Interleukin 22 (IL-22) plays a protective role in T cell-mediated murine hepatitis: IL-22 is a survival factor for hepatocytes via STAT3 activation

Authors

  • Svetlana Radaeva,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Rui Sun,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Hong-na Pan,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Feng Hong,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Bin Gao

    Corresponding author
    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    • Section on Liver Biology, NIAAA/NIH, Park Bldg. Rm 120, 12420 Parklawn Drive, MSC 8115, Bethesda, MD 20892
    Search for more papers by this author
    • fax: 301-480-0257


Abstract

The central role of T cell activation in hepatocellular injury has been well documented. In this article, we provide evidence suggesting that T cells may also play a protective role in liver disease by releasing interleukin-22 (IL-22), a recently identified T cell-derived cytokine whose biological significance is unclear. IL-22 messenger RNA and protein expression are significantly elevated in T cell-mediated hepatitis induced by concanavalin A (ConA) but are less extensively elevated in the carbon tetrachloride-induced liver injury model. Activated CD3+ T cells are likely responsible for the production of IL-22 in the liver after injection of ConA. The IL-22 receptor is normally expressed at high levels by hepatocytes and further induced after ConA injection. IL-22 blockade with a neutralizing antibody reduces signal transducer and activator of transcription factor 3 (STAT3) activation and worsens liver injury in T cell-mediated hepatitis, whereas injection of recombinant IL-22 attenuates such injury. In vitro treatment with recombinant IL-22 or overexpression of IL-22 promotes cell growth and survival in human hepatocellular carcinoma HepG2 cells. Stable overexpression of IL-22 in HepG2 cells constitutively activates STAT3 and induces expression of a variety of antiapoptotic (e.g., Bcl-2, Bcl-xL, Mcl-1) and mitogenic (e.g., c-myc, cyclin D1, Rb2, CDK4) proteins. Blocking STAT3 activation abolishes the antiapoptotic and mitogenic actions of IL-22 in hepatic cells. In conclusion, the T cell-derived cytokine IL-22 is a survival factor for hepatocytes; this suggests that T cell activation may also prevent and repair liver injury by releasing hepatoprotective cytokine IL-22 in addition to its previously documented central role in hepatocellular injury. (HEPATOLOGY 2004;39:1332–1342.)

The central role of T cell activation in hepatocellular injury in a variety of human liver disorders1–9 and animal models10, 11 has been well documented. Evidence suggests that T cell-induced liver injury is mediated through multiple mechanisms, including cytolytic pathways (e.g., Fas, perforin) and the release of a wide variety of proinflammatory cytokines such as interferon γ, tumor necrosis factor α, and interleukin (IL) 4.1–11 Conversely, activated T cells also produce anti-inflammatory cytokines such as IL-10 and antiapoptotic cytokines such as IL-6, suggesting that activated T cells may play an important role in repairing liver injury in addition to their central role in hepatocellular injury.12, 13 However, the protective role of T cells in controlling the progression of liver injury has largely been unexplored.

Interleukin-22 (IL-22)/IL-10-related T-cell-derived inducible factor (IL-TIF) is a recently identified cytokine originally isolated by complementary DNA subtraction as a gene specifically induced by IL-9 in mouse T cells14, 15 and later classified as IL-22 in the Lifeseq EST (Incyte Pharmaceuticals, Palo Alto, CA) database according to its significant similarity to IL-10.16 IL-22 belongs to the IL-10 family of cytokines, which also includes IL-10, IL-19, IL-20, and IL-24.17 IL-22 exerts its functions by binding to a receptor complex composed of 2 chains: IL-10Rβ and IL-22R. The latter is mainly expressed at the highest levels in the pancreas, followed by the small intestine, colon, kidney, and liver.16, 18 IL-22 has been shown to activate multiple signaling pathways, including the Janus kinase-signal transducer and activator of transcription factor (STAT) and mitogen-activated protein kinase pathways in hepatic cells.19

Although it has been clearly demonstrated that IL-22 targets the IL-10Rβ and IL-22R complex that is expressed in various tissues and cell lines, the functions of IL-22 remain obscure. It has been shown that IL-22 up-regulates expression of acute-phase proteins in the liver15 and of pancreatitis-associated proteins in pancreatic acinar cells;20 this suggests that IL-22 plays a role in inflammatory responses. In the present article, we demonstrate for the first time that IL-22 plays a protective role in T cell-mediated hepatitis and is a survival factor for hepatocytes.

Abbreviations:

IL, interleukin; cDNA, complementary DNA; STAT, signal transducer and activator of transcription factor; ERK, extracellular signal-regulated protein kinase; RT-PCR, reverse-transcriptase polymerase chain reaction; LDH, lactate dehydrogenase; MNC, mononuclear cell; FACS, fluorescence-activated cell sorter; ConA, concanavalin A; ALT, alanine transaminase; AST, aspartate transaminase; mRNA, messenger RNA.

