Programmed death-1/B7-H1 negative costimulation protects mouse liver against ischemia and reperfusion injury


  • Haofeng Ji,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Xiuda Shen,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Feng Gao,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Bibo Ke,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Maria Cecilia S. Freitas,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Yoichiro Uchida,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Ronald W. Busuttil,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Yuan Zhai,

    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    Search for more papers by this author
  • Jerzy W. Kupiec-Weglinski

    Corresponding author
    1. Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, Department of Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA
    • Dumont-UCLA Transplant Center, 77-120 CHS, 10833 Le Conte Avenue, Los Angeles, CA 90095
    Search for more papers by this author
    • fax: 310-267-2358

  • Supported by National Institutes of Health Grants RO1 DK062357 and AI23847 (to J.W.K.-W.) and The Dumont Research Foundation. H.J. is a recipient of the American Society of Transplantation Fellowship Grant.

  • Potential conflict of interest: Nothing to report.


Programmed death-1 (PD-1)/B7-H1 costimulation acts as a negative regulator of host alloimmune responses. Although CD4 T cells mediate innate immunity-dominated ischemia and reperfusion injury (IRI) in the liver, the underlying mechanisms remain to be elucidated. This study focused on the role of PD-1/B7-H1 negative signaling in liver IRI. We used an established mouse model of partial liver warm ischemia (90 minutes) followed by reperfusion (6 hours). Although disruption of PD-1 signaling after anti–B7-H1 monoclonal antibody treatment augmented hepatocellular damage, its stimulation following B7-H1 immunoglobulin (B7-H1Ig) fusion protected livers from IRI, as evidenced by low serum alanine aminotransferase levels and well-preserved liver architecture. The therapeutic potential of B7-H1 engagement was evident by diminished intrahepatic T lymphocyte, neutrophil, and macrophage infiltration/activation; reduced cell necrosis/apoptosis but enhanced anti-necrotic/apoptotic Bcl-2/Bcl-xl; and decreased proinflammatory chemokine/cytokine gene expression in parallel with selectively increased interleukin (IL)-10. Neutralization of IL-10 re-created liver IRI and rendered B7-H1Ig–treated hosts susceptible to IRI. These findings were confirmed in T cell–macrophage in vitro coculture in which B7-H1Ig diminished tumor necrosis factor-α/IL-6 levels in an IL-10–dependent manner. Our novel findings document the essential role of the PD-1/B7-H1 pathway in liver IRI. Conclusion: This study is the first to demonstrate that stimulating PD-1 signals ameliorated liver IRI by inhibiting T cell activation and Kupffer cell/macrophage function. Harnessing mechanisms of negative costimulation by PD-1 upon T cell–Kupffer cell cross-talk may be instrumental in the maintenance of hepatic homeostasis by minimizing organ damage and promoting IL-10–dependent cytoprotection. (HEPATOLOGY 2010.)

Liver ischemia and reperfusion injury (IRI), an exogenous antigen-independent inflammatory event, occurs in multiple clinical settings, including partial hepatectomy, trauma, and transplantation. IRI remains one of the most critical problems in liver transplant recipients, causing up to 10% of early graft failure, in turn leading to a higher incidence of acute and chronic rejection and contributing to acute donor liver shortage.1, 2 Although its mechanism has not been fully elucidated, IR-triggered generation of reactive oxygen species inflicts tissue damage, which initiates circulatory disturbances, local inflammation, cell death, and organ failure. In 2003, we proposed that liver damage due to reperfusion following prolonged ischemia should be considered as an innate immunity-dominated inflammation response.2 Our group was among the first to document that activation of Toll-like receptor (TLR) 4 was required for the induction of IR-triggered hepatic inflammation and damage.3 By releasing inflammatory mediators such as tumor necrosis factor-α (TNF-α), interleukin (IL)-6, and CXCL-10 downstream of TLR4 signaling, we have identified Kupffer cells as critical players in the mechanism of IRI.4, 5

