal. Therefore, we postulate that unusually high concentrations of the reduced form of PDC-E2 following intrahepatic cholangiocyte damage and death are in part responsible for the activation of autoreactive B cells in patients with PBC. Additional factors clearly regulate induction of high titer IgG anti-PDC-E2 Ab production since other conditions involving increased cholangiocyte cell death are not associated with the production of autoantibodies specific for PDC-E2.

The bile duct damage in PBC is considered more likely to be mediated by activated T cells than by autoantibodies (42, 43). T cell activation requires the MHC-restricted presentation of peptides by antigen-presenting cells (APCs), such as macrophages and dendritic cells. Previous studies of two non-PBC autoantigens, insulin (44) and beta2-glycoprotein I (45) have shown that APC processing of only the sulfhydryl-reduced forms of these autoantigens leads to activation of autoreactive T cells. While PDC-E2 is endogenously expressed by APCs, APCs have been shown to present different peptides of the same antigen, depending on whether the antigen is expressed endogenously or engulfed from an exogenous source, such as apoptotic cells (46). Determining whether or not the oxidative state of PDC-E2 sulfhydryl groups also affects activation of autoreactive T cells in patients with PBC may further aid in understanding the pathogenesis of this disease.

The suggestion that apoptotic cholangiocytes are a source of immunogen in PBC implies that the activation of autoreactive lymphocytes in PBC is an epiphenomenon following bile duct damage. Yet, autoantibodies are typically detected in PBC patients with early-stage disease when there is little or no evidence of bile duct damage. Interestingly, the aberrant expression of MHC class II molecules on the basolateral surface of intrahepatic cholangiocytes in patients with PBC indicates that the cholangiocytes, themselves, act as APCs (47). Presentation by healthy cholangiocytes of immunogenic PDC-E2 peptides derived from reduced PDC-E2 in engulfed apoptotic cholangiocytes could lead to further cholangiocyte destruction by autoreactive T cells. Thus, a self-sustaining cycle of cholangiocyte destruction might develop after an initial, unidentified triggering event in susceptible individuals. Whether or not healthy cholangiocytes engulf apoptotic cells, as shown for hepatocytes (48), has not been reported. Previous immunohistochemical studies have detected aberrant staining by PDC-E2–specific Ab’s of the apical cell surface and subapical region of PBC patient cholangiocytes (49). Engulfed apoptotic debris is a potential source of this aberrant staining.

Previous studies have directly demonstrated the glutathiolation of a limited subset of soluble cellular proteins in response to increased ROS production after TNF-α stimulation (50). Similar techniques may be useful to confirm directly that PDC-E2 as well as other PBC autoantigens are glutathiolated during apoptosis in most cell types. Sulfhydryl groups are often critical residues within the functional domains of proteins. Thus, protein glutathiolation may play a role in cell signaling by regulating the function of proteins with oxidant-sensitive sulfhydryl groups. With regard to apoptosis signaling, Costantini et al. (51) recently demonstrated that oxidation of a critical sulfhydryl group of the mitochondrial adenine nucleotide transporter (ANT) is sufficient to facilitate mitochondrial membrane permeability transition, which enhances the release from the mitochondria of cytochrome c (52, 53) and apoptosis-inducing factor (54). The release of these factors results in the activation of effector caspases, changes in nuclear morphology, and increased ROS production. Coincidentally, the liver-specific isoform of ANT has been identified previously as a PBC patient autoantigen (55).

