Potential conflict of interest: Nothing to report.
The CD154–CD40 interaction is a critical costimulatory pathway modulating the cellular immune response. Moreover, fulminant hepatitis of various etiologies is characterized by a hepatic influx of CD154-expressing T cells and an upregulation of CD40 expression on Kupffer cells and hepatocytes, implicating this pathway in the pathogenesis of fulminant hepatitis. In this study, we used a murine model of fulminant hepatitis induced by concanavalin A (con A) and documented a significant influx of CD154-expressing T cells into the livers of mice treated with con A, in association with markedly increased expression of CD40 restricted mainly to hepatocytes in damaged areas of the liver. Furthermore, con A hepatitis in CD154-deficient mice was significantly attenuated compared with that in wild-type controls and was associated with a decrease in hepatic tumor necrosis factor α (TNF-α) levels and hepatocyte death. We next determined the role of the CD154–CD40 pathway in hepatocyte death in vitro. These in vitro studies demonstrated that TNF-α induces CD40 expression in hepatocytes and that subsequent activation of CD40 results in hepatocyte apoptosis mediated at least in part by enhanced hepatocyte expression of FasL. In conclusion, CD154 stimulation of CD40 plays a central role in hepatocyte death in fulminant hepatitis through direct and indirect pathways that may have direct therapeutic implications in humans. Supplementary material for this article can be found on the HEPATOLOGY website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). (HEPATOLOGY 2005;42:372–380.)
Fulminant hepatitis (FH) is a devastating liver disease associated with an extremely high mortality rate.1–3 It is unclear why some patients develop FH while others exposed to the same hepatic insult do not. However, it has been suggested that uncontrolled systemic immune activation coupled with aberrant antigen presentation and T-cell activation play crucial roles in driving FH.4 Efficient T-cell activation requires recognition of two signals, including T-cell receptor (TCR) activation upon recognition of major histocompatibility complex (MHC)-peptide complex, and costimulatory signals.5 The CD154–CD40 system is an important costimulatory pathway.6, 7 Cumulative studies have revealed the critical role of this system in immunoinflammatory disorders.8–10 Moreover, a marked upregulation of CD40 has been observed in the livers of patients with FH coupled with an influx of CD154-expressing lymphocytes.11 Therefore, it is very likely that the CD154–CD40 pathway is involved in the pathogenesis of FH.
The study of FH in humans is hampered by the lack of available clinical specimens as well as the difficulty in defining the critical early events in FH.2 Therefore, the study of early events that may drive FH requires the use of animal models. One animal model that closely mimics FH is acute murine hepatitis caused by the intravenous injection of concanavalin A (con A).12–15 Specifically, intravenous administration of con A in mice mimics FH in humans in the following ways: (1) systemic immune activation and recruitment of activated T cells to the liver and stimulation of hepatic antigen-presenting cells (APCs) (i.e., Kupffer cells);12, 16, 17 (2) acute dose-dependent liver injury, which ranges in spectrum from mild hepatitis to liver failure;2, 3 (3) hepatocyte death mainly due to apoptosis and driven by both FasL/Fas and tumor necrosis factor α (TNF-α) pathways;18–20 (4) a pivotal role of the cytokines TNF-α and interferon γ (IFN-γ) in the development of hepatitis;14, 21, 22 and (5) upregulation of hepatic C-C chemokine expression.23, 24
Therefore, given the central role played by CD154–CD40 interactions in regulating the cellular immune response and the observation of an upregulation of CD40 expression and infiltration of CD154-expressing lymphocytes in the livers of patients with FH, we investigated the potential role of CD154 in the development of con A–induced hepatitis.
Male (aged 10-12 weeks) wild-type (WT) and CD154 knockout (KO) mice (all on C57BL/6 background) were purchased from the Jackson Laboratory (Bar Harbor, ME). Mice were group-housed under controlled temperature (22°C) and photoperiod (12:12-hour light/dark cycle) and allowed unrestricted access to standard mouse chow and tap water. Protocols were approved by the University of Calgary Animal Care Committee, and all procedures were conducted in accordance with the regulations of the Canadian Council on Animal Care.