In agreement with others,6 we have reported that T lymphocytes, particularly of the CD4 phenotype, represent the key mediators in IR-triggered liver IRI.7 We have highlighted the role of CD154 costimulation and documented the benefit of disrupting the CD154–CD40 pathway.8-10 Our studies have shown that CD4 T cells can function by way of CD154 without de novo antigen-specific activation, and innate immunity-induced CD40 may trigger CD154–CD40 engagement to facilitate tissue inflammation and injury.11 Our recent study has focused on the distinctive features of newly identified TIM-1–TIM-4 signaling in liver IRI.12 Collectively, these studies have documented a previously unrecognized mechanism through which a CD4 T cell–generated positive costimulation signal can amplify Kupffer cell activity/function and cross-talk to facilitate IR-triggered immune cascade.

Programmed death-1 (PD-1; CD279) is the CD28 homologue expressed selectively by activated T, B, and myeloid cells.13 When cross-linked with PD-L1 (B7-H1; CD274) on hemopoetic and many nonhemopoetic tissues, the PD-1/B7-H1 interaction delivers a potent negative signal that inhibits T and B cell activation and may promote immune tolerance. This study is the first to examine the putative role of PD-1/B7-H1 in the pathophysiology of liver IRI. Our results demonstrate that stimulating PD-1 negative signals ameliorates local inflammation and liver damage and suggest that engaging PD-1/B7-H1 costimulation is required for maintaining liver homeostasis during IR-induced insult.


AU, absorbance units; B7-H1Ig, B7-H1 immunoglobulin; BMM,bone marrow–derived macrophage; H&E, hematoxylin-eosin; IFN-γ, interferon-γ; IL, interleukin; IRI, ischemia and reperfusion injury; mAb, monoclonal antibody; PD-1, programmed death-1; sALT, serum alanine aminotransferase; TLR, Toll-like receptor; TNF-α, tumor necrosis factor α; TUNEL, terminal deoxynucleotidyl transferase –mediated dUTP nick-end labeling; WT, wild-type.

Materials and Methods


Male C57BL/6 wild-type (WT) mice (8-12 weeks old) (Jackson Laboratory, Bar Harbor, ME) were housed in the University of California Los Angeles animal facility under specific pathogen-free conditions and received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals (prepared by the National Academy of Sciences; National Institutes of Health publication 86-23, revised 1985).

Mouse Model of Liver IRI

We used a mouse model of warm partial hepatic IRI.3-5, 7-12 Mice were anesthetized, injected with heparin (100 U/kg intraperitoneally), and the arterial and portal venous blood supply to the cephalad lobes was interrupted by an atraumatic clip. After 90 minutes of local ischemia, the clip was removed. Animals were sacrificed after 6 or 24 hours of reperfusion. Sham-operated mice underwent the same procedure, but without vascular occlusion.

Protein and Monoclonal Antibody Therapy

Rat anti–B7-H1 monoclonal antibody (mAb) (10F.9G2; Bio X Cell, West Lebanon, NH), recombinant B7-H1 immunoglobulin (B7-H1Ig), a dimeric B7-H1 immunoglobulin fusion protein14 (courtesy of Dr. Xian C. Li, Harvard Medical School, Boston, MA), or control Ig (Bio X Cell) was administered intravenously prior to the onset of ischemia (0.25 mg at day −1 and 0.5 mg at day 0). Rat anti-mouse IL-10 mAb (JES5-2A5; Bio X Cell) was administered (0.5 mg at day −1 and 0.5 mg at day 0) with or without B7-H1Ig.

Hepatocellular Damage

Serum alanine aminotransferase (sALT) levels, an indicator of liver injury, were measured with an autoanalyzer (ANTECH Diagnostics, Los Angeles, CA).


Liver specimens (4 μm) were stained with hematoxylin-eosin (H&E) and then analyzed blindly by modified Suzuki's criteria as described.7-10


Primary mAb against mouse T cells CD3 (17A2; BD Biosciences, San Jose, CA), neutrophils Ly-6G (1A8; BD Biosciences) and macrophages F4/80 (FA-11; AbD Serotec, Raleigh, NC) were used as described.12 Liver sections were evaluated blindly by counting labeled cells in 10 high-power fields.

