AMA production in primary biliary cirrhosis is promoted by the TLR9 ligand CpG and suppressed by potassium channel blockers


  • Potential conflict of interest: Nothing to report.


We previously reported that peripheral blood mononuclear cells (PBMCs) from patients with primary biliary cirrhosis (PBC) produce significantly higher levels of polyclonal IgM than controls after exposure to CpG. Furthermore, the prevalence and unusually high levels of antimitochondrial antibodies (AMAs) in patients with PBC suggest a profound loss of B cell tolerance. We have addressed the issue of whether CpG will promote the production of AMAs and whether new experimental agents that inhibit the lymphocyte potassium channels Kv1.3 and KCa3.1 can suppress CpG-mediated B cell activation and AMA production. PBMCs were stimulated with and without CpG and were subsequently analyzed for phenotype, including expression of TLR9, CD86, and KCa3.1 concurrent with measurements of AMA and responses to a control antigen, tetanus toxoid, in supernatants. Additionally, K+ channel expression on B cells from PBC patients and controls was studied using whole-cell patch-clamp technology. In patients with PBC, CpG induces secretion of AMAs in PBMCs and also up-regulates B cell expression of TLR9, CD86, and KCa3.1. Additionally, K+ channel blockers suppress secretion of AMA without a reduction of CpG-B–enhanced IgM production. Furthermore, there is diminished up-regulation of TLR9 and CD86 without affecting proliferation of B cells, B cell apoptosis, or viability. Conclusion: These data suggest that the hyperresponsiveness of B cells in PBC accelerates B cell–mediated autoimmunity. (HEPATOLOGY 2007;45:314–322.)

The serological hallmark of primary biliary cirrhosis (PBC) is the presence of antibodies to mitochondrially located antigens identified as the E2 subunits of the pyruvate dehydrogenase complex and related enzymes.1, 2 The degree to which anti–PDC-E2 is directly involved in the pathogenesis of PBC is uncertain, but the very high prevalence and levels of such antimitochondrial autoantibodies (AMAs) indicate a profound loss of B cell tolerance. One pathway that has been implicated in this loss of tolerance, hitherto identified in rheumatoid arthritis and systemic lupus erythematosus,3–5 is activation of the toll-like receptor TLR9 on autoreactive B cells by either bacterial CpG or CpG sequences in self DNA. Patients with PBC have high levels of serum IgM, and we recently demonstrated that peripheral blood mononuclear cells (PBMCs) from such patients produced significantly higher levels than controls of polyclonal IgM after exposure to CpG.6 In the present study, we addressed the important question of whether CpG promotes the production of AMA. In addition, the need for more effective treatments for PBC led us to explore a newly developed in vitro assay to determine whether experimental agents that inhibit the lymphocyte potassium channels Kv1.3 and KCa3.1 could suppress CpG-mediated B cell activation and AMA production.

The voltage-gated Kv1.3 and the calcium-activated KCa3.1, also called IKCa1 or SK4, have been proposed as attractive targets for immunomodulation because of their differential expression in human T and B cell subsets.7 Activated T cells of either the naïve or the central memory phenotype, as well as IgD+ B cells, predominantly express KCa3.1. In contrast, effector memory T cells and class-switched IgD-CD27+ memory B cells express high levels of Kv1.3.8, 9 Accordingly, K+ channel blockers specific for Kv1.3 or KCa3.1 differentially affect the function of T and B cell subsets and effectively suppress T cell mediated reactions, including delayed type hypersensitivity, inflammatory periodontal bone resorption and experimental autoimmune encephalomyelitis.10–13 However, it is still unknown whether these blockers influence the activation and differentiation of B cells, or autoantibody secretion from B cells. We report herein that in patients with PBC, AMA production, B cell activation, and up-regulation of TLR9 expression in CpG-B–stimulated B cells were all significantly suppressed by K+ channel blockers.


AMA, antimitochondrial antibody; CFSE, 5-carboxyfluorescein diacetate, succinimidyl ester; MFI, mean fluorescence intensity; OD, optical density; PBC, primary biliary cirrhosis; PBMC, peripheral blood mononuclear cell; PBS, phosphate-buffered saline; ShK, Stichodactyl helianthus.

Patients and Methods

Clinical Samples.