Our studies do not indicate which PDC-E2 sulfhydryl group(s) is potentially oxidized during apoptosis. mAb’s directed at different epitopes of PDC-E2 may be useful in this regard, as well as to confirm our cell fractionation results (see Figure 5) showing that PDC-E2 levels remain constant in the mitochondria of apoptotic cells even when the epitope recognized by PBC patient autoantibodies is lost. The major PDC-E2 epitope recognized by PBC patient autoantibodies maps to the inner lipoyl domain of PDC-E2, which contains sulfhydryl groups involved in enzyme function (56, 57). Critical sulfhydryl groups are also present within the antigenic sites of several other PBC autoantigens (58, 59). Additionally, PBC patient autoantibodies inhibit PDC activity in vitro (60). Therefore, glutathiolation of PDC-E2 would likely inhibit its enzymatic activity leading to decreased production of acetyl CoA and citric acid cycle activity. Since loss of PDC-E2 staining was not observed in apoptotic HeLa cells without concomitant changes in nuclear morphology, glutathiolation of PDC-E2 likely followed cytochrome c release and did not contribute to the induction of apoptosis. Further studies are needed to determine if rapid glutathiolation of PDC-E2 after a toxic stimulus might favor necrotic rather than apoptotic cell death.

PDC-E2 is the first major autoantigen for which autoantibody recognition of its antigenic epitope has been shown to be lost after apoptosis due to oxidation of a sulfhydryl group(s). Autoantibody recognition of many other autoantigens, such as GAD65 (61), a major autoantigen in patients with type I diabetes mellitus, and fibrillarin (62), associated with scleroderma, is also dependent on the oxidative state of the autoantigen. In type I diabetes mellitus, pancreatic β islet cell apoptosis is involved in the pathogenesis of the disease (63–65). Whether or not GAD65 is oxidized during apoptosis of pancreatic β islet cells or any other cell type has not been determined. Our findings with PDC-E2 suggest that autoantigen oxidation during apoptosis may be cell-type specific.

These studies were supported by NIH grants T32DK-07832 (J.A. Odin), AR-44684 (L. Casciola-Rosen), DE-12354 (A. Rosen), and DK-24031 (N.F. LaRusso). J.A. Odin is also supported by the Artzt Family Foundation. A. Rosen is also supported by the SLE Foundation and a Translational Research award from the Burroughs Welcome Fund.

See the related Commentary beginning on page 187.

1. Kaplan, MM. Primary biliary cirrhosis. N Engl J Med 1996. 335:1570-1580.
   View this article via: PubMed CrossRef Google Scholar
2. Culp, KS, Fleming, CR, Duffy, J, Baldus, WP, Dickson, ER. Autoimmune associations in primary biliary cirrhosis. Mayo Clin Proc 1982. 57:365-370.
   View this article via: PubMed Google Scholar
3. Poupon, RE, et al. Combined analysis of randomized controlled trials of ursodeoxycholic acid in primary biliary cirrhosis. Gastroenterology 1997. 113:884-890.
   View this article via: PubMed CrossRef Google Scholar
4. Angulo, P, et al. Long-term ursodeoxycholic acid delays histological progression in primary biliary cirrhosis. Hepatology 1999. 29:644-647.
   View this article via: PubMed CrossRef Google Scholar
5. Pares, A, et al. Long-term effects of ursodeoxycholic acid in primary biliary cirrhosis: results of a double-blind controlled multicentric trial. UDCA- Cooperative Group from the Spanish Association for the Study of the Liver. J Hepatol 2000. 32:561-566.
   View this article via: PubMed CrossRef Google Scholar
6. Gershwin, ME, Mackay, IR, Sturgess, A, Coppel, RL. Identification and specificity of a cDNA encoding the 70 kd mitochondrial antigen recognized in primary biliary cirrhosis. J Immunol 1987. 138:3525-3531.
   View this article via: PubMed Google Scholar
7. Yeaman, SJ, et al. Primary biliary cirrhosis: identification of two major M2 mitochondrial autoantigens. Lancet 1988. 1:1067-1070.
   View this article via: PubMed Google Scholar
8. Surh, CD, Coppel, R, Gershwin, ME. Structural requirement for autoreactivity on human pyruvate dehydrogenase-E2, the major autoantigen of primary biliary cirrhosis. Implication for a conformational autoepitope. J Immunol 1990. 144:3367-3374.