Myeloperoxidase Activity Assay

The presence of myeloperoxidase was used as an index of neutrophil accumulation in the liver.7-10 One absorbance unit of myeloperoxidase activity was defined as the quantity of enzyme degrading 1 mol peroxide per minute at 25°C per gram of tissue.

Quantitative Real-Time Polymerase Chain Reaction

Quantitative polymerase chain reaction was performed with a platinum SYBR green quantitative polymerase chain reaction kit (Invitrogen, Carlsbad, CA) using the Chromo4 detector (MJ Research, Waltham, MA). The primers used to amplify specific gene fragments are listed in Supporting Table 1. Target gene expressions were calculated by their ratios to the housekeeping gene hypoxanthine-guanine phosphoribosyl transferase.

Western Blots

Western blots were performed with liver proteins (30 μg/sample) and rabbit anti-mouse cleaved caspase-3, Bcl-2, Bcl-xl, and β-actin mAbs (Cell Signaling Technology, Danvers, MA) as described.8-10, 12 Relative quantities of protein were determined with a densitometer and are expressed in absorbance units (AU).

Terminal Deoxynucleotidyl Transferase–Mediated dUTP Nick-End Labeling Assay

DNA fragments in liver sections resulting from oncotic necrosis and apoptosis were detected by way of terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL) assay (In Situ Cell Death Detection Kit, Roche, Indianapolis, IN) as described.7-9, 12 TUNEL-positive cells were counted in 10 high-power fields/section under light microscopy (×400).

Cell Cultures

Spleen T cells from C57BL/6 mice were incubated for 24 hours by addition of anti-CD3 (145-2C11, BD Biosciences; 0.5 μg/mL) with B7-H1 or control Ig (20 μg/mL). Supernatants were evaluated for interferon-γ (IFN-γ)/IL-10 levels by way of enzyme-linked immunosorbent assay (eBioscience, San Diego, CA). Bone marrow–derived macrophages (BMMs) separated from the femurs and tibias of C57BL/6 mice were cultured (5 × 106/well) with 10% L929-conditioned medium for 6 days. The cell purity was assayed to be 94%-99% CD11b+. In some experiments, BMMs were cocultured with spleen T cells at responder/stimulator ratios of 1:5,12 incubated for 24 hours using anti-CD3 (0.5 μg/mL) with B7-H1Ig or control Ig ± anti–IL-10 mAb (20 μg/mL). Cell-free supernatants were assayed for TNF-α/IL-6 levels by enzyme-linked immunosorbent assay (eBioscience).

Statistical Analysis

All values are expressed as the mean ± SD. Data were analyzed with an unpaired, two-tailed Student t test. P < 0.05 was considered statistically significant.


PD-1/B7-H1 Signaling Prevents Hepatocellular Damage and Ameliorates Liver IRI

We analyzed the hepatocellular damage in our model of 90-minute partial liver warm ischemia followed by reperfusion. As shown in Fig. 1A, sALT levels increased as early as 1 hour of reperfusion, peaked at 6 hours, and decreased thereafter. To determine the function of PD-1/B7-H1 negative signaling in the acute phase of liver IRI in this model, WT mice treated with B7-H1Ig were subjected to 90 minutes of partial liver warm ischemia followed by 6 hours or 24 hours of reperfusion. Unlike WT mice given control Ig, those conditioned with B7-H1Ig were resistant against IR-mediated hepatocellular damage, as evidenced by reduced sALT levels (163 ± 30 U/L versus 845 ± 166 U/L [6 hours], P < 0.01; 65 ± 2 U/L versus 216 ± 113 U/L [24 hours], P < 0.05) (Fig. 1B). These data correlated with histological criteria of liver damage at the peak of 6 hours postreperfusion (Fig. 1C). Control livers revealed severe lobular edema, congestion, ballooning, and hepatocellular necrosis (Suzuki score 2.33 ± 0.29 [6 hours]). In contrast, the B7-H1Ig group showed well-preserved liver architecture and histological detail without edema, vacuolization, or necrosis (Suzuki score 0.33 ± 0.58 [6 hours], P < 0.01). To determine whether this effect was dependent on stimulation of PD-1/B7-H1, a separate group of WT mice was treated with anti–B7-H1 mAb. Indeed, PD-1/B7-H1 blockade re-created IR-triggered hepatocellular injury and augmented liver damage compared with controls, as evidenced by increased sALT levels (1,497 ± 164 U/L [6 hours], 737 ± 264 U/L [24 hours], P < 0.05) (Fig. 1B), and deterioration of liver histology (Fig. 1C), reflected by the Suzuki score (3.83 ± 0.29 [6 hours], P < 0.01).

