Paradoxical effects of short- and long-term interleukin-6 exposure on liver injury and repair


  • Potential conflict of interest: Nothing to report.


Interleukin-6 (IL-6) is an important mediator of liver regeneration and repair that is also elevated in chronic liver diseases, including fatty liver of obesity and cirrhosis. IL-6 has been reported both to delay and accelerate liver regeneration. We examined the effects on liver injury and regeneration of a continuous administration of exogenous IL-6 to mice by injection of an IL-6–expressing CHO-cell line in athymic nude mice and by osmotic mini-pump delivery of recombinant murine IL-6. Short-term IL-6 administration (1-2 days) accelerated early recovery of liver mass, whereas more long-term administration (5-7 days) markedly impaired liver regeneration. Similarly, short-term IL-6 treatment increased hepatic resistance to the lethal effects of the Fas agonist Jo-2, but on more prolonged IL-6 exposure the Jo-2 resistance vanished. IL-6 administration initially induced expression of the anti-apoptotic proteins Bcl-2 and Bcl-xL, correlating with protection against Fas-mediated cell death. More prolonged IL-6 administration, however, resulted in marked induction of the pro-apoptotic protein Bax. This result coincided with increased activation of the type II or intrinsic, mitochondrial path to cell death, manifested by increased caspase-9 activation and increased cytochrome c release after Jo-2 exposure. These data demonstrate that IL-6 can function acutely to improve hepatic regeneration and repair, but that more chronic exposure not only abolishes the protective effects of IL-6, but actually sensitizes the liver to injury and death. In conclusion, elevated IL-6 in certain chronic liver diseases contributes to an increased likelihood of liver failure after injury. (HEPATOLOGY 2006;43:474–484.)

Interleukin-6 (IL-6) is a multifunctional pro-inflammatory cytokine.1, 2 In the liver, IL-6 is a major inducer of the acute phase response after infection and is secreted in response to tumor necrosis factor.2 As well, IL-6 is required for normal liver repair and regeneration after injury or 70% partial hepatectomy (PH).3 Mice bearing targeted deletions of the tumor necrosis factor type I receptor and IL-6 null (Il6−/−) mice show increased mortality and severely impaired hepatocyte DNA synthesis after PH.4 Moreover, Il6−/− mice display abnormal liver regeneration after other injuries, including carbon tetrachloride, bile duct ligation, and ethanol.5–7 In all cases, acute administration of IL-6 rescues normal liver regeneration in Il6−/− mice. Based on such data, IL-6 has been recognized as an important early signal for liver regeneration after acute injury.

Absence of IL-6 signaling in repetitive injury models also results in increased hepatocyte injury. Il6−/− mice exhibit increased apoptosis after either short-term or prolonged (8 weeks) exposure to the DeCarli-Lieber ethanol diet.8 As well, mice with liver-specific deletion of the IL-6 co-receptor gp130 exhibit increased injury in response to chronic carbon tetrachloride exposure. These reports suggest an important role of IL-6 in acute liver injury and a beneficial effect of IL-6 production in repetitive liver injury.9–13

IL-6 is both pro-mitogenic and anti-apoptotic for hepatocytes. We have demonstrated that prolonged exposure to continuous, supraphysiological levels of IL-6 induces hepatocyte proliferation in vivo.14 IL-6 over-expression from a retroviral vector in primates induced marked hepatocyte proliferation in vivo.15 Administration of an IL-6/soluble IL-6Rα fusion protein16 accelerates hepatocyte regeneration, while IL-6 pretreatment protects mice and isolated hepatocytes from Jo-2–mediated cell death.5

Serum IL-6 is elevated in animal models of and patients with chronic liver diseases.8, 10, 11 In long-term ethanol exposure, the persistently increased serum IL-6 is also associated with an exaggerated production of IL-6 after PH.17, 18 Similarly, leptin-deficient ob/ob mice, which exhibit obesity, diabetes, steatosis, and abnormal liver regeneration after PH, exhibit elevated IL-6 baseline concentrations, an exaggerated IL-6 response to resection, as well as chronic STAT3 nuclear translocation.13, 19 These results suggest similar IL-6 responses to liver injury in alcoholic liver disease and obesity-related liver disease in both humans and mice.9–13 The consequences of chronic IL-6 over-expression and persistent signaling in alcoholic and obesity-related hepatocyte dysfunction are unknown, and the existing data are conflicting. Although administration of IL-6 improved the regenerative response in obese Zucker rats, hyperstimulation with IL-6 administered to transgenic mice expressing soluble human IL-6Rα actually repressed cell cycle progression after hepatectomy, suggesting that IL-6 can inhibit liver growth after injury.20

