The β2 integrin CD11b attenuates polyinosinic:Polycytidylic acid–induced hepatitis by negatively regulating natural killer cell functions

Authors

  • Minggang Zhang,

    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    2. Institute of Immunology, Zhejiang University School of Medicine, Hangzhou, China
    Search for more papers by this author
  • Yanmei Han,

    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Chaofeng Han,

    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Sheng Xu,

    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Yan Bao,

    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Zhubo Chen,

    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Yan Gu,

    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Dajing Xia,

    1. Institute of Immunology, Zhejiang University School of Medicine, Hangzhou, China
    Search for more papers by this author
  • Xuetao Cao

    Corresponding author
    1. National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China
    2. Institute of Immunology, Zhejiang University School of Medicine, Hangzhou, China
    • National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, 800 Xiangyin Road, Shanghai 200433, China
    Search for more papers by this author
    • fax: (86)-21-6538-2502.


  • Potential conflict of interest: Nothing to report.

Abstract

The β2 integrins play a key role in inflammation and immune responses. The β2 integrin CD11b has been shown recently to be important in the maintenance of tolerance; however, the underlying mechanisms remain to be fully understood. Natural killer (NK) cells are an important effector of innate immunity but are also a regulator of adaptive immune response. How the activating and inhibitory signals are balanced to determine NK cell function needs to be further identified. CD11b expression was dramatically up-regulated on NK cells once they matured and became activated; therefore, we investigated the role of inducible CD11b in the regulation of NK cells. Neutralizing anti-CD11b antibody enhanced cytotoxicity, interferon-γ (IFN-γ) and granzyme B production of Toll-like receptor 3 (TLR3)-triggered NK cells. CD11b-deficient NK cells stimulated with or without the TLR3 ligand polyinosinic:polycytidylic acid [poly(I:C)] exhibited more potent cytotoxicity, and higher production of IFN-γ and granzyme B. Through in vivo depletion of NK cells and adoptive transfer of CD11b-deficient NK cells, we demonstrated that CD11b-mediated suppression of NK cell function was responsible for attenuation of poly(I:C)-induced acute hepatitis by CD11b. Conclusion: Our findings demonstrate that CD11b negatively regulates NK cell activation and thus attenuates poly(I:C)-induced acute hepatitis. Our study provides a new mechanistic explanation for maintenance of tolerance and control of inflammation by CD11b. (HEPATOLOGY 2009.)

Natural killer (NK) cells, the major effector of the innate immune system, are critical for host defense against infection and immune surveillance through secretion of cytokines and lysis of transformed or malignant cells.1 In addition, NK cells play important roles in the induction and regulation of adaptive immune response. NK cells have been found to be multifunctional; for example, NK cells have been shown recently to be involved in tolerance induction. However, the mechanisms underlying the regulation of NK cell functions are not fully understood.2 NK cells employ several classes of activating receptors, including KIR2S, KIR3S, Ly49D, CD94/NKG2C, CD94/NKG2E, and NKG2D.3, 4 In concert with these, NK cells express a set of inhibitory receptors, such as KLRG1, most members of the Ly49 family, KIR2DL, KIR3D, transforming growth factor-β receptor, and CD94/NKG2A.5, 6 It is well-accepted that the delicate balance between activating and inhibitory signals decides NK cell reactivity.3, 7, 8 How to activate NK cells for immune defense against infectious diseases and cancer has attracted much attention in recent years; however, how to control NK cell function for the immune homeostasis needs to be investigated further. Identification of new negative regulators for NK cell function will contribute to a better understanding of why NK cells are tightly controlled in the physiological process and how the disordered NK cells are involved in the immunopathological injury of host tissue during the inflammatory response.

