|Home | About | Journals | Submit | Contact Us | Français|
NK cells represent a large proportion of the lymphocyte population in the liver and are involved in early innate immunity to pathogen infection. As a result of liver endothelial cell fenestrations, parenchymal cells are not separated by a basal membrane, and thereby pathogen-infected hepatocytes are extensively capable of interacting with innate immune cells including NK cells. In addition, hepatic NK cells interact with surrounding DC and alter their differentiation and function. Recent studies reveal that NK cells exhibit a regulatory function that modulates T cell responses through their interaction with DC and/or direct effect on T cells. Thus, NK cells play a central role, not only in innate immunity, but also in shaping the adaptive immune response. During pathogen infection, there is a remarkable increase of hepatic NK cells, possibly due to the expansion of resident liver NK cells and/or recruitement of NK cells from the blood. The liver microenvironment is believed to modulate hepatic NK cell function through the induction of activating/inhibitory receptor expression and inflammatory cytokine secretion. Particularly, the liver maintains intrahepatic NK cells in a functionally hyporesponsive state compared to splenic NK cells: liver NK cells displayed a dampened IFN-γ response to IL-12/IL-18 stimulation. Notably, the liver contains a significant population of functionally hyporesponsive NK cells that express high levels of the inhibitory receptor NKG2A and lack expression of MHC class I-binding Ly49 receptors. Importantly, adoptively transferred splenic NK cells that migrate to the liver displayed phenotypic and functional changes, supporting a view that the liver environment modifies NK cell receptor expression and functional responsiveness. In this article, we will review studies on the regulation of NK cell repertoire and function in the hepatic environment and the impact of liver NK cell immunoregulatory function on influencing adaptive immunity.
The liver represents a unique immunological environment, as it is able to endure daily exposure to gut-derived foreign antigen without causing inflammation.1,2 The immune tolerance of the liver has been noted for having successful allograft transplants without the administration of immunosuppressive agents. Despite this tolerogenic environment, the liver rapidly and effectively switches from an environment where T cell engagement typically results in immune tolerance and/or cell death to a strong proinflammatory site upon infection, leading to pathogen clearance. However, several pathogens, including hepatitis B and hepatitis C viruses, and malaria exploit liver tolerance mechanisms in attempts to avoid immune clearance and establish persistent infection. An in-depth understanding of the regulatory mechanisms responsible for the initial state of immune tolerance may shed light on identifying potential factors targeted by persistent pathogens.
Natural killer (NK) cells constitute a large proportion of the lymphocyte population in the liver and are involved in regulating hepatic immunity to pathogen infection. Importantly, the phenotype and function of hepatic NK cells (both resident and recruited/blood-derived NK cells) are altered in the liver microenvironment: hepatic NK cells exhibit unique NK cell repertoire and cytokine profiles.3 Therefore, the alteration of hepatic NK cell subsets and their function may contribute to the establishment/maintenance of persistent infections. Furthemore, through cell-to-cell interaction and secretion of cytokines, NK cells can regulate the differentiation and function of dendritic cells (DCs), thereby affecting T cell responses. In this review, we will provide a brief discussion of liver NK cell subsets with distinct effector functions and describe the potential immunoregulatory role of NK cells in shaping the adaptive immunity.
NK cells were initially defined as lymphocytes that could respond nonspecifically to transformed or virally infected target cells without prior sensitization. The majority of NK cells develop in the bone marrow from NK cell precursors (NKPs) while the development of NK cells also appears to occur at other lymphoid sites, such as the thymus and lymph nodes.4 NKPs are able to differentiate solely into NK cells, and both immature and mature NK cells express NK1.1. The earliest committed NKP population in the mouse bone marrow is defined as CD49b−CD161−CD122+.5 NKPs develop into immature NK cells, which express CD161 (NK1.1) and maintain CD122 expression.6,7 Immature NK cell development follows a stepwise acquisition of phenotypic markers, including the expression of CD94/NKG2 receptors, followed by Ly49 receptors and CD49b (Dx5). In addition, the expression of CD11b and CD43 increases as the maturation of NK cells progresses.8,9,10
Immature NK cells that develop into mature NK cells have recently been characterized as either educated (also termed “licensed”) or uneducated.11 An educated NK cell is defined as possessing full functional capacity; that is, NK cells respond to stimulation of activation receptor triggering. NK cell education appears to be regulated by signaling through MHC class I receptors, as NK cells lacking the expression of self-specific MHC class I receptors are functionally hyporesponsive.12,13 Indeed, NK cells isolated from MHC class I deficient mice display a lower functional capacity compared to NK cells from wild-type mice. Furthermore, the ITIM (located on the cytoplasmic tail of MHC class I inhibitory receptors) has been demonstrated to be critical for full functional capacity of NK cells.13 The engagement of self-specific MHC class I receptors is critical in maintaining NK cell tolerance in the face of activating receptor signaling. Taken together, host MHC class I molecules are crucial for regulating NK cell responsiveness and receptor repertoire formation.