Peripheral blood was obtained from 52 female PBC patients (55.0 ± 9.5 yr [mean ± SD]) and 33 healthy female controls (49.9 ± 12.7 yr). The diagnosis of PBC was based on internationally accepted criteria.14 Patients without fibrosis at liver biopsy (i.e., those with stage I and II fibrosis according to Ludwig et al.15) were considered to have early-stage disease. Patients with stage III or IV fibrosis or cirrhosis, or those with a history of major complications from cirrhosis (e.g., ascites requiring diuretic therapy, gastrointestinal bleeding due to portal hypertension, hepatic encephalopathy, or HCC) were considered to have advanced disease. In our series, 25/52 (48%) patients with PBC were stage III or IV. Fifty of 52 patients were receiving ursodeoxycholic acid. All control sera were negative for AMA, whereas 94% of PBC patients were AMA-positive. The number of patients and controls studied for each of the variables is presented below. PBMCs were isolated by density gradient using Histopaque-1077 (Sigma Chemical Co., St. Louis, MO), washed, and resuspended in phosphate-buffered saline (PBS) (Mediatech Inc., Herndon, VA) containing 0.5% bovine serum albumin (Fraction V, OmniPur; EMD Chemicals Inc., Gibbstown, NJ) and 0.05% EDTA (Sigma Chemical Co.). Cells were counted via hemocytometry. Cell viability, always more than 98%, was assessed via trypan blue exclusion.

Culture of PBMCs.

We seeded 106 PBMCs in 48-well flat-bottom plates in 500 μL medium consisting of RPMI 1640 (Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum (GIBCO-Invitrogen Corp., Grand Island, NY), 100 μg/mL streptomycin, 100 U/mL penicillin (Invitrogen) with or without 2 μM CpG ODN, and cultured it for 4 days at 37°C in a 5% CO2 humidified atmosphere. Two types of CpG ODN were used: CpG-B (CpG ODN 2006; TCG TCG TTT TGT CGT TTT GTC GTT) and CpG-A (CpG ODN 2216; GGG GGA CGA TCG TCG GGG GG) (Invivogen, San Diego, CA). In tests for suppression of AMA synthesis, the Kv1.3 blocker Stichodactyl helianthus (ShK) toxin (Bachem Bioscience, PA) and/or the specific KCa3.1 inhibitor TRAM-3416 were added in optimal amounts (100 nM ShK and/or 1 μM TRAM-34) to the conditioned medium. For cell division determination, PBMCs were washed with PBS and labeled with 0.5 μM 5-carboxyfluorescein diacetate, succinimidyl ester (CFSE; Molecular Probes, Eugene, OR) for 15 minutes at 37°C. The cells were resuspended in prewarmed medium for 30 minutes and washed. CFSE-labeled PBMCs were then seeded at 1 × 106 cells/well with 2 μM CpG-B in the presence or absence of ShK and TRAM-34, and cultured for 4 days. Supernatants of the cell culture media were collected at day 4 and stored at −70°C.

AMA Reactivity, IgM Quantification, and ELISA for Tetanus Toxoid.

An ELISA for AMA was performed for both plasma and supernatants of cell culture media using our previously described triple hybrid recombinant molecule containing each of the autoepitopes of the E2 subunits of pyruvate dehydrogenase (PDC-E2), branched-chain 2-oxo-acid dehydrogenase (BCOADC-E2), and 2-oxo-glutarate dehydrogenase (OGDC-E2).17, 18 Briefly, purified recombinant MIT-3 at 10 μg/mL in carbonate buffer (pH 9.6) was coated onto 96-well ELISA plates at 4°C overnight, washed 5 times with PBS containing 0.05% Tween-20, and blocked with 3% skim milk in PBS for 30 minutes. The plates were then incubated with 100 μL of diluted plasma (1:250) or undiluted supernatants for 1 hour at room temperature and washed; 100 μL horseradish peroxidase–conjugated anti-human immunoglobulin (1:2,000) (Zymed, San Francisco, CA) was then added to each well for 1 hour at room temperature and washed. Immunoreactivity was detected by measuring the optical density (OD) at 405 nm after incubating for 15 minutes with 100 μL 40.0 mM 2,2 azinobis 3-ethylbenzthiazoline sulfonic acid containing 0.05 M hydrogen peroxide in citrate buffer. An OD value >3 SD above the mean for control plasma was considered positive. Antibody reactivity in cell culture supernatants was also examined via immunoblotting with known positive and negative controls used throughout. IgM levels in the supernatants were also examined using the Immuno-Tek human IgM ELISA Kit (ZeptoMetrix Corp., Buffalo, NY).