Materials and Methods

Materials.

Anti-STAT3, anti-phospho-STAT3, anti-phospho-STAT1, and anti-phospho-extracellular signal-regulated protein kinase 1/2 (ERK1/2) antibodies were obtained from Cell Signaling (Beverly, MA). Other antibodies used included Mcl-1, Bcl-2, and c-myc antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) cyclin D1, Bcl-xL, PCNA, CDK2, CDK4, Rb2, and Fas ligand/CD95L antibodies (PharMingen, San Diego, CA); and anti–IL-22 antibodies (R&D Systems, Minneapolis, MN). Recombinant IL-22 proteins were purchased from Research Diagnostic Inc. (Flanders, NJ). STAT3 inhibitor peptide (Ac-PYpLKTK-OH), which was shown to inhibit specifically STAT3 activation,21, 22 was from Calbionchem-Novabiochem Corp. (San Diego, CA).

IL-22 Expression Vector.

The pcDNA3-IL-22 expression vector was constructed by inserting murine IL-22 cDNA into the BamHI/EcoRI site of the mammalian expression vector pcDNA3.1(+) (Invitrogen, Carlsbad, CA). The cDNA-encoding full-length IL-22 was obtained from the thymus of B6 mice by reverse-transcriptase polymerase chain reaction (RT-PCR). The following 2 primers, which included BamHI and EcoRI sites (underlined), were used: 5′-GTGGGA TCCCTGATGGCTGTCCTGCAG-3′ and 5′-AGCGAATTCTCGCTCAGACTGCAA GCAT-3′.

Stable Transfection of IL-22 cDNA in HepG2 and Hep3B Cells.

HepG2 and Hep3B cells (ATCC, Rockville, MD) were grown in 6-well culture plates to about 50%-60% confluence. Both the IL-22 expression vector and the pcDNA3 vector control DNA were separately transfected into HepG2 or Hep3B cells using lipofectin (Invitrogen). After 16 hours, the medium was replaced with fresh normal-growth medium and cells were grown for an additional 48 hours. The cells were then exposed to a selective concentration of 800μg/mL of geneticin (Clontech, Palo Alto, CA) to isolate stably transfected cells. Positive colonies were confirmed by Western blot analysis for IL-22 protein expression. In a comparison with purified IL-22 proteins by Western blot analysis, IL-22 concentration in the supernatant of stably transfected HepG2 cells was estimated to be less than 5 ng/mL.

Cell Extraction, Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis, and Western Blotting.

Cell extract preparation and Western blot analysis were performed as described previously.23

Cell Culture and Proliferation.

Primary mouse hepatocyte culture and [3H]thymidine (1 μCi per well) uptake were performed as described previously.23 Cell proliferation was also determined by BrdU labeling, measured by a cell proliferation enzyme-linked immunosorbent assay kit (Roche, Penzberg, Germany). The relative BrdU incorporation was expressed as absorbance (A370nm-A492nm). Cell viability was measured by lactate dehydrogenase (LDH) release using a cytotoxicity assay kit (Promega, Madison, WI). The percentage of cytoxicity was expressed as experimental LDH release (optical density490)/maximum LDH release (optical density490).

Flow Cytometric Analysis and DNA Fragmentation Assay of Apoptosis.

Flow cytometric analysis of cell apoptosis with propidium iodide, and DNA fragmentation assay of apoptosis were performed as described previously.23

Colony Growth in Soft Agar and Tumor Growth in Nude Mice.

HepG2 and Hep3B cells transfected with pcDNA3-IL-22 or pcDNA3-neo were plated on top of a layer of 0.3% agar laid over a lower layer of 0.6% agar in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and G418 at a density of 1 × 104 cells per well in 6-well plates. After 3 or 6 weeks, colonies were stained with crystal violet (0.005%) and scored under the dissection microscope (colonies larger than ≈0.1 mm in diameter were scored as positive). To measure the oncogenicity of IL-22 transfectants, 1 × 106 cells in 0.1 mL saline were injected subcutaneously into the right gluteal region of nude mice using a 26-gauge needle. Mice were monitored at 1- to 2-day intervals for the formation of tumors over a period of 4 to 6 weeks. The tumors were then dissected, weighed, and further analyzed.

Isolation of Liver Mononuclear Cells (MNCs) and CD3+ T Cells, and Fluorescence-Activated Cell Sorter (FACS) Analysis of CD3+ T cells.