   View this article via: PubMed Google Scholar
9. Van de Water, J, et al. Evidence for the targeting by 2-oxo-dehydrogenase enzymes in the T cell response of primary biliary cirrhosis. J Immunol 1991. 146:89-94.
   View this article via: PubMed Google Scholar
10. Jones, DE, Palmer, JM, Yeaman, SJ, Bassendine, MF, Diamond, AG. T cell responses to natural human proteins in primary biliary cirrhosis. Clin Exp Immunol 1997. 107:562-568.
    View this article via: PubMed CrossRef Google Scholar
11. Caldwell, SH, et al. Antimitochondrial antibodies in kindreds of patients with primary biliary cirrhosis: antimitochondrial antibodies are unique to clinical disease and are absent in asymptomatic family members. Hepatology 1992. 16:899-905.
    View this article via: PubMed CrossRef Google Scholar
12. Casiano, CA, Martin, SJ, Green, DR, Tan, EM. Selective cleavage of nuclear autoantigens during CD95 (Fas/APO-1)-mediated T cell apoptosis. J Exp Med 1996. 184:765-770.
    View this article via: PubMed CrossRef Google Scholar
13. Casciola-Rosen, LA, Anhalt, GJ, Rosen, A. DNA-dependent protein kinase is one of a subset of autoantigens specifically cleaved early during apoptosis. J Exp Med 1995. 182:1625-1634.
    View this article via: PubMed CrossRef Google Scholar
14. Casciola-Rosen, LA, Anhalt, G, Rosen, A. Autoantigens targeted in systemic lupus erythematosus are clustered in two populations of surface structures on apoptotic keratinocytes. J Exp Med 1994. 179:1317-1330.
    View this article via: PubMed CrossRef Google Scholar
15. Utz, PJ, Hottelet, M, Schur, PH, Anderson, P. Proteins phosphorylated during stress-induced apoptosis are common targets for autoantibody production in patients with systemic lupus erythematosus. J Exp Med 1997. 185:843-854.
    View this article via: PubMed CrossRef Google Scholar
16. Huang, FP, et al. A discrete subpopulation of dendritic cells transports apoptotic intestinal epithelial cells to T cell areas of mesenteric lymph nodes. J Exp Med 2000. 191:435-444.
    View this article via: PubMed CrossRef Google Scholar
17. Sauter, B, et al. Consequences of cell death: exposure to necrotic tumor cells, but not primary tissue cells or apoptotic cells, induces the maturation of immunostimulatory dendritic cells. J Exp Med 2000. 191:423-434.
    View this article via: PubMed CrossRef Google Scholar
18. Casciola-Rosen, L, Andrade, F, Ulanet, D, Wong, WB, Rosen, A. Cleavage by granzyme B is strongly predictive of autoantigen status: implications for initiation of autoimmunity. J Exp Med 1999. 190:815-826.
    View this article via: PubMed CrossRef Google Scholar
19. Koga, H, Sakisaka, S, Ohishi, M, Sata, M, Tanikawa, K. Nuclear DNA fragmentation and expression of Bcl-2 in primary biliary cirrhosis. Hepatology 1997. 25:1077-1084.
    View this article via: PubMed CrossRef Google Scholar
20. Harada, K, Ozaki, S, Gershwin, ME, Nakanuma, Y. Enhanced apoptosis relates to bile duct loss in primary biliary cirrhosis. Hepatology 1997. 26:1399-1405.
    View this article via: PubMed CrossRef Google Scholar
21. Graham, AM, Dollinger, MM, Howie, SE, Harrison, DJ. Bile duct cells in primary biliary cirrhosis are ‘primed’ for apoptosis. Eur J Gastroenterol Hepatol 1998. 10:553-557.