Figure 1.

Distinct effects of blockade versus activation of the PD-1/B7-H1 negative costimulation pathway in liver IRI (after 90 minutes of warm ischemia). (A) Liver damage [sALT (U/L)] peaked at 6 hours of reperfusion. (B) Hepatocellular function [sALT (U/L)] was impaired after anti–B7-H1 mAb treatment (1497 ± 164 [6 hours], 737 ± 264 [24 hours], ***P < 0.05) but improved after B7-H1Ig treatment (163 ± 30 [6 hours], 65 ± 2 [24 hours], **P < 0.01, ***P < 0.05) compared with controls (845 ± 166 [6 hours], 216 ± 113 [24 hours], n = 6-8/group). These data correlated with (C) Liver histology (H&E staining; magnification ×100 and ×400) of ischemic liver lobes (6 hours).

B7-H1Ig Engagement Attenuates T Cell, Neutrophil, and Macrophage Infiltration

We performed immunohistochemical staining of liver infiltrating cells at 6 hours of reperfusion following 90 minutes of warm ischemia. Treatment with B7-H1Ig diminished the number of cells per high-power field for T cells (0.5 ± 0.58 versus 2.0 ± 0.82, P < 0.05) ( Fig. 2A), neutrophils (2.0 ± 1.0 versus 53.7 ± 3.2, P < 0.001) (Fig. 2B), and macrophages (4.3 ± 1.0 versus 78.0 ± 2.0, P < 0.001) (Fig. 2) in the ischemic liver lobe compared with controls. Myeloperoxidase assay revealed that liver neutrophil activity was also depressed in the treatment group compared with controls (0.03 ± 0.2 U/g versus 1.13 ± 0.3 U/g, P < 0.05) (Supporting Information Fig. 1).

Figure 2.

T lymphocytes, neutrophils, and macrophages in IR livers following activation of the PD-1/B7-H1 pathway. Diminished infiltration of (A) CD3+ T lymphocytes (0.5 ± 0.6 versus 2.0 ± 0.8, ***P < 0.05), (B) Ly-6G+ neutrophils (2.0 ± 1.0 versus 53.7 ± 3.2, *P < 0.001), and (C) F4/80+ macrophages (4.3 ± 1.0 versus 78.0 ± 2.0, *P < 0.001) in B7-H1Ig–treated liver lobes compared with controls (magnification ×100 and ×400).

B7-H1Ig Treatment Prevents IR-Triggered Liver Cytokine/Chemokine Programs

To assess the immunoregulatory mechanism of PD-1/B7-H1 activation, we contrasted intrahepatic T lymphocyte–related cytokine expression patterns in our model. B7-H1Ig significantly (P < 0.05) suppressed gene induction of Th1-type IFN-γ/granzyme B and largely abolished otherwise enhanced gene transcript levels of neutrophil/monocyte-derived proinflammatory chemokines (CXCL-1, CXCL-5, CCL-2, CXCL-10) and cytokines (TNF-α, IL-1β, IFN-β, IL-6) (P < 0.01) ( Fig. 3A,B). In contrast, Th2-type IL-10 but not IL-4 expression significantly (P < 0.01) increased after B7-H1Ig engagement (Fig. 3C).

Figure 3.