The β2 integrins (CD11/CD18) are heterodimeric leukocyte adhesion molecules expressed on hematopoietic cells. The common β2 chain (CD18) associates with four distinct α chains (αL, αM, αX, and αD), forming leukocyte functional antigen-1 (LFA-1, CD11a/CD18), Mac-1 (CD11b/CD18, CR3), gp150,95 (CD11c/CD18), and CD11d/CD18.9 Among them, Mac-1 is expressed on dendritic cells, granulocytes, monocytes/macrophages, and NK cells and participates in cell activation, chemotaxis, cytotoxicity, and phagocytosis. It has been shown recently that active CD11b molecules expressed on antigen-presenting cells can inhibit full T cell activation directly10 and might facilitate the development of peripheral tolerance by suppressing T helper 17 differentiation.11 Also, ligation of Mac-1 could reduce interleukin (IL)-12p70 induction and subsequently impair clearance of Porphyromonas gingivalis.12 In addition, CD11b could attenuate dextran sodium sulfate-induced colitis.13 These reports indicate that CD11b may mediate negative signals to control immune response and inflammation. However, the mechanisms by which CD11b-deficient mice are prone to induction of autoimmune diseases have not been fully elucidated. Our previous studies show that the stromal microenvironment of the spleen, liver, and lung can drive the generation of new populations of regulatory dendritic cells with high expression of CD11b but low expression of major histocompatibility complex class II. Organ stroma-educated CD11bhigh regulatory dendritic cells could down-regulate T cell response, contributing to tolerance induction and resulting in the attenuation of asthma and hepatitis.14–17 However, the biological significance of high CD11b expression in these regulatory dendritic cells remains unclear. Therefore, the regulation of immune response by CD11b and the underlying mechanisms need to be investigated further.

NK cells express CD11b, with more pronounced expression of CD11b once they have matured and are activated. However, the regulation of NK cell function by CD11b remains unclear. The purpose of this study was to investigate the role of inducible CD11b in the regulation of NK cell function. We have demonstrated that CD11b can negatively regulate Toll-like receptor 3 (TLR3)-triggered NK cell function through activation of the c-Jun N-terminal kinase (JNK) pathway. Also, CD11b-mediated suppression of NK cell function contributes to the attenuation of polyinosinic:polycytidylic acid [poly(I:C)]-induced acute hepatitis by CD11b. Our results provide a new mechanistic explanation for the maintenance of tolerance and control of inflammation by CD11b.

Abbreviations

ALT, alanine aminotransferase; AST, aspartate aminotransferase; FACS, fluorescence-activated cell sorting; IFN-γ, interferon-γ; IL, interleukin; JNK, c-Jun N-terminal kinase; MAPK, mitogen-activated protein kinase; MFI, mean fluorescence intensity; NK, natural killer; PMA, phorbol 12-myristate 13-acetate; poly(I:C), polyinosinic:polycytidylic acid; TLR3, Toll-like receptor 3; WT, wild-type.

Materials and Methods

Materials.

C57BL/6 (CD45.2+) and BALB/c mice were obtained from Joint Ventures Sipper BK Experimental Animal Co. (Shanghai, China). CD11b-deficient (CD11b−/−, C57BL/6 background) mice, CD1d-deficient (CD1d−/−, BALB/c background) mice, and C57BL/6 (CD45.1+) mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Animal experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals with the approval of the Scientific Investigation Board of the Second Military Medical University, Shanghai, China. Information on the cell lines and reagents was described in the Supporting Materials section.

Preparation of NK Cells and Hepatocytes.

Murine liver mononuclear cells were isolated and purified as described previously.18 To purify NK cells, splenocytes or liver MNC were stained with antibodies, and CD3NK1.1+ cells, CD3NK1.1+CD11b+ or CD3NK1.1+CD11b cells were sorted using a MoFlo high-speed cell sorter (DakoCytomatix), the purity of which was confirmed to be >98%. Splenic NK cells were used for all experiments unless mentioned specially. Primary hepatocytes were isolated as described.19

Fluorescence-Activated Cell Sorting Analysis.

Cells were stained with monoclonal antibodies as described.18 Flow cytometric analysis was performed on a FACS LSRII with FACSDiva software (BD Biosciences).

Assay for NK Cell Cytotoxicity.

NK cell cytotoxicity was analyzed using flow cytometry as described.18, 20

Western Blot Analysis.

NK cells stimulated with or without 20 μg/mL poly(I:C) for the indicated time, harvested, lysed, and then blotted as described.21

Acute Liver Injury.

Mice were intraperitoneally injected with poly(I:C) (20-30 μg/g body weight), and liver histology, serology, and survival were assessed to represent disease severity. Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were detected according to manufacturer protocols (Rongsheng Company, Shanghai, China), and serum interferon-γ (IFN-γ) was assayed using an enzyme-linked immunosorbent assay kit (R&D Systems, Minneapolis, MN) as described.16

Statistical Analysis.

The statistical significance of differences between experimental groups was analyzed using the Student t test for paired samples. P < 0.05 and P < 0.01 were taken to indicate significance.

Results

CD11b Expression Is Up-regulated During NK Cell Differentiation and Is Dramatically Induced on TLR3-Triggered NK Cells.