Anti-mouse FcγR III/II monoclonal antibody (mAb) (2.4G2), fluorochrome-conjugated anti-mouse NK1.1 mAb (PK136), anti-mouse CD3e mAb (145-2C11), anti-mouse CD4 mAb, anti-mouse CD40 (3/23), anti-mouse CD154 mAb (MR1), anti-mouse CD69 mAb (H1.2F3), anti-mouse FasL mAb (MFL3), and anti-mouse Fas mAb (Jo2) as well as isotype-matched control antibodies were obtained from BD Pharmingen (San Diego, CA). Rat anti-mouse CD40 mAb (3/23) and biotinylated goat anti-rat immunoglobulin G were also obtained from BD Pharmingen. Rabbit anti-mouse Fas antibody and biotinylated anti-rabbit immunoglobulin G were obtained from Calbiochem (San Diego, CA), and rabbit anti-mouse FasL antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA).
Generation of Con A–Induced Hepatitis.
Con A (Sigma, St. Louis, MO) or phosphate-buffered saline (PBS) was injected intravenously at a dose of 13.5 mg/kg (this dose was chosen to give severe hepatitis without mouse mortality).23 Mice were sacrificed 2 and 8 hours following con A administration. Extent of hepatitis was determined biochemically (ALT kit, Sigma) and histologically (hematoxylin-eosin staining).
Cytokine Determination via ELISA.
Hepatic TNF-α, interleukin (IL)-4, IFN-γ, and IL-12 levels were determined via ELISA as previously described23, 25 and results expressed pg/mg protein (Bio-Rad, Mississaugua, ON, Canada).
Isolation of Hepatic Lymphocytes, Macrophages, and Natural Killer T Cells.
The isolation of hepatic mononuclear cells was performed as previously described.23 The number of CD4+ T and natural killer T (NKT) cells were determined via staining with flurorescence-conjugated anti-CD4 mAb and anti-NK1.1 mAb followed by fluorescence-activated cell sorting (FACS). Isolated hepatic mononuclear cells were also used for staining of CD154 expression. Hepatic neutrophil infiltration was determined via esterase staining of liver sections and the number of neutrophils counted as previously described.26
Enrichment of Splenic CD4+ T Cells.
Spleen cells were isolated using routine methods.27 Splenic CD4+ T cells were enriched via magnetic cell sorting (autoMACS; Miltenyi Biotec, Auburn, CA) using a commercial mouse CD4+ T-cell–negative selection kit (Miltenyi Biotec). Purified CD4+ T cells or splenocytes were cultured in the presence of con A (10 μg/mL [Sigma]) at different time points as described for determination of CD154 and CD69 expression via FACS.
Immunohistochemical Analysis of Hepatic CD40 Expression.
Hepatic CD40 expression was determined via an indirect immunoperoxidase technique on cryostat liver sections as previously described28, 29 using primary antibody (rat anti-mouse CD40 mAb, clone 3/23 [BD Pharmingen]), biotin-conjugated goat anti-rat immunoglobulin polyclonal antibody (BD Pharmingen) and strepavidin-horseradish peroxidase (BD Pharmingen). Expression of Fas or FasL in liver tissues was also determined with the same method using paraffin-embedded liver sections.
Preparation of Mouse Hepatocytes.
Mouse hepatocytes were prepared via a two-step in situ perfusion method using a commercial kit from Invitrogen/Gibco (Burlington, ON, Canada) as previously described.30 The number and viability of isolated cells were determined via trypan blue exclusion.
Determination of the Effect of TNF-α on CD40 Expression on Hepatocytes.