    View this article via: PubMed Google Scholar
22. Hockenbery, DM, Oltvai, ZN, Yin, XM, Milliman, CL, Korsmeyer, SJ. Bcl-2 functions in an antioxidant pathway to prevent apoptosis. Cell 1993. 75:241-251.
    View this article via: PubMed CrossRef Google Scholar
23. Tsujimoto, Y, Shimizu, S, Eguchi, Y, Kamiike, W, Matsuda, H. Bcl-2 and Bcl-xL block apoptosis as well as necrosis: possible involvement of common mediators in apoptotic and necrotic signal transduction pathways. Leukemia 1997. 11(Suppl. 3):380-382.
    View this article via: PubMed Google Scholar
24. Tyurin, VA, et al. Glutamate-induced cytotoxicity in PC12 pheochromocytoma cells: role of oxidation of phospholipids, glutathione and protein sulfhydryls revealed by bcl-2 transfection. Brain Res Mol Brain Res 1998. 60:270-281.
    View this article via: PubMed CrossRef Google Scholar
25. Charlotte, F, et al. Immunohistochemical detection of bcl-2 protein in normal and pathological human liver. Am J Pathol 1994. 144:460-465.
    View this article via: PubMed Google Scholar
26. Nakopoulou, L, et al. Bcl-2 protein expression in acute and chronic hepatitis, cirrhosis and hepatocellular carcinoma. Pathol Res Pract 1999. 195:19-24.
    View this article via: PubMed Google Scholar
27. Casciola-Rosen, LA, Miller, DK, Anhalt, GJ, Rosen, A. Specific cleavage of the 70-kDa protein component of the U1 small nuclear ribonucleoprotein is a characteristic biochemical feature of apoptotic cell death. J Biol Chem 1994. 269:30757-30760.
    View this article via: PubMed Google Scholar
28. Vroman, B, LaRusso, NF. Development and characterization of polarized primary cultures of rat intrahepatic bile duct epithelial cells. Lab Invest 1996. 74:303-313.
    View this article via: PubMed Google Scholar
29. Watson, AJ, Levine, S, Donowitz, M, Montrose, MH. Kinetics and regulation of a polarized Na(+)-H+ exchanger from Caco-2 cells, a human intestinal cell line. Am J Physiol 1991. 261:G229-G238.
    View this article via: PubMed Google Scholar
30. Greidinger, EL, Miller, DK, Yamin, TT, Casciola-Rosen, L, Rosen, A. Sequential activation of three distinct ICE-like activities in Fas-ligated Jurkat cells. FEBS Lett 1996. 390:299-303.
    View this article via: PubMed CrossRef Google Scholar
31. Shirasuna, K, Sato, M, Miyazaki, T. A neoplastic epithelial duct cell line established from an irradiated human salivary gland. Cancer 1981. 48:745-752.
    View this article via: PubMed CrossRef Google Scholar
32. Voehringer, DW, McConkey, DJ, McDonnell, TJ, Brisbay, S, Meyn, RE. Bcl-2 expression causes redistribution of glutathione to the nucleus. Proc Natl Acad Sci USA 1998. 95:2956-2960.
    View this article via: PubMed CrossRef Google Scholar
33. Alpini, G, et al. Heterogeneity of the proliferative capacity of rat cholangiocytes after bile duct ligation. Am J Physiol 1998. 274:G767-G775.
    View this article via: PubMed Google Scholar
34. Griffith, OW. Determination of glutathione and glutathione disulfide using glutathione reductase and 2-vinylpyridine. Anal Biochem 1980. 106:207-212.
    View this article via: PubMed CrossRef Google Scholar
35. Mancini, M, et al. The caspase-3 precursor has a cytosolic and mitochondrial distribution: implications for apoptotic signaling. J Cell Biol 1998. 140:1485-1495.
    View this article via: PubMed CrossRef Google Scholar
36. Mendel-Hartvig, I, Nelson, BD, Loof, L, Totterman, TH. Primary biliary cirrhosis: further biochemical and immunological characterization of mitochondrial antigens. Clin Exp Immunol 1985. 62:371-379.
