A novel role for adiponectin in regulating the immune responses in chronic hepatitis C virus infection

Authors


  • Potential conflict of interest: Nothing to report.

  • Presented in part at the 2006 and 2007 Annual Meetings of the American Association for the Study of Liver Diseases.

Abstract

Adipose tissue releases pro-inflammatory and anti-inflammatory mediators, including adiponectin, which elicit a broad range of metabolic and immunological effects. The study aim was to determine in subjects infected with chronic hepatitis C virus (HCV) the effects of total adiponectin and its high-molecular-weight (HMW) and low-molecular-weight isoforms on HCV-specific immune responses. Serum levels of total adiponectin and its isoforms were determined by immunoassay. The ex vivo effect of adiponectin on the HCV-specific T-cell response was examined by interferon gamma (IFN-γ) enzyme-linked immunosorbent spot and enzyme-linked immunosorbent assay cytokine assays. The role of the mitogen-activated protein kinase (MAPK) signaling pathway in mediating the adiponectin effect on T cells was also evaluated. We found that serum levels of total and HMW adiponectin were significantly decreased in subjects with chronic HCV and increased body mass index (BMI) compared with HCV-infected lean subjects. The presence of an anti-HCV specific immune response was strongly associated with lower BMI (P = 0.004) and higher serum total (P = 0.01) and HMW (P = 0.02) adiponectin. In ex vivo assays, total adiponectin and the HMW adiponectin isoform enhanced HCV-specific IFN-γ production (P = 0.02 and 0.03, respectively). Adiponectin-R1 receptors were expressed on T cells and monocytes. In depletion experiments, the IFN-γ response to adiponectin was entirely dependent on the simultaneous presence of both CD4 and CD8 T cells, and to a lesser extent, natural killer cells. Selective inhibition of p38MAPK activity by SB203580 abrogated the IFN-γ response to adiponectin, whereas extracellular signal-regulated kinase 1/2 inhibition by PD98059 did not affect the response. Conclusion: In chronic HCV, a reciprocal association exists between BMI, adiponectin, and the anti-HCV immune responses, emphasizing the important role played by adiposity in regulating the immune response in HCV infection. (HEPATOLOGY 2008;48:374–384.)

Chronic hepatitis C virus (HCV)–related end-stage liver disease is the leading indication for liver transplantation and a rising cause of liver cancer in the Western world.1 There is increased awareness of the detrimental impact of increased body mass index (BMI) on the course of chronic HCV infection, with obese patients developing more aggressive liver injury and having resistance to antiviral therapy.2–4 The reasons for these poorer outcomes in obese patients are unknown.

In HCV infection, T lymphocytes and immunoregulatory cytokines are critical determinants of disease outcomes. In the chimpanzee model of HCV infection, high levels of interferon (IFN)-γ production by effector cells are associated with viral clearance.5, 6 In contrast, type 2 cytokines, such as interleukin (IL)-4 and IL-10, down-regulate the immune response and also modulate hepatic fibrogenesis.7, 8

Adipose tissue releases a variety of pro-inflammatory and anti-inflammatory cytokines, including the adipokines: adiponectin, leptin, and resistin. Adiponectin, a 30-kDa plasma protein, is the most abundant adipokine. Paradoxically, serum levels of adiponectin are decreased in subjects with increased BMI, insulin resistance, or type 2 diabetes. Circulating adiponectin exists in several isoforms, including low-molecular-weight (LMW), medium-molecular-weight, and high-molecular-weight (HMW) forms.9 Adiponectin is an important regulator of cytokine responses, and this effect is isoform-specific. Thus, HMW adiponectin increases IL-6 secretion in human monocytes and human monocytic leukemia cell line cells but does not suppress lipopolysaccharide (LPS)-induced IL-6 secretion. In contrast, LMW adiponectin reduces LPS-mediated IL-6 release and also stimulates IL-10 secretion.10 Because these cytokines play a central role in the inflammatory and immune responses in chronic HCV infection, the potential for adiponectin to alter HCV outcomes is evident.

The metabolic and inflammatory effects of adiponectin are broad, and some are mediated through the mitogen-activated protein kinase (MAPK) signaling pathway.11 Relevant to the theme of HCV infection, the MAPK pathway has been shown to play a crucial role in regulating IFN-α and IFN-γ signaling in viral infections.12, 13

Thus, the broad aims of this study were to evaluate the effect of adiponectin on the immune responses in patients with chronic HCV infection. Specifically, we aimed to: (1) examine the effect of adiponectin and its isoforms on the anti-HCV cellular immune response in vivo and ex vivo; (2) examine the effect of adiponectin and its isoforms on type 1 and type II cytokine production; and (3) determine whether the MAPK signaling pathway plays a role in mediating the effects of adiponectin on the immune response in chronic HCV.

