SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Our understanding of primary biliary cirrhosis (PBC) has been significantly enhanced by the rigorous dissection of the multilineage T and B cell response against the immunodominant mitochondrial autoantigen, the E2 component of the pyruvate dehydrogenase complex (PDC-E2). PDC-E2 is a ubiquitous protein present in mitochondria of nucleated cells. However, the damage of PBC is confined to small biliary epithelial cells (BECs). We have previously demonstrated that BECs translocate immunologically intact PDC-E2 to apoptotic bodies and create an apotope. To define the significance of this observation, we have studied the ability of biliary or control epithelial apotopes to induce cytokine secretion from mature monocyte-derived macrophages (MDMϕs) from either patients with PBC or controls in the presence or absence of anti-mitochondrial antibodies (AMAs). We demonstrate that there is intense inflammatory cytokine production in the presence of the unique triad of BEC apotopes, macrophages from patients with PBC, and AMAs. The cytokine secretion is inhibited by anti-CD16 and is not due to differences in apotope uptake. Moreover, MDMϕs from PBC patients cultured with BEC apoptotic bodies in the presence of AMAs markedly increase tumor necrosis factor–related apoptosis-inducing ligand expression. Conclusion: These results provide a mechanism for the biliary specificity of PBC, the recurrence of disease after liver transplantation, and the success of ursodiol in treatment. They further emphasize the critical role of the innate immune system in the perpetuation of this autoimmune disease. (HEPATOLOGY 2010;)

There have been significant advances in our understanding of primary biliary cirrhosis (PBC),1 including the dissection of the autoreactive CD4 and CD8 responses2-5 and the molecular characteristics of anti-mitochondrial antibodies (AMAs).6, 7 The results of these studies suggest that PBC ensues from a multilineage loss of tolerance to the E2 component of the pyruvate dehydrogenase complex (PDC-E2), the immunodominant autoantigen of PBC.8, 9 Although the mechanisms involved in the loss of tolerance remain unknown, the roles of genetic susceptibility10 and environmental factors that modify the autoantigen motif and contribute to the breakdown of tolerance have gained attention.11, 12 However, a major unanswered question regarding the pathogenesis of PBC is the specific targeting of the small biliary duct epithelial cell. All nucleated cells have mitochondria, yet only small biliary epithelial cells (BECs) and, to a lesser extent, salivary gland cells are the targets of the autoimmune attack in PBC.

Apoptotic cells are normally efficiently cleared after engulfment by professional phagocytes followed by an anti-inflammatory response.13, 14 When such uptake is impaired, cell lysis can release intracellular components that are a potential source of autoantigenic stimulation and autoimmunity onset.15 We have recently demonstrated that, in contrast to other epithelial cells, small BECs translocate immunologically intact PDC-E2 to apoptotic bodies formed during apoptosis.16 We and others have called the epitope expressed on apoptotic cells an apotope,17, 18 and we submit that the biliary apotope has important biological and clinical significance in PBC. To investigate the unique target cell specificity of PBC, we have studied the ability of apoptotic bodies from either small BECs or control epithelial cells to induce cytokine secretion from macrophages of either patients with PBC or control subjects in the presence or absence of AMAs.

We report that the unique triad consisting of macrophages from patients with PBC, BEC apotopes, and AMAs leads to a burst of inflammatory cytokines. This finding implies that AMAs contribute to bile duct damage and explains the tissue specificity of PBC. Moreover, our results offer an explanation for the recurrence of PBC after liver transplantation,19 the relative failure of immunosuppressive drugs to modify what is considered a model autoimmune disease,20 and the efficacy in PBC of ursodiol, a drug that has antiapoptotic properties.21

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Subjects.

Fresh heparinized peripheral blood samples were obtained from patients diagnosed with PBC (n = 25), unaffected controls (n = 20), subjects with primary sclerosing cholangitis (PSC; n = 6), and subjects with systemic lupus erythematosus (SLE; n = 3).1, 22, 23 All patients with PBC were women and had detectable AMAs. The mean age was 56 years (range = 43-66 years), and 70% were taking ursodiol. Patients with PBC had histological stage I disease (n = 7), stage II disease (n = 15), or stage III disease (n = 3), and they were excluded if they had histological stage IV disease. Subjects were also excluded if they had malignancies or were using immunosuppressive drugs. Patients and controls were matched for age and sex, and separate unrelated donors were used for each independent experiment. The institutional review board of the University of California at Davis approved the study protocol, and all subjects provided written, informed consent.

