Natural killer group 2 member D, its ligands, and liver disease: Good or bad?


  • Bin Gao M.D., Ph.D.

    Corresponding author
    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    • Section on Liver Biology, NIAAA/NIH, Room 2S-33, 5625 Fishers Lane, Bethesda, MD 20892
    Search for more papers by this author
    • fax: 301-443-0000.

  • See Article on Page 92

  • Potential conflict of interest: Nothing to report.

Liver lymphocytes are enriched in natural killer (NK) cells, which play an important role in host defenses against microbial infection and tumor transformation in the liver.1 Generally, it is believed that the cytotoxicity of NK cells against target cells is controlled by the opposing signals from inhibitory and stimulatory receptors on NK cells interacting with their corresponding ligands expressed on target cells.2, 3 The NK cell inhibitory receptors include CD94/NKG2, Ly49A, and the immunoglobulin-like killer inhibitory receptor, which interact with inhibitory ligands (e.g., self major histocompatibility complex [MHC] class I molecules) expressed on target cells and inactivate NK cell function. The stimulatory receptors include NKp46, NKp30, NKp44, NKG2D, and DNAX accessory molecule 1 (CD226). Among these, the best-defined receptor is NKG2D (natural killer group 2, member D), a highly conserved C-type lectin-like membrane glycoprotein that is also one of the major activating receptors on NK cells.2, 3 Expressed on essentially all NK cells, as well as on γδ-T cell receptor (TcR)+ T cells and αβ-TcR+ CD8+ T cells, NKG2D is found in both humans and mice. In humans, known NKG2D ligands include MHC class I–related chain A and B (MICA/B) and UL16-binding protein 1, 2, 3, 4 (ULBP1, ULBP2, ULBP3, ULBP4). In mice, known NKG2D ligands include retinoic acid early inducible gene-1 (RAE-1), minor histocompatibility H60, and murine UL16-binding protein–like transcript 1 (MULT1). The expression of these NKG2D ligands is usually up-regulated on microbe-infected, transformed, or stressed cells. The interaction between these ligands and NKG2D on NK cells leads to NK cell activation, thereby playing an important role in host defenses against viral infection and tumor transformation. In addition, CD8+ T cells also express NKG2D, which serves as a costimulatory signal to activate CD8+ T cells.2, 3


HBV, hepatitis B virus; HCV, hepatitis C virus; HLA-C1; human leukocyte antigen C group1; HSC, hepatic stellate cell; IFN, interferon; MHC, major histocompatibility complex; MICA/B, MHC class I–related chain A and B; MULT-1, murine UL16-binding protein–like transcript 1; NASH, nonalcoholic steatohepatitis; NAFL, nonalcoholic fatty liver; NK, natural killer cell; NKG2D, natural killer group 2, member D; poly I:C, polyinosinic:polycytidylic acid; RAE-1, retinoic acid early inducible gene-1; ULBP1, UL16-binding protein 1.

The important role of NK cells in the clearance of early hepatitis C virus (HCV) infection is suggested by the results of several genetic studies on the interaction between NK cell receptors and their ligands.4, 5 For instance, Khakoo et al.4 reported that patients with the inhibitory NK cell receptor (KIR2DL3) and its ligand (human leukocyte antigen C group1 [HLA-C1]) had a better chance of spontaneous recovery from acute HCV infection. This is likely due to weak inhibitory KIR2DL3–HLA-C1 interaction, which results in the lack of strong NK cell inhibition and subsequent induction of strong NK cell functions that contribute to HCV clearance. However, the role of NK cell activating receptor NKG2D and its ligands in controlling HCV infection remains largely unknown. Recently, several studies have shown that NKG2D+NK cells are highly enriched in intrahepatic compartments in patients with chronic HCV infection, which correlates with hepatocellular damage.6 Although the expression of NKG2D ligands on HCV-infected or HBV-infected hepatocytes in humans has not yet been explored, it is expected to be elevated because in several murine models of liver injury, up-regulated ligands have been detected on stressed hepatocytes (see below) (Fig. 1).

Figure 1.