At present, multiple models have been postulated for the functional maturation of NK cells (see Fig. 1).14,15 Two of these models, arming and disarming, are based on the idea that NK cell education, or licensing, is aimed at ensuring that a functional NK cell is equipped with an appropriate inhibitory receptor whose signaling acts as a balance to regulate signaling via activating receptors. The arming model suggests that direct signaling through a MHC class I receptor supplies a signal that conveys functional capacity on the NK cell. The disarming model, similar to the concept of T cell exhaustion, suggests that in the absence of inhibitory signaling via the MHC class I receptors, continual signaling through activating receptors renders the NK cell hyporesponsive. A more recently proposed third model, rheostat, satisfies both arming and disarming by suggesting that NK cells can be educated through the strength of inhibitory receptor signaling (Fig. 1). Tuning the NK cell with a higher inhibitory input can lead to better IFN-γ production, whereas potential granzyme release can be triggered at lower input.16 To date, the data is insufficient to determine the precise role of each model in acquring NK cell effector function.
NK cells are relatively enriched in the liver and are embedded in the endothelial lining of the liver sinusoids. Early reports described a liver-specific NK cell population from rat as “pit cells,” due to the presence of a large number of intracellular granules. In addition, rat liver NK cells have been subclassified based on the density and size of their granules as 1) high density-large granules, and 2) low density-small granules. It appears that NK cells with high density-large granules derived from the blood migrate into the liver and further differentiate into low density-small granular hepatic NK cells.17 However, the mechanism(s) for NK cell migration to the liver and factor(s) involved in altering NK cell function are not well understood. Further studies are needed to define these questions.
Importantly, differential functional capacity has been attributed to distinct human NK cell subsets (CD16−CD56Bright or CD16+CD56Dim NK cells) based on the expression of CD16 and CD56. CD16-CD56Bright NK cells may give rise to CD16+CD56Dim NK cells, as suggested by shorter telomere length in CD16+CD56Dim NK cells.18 CD16+CD56Dim NK cells are highly cytotoxic, and make up the bulk of the NK cell population in human peripheral blood (>90%). In contrast, CD16−CD56Bright NK cells are poorly cytotoxic but produce a broad range of cytokines (IFN-γ, TNF-α, GM-CSF, IL-10).19,20 Recently, the surface expression of CD11b and CD27 have been used to define the maturation status of mouse NK cells.21 Adoptive transfer experiments indicated a linear progression of maturation from CD11b−CD27+, to CD11b+CD27+, to CD11b+CD27−. With regard to function, CD11b−CD27+ NK cells produce higher amounts of cytokines and show decreased cytotoxicity compared to CD11b+CD27+ NK cells. In contrast, CD11b+CD27− NK cells express higher levels of killer-cell lectin-like receptor subfamily G, member 1 (KLRG1), which is induced following NK cell activation and proliferation.22 In addition, CD11b+CD27− NK cells are predominant in the blood, while CD11b−CD27+ NK cells are found at a greater proportion in the liver.3
However, it is not known how unique NK cell subsets exist at different tissue sites, particularly in the liver compartment. Thus, future studies on determining factor(s) involved in NK cell differentiation and maturation at various tissue sites would be important for advancing our understanding of NK cell biology and their impact on mounting the immune response. In addition, it will provide useful insight into designing immunotherapeutics for treatment of infectious disease and cancer patients.