    View this article via: PubMed Google Scholar
37. Cai, J, Jones, DP. Mitochondrial redox signaling during apoptosis. J Bioenerg Biomembr 1999. 31:327-334.
    View this article via: PubMed CrossRef Google Scholar
38. Celli, A, Que, FG, Gores, GJ, LaRusso, NF. Glutathione depletion is associated with decreased Bcl-2 expression and increased apoptosis in cholangiocytes. Am J Physiol 1998. 275:G749-G757.
    View this article via: PubMed Google Scholar
39. Que, FG, Gores, GJ, LaRusso, NF. Development and initial application of an in vitro model of apoptosis in rodent cholangiocytes. Am J Physiol 1997. 272:G106-G115.
    View this article via: PubMed Google Scholar
40. Pablos, JL, Carreira, PE, Serrano, L, Del Castillo, P, Gomez-Reino, JJ. Apoptosis and proliferation of fibroblasts during postnatal skin development and scleroderma in the tight-skin mouse. J Histochem Cytochem 1997. 45:711-719.
    View this article via: PubMed Google Scholar
41. Meredith, MJ, et al. Expression of Bcl-2 increases intracellular glutathione by inhibiting methionine-dependent GSH efflux. Biochem Biophys Res Commun 1998. 248:458-463.
    View this article via: PubMed CrossRef Google Scholar
42. Lohr, H, et al. Autoreactive liver-infiltrating T cells in primary biliary cirrhosis recognize inner mitochondrial epitopes and the pyruvate dehydrogenase complex. J Hepatol 1993. 18:322-327.
    View this article via: PubMed CrossRef Google Scholar
43. Van de Water, J, et al. The role of T cells in primary biliary cirrhosis. Semin Liver Dis 1997. 17:105-113.
    View this article via: PubMed Google Scholar
44. Jensen, PE. Reduction of disulfide bonds during antigen processing: evidence from a thiol-dependent insulin determinant. J Exp Med 1991. 174:1121-1130.
    View this article via: PubMed CrossRef Google Scholar
45. Hattori, N, et al. T cells that are autoreactive to beta2-glycoprotein I in patients with antiphospholipid syndrome and healthy individuals. Arthritis Rheum 2000. 43:65-75.
    View this article via: PubMed CrossRef Google Scholar
46. Barlow, AK, He, X, Janeway, C. Exogenously provided peptides of a self-antigen can be processed into forms that are recognized by self-T cells. J Exp Med 1998. 187:1403-1415.
    View this article via: PubMed CrossRef Google Scholar
47. Ballardini, G, et al. Aberrant expression of HLA-DR antigens on bile duct epithelium in primary biliary cirrhosis: relevance to pathogenesis. Lancet 1984. 2:1009-1013.
    View this article via: PubMed Google Scholar
48. Dini, L. Recognizing death: liver phagocytosis of apoptotic cells. Eur J Histochem 2000. 44:217-227.
    View this article via: PubMed Google Scholar
49. van de Water, J, et al. Molecular mimicry in primary biliary cirrhosis. Evidence for biliary epithelial expression of a molecule cross-reactive with pyruvate dehydrogenase complex-E2. J Clin Invest 1993. 91:2653-2664.
    View this article via: JCI PubMed CrossRef Google Scholar
50. Sullivan, DM, Wehr, NB, Fergusson, MM, Levine, RL, Finkel, T. Identification of oxidant-sensitive proteins TNF-alpha induces protein glutathiolation. Biochemistry 2000. 39:11121-11128.
    View this article via: PubMed CrossRef Google Scholar
51. Costantini, P, et al. Oxidation of a critical thiol residue of the adenine nucleotide translocator enforces Bcl-2-independent permeability transition pore opening and apoptosis. Oncogene 2000. 19:307-314.