Abbreviations

Adipo-R1, adiponectin receptor-R1; BMI, body mass index; CEF, cytomegalovirus, Epstein-Barr virus, and influenza (flu) virus; ELISA, enzyme-linked immunosorbent assay; ELISPOT, enzyme-linked immunosorbent spot; ERK1/2, extracellular signal-regulated kinase 1/2; HCV, hepatitis C virus; HMW, high molecular weight; IFN, interferon; IL, interleukin; LMW, low molecular weight; LPS, lipopolysaccharide; MAPK, mitogen-activated protein kinase; NK, natural killer; PBMC, peripheral blood mononuclear cell; PBS, phosphate-buffered saline; TNF-α tumor necrosis factor alpha.

Patients and Methods

Patients.

The study cohort included 35 patients with chronic HCV who underwent evaluation for antiviral therapy. Patients were excluded if they had human immunodeficiency virus or chronic hepatitis B coinfections. Each patient had a liver biopsy and a venous blood sample collected. Venous blood from four healthy donors was used for control studies. The study was approved by our institutional review board, and all patients provided written informed consent.

Measurement of Serum Total Adiponectin, Adiponectin Isoforms, Leptin, and Resistin.

For serum concentrations of total adiponectin and the adiponectin isoforms, LMW, medium molecular weight, and HMW were assayed using the adiponectin multimeric enzyme-linked immunosorbent assay (ELISA) kit (Alpco Diagnostics, Windham, NH) according to the manufacturer's protocol. Serum leptin and resistin levels were assayed using a multiplex immunoassay kit (Linco Research, Inc., St. Charles, MO). Readings were obtained on the Bio-PlexTM HTF system (Luminex Corp., Austin, TX).

Preparation of Adiponectin Isoforms.

For the experiments in this report, recombinant human adiponectin (R&D Systems, Minneapolis, MN) was used. The total recombinant adiponectin was expressed in a mouse myeloma cell line, NSO with endotoxin level <1.0 EU per microgram determined by the Limules amebocyte lysate (LAL) method. The HMW (>440 kDa) and LMW adiponectin isoforms were fractionated from total recombinant adiponectin using microsep 100K OMEGA centrifugal devices spun at 1000g for 10 to 20 minutes at room temperature. LMW fractions were further concentrated using microsep 1K devices. Separation was confirmed by western blot analysis following a 5% native polyacrylamide gel electrophoresis. Isoforms were detected with a primary monoclonal antibody to adiponectin (Axxora, San Diego, CA) and horseradish peroxidase conjugated rabbit antimouse secondary antibody. Signals were obtained using enhanced chemiluminescence technology according to the manufacturer's protocol (Amersham Biosciences, Germany).

IFN-γ Enzyme-Linked Immunosorbent Spot Assay.

Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Hypaque (Amersham Biosciences, NJ). Cells were either used fresh or cryopreserved in dimethylsulfoxide following standard procedures. HCV-specific T cell responses were determined by pulsing PBMCs with 10 pools of overlapping HCV peptides (18-mers overlapping by 11 to stimulate both CD4 and CD8 T cell responses) corresponding to each of the HCV proteins (from NIH AIDS Reference and Reagent Program) as previously described.14 IFN-γ production was also determined in response to stimulation of PBMCs with overlapping peptides from cytomegalovirus, Epstein-Barr, and influenza (flu) virus (CEF) (NIH AIDS R&R Program). The spots were counted with an AID enzyme-linked immunosorbent spot (ELISPOT) plate reader, version 3.5 (Strassberg, Germany). Mean triplicate data were analyzed and expressed as IFN-γ spot-forming units/106 cells after background subtraction. A positive IFN-γ response was taken as 25 spot-forming units/105 PBMCs and more than twice background (unstimulated PBMCs).

Cell Stimulation.

PBMCs were cultured in 96-well HTS plates (Millipore, Australia) and pulsed with overlapping HCV peptide pools with or without total adiponectin, LMW adiponectin, or HMW adiponectin at 10 μg/mL for 24 hours. The stimulated cells were washed three times in the culture medium before being used in ELISPOT or cytokine assays. To confirm viability and as a positive control, PBMCs were also cultured in separate wells with phytohemagglutinin 10 μg/mL.

Depletion of Lymphocyte Subpopulations and Flow Cytometric Analysis.

