It is widely recognized that there is a weak adaptive immune response to hepatic antigens, resulting in suboptimal T-cell priming, persistence of hepatotropic pathogens, and relative tolerance of liver allografts. Less emphasized is the fact that the liver possesses a strong innate immune inflammatory response. This is seen experimentally and clinically with liver inflammation due to toxic and metabolic stress, sepsis, and ischemia. The hepatic inflammatory response is much stronger than in other organs given a comparable insult, which gives understanding of it great importance.
The cytosolic sensor NLR family pyrin domain containing 3 (NLRP3) is required for the development of inflammation and fibrosis in response to a diverse range of insults in many organs. Demonstration of a role for the NLRP3 inflammasome in the liver has kept up with the fast pace of discoveries in the field of inflammasome biology, and in the case of fibrosis has preceded findings in other organs.[2-5] The NLRP3 inflammasome has now been demonstrated to have an important role by multiple groups in inflammatory conditions including drug-induced liver injury (DILI), alcoholic steatohepatitis (ASH), nonalcoholic steatohepatitis (NASH), fibrosis, and ischemia reperfusion, which is consistent with its role in other organs.[6-9] In this issue Wree et al. add an important dataset to this topic by demonstrating that mice expressing a constitutively active form of NLRP3 under control of the endogenous NLRP3 promoter develop extensive liver inflammation, hepatocyte pyroptosis, stellate cell activation, and liver fibrosis. The studies in the liver to date have examined NLRP3-deficient mice, and have informed us that NLRP3 is necessary for hepatic inflammation and fibrosis. They do not inform us of the developmental stage when NLRP3 is required, the cell type it is required in, or even if it is directly required or simply provides a necessary cellular environment for inflammation to occur. The dataset from Wree et al. completes the argument by demonstrating that NLRP3 is not only required for hepatic inflammation and fibrosis, but that NLRP3 activation is sufficient for hepatic inflammation and fibrosis. This is very important for a number of reasons. First, Wree et al. use a different experimental approach to reach the same conclusions as the earlier studies. Particular experimental models and specific human diseases will clearly still need to be evaluated on an individual basis, but these findings remove any doubt about the importance of NLRP3 in the general phenomenon of hepatic inflammation and fibrosis. Additional importance of their findings is due to the fact that very large number of stimuli can activate NLRP3. These include changes in extracellular adenosine triphosphate (ATP), Ca++, oxidative stress, metabolic stress, fatty acids, and places NLRP3 in the path of ASH and NASH pathogenesis.[7, 8]
The current study details the development of liver pathology in total body constitutively active NLRP3 mice. Such broad NLRP3 activation results in a systemic inflammatory phenotype which in addition to liver pathology includes neutrophilia and inflammation in many tissues, including the skin and large joints. The systemic inflammatory response does not invalidate the main findings of the study by Wree et al., but requires caution in interpreting them as entirely due to liver intrinsic NLRP3 activation. Certainly the premature death of animals at 2-3 weeks of age may not be due only to liver pathology. Is it possible that the major findings of liver inflammation, hepatocyte pyroptosis, stellate cell activation, and liver fibrosis are due to extrahepatic myeloid cell activation with subsequent accumulation in the liver? This is possible, but due to the significant trafficking of activated immune cells to the liver is very difficult to control for. The stronger phenotype in the total body NLRP3 active mice compared to the myloid-specific NLRP3 active mice suggests that the difference is due to hepatocyte-specific NLRP3 activation, and this can be directly confirmed by examining hepatocyte-specific NLRP3 active mice.
In addition to demonstrating a generic role for NLRP3 in hepatic inflammation and fibrosis, the study by Wree et al. provided some important details. As expected, liver tissue levels of interleukin (IL)-1β and IL-18, the cytokines immediately generated by NLRP3 inflammasome activation, were elevated. Both these cytokines are proximal in the inflammatory cascade and by way of their receptors can initiate a number of other important proinflammatory mediators. This was associated with increased levels of tumor necrosis factor alpha (TNF-α), chemoattractants CXCL1, CXCL2, CXCL5, and infiltration of cells which were predominantly CD45, CD11b and myeloperoxidase-positive, consistent with a neutrophilic infiltrate. In addition to inflammation there were two other important consequences of NLRP3 activity: hepatocyte death and stellate cell activation. Hepatocyte death was of a particular type termed pyroptosis, which has many of the features of apoptosis with DNA damage and terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) positivity. However, in stark contrast to apoptosis, where the intracellular contents are retained during cell death, in pyroptsosis membrane pores develop and intracellular contents are released. This has very important functional consequences. Just as apoptosis is typically considered a noninflammatory silent form of cell death, pyroptosis by release of intracellular contents stimulates, and perhaps sustains, an inflammatory cycle. From a disease perspective pyroptosis has been demonstrated to be important in the killing of intracellular bacteria. Whether pyroptosis has an important role in human liver diseases, or is simply a feature of this experimental system, remains to be determined.