Mouse liver MNCs were isolated as described previously.24 The isolated hepatic MNCs were resuspended in saline for RT-PCR or further purification of CD3+ T cells. For purification of CD3+ T cells, liver MNCs were stained with fluorescein isothiocyanate-conjugated anti-CD3 antibody (Pharmingen) and then incubated with anti-fluorescein isothiocyanate microbeads (Miltenyi Biotec, Auburn, CA) for 15 minutes at 4°C. CD3+ T cells were enriched by positive magnetic cell sorting according to the manufacturer's protocol. Approximately 92.12% of magnetic cell sorting-purified cells were CD3 positive as confirmed by FACS analysis. Hepatic CD3+T cells were determined by FACS analysis with the anti-CD3 antibody (PharMingen) as described previously.24

RT-PCR.

RT-PCR was performed as described previously.23 Primer sequences used in the study are listed in Table 1. Genomic DNA contamination was ruled out because polymerase chain reaction (PCR) using RNA without reverse transcription failed to yield amplicons. PCR bands were scanned using Storm PhosphoImager (Molecular Dynamics, Piscataway, NJ).

Table 1. PCR Primers Used in This Study
GenesSequencesSizes (bp)
  1. Abbreviation: bp, base pair.

Murine IL-22F (5′-3′) GTGGGATCCCTGATGGCTGTCC-TGCAG560
 R (5′-3′) AGCGAATTCTCGCTCAGACTGC-AAGCAT 
Murine IL-22RF (5′-3′) CTACGTGTGCCGAGTGAAGA189
 R (5′-3′) AAGCGTAGGGGTTGAAAGGT 
Murine IL-10RβF (5′-3′) GCCAGCTCTAGGAATGATTC400
 R (5′-3′) AATGTTCTTCAAGGTCCAC 
Human IL-22RF (5′-3′) CCTGAGCTACAGATATGTCAC-CAAG78
 R (5′-3′) GGCTGGAAAGTCAGGACTCG 
Human IL-10RβF (5′-3′) GCCTGTCTGTGAGCAAACAA242
 R (5′-3′) CCGACAATGGAAAGGAGAAA 

Examination of Liver Injury.

Liver injury was quantified by measuring plasma enzyme activities of alanine transaminase (ALT) and aspartate transaminase (AST) using a kit from Sigma Chemical Co. (St. Louis, MO) and determined by hematoxylin-eosin staining.

Gel Shift Mobility Assay.

The gel mobility shift assay was performed as described previously.23 The double-stranded oligonucleotide-containing (5′-GTC GAC ATT TCC CGT AAA TCG TCG A 3′) STAT3-binding site was used as a probe.

Statistical Analysis.

Statistical analysis was performed using the Student t test. All P values were 2-tailed; P less than .05 was taken as statistically significant.

Results

Expression of IL-22 and IL-22R in the Liver is Significantly Induced in T Cell-Mediated Hepatitis.

Expression of IL-22 in the liver was examined in 3 murine models of liver injury. After injection of concanavalin A (ConA), IL-22 protein expression was significantly induced in the liver (Fig. 1A), peaking between 1 and 9 hours and returning to basal levels at 24 hours. IL-22 protein expression was also induced in murine liver injury models with lipopolysaccharide (LPS)/D-galactosamine and CCl4, but was less extensive (Fig. 1A).

Figure 1.

Expression of IL-22, IL-22R, and IL-10Rβ in murine models of liver injury. (A) C57/BL6 mice (7-8 weeks) were injected with ConA (12 μg/g intravenously), LPS (0.5 ng/g intraperitoneally) with D-galactosamine (D-Gal; 800 μg/g intraperitoneally), or CCl4 (1 μL/g orally). After various time periods, total liver extracts were then subjected to Western blotting using anti–IL-22 and anti–β-actin antibodies. (B) C57/BL6 mice (7-8 weeks) were injected with ConA (12 μg/g intravenously) for various time periods. Hepatoctyes, liver MNCs, and CD3+ T cells were purified and subjected to RT-PCR analysis with primers for IL-22, IL-22R, IL-10Rβ, and β-actin. (C) C57/BL6 mice were treated identically to those in panel A. After various time periods, liver MNCs were isolated and subjected to FACS analysis of CD3+ T cells. Total numbers of CD3+ T cells were obtained by total liver MNCs × % CD3+ T cells. Values represent means ± SEM. for 4 mice in each group. *P < .05; **P < .01 versus the corresponding control group at 0 h.

To determine which cell types were responsible for inducing IL-22 expression, hepatocytes, hepatic MNCs, and CD3+ T cells were isolated. As shown in Fig. 1B, ConA injection significantly induced IL-22 messenger RNA (mRNA) expression in the intact liver, hepatic MNCs, and purified CD3+T cells, but not in isolated hepatocytes. Moreover, RT-PCR showed that the liver and isolated hepatocytes expressed high levels of IL-22R, which were further elevated after injection of ConA. In contrast, expression of IL-22R was undetectable in hepatic MNCs and CD3+ T cells. Expression of IL-10Rβ, which is also required for IL-22 signaling,16, 18 was detected in the liver and in hepatocytes, MNCs, and CD3+ T cells.