PBMCs were washed and resuspended in staining buffer composed of endotoxin free PBS, supplemented with 1 mM ethylenediaminetetraacetic acid and 10% fetal bovine serum. Cellular subpopulations were separated by a magnetic antibody labeling system (Miltenyi Biotech, NSW, Australia), using monoclonal antibodies directed against CD4 and CD8 (T cells), CD14 (monocytes), CD22 (B-lymphocytes), and CD56 (natural killer cells). Magnetic separations were conducted on columns placed in a MACS separator magnetic field (Miltenyi Biotec). Separated cell purity was confirmed by fluorescence-activated cell sorting analysis using fluorescence-labeled antibodies. The cells were first washed and suspended in phosphate-buffered saline (PBS)-0.5% bovine serum albumin and labeled with the following monoclonal antibodies: anti-CD14-antigen-presenting cells, anti-CD4-phycoerythrin, anti-CD8-phycoerythrin, and anti-CD56-fluorescein isothiocyanate (Becton Dickinson, Germany) at room temperature for 30 minutes. The cells were washed twice in PBS-0.5% bovine serum albumin and fixed in 1% paraformaldehyde made up in PBS-0.5% bovine serum albumin.

Cytokine ELISA Assays.

The cytokines IFN-γ, tumor necrosis factor alpha (TNF-α), IL-6, and IL-10 were detected in cultures of PBMCs stimulated with either HCV peptides alone, adiponectin, or HCV peptides in the presence of adiponectin by using DuoSet ELISA kits (R&D Systems).

Western Blots.

PBMCs were washed twice in cold PBS and lysed in extraction buffer. Buffers were supplemented with complete protease inhibitor (Roche Diagnostics, Victoria, Australia). Protein concentrations were determined by employing the Bradford Protein assay (Bio-Rad), denatured at 95°C in standard Laemmli buffer, separated on a 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis gel and electro-transferred (Bio-Rad) onto nitrocellulose membranes (Amersham Biosciences). Membranes were incubated overnight at 4°C with polyclonal antibody raised against adiponectin-R1 (adipo-R1) (Axxora, San Diego, CA), adiponectin-R2 (adipo-R2) (Alpha Diagnostic, San Antonio, TX), total p38MAPK (Cell Signalling Tech), active p38MAPK (Promega), or total extracellular signal-regulated kinase (ERK) 1/2 or active ERK1/2 (Cell Signalling Tech). Beta-actin antibody (Sigma) was housekeeping control where applicable. Membranes were then washed and incubated with horseradish peroxidase–conjugated goat anti-mouse (Alpha Diagnostic), or goat anti-rabbit antibodies (Bio-Rad) at room temperature for 1 hour. A dual-color precision plus protein standard (Bio-Rad) was used as size markers. Signals were detected by employing enhanced chemiluminescence technology according to instructions (Amersham, Germany).

Inhibition Assays.

PBMCs were pretreated with 20 μM PD98059 or 10 μM SB203850 for 30 minutes. As a negative vehicle control, cells were treated with dimethylsulfoxide at the percentage (vol/vol) for inhibitor treatment. Adiponectin was then added to stimulate the cells for 18 to 24 hours and functional activity assessed by ELISPOT assay.

Statistical Analysis.

Results are presented as means ± standard deviation. Unpaired Student t test was used to compare quantitative variables. The Pearson's test was used for testing the correlations between variables. Statistical analysis was conducted with SPSS 11.0 software (SPSS Inc, IL).

Results

Serum Total Adiponectin and HMW Adiponectin Levels Are Positively Associated with Anti-HCV Cellular Immune Responses.

We evaluated 35 chronic HCV treatment-naïve patients for an association between serum levels of total adiponectin, leptin, or resistin and the anti-HCV cellular immune response using the IFN-γ ELISPOT assay. The clinical characteristics of the group are shown in Table 1. A strong association was evident between serum adiponectin levels and the presence or absence of anti-HCV cellular immune responses, with a mean total adiponectin level of 6 (±2) μg/mL in those with HCV-specific T cell responses (n = 15) versus 4.2 (±2) μg/mL in those without (n = 20) (P = 0.01) (Table 1 and Fig. 1). Among the different adiponectin isoforms (Table 1), only HMW adiponectin was associated with a positive anti-HCV immune response mean HMW adiponectin 3 (±1.5) μg/mL versus 1.9 (±1.2) μg/mL in negative responders (P = 0.02).

Table 1. Differences in Clinical Characteristics of the Studied Cohort According to IFN-γ ELISPOT Response (Positive Anti-HCV Responders Versus Negative Anti-HCV Responders)
VariablePositive Anti-HCV Responders (n = 15)Negative Anti-HCV Responders (n = 20)P Value
  1. Abbreviation: HAI, hepatic activity index.