Thus, whether persistent IL-6 signaling is protective for hepatocytes or contributes to injury is unknown. IL-6 null models do not address the effects of excess IL-6 on liver function. To examine directly the effect of excess IL-6 on liver homeostasis in vivo, we used our previously published CHO-IL-6/athymic nude mouse model and osmotic mini-pumps to continuously deliver recombinant IL-6. This report examines the effects of sustained high-dose IL-6 on liver regeneration after PH or apoptosis induced by administration of the CD95/Fas agonist, Jo2.


IL-6, interleukin-6; PH, partial hepatectomy; EGFR, epidermal growth factor receptor; TUNEL, transferase–mediated dUTP-nick-end labeling; PARP, poly (ADP-ribose) polymerase; AIF, apoptosis-inducing factor; rmIL-6, recombinant murine IL-6; Ref-1, redox factor-1.

Materials and Methods

Reagents and Antibodies.

Reagents were as follows: recombinant murine IL-6 (PeproTech, Rocky Hills, NJ); alpha–minimum essential medium, dialyzed fetal bovine serum (Gibco, Grand Island, NY); Na-deoxycholate, BrdU, methotrexate (amithopterin), sodium orthovanadate, NaF, phenylmethyl sulfonyl fluoride (Sigma, St. Louis, MO); Quantikine human IL-6 assay (R&D Systems, Minneapolis, MN); protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany); DAKO CSA System, peroxidase (Dako, Carpinteria, CA). Antibodies included anti-Bax, anti-Bad, monoclonal anti-phospho-Bad, anti-Bcl-xL, anti-cleaved caspase-3, anti-Cleaved caspase-7, anti-caspase-9, monoclonal anti-p53, anti-BID, anti-Stat3, anti-phospho-Stat3, anti-cleaved PARP (Cell Signaling Technology, Danvers, MA); anti– apoptosis-inducing factor (AIF), anti-Bcl-2, anti-FLIPS/L, anti redox factor-1 (Ref-1), anti-gp130 (Santa Cruz Biotechnology, Santa Cruz, CA); purified hamster anti-mouse FAS monoclonal antibody (Clone: Jo2), anti-cytochrome c monoclonal antibody (BD Biosciences, San Diego, CA); monoclonal anti-cytochrome c oxidase subunit Vb (COX) (Molecular Probes, Eugene, OR); anti-Bak, monoclonal anti-phospho-epidermal growth factor receptor (EGFR), monoclonal anti-EGFR (Upstate Biotechnology, Lake Placid, NY); monoclonal anti-proliferating cell nuclear antigen, monoclonal anti-beta actin (Sigma); ImmunoPure peroxidase conjugated goat anti-mouse and donkey anti-rabbit IgG, Super Signal West Pico Chemiluminescent Substrate, Coomassie protein assay kit (Pierce Chemical, Rockford, IL).

Cell Lines.

A CHO cell line stably expressing IL-6 as well as a control line (CHO-control) were cultured as described.14

Animal Models.

Mouse experiments were approved by the Institutional Animal Care and Use Committees at both the University of Miami and the University of Rochester Schools of Medicine. Mice received humane care in accordance with NIH publication 86-23 “Guide for the Care and Use of Laboratory Animals.” Female athymic nude mice (Harlan, Indianapolis, IN or Charles River Laboratory, Wilmington, MA), 6 to 10 weeks old, were injected in the thigh with 106 cells in 0.1 mL phosphate-buffered saline. Alzet osmotic mini-pumps (Durect Corporation, Cupertino, CA) containing recombinant murine IL-6 in sterile saline or control pumps with saline only were placed intraperitoneally in 6- to 8-week old male C57BL6/J mice (Jackson Laboratory, Bar Harbor, ME) through a midline incision. Athymic nude mice injected with CHO-control or IL-6 cells were subjected to PH or sham surgery as described21 or injected intraperitoneally with hamster monoclonal anti-mouse CD95/Fas antibody.

Western Blotting Analysis.