Generation of Monocyte-Derived Macrophages (MDMϕs).

Human mononuclear cells were isolated under endotoxin-free conditions from buffy coats of centrifuged peripheral blood, and this was followed by anti-CD14 microbead-assisted magnetic cell sorting (Miltenyi Biotec). The purity of the monocytic population was >95% as assessed by flow cytometry. Aliquots of monocytes (0.5 × 106/mL) were then resuspended in Roswell Park Memorial Institute (RPMI) culture medium containing 10% heat inactivated fetal bovine serum, 90 U/mL penicillin, and 90 μg/mL streptomycin supplemented with 100 ng/mL granulocyte-macrophage colony-stimulating factor (GM-CSF), and they were cultured for at least 5 days in a humidified 5% CO2 incubator. Viability was confirmed by trypan blue dye exclusion. The only preparations used were those in which the percentage of CD14+ cells was >95% and the viability was >90%.

Cell Culture.

Human intrahepatic biliary epithelial cells (HIBECs) from two healthy donors were purchased from ScienCell. Data from both sets of healthy donors were similar and were combined. Primary cultures of human keratinocytes (a gift from Dr. Rivkah Isseroff, University of California at Davis) were used as controls on the basis of our previous data.16 Previous research on apoptosis also used bronchial epithelial cells, HeLa cells, and CaCo-2 cells. Both the HIBECs and the keratinocytes were obtained from normal human tissue and were cryopreserved immediately after isolation. All experiments were performed with cells between passages 2 and 8. HIBECs were cultured in a sterile epithelial cell medium (ScienCell) supplemented with 2% fetal bovine serum, an epithelial cell growth supplement (ScienCell), and 1% penicillin. Keratinocytes were cultured in EpiLife medium (Gibco-Invitrogen) supplemented with a bovine pituitary extract, bovine insulin hydrocortisone, bovine transferrin, and human epidermal growth factor. Cells were cultured at 37°C in a humidified 5% CO2 incubator.

Apoptosis Induction and Isolation of Apoptotic Bodies.

Apoptosis was induced with bile salts as described.16 After induction of apoptosis, supernatants were collected and centrifuged twice (at 500g for 5 minutes) for the removal of intact cells. The supernatant fluid was then passed through a 1.2-μm nonpyrogenic, hydrophilic syringe filter. After centrifugation at 100,000g for 45 minutes, the pellet containing apoptotic bodies was resuspended in RPMI medium. The number of apoptotic bodies was determined, and the suspension was immediately used for coculture with MDMϕs.

Antibody Reagents and AMA Purification.

Human immunoglobulin G (IgG) was purified from sera of PBC patients with high levels of AMAs with a protein G column (Pierce). Similarly, serum IgG from unaffected controls was purified. AMA immunoglobulin A (IgA) from a patient diagnosed with PBC and POEMS syndrome (polyneuropathy, organomegaly, endocrinopathy/edema, M-protein, and skin abnormalities) was purified with jacalin-agarose beads (Pierce); serum IgA from normals was used as a control. All experiments were conducted under sterile conditions, and solutions were filtered through a 0.22-μm filter.

Culture Assay.

Prior to coculturing, all cells were washed in RPMI medium. Aliquots of 105 MDMϕs were suspended in 200 μL of RPMI-1640 medium with 10% fetal bovine serum and were plated in a 96-well culture plate with apoptotic bodies at a concentration of 10 apoptotic bodies per cell. At the same time, AMAs or control antibodies were added to the culture (100 μg/mL). Controls for each experiment included apoptotic bodies prepared from keratinocytes, control (non-AMA) antibodies, and appropriate isotype-matched immunoglobulin. For a positive control, lipopolysaccharide was included in the culture medium at a final concentration of 10 μg/mL. MDMϕs cultured alone were used as a negative control. Other controls included cultures of MDMϕs with apoptotic bodies alone and MDMϕs with antibody alone. Anti-CD16 antibody (clone 3G8) and anti-CD16/32 were purchased from BD Pharmingen and Miltenyi Biotec, respectively, and used to block the fragment crystallizable receptor (FcR; 10 μg/mL). The treatment of MDMϕs with anti-CD16 was initiated 30 minutes before the addition of apoptotic bodies. For full M1 polarization, MDMϕs from healthy controls were stimulated with 100 U/mL interferon gamma (IFN-γ; Boehringer Ingelheim) and 1 ng/mL lipopolysaccharide (Cambrex) during the last 24 hours of culture. All cultures were performed in RPMI-1640 medium for 24 hours at 37°C. Cell culture supernatants were collected, centrifuged for the removal of cell debris, harvested, and stored at −80°C until use. All samples were run twice.