Following isolation, hepatocytes were seeded in collagen I–coated six-well plates and cultured in WE medium (Invitrogen/Gibco) containing 10% fetal bovine serum for 3 hours before replenishing medium, followed by 24 hours culture for recovery before treatment with 10 ng/mL recombinant murine TNF-α (rmTNF-α) (R&D Systems, Minneapolis, MN). After 12 hours, cells were collected and stained with phycoerythrin-conjugated anti-CD40 mAb followed by FACS analysis as detailed below.
Evaluation of Functional Consequences of CD40 Activation on Hepatocytes In Vitro.
After isolation, hepatocytes were seeded in 6-well plates and 8-well chamberslides. Cultured cells were pretreated with rmTNF-α (10 ng/mL) in the presence or absence of 50 ng/mL anti-FasL antibody31 for 12 hours followed by stimulation with 1 μg/mL anti-CD40 antibody (BD Pharmingen) for a further 4 hours. Control groups included cells with no stimulation, cells treated with TNF-α (10 ng/mL × 16 hr) alone, or cells treated with agonistic anti-CD40 antibody (1 μg/mL × 4 hr) alone. The expression of Fas/FasL on hepatocytes was determined via FACS. Hepatocyte apoptosis was determined via immunofluorescent staining with FITC-conjugated Annexin V based on the supplied protocol (BD Pharmingen).
Evaluation of Functional Consequences of CD40 Activation in Con A–Treated CD154 KO Mice In Vivo.
To determine if the attenuating effect of CD154 deficiency on con A hepatitis could be reversed by activating CD40, we administered agonistic anti-CD40 antibody (200 μg/mouse or isotype control mAb [BD Pharmingen]) to con A–treated CD154 KO mice and measured serum alanine aminotransferase (ALT) levels 8 hours later.
Flow Cytometry Analysis.
To analyze the expression of CD154 and CD69 on CD4+ T cells, and CD40 or Fas/FasL on hepatocytes, FACS analysis was performed as previously described.23, 32, 33
All data are expressed as the mean ± SEM. Comparisons between two experimental groups of data were performed using an unpaired Student t test. Comparison among three or more experimental groups of data was performed using one-way ANOVA. Post hoc comparisons were performed using Bonfferonni multiple comparison tests. A P value of less than .05 was considered significant.
CD154 and Con A–Induced Liver Damage.
Con A–treated WT mice displayed severe hepatic necrosis and significant infiltration of inflammatory cells 8 hours after con A injection. In contrast, CD154 KO mice exhibited significantly reduced liver damage (Supplementary Fig. 1). Consistent with histological findings, serum ALT levels were significantly lower in CD154 KO mice 8 hours after con A treatment compared with WT mice. Serum ALT levels were not significantly elevated in WT or KO mice 2 hours after con A administration (Fig. 1).
CD154 and Intrahepatic Cytokine Levels in Con A Hepatitis.
Con A–induced liver damage is mainly driven by the infiltration of CD4+ T cells and activation of hepatic NKT cells and macrophages, as well as the specific cytokines produced by these cells—mainly TNF-α, IL-4, IFN-γ, and IL-12.12, 13, 34 Therefore, we determined the hepatic levels of these four cytokines in con A–treated mice. Con A treatment resulted in significant increases in hepatic levels of all four cytokines in WT mice 8 hours after con A injection. However, hepatic levels of TNF-α and IL-4 were significantly lower in con A–treated CD154 KO mice compared with WT mice at this time point (P < .05) (Supplementary Fig. 2). In contrast, hepatic IFN-γ and IL-12 levels were similar in WT and KO mice 8 hours after con A administration (IFN-γ: WT-PBS 4.1 ± 1.2 and KO-PBS 3.8 ± 1.0 pg/mg protein vs. WT-con A 8.6 ± 1.8 and KO-con A 10.5 ± 3.4 pg/mg protein [P < .05 vs. respective controls]; IL-12: WT-PBS 3.1 ± 0.7 and KO-PBS 3.2 ± 0.7 pg/mg protein vs. WT-con A 9.5 ± 2.5 vs. KO-con A 8.5 ± 3.8 pg/mg protein [P < .05 vs. respective controls]) (n = 6 mice per group). Furthermore, lower hepatic TNF-α and IL-4 levels were also observed in con A–treated CD154 KO mice compared with con A–treated WT controls 2 hours after con A treatment (Supplementary Fig. 2).