Because CD3+T cells appear to be the major source of IL-22, we compared infiltration of CD3+ T cells in these 3 models of liver injury. As shown in Fig. 1C, the number of CD3+ T cells was significantly increased after injection of ConA and slightly increased after injection of LPS/D-galactosamine, but it declined significantly after injection of CCl4.

IL-22 Plays a Protective Role in ConA-Induced Liver Injury.

To define the role of IL-22 in T cell-mediated hepatitis, we explored the effects of neutralizing IL-22 antibodies (IL-22mAb) on liver injury. As shown in Fig. 2A, injection of a modest dose of ConA (10 μg/g) induced a significant elevation in ALT and AST and spotted necrosis in the liver, whereas injection of ConA combined with neutralizing IL-22 antibodies markedly enhanced ALT and AST levels and induced massive necrosis in the liver. In contrast, pretreatment of mice with recombinant IL-22 protein (rIL-22) before ConA injection almost completely prevented elevations in ALT and AST and the massive necrosis induced by injection of a high dose of ConA (12 μg/g; Fig. 2B).

Figure 2.

Protective effects of IL-22 in murine models of liver injury. (A) C57/BL 6 mice were injected with ConA (10 μg/g intravenously). Thirty minutes later, mice were treated (intravenously) with an IL-22 neutralizing antibody (IL-22mAb 2.5 μg/g) or control immunoglobulin G (2.5 μg/g). After 16 hours, serum was collected for measuring ALT and AST levels, and the livers collected for hematoxylin-eosin staining. (B) C57/BL 6 mice were injected with recombinant IL-22 protein (rIL-22) (0.25 μg/g intravenously). After 6 hours, mice were administered ConA (12 μg/g intravenously) for 16 hours. Serum was collected for measuring ALT and AST levels, and the livers collected for hematoxylin-eosin staining. (C and D) C57/BL6 mice (7-8 weeks) were injected with LPS (0.5 ng/g intraperitoneally) with D-galactosamine (800 μg/g, i.p.) or CCl4 (1μl/g orally). Thirty minutes later, mice were treated intravenously with an IL-22 neutralizing antibody (2.5 μg/g) or control immunoglobulin G (IgG; 2.5 μg/g). After 16 hours, serum was collected for measuring ALT and AST levels. (A-D) Bars represent means ± SEM. for 4 mice in each group. Data in panels A-D represent 2 independent experiments. *P < .05; **P < .01 versus the corresponding saline-treated group.

The role of IL-22 in LPS/D-galactosamine- and CCl4-induced liver injury was also examined. As shown in Figs. 2C and 2D, injection of neutralizing IL-22 antibodies significantly worsened LPS/D-galactosamine-induced elevation of ALT and AST but had a minimal effect on CCl4-induced liver injury.

IL-22 Is Partially Responsible for Hepatic STAT3 Activation in T Cell-Mediated Hepatitis.

IL-22 has been shown to activate multiple signaling pathways in H4IIE rat hepatoma cells.19 In the present study, we examined IL-22 activation of signals in the liver in vivo. As shown in Fig. 3A, injection of rIL-22 significantly activated STAT3 in the liver with peak effect at 1 hour and return to basal levels at 3 hours after injection. Injection of IL-22 only induced marginal activation of ERK1/2 and did not induce STAT1 activation.

Figure 3.

IL-22 is partially responsible for hepatic STAT3 activation in T cell-mediated hepatitis. (A) C57/BL 6 mice were injected with IL-22 (0.2 μg/g intravenously). After various time periods, liver extracts were prepared and subjected to Western blot analyses with various antibodies. (B) C57/BL 6 mice were injected with ConA (10 μg/g intravenously). Thirty minutes later, mice were treated intravenously with an IL-22 neutralizing antibody (2.5 μg/g) or control immunoglobulin G (IgG; 2.5 μg/g). After 1 and 3 hours, livers were collected for Western blot analyses. (C) The pSTAT3 protein levels in panel B were quantified using ImageQuant software (Molecular Dynamics) and normalized to STAT3 protein levels at each time point. Fold induction is the relative induction compared with untreated wild-type control mice. Values represent means ± SEM for 4 mice in each group. (D) C57/BL6 mice were treated identically to those in panel B. After 1 and 3 hours, spleens were collected for Western blot analyses. *P < .05; **P < .01 versus the corresponding control group at 0 h.

Next, we examined whether IL-22 was involved in STAT3 activation in ConA-induced T cell hepatitis. As shown in Fig. 3B, treatment with IL-22 neutralizing antibodies markedly attenuated ConA-induced STAT3 activation in the liver without inhibition of STAT1 and ERK1/2 activation. Densitometry analysis showed that neutralizing IL-22 caused an approximate 50% and 40% reduction of STAT3 activation at 1 hour and 3 hours after injection of ConA, respectively (Fig. 3C). In contrast, IL-22 blockade did not affect ConA-induced STAT3 activation in the spleen (Fig. 3D). This suggests that IL-22 is not involved in STAT3 activation in the spleen after injection of ConA, probably because the spleen does not express IL-22R.16, 18

IL-22 Is a Survival Factor for Hepatocytes.