Cytokine Detection.

Culture supernatants were analyzed on the Luminex multiplex platform with a human cytokine antibody bead kit (Millipore). The assay kit allowed for the detection of interleukin-2 (IL-2), IL-4, IL-6, IL-10, IL-12p40, IFN-γ, tumor necrosis factor alpha (TNF-α), GM-CSF, and macrophage inflammatory protein 1 beta (MIP-1β). Briefly, analyte-specific antibody conjugated beads were incubated with 50 μL of a biotinylated detection antibody solution and 50 μL of the standard or 25 μL of a sample in 25 μL of an assay diluent for 3 hours at room temperature in the dark on an orbital shaker (500-600 rpm). After the wells were washed and aspirated by vacuum manifold aspiration, 100 μL of streptavidin-conjugated R-phycoerythrin (PE) was added to each well and incubated for 30 minutes at room temperature in the dark on an orbital shaker (500-600 rpm). PE fluorescence was measured with the Luminex 100 LabMAP system. Cytokine concentrations from the obtained mean fluorescence values were calculated from standard curves of each tested cytokine with Bio-Plex Manager software. Macrophage colony-stimulating factor (M-CSF) was measured in sera with an enzyme-linked immunosorbent assay kit (Antigenix America).24

Evaluation of Cell Phenotypes.

Phenotypic analysis of peripheral blood mononuclear cells (PBMCs) and MDMϕs was assessed on a FACScan flow cytometer upgraded by Cytec Development with fluorochrome-conjugated monoclonal antibodies against cell surface markers CD3 (BD Biosciences), CD1a (Biolegend), CD14 (Caltag), CD16 (Biolegend), chemokine (C-C motif) receptor 2 (CCR2; RD Systems), chemokine (C-X3-C motif) receptor 1 (CX3CR1; Biolegend), CD80 (eBioscience), CD83 (Caltag), and human leukocyte antigen DR (HLA-DR; Pharmingen). The appropriate fluorochrome-conjugated isotype control antibodies were used as negative controls. Acquired data were analyzed with the CellQuest Pro (BD Immunocytometry Systems) and FlowJo (Tree Star, Inc.) software packages.

Immunofluorescence and Confocal Microscopy.

Uptake of apoptotic bodies by MDMϕs was evaluated by immunofluorescence microscopy. The apoptotic bodies used in these assays were isolated as follows. HIBECs or control keratinocytes were first stained with carboxyfluorescein diacetate succinimidyl ester (CFSE; 3 μM; Molecular Probes) for 20 minutes at 37°C. CFSE-labeled cells were then induced to undergo apoptosis with bile salts. After 3 hours of incubation, apoptotic bodies were isolated from cellular supernatants as previously described. MDMϕs were coincubated for 24 hours with CFSE-labeled apoptotic bodies from HIBECs or the keratinocyte cultures in the presence of human AMA IgG or control IgG (100 μg/mL). When anti-CD16 was used, it was added to the culture system 30 minutes prior to AMAs. Cells were then fixed in 3.7% formaldehyde (5 minutes at room temperature) and stained with PE-conjugated anti-human CD14 antibody diluted 1:200 overnight at 4°C. These cells were incubated with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen) to stain the nucleus. Controls consisting of incubating cells with the secondary antibody alone did not demonstrate any detectable staining under the same conditions. For the phagocytosis experiments, MDMϕs were cultured in glass-bottom dishes, and apoptotic bodies were added to the culture medium at a constant ratio. Immunofluorescence-labeled samples were examined with an LSM5 Pascal confocal laser scanning microscope (Zeiss) with a 100× oil-immersion objective.

Phagocytosis was evaluated by random counting of 100 MDMϕs per well at ×100 with an oil-immersion objective. The percentage of phagocytosis was calculated via the counting of the number of macrophages that had ingested at least one apoptotic body. The phagocytic index was expressed as the percentage of phagocytosis multiplied by the mean number of phagocytosed bodies per macrophage and was evaluated at 0, 4, and 24 hours of incubation.

Tumor Necrosis Factor–Related Apoptosis-Inducing Ligand (TRAIL) and Fas Ligand (FasL) Expression.