CD154 Expression on CD4+ T Cells Is Upregulated by Con A.
There was no expression of CD154 on hepatic CD4+ T cells isolated from naive mice (Fig. 2). However, CD154 expression on isolated hepatic CD4+ T cells was significantly increased 2 and 4 hours after con A administration but had returned to baseline 8 hours afterward (Fig. 2A). Consistent with our in vivo observations, con A stimulation of isolated splenic CD4+ T cells in vitro resulted in a nearly four-fold increase of CD154 expression compared with unstimulated controls (Fig. 2B). To determine whether CD154 expression was restricted only to hepatic CD4+ T cells after con A treatment, we isolated hepatic NKT cells and macrophages from naive and con A–treated WT mice. We found that con A induced a significant increase in CD154 expression on both hepatic CD4+ NKT cells and macrophages (Fig. 2C-D). In addition, there was a modest but significant decrease in CD69 expression on CD4+ T cells isolated from CD154 KO mice compared with WT mice after con A treatment (Supplementary Fig. 3).
Effects of CD154 Deficiency on Hepatic Cellular Infiltrate After Con A Treatment.
Con A–mediated hepatitis is driven by CD4+ T cells, neutrophils, and NKT cells within the liver, whereas hepatic NK and CD8+ T-cell numbers do not change after con A administration,35 and these two types of cells do not appear to be pathogenic in this model.13, 33 Therefore, we determined the effect of CD154 deficiency on the hepatic cellular infiltrate in con A hepatitis. CD154 KO mice demonstrated similar numbers of hepatic neutrophils and NKT cells after con A administration (Supplementary Fig. 4A-B); however, a small but significant reduction in hepatic CD4+ T-cell infiltration was noted (Supplementary Fig. 4C).
In Situ Hepatic CD40 Expression by Immunohistochemistry in Con A–Treated Mice.
CD40 expression was not observed in naive mice. By 2 hours after con A treatment, CD40-expressing hepatocytes were observed in WT but not CD154 KO mice (Fig. 3A,B). However, by 8 hours after con A treatment, strong CD40 expression was documented on hepatocytes in WT mice (restricted mainly to damaged areas of liver), but only weak CD40 expression was noted in CD154 KO mice (Fig. 3C-D).
CD40 Expression on Isolated Primary Hepatocytes Is Upregulated by TNF-α.
To link the enhanced expression of CD40 on hepatocytes with the increased intrahepatic levels of TNF-α in con A–treated WT mice, we next determined the effect of rmTNF-α on modulating CD40 expression on hepatocytes in vitro. As shown in Fig. 4, rmTNF-α treatment resulted in a two-fold increase of CD40 expression on hepatocytes compared with unstimulated controls. In contrast, recombinant murine IL-1β did not alter hepatocyte CD40 expression (data not shown).
Functional Consequences of CD40 Activation on Hepatocytes.