To investigate the antiapoptotic effects of IL-22 in hepatocytes, we generated a line of HepG2 cells stably transfected with IL-22 cDNA. As shown in Fig. 4A, 3 positive clones were selected based on their high IL-22 protein expression levels. RT-PCR also showed that HepG2 cells and stably transfected HepG2 cells expressed high levels of IL-22R mRNA and IL-10Rβ mRNA. Next, the effects of IL-22 on serum starvation-induced apoptosis of HepG2 cells were examined. As shown in Fig. 4B, visual observations by light microscopy showed that serum starvation for 3 days caused death in 90% of HepG2 cells transfected with empty vector (G2-neo). In G2-neo cells treated with recombinant IL-22, serum starvation-induced cell death was significantly reduced to about 70%. G2-IL-22-4 and G2-IL-22-5 cells, which were stably transfected with IL-22, were completely resistant to serum starvation-induced cell death; less than 5% of cell death was observed. These observations were further confirmed by the trypan blue exclusion test (Fig. 4C), which showed that after 3 days in serum-free medium, more than 90% of G2-neo cells died, but less than 10% of HepG2 cells stably transfected with IL-22 were even affected. Furthermore, to determine whether cell death was attributable to apoptosis, FACS analysis and DNA fragmentation were performed. As shown in Fig. 4D, after 3 days of serum starvation, 53.24% of G2-neo cells were apoptotic (M1, sub-G1 phase). G2-neo cells treated with IL-22 reduced the number of apoptotic cells to 40.64%, and only 6.35% and 19.75% of G2-IL-22-4 and G2-IL-22-5 cells, respectively, were apoptotic. In serum-starved G2-neo cells, significant DNA fragmentation was detected; this was markedly attenuated by the addition of IL-22 (Fig. 4E). DNA fragmentation was completely absent in G2-IL-22-4 and G2-IL-22-5 cells (Fig. 4E).

Figure 4.

IL-22 protects against serum starvation-induced HepG2 cell apoptosis. (A) HepG2 cells were stably transfected with IL-22 cDNA. Three positive clones were selected. Cell extracts and total RNA from empty vector-transfected clones and IL-22-transfected clones were subjected to Western blot analyses (for IL-22 and β-actin) and RT-PCR analyses (for IL-22R, IL-10Rβ, and β-actin), respectively. (B-E) Empty vector-transfected clones and IL-22-transfected clones were serum-starved for several days; cell death was determined by (B) light microscopy, (C) trypan blue exclusion test, (D) FACS analysis of propium iodide staining, and (E) DNA fragmentation. Recombinant IL-22 (10 ng/mL) was also included in some wells of serum-starved, empty vector-transfected HepG2 cells. (D) M1 (sub-G1 peak) represents apoptotic cells. Data in panels B-E are representative of 3 or more independent experiments.

IL-22 Promotes Growth of HepG2 and Hep3B Cells in Soft Agar and Tumor Formation in Nude Mice.

The effects of IL-22 overexpression on anchorage-independent growth were tested on soft agar. After 5 weeks of growth, G2-IL-22-5, G2-IL-22-4, and G2-IL-22-27 yielded much larger colonies, and in higher numbers, than G2-neo cells (Fig. 5A). Similarly, Hep3B cells transfected with IL-22 yielded much higher numbers of colonies after 6 weeks of growth on soft agar compared to 3B-neo cells (Fig. 5B).

Figure 5.

Growth of HepG2 and Hep3B cells stably transfected with IL-22 in soft agar and nude mice. (A) HepG2-neo cells and the 3 clones of HepG2 cells stably transfected with IL-22 (G2-IL-22-5, G2-IL-4, and G2-IL-22-27) were grown in soft agar for 21 days. The number of colonies was scored. (B) Hep3B-neo cells and the three clones of Hep3B cells stably transfected with IL-22 (3B-IL-22-26, 3B-IL-22-23, and 3B-IL-22-7) were grown in soft agar for 35 days, and the number of colonies was scored. (A and B) Bars represent means ± SEM from triplicates in 1 experiment. Three independent experiments were repeated and similar results were obtained. (C) Nude mice were implanted with 1 × 106 of HepG2-neo, G2-IL-22-5, and G2-IL-4 cells (10 mice for each group). Five weeks later, tumors were dissected and weighed. (D) Protein extracts were prepared from the tumor tissues and subjected to Western blot analysis using various antibodies as indicated.

