Rowan van Golen, Department of Experimental Surgery, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. Email: email@example.com
Ischemia and reperfusion (I/R) injury is an often unavoidable consequence of major liver surgery and is characterized by a sterile inflammatory response that jeopardizes the viability of the organ. The inflammatory response results from acute oxidative and nitrosative stress and consequent hepatocellular death during the early reperfusion phase, which causes the release of endogenous self-antigens known as damage-associated molecular patterns (DAMPs). DAMPs, in turn, are indirectly responsible for a second wave of reactive oxygen and nitrogen species (ROS and RNS) production by driving the chemoattraction of various leukocyte subsets that exacerbate oxidative liver damage during the later stages of reperfusion. In this review, the molecular mechanisms underlying hepatic I/R injury are outlined, with emphasis on the interplay between ROS/RNS, DAMPs, and the cell types that either produce ROS/RNS and DAMPs or respond to them. This theoretical background is subsequently used to explain why current interventions for hepatic I/R injury have not been very successful. Moreover, novel therapeutic modalities are addressed, including MitoSNO and nilotinib, and metalloporphyrins on the basis of the updated paradigm of hepatic I/R injury.
Liver resection or transplantation often constitutes the only curative treatment option for patients who suffer from a hepatic malignancy or end-stage liver disease. Irrespective of the underlying disease, intraoperative cessation of hepatic blood supply (i.e. ischemia) is sometimes necessary during resection or is inherent to the transplantation procedure, and causes further impairment of organ function. Following resection or transplantation, ischemia is alleviated by the restoration of blood flow to the organ (i.e. reperfusion), which triggers a cascade of molecular events that entails oxidative stress, induction of an immune response, and inflammation.[1, 2] The damaging effects that result from these events, which are highlighted in this paper, are known as ischemia and reperfusion (I/R) injury and collectively determine the postoperative outcome.
The need for major liver surgery is likely to increase as the number of surgical patients in which a malignancy coexists with parenchymal liver disease is expected to rise.[3, 4] Given that the severity of I/R injury correlates positively with the extent of preexisting liver disease, for example non-alcoholic fatty liver disease, this trend will not only result in an increased number of hepatic resections, but also in an increased percentage of high-risk procedures. The same applies to liver transplantation, where the gap between organ supply and demand has been partially filled by the utilization of marginal (e.g. steatotic) grafts. The use of suboptimal organs, however, remains associated with dissatisfying outcomes, while the already insufficient donor pool suffers from a deteriorating quality of grafts. As a result, hepatic I/R injury will most likely become even more prevalent, necessitating the development of effective treatment strategies.
In spite of these cues, few options are currently available to prevent or treat hepatic I/R injury in patients. The modalities that have been clinically evaluated can roughly be divided into surgical and pharmacological interventions. Surgical approaches have mainly focused on local and remote ischemic preconditioning and the fine-tuning of portal triad clamping regimens, which has yielded modest improvements at best.[9, 10] Pharmaceutical approaches have primarily addressed the use of antioxidants (e.g. vitamin E), vasodilators (e.g. prostaglandin E1), and volatile anesthetics (e.g. sevoflurane). The testing of these pharmaceuticals, however, was predicated on meager preclinical evidence (anesthetics) or insufficiently proven efficacy (antioxidants) as a result of which the pharmacological approach has also yielded inconsistent and unsatisfactory results.[8, 12]
Unfortunately, the majority of trials currently entered into the clinical trial database (http://www.clinicaltrials.gov) can be fitted into the framework of interventions described above. It, therefore, appears that a therapeutic plateau has been reached, which in most instances can only be breached by changing strategies. Accordingly, a critical reevaluation of the molecular and cellular basis of hepatic I/R injury is needed to propel interventional strategies into a more appropriate and hopefully more effective direction.
