Mature peritoneal macrophages take an avascular route into the injured liver and promote tissue repair.

Abstract

Sterile liver injury triggers a complex host immune response during which macrophages contribute to both tissue inflammation and tissue regeneration. Two macrophage populations have been identified: liver-resident macrophages (Kupffer cells), which are located in the sinusoids, and macrophages, which differentiate from recruited blood monocytes. Wang and Kubes1 now identify peritoneal macrophages as a third population to respond to liver injury. These cells do not depend on blood vessels but migrate directly through the mesothelium that covers the liver. They arrive at the injury site within 1 hour and contribute to liver regeneration by dismantling the nuclei of necrotic cells, thereby releasing DNA to cover the necrotic area. Due to its location and unique anatomy (Fig. 1), the liver is readily accessible to immune cells. It is a heavily perfused organ that receives arterial blood from the abdominal aorta and venous blood from the peribiliary plexus and the splanchnic organs. About 30% of the total blood passes through the liver every minute, carrying about 108 peripheral blood lymphocytes in 24 hours.2 Extravasation of immune cells from the blood is facilitated by the low pressure and slow blood flow in the liver sinusoids and by fenestrations in the sinusoidal endothelial cells and the lack of a basal membrane. It is therefore not surprising that any type of liver injury results in the rapid recruitment of innate immune cells from the blood. In their model of sterile liver injury, Wang and Kubes touch the liver of anesthetized mice with a thermal probe to create a sterile lesion of about 0.02 mm3 size. Hepatocyte necrosis results in the release of damage-associated molecular patterns such as adenosine triphosphate (ATP) from the cytosol and DNA and N-formyl peptides from disrupted mitochondria that trigger sterile innate immune responses. Using spinning disk confocal microcopy for intravital imaging and fluorescent reporter mice to track and visualize individual cells, the authors have previously shown that Kupffer cells, the liver's resident macrophages located in the sinusoids, play an essential role in this process.3 Kupffer cells respond to extracellular ATP with inflammasome activation and release of chemokines and proinflammatory cytokines such as interleukin-1β (IL-1β) (Fig. 1B). IL-1β induces expression of intercellular cell adhesion molecule 1 on sinusoidal cells, which triggers neutrophil adherence. Neutrophils then crawl along the perfused sinusoids on a gradient of chemokines and then migrate into the tissue within 3-4 hours of the injury.3 Neutrophils clear debris and undergo apoptosis, which results in the formation of neutrophil extranuclear traps.4 Bone marrow–derived proinflammatory monocytes are also recruited from the sinusoidal blood into the liver, but it takes about 8-12 hours until they arrive at the injury site and form a ring-like structure around the necrotic tissue.5 It takes even longer (>72 hours) for them to convert from a classic proinflammatory phenotype into a nonclassical or alternative phenotype that facilitates wound healing. Wang and Kubes now find that wound healing and tissue regeneration are aided by a specialized population of peritoneal macrophages that moves directly through the visceral endothelium of the liver and not through the vasculature. The story starts with the authors' observation that intravascular injection and topical application of an antibody against the macrophage marker F4/80 yield different results. Intravascular injection of the F4/80 antibody shows that Kupffer cells are completely depleted from the sinusoids within a 600-μm radius of the thermal injury site and that neither Kupffer cells outside that range nor any other F4/80+ cells move through the vasculature toward the injury site during the first 72 hours. In contrast, topical application of the F4/80 antibody showed that F4/80+ cells appeared within 1 hour at the injury site. What is the origin of these cells? Flow cytometry confirmed that they were indeed different from Kupffer cells. Rather, they expressed markers such as the adhesion molecule cluster of differentiation 102 (CD102) and the transcription factor GATA6, which are both characteristic for a subpopulation of peritoneal macrophages. Depletion of peritoneal macrophages prevented the early influx of F4/80+ cells to the injury site. Even more intriguingly, peritoneal macrophages were recruited to the injury site only when they were injected intraperitoneally, not when they were injected into the vasculature. Their migration did not depend on chemokines because inhibition of Gαi protein–coupled chemokine receptors with pertussis toxin did not affect their recruitment. Likewise, blocking of β1 or β2 integrins did not affect the cells' recruitment. The authors show that extracellular ATP represents the first signal to peritoneal macrophages. The authors then show that hyaluronan is exposed at the injury site and facilitates macrophage adhesion through its ligand CD44. Consistent with these findings, injection of CD44-specific antibodies or pretreatment of mice with hyaluronidase prevented recruitment of peritoneal macrophages. This is interesting because hyaluronan–CD44 interaction contributes to neutrophil recruitment solely during infection and not during sterile inflammation. What happens after recruitment of peritoneal macrophages to the injury site? A characteristic of the recruited peritoneal macrophages is expression of the zinc finger transcription factor GATA6, which controls proliferation, survival, and metabolism. Indeed, peritoneal macrophages underwent rapid expansion at the injury site, as shown by up-regulation of Ki67 and loss of bromodeoxyuridine after pulse labeling. Further, they up-regulated CD273, CD206, and arginase 1 expression along with the transcriptional profile (Chil3, Mrc1, Retnla, Il10) of tissue repair macrophages. Imaging studies revealed that peritoneal macrophages dismantled intact nuclei of dead cells by pulling off vesicles with nuclear DNA. The released DNA completely covered the wound 12 hours after the injury and facilitated tissue repair. This resulted in a 75% reduction of the injured area by day 7 in wild-type mice, whereas macrophage recruitment and tissue regeneration were delayed in GATA6-deficient mice. This new paradigm of avascular recruitment of mature macrophages opens many avenues for future research. Wang and Kubes demonstrate that their findings are not limited to thermal injury close to the surface of the liver but that they extend to acute, generalized necroinflammatory liver injury after carbon tetrachloride ingestion. This prompts further questions: How far and how quickly can the cells travel through liver tissue if they need to reach the center of the liver? Can they avoid entering sinusoids and “getting lost,” as observed after intravascular injection? Do the results extend to acute sterile inflammation in other intraperitoneal organs such as stomach, pancreas, and gut? Conversely, do retroperitoneal organs such as the kidney lack access to these cells? These questions are important if one considers therapeutic applications. Acetaminophen-induced liver failure, for example, is characterized by a marked increase in inflammatory macrophages in areas of hepatic necrosis, along with a reduction of the number of monocytes in the blood.6 Morbidity and mortality may be improved by an increased number of peritoneal macrophages at the injury site. Given that peritoneal macrophages can readily be harvested from the peritoneal cavity, they may be expanded in vitro and cryopreserved for treatment of acute diseases such as drug-induced and alcohol-induced acute liver injury and ischemia–reperfusion injury. Conversely, it is important to study peritoneal macrophages in chronic liver disease. Chronic sterile inflammation, as in nonalcoholic fatty liver disease, is associated with constant release of damage-associated molecular patterns.7 Do continuous activation and recruitment of peritoneal macrophages contribute to fibrogenesis under these conditions? How do infections such as bacterial peritonitis that are common in patients with liver cirrhosis affect the number and function of peritoneal macrophages? As this brief list of questions exemplifies, the study by Wang and Kubes likely represents the first of many publications—a landmark study identifying peritoneal macrophages as highly mobile cells with a specialized function in tissue repair. Barbara Rehermann, M.D. Immunology Section, Liver Diseases Branch National Institute of Diabetes and Digestive and Kidney Diseases National Institutes of Health Bethesda, MD