Sex (M:F)13: 217:3NS
Age (years)36 ± 836 ± 6NS
HCV RNA1.2 × 106 ± 2 × 1052 × 106 ± 3 ×105NS
Genotype 11215NS
Genotype 335 
ALT (U/L)89 ± 20129 ± 740.05
Liver histology score   
 Portal inflammation1.8 ± 0.762.26 ± 0.730.07
 Lobular inflammation1.5 ± 0.511.63 ± 0.55NS
 HAI3.27 ± 0.973.99 ± 0.930.06
 Fibrosis1.67 ± 1.042.0 ± 1.3NS
BMI (kg/m2)23 ± 2.527 ± 40.004
Adiponectin (μg/mL)   
 Total6 ± 24.2 ± 20.01
 HMW3 ± 1.51.9 ± 1.20.02
 MMW1.3 ± 0.51.2 ± 0.8NS
 LMW1.5 ± 0.51.4 ± 0.7NS
Leptin (ng/mL)5.5 ±3.07 ± 5.7NS
Resistin (ng/mL)3.7 ± 2.14.6 ± 6.9NS
Figure 1.

Relationship between BMI serum total and HMW adiponectin levels and the anti-HCV immune response. Anti-HCV T cell responses were assessed in 35 patients using the IFN-γ ELISPOT assay. (A) A strong association was evident between BMI and the presence (n = 15) or absence (n = 20) of anti-HCV cellular immune responses (P = 0.004). (B) Increased serum total and HMW adiponectin levels were also found in patients with positive anti-HCV T cell responses (P = 0.01 and P = 0.02, respectively).

High BMI Is Negatively Associated with Anti-HCV Cellular Immune Response.

HCV-infected subjects with positive anti-HCV cellular immune responses had significantly lower BMIs of 23 (±2.5) kg/m2 compared with negative responders with BMIs of 27 (±4) kg/m2 (P = 0.004) (Table 1). Age, sex, and HCV RNA levels were not different between subjects with positive and negative HCV immune responses. The mean serum alanine aminotransferase level, however, was lower in subjects with a positive anti-HCV immune response (89 ± 20 U/L) versus negative anti-HCV immune response (129 ± 78 U/L) (P = 0.05). Similarly, comparison of necroinflammatory activity in liver biopsies showed that subjects with positive anti-HCV immune responses tended to have a lower hepatic activity index (3.27 ± 0.97) compared with negative anti-HCV immune responders (3.99 ± 0.93) (P = 0.06).

Relationship Between BMI, Serum Adiponectin Levels, and Anti-HCV Cellular Immune Responses.

BMI correlated negatively with serum levels of both total adiponectin (r = −0.5, P = 0.004) and HMW adiponectin (r = −0.42, P = 0.012). A forward stepwise logistic regression was then performed to assess the relationship between age, sex, HCV genotype, BMI, total adiponectin, hepatic histology, and the anti-HCV immune response. BMI was the most significant predictor of the anti-HCV immune response (P = 0.02, odds ratio 1.54, 95% confidence interval 1.1-2.2), followed by total adiponectin (P = 0.05, odds ratio 2.4, 95% confidence interval 0.95-3.8). Adiponectin and its isoforms did not correlate with histological markers of liver injury.

Total Adiponectin Induces IFN-γ Production by PBMCs from Subjects with Chronic HCV.

Furthermore, we aimed to assess whether the observed association between adiponectin and the anti-HCV cellular immune response is an epiphenomenon or a direct effect of adiponectin on the peripheral immune response. We examined IFN-γ production in ELISPOT assays from HCV-infected patients with recombinant human adiponectin (n = 5) and healthy controls (n = 3). Among the five HCV subjects, three were infected with HCV genotype 1 and two others with genotype 3. The mean BMI of the HCV subjects was 25 (±3) kg/m2 compared with normal controls of 23 (±2) kg/m2. In the HCV group, the mean serum adiponectin level was 5.6 (±1.94) μg/mL versus controls of 5.8 (±1.72) μg/mL. IFN-γ ELISPOT responses were determined after stimulation of PBMCs with adiponectin alone, overlapping HCV peptides or, both. To examine the antigen specificity of the phenomenon, IFN-γ ELISPOT responses were also measured in response to PBMC stimulation with CEF peptides with or without adiponectin. We observed that in HCV-infected individuals, stimulation of PBMCs with adiponectin alone resulted in a substantial induction of IFN-γ production (Fig. 2). However, in the presence of both HCV peptides and adiponectin, IFN-γ production was significantly increased compared with stimulation with either HCV peptides or adiponectin alone (P < 0.05) (Fig. 2). A similar effect was evident, although to a lesser degree, in samples stimulated by CEF peptides and adiponectin (Fig. 2, column 5) and in PBMCs from normal subjects (data not shown).

Figure 2.

Adiponectin enhances anti-HCV IFN-γ production. Stimulation of PBMCs with adiponectin alone resulted in a substantial induction of IFN-γ production (column 1). In the presence of both HCV peptides and adiponectin (column 3), IFN-γ production was significantly increased compared with stimulation with either HCV peptides alone (column 2) or adiponectin alone (column 1) (*P < 0.05). Stimulation of PBMCs with CEF peptides also enhanced IFN-γ production (column 4). Adiponectin, however, did not significantly enhance the effect of CEF peptides (column 5)

In dose–response experiments, adiponectin concentrations below 5 μg/mL elicited no response, and the peak effect was evident at 10 μg/mL (data not shown). To confirm that the observed effect was attributable to adiponectin, the experiments were repeated after inactivation of adiponectin by treatment with proteinase K, which completely abolished the IFN-γ response (data not shown). Moreover, the potential contribution of endotoxin contamination of the recombinant adiponectin was excluded by neutralization with anti-CD14 (Beckman Coulter) and anti-TLR4 antibodies (Imgenex Corp) (data not shown).