OD value of antibodies to tetanus toxoid in plasma (1:100 dilution) and supernatants (undiluted) derived from cultured cells of PBC and controls were measured using the Tetanus ELISA Kit (IBL-America Inc., Minneapolis, MN).

Flow Cytometry.

Four-day cultured PBMCs were resuspended in staining buffer (0.5% bovine serum albumin, 0.04% EDTA, 0.05% sodium azide in PBS), and distributed into 25-μL aliquots with subsequent preincubation with anti-human FcR blocking reagent (Miltenyi Biotech Inc.) for 15 minutes at 4°C. B cells were identified by staining with FITC or allophycocyanin-Cy7–conjugated anti-human CD19 (eBioscience, San Diego, CA). B cell activation and expression of costimulatory molecules were analyzed by staining with TRI-COLOR–conjugated CD86 (Caltag, Burlingame, CA), or mouse IgG1 (Caltag) as an isotype control for 15 minutes at 4°C. For TLR9 expression, CD19-stained cells were fixed and permeabilized with BD Cytofix/Cytoperm solution (BD Biosciences) for 15 minutes at 4°C and washed once. Subsequently, intracellular staining was performed with PE-conjugated anti-human TLR9 (eBioscience) or a rat IgG2a isotype control (eBioscience). After staining, the cells were washed and fixed with 1% paraformaldehyde in PBS. For analysis of B cell apoptosis, cells were stained with FITC-conjugated anti-human CD19 (eBioscience) and PE-conjugated Annexin V (BD Biosciences), and stained cells were counted on a FACScan flow cytometer (BD Immunocytometry Systems) that had been upgraded by Cytek Development (Fremont, CA) to allow for 5-color analysis. The acquired data were analyzed with Cellquest PRO software (BD Immunocytometry Systems). The values of mean fluorescence intensity (MFI) of TLR9 and CD86 obtained from analyzed data were compared in cells from PBC patients and healthy controls or within paired samples of PBC.


K+ channel expression on B cells from PBC patients or healthy controls was studied in the whole-cell mode of the patch-clamp technique9 with an EPC-10 HEKA amplifier 4 days after activation of PBMCs with CpG-B as described above. To identify B cells, PBMCs were stained with Alexa-488–conjugated anti-human CD19 monoclonal antibody (BD Pharmingen) on ice in complete RPMI, washed, put on poly-L-lysine–coated cover slips, and kept in the dark at 4°C for 10-30 minutes to attach. CD19+ cells were then visualized with fluorescence microscopy and patch-clamped with an aspartate-based pipette solution containing 1 μM free Ca2+ to activated KCa3.1 currents. The pipette solution contained 145 mM K+ aspartate, 2 mM MgCl2, 10 mM HEPES, 10 mM K2EGTA, and 8.5 mM CaCl2 (1 μM free Ca2+) (pH 7.2, 290 mOsm). To reduce chloride “leak” currents, we used a Na+ aspartate external solution containing 160 mM Na+ aspartate, 4.5 mM KCl, 2 CaCl2, 1 mM MgCl2, and 5 mM HEPES (pH 7.4, 300 mOsm). K+ currents were elicited with voltage ramps from −120 to 40 mV of 200-millisecond duration applied every 10 seconds. Whole-cell KCa3.1 conductances were calculated from the slope of KCa3.1 current at −80 mV, where the KCa3.1 currents are not “contaminated” by Kv1.3, which activates only at voltages above −40 mV. The KCa3.1 whole-cell conductance was then divided by the KCa3.1 single-channel conductance (11 pS) to determine the KCa3.1 channel number per cell. Cell capacitance, a direct measurement of cell surface area, was constantly monitored during recording and only cells with membrane capacitances >4 pF (diameter >11 μm) were studied to ensure that only clearly activated cells were analyzed.

Statistical Analysis.

Values for the percentage and absolute number of B cells; MFI of TLR9, CD86, and IgM; and OD of AMA in the presence of CpG-B (expressed as the mean ± SEM) in PBC were compared with data in the absence of CpG-B or CpG-A or data in the presence of CpG-A, CpG-B with ShK, CpG-B with TRAM-34, or CpG-B with both ShK and TRAM-34 using a paired, two-tailed Wilcoxon matched-pairs signed-rank test. MFI of TLR9 and CD86, OD of AMA, and level of KCa3.1 channels expression on B cells in PBC were compared with data for healthy controls using an unpaired, two-tailed Mann-Whitney U test. A P value of less than .05 was considered statistically significant.