The inflammatory environment during hepatic I/R
It has become clear that (over)activation of the immune system is a critical factor in hepatic I/R injury. Although liver-resident macrophages (Kupffer cells [KCs]) and chemoattracted neutrophils have long been implicated as the chief culprits in I/R injury, the question as to how immune cells are stimulated in a pathogen-free surgical setting has only recently been answered. It all starts with intrahepatic oxidative/nitrosative stress. During reperfusion, the first wave of reactive oxygen and nitrogen species (ROS and RNS) generation by mitochondria poses an acute threat to the viability of hepatocytes. Oxidatively/nitrosatively stressed and dying hepatocytes subsequently leak cellular components known as damage-associated molecular patterns (DAMPs) into the circulation to notify the host of tissue damage. These self-antigens are functional components of healthy cells (e.g. histones), but become potent immunostimulators in the extracellular compartment. Circulating DAMPs are detected by antigen-presenting cells, such as KCs, which translate the alarm signal into an overt inflammatory response, for example through cytokine production. The array of inflammatory cues released into the circulation in turn governs the chemoattraction of various nonresident leukocytes that initiate a second wave of ROS/RNS production. Of the chemoattracted cell types, neutrophils and monocytes possess the greatest ROS/RNS-generating potential. Inasmuch as these inflammatory cells are programmed to eliminate pathogens through ROS/RNS production, the sterile nature of hepatic I/R injury causes the produced oxidants to become directed against self (liver parenchyma) instead of non-self (microbes), resulting in profuse tissue destruction. Furthermore, additional leukocyte subsets (e.g. T cells, natural killer [NK] cells) are recruited to the liver and exacerbate I/R injury.
The cellular constituents that mediate hepatic I/R injury are summarized in Figure 1. Accordingly, the cytokine networks that relay DAMP-derived danger signals to effector leukocytes contribute considerably to the severity of hepatic I/R injury, adding to the damage caused by the first wave of ROS/RNS generation.
New paradigm; new leads
The adapted view on hepatic I/R injury has exposed several signaling routes that likely are involved in hepatic I/R injury and could prove viable targets for intervention, but need to be further investigated (Fig. 1, gray text). One of the cytokines that exacts additional attention is interleukin 17A (IL-17A), which reportedly mediates reperfusion injury by driving neutrophil accumulation. IL-17A is best known as the signature cytokine of T-helper 17 (Th17) cells, which are typified by the lineage-specific transcription factor retinoic acid receptor related orphan receptor gamma (RORγ(t)) and are activated by a combination of IL-23, IL-6, and transforming growth factor-β. Normally, Th17 cells polarize and expand over 3–5 days in response to pathogens or an autoimmune challenge. However, a small portion of IL-17A-producing lymphocytes serve an innate-type role and express RORγ(t) within hours after detecting IL-23 and IL-1β, which better fits the time-course of hepatic I/R injury. Of the early responders, γδ T cells have been found in murine livers following I/R, which also holds for the chief activators of RORγ(t), namely IL-1β and IL-23. However, detailed information on the interplay between IL-23, (innate-type) Th17 cells, and IL-17A during hepatic I/R was not provided in these reports.
The significance of IL-23 and IL-17A in I/R injury of wild-type livers is not unequivocal and has recently been contested. In contrast to the abovementioned, it was shown that the expression of IL-23 and IL-17A is exclusively upregulated in mice that are deficient in the transcription factor interferon regulatory factor 3 (IRF3). IRF3 reportedly blocks IL-17A-mediated liver injury in wild-type animals by propagating the production of IL-27, which is known to suppress Th17 development. Because IRF3 activity is controlled by toll-like receptor 4 (TLR-4) signaling, exogenous lipopolysaccharide (LPS) was infused as a TLR-4 agonist to determine the cellular origin of IL-23 and IL-17A in IRF3-/- animals. In doing so, KCs, which released IL-23, and γδ T cells/NK T cells, which released IL-17A, were identified as cellular mediators of the LPS-induced Th17 response. These results, however, contradict an earlier report, in which it was established that IRF3 deficiency actually protects mouse livers from I/R injury by short-circuiting the TLR-4 signaling axis. Moreover, the infusion of LPS defies the sterile nature of I/R injury, so it remains to be elucidated if and to what extent the IL-23/IL-17A axis is involved in sterile I/R injury of wild-type animals, which better reflects the clinical situation.