Quantitative real-time polymerase chain reaction-assisted detection of (A) chemokines (CXCL-10, CXCL-1, CXCL-5, and CCL-2), (B) cytokines (IFN-γ, granzyme B, TNF-α, IL-1β, IFN-β, IL-6), and (C) IL-4 and IL-10. Selectively increased IL-10 expression after activation of PD-1/B7-H1 signaling. Data were normalized to hypoxanthine-guanine phosphoribosyl transferase gene expression (**P < 0.01, ***P < 0.05, n = 6-8/group).

B7-H1Ig Inhibits IR-Induced Liver Necrosis and Apoptosis

We performed TUNEL assay to detect IR-triggered oncotic necrosis and apoptosis. B7-H1Ig treatment diminished otherwise abundant hepatocellular necrosis and apoptosis in IR-injured livers (2.3 ± 0.6 versus 38.0 ± 2.0; P < 0.001) ( Fig. 4A,B). In parallel, western blot analysis revealed selectively decreased expression (AU) of cleaved caspase-3 and increased anti-necrotic/apoptotic Bcl-2/Bcl-xl proteins in the B7-H1Ig group (control Ig versus B7-H1Ig: 1.84 ± 0.041 versus 0.07 ± 0.020 [cleaved caspase-3], 0.20 ± 0.081 versus 2.12 ± 0.086 [Bcl-2], 0.29 ± 0.064 versus 2.08 ± 0.120 [Bcl-xl]) (Fig. 4C).

Figure 4.

Necrotic/apoptotic and anti-necrotic/apoptotic pathways in IR livers following activation of PD-1/ B7-H1. (A) TUNEL-assisted detection of hepatic necrosis/apoptosis (arrows) in ischemic lobes (magnification ×400). (B) Lower frequency of hepatic TUNEL+ cells in the B7-H1Ig–treated group compared with controls (2.3 ± 0.6 versus 38.0 ± 2.0, *P < 0.001). (C) Western blot-assisted detection of cleaved caspase-3, Bcl-2, and Bcl-xl after B7-H1Ig treatment compared with controls (0.07 ± 0.020 AU versus 1.84 ± 0.041 AU [cleaved caspase-3], 2.12 ± 0.086 AU versus 0.20 ± 0.081 AU [Bcl-2], 2.08 ± 0.120 AU versus 0.29 ± 0.064 AU [Bcl-xl], **P < 0.01). Data are representative of n = 4.

IL-10 Neutralization Restores Liver IRI in B7-H1Ig–Treated Mice

As liver inflammation response to IR in B7-H1Ig–treated mice was characterized by selectively increased IL-10 ( Fig. 3C), the question of whether IL-10 played a cytoprotective function was addressed by neutralizing IL-10. Indeed, significant increase in liver injury was observed after infusion of B7-H1Ig–treated mice with anti–IL-10 mAb, as shown by sALT levels (1,656.7 ± 358 versus 163 ± 30 U/L after B7-H1Ig monotherapy, P < 0.001) (Fig. 5A) and liver histology (Fig. 5B). Livers in B7-H1Ig–treated mice in which IL-10 was neutralized were characterized by zonal/panlobular parenchyma necrosis (Suzuki score 3.88 ± 0.25), which was comparable with controls (Fig. 1B). Infusion of anti–IL-10 mAb triggered a significant (P < 0.01) increase in the inflammatory gene expression programs (CXCL-10, TNF-α, and IL-6). Thus, IL-10 neutralization re-created liver IRI, rendered B7-H1Ig–treated hosts susceptible to IR, and confirmed the pivotal cytoprotective role of IL-10 produced by B7-H1Ig engagement.

Figure 5.

Functional significance of IL-10 for liver cytoprotection after activation of the PD-1/ B7-H1 pathway. The neutralization of IL-10 restored liver injury in B7-H1Ig–treated mice, as evidenced by (A) sALT levels (1,656.7 ± 358 U/L versus 163 ± 30 U/L in the B7-H1Ig monotherapy group, **P < 0.01, n = 6/group), (B) liver histology (representative H&E staining [magnification ×100 and ×400], Suzuki score 3.88 ± 0.25 versus 0.33 ± 0.58 in the B7-H1Ig group [ Fig. 1B], **P < 0.01), and (C) increased CXCL-10, TNF-α, and IL-6 expression.