CpG Stimulation Induces Secretion of AMA by PBMCs.

PBMCs from PBC patients and controls cultured with or without addition of 2 μM of CpG-A or CpG-B and 4-day supernatants were tested for secreted AMA via ELISA. Under all culture conditions, PBMCs from PBC patients (n = 14) but not controls (n = 8) secreted AMA, and stimulation with either CpG-A or CpG-B significantly enhanced the levels of secreted antibody from PBMCs of PBC compared with controls (Fig. 1A). The presence of AMA reactivity in the conditioned medium of stimulated PBMC from patients but not controls shown via ELISA was confirmed via immunoblotting (data not shown). There was a significantly greater induction of AMA secretion from PBC cells with CpG-B compared with CpG-A (Fig. 1A); thus, only CpG-B was used in subsequent experiments. A positive ELISA was defined as an OD value >3 SD above the mean for CpG-B–stimulated supernatants from 33 controls. Using these criteria, and at day 4 of culture, 34/52 (65.4%) CpG-B–stimulated PBC samples were AMA-positive, whereas none of the controls was positive (Fig. 1B). Responses to tetanus toxoid were used as a further control in supernatants with or without CpG-B stimulation; all subjects were up to date on tetanus toxoid vaccination. There was no significant difference in tetanus toxoid antibody in plasma of PBC versus controls (1.436 ± 0.292 vs. 0.819 ± 0.384), and, in contrast to AMA production, no significant increase in tetanus toxoid responses in either group was detected in supernatants following CpG-B stimulation.

Figure 1.

(A) Secretion of AMA in cultured PBMCs via ELISA on recombinant MIT3 of 4-day cultures of PBMCs in the presence or absence of CpG-A or CpG-B. OD readings for AMA in PBC (n = 14) were significantly higher than in controls (n = 8). AMA secretion was significantly enhanced by CpG, and more so by CpG-B than CpG-A. **P < 0.01. ***P < 0.001. Although OD values are enhanced by CpG-B stimulation in controls, there was no difference compared with CpG-A, and AMA-positive samples were only detected in PBC. (B) Using our rigorous definition of positive (3 SD above controls), supernatants of 34/52 (65.4%) PBC samples were AMA-positive following CpG-B stimulation compared with 0/33 controls.

In accordance with previous studies, 94% (49/52) of plasma samples from patients with PBC, but no control samples, were positive for AMA via ELISA. The mean OD values for early-stage PBC (stage I-II; n = 27) or late-stage PBC (stage III-IV; n = 25) were similar (1.357 ± 0.126 versus 1.304 ± 0.139). Of interest, levels of AMA in plasma as judged by OD values correlated with levels of AMA secreted from nonstimulated (P < 0.0001, r = 0.592, n = 52) (Fig. 2A) or CpG-B–stimulated (P < 0.05, r = 0.347, n = 52) (Fig. 2C) PBMCs in 4-day cultures; however, OD data of plasma compared with CpG-A stimulation data did not correlate (P = 0.0996, n = 14) (Fig. 2B). These data suggest that secretion of AMA by B cells in vitro reflects quantitative aspects of antibody production in vivo, with potential for use of the culture system as a model for AMA production.

Figure 2.

Correlation of ELISA readings for AMA in plasma with readings for supernatants in the presence or absence of CpG-A or CpG-B. Points indicate individuals with PBC [n = 52 in (A,C); n =14 in (B)]. Plasma AMA levels correlated with those of unstimulated medium and supernatant of CpG-B–stimulated PBMCs, but not with those after CpG-A stimulation.

CpG-B Up-regulates B Cell Expression of TLR9, CD86, and KCa3.1.