Liver sections were homogenized in ice-cold modified RIPA buffer (50 mmol/L Tris-HCl, 1% NP-40, 0.25% Na-deoxycholate, 150 mmol/L NaCl, 1 mmol/L EDTA, 1mmol/L phenylmethyl sulfonyl fluoride, 1 mmol/L sodium orthovanadate, 1 mmol/L NaF, protease inhibitor cocktail, pH 7.4). Isolation of mitochondria was performed as described.21 Densitometry and quantitation of protein expression was performed using Kodak 1D software (Kodak, Rochester, NY). Statistical analysis of expression levels was performed by Student's t test using InStat or Prism software (GraphPad). Survival analysis was performed using the log-rank test (SPSS, Chicago, IL).


Immunohistochemical staining for cleaved caspase-3 was performed with the DAKO CSA Peroxidase System using a 1:200 dilution of antibody. Terminal deoxynucleotidyl transferase–mediated dUTP-nick-end labeling (TUNEL) was performed using the in situ cell death detection kit, POD (Roche).


Effects of IL-6 Pretreatment on Liver Regeneration.

Previously we showed that intramuscular injection of CHO cells expressing recombinant human IL-6 (CHO-IL-6) into athymic nude mice results in sustained serum levels of 40 to 80 ng/mL.14 This was associated with massive hepatocyte proliferation, which over a 12- to 15-day period resulted in livers twofold larger than livers of mice injected with a similarly selected control cell line (CHO-control).

To determine whether continuous IL-6 administration would accelerate liver regeneration, we performed PH on athymic nude mice 48 hours after injection of CHO-IL-6 or CHO-control cells. Although starting liver mass (measured as total weight of the resected liver lobes) was not different in IL-6–treated and control mice on the day of PH (data not shown), IL-6 administration clearly accelerated the early recovery of liver mass. IL-6–treated mice exhibited increased accumulation of liver mass at both 24 and 48 hours after PH versus controls; however, by 4 days after hepatectomy, no significant difference was observed (Fig. 1A).

Figure 1.

Paradoxical effects of IL-6 administration on recovery after hepatectomy. (A) IL-6 pre-treatment accelerated accumulation of liver mass after PH. Mice injected with CHO-IL-6 cells 48 hours before PH showed increased accumulation of liver mass versus CHO-controls at 24 and 48 hours but not at 96 hours post-PH (*P < .05, **P < .01, n = 4-12 per group). (B) Markedly increased 48-hour operative mortality after PH with 5 or 7 days IL-6 pretreatment. Mice were injected with CHO-IL-6 or CHO-control cells either 1, 5, or 7 days before 70% hepatectomy. Graph demonstrates percent survival (to moribund) 48 hours post hepatectomy. (***P < .0001 by Fisher's exact test; n = 11 or 12 per group). (C) TUNEL of livers 6 hours after PH 2 and 7 days after IL-6 pretreatment or control. (D) Quantitation of TUNEL staining in (E) demonstrates dramatically decreased TUNEL positive nuclei with 2 days of IL-6 pretreatment, but paradoxically increased TUNEL positive nuclei after 7 days of IL-6 pretreatment. Each point represents four separate animals with four quantified fields per animal (***P < .001). (E-F). Western blotting analysis and quantitation demonstrates that 2-day but not 7-day IL-6 pretreatment is associated with decreased cleaved caspase-7 and 9 (*P < .05, **P < .01, ***P < .0001, n = 3 per group). IL-6, interleukin-6; PH, partial hepatectomy; TUNEL, transferase–mediated dUTP-nick-end labeling.