First, we analyzed CD11b expression during differentiation of NK cells. As shown in Fig. 1A, little CD11b expression was observed on NK precursors (LinCD122+NKG2D+NK1.1); however, CD11b expression on immature NK cells (CD3NK1.1+DX5) increased significantly. Once NK cells matured (inactive mature NK cells, CD69CD43+CD3DX5+) or were activated (CD69+CD3DX5+),22 CD11b expression increased dramatically. In addition, NK cells from newborn mice expressed little CD11b, but CD11b expression on NK cells increased along with age, and up to about 80% of NK cells were CD11b-positive when mice grew up to 8 weeks (Fig. 1B). The data indicated that CD11b increases along with NK cell maturation and activation. Then we stimulated the purified NK cells with poly(I:C), phorbol 12-myristate 13-acetate (PMA) plus ionomycin, IL-12, or IL-15, respectively, and found that poly(I:C) and PMA plus ionomycin could dramatically increase CD11b expression on NK cells (Fig. 1C). Consistent with in vitro observations, CD11b expression increased rapidly and dramatically on NK cells in mice intraperitoneally injected with poly(I:C) (Fig. 1D). Combining in vitro and in vivo data, we showed that CD11b expression is dramatically induced on activated NK cells.

Figure 1.

CD11b expression is up-regulated dramatically in mature and activated NK cells. (A) CD11b expression on NK cells during differentiation of NK cells. NK precursors (LinCD122+NKG2D+NK1.1), immature NK cells (CD3NK1.1+DX5), inactivated mature NK cells (CD69CD43+CD3DX5+), and active NK cells (CD69+CD3DX5+) were sorted from splenocytes of C57BL/6 mice with a MoFlo high-speed cell sorter (DakoCytomatix). CD11b expression was analyzed by way of fluorescence-activated cell sorting (FACS). (B) CD11b expression on splenic NK cells of C57BL/6 mice with different ages, including newborn (0 d), 1 week (1 w), 4 weeks (4 w), and 8 weeks (8 w) (n = 6 mice per group with different age tested). (C) Purified splenic NK cells were from 4-week-old C57BL/6 mice, and the mean fluorescence intensity (MFI) folds represent the up-regulation of CD11b expression induced by poly(I:C), PMA plus ionomycin, IL-12, or IL-15 in vitro, compared with that of purified NK cells cultured at the same time but without stimuli; the CD11b MFI of the latter was set to 1. (D) CD11b expression on NK cells derived from 4-week-old C57BL/6 mice (n = 6) intraperitoneally injected with TLR3 ligand poly(I:C) (20μg/g body weight) for the time indicated (0-48 hours). Percentage of CD11b+ NK cells in total splenic NK cells and increase of CD11b MFI folds were shown for the in vivo up-regulation of CD11b expression on NK cells activated by poly(I:C). The increase of CD11b MFI folds was calculated as follows: CD11b MFI of splenic NK cells from mice with no stimulation (i.e., at 0 hours) was set to 1, and then CD11b MFI of NK cells with stimulation was compared with it. All experiments were performed five times with similar results.

Phenotypic and Functional Analysis of CD11b+ NK Cells and CD11b NK Cells.

CD11b expression increases rapidly and dramatically on NK cells upon maturation and activation. We wonder what role of CD11b in the regulation of these CD11b+ mature or activated NK cells. This inspired us to examine the characteristics of CD11b+ and CD11b NK cells. First, we observed the phenotype of CD11b+ and CD11b NK cells, and found that more expression of CD43, NKG2A/C/E, NKG2D, Ly49A, Ly49C/I, and Ly49D on CD11b+ NK cells than that on CD11b NK cells (Fig. 2A). Accordingly, more significant expression of IFN-γ, granzyme B, and perforin (Fig. 2B) and more potent cytotoxicity (Fig. 2C) of CD11b+ NK cells were found than those of CD11b NK cells. Then, we transferred the sorted CD11b+ or CD11b NK cells with the CD45.1 marker to recipients (CD45.2+), and CD11b expression on CD45.1+NK1.1+CD3 splenocytes of the recipients was analyzed 1 week later. Interestingly, donor CD11b NK cells could develop to CD11b+ NK cells, but not vice versa (Fig. 2D). Taken together with the above data, these findings strongly suggest that CD11b is an inducible molecule of mature and activated NK cells.

Figure 2.