Another immunological phenomenon that has been implicated in hepatic I/R injury is purinergic signaling, which modulates immune cell function by adenosine triphosphate (ATP) and its catabolites (adenosine diphosphate [ADP], adenosine monophosphate [AMP], and adenosine). It has been shown that ATP accumulates extracellularly following hyperthermia-induced sterile liver inflammation in mice. In addition to leakage from dying and dead cells, circulating purines can be derived from activated leukocytes, which release ATP to strengthen their pro-inflammatory effects in an autocrine loop. When derived from dying or dead cells, ATP acts as a DAMP that is recognized by the P2X7 receptor on KCs, leading to the activation of the NALP3 inflammasome and the release of IL-1β and IL-18. The interleukins in turn drive neutrophil accumulation by triggering production of the chemokine CXCL2 and increasing the endothelial expression of the neutrophil receptor, intercellular adhesion molecule 1. Additionally, Beldi and colleagues demonstrated that, following liver I/R in mice, NK cells metabolize extracellular ATP to ADP and AMP using the cell-surface endonucleotidase CD39. The purines signal through one of the five purinergic receptors on the NK cell surface (see Fig. 1, bottom left) to amplify interferon gamma production and boost the inflammatory response. These data suggest a broad role for extracellular purines in hepatic I/R injury, as has recently been claimed for several other liver pathologies.
Potential therapeutic approaches
The characterization of hepatic I/R injury as a sterile inflammatory disorder may unlock a novel therapeutic avenue in which specific stages of inflammation can be targeted to preserve liver function. As alluded to before, antioxidant therapy has not proven very successful to date, which necessitates the (clinical) evaluation of more sophisticated second-generation compounds capable of neutralizing ROS/RNS. Additionally, the inflammatory cascade can be inhibited at various biochemical intersections to ameliorate the recruitment of ROS/RNS-producing leukocytes and with it the second wave of ROS/RNS generation.
As for the most proximal stages of reperfusion injury, administration of the MitoSNO has proven effective in dampening the early mitochondrial ROS burst in a mouse model of cardiac I/R. By S-nitrosating cysteine-39 of complex I (CysSH + MitoSNO → CysSNO + MitoSH) in the mitochondrial electron transport chain, MitoSNO retains complex I in a less active conformation, thereby slowing down electron transport and limiting mitochondrial ROS production by complex I during the reperfusion phase (Michael Murphy, pers. comm., 2012). As the S-nitrosothiol on complex I is relatively rapidly reduced back to a free thiol by the endogenous thiol reductant systems (CysSNO → CysSH, half life of ± 5 min), the suppressive effect on electron transport and ROS production gradually dissipates within 5–10 min of reperfusion. As a result, excessive mitochondrial damage and corollary DAMP release can be prevented during the (hyper)acute reperfusion phase, and resumption of oxidative phosphorylation and repletion of ATP levels can occur in a timely fashion (Table 1). Whether this also holds true for hepatic I/R injury remains to be tested.
Table 1. Therapeutic compounds that (potentially) tackle the inflammatory component of hepatic ischemia and reperfusion injury
Regarding the later phases of inflammation, treatment with the receptor tyrosine kinase inhibitor nilotinib was shown to ameliorate hepatic I/R injury in mice by inhibiting the pro-inflammatory actions of p38 mitogen-activated protein kinase (MAPK) and c-Jun N-terminal kinase (JNK). The observed results were cell type-specific: whereas p38 MAPK inhibition diminished DAMP-induced cytokine release by non-parenchymal cells and hampered further leukocyte chemoattraction, inhibition of JNK phosphorylation predominantly boosted the survival of hepatocytes. In doing so, nilotinib is the first compound to multifactorially play in on the inflammatory context of hepatic I/R injury by tackling both DAMP release (i.e. hepatocyte death) as well as DAMP signaling (Table 1). Due to the promising results and the fact that nilotinib is generally well tolerated, the clinical application of nilotinib for the treatment of hepatic I/R injury has been readily advocated.