We next sought to determine whether longer-term administration of IL-6 would further increase the rate of regeneration. Mice treated for 5 or 7 days with CHO-IL-6 cells exhibited no decrease in total body weight compared with controls, although total liver weight was increased 20% to 40% at 7 days, and total white adipose tissue weight was decreased. Histological analysis of intact livers from IL-6–treated mice without hepatectomy at 5d after injection demonstrated mitotic figures without evidence of apoptosis, reflecting ongoing hepatocyte proliferation as previously described.14 We hypothesized, therefore, that PH at 5 days in the setting of such active pre-resective hyperplasia would result not only in accelerated regeneration because of enhanced hepatocyte proliferation, but also in shorter time to recovery of starting liver mass because of the larger post-resection remnant liver mass. We found, however, that 5-day or 7-day IL-6–treated mice did not tolerate PH, and most died within 36 hours of the procedure (Fig. 1B). We hypothesized that the markedly increased mortality rate after 5 days' or 7 days' IL-6 exposure and subsequent PH might be attributable to abnormalities of apoptosis, consistent with the abnormal regenerative response elicited by hyper–IL-6.20 TUNEL staining of 5-day control and IL-6–treated livers was performed. We found a marked decrease in TUNEL-positive nuclei in livers pretreated with IL-6 for 2 days; however, after more long-term IL-6 treatment slightly more TUNEL-positive nuclei were observed in the IL-6 versus the control group (Fig. 1C-D). These data suggest that the anti-apoptotic effects of IL-6 were effective only after short-term exposure. We next examined the production of cleaved caspases and consistent with the TUNEL result found that there was less cleaved caspase-7 and 9 after 2-day IL-6 treatment but paradoxically more after longer pretreatment (Fig. 1E-F). No increased caspase-3 production was observed in any livers after PH (data not shown), suggesting that the mechanism for apoptosis involved activation of caspase-7 and 9. Given that liver regeneration involves a complex interaction of both mitogenic as well as anti-apoptotic signals, we were concerned that the increased liver apoptosis likely is only one of many potential factors that are dysregulated in PH after chronic IL-6 exposure. Likely insulin resistance, production and effects of acute phase proteins, and loss of protein and fat stores for the regenerative response also might be contributing to the observed increase mortality. To better understand the specific pathophysiological changes in the liver that are occurring after acute and chronic IL-6 exposure, we sought to examine the anti-apoptotic effects of short (2 days) and longer-term (7 days) IL-6 treatment on a second more targeted model of hepatocyte apoptosis. Thus, we examined the effects of in vivo administration of Jo2 antibody, a direct activator of CD95/Fas.5, 22

Effects of IL-6 on Jo-2–Mediated Apoptosis In Vivo.

Short-term IL-6 exposure is known to increase resistance to apoptotic signals.5, 23–25 As well, we have shown that CHO-IL-6 treatment of mice inhibits liver caspase-3 cleavage and activity.23 Thus, we hypothesized that continuous IL-6 treatment would inhibit apoptosis resulting from Jo2 administration. Indeed, 2 days of IL-6 pretreatment markedly decreased mortality in response to Jo2 administration (0.4 μg/g body weight); however, 7 days of CHO-IL-6 pre-treatment had no significant protective effect against Jo2-mediated death over 2 days of observation (Fig. 2A). Furthermore, we examined the effect of a sub-lethal (low) dose of Jo-2 (0.2 μg/g body weight) on concentrations of the serum markers of liver injury, aspartate aminotransferase and alanine aminotransferase; again a transient anti-apoptotic effect was observed (Fig. 2B). A significant survival benefit was observed after short-term IL-6 exposure but not after more prolonged exposure (Fig. 2C). Of note, 7-day IL-6–treated mice had proportionately larger livers than both 2-day IL-6–treated mice and all CHO-controls. Liver size had increased approximately 20% to 40% over controls; however, the Jo2 dose used was based on body weight. Thus, the Jo2 dose per gram of liver tissue given to the 7-day IL-6–treated mice was lower than that given to the 7-day control mice. In fact, slight increases in dose as would have occurred if dose were calculated proportional to liver mass resulted in substantially increased mortality, suggesting there may even have been increased sensitivity to the lethal effects of Jo2 in the 7-day IL-6–treated mice (data not shown).

Figure 2.

Short-term but not long-term IL-6 treatment protects mice from Fas-mediated liver injury and apoptosis. (A) Enhanced survival with 2 days of IL-6 pretreatment. Mice were injected with CHO-IL-6 or CHO-control cells 2 or 7 days before challenge with intraperitoneal injection of Jo2 antibody. Markedly increased 48-hour survival (to moribund) was observed in 2-day IL-6–treated mice vs. controls. A statistically significant decrease in survival was noted for mice treated for 7 days with IL-6. (***P < .0001, n = 11-13 per group). (B) Reduced serum AST and ALT with 2-day but not 7-day pretreatment with IL-6 prior to sublethal dosage of Jo2 (0.2 μg/g body weight) (***P < .001, n = 5 per group) (C) Kaplan-Meier survival curves reveal reduced Jo2 mortality after 2-day but not 7-day pretreatment with IL-6 (***P < .001, log rank test, n = 12 per group). (D) Inhibition of caspase-3 cleavage by 2-day IL-6 pretreatment. Western blotting analysis of liver lysates from mice pretreated the indicated number of days with CHO-control or CHO-IL-6 cells, then challenged with Jo2 antibody and killed 6 hours later, demonstrates near-complete inhibition of caspase-3 cleavage in 2-day CHO-IL-6 livers vs. controls. In contrast, robust caspase-3 cleavage was observed in 7-day CHO-IL-6 Jo2 livers. (E) Histological analysis demonstrates protective effects of 2-day IL-6 pretreatment on Jo2-induced hepatocyte apoptosis. Hematoxylin-eosin staining of liver sections shows reduced hemorrhage, nuclear break, and apoptotic cell bodies in 2-day CHO-IL-6 livers vs. controls. Figure labels indicate original magnification.