Characteristics of CD11b+ NK cells and CD11b NK cells. (A) Phenotype of primary CD3NK1.1+CD11b+ (CD11b+ NK) and CD3NK1.1+CD11b (CD11b NK) NK cells from spleens of C57BL/6 mice. Isotype represents the phenotype of CD3NK1.1+ NK cells. (B) IFN-γ, perforin, and granzyme B expression of CD11b+ NK cells and CD11b NK cells. Splenic CD11b+ NK cells and CD11b NK cells were stimulated with poly(I:C) (20 μg/mL) in the presence of Brefeldin A for 6 to 8 hours in vitro, then IFN-γ, perforin, and granzyme B were intracellularly stained and analyzed by way of FACS. (C) More potent cytotoxicity of the purified CD11b+ NK cells against YAC-1 cells. Purified CD11b+ or CD11b NK cells were seeded in 96-well plates and cocultured with target cells for 4 hours at 37°C. Cells were then collected, stained with anti-NK1.1 allophycocyanin and annexin V–fluorescein isothiocyanate, stained with 7-aminoactinomycin D after washing, and analyzed by FACS. Live target cells were assayed by gating on NK1.1annexin V7-aminoactinomycin population. Cytotoxicity (%) = (NK1.1 cells − NK1.1annexin V7-aminoactinomycin cells)/(NK1.1 cells) ×100. E:T, effector:target ratio. **P < 0.01 versus CD11b NK cells. (D) Differentiation of CD45.1+CD11b NK cells to CD45.1+CD11b+ NK cell in vivo. CD45.1+CD11b+ or CD45.1+CD11b NK cells from spleens of C57BL/6 (CD45.1+) mice were adoptively transferred to recipient (C57BL/6, CD45.2+) mice, then CD11b expression on splenic CD45.1+CD3NK1.1+ NK cells was analyzed 1 week later. All experiments above were performed five times with similar results.

CD11b Negatively Regulates Function of TLR3-Triggered Mature NK Cells In Vitro.

In order to elucidate the role of CD11b in NK cell function, we sorted CD11b+ NK cells and determined their cytotoxicity when stimulated with or without poly(I:C) in the presence or absence of neutralizing anti-CD11b antibody. The cytotoxicity of CD11b+ NK cells increased once activated by poly(I:C) in vitro. Additionally, neutralizing anti-CD11b antibody could significantly enhance the cytotoxicity of CD11b+ NK cells activated by poly(I:C) in vitro (Fig. 3A). Consistently, CD11b-deficient NK cells exhibited more potent cytotoxicity than wild-type (WT) NK cells after stimulation with poly(I:C) (Fig. 3B). The results suggest that CD11b can negatively regulate the cytotoxicity of TLR3-triggered NK cells. Accordingly, neutralizing anti-CD11b antibody could significantly enhance the expression of IFN-γ and granzyme B (Fig. 3C and Supporting Fig. 1A,B), but could not affect the expression of perforin (Fig. 3C) on CD11b+ NK cells activated by poly(I:C) in vitro. Consistently, CD11b-deficient NK cells expressed more IFN-γ and granzyme B (Fig. 3D and Supporting Fig. 1C,D) than WT NK cells after stimulation with poly(I:C) in vitro, but CD11b deficiency did not affect the perforin expression in NK cells as above (Supporting Fig. 2). Thus, the inducible CD11b can negatively regulate the cytotoxicity of TLR3-triggered mature NK cells through suppression of IFN-γ and granzyme B expression.

Figure 3.

CD11b negatively regulates mature NK cell function. (A-B) NK cell cytotoxicity was assayed as described in Fig. 2C. (A) Cytotoxicity against YAC-1 cells of sorted C57BL/6 splenic CD11bhigh NK cells stimulated with or without poly(I:C) (20 μg/mL) in the presence or absence of neutralizing anti-CD11b antibody. (B) CD11b-deficient (CD11b−/−) splenic NK cells exhibited more potent cytotoxicity against YAC-1 cells than C57BL/6 (WT) splenic NK cells stimulated with or without poly(I:C) (20 μg/mL). (C) Intracellular staining assay for IFN-γ, granzyme B, and perforin production from sorted C57BL/6 splenic CD11bhigh NK cells stimulated with poly(I:C) (20 μg/mL) for 6 hours in the presence or absence of neutralizing anti-CD11b or isotype antibody in vitro. (D) CD11b−/− splenic NK cells produced more IFN-γ and granzyme B than C57BL/6 (WT) splenic NK cells stimulated with or without poly(I:C) (20 μg/mL) for 6 hours in vitro, as detected by intracellular staining. All data are representative of three independent experiments.

CD11b Negatively Regulates Function of TLR3-Triggered Liver NK Cells In Vivo.