The tumorigenicity of HepG2 cells transfected with IL-22 was further examined in nude mice by injection of 106 cells. Two weeks after injection, 9 of 10 mice injected with G2-IL-22-4 cells and 7 of 10 mice injected with G2-IL-22-5 cells had tangible tumors at the injection site, whereas only 2 of 10 mice injected with G2-neo cells had tumors. After 5 weeks, mice injected with G2-IL-22-4 cells and G2-IL-22-5 cells had much bigger tumors than those injected with G2-neo cells (Fig. 5C). Western blot analysis showed that high levels of IL-22 were detected in tumors from mice injected with G2-IL-22-4 and G2-IL-22-5 cells but not in mice injected with G2-neo cells (Fig. 5D). Interestingly, tumors from G2-IL-22-4–injected mice had higher levels of IL-22 and larger tumors than those from G2-IL-22-5 cells, suggesting that tumor sizes correlated with levels of IL-22. Furthermore, higher levels of constitutively activated STAT3, Bcl-xL, and Rb2 proteins were detected in tumors from G2-IL-22-4 and G2-IL-22-5 cells compared to those from G2-neo cells. Similarly, injection of Hep3B cells transfected with IL-22 had much higher incidence of tumors compared to injection of 3B-neo cells (data not shown).

Effects of IL-22 on Hepatocyte Proliferation.

During the course of this study, we noticed that HepG2 cells stably transfected with IL-22 grew much faster than G2-neo cells cultured in the medium containing 10% serum. As shown in Fig. 6A, the doubling time for G2-neo cells in serum-containing medium was 18 hours, whereas G2-IL-22-4 and G2-IL-22-5 required 11 hours and 10 hours, respectively, to double in number. Furthermore, significantly higher BrdU uptake was detected in HepG2 cells stably transfected with IL-22 compared to HepG2-neo cells, while similar levels of LDH release were detected in the supernatants of these cells (Fig. 6B).

Figure 6.

The mitogenic effect of IL-22 from stably transfected HepG2 cells and recombinant IL-22 on hepatocytes. (A-B) G2-neo, G2-IL-22-4, G2-IL-22-5, and G2-IL-22-27 were cultured in growth medium containing 10% serum. (A) After various time periods, cell numbers were counted. (B) BrdU labeling and LDH release were measured 1 day after culture. (C) HepG2 cells, Hep3B cells, and primary mouse hepatocytes were serum-starved for 4 hours, followed by the addition of 0.2% FBS and various concentrations of IL-22 or epidermal growth factor (EGF). After 24 hours, [3H]thymidine was added for an additional 2 hours, and [3H]thymidine uptake was measured. Bars represent means ± SEM from 20 independent experiments for HepG2 and Hep3B cells, and from 5 independent experiments for primary hepatoctyes. *P < .05; **P < .01; ***P < .001 versus the nontreated group.

The effect of recombinant IL-22 on hepatocyte proliferation in vitro was examined. Unlike the mitogenic effects of IL-22 overexpression in stably transfected HepG2 cells, which were very consistent and repeatable (Figs. 6A and B), the effects of recombinant IL-22 on hepatocyte proliferation were variable. Sometimes, it was observed that treatment with recombinant IL-22 significantly stimulated [3H]thymidine uptake in hepatocytes; at other times, only marginal effects were observed. In order to see the true effects of IL-22 on hepatocyte proliferation, a large number of independent experiments were performed. Fig. 6C is a summary of 20 independent experiments for HepG2 and Hep3B cells, and 5 independent experiments for primary mouse hepatoctyes, which showed that IL-22 induced significant [3H]thymidine uptake in HepG2 cells, Hep3B cells, and primary mouse hepatocytes. Treatment with IL-22 significantly stimulated cell proliferation of HepG2 and Hep3B cells with peak effect at 1 to 10 ng/mL, and IL-22 stimulated proliferation of primary mouse hepatocytes with peak effect at 10 to 50 ng/mL (Fig. 6C). The mitogenic effect of IL-22 in primary mouse hepatocytes was less effective compared to epidermal growth factor stimulation. The effects of recombinant IL-22 on BrdU labeling in hepatocytes were also examined, but only a marginal effect was obtained (data not shown).

The Antiapoptotic and Mitogenic Functions of IL-22 Are Mediated by Activation of STAT3 and Subsequent Induction of Antiapoptotic and Proliferation-Associated Proteins.

As shown in Fig. 7A, IL-22 treatment induced strong activation of STAT3 in both HepG2 and Hep3B cells, and a modest activation of STAT3 in primary mouse hepatocytes. A weak activation of ERK1/2 and STAT1 was also observed. In contrast, IL-22 induced a very weak or undetectable activation of p38 mitogen-activated (MAP) kinase and c-Jun NH2-terminal kinase (JNK) in these cells (data not shown).

Figure 7.