Western blotting analysis of liver extracts from Jo2-treated mice showed no induction of caspase-7 as was observed after hepatectomy. Rather, reduced cleaved caspase-3 protein in 2-day IL-6–treated mice versus 2-day controls was observed (Fig. 2D). In stark contrast, cleaved caspase-3 levels in Jo2-treated 7-day CHO-IL-6 livers were comparable to controls, although the 17-kd fragment was more abundant in the IL-6–treated mice than in controls. Histological analysis demonstrated decreased cellular injury and fewer apoptotic figures in the 2-day IL-6–treated group relative to 2-day controls (Fig. 2D). TUNEL-staining for DNA fragmentation and immunohistochemistry for cleaved caspase-3 (the active form of the enzyme) demonstrated fewer TUNEL-positive nuclei and markedly less cleaved caspase-3 immunoreactivity in 2-day IL-6–treated livers over 2-day controls. In contrast, 7-day IL-6–treated Jo2 livers showed similar cellular injury, apoptotic figures, TUNEL staining and cleaved caspase-3 immunoreactivity as the Jo2 controls, consistent with a lack of protection with 7 days of IL-6 treatment.

Sustained IL-6 Administration Inhibits Apoptotic Pathways.

We next examined the effects of continuous IL-6 administration on activation of the gp130/STAT3 signaling pathway and markers of apoptosis. Western blotting analysis showed persistent STAT3 phosphorylation in CHO-IL-6–treated livers as early as 2 days after CHO-IL-6 injection (Fig. 3A). Expression of the IL-6 co-receptor gp130 was increased by CHO-IL-6 treatment at all times examined. Consistent with our previous observations, cleavage of the caspase-3 substrate poly (ADP-ribose) polymerase (PARP) was chronically and progressively reduced by CHO-IL-6 treatment. These results suggest that chronic IL-6 administration led to persistent STAT3 activation, upregulation of the IL-6 signaling pathway, and inhibition of the apoptotic cascade (Fig. 3A).

Figure 3.

Modulation of IL-6 signaling and apopotic pathway members by IL-6 treatment. (A) Western blotting analysis of liver extracts from mice treated with CHO-control or CHO-IL-6 cells for the indicated number of days before killing demonstrates increased phospho-STAT3 and gp130 levels at all times examined. In contrast, cleavage of the caspase-3 substrate PARP was progressively reduced with longer IL-6 treatment. Western blotting for β-actin indicates equivalent loading. (B) Early induction of anti-apoptotic and later induction of pro-apoptotic Bcl-2 family members by sustained IL-6 treatment. Western blotting analysis of same extracts as in A shows increased expression of Bcl-2 and Bcl-xl beginning at day 2 of IL-6 treatment and continuing through 7 days. Increased expression of FLIP and Ref-1 were not observed at any time. In contrast to the early and consistently elevated expression of Bcl-2 and Bcl-xL, expression of the pro-apoptotic protein Bax was modestly elevated from 2 to 5 days of IL-6 treatment, then dramatically increased by 7 days. Expression of Bid was essentially unchanged. Expression of the BH3-only protein Bad and its phosphorylated form were both increased by IL-6 treatment. (C) Densitometry shows Bax expression was induced 2- to 3-fold from 2 to 5 days of IL-6 treatment; by 7 days, Bax expression was approximately 7-fold normal levels (*P < .05, **P < .0001, n = 3-4 mice per point.). Results representative of two or three experiments). IL-6, interleukin-6; PARP, poly (ADP-ribose) polymerase.

To determine whether the inhibition of Jo-2–induced apoptosis might be attributable to induction of anti-apoptotic proteins by IL-6 treatment, we performed Western blotting analysis of liver lysates from mice treated with CHO-control or CHO-IL-6 cells. Indeed, both Bcl-2 and Bcl-xL demonstrated a sustained induction of roughly 3-4 fold over CHO-controls on days 2 through day 7 of CHO-IL-6 treatment (Fig. 3B). No induction of either the anti-apoptotic bifunctional endonuclease/redox factor Ref-1 or the caspase-8/10 inhibitor FLIP (FLICE-like inhibitory protein) was observed at any time point (Fig. 3B).25–29 Thus, chronic IL-6 administration led to persistent STAT3 activation and sustained induction of the anti-apoptotic proteins Bcl-2 and Bcl-xL.