As previously reported, and confirmed herein, CpG-B up-regulates B cell activity. PBMCs from 21 PBC patients and 17 controls were cultured for 4 days without or with CpG-B, and stained for expression of B cell markers including TLR9, CD86, and CD19. Representative histograms of MFI are shown in Fig. 3A-B. The MFI of TLR9 and CD86 are summarized in Fig. 3C-D. The up-regulated expression of TLR9 and CD86 by CpG-B was significantly greater for cells from PBC patients compared with controls (P < 0.05, P < 0.01, respectively) (Fig. 3C,D); no statistical differences were found in unstimulated cells. In addition, B cells in 4-day cultures of PBMCs were stained with Alexa-488–fluorescence-conjugated CD19 and patch-clamped (Fig. 4A). Cells were dialyzed with a pipette solution containing 1 μM of free calcium, and K+ currents elicited with ramp pulses. Under such conditions, the current visible between −120 mV and −50 mV is solely carried by the calcium-dependent KCa3.1 channels, whereas the current above −40 mV is a combination of both KCa3.1 and Kv1.3. In the resting state, B cells among PBMCs expressed very few KCa3.1 channels (4≈10 per cell) with no difference for B cells from 6 PBC patients or 6 controls (Fig. 4B). However, for activated B cells from 4-day cultures of PBMCs stimulated with CpG-B, KCa3.1 channels were significantly up-regulated in B cells from 4 PBC patients (178.7 ± 18.2 channels per cell, n = 39 cells) compared with B cells from 6 controls (64.7 ± 9.9 channels per cell, n = 62 cells, P < 0.0001) (Fig. 4B). The significantly larger KCa3.1 current component observed in CpG-B–activated B cells from PBC patients was blocked by the specific inhibitor TRAM-34 with an EC50 of approximately 250 nM, confirming the molecular identity of the channel.

Figure 3.

Up-regulated expression of TLR9 and CD86 by CpG-B. (A,B) Representative histograms for expression of TLR9 and CD86 accompanied by MFI for TLR9 and CD86 analyzed via flow cytometry from CD19+ B cells of 4-day PBMC cultures. (C,D) MFI for TLR9 and CD86 were obtained from 21 patients with PBC and 17 controls. Values for cells stimulated with CpG-B were significantly higher in patients with PBC than in controls for both TLR9 and CD86. *P < 0.05. **P < 0.01.

Figure 4.

KCa3.1 expression is elevated on CpG-B–stimulated B cells from PBC patients compared with controls. (A) B cells stained with Alexa-488–conjugated antibody to CD19 were observed via fluorescence microscopy (top) and then patch-clamped (bottom). (B) KCa3.1 channel numbers per B cell were determined by dividing the KCa3.1 whole-cell conductance by the KCa3.1 single-channel conductance (11 pS). KCa3.1 expression by single B cells from 4-day CpG-B–stimulated PBMCs was significantly greater in PBC compared with controls, whereas there was no difference between unstimulated cells of PBC and controls. ***P < 0.001. (C) Representative K+ currents from CD19+ B cells from a control and a patient with PBC 4 days after activation with CpG-B. The current seen at a voltage below −40 mV is carried by KCa3.1, whereas the current at more positive voltages represents a combination of KCa3.1 and Kv1.3 as shown in (C). The significantly larger KCa3.1 current component in a B cell from a patient with PBC was blocked by 250 nM TRAM-34.

Taken together, these results suggest that the expression levels of KCa3.1 in B cells can be increased by CpG-B activation of the TLR9 signaling pathway, and that this pathway is more active in PBC patients than in healthy controls, thus facilitating B cell activation. Furthermore, blockade of KCa3.1 may interrupt this signaling pathway and thereby prevent the downstream secretion of autoantibody.

K+ Channel Blockers Suppress Secretion of AMA From CpG-Stimulated B Cells.

We determined whether K+ channel blockers affected AMA secretion by stimulating PBMCs from 27 PBC patients with CpG-B alone, or with CpG-B together with either 100 nM of the Kv1.3 blocker ShK, 1 μM of the KCa3.1 blocker TRAM-34, or a combination of both. According to the criteria described above, 17 of the 27 CpG-B–stimulated PBMCs from PBC patients were AMA-positive and were examined further. CpG-B–induced production of AMA was significantly decreased by ShK (P < 0.05) and TRAM-34 (P < 0.01) and by the combination of ShK and TRAM-34 (P < 0.001). The suppressive effect of the combination was significantly greater (P < 0.05) than that of TRAM-34 alone, and TRAM-34 alone was greater than that of ShK alone (P < 0.05) (Fig. 4A).

K+ Channel Blockers Suppress CpG-B–Induced Up-regulation of TLR9 and CD86.