HMW But Not LMW Adiponectin Induces IFN-γ Production by PBMCs from Subjects with Chronic HCV.

Next, we sought to dissect out in vitro the effect of various adiponectin isoforms on these responses. HMW and LMW adiponectin isoforms were prepared and the purity of each confirmed by western blot (Fig. 3). HMW, but not LMW, adiponectin enhanced IFN-γ production by PBMCs when used alone or in combination with HCV peptides (Fig. 4, columns 2 and 4, respectively). These observations are consistent with the in vivo finding of a positive association between the level of HMW adiponectin and the anti-HCV immune response.

Figure 3.

Western blot of recombinant full-length adiponectin and its isoforms. HMW and LMW adiponectin isoforms were resolved from recombinant human full-length (total) adiponectin as outlined in the Patients and Methods section. Lanes: (A) total adiponectin, (B) HMW adiponectin isoforms (>440 kDa), and (C) LMW adiponectin isoform (140 kDa). Molecular weight markers (kDa) are shown at left.

Figure 4.

HMW adiponectin enhances anti-HCV IFN-γ production. Stimulation of PBMCs with either HCV peptides alone (column 1) or HMW adiponectin (column 2) but not LMW adiponectin (column 3) induced substantial IFN-γ production. In the presence of HCV peptides and HMW adiponectin (column 4), IFN-γ production was significantly increased compared with other groups (*P < 0.05). The addition of LMW adiponectin to HCV peptides did not alter IFN-γ production (column 5).

Leukocyte Subpopulations Express Adipo-R1.

Western blots of protein extracts from leukocyte subpopulations (CD4, CD8, and monocytes) were performed with polyclonal antibodies against adipo-R1 and adipo-R2. Adipo-R1 (approximately 47.5 kDa) was expressed in each of the cell populations (Fig. 5). Adipo-R2 was not detected on leukocytes but was abundant on protein extracts of adipose tissue (Fig. 5).

Figure 5.

Adipo-R1 and adipo-R2 expression on leukocyte subpopulations. Expression of (A) Adipo-R1 (47.5 kDa) and (B) Adipo-R2 (35.4 kDa) on various cell populations. Lanes: (a) adipose tissue, (b) PBMCs, (c) purified CD4 T cells, (d) purified CD8 T cells, (e) purified monocytes, and (f) human monocytic leukemia cell line. Beta-actin served as loading control. Adipo-R1 was expressed on each of the cell populations. Adipo-R2 was not detected on leukocytes but was abundant on adipose tissue protein extracts

T Cells Are the Cellular Source of IFN-γ in Response to Adiponectin Stimulation.

We then sought to determine the cellular basis of the adiponectin-induced IFN-γ response. Purified CD4 and CD8 T cell populations were isolated from PBMCs of patients with chronic HCV (n = 5). IFN-γ responses from the sorted populations after total adiponectin stimulation were determined by ELISPOT. We observed that neither purified CD4 nor CD8 T cells from HCV-infected individuals were able to produce IFN-γ in response to adiponectin stimulation (Fig. 6). Compared with nondepleted PBMCs, depletion of either CD4 or CD8 T cells resulted in a significant reduction in IFN-γ production in response to adiponectin (Fig. 6, columns 2 and 3) (P < 0.01 for each). Importantly, simultaneous depletion of CD4 and CD8 T cells resulted in complete inhibition of the IFN-γ response to adiponectin (Fig. 6, column 4).

Figure 6.

T cells are the cellular source of IFN-γ response to adiponectin. Stimulation of PBMCs with adiponectin enhanced the IFN-γ response (column 1). Depleting CD4 or CD8 T cells from the PBMC pool attenuated the IFN-γ response to adiponectin, but did not abolish it (columns 2 and 3, respectively) (P < 0.01 for each). In contrast, simultaneous depletion of both CD4 and CD8 T cells from the PBMC pool completely abrogated the IFN-γ response to adiponectin (column 4).

Natural Killer Cells Also Support the T Cell Response to Adiponectin.

