Shivakumar P, Sabla GE, Whitington P, Chougnet CA, Bezerra JA. Neonatal NK cells target the mouse duct epithelium via Nkg2d and drive tissue-specific injury in experimental biliary atresia. J Clin Invest 2009;119:2281–2290. (Reprinted with permission).
Biliary atresia is a neonatal obstructive cholangiopathy that progresses to end-stage liver disease. Although the etiology is unknown, a neonatal adaptive immune signature has been mechanistically linked to obstruction of the extrahepatic bile ducts. Here, we investigated the role of the innate immune response in the pathogenesis of biliary atresia. Analysis of livers of infants at diagnosis revealed that NK cells populate the vicinity of intrahepatic bile ducts and overexpress several genes involved in cytotoxicity. Using a model of rotavirus-induced biliary atresia in newborn mice, we found that activated NK cells also populated murine livers and were the most abundant cells in extrahepatic bile ducts at the time of obstruction. Rotavirus-primed hepatic NK cells lysed cholangiocytes in a contact- and Nkg2d-dependent fashion. Depletion of NK cells and blockade of Nkg2d each prevented injury of the duct epithelium after rotavirus infection, maintained continuity of duct lumen between the liver and duodenum, and enabled bile flow, despite the presence of virus in the tissue and the overexpression of proinflammatory cytokines. These findings identify NK cells as key initiators of cholangiocyte injury via Nkg2d and demonstrate that injury to the duct epithelium drives the phenotype of experimental biliary atresia.
Current theories of the pathogenesis of biliary atresia propose that the initial bile duct injury is likely to be the consequence of perinatal infection with a cholangiotropic virus. Resultant alteration of the bile duct or intrahepatic microenvironment and/or cross-reactivity between viral and self-antigens triggers secondary autoimmunity and inflammation, which ultimately progresses to ductopenia, fibrosis, and biliary obstruction.1 Reovirus, rotavirus, and herpesviruses including cytomegalovirus have all been considered possible candidates for the role of initiating agent. Clinical studies aiming to detect serological evidence of exposure to these viruses, or their continued presence in the tissues of affected infants at the time of diagnosis, have yielded conflicting results. In contrast, murine neonatal infection with either reovirus serotype 3 or rhesus group A rotavirus (RRV) reproduces biliary obstruction with histological features recapitulating those of human disease. Of note, susceptibility to the development of biliary inflammation and obstruction after infection is restricted to the early neonatal period, and by the time the clinical picture of obstructive jaundice is evident, the virus is no longer detectable in the liver.1
Both human and murine experimental biliary atresia share a similar inflammatory signature, with a mixed lymphocytic infiltrate and predominant expression of Th1 cytokines such as interferon-gamma (IFN-γ) and interleukin-12 (IL-12), and chemokines including chemokine (CXC motif) ligand 9 (CXCL9) and CXCL10, which are important in the recruitment of cytotoxic effectors. T cell receptor Vβ usage analysis of infiltrating lymphocytes suggests an oligoclonal expansion in response to unknown biliary epithelial antigens.2 Both CD8+ and CD4+ T cells proliferate and produce IFN-γ after experimental rotavirus infection, and macrophages are activated to produce effector molecules including tumor necrosis factor-α and inducible nitric oxide synthase.3 Depletion of CD4+ cells from RRV-exposed mice has little effect upon disease progression, whereas CD8+ cell depletion after exposure prevents the development of biliary obstruction in the majority of mice.4 Adoptive transfer of virus-primed CD8+ T cells leads to cholangitis in the recipient, while not reproducing the phenotype of biliary atresia,4 suggesting that these cells are necessary, but not sufficient for the full expression of disease.