Divergent roles of NKG2D ligands in the pathogenesis of liver disease. Expression of NKG2D ligands is up-regulated in hepatocytes, cholangiocytes, stellate cells, and Kupffer cells by a variety of stimuli. These ligands can bind NKG2D on NK cells and subsequently trigger NK cell activation to produce a variety of cytotoxic mediators. These mediators can kill hepatocytes and cholangiocytes to induce liver damage and kill activated HSCs to reduce liver fibrosis. Elevated NKG2D ligands on Kupffer cells can also trigger NK cell activation and induce hepatocellular damage. Hepatocelluar carcinoma and cholangiocarcinoma cells also express NKG2D ligands, which can trigger NKG2D-dependent antitumor immune responses to kill these tumor cells. However, NKG2D ligands (MICA/B) can also be cleaved proteolytically from hepatocellular carcinoma cells to generate soluble forms of MICA/B that down-regulate NKG2D expression and inhibit NKG2D-mediated antitumor effects. Chronic alcohol drinking down-regulates NKG2D expression whereas treatment with poly I:C and type I IFN induces its expression. DILI, drug-induced liver injury.

The expression of RAE-1, MULT-1, and H60 is not detected on normal mouse hepatocytes; however, it is detected at high levels on hepatocytes from bile duct-ligated mice,7 hepatitis B virus (HBV) transgenic mice,8, 9 and mice with drug-induced liver injury.10 Elevated levels of these ligands trigger activation of NK cells, as well as natural killer T (NKT) cells, to kill hepatocytes, resulting in hepatocellular damage.7–10 Induction of RAE-1 expression has also been reported on Kupffer cells in mice treated with polyinosinic:polycytidylic acid (poly I:C) plus D-galactosamine (D-GalN).11 The interaction between NKG2D and RAE-1 stimulates NK cells to produce interferon-gamma (IFN-γ), which then acts together with Kupffer cell-derived tumor necrosis factor-α to synergistically induce fulminant hepatitis.11 In addition to triggering hepatocyte damage, the interaction between NKG2D and corresponding ligands is also involved in NK cell-mediated cholangiocyte injury in a murine model of biliary atresia induced by rotavirus infection.12 In this model, NK cells accumulate in extrahepatic bile ducts and hepatic expression of RAE1, H60, MULT-1 messenger RNAs is markedly up-regulated. Blockade of NKGD2 prevents both epithelial cell injury and the development of the atresia phenotype. In vitro, NK cells lyse cholangiocytes in a contact-dependent and NKG2D-dependent manner.

Although the contribution of NKG2D ligands to hepatocyte and cholangiocyte injury is well documented in rodent models (Fig 1), little is known about their roles in the pathogenesis of human liver disease. In this issue of HEPATOLOGY, Kahraman et al.13 report that hepatic NK cells, expression of NK cell–associated cytotoxic mediators (such as tumor necrosis factor–related apoptosis-inducing ligand [TRAIL]), NKG2D, and MICA/B messenger RNAs in the liver are significantly elevated in nonalcoholic steatohepatitis (NASH) and, to a lesser extent, in nonalcoholic fatty liver (NAFL), when compared to normal healthy control livers. Immunohistochemical analyses reveal that MICA/B proteins are detected from hepatocytes in patients with NASH and NAFL. Moreover, the expression of MICA/B messenger RNAs are positively correlated with NAS score and hepatocyte apoptosis in NASH. These findings suggest that the MICA/B protein levels are up-regulated in hepatocytes from patients with NASH through mechanisms that have yet to be identified. This is then followed by activation of hepatic NK cells that release TRAIL as a means of killing hepatocytes, and thereby inducing hepatocelluar damage in these patients. It is not clear whether NK cells only kill MICA/B-positive hepatocytes or if they also kill MICA/B-negative cells. This could be confirmed by performing double-staining to determine whether MICA/B protein and terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL+) hepatocytes are colocalized. Whereas the molecular mechanism underlying up-regulation of hepatic MICA/B in NASH and NAFL patients was not explored in the study by Kahraman et al.,13 it has been well documented that expression of NKG2D ligands are up-regulated through a variety of stimuli that have been collectively termed “cellular stress”, such as cellular transformation, viral infection, and/or DNA damage.2, 3 It is known that fat accumulation can generate oxidative stress-induced DNA damage in hepatocytes,14 which may contribute to the up-regulation of hepatic MICA/B in patients with NASH and NAFL.

