Liver regeneration is a highly integrated process involving a wide array of cellular interactions occurring between the different liver cell types, but also between the liver and the extrahepatic environment. Besides the main growth factors and cytokines acting in a paracrine way, diverse endocrine and neuroendocrine interactions, as well as the recruitment of blood cells, affect liver regeneration. This entire network acts in concert to achieve a fine-tuning of proliferative and hepatoprotective cascades, ultimately leading to liver mass restoration after injury or resection.1 A study published in this issue of HEPATOLOGY,2 sheds new light on the complex interactions between immune cells and liver regeneration processes, with the purinergic system as one possible regulator.
Among modulators of liver regeneration, the innate immune system constitutes a complex network of interacting cells and cytokines, of which the resulting effect on liver regeneration is not entirely understood. Surprisingly, the complement system3 as well as cytotoxic cytokines, like tumor necrosis factor alpha,1 are reported as positive regulators for hepatocyte regeneration. The role of Kupffer cells, the hepatic resident macrophages, is still controversial, because these cells secrete numerous cytokines, stimulating or inhibiting hepatocyte proliferation, after liver injury.1 As to natural killer (NK) and natural killer T (NKT) cells, the main lymphoid population in human and mouse liver, although the majority of studies reported on their negative effect on liver regeneration,4, 5 their role still remains controversial.6-8 NK cells would inhibit regeneration, in particular through the secretion of interferon-gamma (IFN-γ), whereas NKT cells would play a minor role in the normal liver after partial hepatectomy (PH).5 However, NK and NKT cells may have a positive effect on oval cell-dependent regeneration after acute liver injury when hepatocytes cannot replicate.9 The study by Graubardt et al. revisited the controversy on NK cells during liver regeneration by introducing a new parameter: the purinergic system, currently recognized as a major component of the inflammatory response after injury.10
Signaling by extracellular adenosine triphosphate (ATP) acting on plasma membrane purinergic receptors, that is, P2Y1,2,4,6,11-14 (G-protein-coupled receptors) and P2X1-7 (ATP-gated channels with high Ca2+ permeability), is vital not only in excitable, but also in nonexcitable cells and tissues.11 Vascular shear stress, organ distension, or cellular injury are known to trigger ATP release from endothelial, epithelial, and other cell types.12 Once in the extracellular medium, ATP is degraded by the powerful ecto-ATPases expressed on cell surfaces to avoid excessive activation (and desensitization) of purinergic receptors. Adenosine, the breakdown product of ATP, can bind P1 purinergic receptors, with specific downstream biological responses.11 Thus, in a given tissue, the final purinergic input will be determined by the complex combination of numerous parameters, including mainly mechanisms of ATP release, purinergic receptors expression and distribution, and ecto-ATPases expression.11
In the liver, each cell type expresses its own repertoire of purinoceptors and ecto-ATPases, and evidence for a crucial effect of purinergic signaling in liver physiology is growing, including modulation of bile secretion and ischemia protection.13-15 In vitro studies suggested that extracellular ATP had a positive effect on proliferation of primary rat hepatocytes.16 Recently, we showed that an extracellular ATP release from the liver occurred immediately after PH, contributing to liver regeneration in the rat.15 We also observed an immediate ATP release from the liver after PH for living donor transplantation, suggesting that purinergic signaling may be also involved during human liver regeneration.15 However, the purinoceptors and precise mechanisms involved in ATP-mediated effect on liver regeneration still remain to be defined.
The authors of the present study previously reported on a coordinating role of the ectonucleotidase CD39-mediated ATP and adenosine diphosphate degradation after PH in mice, suggesting that both hepatocytes and endothelial cells would need, in their microenvironment, a fine-tuned balance between the different nucleotide species to adequately proliferate after PH.17 The same group also showed that the lack of CD39 protected mice from liver injury in models known to be, at least in part, mediated by NK and NKT cells.14, 18 In the absence of CD39, ATP-induced NKT cell apoptosis18 and inhibition of IFN-γ secretion14 cannot be refrained, thereby limiting liver injury.
In their study, Graubardt et al.2 propose that hepatic NK cell activation is required for optimal liver regeneration in a process that would depend on extracellular ATP hydrolysis. The authors reported an impaired regeneration and increased liver injury after PH in mice lacking T, B, NKT, and NK cells (Rag2/common gamma-null), whereas this was not observed in mice lacking only B, T, and NKT lymphocytes (Rag1-null). In wild-type mice, the authors also showed that PH increased cytotoxicity and immaturity of hepatic NK cells. Because ATP is released immediately after PH,15 and based on their previous studies,14, 18 the authors asked whether hepatic NK cell activity was dependent on extracellular ATP hydrolysis during liver regeneration. They observed that, when ATP was hydrolyzed by apyrase (soluble ectonucleotidase) treatment in vivo, hepatocyte entry in S phase after PH was enhanced, and that this effect was dependent on NK cells. Also, ATP clearance by apyrase treatment further increased hepatic NK cell cytotoxicity after PH. In vitro, the authors showed that ATP stimulation decreases NK cell cytotoxicity, possibly through P2X3 receptors. They propose that NK cells, which cytotoxicity is enhanced after PH, promote liver regeneration in a process inhibited by extracellular ATP and enhanced by its hydrolysis.
In light of the literature, this interesting study adds complexity on top of controversial questions. First, in contradiction with a number of previous reports, in this study, NK cells appear to be beneficial for liver regeneration; the more cytotoxic they are, the best it seems to be for the liver after PH. Although this apparent paradox during liver regeneration had already been reported on for other cytotoxic components of the innate immune system,1, 3 as well as for Fas receptor engagement,19 the mechanisms by which cytotoxic NK cells would enhance hepatocyte regeneration remain to be defined. It is also important to note that, because Graubardt et al. used mice lacking not only NK, but also NKT, B, and T cells, the observed inhibition of liver regeneration may not result only from the absence of NK cells, but also from the lack of other immune cells, which interaction with NK cells may affect the final response to injury. Second, it may be important to consider that, besides ATP clearance, in vivo experiments with the soluble ectonucleotidase apyrase also provide the liver with ATP degradation products, such as adenosine, which could contribute to the liver response to PH.11 Third, it is not possible, at present, to define unequivocally a global resulting effect of ATP release during liver repair. Although, after PH, extracellular ATP can directly or indirectly promote hepatocyte proliferation,15, 16 it could also indirectly inhibit it through NK cell inactivation, as highlighted in this study. However, in the context of liver injury (toxic, ischemic, or autoimmune), extracellular ATP would be globally beneficial, whereas ectonucleotidase-mediated ATP clearance would be deleterious, although this view could be challenged by considering the differential effect of extracellular ATP on the different liver cell types and their specific purinoceptor repertoires.20 Here, there is obviously still room for further research about these apparently opposite endpoints of purinergic signaling in liver regeneration before any therapeutic strategy could be rationally anticipated. 1