Total RNA was extracted from MDMϕs with the Qiagen RNeasy mini kit. One microgram of total RNA was reverse-transcribed and quantified by real-time PCR on an ABI-Prism 7900HT sequence detection system. Amplification was performed for 40 cycles in a total volume of 12 μL, and products were detected with SYBR Green. The relative expression level of each target gene was determined by normalization of its messenger RNA (mRNA) level to the internal control glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The primers were as follows: FasL, 5′-AAAGTGGCCCATTTAACAGGC-3′ (forward) and 5′-AAAGCAGGACAATTCCATAGGTG-3′ (reverse); TRAIL, 5′-ATGGCTATGATGGAGGTCCAG-3′ (forward) and 5′-TTGTCCTGCATCTGCTTCAGC-3′ (reverse); and GAPDH, 5′-CCATGGAGAAGGCTG GGG-3′ (forward) and 5′-CAAAGTTGTCATGGAT GACC-3′ (reverse).

Statistical Analysis.

Statistical differences between groups were determined with a two-tailed Mann-Whitney nonparametric test with the 95% confidence interval (CI). All results were expressed as means and standard errors of the mean. The percentages of phagocytosis in different cell populations were compared with Fisher's exact test. We used the Prism statistical package (GraphPad Software, Inc.) to carry out the analyses.

Results

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Macrophages from Patients with PBC Cultured with Apoptotic Bodies from HIBECs Secrete Proinflammatory Cytokines Only in the Presence of AMAs.

To determine factors that stimulate cytokine secretion in PBC, we incubated MDMϕs from patients with PBC together with AMAs and apoptotic bodies from HIBECs; this led to increased secretion of proinflammatory cytokines. In the presence of AMAs, there was a 10-fold higher level of secreted TNF-α and 2-fold higher levels of secreted IL-6, IL-10, MIP-1b, and IL-12p40 in comparison with cytokine secretion in the presence of control IgG (P < 0.001; Fig. 1A). Although GM-CSF was added to the monocyte cultures to stimulate macrophage differentiation, its concentration in the culture supernatant was significantly higher in patients with PBC versus controls (P < 0.001), and this suggested supplemental production of GM-CSF. The patient's histological stage did not affect the cytokine levels or profiles secreted from the cultured cells. The level of cytokine release from MDMϕs of unaffected controls was independent of the presence of the antibody (either AMAs or control) or the source of the apoptotic bodies (HIBECs or control; Fig. 1B). Other controls included cultures of MDMϕs from patients with PBC (n = 25) that were incubated only with apoptotic bodies from either HIBECs or control epithelial cells, which showed no significant level of cytokine production. This was also true when apoptotic bodies were incubated with MDMϕs from 20 unaffected donors (data not shown).

thumbnail image

Figure 1. Macrophages from patients with PBC, cultured with apoptotic bodies from HIBECs, secreted proinflammatory cytokines in the presence of AMAs. MDMϕs from (A) patients with PBC and (B) unaffected controls were cultured with apoptotic bodies and antibodies (AMA IgG or control IgG). The triad of MDMϕs from patients with PBC (but not from controls), apoptotic bodies from HIBECs, and AMAs led to significantly increased secretion of TNF-α, IL-6, IL-10, MIP-1b, and IL-12p40 in comparison with the IgG control. Each experiment was performed twice. ***The level of significance was P < 0.001 (based on a two-tailed Mann-Whitney test with the 95% CI).

Download figure to PowerPoint

IgA and IgG AMA isotypes equally stimulated secretion of cytokines from MDMϕs of patients with PBC in the presence of HIBEC apoptotic bodies (Fig. 2). We also evaluated the inflammatory response of MDMϕs from patients with PSC (n = 6) or SLE (n = 3) under the same conditions in the presence of AMAs and HIBEC apoptotic bodies. The MDMϕ cells from patients with PBC demonstrated significantly higher levels of cytokine production versus those from patients with PSC (Fig. 3) or SLE (data not shown). Hence, AMA-stimulated cytokine secretion was specific for PBC in the presence of HIBEC apoptotic bodies.

thumbnail image

Figure 2. Both IgA and IgG isotypes stimulated MDMϕs from patients with PBC. MDMϕs from 10 patients with PBC and from 10 controls were cultured with HIBEC apoptotic bodies in the presence of two different AMA isotypes (IgG and IgA). Supernatants were collected after 24 hours of incubation, and the cytokine concentration was determined. Although cytokine levels were markedly increased (see Fig. 1A), there was no significant difference in levels of TNF-α, IL-6, IL-10, MIP-1b, IL-12p40, and GM-CSF between AMA IgG and IgA. Abbreviation: NS, not significant.