    View this article via: PubMed CrossRef Google Scholar
52. Scarlett, JL, Murphy, MP. Release of apoptogenic proteins from the mitochondrial intermembrane space during the mitochondrial permeability transition. FEBS Lett 1997. 418:282-286.
    View this article via: PubMed CrossRef Google Scholar
53. Kantrow, SP, Piantadosi, CA. Release of cytochrome c from liver mitochondria during permeability transition. Biochem Biophys Res Commun 1997. 232:669-671.
    View this article via: PubMed CrossRef Google Scholar
54. Susin, SA, et al. Molecular characterization of mitochondrial apoptosis-inducing factor. Nature 1999. 397:441-446.
    View this article via: PubMed CrossRef Google Scholar
55. Schultheiss, HP, Berg, P, Klingenberg, M. The mitochondrial adenine nucleotide translocator is an antigen in primary biliary cirrhosis. Clin Exp Immunol 1983. 54:648-654.
    View this article via: PubMed Google Scholar
56. van de Water, J, Gershwin, ME, Leung, P, Ansari, A, Coppel, RL. The autoepitope of the 74-kD mitochondrial autoantigen of primary biliary cirrhosis corresponds to the functional site of dihydrolipoamide acetyltransferase. J Exp Med 1988. 167:1791-1799.
    View this article via: PubMed CrossRef Google Scholar
57. Quinn, J, et al. Lipoylated and unlipoylated domains of human PDC-E2 as autoantigens in primary biliary cirrhosis: significance of lipoate attachment. Hepatology 1993. 18:1384-1391.
    View this article via: PubMed Google Scholar
58. Moteki, S, et al. Epitope mapping and reactivity of autoantibodies to the E2 component of 2-oxoglutarate dehydrogenase complex in primary biliary cirrhosis using recombinant 2-oxoglutarate dehydrogenase complex. Hepatology 1996. 23:436-444.
    View this article via: PubMed CrossRef Google Scholar
59. Dubel, L, et al. Autoepitope mapping and reactivity of autoantibodies to the dihydrolipoamide dehydrogenase-binding protein (E3BP) and the glycine cleavage proteins in primary biliary cirrhosis. Hepatology 1999. 29:1013-1018.
    View this article via: PubMed CrossRef Google Scholar
60. Teoh, KL, Mackay, IR, Rowley, MJ, Fussey, SP. Enzyme inhibitory autoantibodies to pyruvate dehydrogenase complex in primary biliary cirrhosis differ for mammalian, yeast and bacterial enzymes: implications for molecular mimicry. Hepatology 1994. 19:1029-1033.
    View this article via: PubMed Google Scholar
61. Rowley, MJ, et al. Diabetic sera react with the glutamic acid decarboxylase molecule in a dimeric-oligomeric form. Clin Exp Immunol 1996. 106:323-328.
    View this article via: PubMed CrossRef Google Scholar
62. Pollard, KM, et al. The autoimmunity-inducing xenobiotic mercury interacts with the autoantigen fibrillarin and modifies its molecular and antigenic properties. J Immunol 1997. 158:3521-3528.
    View this article via: PubMed Google Scholar
63. Kurrer, MO, Pakala, SV, Hanson, HL, Katz, JD. Beta cell apoptosis in T cell-mediated autoimmune diabetes. Proc Natl Acad Sci USA 1997. 94:213-218.
    View this article via: PubMed CrossRef Google Scholar
64. Grey, ST, Arvelo, MB, Hasenkamp, W, Bach, FH, Ferran, C. A20 inhibits cytokine-induced apoptosis and nuclear factor kappaB-dependent gene activation in islets. J Exp Med 1999. 190:1135-1146.
    View this article via: PubMed CrossRef Google Scholar
65. Trudeau, JD, et al. Neonatal beta-cell apoptosis: a trigger for autoimmune diabetes? Diabetes 2000. 49:1-7.
    View this article via: PubMed CrossRef Google Scholar