PD-1/B7-H1 Pathway Regulates T Lymphocyte–Macrophage Cross-talk

We analyzed the immunomodulatory function of PD-1/B7-H1 signaling in a well-controlled cell culture system, designed to mimic liver IRI. First, we screened anti-CD3 mAb-mediated activation of T cells with control Ig/B7-H1Ig by enzyme-linked immunosorbent assay ( Fig. 6A). Addition of B7-H1Ig decreased IFN-γ levels (88.3 ± 21 versus 1267.8 ± 30 pg/mL, P < 0.001) yet increased IL-10 levels (641.8 ± 42 versus 302.1 ± 72 pg/mL, P < 0.05) compared with control Ig cultures. These data confirm our in vivo finding (Fig. 3) that activation of the PD-1/B7-H1 pathway preferentially induces T cell–derived IL-10.

Figure 6.

The effects of B7-H1Ig on T cell/macrophage activation in vitro. (A) IFN-γ depression (88.3 ± 21 versus 1,267.8 ± 30 pg/mL) and IL-10 increase (641.8 ± 42 versus 302.1 ± 72 pg/mL) after treatment of anti-CD3–stimulated T cells with B7-H1Ig (*P < 0.001, n = 3/group). (B) Up-regulated TNF-α/IL-6 production in anti-CD3–stimulated T cell + BMM cocultures. Addition of B7-H1Ig suppressed TNF-α (62.0 ± 6 versus 174.6 ± 11 pg/mL) and IL-6 (129.2 ± 8 versus 653.4 ± 7 pg/mL) levels (**P < 0.01, n = 3/group). Concomitant neutralization of IL-10 re-created TNF-α (123.0 ± 3 pg/mL) and IL-6 (356.5 ± 9 pg/ml) levels (**P < 0.01, n = 3/group).

The cross-talk between T lymphocytes and macrophages is essential for the progression of liver injury in the early phase of IRI.6, 15 To address the mechanism by which B7-H1 engagement may affect macrophage priming, we cultured mouse BMMs plus anti-CD3 mAb-stimulated T cells with control Ig, B7-H1Ig, or B7-H1Ig plus anti–IL-10 mAb ( Fig. 6B). Anti-CD3–activated T lymphocytes primed BMMs in this coculture system, as evidenced by increased TNF-α/IL-6 elaboration (P < 0.01). Interestingly, B7-H1Ig suppressed macrophage-induced TNF-α and IL-6 levels (62.0 ± 6 versus 174.6 ± 11 pg/mL [TNF-α], 129.2 ± 8 versus 653.4 ± 7 pg/mL [IL-6]; P < 0.01). However, concomitant anti–IL-10 mAb re-created BMM activation, as evidenced by augmented TNF-α (123.0 ± 3 pg/mL) and IL-6 (356.5 ± 9 pg/mL) expression. These results document the key regulatory function of PD-1/B7-H1 signaling and suggest that the defect of B7-H1Ig–activated T cells to activate macrophages was IL-10–dependent.


In this study, we demonstrate for the first time that PD-1 negative T cell costimulation regulates local innate immunity-driven inflammation response leading to liver IRI. Indeed, although disruption of PD-1 signaling augmented the hepatocellular damage, its deliberate stimulation following B7-H1 engagement protected livers from fulminant IRI through a local IL-10–mediated mechanism. These data suggest that engaging a negative PD-1/B7-H1 signal is required for maintaining liver homeostasis during IR-mediated hepatocellular insult.