Download figure to PowerPoint

thumbnail image

Figure 3. MDMϕs from PBC patients secreted significantly higher levels of cytokines in comparison with MDMϕs from PSC patients. Cytokine secretion was studied in supernatants that were collected at 24 hours from MDMϕs isolated from patients with PBC (n = 25) or PSC (n = 6) and cultured with HIBEC apoptotic bodies in the presence of AMAs. Levels of TNF-α, IL-6, IL-10, MIP-1b, and IL-12p40 were significantly higher in cultures with MDMϕs from PBC patients versus PSC patients. Statistical differences (***P < 0.001) were analyzed with a two-tailed Mann-Whitney test (95% CI). All samples were run twice.

Download figure to PowerPoint

Phenotype of Monocytes and MDMϕs from Patients with PBC and Controls.

To address the question of whether cytokine secretion was a result of a different state of MDMϕ activation in patients with different diseases, we determined the phenotypes of monocytes from PBMCs and MDMϕs. Freshly isolated PBMCs and MDMϕs were stained with fluorochrome-conjugated antibodies to CD14, CD16, CCR2, CX3CR1, HLA-DR, CD80, and CD83, and the phenotypes were assessed by flow cytometry (Fig. 4). Monocytes from freshly isolated PBMCs were gated on the basis of CD14 expression into CD14high and CD14low subpopulations (Fig. 4A).25, 26 These gated subsets were then analyzed for the expression of CD16 and chemokine receptors, and two classical subpopulations of monocytes were identified: classical CD14high CD16 CCR2+ (87%) and a minor subset of CD14low CD16+ CX3CR1+ (13%). This monocyte phenotype from PBC patients was essentially the same as the monocyte phenotype from control subjects. CD14+ monocytes were then stimulated with GM-CSF for 5 days and cultured in vitro with apoptotic bodies. Stimulated MDMϕs expressed higher levels of CD83, CD80, and HLA-DR in comparison with prestimulated monocytes (Fig. 4B and Table 1).25, 27 The same profile of increased CD80, CD83, and HLA-DR was also noted in MDMϕs from control donors (Table 1) and from patients with PSC and SLE (data not shown). For the assessment of whether M-CSF determines different states of MDMϕ activation, serum levels of M-CSF from patients with PBC and controls were analyzed; however, no differences were revealed (data not shown).

thumbnail image

Figure 4. Phenotype of circulating monocytes and MDMϕs as determined by a fluorescence-activated cell sorting analysis of monocyte subsets and phenotype of monocytes and MDMϕs in patients with PBC and in controls. Cells were stained with fluorochrome-conjugated antibodies to CD14, CD16, CCR2, CX3CR1, HLA-DR, CD80, and CD83 (A) Monocytes and lymphocytes from PBMCs were gated on the basis of forward and side scatter profiles, and two monocyte subpopulations were identified: classical CD14high CD16 CCR2+ (87%) and a minor subset of CD14low CD16+ CX3CR1+ (13%). (B) The enriched population of CD14+ monocytes was stimulated to develop MDMϕs and was cultured with apoptotic bodies. No differences in the phenotypes of PBMCs and MDMϕs were observed between PBC patients and HCs. Abbreviations: Cy7, cyanine 7; FITC, fluorescein isothiocyanate; HC, healthy control.

Download figure to PowerPoint

Table 1. Phenotypes of Monocytes from PBMCs and MDMϕs from Patients with PBC and Healthy Controls
 PBMCsMDMϕs
Patients with PBCHealthy ControlsPPatients with PBCHealthy ControlsP
  1. Abbreviations: MFI, median fluorescence intensity; NA, not available; NS, not significant.

Total CD14 (106/mL)1.1 ± 0.181.34 ± 0.13NS0.68 ± 0.130.66 ± 0.12NS
CD14+ CD16low CCR2+ (106/mL)0.32 ± 0.090.37 ± 0.13NSNA 
CD14low CD16+ CX3CR1+ (106/mL)0.06 ± 0.020.08 ± 0.02NSNA 
CD14+ CD83+ (MFI)17.07 ± 1.5613.89 ± 0.87NS114.01 ± 6.53121.59 ± 4.32NS
CD14+ CD80+ (MFI)6.07 ± 0.185.87 ± 0.18NS137.45 ± 11.43138.84 ± 8.50NS
CD14+ HLA-DR+ (MFI)218.99 ± 28.81242.31 ± 16.88NS776.62 ± 88.02964.46 ± 88.08NS
Fragment Crystallizable Receptor for Immunoglobulin G (FcγR) Mediates Cytokine Production.