Because T cells alone were unable to respond to adiponectin yet their complete depletion resulted in abrogation of the IFN-γ response, we proceeded to characterize the key leukocyte subpopulation(s) involved in supporting the T cell response to adiponectin. IFN-γ ELISPOT assays were performed after a series of depletion experiments examining the role of monocytes, B cells, and natural killer (NK) cells. Depletion of CD14 (monocytes) or CD22 (mature B cells) from PBMCs did not affect IFN-γ production. In contrast, depletion of the CD56 (NK cell) population significantly reduced IFN-γ production compared with nondepleted PBMCs (P = 0.03) (Fig. 7, column 2).

Figure 7.

NK cells support IFN-γ response to adiponectin. Depleting PBMCs of NK cells resulted in reduced IFN-γ response (column 2) compared with nondepleted cells (column 1) (*P = 0.03). Replenishing PBMCs with NK cells restored the IFN-γ response to adiponectin (column 3). However, NK cells alone did not respond to adiponectin stimulation (column 4).

Adiponectin Affects the Expression of Other Cytokines in Subjects with Chronic HCV.

To further elucidate the effect of adiponectin on the cellular immune response, we evaluated the effect of adiponectin on type 1 and type 2 cytokine profiles in PBMCs from patients with chronic HCV (n = 10). The studied group mean BMI was 23 (± 2) and mean serum adiponectin level was 4 (± 1.9) μg/mL. Six subjects were infected with HCV genotype 1, whereas four had genotype 3. PBMCs were stimulated with either adiponectin alone or HCV peptides alone, or were pretreated with adiponectin (10 μg/mL) before stimulation with HCV peptides. Stimulation of PBMCs with HCV peptides enhanced the production of these cytokines, whereas adiponectin alone enhanced substantially the production of IFN-γ and to a lesser extent IL-6, TNF-α, and IL-10 (Fig. 8). Consistent with the ELISPOT data set, adiponectin pretreatment resulted in a significant increase in the HCV peptide-induced expression of IFN-γ (P = 0.03). In contrast, adiponectin pretreatment attenuated the HCV-induced IL-6 response (P = 0.04), but did not significantly affect either TNF-α or IL-10 secretion (Fig. 8). No specific clinical variables (age, sex, BMI, serum adiponectin, or HCV genotype) predicted the effect of adiponectin on cytokine responses.

Figure 8.

Adiponectin affects the expression of other anti-HCV specific cytokines. PBMCs from HCV-infected individuals (n = 10) were stimulated with either adiponectin (10 μg/mL) or HCV peptides alone, or were pretreated with adiponectin before stimulation with HCV peptides. Stimulation with either HCV peptides or adiponectin alone enhanced the production of IFN-γ, IL-6, TNF-α, and IL-10. Adiponectin pretreatment before HCV peptide stimulation significantly increased IFN-γ expression (*P = 0.03). In contrast, adiponectin pretreatment attenuated the HCV-induced IL-6 response (*P = 0.04), but did not affect either TNF-α or IL-10 secretion.

Adiponectin Activates the p38MAPK Signaling Pathway.

To determine the molecular pathways involved in mediating the effect of adiponectin on the immune response in chronic HCV, we examined the MAPK signaling pathways (ERK1/2 and p38MAPK), which commonly mediate the metabolic effects of adiponectin. We initially evaluated the expression of total ERK1/2 and p38MAPK in PBMCs and the induction of ERK1/2 and p38MAPK phosphorylation (indicative of activation) in response to adiponectin. Total ERK1/2 and total p38MAPK were detected in PBMCs by western blotting. Adiponectin induced phosphorylation of p38MAPK but not ERK1/2 (Fig. 9). Phosphorylation of p38MAPK occurred within 5 minutes of PBMC stimulation by adiponectin; the effect peaked at 30 minutes and was sustained at 60 minutes after stimulation (Fig. 9).

Figure 9.

Adiponectin activates p38MAPK but not ERK1/2. (A) Total p38MAPK was detected in PBMCs. Stimulation of PBMCs with adiponectin 10 μg/mL led to the activation (phosphorylation) of p38MAPK (P-p38MAPK). (B) Total ERK1/2 was detected in PBMCs. Stimulation of PBMCs with adiponectin (10 μg/mL) was not associated with activation (phosphorylation) of ERK1/2 (P-ERK1/2). Positive control was PBMCs stimulated with 10 μg/mL LPS for 15 minutes.

P38MAPK Signaling Pathway Mediates IFN-γ Response to Adiponectin.

Next, we further elucidated the role of the p38MAPK pathway in mediating the effect of adiponectin on the immune response in chronic HCV infection. IFN-γ production in response to adiponectin stimulation of PBMCs was assessed in subjects with chronic HCV (n = 8) after selective inhibition of p38MAPK activity by the specific p38MAPK-inhibitor 4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4- pyridyl)-1 H-imidazole (SB203580; Sigma). Pre-incubation of PBMCs with SB203580 for 30 minutes before adiponectin stimulation significantly abrogated the IFN-γ response, inhibiting it completely in some cases (Fig. 10) (P < 0.001). There were no specific clinical factors (age, sex, BMI, serum adiponectin, or HCV genotype) that predicted the inhibitory response of SB203580.