Besides the critical role for NK cell education through MHC class I receptors, NK cell activation is regulated by the balance of anti- and pro-activation signals mediated through the expression of inhibitory and activating receptors (Ly49 and NKG2) (see Fig. 2). In mice, several NK cell receptors have been identified in the Ly49 receptor family.23 Ly49 receptors are type II transmembrane glycoproteins that are expressed on NK cells and form disulfide linked homodimers, with the majority of NK cells expressing one to four individual Ly49 receptors.24 Most of the 19 Ly49 receptors identified are inhibitory receptors (incorporating one ITIM) and recognize MHC class I and MHC class I-like molecules.25 In contrast, many of the ligands for the activating Ly49 receptors remain unknown. While structurally different from Ly49 receptors, a family of functional homologues of the Ly49 receptor molecules is expressed on human NK cells, referred to as the killer cell immunoglobulin-like receptor (KIR) family, and includes both activating and inhibitory receptors. Importantly, there is a high genetic diversity in the KIR and Ly49 receptor families, where the genes vary in number, genomic organization, and allelic polymorphism among individual haplotypes.26
The NKG2 group of receptors are C-type lectin-like receptors that are present in both humans and mice.27 The NKG2 receptors include NKG2A, C, and E, (with NKG2F also expressed in humans) and are expressed at the surface as heterodimers with CD94. NKG2A is an inhibitory receptor, with two ITIMs present in the cytoplasmic tail, while NKG2C and E receptor engagement results in an activating signal via the adaptor protein DAP12.28 In mice, NKG2A is expressed on most NK cells, while NKG2C and E are expressed at significantly lower frequencies. The ligand for NKG2A is the nonclassical MHC class I molecule Qa-1b (human HLA-E), whose expression is dependent on peptides derived from classical MHC class I molecules and thus acts as an overall measurement of MHC class I expression.29,30
NKG2 receptor expression is regulated by the inflammatory stimuli. In particular, IL-10 and TGF-β are able to induce the expression of NKG2A.31 Thereby, the local microenvironment may influence NK cells by modulating their response to infection. Within the liver, such NK cell modulation may contribute to liver tolerance, and may facilitate persistent infection by certain liver-tropic pathogens. Indeed, studies on the functional competence of liver NK cells by the expression of various NK inhibitory and activating receptors demonstrate that the liver contains a prominent subset of NKG2A+ NK cells that lack Ly49 receptor expression. Importantly, this NKG2A+Ly49− NK cell subset is hyporesponsive to IL-12/IL-18 stimulation in the liver compartment. Additionally, when adoptively transferred splenic NK cells migrate to the liver, they more closely resemble the phenotype and function of liver resident NK cells.3 These results support a view that the liver environment can modify NK cell receptor expression and responsiveness to cytokine stimulation, and/or may preferentially retain immature NK cells within the liver.
Although initially described as being part of the NKG2 group of receptors, NKG2D has very little homology to the other NKG2 receptors. NKG2D does not bind CD94, but is expressed as a homodimer on both mouse and human NK cells.32 NKG2D is an activating receptor, and recognizes ligands induced by cellular stress on virally infected or transformed cells.33 In mice, two isoforms of NKG2D are present, characterized by either a long or short cytoplasmic tail.34 The long-tailed isoform associates with the adaptor protein DAP10, while the short-tailed isoform can associate with DAP10 or DAP12. However, the expression of only the long-tailed isoform of NKG2D has been reported in humans. Notably, NKG2D is expressed by most NK cells and the cytokine environment can modulate the level of expression: IL-15 and TNF-α increase the expression of NKG2D, whereas TGF-β downregulates NKG2D expression. Importantly, the KIR, Ly49, and NKG2 receptors on NK cells are stochastically expressed, resulting in heterogenous expression of NK cell receptors on NK cells.35,36,37
The natural cytotoxicity receptors (NCR) are activating Ig-like transmembrane glycoproteins. While the NCRs consist of NKp30, NKp44, and NKp46 in humans, mice express only NKp46 and have a NKp30/Ncr3 pseudogene. NKp46 has two extracellular C2-type Ig-like domains, and a positively charged transmembrane region that associates with the ITAM containing CD3ζ and FcR′γ.38,39,40 It appears that the majority of normal resting target cells either express low-level or no NCR-ligands.39 Although no cellular NCR ligands have been formally identified, the extracellular form of the intermediate filament protein vimentin may be associated with NKp46 cross-linking.41,42,43 NKp46 may also bind membrane-associated heparan sulfate proteoglycans and has been shown to bind hemagglutinin-neuraminidase from influenza and sendai viruses.39,44
The main NK cell effector functions include direct cell killing of infected or transformed cells by perforin, granzymes, and TRAIL as well as the secretion of a variety of pro- and anti-inflammatory cytokines.45,46 Notably, cytokine production by NK cells can be regulated through both activating and inhibitory receptors: activating receptor engagement leads to production of IFN-γ, which plays a critical role in shaping the subsequent adaptive immune response. In addition to the cross-linking of activating receptor, NK cell activation can also be induced by various cytokines, including type I IFNs, IL-2, IL-12, IL-15, and IL-18. Distinct NK cell functions have been attributed to stimulation with different cytokines having some synergistic effects. Type I IFNs lead to NK cytotoxic activity, while IL-12 and IL-18 are potent inducers of IFN-γ production.34,47 In addition, IL-15 is involved in NK cell proliferation and appears to be important in NK cell “priming” through trans-presentation of membrane-bound IL-15 on the surface of DCs.