In addition to inflammation and hepatocyte death NLRP3 activation resulted in stellate cell activation and fibrosis. This result is not unexpected, as many other models with inflammation and hepatocyte death have been associated with stellate cell activation and fibrosis. Wree et al., however, went on to develop a mouse line with myeloid-specific constitutive NLRP3 activation which also resulted in liver inflammation, stellate cell activation, and fibrosis but not pyroptotic hepatocyte death. The replication of most of the phenotype in the myeloid-specific mutant strongly suggests that an immune cell localization and activation of NLRP3 is sufficient to drive liver fibrosis. This is consistent with in vitro data on the ability of IL-1β and IL-18 to activate HSC and tissue fibroblasts.[5, 8, 14] It is also consistent with findings in experimental models of DILI, NASH, and ASH.[7, 8]
The ability of IL-1R antagonism to block liver inflammation but not HSC activation and fibrosis in the current model is interesting. Does it predict that we will see similar partial efficacy in human diseases such as NASH? The positive finding of blocking inflammation confirms the importance of IL-1β in driving inflammation due to inflammasome activation. The negative finding of not limiting fibrosis is in contrast to the ability of anakinra to limit fibrosis in other models. This is likely due to the fact that constitutive NLRP3 activation in hepatocytes is directly inducing hepatocyte death by way of pyroptosis, and hepatocyte death can drive HSC activation. In some experimental models hepatocyte death is driven by the inflammatory response and can be limited by IL1R blockade. The inflammatory response is probably a major mechanism of hepatocyte death in many liver diseases, and limiting it will likely be therapeutic for hepatocyte death as well as fibrosis.
Now that the bookends of NLRP3 being required and sufficient for liver inflammation are in place, what is left in the middle? Much still needs to be clarified regarding the cells in which NLRP3 activation is occurring in different disease states. This, and other studies, point towards the bone marrow-derived cell compartment and particularly Kupffer cells. The consequences of NLRP3 activation in hepatic natural killer, natural killer T, and dendritic cells are also of interest and not yet studied. Of great importance is the issue of NLRP3 regulation. Signals 1 and 2 for NLRP3 activation result in a burst of activation, which starts in minutes and has peaked by 24 hours. What is very striking is that attempts to drive the inflammasome by repeated or sustained exposure of macrophages to signals 1 and 2 is not just ineffective, it is counterproductive. Exposure to signal 1, archetypically by lipopolysaccharide (LPS), results in tolerance to LPS and other Toll-like receptor (TLR) ligands (LPS tolerance and heterotolerance, respectively). Sustained exposure to signals 2, such as ATP or crystals, results in cell death. The importance of the inability to drive NLRP3 inflammasome in a sustained manner is that although some diseases such as gout and acetaminophen (APAP) toxicity are acute, many such as atherosclerosis, NASH and ASH are chronic. This suggests that there are important regulators, particularly positive regulators of NLRP3 inflammasome activity, which remain to be identified. Preliminary data suggests that HIF-1α is likely an important proximal regulator. This is consistent with the hypoxic environment in inflamed tissues and of ischemia being conducive to inflammation and fibrosis. Important candidates as regulators are emerging from immunoprecipitation and imaging studies with interferon inducible GTPase such as GBP5 and beta-arresting having direct interaction with NLRP3.[17, 18]
In summary, NLRP3 activation is now confirmed to be required in a number of inflammatory liver models, and also to be sufficient to induce hepatic inflammation and fibrosis. This provides confidence in NLRP3, and molecules along the NLRP3 inflammasome pathway, as therapeutic targets for sterile hepatic inflammatory and fibrotic diseases. For the development of therapies and understanding pathogenesis it shifts the focus to the mechanisms which regulate NLRP3 inflammasome activation, with particular attention to the question of how in many disease states inflammasome activation is sustained beyond a short burst of activity.
Wajahat Z. Mehal, M.D., Ph.D.
From the Section of Digestive Diseases
New Haven, CT