To determine the role of FcγR in cytokine production, we evaluated the responses of MDMϕs preincubated with or without anti-CD16 monoclonal antibody to the blocking of FcγRIII; this was followed by the addition of the apoptotic bodies and IgG AMA or control sera. MDMϕs from patients with PBC had markedly reduced cytokine production in the presence of anti-CD16 (P < 0.001; Fig. 5A). After FcγR was blocked, the cytokine levels were similar to those secreted by MDMϕs from patients with PBC incubated with HIBEC apoptotic bodies in the presence of control IgG.

thumbnail image

Figure 5. Stimulation of PBC macrophages in the presence of AMAs and HIBEC apoptotic bodies was mediated by FcR and was a result of M1 polarization. (A) Comparison of cytokine secretion from MDMϕs from patients with PBC (n = 8) that were cultured with HIBEC apoptotic bodies in the presence of AMAs with (white bars) or without (black bars) pretreatment with CD16 monoclonal antibodies (clone 3G8). Cytokine levels were markedly decreased when FcγR was blocked (***P < 0.001). (B) Comparison of TNF-α secretion from MDMϕs from patients with PBC (n = 8) and from HCs after M1 polarization and untreated HCs (n = 8) that were cultured with HIBEC apoptotic bodies in the presence of AMAs (***P < 0.001, *P < 0.05). Abbreviations: HC, healthy control; NS, not significant.

Download figure to PowerPoint

M1-Polarized MDMϕs from Healthy Controls Show Strong Reactivity Against AMAs and Apoptotic Bodies from Biliary Cells.

We hypothesized that the proinflammatory cytokine response of MDMϕs from PBC patients but not controls might have resulted from M1 polarization of macrophages in PBC patients. When MDMϕs from healthy controls were polarized toward an M1 phenotype with IFN-γ and lipopolysaccharide and cocultured with AMAs and HIBEC apoptotic bodies, they secreted significantly greater amounts of TNF-α in comparison with untreated MDMϕs from the same subjects but not in comparison with MDMϕs from PBC patients (Fig. 5B).

Engulfment of Apoptotic Bodies by Monocyte-Derived Phagocytes Is Reduced in Patients with PBC.

To assess whether MDMϕ cytokine production was associated with the ability to phagocytose apoptotic bodies, MDMϕs from patients with PBC and controls were coincubated for 24 hours with CFSE-labeled apoptotic bodies from HIBECs in the presence of human AMA IgG or control IgG, and the percentage of phagocytosis and the phagocytic index were determined. MDMϕs from PBC and all control groups phagocytosed apoptotic bodies during the 24-hour incubation period (Fig. 6A). However, MDMϕs from PBC demonstrated reduced uptake (17% versus 47%, P < 0.001) of apoptotic bodies (Fig. 6B) associated with a decrease in the phagocytic index at 4 and 24 hours (Fig. 6C). Blocking FcγR with anti-CD16 did not modify the ability of MDMϕs from patients with PBC or controls to take up apoptotic bodies (Fig. 6D), and this suggests that cytokine production and phagocytosis follow independent pathways.

thumbnail image

Figure 6. Uptake of apoptotic bodies. (A) Confocal imaging of MDMϕs from a representative patient with PBC and an unaffected control that were coincubated for 24 hours with CFSE-labeled apoptotic bodies from HIBECs in the presence of human AMA IgG or control IgG. After culturing, the cells were fixed and stained with PE-conjugated anti-human CD14 (red); DAPI was used to stain the nucleus (blue), and CFSE-stained apoptotic bodies are shown in green. In all four conditions, apoptotic bodies (green) were located inside the cells, and this indicated that MDMϕs actively engulfed apoptotic bodies after 24 hours (arrows). The scale bars represent 20 μm. (B) Comparison of phagocytic efficacy, expressed as a percentage of phagocytosis, of MDMϕs from PBC and control patients. MDMϕs from PBC patients had a reduced uptake of HIBEC apoptotic bodies in the presence of AMAs (17% versus 47%, ***P < 0.001); this difference was not noted when an IgG control was used (34% versus 49%, P = not significant). The percentage of phagocytosis was calculated by the counting of the number of macrophages that had ingested at least one HIBEC apoptotic body (Fisher's exact test). (C) The phagocytic index was expressed as the percentage of phagocytosis multiplied by the mean number of phagocytosed bodies per macrophage (ABMϕ): Phagocytic index = Phagocytosis percentage × Mean ABM ϕ/100 It was evaluated at 0, 4, and 24 hours of incubation (**P < 0.01). (D) The ability of MDMϕs from PBC and control patients to uptake HIBEC apoptotic bodies, expressed as the percentage of phagocytosis, was studied in the presence of AMAs with or without FcγR blocking. No significant difference in the percentage of phagocytosis was observed when FcγR was blocked in both patients and controls. Statistical differences were determined with Fisher's exact test. Abbreviations: HC, healthy control; NS, not significant.