To determine the potential effects of CD40 activation on hepatocytes, further in vitro experiments were conducted. CD40 activation with an agonistic CD40 antibody in TNF-α–treated hepatocytes in vitro significantly increased hepatocyte Fas/FasL expression (Fig. 5). In addition, CD40 activation also increased the number of apoptotic hepatocytes, as determined by Annexin V staining. Moreover, hepatocyte apoptosis induced by CD40 activation was significantly reduced by pretreatment of cells with a neutralizing anti-FasL antibody (Fig. 6). No significant increase in the number of apoptotic cells was observed in hepatocytes treated with TNF-α or anti-CD40 antibody alone. We confirmed these in vitro results, demonstrating an important effect of CD40 activation on hepatocytes upon subsequent hepatocyte death in vivo. Specifically, con A–treated CD154 KO mice demonstrated significantly lower serum ALT levels 8 hours after con A treatment compared with con A–treated WT mice (Fig. 7). Treatment of CD154 KO mice with an agonistic CD40 mAb before con A treatment completely abrogated the effects of CD154 deficiency with regard to con A hepatitis (Fig. 7).
To determine whether there are similar changes in hepatocyte Fas/FasL expression in con A–treated mice in vivo, we determined the expression of Fas/FasL by immunohistochemistry. Hepatic expression of Fas/FasL in con A–treated WT mice were strikingly increased compared with CD154 knockout mice. As shown in Fig. 8, strongly positive staining hepatocytes for Fas/FasL were observed in damaged areas of liver in WT mice 8 hours after con A treatment, whereas there were only a few positively stained cells in the livers of con A–treated CD154 KO mice.
In this study, our results demonstrate that the CD154–CD40 pathway plays a major role in the development of con A–induced hepatitis. Specifically, CD154 deficiency significantly attenuates con A hepatitis—an effect completely abrogated by the exogenous administration of an agonistic CD40 antibody.
It is well established that con A–induced hepatitis is driven mainly by the infiltration of CD4+ T cells into the liver that follows hepatic neutrophil infiltration,26 as well as the activation of hepatic NKT cells and macrophages.12–14, 36 Moreover, specific cytokines produced by these cells (i.e., IFN-γ by CD4+ T cells, IL-4 by NKT cells, and IL-12 and TNF-α by macrophages) are critical in the development of con A hepatitis.12–14 We presumed that the lack of CD154 activation of CD40 expressed on hepatic macrophages would result in a reduction of con A–mediated hepatitis by decreasing macrophage production of TNF-α and IL-12. Our results clearly show that con A–induced liver damage in CD154 KO mice is significantly reduced and is accompanied by lower hepatic levels of TNF-α, but not IL-12. Given that the ligation of CD154 with CD40 on macrophages is important for the production of TNF-α,37 and that there is documented clustering of Kupffer cells and T cells in the liver during con A hepatitis,12 and that CD154-expressing T cells infiltrate the liver in con A hepatitis, our results are consistent with a major role of CD154–CD40 interaction–driven Kupffer cell–related TNF-α secretion in con A hepatitis. However, our findings suggest that hepatic IL-12 production during con A hepatitis appears to be CD154–CD40-independent. The observation of similar hepatic IL-12 levels in con A–treated CD154 KO and WT mice is consistent with our observation of similar hepatic IFN-γ levels in these mice.38 Moreover, these findings are consistent with a major role of TNF-α release within the liver, driven by CD154–CD40 interactions, during con A hepatitis. Interestingly, we also documented that hepatic NKT cells and macrophages themselves express CD154 in con A hepatitis, implicating these cell types in the activation of CD40 on antigen-presenting cells such as macrophages in this model, with resultant TNF-α secretion. Macrophage expression of CD154 has previously been identified as an important driver in liver cell death in the setting of liver transplant rejection in humans.31 Hepatic NKT cells are the main source of IL-4 in con A hepatitis,33, 39 and we observed lower hepatic IL-4 levels (proinflammatory in the liver33, 36) in con A–treated CD154 KO versus WT mice, suggesting decreased hepatic NKT cell production of IL-4 in con A hepatitis in the KO mice. Given that we could not document CD40 expression on hepatic NKT cells (data not shown), this finding is of interest and warrants further investigation. Our preliminary studies in this regard have indicated that TNF-α can directly stimulate NKT cell IL-4 production (data not shown).