Triggering negative signals through PD-1/B7-H1 in mice has been shown to promote corneal, skin, and cardiac allograft survival16-18 and peripheral transplantation tolerance.19-22 In addition, PD-1/B7-H1 interaction is essential for the spontaneous acceptance of mouse liver allografts.23 The necessity for PD-1/B7-H1 costimulation in hepatic defense against IR insult became evident after treatment of WT mice with anti–B7-H1 mAb. PD-1 blockade increased sALT levels and histological Suzuki grading of liver injury. We have reported similar findings in mice deficient in antioxidant heme oxygenase-1, in which decreased basal heme oxygenase-1 levels exacerbated IR-mediated liver damage.24 Similar to the cytoprotection facilitated by heme oxygenase-1,25 we asked whether stimulating PD-1/B7-H1 signals might improve liver function. We chose the approach of Freeman et al.26 by engaging the negative receptor PD-1 with a dimeric recombinant fusion protein consisting of the extracellular domain of B7-H1 and the Fc portion of IgG. This construct has been used in mouse islet14 and cardiac18 allograft models. In our series, stimulation of PD-1 signaling decreased sALT levels and ameliorated the cardinal histological features of liver injury. The therapeutic potential of PD-1 stimulation was also evident by diminished local T lymphocyte, neutrophil, and macrophage infiltration/activation; reduced parenchyma cell necrosis/apoptosis but enhanced anti-necrotic/apoptotic Bcl-2/Bcl-xl protein levels; and decreased inflammatory chemokine/cytokine gene programs in parallel with increased IL-10. Strikingly, neutralization of IL-10 recreated liver IRI and rendered IR-resistant B7-H1Ig pretreated hosts fully susceptible to the panoply of hepatic proinflammatory cascade.

In addition to Kupffer and epithelial cells, liver sinusoidal endothelial cells constitutively express B7-H1.27-30 Hence, PD-1/B7-H1 negative signaling might act as a local traffic regulator to suspend the pathological cell sequestration in the target tissue. Indeed, B7-H1 fusion protein has been shown to determine the accumulation of intrahepatic CD8+ T cells.31 As in our previous studies,12 relatively few CD3+ and CD4+ cells could be found in IR livers, consistent with activation/recruitment of CD4+ T cells within the first hour of reperfusion. Of note, B7-H1 ligation further diminished inflammatory T lymphocyte infiltration/activation, as evidenced by immunohistology and messenger RNA coding for IFN-γ, granzyme B, and CXCL-10. Interestingly, the levels of anti-inflammatory IL-10 (but not IL-4) were selectively heightened, in agreement with the ability of B7-H1Ig–treated T cells to preferentially secrete IL-10,32 and increased PD-1 and IL-10 levels were found in liver transplantation patients at high risk for CMV disease.33 Moreover, PD-1–induced IL-10 may impair CD4+ T cell activation during HIV infection.34 Such an altered local inflammation was responsible for liver protection, because IL-10 neutralization restored inflammation and hepatocellular damage. In support of this notion, we have reported that IL-10 was required for liver protection in mice deficient in CXCL-10,4 and that viral IL-10 gene transfer in WT recipients prevented hepatic IR insult in association with depressed Th1 cytokine/chemokine programs.35 It is plausible that by virtue of selective IL-10 expression, B7-H1Ig might raise the defensive threshold to inflammatory response in IR-exposed livers.

Our results suggest that PD-1/B7-H1 interaction mediates local inflammatory cell infiltration and activation. In the first phase of IR-mediated inflammation response, activation of macrophages and Kupffer cells results in the release of TNF-α, IL-1β, IL-6, CXCL-10, and CCL-2, the signature markers of liver IRI.1-5 These cytokines and chemokines also influence T cell and macrophage trafficking patterns, as evidenced by increased numbers of infiltrating CD3+ cells and F4/80+ cells. However, stimulating PD-1 signals blunted the number of macrophages sequestered in the liver and their inflammation/chemotactic expression programs. In the second phase of IRI, activated neutrophils dominate local damage cascade.1, 2 We observed a marked increase in Ly-6G+ neutrophil infiltration and myeloperoxidase activity in control livers compared with sham controls. Unlike the control group, livers in B7-H1Ig–treated mice were characterized by decreased neutrophil sequestration, along with diminished CXCL-1 and CXCL-5, the key chemoattractants facilitating neutrophil recruitment in hepatic IR inflammation. As T helper 1–derived IFN-γ acts directly on neutrophils to enhance their sequestration in the liver, B7-H1 cross-linking can regulate neutrophil function through cytokine/chemokine networks.