Download figure to PowerPoint

TRAIL Expression Is Increased in MDMϕs from PBC Patients Cultured with Apoptotic Bodies from HIBECs in the Presence of AMAs.

TRAIL and FasL are proapoptotic ligands,28 and TRAIL can induce BEC cell damage.29 Because of the high levels of TNF-α secreted by MDMϕs from PBC patients when MDMϕs are cocultured with HIBEC apoptotic bodies in the presence of AMAs, we quantified mRNAs encoding TRAIL and FasL in MDMϕs after incubation with apoptotic bodies and antibodies. TRAIL mRNA expression was markedly increased in MDMϕs from PBC patients when MDMϕs were cocultured with HIBEC apoptotic bodies in the presence of AMAs (Fig. 7). In contrast, no significant differences in the expression of FasL were noted (data not shown). These results suggest that TRAIL but not FasL plays a role in the proinflammatory response of MDMϕs in PBC.

thumbnail image

Figure 7. TRAIL expression was increased in MDMϕs from PBC patients that were cultured with apoptotic bodies from HIBECs in the presence of AMAs. The relative expression levels of TRAIL in MDMϕs were compared with respect to MDMϕs, AMAs, apoptotic body cell types, and the use of anti-CD16 TRAIL expression was increased at least 4-fold (**P < 0.01) in PBC MDMϕs, AMAs, and apoptotic bodies from HIBECs and was significantly reduced (**P < 0.01) when the cells were pretreated with anti-CD16 Abbreviation: HC, healthy control.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Our present results and those of our previous studies provide insights into why the autoimmune damage in PBC is primarily confined to cells of the biliary epithelial tract. We previously reported that PDC-E2, the immunodominant autoantigen of PBC, remains immunologically intact in BECs after apoptosis and localizes within apoptotic bodies, where it is accessible to AMAs.16, 30 This phenomenon was specific to BECs and did not occur in other epithelial cell types tested. We now demonstrate that the triad of MDMϕs from patients with PBC, AMAs, and apoptotic bodies from BECs can stimulate a burst of proinflammatory cytokines. We hypothesize that within the intrahepatic biliary environment, the normal turnover of BECs becomes the source of the PDC-E2 apotope. PDC-E2 in the apoptotic bodies is recognized by circulating AMAs, and the apotope-AMA complex then stimulates macrophages from subjects with a susceptible genetic background. This eventually leads to a localized burst of proinflammatory cytokines, and the ensuing inflammation is associated with increased apoptosis of surrounding cells, including BECs, and perpetuation of local inflammation with chronic damage to the biliary tract. These data complete previous work from our group demonstrating that immune complexes including autoantibodies against PDC-E2 can be efficiently internalized, processed, and presented by antigen-presenting cells (APCs) to induce specific cytotoxic T lymphocytes in PBC patients.31 We emphasize that the BECs used in our experiments were from normal donors. Hence, our results imply that, on the basis of the apotope, the BECs in patients with PBC are innocent victims, and thus there is no specific phenotype of BECs in patients with PBC.