We determined the number of NK cells in mouse liver after in vivo injection of poly(I:C). The percentage of NK cells in the liver mononuclear cells of normal C57BL/6 mice was 10%; injection with poly(I:C) dramatically elevated this percentage to about 35% at 24 hours (Fig. 4A). There were more NK cells significantly accumulated in the livers of CD11b-deficient mice after poly(I:C) injection, indicating that CD11b expression affects NK cell accumulation in livers induced by poly(I:C).

Figure 4.

CD11b deficiency increases NK cell function in the livers of mice injected with poly(I:C). (A) NK cell number in the livers of C57BL/6 (WT) and CD11b-deficient mice (CD11b−/−) after injection of poly(I:C) (20 μg/g body weight). *P < 0.05. **P < 0.01 versus WT mice. (B) Effect of CD11b expression on the cytotoxicity of liver NK cells against YAC-1 cells or primary hepatocytes of C57BL/6 mice. Liver NK cells were freshly sorted from C57BL/6 mice (WT) or CD11b−/− mice with or without poly(I:C) injection (20 μg/g body weight) for 18 hours, and their cytotoxicity was assayed against primary hepatocytes or YAC-1 cells as described in Fig. 2C. E:T, effector:target ratio. (C) Effect of CD11b expression on granzyme B and IFN-γ production of liver NK cells stimulated with or without poly(I:C). Granzyme B and IFN-γ production in NK cells derived from C57BL/6 (WT) or CD11b−/− mice 18 hours after poly(I:C) injection (20 μg/g body weight) were detected by intracellular staining with FACS. All experiments above were performed three times with similar results.

In order to directly elucidate the role of CD11b in the regulation of function of NK cells accumulated in the liver after poly(I:C) injection, we sorted NK cells from mouse livers 18 hours after poly(I:C) injection and found that CD11b-deficient NK cells showed more potent cytotoxicity against YAC-1 cells and primary hepatocytes (Fig. 4B) than NK cells from WT mice. We also observed that granzyme B and IFN-γ production were significantly increased in CD11b-deficient liver NK cells compared with that in WT liver NK cells 18 hours after poly(I:C) injection (Fig. 4C). The data indicated that CD11b could inhibit the cytotoxicity of TLR3-triggered liver NK cells to target cells including hepatocytes in vivo.

CD11b Negatively Regulates TLR3-Triggered NK Cell Function by Activating the JNK Pathway.

We investigated the underlying mechanisms for the negative regulation of TLR3-triggered NK cell function by CD11b. The mitogen-activated protein kinase (MAPK) pathway was analyzed in TLR3-triggered WT NK cells in the presence or absence of neutralizing anti-CD11b antibody (Fig. 5A). We found that CD11b blockade could inhibit JNK activation in NK cells stimulated with poly(I:C). Accordingly, poly(I:C)-induced JNK activation was impaired in CD11b-deficient NK cells compared with that in WT NK cells. The data suggest that CD11b could promote JNK activation in TLR3-triggered NK cells. Wild-type NK cells were pretreated with MAPK specific inhibitors, then stimulated with poly(I:C). Consistently, SP600125, a kind of JNK pathway specific inhibitor, could dramatically promote granzyme B generation (Fig. 5B,C) and IFN-γ production (Fig. 5D) in TLR3-triggered NK cells. These results suggest that CD11b may inhibit TLR3-triggered activation of NK cells through activation of JNK pathway.

Figure 5.

CD11b controls TLR3-triggered granzyme B and IFN-γ production by activation of JNK in NK cells. (A) Western blot analysis of MAPK pathway in TLR3-triggered NK cells. C57BL/6 (WT) splenic NK cells were pretreated with anti-CD11b antibody or isotype antibody for 15 minutes, and then stimulated with poly(I:C) (20 μg/mL) for the time indicated, and NK cells purified from CD11b−/− mice were stimulated with poly(I:C) (20 μg/mL) for the time indicated. The cells were harvested, lysed, and quantified for protein concentration and blotted with equal protein quantity. (B) Effects of MAPK inhibitors on granzyme B expression in TLR3-triggered NK cells. C57BL/6 (WT) splenic NK cells were pretreated with PD98059 (extracellular signal-regulated kinase–specific inhibitor), SP600125 (JNK-specific inhibitor), SB203580 (p38-specific inhibitor) for 30 minutes, respectively, and then stimulated with poly(I:C) (20 μg/mL) in the presence of Brefeldin A for 6 hours. Granzyme B expression in TLR3-triggered NK cells was analyzed by way of FACS. (C) The statistics of data in (B). Percentage of granzyme B+ NK cells in total NK cells and granzyme B MFI were shown. *P < 0.05. **P < 0.01 versus dimethyl sulfoxide group. (D) Effects of MAPK inhibitors on IFN-γ expression in TLR3-triggered NK cells. The C57BL/6 (WT) splenic NK cells were treated as in (B), and IFN-γ expression in TLR3-triggered NK cells was analyzed by way of FACS. All data are representative of five independent experiments.