IL-6 Induces Bax, Bad, and Mitochondrial Destabilization.

Based on these results, we found no evidence for late downregulation of anti-apoptotic proteins, but rather found evidence for sustained anti-apoptotic signaling in unchallenged, IL-6–treated livers as previously reported.5 Indeed, IL-6 exposure maintained increased Bcl-2 and Bcl-xL levels regardless of duration of exposure. Thus we examined the alternative explanation for the paradoxically increased sensitivity to apoptosis upon more prolonged IL-6 treatment, namely, upregulation of pro-apoptotic proteins. BH3-only members of the Bcl-2 super-family function as upstream sentinels that respond to specific, proximal death and survival signals.30 One of these BH2-only proteins, Bid, is activated by caspase-8 cleavage after engagement of cell surface death receptors. When cleaved, the tBid (truncated Bid) translocates to the mitochondria, where it triggers Bak oligomerization with subsequent mitochondrial damage and activation of the intrinsic path to cell death.30 Western blotting analysis of liver lysates from control and IL-6–treated mice, however, demonstrated no increase in levels of full-length Bid (Fig. 3B). Truncated Bid was undetectable. A second BH3-only protein, Bad, determines life or death for the cell depending on the presence or absence of survival factors and glucose. Growth factors and glucose result in phosphorylation and inactivation of Bad, preventing its activation of the pro-death Bcl-2 family member Bax.30 Although Western blotting analysis showed that IL-6 treatment markedly increased total Bad protein over CHO-controls, total phospho-Bad protein also increased, suggesting that the protein was inactive in the unstimulated, CHO-IL-6–treated livers (Fig. 3B).

BH3-only proteins activate Bak and Bax, causing their dissociation from anti-apoptotic proteins such as Bcl-2, Bcl-xL and MCL-1, and translocation to the endoplasmic reticulum and mitochondrion, triggering cytochrome c release, formation of the apoptosome, and activation of effector caspases to induce apoptosis. Western blotting analysis showed that a tBid target protein, Bak, was not increased by chronic IL-6 treatment versus controls (Fig. 3B). In stark contrast, expression of the Bcl-2 family member death agonist Bax was markedly induced by IL-6 treatment, with expression peaking at later times precisely when CHO-IL-6 mice exhibit both decreased survival after PH and lack of protection from Jo2-induced apoptosis (Fig. 3B-C).

Chronic IL-6 Leads to Increased Jo-2–Mediated Mitochondrial Bax, Cytochrome c and AIF Release, and Activation of Caspase-9.

The observed upregulation of Bax suggested that the cause of chronic IL-6 exposure-associated hepatocyte sensitization to death might be enhanced activation of the intrinsic (mitochondrial) pathway to cell death. If this hypothesis is correct, increased mitochondrial Bax, increased cytochrome c release, and enhanced caspase-9 cleavage would be observed after an extrinsic apoptotic stimulus. We therefore treated mice with CHO-IL-6 or CHO-control cells for 2 or 7 days and analyzed the intrinsic pathway to apoptosis by Western blotting analysis of total and fractionated liver lysates. In the absence of Jo2 administration, liver samples from mice subjected to 7-day CHO-IL-6 treatment demonstrated biochemical features of mitochondrial destabilization (Fig. 4), despite histological evidence of robust hepatocyte mitosis and no apparent apoptosis (data not shown). Specifically, livers from 7-day CHO-IL-6 mice showed substantial mitochondrial Bax association and increased cytosolic apoptosis-inducing factor (AIF),31 a mitochondrial effector of apoptotic cell death. Despite the presence of these two key biochemical features of apoptosis, the intrinsic pathway to cell death was not activated in 7-day CHO-IL-6 livers as indicated by the fact that cytosolic cytochrome c, caspase-9 cleavage, caspase-3 cleavage, and PARP cleavage were actually decreased or unchanged relative to CHO-controls.

Figure 4.