IL-22 protects against apoptosis and promotes cell proliferation of hepatocytes in a STAT3-dependent manner. (A) HepG2 cells, Hep3B cells, and primary mouse hepatocytes (PMH) were serum starved overnight, followed by stimulation with IL-22 (10 ng/mL) for various time points as indicated. Cell extracts were then prepared and subjected to Western blotting using various antibodies. (B) G2-neo and 3 clones of HepG2 cells stably transfected with IL-22 (G2-IL-22-4, G2-IL-22-5, and G2-IL-22-27) were cultured in serum-free medium for 1 or 2 days. Cell extracts were then prepared and subjected to Western blot analysis using various antibodies. (C) Serum-starved HepG2 cells were treated with the STAT3 inhibitor peptide (Ac-PYpLKTK-OH) (ST3 inh) for 30 minutes, followed by IL-22 (10 μg/mL) stimulation for 30 minutes. Gel shift assay was then performed to analyze STAT3 binding. Ten μg liver nuclear extracts were used. In lane 2, 100-fold cold probe was included in the binding reaction. In lane 5, HepG2 cells were treated with 200 μmol STAT3 inhibitor peptide. (D) G2-neo, G2-IL-22-4 (upper panel), and G2-IL-22-5 (lower panel) cells were cultured in growth medium containing 10% serum. After changing to serum-free medium, STAT3 inhibitor (open triangle) 160 μmol or (open circle) 80 μmol) was added. Viable cells were determined by the trypan blue exclusion test.

Next, we examined activation of signaling pathways in HepG2 cells stably transfected with IL-22. To rule out the effect of serum, signaling pathways were examined after 1 to 2 days of serum starvation. As shown in Fig. 7B, high levels of constitutively activated STAT3 were detected in all 3 clones of stably transfected HepG2 cells compared to G2-neo cells, while constitutive activation of ERK1/2, p38 MAP kinase, JNK, and protein kinase B (Akt) were not detected in these stably transfected cells (data not shown). To define the role of STAT3 in the antiapoptotic and mitogenic effects of IL-22, we used a cell-permeable analog of the STAT3-SH2 domain-binding phosphopeptide that has been shown to act as a highly selective and potent blocker of STAT3 activation.21, 22 As demonstrated by the gel mobility shift assay in Fig. 7C, the inhibitor efficiently blocked IL-22 activation of STAT3 in HepG2 cells (lane 5 vs. lanes 3 and 4). Fig. 7D showed that G2-neo cells started to die and stopped proliferating after serum starvation, but G2-IL-22-4 and G2-IL-22-5 cells continued to grow and survive. Treatment with a phospho-STAT3 inhibitor peptide caused cell death and suppressed cell proliferation of G2-IL-22-4 and G2-IL-22-5 cells (Fig. 7D). This suggests that constitutive activation of STAT3 is responsible for the antiapoptotic and mitogenic actions of IL-22 in hepatic cells.

To further explore the mechanism underlying IL-22/STAT3-mediated antiapoptotic and mitogenic effects, expression of a variety of apoptosis- and proliferation-associated proteins were examined. As shown in Fig. 7B, several antiapoptotic proteins (Bcl-xL, Bcl-2, Mcl-1) and several mitogenic proteins (c-myc, cyclin D1, Rb2, CDK4) were markedly induced in stably transfected HepG2 cells 1 and 2 days after serum starvation compared to induction in G2-neo cells. Expression of CDK2 and Fas ligand remained unchanged in HepG2 cells transfected with IL-22 compared to G2-neo cells. The p21cip1/WAF protein was also induced in stably transfected HepG2 clones after 1 day of serum starvation (Fig. 7B).

Discussion

In the present study, we demonstrated that IL-22 expression was significantly induced in T cell-mediated hepatitis (Fig. 1) and that IL-22 blockade markedly enhanced liver injury in this model, while administration of recombinant IL-22 prevented ConA-induced liver injury (Fig. 2). These findings suggest that IL-22 acts as a protective cytokine to attenuate liver injury in T cell-mediated hepatitis. Overexpression of IL-22 in human HepG2 hepatoma cells completely prevented serum starvation-induced cell apoptosis (Fig. 4); this indicates that IL-22 is a survival factor for hepatocytes in vitro. The fact that recombinant IL-22 also protects hepatocytes from serum starvation-induced cell death further supports this conclusion (Fig. 4). Previously, we showed that STAT3 was rapidly activated after injection of ConA and that IL-6 was only partially responsible for this activation since significant STAT3 activation was still detected in IL-6-deficient mice after injection of ConA.24 In the present study, we showed that injection of IL-22 rapidly induced STAT3 activation in the liver, and IL-22 blockade significantly reduced hepatic STAT3 activation in T cell-mediated hepatitis (Fig. 3), indicating that IL-22 is also partially responsible for hepatic STAT3 activation in this model. Activation of STAT3 by IL-6 has previously been implicated in the ability of IL-6 to protect against various forms of liver injury.24–34 Thus, activation of STAT3 by IL-22 is likely responsible for the protective role of IL-22 in hepatoctyes. Indeed, blocking STAT3 with a STAT3 phosphopeptide markedly blocked the antiapoptotic effects of IL-22 (Fig. 7D). It has been suggested that the antiapoptotic actions of STAT3 are mediated by induction of several downstream genes, including Bcl-xL, Bcl-2, and Mcl-1,34 which may also be important downstream factors in the protective role of IL-22 in the liver because overexpression of IL-22 significantly induced expression of these genes in hepatocytes (Fig. 7B).

