Natural killer T cells: Novel players in biliary disease?

Potential conflict of interest: Nothing to report. This work was supported by the Deutsche Forschungsgemeinschaft (SFB841 and AR 645/3‐1) and the YAEL Foundation. See Article on Page 1249. The pathogenesis of cholestatic liver diseases and, in particular, the mechanisms leading to immune‐mediated bile duct injury as seen in the autoimmune liver diseases, primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC), are incompletely understood (reviewed in a previous work1). The epithelial lining of bile ducts represents a large mucosal barrier toward bile fluid, which directly connects the gut with the liver. On the other side, cholangiocytes are in contact with various nonparenchymal liver cells and portal venous blood draining the intestine. Whereas it has formerly been believed that cholangiocytes are mere victims of immune‐mediated injury, it is now known that they play an active role in regulating the mucosal barrier and the immunological microenvironment. In that respect, cholangiocytes are able to express proinflammatory cytokines, chemokines, Toll‐like receptors (TLRs), as well as human leukocyte antigen class I and II molecules. Natural killer T (NKT) cells are an immune cell population abundant in mouse liver and present in human liver. NKT cells have features of adaptive and innate immunity and may play a role in different human diseases (reviewed in a previous work2). They express a T‐cell receptor (TCR) and rapidly produce cytokines, such as interferon‐gamma (IFN‐γ) and interleukin‐4, poststimulation. NKT cells are best defined by their CD1d‐restricted antigen recognition of lipid antigens, such as glycolipids and phospholipids. CD1d is a major histocompatibility complex class I–like molecule and is expressed by a variety of cells, including dendritic cells, hepatocytes, and B cells. As has been shown for infectious diseases, NKT cells are part of an early line of defense against pathogens. Additionally, NKT cells have also been investigated in cancer and autoimmune diseases.2 A technical limitation of many studies on NKT cells is that they lack specific staining by lipid‐loaded CD1d multimers and are thus often based on the simultaneous detection of NK cell markers and the TCR. However, not all NKT cells express NK markers, such as CD56.3 Another characteristic of NKT cells, which may complicate the interpretation of studies, is their heterogeneity. They can at least be divided into the subgroups of type I or invariant NKT (iNKT) and type II NKT cells with partly diverging biological function.4 iNKT express a TCR with an invariant α‐chain and a limited set of β‐chains. This semi‐invariant TCR of iNKT recognizes the stimulating glycosphingolipid, α‐galactosylceramide (αGalCer).2 It is not fully clear whether αGalCer mimics a relevant foreign or self‐antigen. It was isolated in search for molecules with antitumor activity and then discovered to be specifically recognized by iNKT cells. Type II NKT cells possess a more diverse TCR repertoire and recognize the self‐glycolipid 3‘‐sulfated β‐galactosyl ceramide (called sulfatide) by CD1d. Murine studies point to a rather proinflammatory role of iNKT and a rather regulatory role of type II NKT cells.4 Type II NKT cells were described later than type I NKT cells and therefore more knowledge has thus far accumulated on the function of iNKT cells. NKT cells have been described to contribute to the pathogenesis of diverse liver pathologies, such as viral hepatitis3 and hepatic ischemic reperfusion injury.5 For instance, in the mouse model of ischemic liver reperfusion injury, type I NKT cells aggravated, whereas type II NKT cells protected from reperfusion injury by inactivation of type I NKT cells, inhibition of IFN‐γ production and recruitment of myeloid‐derived inflammatory and NK cells.5 In this issue of Hepatology, Schrumpf et al. present a study investigating the CD1d expression on different human and murine cholangiocyte cell lines and on human primary cholangiocytes.6 The researchers show that CD1d is expressed on cholangiocytes and is able to stimulate iNKT cells by their prototypical antigen, αGalCer. Explant livers of patients with PBC, PSC, alcoholic liver disease, and healthy control tissue obtained from resected livers were analyzed with regard to their expression of CD1d on large and small bile ducts. All explanted cirrhotic livers revealed a lower expression of CD1d, compared to healthy control livers. However, there were differences in the expression patterns of CD1d on bile ducts between PSC and PBC patients, with stronger CD1d signals on large bile ducts in PSC. The researchers could not show a correlation between total bilirubin and CD1d expression, but an analysis of the influence of bile acids on CD1d expression was not performed. There have only been few studies investigating NKT cells and CD1d expression in human cholestatic liver disease. In early PBC, the expression of CD1d on small bile ducts was reported to be enhanced, compared to PSC and normal livers.7 These results are not in line with the current study by Schrumpf et al.6 One explanation for this discrepancy might be that Schrumpf et al. analyzed explanted end‐stage livers in contrast to early disease stage in the former study. Kita et al. quantified iNKT cells in PBC livers and blood using tetramer staining.8 The intrahepatic frequency of iNKT was found to be raised in end‐stage PBC patients, compared to healthy livers, whereas there were no differences of iNKT numbers in peripheral blood between PBC patients and healthy controls. Specific analyses of NKT cells in human PSC have not been conducted as yet. The study by Schrumpf et al. clearly advances our knowledge on the potential role of CD1d and NKT cells in immune‐mediated cholestatic liver diseases.6 However, important questions remain to be answered in order to appreciate the role of these cells and the function of CD1d in human disease: What kind of lipid antigens (bacterial, alimentary, or toxic metabolites) are presented by cholangiocytes on CD1d? How is the expression of CD1d regulated, including the potential influence of bile acids? How do NKT cells respond to stimulation by cholangiocytes (e.g., cytokines, homing, or kinetics)? Do activated NKT cells contribute to bile duct destruction directly or indirectly, that is, by attracting other inflammatory cells? Is there a differential contribution of type I and II NKT cells in human disease? To date, many of these questions are difficult to answer owing to technical limitations, such as the inability to stain the different types of NKT cells immunohistochemically. If NKT cells are of pathological relevance in cholestatic liver diseases, they may harbor the potential for therapeutic immune modulation. Interestingly, in the mouse model of concanavalin A–induced hepatitis, administration of a synthetic agonist of the nuclear bile acid receptor, farnesoid X receptor (FXR), attenuated liver injury.9 This protection was mediated by FXR‐induced down‐regulation of osteopontin‐expression in NKT cells. It is tempting to speculate that FXR agonists, such as obeticholic acid, which are currently evaluated in clinical trials for treatment of PBC and nonalcoholic steatohepatitis, might be able to modulate the function of NKT cells in cholestatic liver disease. Notwithstanding the interest in NKT cells in human liver diseases, one should bear in mind that this cell type shows a much lower abundance in healthy human liver (approximately 1% of liver lymphocytes) than in murine liver (approximately 30%).2 Still, it seems possible that even a few NKT cells might trigger liver inflammation.10 The study by Schrumpf et al. is to be applauded for investigating human samples,6 but specific quantification and subgroup analyses of NKT cells in human livers are rare and clearly require further studies. A better understanding of CD1d‐mediated antigen presentation by cholangiocytes and their effect on the various NKT cell subtypes might help to elucidate the pathogenesis of cholestatic liver diseases. Author names in bold designate shared co‐first authorship.