Evidence of increased mitochondrial destabilization and mitochondria-mediated apoptosis with chronic IL-6 treatment. Mice were treated for 2 days or 7 days with CHO-control or CHO-IL6 cells, then subjected to intraperitoneal injection of Jo2 antibody or carrier and killed 6 hours later. Liver samples were homogenized or separated into cytosolic and mitrochondrial fractions and subjected to Western blotting analysis. IL-6, interleukin-6.

Challenge of CHO-control and CHO-IL-6–injected mice with Jo2 resulted in biochemical evidence for inhibition of apoptosis in 2-day CHO-IL-6 mice versus 2-day CHO-controls. Mitochondrial Bax was increased in 2-day CHO-IL-6 mice versus 2-day CHO-controls; however, mitochondrial injury was diminished in 2-day CHO-IL-6 livers. Specifically, Western blotting analysis of cytosolic fractions demonstrated decreased cytochrome c and decreased AIF in 2-day CHO-IL-6 Jo2 livers versus 2-day CHO-control Jo2 livers. Moreover, capase-9 and caspase-3 cleavage in 2-day CHO-IL-6 Jo2 livers were also dramatically reduced over 2-day CHO-control Jo2 livers.

In contrast to the protection against apoptosis produced by 2 days of IL-6 treatment, 7 days of IL-6 treatment did not diminish activation of apoptotic pathways. Analysis of extracts from 7-day CHO-IL-6 Jo2-treated livers demonstrated equivalent caspase-9 activation, similar caspase-3 cleavage, and comparable cytosolic cytochrome c levels as 7-day CHO-control Jo2-treated livers. Strikingly, both mitochondrial Bax and cytosolic AIF were increased in 7-day CHO-IL-6 Jo2-treated livers. Taken together, these results suggest that long-term IL-6 treatment results in increased Bax expression, which tilts the balance of life or death toward apoptosis in the presence of an apoptogenic agent such as PH or Jo2 administration.

IL-6 Administered Through Osmotic Mini-pump Increases Bax.

To confirm that the observed induction of Bax was not a phenomenon of the athymic nude mouse/CHO model system, we also delivered recombinant murine IL-6 (rmIL-6) to immuno-competent C57BL6/J mice through intraperitoneally-implanted osmotic mini-pumps. Control mice were implanted with pumps containing carrier only. The osmotic mini-pumps delivered 10 μg of IL-6 per day for 7 days. Serum rmIL-6 concentrations roughly 80 times normal or carrier-only pump levels were measured at the time of necropsy (2,182 ± 1427 pg/mL vs. 27.35 ± 22.49 pg/mL). Consistent with our previous report of IL-6–induced hepatomegaly, increased liver mass and increased hepatocyte mitotic figures were observed, with no evidence of apoptosis (unpublished data). Western blotting analysis of liver lysates from mice treated for 7 days with rmIL-6 pumps demonstrated sustained STAT-3 phosphorylation, increased gp130 expression, increased proliferating cell nuclear antigen expression, decreased PARP cleavage, and increased Bax expression as compared with carrier-only control livers (Fig. 5 and Table 1). These results confirm in a second model system the pro-mitogenic and anti-apoptotic effects of chronic IL-6 administration, despite the induction of Bax.

Figure 5.

Adminstration of recombinant murine IL-6 by osmotic mini-pump induces STAT3 signaling, proliferation, and suppression of apoptotic markers in the presence of increased Bax. Western blotting analysis of liver lysates from mice treated for 7 days with saline or rmIL-6 delivered by intraperitoneal osmotic mini-pump shows that this delivery system recapitulates the effects of IL-6 delivered in the CHO-nude mouse model, including Bax induction. Results representative of two experiments, each with n = 4 or 5 per group. IL-6, interleukin-6.

Table 1. Relative Protein Expression in Livers of Pump-Bearing Mice
  1. Densitometry and statistical analysis of Western blotting results shown in Fig. 5 (N = 4 per condition).

pSTAT31 ± 0.647.18 ± 2.02<.01
STAT31 ± 0.050.80 ± 0.16NS
GP1301 ± 0.181.39 ± 0.12<.05
PCNA1 ± 0.051.47 ± 0.20<.05
cPARP1 ± 0.080.26 ± 0.08<.001
BAX1 ± 0.412.12 ± 0.31<.05
C-MYC1 ± 0.432.41 ± 0.62<.05
SOCS31 ± 0.412.34 ± 0.87<.05