In addition to preventing leukocyte accumulation and activation, it is also possible to directly neutralize ROS/RNS. Until now, most compounds administered to patients undergoing portal triad clamping specifically targeted free radicals, that is compounds with an unpaired electron. Examples are mannitol, which targets hydroxyl radicals; vitamin E, which reacts with lipid (peroxyl) radicals; and α-lipoic acid, which scavenges various free radicals. Due to their high reactivity, however, free radicals tend to react with the first oxidation-sensitive biomolecule they encounter rather than the administered antioxidant, which could account for the marginal efficacy of the aforementioned compounds. Moreover, most free radicals are derived from the more non-radical oxidants hydrogen peroxide and peroxynitrite. Because hydrogen peroxide and peroxynitrite are more stable than free radicals, their diffusion distance is considerably larger. This stability allows them to react with preferred substrates such as carbon dioxide or transition metals (e.g. ferrous iron), leading to the formation of the highly reactive nitrogen dioxide, hydroxyl radical, and/or carbonate radical anion. Considering that transition metals in biological systems are commonly bound to biomolecules, it is likely that transition metal-catalyzed reactions are how hydrogen peroxide and peroxynitrite exert their damaging effects, especially since antioxidants cannot prevent these site-specific processes. Consequently, these considerations plead for the use of compounds that neutralize non-radical ROS/RNS instead. In that regard, high-affinity peroxynitrite decomposition catalysts, such as ferric and manganese porphyrins, might be of interest. One such compound, Mn(III)mesotetrakis(N-N-diethylimidazolium-2-yl)porphyrin (Table 1), has already undergone phase I clinical testing and was well tolerated by patients who suffer from amyotrophic lateral sclerosis. This new generation of compounds may, therefore, reinvigorate the use of antioxidants for the treatment of hepatic I/R injury.
Hepatic I/R injury is a surgery-induced problem that manifests itself as a sterile inflammatory disorder and poses a direct threat to organ viability. In spite of the consequences and in the absence of effective liver assist therapies, mere supportive care still remains the only treatment option for operated patients. The novel insights into the inflammatory component of hepatic I/R injury, however, could spur the development of second-generation intervention modalities. As most proximal triggers, neutralizing DAMPs directly should prevent the onset of inflammation as well as the second wave of oxidative/nitrosative stress, although the vital role of DAMPs in healthy cells could hinder this approach. Alternatively, more targeted (e.g. antibody) therapies could be employed to prevent DAMP release or neutralize the pro-inflammatory effects of selective DAMPs, thereby deterring the self-amplifying cycle of cell death, DAMP release, leukocyte activation, and ROS/RNS generation. In that respect, it could be feasible to block the function of DAMP receptors. Further downstream, defining the cytokines that attract and activate effector leukocytes could also reveal viable targets for intervention. Lastly, it could be worthwhile to directly neutralize ROS/RNS with second-generation compounds that slow down the rate of ROS/RNS production during the (hyper)acute reperfusion phase or that target physiologically relevant non-radical ROS/RNS that serve as templates for the formation of free radicals. In any case, the field of liver surgery needs a novel set of treatment modalities to replace the currently available options (Table 1), which have largely proven inadequate.
The authors are grateful to Dr Michael Murphy from the MRC Mitochondrial Biology Unit, Cambridge, UK, for providing detailed information on MitoSNO, and to Inge Kos from the Medical Illustration Service for the artwork. This work was supported by a PhD scholarship from the Academic Medical Center.