CD11b Attenuates Acute Liver Injury Mediated by TLR3-Triggered NK Cells.

On the basis of observations that more CD11b-deficient NK cells could be recruited to the liver and the function of liver NK cells in CD11b-deficient mice increased more significantly after in vivo injection of poly(I:C), we finally investigated the biological significance of the CD11b in the negative regulation of NK cell function with the mouse model of acute liver injury induced by poly(I:C). We observed that the levels of serum ALT, AST, and IFN-γ increased significantly in WT mice injected with poly(I:C), and the levels of serum ALT, AST, and IFN-γ increased more significantly and rapidly in CD11b-deficient mice than in WT mice after injection with poly(I:C) (Fig. 6A-C). Serum ALT and AST levels reached their peak in CD11b-deficient mice 4 hours earlier than WT mice. In the CD11b-deficient mice injected with poly(I:C), the peak time of serum IFN-γ was at 8 hours, 8 hours earlier than the peak time of ALT (16-20 hours) and AST (20-24 hours), suggesting that NK cell–derived IFN-γ might be responsible for liver damage with increased production of ALT and AST. Histological analysis also revealed the inflammation and necrosis of liver were significantly more severe in CD11b-deficient mice than in WT mice after injection with poly(I:C) (Fig. 6D). Importantly, CD11b deficiency could shorten the survival of mice injected with a lethal dosage of poly(I:C) (Fig. 6E). These data suggest that CD11b may attenuate poly(I:C)-induced acute liver injury.

Figure 6.

CD11b deficiency promotes poly(I:C)-induced acute liver injury in vivo. (A-C) Increased levels of serum ALT, AST, and IFN-γ in the CD11b-deficient mice injected with poly(I:C). WT C57BL/6 mice (n = 10) and CD11b−/− mice (n = 10) were injected with poly(I:C) (20 μg/g body weight), and serum ALT (A), AST (B), and IFN-γ (C) were assayed at the indicated times. Sera were collected for assays of ALT, AST, and IFN-γ according to the manufacturer's instructions. *P < 0.05 and **P < 0.01 versus C57BL/6 (WT) mice. (D) More severe damage of liver tissue in CD11b-deficient mice injected with poly(I:C). C57BL/6 (WT) and CD11b−/− mice were injected with poly(I:C) (20 μg/g body weight) for 18 hours. Livers were then fixed, embedded, and sliced, and the liver sections were stained with hematoxylin-eosin. Arrows indicate the liver necrosis focus. (E) Shortened survival of CD11b-deficient mice injected with high doses of poly(I:C). Survival of C57BL/6 (WT) and CD11b−/− mice (n = 10) was observed every hour after injection of poly(I:C) (30 μg/g body weight). All experiments above were performed five times with similar results.

In order to determine whether poly(I:C)-induced acute liver injury was mediated by TLR3-triggered NK cells and whether the suppression of NK cell function by CD11b would contribute to the attenuation of poly(I:C)-induced acute liver injury by CD11b, we deleted NK cells through in vivo injection of PK136 monoclonal antibody, then injected poly(I:C) 24 hours later. As shown in Fig. 7A, the depletion of NK cells effectively eliminated the poly(I:C)-induced elevation of serum ALT, AST, and IFN-γ in both WT and CD11b-deficient mice. As we know, PK136 is a kind of monoclonal antibody against mouse NK1.1 antigen. In order to know whether NK T cells (NK1.1+CD3+) were also involved in poly(I:C)-induced acute liver injury, CD1d-deficient (CD1d−/−) mice lacking NK T cells were injected with poly(I:C); however, no significant difference in serum ALT and AST was observed between WT mice and CD1d−/− mice (Supporting Fig. 3), suggesting that NK cells, but not NK T cells, contributed to the pathogenesis of this model with poly(I:C)-induced acute liver injury.

Figure 7.