A number of reports have identified an important role for IL-6 in the hepatic regenerative response after injury or resection.4–7 More recently, investigators have also examined the potential therapeutic application of IL-6 and a linked IL-6 co-receptor complex (termed “super-IL-6”) in liver disease.16 As demonstrated herein, exogenous IL-6 may accelerate recovery of liver mass in vivo, consistent with findings that IL-6 may have therapeutic application in the prevention of ischemic injury during organ preservation before transplantation and in amelioration of hepatic steatosis.32–35 The role and effects of the persistently increased IL-6 present in a wide variety of disease states, however, remains poorly defined. Several reports suggest that this chronic signaling may be protective against injury. We and others have reported, however, that IL-6 may alter glucose resistance in vivo and may actually contribute to the insulin resistance observed in diabetes.36, 37

We show that short-term IL-6 exposure is anti-apoptotic and accelerates recovery of liver mass after hepatecomy and confirm the findings of Kovalovich et al.5 that IL-6 decreases hepatic sensitivity to Jo2-induced hepatocyte apoptosis. This resistance is associated with decreased activation of the mitochondrial apoptotic pathway with increased expression of the anti-apoptotic factors Bcl-2 and Bcl-xL. Less impressive induction of FLIP or Ref-1 is observed in our model system than has been reported by others.5, 25 Our data demonstrate that acute IL-6 treatment is associated with decreased cleaved-caspase-9 and cytoplasmic release of mitochondrial proteins. These data support and confirm the potential application of IL-6 as a factor to improve liver preservation.33

The acute effects of IL-6 on the liver in vivo are, however, in apparent conflict with other published reports that IL-6 delayed the regenerative response.20 We initially investigated whether longer-term IL-6 would continue to be beneficial for recovery from liver resection and were surprised to find impaired recovery after injury. We report herein the novel observation that prolonged exposure to IL-6 in vivo is anti-regenerative and pro-apoptotic. Prolonged exposure to IL-6 is associated not only with the continued increased production of multiple anti-apoptotic factors but with the induction of mitochondrial-destabilizing pro-apoptotic protein Bax. The induction of Bax sensitizes hepatocytes to the intrinsic pathway to cell death by tipping the balance between anti- and pro-apoptotic signals toward mitochondrial destabilization and death through increased caspase-9 activation. Induction of these pro-apoptotic factors appears to override the early anti-apoptotic effects of IL-6.

Upregulation of hepatic Bax has been observed in chronic liver disease and in liver dysfunction in late sepsis.38–41 These results suggest persistent abnormal signaling in these conditions, possibly because of chronic IL-6 overproduction, may be mediating some of the cellular deficiencies observed in these chronic states. An association of IL-6 and STAT-3 activation with fatty liver disease has been demonstrated,13 with ob/ob mice exhibiting persistent STAT3 signaling as well as profound steatosis and abnormal liver regeneration.

The cellular pathways involved in chronic IL-6–induced upregulation of Bax are unclear. We have previously demonstrated prolonged IL-6 exposure induced decreased baseline c-met and EGFR expression and phosphorylation, suggesting decreased hepatocyte signaling of the important hepatic mitogens hepatocyte growth factor and EGF, respectively.14 Such growth factor withdrawal/inhibition of growth factor signaling is well known to induce Bad production/activation and mitochondrially mediated apoptosis. As well, we have found that persistent IL-6 signaling results in increased expression of suppressor of cytokine signaling-3 (unpublished observation), which is known to inhibit signaling of multiple pro-mitogenic pathways, including the insulin and EGFR pathways. The degree of suppressor of cytokine signaling-3 induction in our system, however, is perhaps 2- to 3-fold and considerably less than that observed during the liver regenerative response.42 Another intriguing possibility is that organ cross-talk secondary to systemic cachexia may be involved in the hepatic Bax response. Future studies are targeted at determining the molecular pathways altered by chronic IL-6 over-expression.

Finally, we found that the terminal caspase activated by PH and caspase-3 differed. PH activated caspase-7 and Jo-2 activated caspase-3. This was observed also in control mice not treated with IL-6. In both models caspase-9 activation occurred, suggesting that both pathways involve a mitochondrial component to the terminal caspase activation. How these pathways ultimately result in the observed different caspase activation, however, is unclear. As well, although IL-6 pre-treatment resulted in an improved early recovery of liver mass after PH, no improved survival after PH was observed, making the benefit of IL-6 unclear. Other questions also remain, including the contribution of IL-6–induced peripheral insulin resistance and acute phase proteins to the dysfunction after PH or Jo-2. Future studies targeting disruption of hepatic STAT3 and its downstream targets are planned.