Importantly, NK cells express TLRs crucial for the recognition of pathogen-associated molecular patterns (PAMPs).48 TLR stimulation on NK cells can lead to IFN-γ production, as well as the up-regulation of chemokines CCL3, CCL4, and CCL5.49 Particularly, TLR-3 activates NK cells and increases NK cytotoxicity along with induction of cytokine production.40 In addition, despite the upregulation of NKG2D by TLR stimulation, the induction of NKG2D ligands on macrophages is specific to TLR-4, but not TLR-3, stimulation.50 Because of the high concentration of incoming LPS from the gut, this could be potentially harmful to the liver and thereby it is likely that the counteractive mechanism would be of importance to keep the low level of NKG2D expression by NK cells in the liver. Recently, the crosstalk between complement and the TLRs has been reported to inhibit IL-12 production and this could play a role in dampening excessive NK cell activation. Given that the liver is a major site for the production of complement protein, the combinatorial effect of both TLR and complement could be of significant importance for influencing NK cell function in the liver.51
As hepatic tissue macrophages (ie, Kupffer cells) are critical in endotoxin removal from the blood, it is likely that Kupffer cells may play an important role in maintaining liver tolerance. Although they can be found throughout the liver, Kupffer cells are mainly localized around the periportal regions. A recent study demonstrated that LPS treatment of Kupffer cells resulted in the higher level of anti-inflammatory cytokine IL-10 production compared to that in polyI:C treatment, while IL-12 was produced at similar levels in response to both TLR ligands.52 This study also indicated an important role for IL-10 in maintaining liver tolerance, and further demonstrated that IL-10 production by Kupffer cells led to decreased NK cell activation. Based on the finding that NK cell killing of microglia in the CNS is regulated via NKG2A/Qa-1b interactions in the EAE mouse model,53 it is tempting to speculate that the high expression of NKG2A could play an equally important role in the regulation of liver NK cell function. Taken together, NK cells are regulated via a balance of potentially continuous sensing of their environment through various activating and inhibitory receptors, with activation occurring when activating signals overwhelm the inhibitory signals, either through decreased inhibitory signaling or increased activating signaling.
Hepatic fibrosis is a would-healing response to injury, commonly seen in chronic liver disease, and most often attributed to cell death within the liver. Fibrosis can be characterized by the formation of new blood vessels (angiogenesis), sinusoidal remodeling, and pericyte (stellate cell) expansion.54 Hepatic stellate cells (HSCs) are activated through the phagocytosis of apoptotic bodies and cell debris, leading to their differentiation, proliferation and pro-fibrotic progression. Therefore, HSC death is believed to deliver anti-inflammatory and anti-fibrotic responses. Primarily, NK cells appear to be involved in HSC death and thereby depletion of NK cells leads to greater observed fibrosis in the murine fibrosis model.
Quiescent HSCs are relatively resistant to apoptotic signals, whereas activated HSCs are more prone to cell death. Of the many pro-apoptotic molecules investigated in activated HSCs, the upregulation of TRAIL receptor may be an essential NK cell-mediated cytolytic ligand.55 Interestingly, activated HSCs also upregulate a variety of intermediate filaments, including desmin and vimentin, whose levels increase strongly between day 2 and 6 in primary cultures, but only vimentin levels remain consistently high.56 If extracellular vimentin is indeed an activating ligand for the NCR NKp46, it would be an intriguing mechanism behind NK cell-mediated killing of HSCs.
Recently, our understanding of the role of NK cells in regulating the overall host immunity has increased dramatically. Through the production of a wide range of cytokines, NK cells contribute to enhance the immune response by triggering the activation of other innate immune cells and inducing the proinflammatory responses. NK cells also play a critical role in shaping the adaptive immune response through optimal T cell activation via direct cell-to-cell interaction with DC. As a result of NK-DC crosstalk, DC mature and secrete cytokines/chemokines, resulting in productive T cell responses.47,57 In contrast, NK cells have been shown to dampen the adaptive immune responses by selective killing of macrophages, DC, and/or T cells.58,59 Thus, NK cells exhibit numerous functions to regulate host immunity. However, it is yet to be determined whether these multiple immunoregulatory functions of NK cells are exerted by specific NK cell subsets or functional changes of overall NK cells. Nevertheless, the understanding of the mechanism(s) involved in the regulation of host immunity by NK cells is crucial for designing immunotherapeutic agents to boost host immunity to viral infection and tumor.