We next examined the effects of K+ channel blockers on the up-regulation of TLR9 and CD86 using CpG-B–stimulated B cells from 12 AMA-positive PBC samples (see above). The expression of TLR9 and CD86 on B cells was analyzed via flow cytometry, and expression levels were compared between stimulated and unstimulated cells. The up-regulated expression of TLR9 on B cells was significantly suppressed by TRAM-34 alone (P < 0.01) or TRAM-34 combined with ShK (P < 0.01), but not by ShK alone (Fig. 5B). The differences between the effects of TRAM-34 alone and those of ShK and TRAM-34 combination were significant (P < 0.05). Similar results were obtained for expression of CD86 in that the MFI for CpG-B–stimulated B cells was significantly decreased by TRAM-34 alone (P < 0.01) or by the combination of the 2 blockers (P < 0.001), but not by ShK alone (Fig. 5C). The combination of ShK and TRAM-34 gave no greater suppression of CD86 expression on CpG-B–stimulated B cells than did TRAM-34 alone.

Figure 5.

Suppression of AMA secretion, TLR9, and CD86 expression by K+ channel blockers (AMA, n = 17; TLR9 and CD86, n = 12). (A) Suppression of AMA was examined in the presence of ShK, TRAM-34, or a combination of the two. To assess CpG-B–stimulated enhancement of AMA secretion, ΔOD of AMA [OD of AMA with blocker (CpG-B with or without K+ channel blockers) − OD of AMA (medium alone)] were compared. AMA secretion from B cells stimulated by CpG-B in 4-day cultures of PBMCs was suppressed by ShK (P < 0.05), TRAM-34 (P < 0.01), and the combination of the two (P < 0.001). The difference in OD of AMA was also significant between CpG-B plus TRAM-34 and CpG-B plus the combination (P < 0.05), as well as between CpG-B plus ShK and CpG-B plus TRAM-34 (P < 0.05). (B) In AMA-positive samples of 4-day cultivated PBMCs from PBC, the up-regulated expression of TLR9 by CpG-B was suppressed by TRAM-34 or the combination. (C) In the same PBC samples as in (B), the up-regulated expression of CD86 was also suppressed by TRAM-34 or the combination of both blockers. *P < 0.05. **P < 0.01. ***P < 0.001.

K+ Channel Blockers Do Not Diminish CpG-B–Induced IgM Production From B Cells.

To examine whether K+ channel blockers influence AMA production from CpG-B–stimulated PBMCs by reducing total IgM production from B cells, levels of IgM in the supernatants under each condition from 17 AMA+ samples after 4-day PBMC culture were compared. In contrast to AMA suppression, CpG-B–induced IgM production was not suppressed by these K+ channel blockers (Table 1).

Table 1. IgM Production Induced by CpG-B in the Presence or Absence of K+ Channel Inhibitors
 IgM (ng/ml)P Value (vs. Control)
  1. NOTE. All values are expressed as the mean ± SEM (n = 17).

  2. Abbreviations: NA, not applicable; NS, not significant.

Control678.1 ± 98.9NA
ShK698.3 ± 101.0NS
TRAM-34645.1 ± 93.8NS
ShK + TRAM-34698.7 ± 108.2NS

K+ Channel Blockers Do Not Affect CpG-B–Induced Proliferation of B Cells.

To examine whether the K+ channel blockers inhibited production of antibody from CpG-B–stimulated PBMCs via suppression of B cell proliferation, stimulated cells were examined for the frequency of B cells with amplified CD19+ expression based on an attenuated CFSE signal. Cultured PBMCs from 12 PBC patients were stimulated with CpG-B in the absence or presence of the K+ channel blockers ShK and TRAM-34, and, after 4 days, the frequency of CFSElow CD19+ B cells was determined via flow cytometry. The lack of any significant difference in the frequency of amplified B cells, with or without the blockers (Table 2), suggests that B cell proliferation in CpG-B–stimulated PBMCs in culture is unaffected by ShK or TRAM-34.

Table 2. The Frequency of CFSElow CD19+ B Cells in CpG-B–Stimulated PBMCs
 Frequency of CFSElow B CellsP Value (vs. Control)
  1. NOTE. All values are expressed as the mean ± SEM.

  2. Abbreviations: NA, not applicable; NS, not significant.

Control79.00 ± 2.24NA
ShK80.15 ± 2.29NS
TRAM-3479.39 ± 2.11NS
ShK + TRAM-3477.27 ± 2.41NS

K+ Channel Blockers Neither Influence B Cell Apoptosis nor Reduce Viability of CpG-B–Stimulated PBMCs.