CD154 is also known to play an important role in cellular recruitment (mainly CD4+ T cells) to inflammatory tissues.40, 41 Therefore, decreased con A–induced hepatitis in CD154 KO mice might also be due to decreased hepatic cellular influx of cell types important in the development of con A hepatitis. CD4+ T cells, neutrophils, and NKT cells are important in this regard.12, 13, 26, 36 CD8+ T cells and NK cells do not appear to play a pathogenic role in this model.13, 33, 35 We documented similar hepatic neutrophil and NKT cell numbers after con A treatment in CD154 KO and WT mice (Supplementary Fig. 4). However, CD154 deficiency was associated with a modest but significant reduction in hepatic CD4+ T-cell recruitment after con A (Supplementary Fig. 4C). This finding suggests that CD154–CD40 interactions may play a role in CD4+ T-cell recruitment in con A hepatitis and that decreased CD4+ T-cell recruitment may contribute, at least in part, to the reduced severity of con A hepatitis in CD154 KO mice. In addition, CD4+ T cells isolated from CD154 KO mice demonstrate a modest defect with regard to con A–induced activation (as reflected by CD69 expression). Although this defect in activation might contribute to decreased hepatitis in con A–treated CD154 KO mice, we feel that this is unlikely given that CD40 stimulation by an agonistic antibody in CD154 KO mice in vivo completely abrogates the effect of CD154 deficiency in con A hepatitis (Fig. 7).
It is well documented that TNF-α plays an essential role in the development of con A–induced hepatitis.13, 14, 42 However, how TNF-α kills hepatocytes remains largely unknown. Our current results demonstrate that rmTNF-α significantly upregulates the expression of CD40 on isolated hepatocytes. Interestingly, the induction of CD40 expression by TNF-α has also been found in other cell types.43–45 Moreover, Afford and colleagues31 reported that enhanced expression of CD40 was observed in human hepatocytes in chronic rejection after liver transplantation, and activation of CD40 was able to induce hepatocyte apoptosis via the Fas pathway. In addition, enhanced CD40 expression has been detected on hepatocytes from patients with fulminant hepatitis,11 and there is increasing evidence indicating the importance of hepatocyte apoptosis induced by the Fas/FasL pathway in the development of con A–induced hepatitis, as well as in fulminant hepatitis.14, 15, 34, 46 Our current data show that CD40 expression is markedly increased on hepatocytes 8 hours after con A treatment, but is present as early as 2 hours after con A administration. In addition, CD40 activation on hepatocytes in vitro results in apoptosis of hepatocytes accompanied by a marked increase of FasL/ Fas expression on hepatocytes. Moreover, blocking Fas/FasL interactions during CD40 activation of hepatocytes partially abrogated apoptosis of these cells. This finding suggests that there may be another non–FasL-driven pathway involved in the CD40-mediated apoptosis of hepatocytes. One such pathway could involve signaling through TRAIL or TNFR1.47, 48 Taken together, our results suggest that TNF-α plays a fundamental role in con A hepatitis by inducing the upregulation of CD40 expression on hepatocytes whose activation induces apoptosis of hepatocytes mediated at least partially by the Fas/FasL pathway. Therefore, based on our findings, we propose a model for the role of the CD154–CD40 costimulatory system in con A–induced hepatitis (Supplementary Fig. 5). We do, however, recognize that this model is an oversimplification of a complex series of events in con A hepatitis.
In conclusion, our studies indicate that the CD154–CD40 costimulatory pathway plays an important proinflammatory role in con A–induced hepatitis by inducing the production of TNF-α within the liver. Subsequently, TNF-α upregulates CD40 expression on hepatocytes, and activation of CD40 induces the apoptosis of hepatocytes through the Fas/FasL pathway. Therefore, we suggest that the CD154–CD40 pathway may be involved in the pathogenesis of FH and that this costimulatory pathway may be a potential target for the development of novel therapeutic approaches to treat patients with fulminant hepatitis.