Suppression of NK cell function contributes to attenuation of poly(I:C)-mediated acute hepatitis by CD11b. (A) Effects of NK cell depletion on the increase of serum ALT, AST, and IFN-γ in mice injected with poly(I:C). WT C57BL/6 mice (n = 6) or CD11b−/− mice (n = 6) were depleted of NK cells through intravenous injection of PK136 (200 μg/mouse). Mouse immunogloblin G2a was used as an isotype control (200 μg/mouse); poly(I:C) (20 μg/g body weight) was injected intraperitoneally 24 hours later. Serum ALT and AST levels were determined 20 hours after poly(I:C) injection and serum IFN-γ was assayed 8 hours after poly(I:C) injection as described in Fig. 6. **P < 0.01. (B) Effects of NK cell transfer on the increase of serum ALT, and AST in the NK cell–depleted mice injected with poly(I:C). C57BL/6 mice (n = 6) were depleted of NK cells as in (A). Forty-eight hours later, the purified CD11b+ NK cells from C57BL/6 mice or purified CD11b-deficient (CD11b−/−) NK cells (4 × 106 cells/mouse) were transferred adoptively into these NK cell–depleted mice, and then injected with poly(I:C) (20 μ/g body weight) 24 hours later. Serum ALT and AST levels were determined 20 hours after injection of poly(I:C). **P < 0.01. Data represent one of three (B) or five (A) experiments with similar results.

Finally, we wanted to confirm whether CD11b could attenuate such acute liver injury by inhibiting NK cell function. We adoptively transferred CD11b-deficient NK cells to mice in which NK cells were depleted 2 days previously and injected poly(I:C) 24 hours later. As shown in Fig. 7B, adoptive transfer of WT NK cells restored liver injury in these mice. More importantly, adoptive transfer of CD11b-deficient NK cells resulted in more severe acute liver injury with significantly higher serum ALT and AST levels. Together, these data demonstrate that CD11b-mediated suppression of NK cells contributes to the attenuation of poly(I:C)-induced acute liver injury by CD11b in vivo.

Discussion

As the important effector at the first line of host innate defense, NK cells are very sensitive to the stimulation of pathogenic components. The recognition of pathogens activates NK cells, which then exert their innate immunity to eliminate the pathogens by releasing cytotoxic substances, cytokines, and chemokines, directly or indirectly activating other kinds of immune cells to cooperate.23 However, more and more recent evidence has shown that NK cells also play a role in the regulation of immune response,23–26 and disorders of NK cells may be involved in the pathogenesis of autoimmune diseases.27, 28 For example, NK cells promote the induction of allograft tolerance for hematopoietic stem cells, islets, and skin,29–32 and play a preventive role for autoimmune diabetes in nonobese diabetic mice.27, 33 NK cell–derived IFN-γ and subsets of NK cells with regulatory function have been proposed to be the reasons for the regulation of immune response by NK cells.26, 33 Therefore, much attention has been devoted recently to the investigation of how NK cells are tightly controlled in the physiological process and what molecules responsible for the negative control of NK cell functions. More negative regulators of NK cell function need to be determined. For example, β-arrestin 2 was recently reported to be a novel inhibitor of NK cell activation and maintain resting status of NK cells under physiological conditions by recruiting the tyrosine phosphatases SHP-1 and SHP-2 to an inhibitory receptor, KIR2DL1.34 Here we provide evidence that CD11b is a negative feedback regulator of NK cell function.

In our preliminary experiments, we found that CD11b expression rapidly increased on NK cells upon activation. Also, TLR3 ligand poly(I:C) injection could increase CD11b expression on NK cells dramatically in vivo. Interestingly, CD11b expression on NK cells increases while mice grow to 8 weeks, which may be due to NK cell maturation with age. However, there are always ∼20% NK cells which are CD11b negative even how old the mice are (data not shown). As for poly(I:C) injection into adult mice, the abundance of CD11b on NK cells will increase significantly compared with that on NK cells without poly(I:C) injection, however, there is always a population of CD11b negative NK cells (Supporting Fig. 4), which may be possibly due to immature NK cells existing, or new NK cells generated from their progenitors. Thus, on the basis of these observations, we went further in this study to investigate the role of CD11b in the regulation of NK cell function. Previous report showed that complement C3b, a kind of ligand of Mac-1 (CD11b/CD18), could inhibit NK cell–mediated antibody-dependent cellular cytotoxicity.35 We found that the neutralizing anti-CD11b antibody could enhance cytotoxicity, IFN-γ and granzyme B production of TLR3-triggered NK cells. Furthermore, CD11b-deficient NK cells showed more potent cytotoxicity and more production of IFN-γ and granzyme B in response to TLR3 ligand poly(I:C). These data strongly suggest that CD11b can significantly control the activation of NK cells triggered by TLR3.