To examine whether K+ channel blockers inhibited production of AMA from CpG-B–stimulated PBMCs by enhancing apoptosis, PBMCs from 5 PBC patients were seeded in tissue culture plates at 1 × 106 cells/well and stimulated for 4 days with CpG-B in the absence or presence of ShK and TRAM-34. B cell viability was defined via annexin V negativity and CD19 positivity using relevant antibodies and flow cytometry as described above. In addition, the absolute number of viable cells in each well was determined with a hemocytometer using trypan blue dye exclusion to calculate absolute numbers of annexin V–negative viable B cells. There were no significant differences with or without the blocking agents in either the absolute number of viable B cells per well or the frequency of annexin V–negative B cells (Table 3), suggesting that the differences in AMA production from CpG-B–stimulated PBMCs in the presence of K+ channel inhibitors was not attributable to differences in the number of viable B cells.

Table 3. Frequency of CD19+ B Cells in Annexin V–Negative Cells and Absolute Number of Viable CD19+ B Cells in CpG-B–Stimulated PBMCs
 Frequency of B CellsViable B Cells (×104)P Value (vs. Control)
  1. NOTE. All values are expressed as the mean ± SEM.

  2. Abbreviations: NA, not applicable; NS, not significant.

Control44.27 ± 4.449.94 ± 4.16NA
ShK45.65 ± 5.258.76 ± 2.44NS
TRAM-3444.78 ± 3.8310.84 ± 3.78NS
ShK + TRAM-3451.42 ± 3.839.72 ± 2.56NS


PBC is characterized by a high prevalence (90%-95%) of autoantibodies to the mitochondrially located E2 subunit of PDC and related enzymes (AMA). A possible role of AMA in disease is suggested by data of secretion into bile ducts via transcytosis, including enhanced susceptibility of bile duct epithelial cells to undergo apoptosis.19 In addition, as proposed for other autoimmune diseases,20 B lymphocytes expressing anti–PDC-E2 as a B cell receptor can potentially function as efficient antigen-capturing and antigen-presenting cells to perpetuate autoimmune reactivity.

Here, we demonstrate that stimulation with CpG-B of PBMCs increases TLR9, CD86, and KCa3.1 expression on B cells from PBC patients compared with healthy age and sex-matched controls, with greatly enhanced AMA production. We further demonstrate that the effects of CpG-B stimulation can be partly reversed by blockers of the lymphocyte potassium channels KCa3.1 and Kv1.3. Taken together with our previous study on the induction of IgM by CpG-B in PBC, these new results demonstrate that CpG motifs activate B cells and induce both IgM and AMA production in PBMCs from PBC patients.

In parallel with increasing expression of the receptor TLR9 and of CD86, stimulation of PBMCs with CpG-B in patients with PBC resulted in a significant up-regulation of KCa3.1. Expression of this channel had been reported to be mediated by AP1, an activation protein that is activated downstream of TLR9 and which in conjunction with nuclear factor κB has been reported to increase immunoglobulin transcription. Furthermore, together with Ikaros-2, AP-1 also binds to the promoter of KCa3.1 and initiates new transcription of the channel gene.16 Increased expression of KCa3.1 thus seems to be part of the CpG motif-mediated activation program in which KCa3.1 would facilitate Ca2+ influx by providing the counterbalancing K+ efflux.11, 21 The fact that activation with CpG-B leads to a stronger induction of KCa3.1 in PBC patients than in controls argues that the TLR9 pathway in B cells might be “hyperactive” or constitutively activated in PBC patients, resulting in a stronger B cell activation than in healthy individuals.

Bacterial infection in various settings has been repeatedly invoked in the etiopathogenesis of PBC.22–27 This is usually linked to the concept of molecular (epitope) mimicry, a resilient explanation for the occurrence of autoimmunity, despite few well-proven examples.28, 29 However, our recent studies provide an alternative explanation, because microbial CpG enhances IgM production in PBMC cultures,6 and CD27+ memory B cells in PBC patients are responsible for this IgM production through TLR9 signaling.6 Thus, chronic bacterial exposure in PBC could increase responsiveness to CpG motifs and contribute to the characteristic hyper-IgM production.30 It is interesting to note the contrast between the anti-PDC and anti–tetanus toxoid antibody response. We suspect the major difference is the sensitivity of autoantigen-specific B cells for CpG-B polyclonal activation. Furthermore, it is likely that tetanus toxoid–specific antibodies are secreted by plasma cells that are not present in the PBMC cultures. These data would further point toward a B cell dysregulation in PBC. Future studies should be directed at this and similar issues dealing with the specificity of the response.