Here, Shivakumar et al.5 demonstrate the presence of abundant, activated natural killer (NK) cells in the livers of infants diagnosed with biliary atresia, as well as those of RRV-infected mice. Either NK cell depletion with anti-asialo GM1 or blockade of the activating receptor NK group 2, member D (NKG2D) soon after RRV injection abrogates bile duct injury and obstruction. In this model, damage to the bile duct epithelium is evident at 3 days after RRV inoculation, and this appears to be a critical checkpoint in NK-dependent disease progression. Depletion of NK cells prior to bile duct damage prevented obstruction, whereas NK cell depletion beyond this point was markedly less effective in disease prevention. Interestingly, depletion of CD8+ cells prevented the intrahepatic accumulation of NK cells at 7 days after RRV exposure, although the effect upon NK cell numbers at the critical early time before day 3 was not reported. CD8+ T cell–NK cell cooperation has been clearly demonstrated in antitumor responses, and may involve both direct interactions and effects mediated via cross-talk with dendritic cells (DCs). Possible mechanisms by which activated cytotoxic T lymphocytes (CTLs) can provide “help” for NK cells include the secretion of chemokines and cytokines which promote NK cell recruitment, proliferation, survival, and cytotoxic capacity, and the augmentation of tissue damage, with further expression of activating stress ligands and release of endogenous ligands for Toll-like receptors, which may also directly activate NK cells.6 Conversely, secretion of IFN-γ by activated NK cells up-regulates IL-12 production by DCs, and removes the requirement for CD4+ T cells in the priming of antitumor CD8+CTLs.7 NK cells infiltrating tumors can potentiate CD8+ T cell cytotoxicity at the local level, independent of DCs, but by a mechanism also involving secretion of IFN-γ.8 NK-stimulated CD8+CTLs in the tumor infiltrate express high levels of perforin and granzyme, and these T cells, rather than NK cells are the principal cytotoxic effectors. Early NK cell depletion can prevent CD8+CTL activation, resulting in tumor overgrowth.8
Differences in the way the immune system reacts to viral infection between the early neonatal period and later stages of life may influence susceptibility to NK cell–dependent cholangiocyte lysis following RRV exposure. One important difference may involve the regulation of expression of NKG2D ligands by cholangiocytes. NKG2D ligand expression on a variety of cell types can be induced by Toll-like receptor (TLR) signaling. Mouse and human macrophages and mouse renal tubular epithelial cells express NKG2D ligands in response to engagement of TLR49 (Chen et al., unpublished data), whereas TLR3 signaling results in NKG2D ligand expression on human muscle cells.9 Mouse and human cholangiocytes also express TLR3, the receptor for double-stranded viral RNA (dsRNA), and both reovirus and rotavirus are members of the dsRNA virus family Reoviridae.10 Neonatal mice display heightened sensitivity to TLR3 stimulation, producing increased amounts of proinflammatory cytokines in response to either stimulation with the synthetic dsRNA analogue poly(I:C), or infection with the dsRNA virus MHV-A59.11 It is tempting to speculate that neonatal, but not adult, infection with Reoviridae leads to strong up-regulation of NKG2D ligands on biliary epithelial cells via engagement of TLR3, rendering them susceptible to NK cell/NKG2D-mediated injury (Fig. 1).
The findings reported by Shivakumar et al.5 represent an important advance in our understanding of the pathogenesis of biliary atresia, but this may not translate directly into a therapeutic strategy. Because the critical NK cell–mediated events occur so early in the course of the disease, it is unlikely that treatments based on NK cell depletion or inhibition would be able to be used in time to interrupt disease progression. Furthermore, although an underlying viral etiology seems likely, the identity of the virus and the exact timing of infection (intrauterine versus neonatal) remain uncertain, making vaccination as a primary prevention strategy unfeasible at present. However, interventions which target the sequelae of NK-dependent injury may be of benefit. Whereas apoptosis of the bile duct epithelium in the early stages of biliary atresia may be largely the consequence of immune attack, other factors, such as the increased exposure to cytotoxic bile acids that accompanies cholestasis, may also contribute to the continued increased level of cholangiocyte apoptosis which culminates in ductopenia.12 Moreover, in both hepatocytes and cholangiocytes there is synergy between the extrinsic and intrinsic apoptotic pathways in bringing about cell death.13 Recent work suggests that biliary damage could be ameliorated by harnessing an innate antiapoptotic signal and an existing therapeutic agent. Cholangiocytes, like pancreatic ductal epithelial cells and β cells express both glucagon-like peptide 1 (GLP-1) and its receptor, and levels of GLP-1 are increased in biliary atresia. Signaling through GLP-1R operates via the phosphotidyl inositol 3-kinase pathway to prevent mitochondrial translocation of Bax and release of cytochrome C, thus preventing intrinsic pathway apoptosis.14 Endogenous GLP-1 is rapidly inactivated by the dipeptidyl peptidase IV (DPPIV) produced by hepatocytes and cholangiocytes15 but synthetic DPPIV-resistant GLP-1 analogs are now available and in clinical use for the treatment of type 2 diabetes. Administration of one of these agents has been shown to prevent both apoptosis in cholangiocytes exposed to glycochenodeoxycholic acid in vitro, and bile duct loss in a rodent model of cholestasis.14 Further studies of these agents in other cholangiopathies is urgently warranted.