Signaling pathways related to β2 integrin are diverse. Our experiments identified the relationship between CD11b molecule and MAPK pathways. CD11b may activate JNK in NK cells, because neutralizing anti-CD11b antibody significantly reduced p-JNK in WT NK cells, and JNK activation was impaired in CD11b-deficient NK cells in response to poly(I:C) stimulation. Accordingly, JNK specific inhibitor, SP600125 could dramatically promote granzyme B and IFN-γ production in NK cells following poly(I:C) stimulation. Thus, CD11b may inhibit TLR3-triggered activation of NK cells through activation of JNK pathway. Heterodimeric Mac-1 is composed of CD11b and CD18, and it is the most confusing member of the β2-integrin family with more than 30 known ligands, such as angiostatin, bacterial and fungal products, coagulation factor X, complement fragment iC3b, denatured protein, fibrinogen, hookworm neutrophil inhibitory factor, ICAM-1 and plasminogen.10, 36 It remains unclear how this single integrin recognizes such multiple structurally distinct ligands and how Mac-1 converts such multiple specific binding into biological signaling. We show that CD11b negatively regulates NK cell function, but the ligand here for CD11b binding and leading to negative signaling in NK cells remain to be further elucidated.

NK cell recruitment and activation in a particular site is regarded as a critical step for NK cell functioning during defense against infectious diseases or tumor metastasis.37 NK cells are quite abundant in the liver compared with in the other peripheral lymphatic system,38 and various artificially synthesized or natural products, such as poly(I:C),39 are able to recruit and activate NK cells in the liver, implying an important role of NK cells in liver biology and physiology. Our data (Fig. 7) showed that NK cell deletion significantly reduced IFN-γ release after poly(I:C) injection, further confirming the previous report that ligands of TLR-2-5 induce high levels of IFN-γ production by the purified resting NK cells, and NK cells are the main source of IFN-γ following stimulation with TLR ligands.40, 41 Interestingly, previous reports have shown that NK cells can eliminate target cells through FasL and TRAIL-mediated pathways.42, 43 In addition, natural killing also involves exocytosis of perforin- and granzyme-containing granules via a metabolically active process. In this study, we observed that TLR3 could trigger more production of IFN-γ and granzyme B, but not perforin (Supporting Fig. 2), in CD11b-deficient NK cells. Notably, CD11b deficiency promoted poly(I:C)-induced liver injury, because the levels of serum ALT, AST and IFN-γ increased more significantly and rapidly in the CD11b-deficient mice than that in WT mice after injection with poly(I:C), and CD11b deficiency dramatically shortened the survival of the mice injected with lethal dosage of poly(I:C). By using NK cell depletion and adoptive transfer of CD11b-deficient NK cells, we show that CD11b on NK cells contributes to attenuation of liver injury mediated by TLR3-triggered NK cells. CD11b is also expressed on macrophages, and we have shown that CD11b is also a negative regulator of macrophages in response to TLR4 ligand (our unpublished data). Although in vivo administration of poly(I:C) can activate macrophage responses, with experiments of NK cell depletion and adoptive transfer of CD11b-deficient NK cells (Fig. 7) and using GdCl344 to delete macrophages in the liver (Supporting Fig. 5), we show that CD11b on NK cells, but not CD11b on macrophages, was responsible for attenuation of liver injury mediated by TLR3-triggered NK cells. Also, NK cell-derived IFN-γ and direct cytotoxicity might be responsible for the liver damage with the increased production of ALT and AST, because the peak of serum IFN-γ (8 hours) was about 8 hours earlier than the peak time of ALT (16-20 hours) and AST (20-24 hours) in the model of acute liver injury. So, the data indicate that CD11b is an important regulator in maintaining immune homeostasis in vivo.

In conclusion, this is the first report about the negative regulation of NK cell function by β2-integrin CD11b both in vitro and in vivo. CD11b molecule, rapidly increased on the activated NK cells, can control TLR3-triggered NK cell functions by inhibiting cytotoxicity and production of IFN-γ and granzyme B through activation of JNK pathway, thus controlling excessive NK cell–mediated immunopathological damage or immune inflammation. Our results provide a new mechanistic explanation for CD11b in the maintenance of immune homeostasis and inducible feedback control of inflammation.

Acknowledgements

We thank Yushi Yao for helpful discussion, Xiaoting Zuo, and Jianqiu Long for technical assistance, and Canrong Ni (Department of Pathology, Changhai Hospital, Shanghai, China) for pathological analysis.

Ancillary

Advertisement