HepG2 cells stably transfected with IL-22 grew much faster than control G2-neo cells (Figs. 5 and 6A), suggesting that IL-22 promotes hepatocyte proliferation. Western blot analysis in Fig. 7B showed that HepG2 cells stably transfected with IL-22 expressed high levels of cyclin D1 and c-myc, which may contribute to the mitogenic action of IL-22 in hepatocytes. It has been shown that overexpression of cyclin D1 promotes hepatocyte replication and growth in vitro35 and in vivo,36 and overcomes rapamycin- or protein deprivation-mediated suppression of hepatocyte proliferation.37, 38 The mitogenic effect of c-myc has been well documented in a wide variety of cell types, including hepatocytes.39, 40 Collectively, induction of both cyclin D1 and c-myc may play an important role in IL-22 stimulation of hepatocyte proliferation. Although we clearly demonstrated that overexpression of IL-22 cDNA stimulated HepG2 cell proliferation, the mitogenic effect of recombinant IL-22 on hepatocyte proliferation was variable, and the reasons for this are unclear. A summary of a large number of independent experiments showed that recombinant IL-22 induced significant [3H]thymidine uptake in HepG2 cells, Hep3B cells, and primary mouse hepatocytes (Fig. 6C). The effects of recombinant IL-22 on BrdU labeling in these cells were marginal (data not shown); this may be due to lesser sensitivity of the BrdU labeling assay compared to the [3H]thymidine uptake assay. The variable data on the effects of recombinant IL-22 on hepatocyte proliferation may be due to the instability of recombinant IL-22 in vitro, since the mitogenic effects of overexpression of IL-22 in HepG2 cells were consistent. Thus, recombinant IL-22 from different sources and different batches may contribute to these variable results in hepatocytes. Interestingly, the effect of IL-22 on HepG2 and Hep3B cell proliferation decreased with increasing cytokine concentration (Fig. 6C); however, the underlying mechanism is not clear. Activation of ERK1/2 has been shown to either promote or inhibit hepatocyte proliferation depending on the density of activated ERK1/2.41, 42 Similar to ERK1/2, activation of STAT3 has also been implicated in both stimulation of cell proliferation43, 44 and induction of cell cycle arrest.45, 46 The dual effects of STAT3 on cell proliferation may be due to STAT3 stimulation of both mitogenic genes (e.g., cyclin D1 and c-myc) and cell cycle arrest-related genes such as p21cip1/WAF (Fig. 7B). Taken together, the instability of recombinant IL-22 and induction of both mitogenic and cell cycle arrest-related genes may contribute to the variable effects of IL-22 on hepatocyte proliferation.

In summary, our findings suggest that the T cell-derived cytokine IL-22 plays a protective role against liver injury in ConA-induced T cell hepatitis, and IL-22 is a survival factor for hepatocytes. During hepatitis virus infection, T cell-mediated immune responses are generally believed to play a key role in viral clearance through direct cytotoxicity and release of antiviral cytokines (such as tumor necrosis factor α and interferon γ), and also play a central role in hepatocellular injury.1–3, 8, 9 Therefore, it is likely that T cell-mediated immune responses in viral hepatitis activate antiviral signal pathways that also play a central role in hepatocellular injury on the one hand, and, on the other hand, also produce the hepatoprotective cytokines (such as IL-22) to prevent and repair liver injury. T cell activation has also been implicated in hepatocellular injury in autoimmune hepatitis and alcoholic liver disease.5–7 It is therefore plausible that activated T cells may also play a protective role in these liver disorders by releasing IL-22. Additionally, we also demonstrate that hydrodynamic gene delivery of IL-22 protects the mouse liver from ConA, CCl4, and Fas ligand-induced injury,47 thus, IL-22 potentially could be a therapeutic drug to treat liver diseases such as acute liver failure. Although we demonstrate that constitutive overexpression of IL-22 enhances the tumorigenicity of hepatocellular carcinoma cells in vitro and in vivo, there is no evidence indicating that short-term treatment with IL-22 induces hepatocellular carcinoma. Thus, short-term treatment of liver disease with IL-22 should be safe.

Ancillary