Given a large proportion of NK cells in the liver lymphocyte populations, NK cells may help maintain liver tolerance through their interactions with various cell types. Interestingly, NK cells cocultured with hepatocytes appear to alter the ability of DC to prime CD4+ T cells, resulting in a regulatory T cell phenotype and function.31 In addition, DC induction of a T cell regulatory phenotype is dependent on NKG2A engagement on NK cells during coculture with hepatocytes.60 NKG2A expression is reportedly increased on NK cells from chronic HCV patients, suggesting a role for NKG2A in persistent viral infection.61,62 These results indicate that liver NK cells might lead to inhibition of antiviral T cell responses and facilitate pathogen persistence.
Although there are conflicting reports, several studies have demonstrated the alteration in NK cell population from chronic HCV patients, including frequency, phenotype and function. It has also been reported that freshly isolated NK cells from HCV patients show significant production of IL-10.63 In addition, NK cells from chronic HCV patients are less capable in activating DCs, compared to control NK cells.64 Recently, it was suggested that NK cells may play a central role in the ability of HCV to establish persistent liver infection, and that the NK-DC interaction may be at the core of the immunomodulatory effects targeted by HCV.64 Furthermore, there has been recent evidence suggesting that NKp30 expressing NK cells may contribute to innate resistance and help reduce HCV infection.65 Given the central role that DC play in driving a strong and competent antiviral immune response, and the ability of NK cells to influence DC maturation, dysregulation of the NK-DC interaction in favor of a weakened T cell response would be an effective means of establishing a persistent infection. It is plausible that HCV and other persistent liver pathogens may take advantage of tolerance mechanisms already in place within the naïve liver, thus emphasizing the importance of investigating such mechanisms that contribute to a tolerogenic liver environment. Therefore, it is possible that NK cells are capable of influencing adaptive immunity via two potential mechanisms: 1) affecting APC (eg, DC) function, 2) direct effect on T cell function (see Fig. 3).
As a first possibility for the T cell modulation by NK cells, several studies have indicated that the liver DC population display an immature phenotype and are poor T cell activators. Human liver DC have been shown to secrete more IL-10, as compared to skin DC, and were found to be less capable in stimulating T cell proliferation, leading to IL-10-producing T cells.66 In contrast, skin DC activation of T cells resulted in greater proliferation and more IFN-γ production. Interestingly, DC isolated from human hepatic lymph nodes have also been characterized by high IL-10 production and low allogenic stimulatory capacity, compared to human inguinal lymph node DC.67 Murine liver DC have also been described as less capable of stimulating T cells, compared to spleen DC.68 This may be due in part to differences in subset composition. Indeed, the liver reportedly contains a large proportion of plasmacytoid DC, as well as a B220-CD11chigh and B220-CD-11clow DC subsets, expressing high and low levels of costimulatory molecules, respectively.69 Secondly, recent evidence has given a direct role of NK cells to the cytolysis of activated T cells. Interestingly, apoptotic cells can upregulate NKp46 expression by NK cells, leading to increased natural cytotoxicity toward vementin+ CD4 T cells, which could be interesting due to the high numbers of apoptotic T cells found in the liver.43 On the other hand, IL-10 is able to block vementin secretion by macrophages.41 However, a direct effect of liver resident NK cells on killing T cells in vivo has yet to be determined.
The tolerogenic nature of the liver allows daily exposure to gut-derived foreign antigen without causing inflammation, but may facilitate persistent infection in the liver. Natural killer (NK) cells play a central role, not only in innate immunity, but also in shaping the adaptive immune response. The liver maintains intrahepatic NK cells in a functionally hyporesponsive state. Compared to splenic NK cells, liver NK cells display a dampened IFN-γ response to IL-12/IL-18 stimulation. In addition, the liver contains a significant population of functionally hyporesponsive NK cells that express high levels of the inhibitory receptor NKG2A and lack expression of MHC class I-binding Ly49 receptors. Based on the finding that adoptively transferred splenic NK cells that migrate to the liver displayed phenotypic and functional changes, the liver environment modifies NK cell receptor expression and functional responsiveness. Future studies need to be done to define key factor(s) contributing to the regulation of liver NK cells in part by maintaining a greater percentage of the hyporesponsive NKG2A+Ly49− NK cells.
We thank the members of Hahn laboratory for helpful discussions. This work was supported by the grants from NIH (R01DK063222, U19AI083024) to YSH.