Figure 10.

Selective inhibition of p38MAPK abrogates the IFN-γ response to adiponectin. IFN-γ response to adiponectin in PBMCs was assessed after pretreatment with either the p38MAPK inhibitor SB203850 (10 μM) for 30 minutes before adiponectin stimulation (white bars) or vehicle (black bars). Inhibition of p38MAPK variably decreased IFN-γ production in all patients. IFN-γ response to adiponectin in PBMCs was assessed after pretreatment with either the p38MAPK inhibitor SB203850 (10 μM) for 30 minutes before adiponectin stimulation (white bars) or vehicle (black bars). Inhibition of p38MAPK variably decreased IFN-γ production in all patients.

In contrast, treating PBMCs with an ERK1/2 inhibitor PD98059 (Calbiochem, CA) did not affect the IFN-γ response to adiponectin (data not shown).

Discussion

The current study provides novel information into the mechanism whereby increased BMI and adiponectin affect the anti-HCV-cellular immune response. Initially we observed that the effect of increased BMI on the immune response in vivo is in part due to reduced circulating levels of full-length total adiponectin and, in particular, the HMW adiponectin isoform. Second, we found that both total and HMW adiponectin increased production of the antiviral cytokine IFN-γ. Furthermore, T cells were the primary cellular source of the IFN-γ response, because removing T cell subpopulations completely abolished the IFN-γ response to adiponectin. However, because T cells in isolation were unable to produce an IFN-γ response to adiponectin, NK cells (but not monocytes or B cells) provide an accessory cell stimulus. The third finding was that adiponectin attenuated IL-6 production in response to HCV peptide stimulation in vitro, while increasing the IFN-γ response. Finally, the effect of adiponectin on the immune response in chronic HCV was in part mediated by the p38MAPK signaling pathway.

There is increasing evidence that interactions between the immune and metabolic systems can potentially affect the course of chronic HCV infection.15, 16 Generally obese subjects appear to be more susceptible to infections.17 Furthermore, obese subjects with chronic HCV are less likely to achieve viral clearance with conventional antiviral therapy and are more likely to develop progressive liver disease.2–7 The current report provides important insights into potential mechanisms for these effects. Thus, the impaired anti-HCV immune response in obese HCV subjects was inversely related to BMI and serum levels of both total and HMW adiponectin isoforms. Patients with impaired anti-HCV cellular immune responses tended to have higher serum alanine aminotransferase levels and increased hepatic necroinflammation. It is plausible that in those obese subjects with reduced adiponectin, more effective evasion by HCV of the host immune responses ensues, thereby favoring continued viral replication, ongoing liver inflammation, and impaired treatment responses. The intricate association between the specific anti-HCV immune responses and obesity-related factors (including adiponectin) are clearly crucial in driving liver injury and response to antiviral therapy in subjects with chronic HCV.

The precise effect of adiponectin on inflammatory responses is intriguing, as reports have described both anti-inflammatory and pro-inflammatory properties. Thus, anti-inflammatory effects attributed to adiponectin include the inhibition of TNF-α production and activity, inhibition of nuclear factor kappaB activation and induction of anti-inflammatory cytokines.18, 19 Also, in contrast to our study, in experimental models of systemic inflammation, adiponectin administration inhibited the production of IFN-γ in response to LPS.18 This inhibitory effect of adiponectin has been partly attributed to its ability to directly bind and possibly inactivate LPS.20 Thus, in adiponectin knockout mice, exogenous adiponectin did not affect LPS inflammatory activity and was not associated with alterations in either cytokine production or responsiveness to LPS.21 Pro-inflammatory effects of full-length adiponectin include increased production of IL-8 and monocyte chemotactic protein-1 (MCP-1) in human endothelial cells and monocytes.22 Similarly, in human synovial fibroblasts and colon tissue, adiponectin induced the production of inflammatory mediators such as matrix metalloproteinase-1 via the p38MAPK pathway.23 In addition, simultaneous anti-inflammatory and pro-inflammatory effects for adiponectin in the same cell line have been reported.24, 25 Thus, in macrophages, adiponectin induced TNF-α, which then increased IL-10 expression and eventual dampening of LPS-mediated inflammatory cytokine production.24 Taken together, this body of contrasting data emphasizes that adiponectin cannot be classified as either an entirely anti-inflammatory or pro-inflammatory molecule. The net effect of adiponectin largely depends on the individual isoform and on the particular cell system being studied.