One of the principal mechanisms by which PD-1/B7-H1 ligation affects host alloimmunity is through modulation of T cell apoptosis.12 B7-H1 but not PD-1 blockade inhibited apoptosis of alloantigen-specific T cells in transplant recipients,20 and B7-H1 was identified as a key protein controlling deletion of hepatic CD8+ T cells.31 During liver IRI, death receptor activation, mitochondrial Ca2+ loading, and reactive oxygen species promote mitochondrial permeability transition, which leads to hepatocellular swelling, rupture of the plasma membrane, and release of cytochrome C and other enzymes, resulting in adenosine triphosphate depletion–dependent oncotic necrosis and caspase-dependent apoptosis.1, 36 Both oncotic necrosis and apoptosis proceed through DNA degradation, which can be detected by way of TUNEL assay.36 In addition to nonparenchymal liver cells, hepatocytes constitutively express low levels of PD-L1, which is strongly enhanced by activated T cells or viral infection and is augmented by stimulation with type I or type II IFNs.28 B7-H1Ig engagement did inhibit necrosis/apoptosis in IR livers, as evidenced by decreased frequency of TUNEL+ cells and consistent with diminished cleaved caspase-3 expression. Simultaneously, we detected increased Bcl-2/Bcl-xl levels, which are known to exert anti-necrotic/apoptotic functions.37 Hence, a cellular and physiological mechanism by which B7-H1 ligation exerts cytoprotection accompanied by enhanced local expression of Bcl-2/Bcl-xl is plausible. Consistent with our findings, increased Bcl-2/Bcl-xl levels prevented cell apoptosis in mouse liver IRI.38

We attempted to mimic an in vivo liver damage scenario by employing B7-H1Ig in anti-CD3 mAb-activated murine T cell cultures. Consistent with published data,32 B7-H1Ig–treated T lymphocytes failed to elaborate IFN-γ, yet their IL-10 levels increased over two-fold. This is in agreement with our in vivo findings wherein PD-1 signals attenuated IFN-γ and promoted IL-10 production. We used BMMs and anti-CD3 mAb-activated T cell cocultures to analyze direct cellular interactions. Although B7-H1Ig failed to affect TNF-α/IL-6 in macrophages, it diminished cytokine elaboration profiles in IL-10–dependent fashion in the coculture system. Thus, PD-1 ligation by B7-H1 regulates T cell–macrophage cross-talk, and IL-10 exerts pivotal cytoprotective function in an innate adaptive cytoprotective feedback mechanism. However, other complementary IL-10–protective mechanisms may be at work. Indeed, IL-10–producing conventional dendritic cells requiring TLR9 can provide protection in a sterile inflammation model of liver IRI.39

Our results document the essential role of the PD-1/B7-H1 pathway in liver inflammation leading to organ damage due to warm IR (Fig. 7). This study is the first to demonstrate that stimulating PD-1 negative signals ameliorates the liver IRI by inhibiting T cell activation and Kupffer cell and macrophage functions. Our results provide evidence that harnessing the physiological mechanisms of negative costimulation by PD-1 upon T cell–Kupffer cell cross-talk may be instrumental in the maintenance of hepatic homeostasis by minimizing organ damage and promoting IL-10–dependent cytoprotection. Targeting PD-1 represents a novel means of improving liver function, expanding the organ donor pool, and improving the overall success of liver transplantation.

Figure 7.

Cross-talk interactions between liver IR and PD-1 signaling. IR-triggered liver damage leads to CD4 T cell activation, which can transmit the signal to macrophage and neutrophil via IFN-γ. PD-1 is being induced on activated CD4 T cells. After B7-H1Ig engagement, preferential IL-10 secretion attenuates activated macrophage and neutrophil function, and ultimately prevents IR-induced hepatic injury.