Interestingly, Kahraman et al.13 also show that levels of MICA/B messenger RNAs correlate positively with stage of fibrosis, suggesting that MICA/B also contribute to the progression of liver fibrosis. However, recent studies using murine models of liver fibrosis suggest that the NKG2D-ligand interaction triggers the killing of activated hepatic stellate cells (HSCs) by NK cells through a TRAIL-dependent mechanism, thereby inhibiting liver fibrosis.15, 16 Liver fibrosis is a common scarring response to virtually all forms of chronic liver injury and is characterized by HSC activation and accumulation of collagen in the liver. In normal healthy livers, HSCs are quiescent, storing large amounts of vitamin A (retinol). During liver injury or culturing on plastic dishes, HSCs become activated and differentiate into myofiboblast cells that produce collagen. Activated HSCs can also become senescent during chronic liver injury or after 9–15 passages in vitro. Quiescent and fully activated HSCs do not express RAE-1, whereas early-activated and senescent-activated HSCs express high levels of RAE-1 or MICA/B and become sensitive to NK cell killing via NKG2D–RAE-1 or NKG2D-MICA/B recognition.15, 16 Additionally, the expression of NK cell inhibitory ligand MHC I is down-regulated,17 whereas TRAIL receptor expression is up-regulated in activated HSCs,18, 19 which may represent additional important mechanisms contributing to increased sensitivity of HCSs to NK cell/TRAIL killing. At present, it is not clear whether expression of NKG2D ligands (MICA/B) is also up-regulated in activated human HSCs during chronic liver diseases. Although Kahraman et al.13 performed immunohistochemical analyses of NASH liver samples with MICA/B antibodies, the expression of MICA/B on the HSCs was not well illustrated. Double-staining with MICA/B antibodies and HSC markers will be required in the future to identify whether MICA/B proteins are expressed on activated HSCs in patients with chronic liver disease.

In addition to having a beneficial effect on liver fibrosis, the interaction of NKG2D and corresponding ligands may also play an important role in immunosurveillance against cholangiocarcinoma and hepatocellular carcinoma. Genetic analyses of NKG2D polymorphisms strongly suggest that interaction of NKG2D-MICA protects against cholangiocarcinoma in patients with primary sclerosing cholangitis.20 The inhibitory effect of NKG2D on liver tumors is likely mediated via activation of NK cells and the subsequent production of TRAIL, which is a potent cytotoxic mediator that kills hepatocellular carcinoma and cholangiocarcinoma cells.21 However, MICA/B proteins are also cleaved proteolytically from hepatocellular carcinoma cells and appear as soluble particles in serum from these patients.22 These soluble MICA/B proteins can down-regulate NKG2D expression and inhibit NKG2D-mediated NK cytotoxicity against liver tumor cells,23 allowing tumor cells to escape from immunity attacks mediated by NKG2D.

In summary, interaction between NKG2D and ligands can trigger NK cell activation, playing an important role in host defenses against hepatitis viral infection and liver cancer development, as well as the inhibition of liver fibrosis (Fig 1). Such activation may also contribute to hepatocellular damage in patients with HCV infection and NAFL/NASH. Additionally, the down-regulation of this activation (e.g., by alcohol drinking) may accelerate the progression of liver disease, including viral infection and liver fibrosis,24, 25 while up-regulation of NKG2D expression on NK cells by poly I:C and IFN-γ can markedly enhance NK cell cytotoxicity against hepatocytes and HSCs, leading to hepatocellular damage and the reduction of liver fibrosis, respectively.11, 15 Interestingly, treating mice with IFN-α also markedly increased expression of NKG2D on liver NK cells (our unpublished data). It will be very interesting to determine whether current treatment protocol using IFN-α in patients with chronic HCV infection also induces NKG2D expression on hepatic and peripheral NK cells. This may be an important mechanism contributing to the well-documented antiviral, antifibrotic, and antitumor effects of IFN-α in patients with chronic liver disease.

Genetic variations in NKG2D and it ligands (such as MICA/B) are known to affect the binding affinity of NKG2D ligands, which can subsequently alter NK cell function. Therefore, genetic variations may be important in explaining spontaneous recovery of acute HCV infection,4 the susceptibility of primary sclerosing cholangitis,26 and cholangiocarcinoma development.20 The study by Kahraman et al.13 highlights an unappreciated mechanism by which the interaction of NKG2D-MICA plays an important role in the pathogenesis of NASH. Therefore, future studies evaluating the association of genetic variants in the NKGD2 and MICA genes with NASH will certainly generate interesting data that could be helpful in the diagnosis and therapeutic treatment of patients with NASH.

Note added in proof:

Since this paper was originally submitted, Ahlenstiel et al27 report that NK cells are activated by IFN-α during chronic HCV infection and contribute to liver damage through TRAIL expression and cytotoxicity. It will be very interesting to investigate whether the interaction of NKG2D-ligand also contributes to NK cell activation during HCV infection.