Human B cells do not express TLR4, and the ability of lipopolysaccharides to induce IgM production in human B cells is limited (unpublished data). Therefore, for the work herein, we selected CpG motifs for B cell stimulation.31 In the present study, CpG-B—in contrast to CpG-A—stimulated strong AMA secretion by B cells among PBMC cultures from PBC patients (Fig. 1A). The frequency of AMA-positive samples was 65.4% in CpG-B–stimulated PBMC cultures and 9.6% of unstimulated cultures based on our rigorous use of 3 SDs above the mean. We emphasize that our work is not proposing a specific bacterial etiology of PBC, but rather a role for bacterial CpG as a modulator of disease. In fact, we submit the possibility that gut bacterial flora may be the principal determinant that leads to the continued development of AMAs and the IgM response.

Evidence for the involvement of TLR9 in the increased AMA production by cultured B cells after CpG-B stimulation is reflected by up-regulation of TLR9 and CD86 in PBC that significantly exceeds data for controls (Fig. 3). This increase in TLR9 expression could result in enhanced signaling after increased CpG-B ligation, while the accompanying CD86 up-regulation would provide additional costimulatory signaling by ligation with high-affinity CD28 or low-affinity CD152 receptors on T cells.32 Our previous work also demonstrated that B cells need non–B cell help for IgM production.6 Accordingly, these data suggest that in PBC, hyperresponsiveness of B cells to bacterial CpG motif could potentially accelerate B cell autoimmunity.

The specific mechanism of the potassium channels Kv1.3 and KCa3.1 in B cell activation and signaling has not been studied in detail. However, similar to T cells both channels are likely involved in regulating the intracellular Ca2+ concentration.7 We have demonstrated that CpG-B had a greater up-regulatory effect on KCa3.1 expression on B cells from PBC patients than on B cells from controls (Fig. 4B), similar to the up-regulated expression of TLR 9 and CD86. This is consistent with KCa3.1 channel involvement in B cell activation and autoantibody production by CpG-B–stimulated B cells. Moreover, we reported that the specific KCa3.1 blocker TRAM-34 (Fig. 4C) could significantly suppress AMA secretion in CpG-B–stimulated PBMC cultures (Fig. 5A). Although ShK, a potent Kv1.3 blocker, was effective in suppressing AMA production, TRAM-34 alone and the combination of both TRAM-34 and ShK demonstrated greater effectiveness in suppression than ShK alone. TRAM-34 and the TRAM-34/ShK combination also inhibited the up-regulation of TLR9 (Fig. 5B), suggesting diminished TLR9 signaling could be one of the mechanisms for decreased secretion of AMA in CpG-B–stimulated PBMC cultures after exposure to K+ channel blockers. Additionally, the inhibition of CD86 up-regulation on CpG-B–stimulated B cells by the blockers (Fig. 5C) may decrease T cell–B cell interaction through CD86 ligation by CD28 and CD152 on T cells, resulting in the suppression of autoantibody production.32

The suppression of AMA production by K+ channel inhibitors did not depend on a reduction of IgM production, a decrease on the proliferative capability of autoreactive B cells, or an increase in the degree of apoptotic B cells. Thus, TRAM-34 or the combination of blockers neither decreased CpG-B–stimulated B cell proliferation as shown by CFSE incorporation nor diminished the absolute number of viable cells as judged by annexin V staining and by trypan blue exclusion for both B cells and CD3+ T cells. These data also suggest that 100 nM ShK and/or 1 μM TRAM-34 did not promote B cell or T cell apoptosis and that the concentration of these blockers is not cytotoxic in 4-day cultured PBMCs. Total IgM levels in the supernatants were unaffected by the K+ channel inhibitors, suggesting that these inhibitors were not broad active suppressants of all B cell immunoglobulin production.

There is a great need for more effective agents in the treatment of PBC.33 Both B cell–T cell interactions and the antigen-presenting function of autoreactive B cells are important in PBC; autoreactive CD4+ and CD8+ T and B cells to PDC-E2 are evident at sites of liver damage in PBC.34–39 Hence, we re-emphasize our findings that the KCa3.1 blocker TRAM-34, particularly in combination with Kv1.3 channel blocker ShK, is capable of suppressing autoantibody production as well as reducing the expression of costimulatory molecules in CpG-B–stimulated B cells. Thus, if TRAM-34 in combination with ShK were to prove effective in suppressing both B cell autoantibody production and T cell activation, these blockers could have therapeutic value in PBC.


We thank Patrick S. C. Leung and Roman Rieger for their technical support on the ELISAs and to Marcy Crees for organizing patient meetings for the sample collections.