Our results can explain several phenomena that occur in PBC. First, they explain why only BECs are targeted in PBC. They further suggest that we should examine the salivary gland epithelium to see whether apoptotic bodies derived from these cells, which are also affected in the disease, are enriched with PBC-specific autoantigens. Moreover, PBC frequently recurs in patients after liver transplantation19 in a situation in which an absence of major histocompatibility complex matching would preclude presentation of PDC-E2 to elements of the adaptive immune system and compromise the ability of CD8+ T cells to attack BECs.5 Indeed, apoptosis of BECs in the newly transplanted liver completes the triad required for inflammatory cytokine release and provides a milieu in which disease can recur. Furthermore, because steroids mainly suppress T cell–mediated immunity,32 our results explain the unsatisfactory response of PBC to classic immunosuppressive agents20 and why ursodiol, an agent that suppresses apoptosis,21 is effective in slowing disease progression in PBC.21 Our results also have broader implications for other organ-specific autoimmune diseases in which the target autoantigen is widely distributed,33-36 such as neonatal cardiomyopathy of anti-Ro–positive mothers.37, 38

Macrophages are essential effector cells that undergo various forms of activation in response to cytokines and microbial signals. In particular, IFN-γ and IL-4 activate distinct functional programs, inducing M1 (or classic) activation and M2 (or alternative) activation, which mirror the T helper 1/T helper 2 dichotomy.39, 40 It is known that the innate immune system of patients with PBC demonstrates higher reactivity than controls. Indeed, peripheral monocytes from patients with PBC secrete high levels of cytokines.41 Moreover, the frequency and absolute number of blood and liver natural killer cells in patients with PBC, as well as their cytotoxic activity and perforin expression, are increased.42 However, very little is known about the state of macrophage activation in autoimmunity. Our data emphasize an essential role of the innate immune system in PBC and suggest that MDMϕs in patients with PBC are polarized, most likely to M1, as a result of the specific microenvironment.43, 44 This would explain the strong reactivity of MDMϕs from patients when they interact with AMAs and apoptotic bodies of HIBECs. A minority of patients with PBC (<5%), when tested with Luminex,45 do not have detectable serum AMAs. According to our model, these patients should have either biliary disease mediated by a different mechanism or antibodies that recognize still unidentified apotopes. In this respect, so-called AMA-negative patients with PBC are more likely to produce antinuclear antibodies than those with detectable AMAs.1 Future studies should test for the possible role of known nuclear autoantigens in the inflammatory cytokine response in PBC.

Recent data from a genome-wide association study46 and animal models47 have identified the IL-12 signaling pathway as a central player in the etiopathogenesis of PBC. In our study, MDMϕs from patients with PBC in the presence of HIBEC apoptotic bodies and AMAs secreted high levels of IL-12p40. Hence, our results provide additional support for a central role of IL-12 signaling in PBC pathogenesis. Our data further demonstrate that activation of MDMϕs from patients with PBC is mediated by FcγRIII. Activating FcγRs may facilitate antigen presentation and dendritic cell maturation,48 whereas in the late phase of the immune response, the inhibitory FcγRIIb may down-regulate B cell activation upon crosslinking with activating receptors.49 Our current results suggest that the immune complex of AMAs and BEC apotopes specifically engages FcγRIII and stimulates cytokine production. FcγRIII-mediated activation of APCs is known to be proinflammatory50; however, it is usually associated with stimulation of phagocytosis.51 We have demonstrated reduced uptake of HIBEC apoptotic bodies by MDMϕs in the presence of AMAs that is not CD16-mediated; this is similar to neonatal cardiomyopathy.38 This could decrease antigen clearance and contribute to the perpetuation of the innate immune response. However, our data also suggest that FcγRIII is not the only mechanism because AMA IgA is also able to stimulate macrophages from patients with PBC in the presence of HIBEC apoptotic bodies.

Macrophage activation may lead to direct damage to BECs, perhaps because of high levels of TNF-α or TRAIL, which belongs to the TNF family of cytokines and signals through receptors that contain cytoplasmic death domains.28 TRAIL can induce BEC cell damage,29 and bile acids can increase TRAIL's effects through the up-regulation of its receptor; the cellular responses to TRAIL have been related to the inflammatory status of the liver, and this indicates a link between cholestasis and TRAIL signaling.52 Our results thus highlight the essential role of the innate immune system in PBC and demonstrate a role of AMAs in BEC damage. Clearly, there are functional differences in the monocytic population from PBC that respond to apoptotic blebs of bile duct cells only in the presence of AMAs. Future studies must now define the specific responsive monocytic population with primary monocytes and plasmacytoid dendritic cells as well as GM-CSF–stimulated or GM-CSF/IL-4–stimulated macrophages and myeloid dendritic cells. Identification of this population will be important for defining functional signaling pathways using differential phosphorylation of specific signal proteins in cells stimulated with blebs/AMAs. These studies, including the potential use of next-generation deep sequencing, will not only determine all genes in the activated signaling pathways but also collectively determine why PBC monocytes react to induce a proinflammatory environment at the site of tissue damage.

References

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References