In our hands, convergence data (ELISPOT and ELISA) confirmed that adiponectin increases the production of IFN-γ by PBMCs from HCV-infected subjects. Despite being a pro-inflammatory molecule, IFN-γ regulation has important ramifications in chronic HCV infection. First, IFN-γ is crucial in HCV infection in the establishment of an antiviral state both, by inhibiting viral replication26 and by the induction of proteins with systemic functions involved in regulating the innate and adaptive immune responses.27 Thus, the resolution of acute HCV infection has been shown to strongly correlate with increased IFN-γ–producing activated T cells. Second, in established chronic HCV infection, induction of IFN-γ and its related genes is a characteristic feature of those patients who achieve HCV clearance with antiviral therapy.28 Furthermore, the degree of viremia has been shown to correlate inversely with the expression of IFN-γ in the livers of HCV-infected persons.29 Finally, as previously discussed, obese subjects display impaired immune responses as evidenced by reduced IFN-γ production by PBMCs in response to LPS stimulation.30 Thus, considering our ex vivo finding of the positive effect of adiponectin on IFN-γ production by T cells in HCV subjects, one of the plausible explanations for the poorer treatment outcomes among obese HCV subjects is a defective IFN-γ response in conjunction with the observed low circulating adiponectin levels.

Further mechanistic studies showed that depletion of T cells completely abolished the production of IFN-γ in response to adiponectin. However, the T cell response to adiponectin required the simultaneous presence of accessory cells, with NK cells being implicated. This observation is in line with recent studies highlighting the role of direct cell–cell interactions between NK and T lymphocytes in supporting T cell proliferation and in maintaining IFN-γ production by CD4 T cells.31, 32 The defective cellular immunity observed in chronic HCV may be a consequence of inhibition of effective interactions between NK cells, dendritic cells, and T cells.33–35 This results in shifting cytokine profiles toward a type 2 response with resultant viral persistence and limited immune control of ongoing viral replication.34, 35 Furthermore, NK cells have been shown to play an important role in controlling HCV replication, again through an IFN-γ–dependent mechanism.36 As such, one of the effective mechanisms whereby HCV evades the immune response is through inhibiting IFN-γ production by NK cells. Thus, the observed T cell–NK cell interaction to produce IFN-γ in response to adiponectin, observed in the current study, may represent an important step in the interface between metabolic factors and the HCV immune response.

Another relevant observation from this study was the ability of adiponectin to attenuate the induction of IL-6 by HCV peptides. This is relevant because IL-6 plasma levels increase with increasing body fat content and are implicated in the pro-inflammatory state of obesity, leading to insulin resistance.37 Also, in chronic HCV, increased IL-6 production is proposed to promote a defective response to IFN therapy by inducing the expression of suppressor of cytokine signaling 3, a negative IFN regulator.38

The final observation in this report is that p38MAPK, a component of the MAPK signaling pathway, is implicated in mediating the effect of adiponectin on IFN-γ production. This observation is supported by other studies that demonstrated that the inflammatory responses in monocytic cell lines in response to full-length adiponectin isoforms are in part mediated by p38MAPK.39 In contrast, the convergence data from our study have shown no effect for the ERK1/2 component of the MAPK signaling pathway in mediating the IFN-γ response to adiponectin. The effect of adiponectin on various components of the MAPK signaling pathway has certainly varied among earlier studies.24, 40 These differences in MAPK signaling pathway responses to adiponectin are likely attributable to both the diverse cell lines studied with varying adipoR1 and adipoR2 expression and the biologically distinct adiponectin components examined. The MAPK signaling pathway is also important in HCV, because inhibition of the pathway by HCV proteins (including core and nonstructural protein-5a) promotes viral replication.41, 42 Moreover, p38MAPK signaling plays an important role in the antiviral responses mediated by IFN-α and IFN-γ. Thus, in an inducible 3T3-L1 clone, inhibition of p38MAPK signaling led to diminished IFN-γ–mediated protection against viral killing.12, 13

What are the clinical implications of these findings in obese patients with chronic HCV? Weight loss, in overweight subjects, is associated with reduced expression of inflammatory cytokines, including IL-6, increased serum levels of adiponectin, and reduced hepatic inflammation.43, 44 Similarly, direct administration of adiponectin in animal models of obesity and liver injury reduced the inflammatory responses and had hepatoprotective effects.45, 46 Thus, in obese subjects with chronic HCV, measures to reduce weight or increase adiponectin levels should impact favorably on the HCV immune response. Clearly this hypothesis is amenable to empirical testing in the clinical setting.

Taken together, the findings from the current study illustrate, for the first time, the importance of obesity and adiponectin in modulating the immune response in chronic HCV infection. Based on our findings, we propose that the detrimental effect of obesity in chronic HCV is, in part, attributable to the influence of adiponectin, particularly the HMW isoform, on cytokine responses in T cells and other immune cells. Overall, the unfolding role of adiposity as an important immune regulator in HCV infection provides new opportunities for developing more effective therapeutic approaches in these patients.

Ancillary

Advertisement