Regulatory Mechanisms of Injury and Repair after Hepatic Ischemia/Reperfusion

Hepatic ischemia/reperfusion injury is an important complication of liver surgery and transplantation. The mechanisms of this injury as well as the subsequent reparative and regenerative processes have been the subject of thorough study. In this paper, we discuss the complex and coordinated responses leading to parenchymal damage after liver ischemia/reperfusion as well as the manner in which the liver clears damaged cells and regenerates functional mass.


Introduction
Hepatic ischemia/reperfusion (I/R) injury is a consequence of vascular in�ow occlusion due to portal vascular clamping during complex liver surgery. I/R injury of the liver is directly related to the duration of liver ischemia and is a major cause of morbidity and mortality from liver transplantation and resection [1][2][3]. ere has been considerable study of the biochemical and cellular changes occurring during I/R which has informed clinical practice. e results of these studies, which are the focus of this paper, have led to advances in our understanding of the pathophysiology of hepatic I/R injury and the development of new therapeutic modalities.

Initiation of Reperfusion Injury
Early work by Jaeschke et al. [4][5][6] established that there are two distinct phases of liver injury aer warm I/R. e initial phase of injury occurring within the �rst couple of hours of reperfusion is characterized by Kupffer cell-induced oxidant stress. Kupffer cell production and release of reactive oxygen species, including superoxide anion and hydrogen peroxide, result in acute hepatocellular injury. Blockade of Kupffer cell activity, accomplished by administration of gadolinium chloride or methyl palmitate, reduces acute hepatocyte damage. In addition, complement activation products are critically important for Kupffer cell activation during the initial phase of injury as depletion of complement reduces Kupffer cell-induced oxidant stress [7]. Despite the contribution of Kupffer cell-derived oxidants, the extent of injury during this initial phase is far less than that observed at later time points. Events occurring during the initial phase of liver injury, including activation of Kupffer cells, initiate a complex in�ammatory cascade leading to the recruitment of various populations of leukocytes to the liver. e �rst population of leukocytes recruited aer reperfusion is CD4 T lymphocytes.

Hepatic Recruitment of CD4 T Cells
Signi�cant involvement of T lymphocytes in hepatic I/R was �rst demonstrated in 1997 in a report that found that T lymphocytes rapidly accumulated in the liver aer reperfusion [8]. is study showed that CD4, but not CD8, T lymphocytes were recruited into the postischemic liver within 1 hour of reperfusion. e briskness of this response is surprising as it preceded the in�ux of innate immune cells to the injured tissue. Later studies by our group con�rmed this rapid recruitment of CD4 T cells [9]. e mechanisms by which T cells are so rapidly recruited to the postischemic liver remain unde�ned. However, there is growing evidence that hepatic expression of chemokines is an important contributor to this process [10].
As mentioned above, CD4 lymphocytes are recruited into postischemic liver long before any appreciable neutrophil accumulation. Both antibody depletion of CD4 T cells and CD4-knockout mice showed reduced liver recruitment of neutrophils aer I/R [8,9]. e mechanism by which CD4 T cells regulate subsequent neutrophil accumulation appears to be related to their release of IL-17. IL-17 is preferentially expressed and secreted by activated CD4 lymphocytes [11]. Furthermore, in a model of peritoneal in�ammation, IL-17 was found to mediate neutrophil recruitment by increasing the production of chemokines by the peritoneal mesothelium [12]. IL-17 has also been shown to induce chemokine production by other cell types, including epithelial cells, �broblasts, osteoblasts, and endothelial cells [13][14][15]. Our studies found that production of neutrophil-attracting chemokines was decreased in CD4-knockout mice and that in wild-type mice treated with anti-IL-17 antibodies, chemokine expression was reduced [9]. In both of these experiments, liver neutrophil accumulation was also reduced. Moreover, adoptive transfer of CD4 lymphocytes into CD4-knockout mice resulted in dramatic increases in the expression of chemokines and the degree of liver neutrophil recruitment [9]. us, it would appear that CD4 lymphocytes are an important regulator of hepatic neutrophil recruitment during liver I/R and that this occurs via their release of IL-17. e question of whether or not T cell involvement in liver I/R is driven by antigenic or nonantigenic mechanisms has not been elucidated. Some studies show that utilization of MHC II blocking antibodies has no effect on serum ALT following hepatic IR [16]. is study suggested that T cells play a bene�cial role not involving the TCR and that lymphocyte actions occur through a nonantigenic mechanism. It is well established that during hepatic I/R in�ammatory cytokines such as IL-12 and IL-18 are rapidly expressed [17,18]. Furthermore, nonnaive as well as unconventional T cells can be functionally activated by these cytokines in a manner independent of TCR engagement [19][20][21]. Taken collectively, these studies suggest the possibility of nonantigenic activation of T cells during the initial stages of I/R in the liver. Alternatively, recent studies in other models of I/R, have discovered the presence of an IgM that reacts with self-antigens generated by damaged tissues [22,23]. ese self-reactive IgMs activate the classical pathway of complement and contribute substantially to the initiation of the injury response. A similar mechanism may be applicable to liver I/R, but to date has not been examined.
In order to successfully mount an immune response to an antigen, T lymphocytes need to receive two different signals. e �rst signal is delivered by the antigen upon its binding to the T-cell receptor (TCR). is antigen-speci�c event is usually termed signal one. e second signal, signal two, is costimulation delivered by antigen presenting cells and is a non-antigen-speci�c event. ere are a large number of different costimulatory molecules and they vary greatly in their expression patterns and function [24]. One of the most widely studied co-stimulatory pathways is the CD40-CD154 pathway. CD40 is a member of the tumor necrosis factor receptor superfamily and is expressed on APCs such as dendritic cells (DCs), macrophages, and B-cells. Ligation of CD40 by its cognate ligand CD154 (which is transiently expressed on activated T helper cells) leads to co-stimulation of the target cell. Speci�cally, during liver I/R, it has been shown that gene therapy-mediated CD154 blockade (Ad-CD40 Ig), antibody-induced systemic CD154 blockade (MR1 mAb), and genetically targeted CD154 absence (CD154 KO mice) ameliorated otherwise fulminant injury in a warm liver I/R model [25]. ese bene�cial effects resulting from the disruption of CD154-CD40 signaling were accompanied by (1) diminished liver T-cell sequestration; (2) decrease of VEGF expression (3) inhibition of TNF-and T-helper () type 1 cytokine production and (4) induction of antiapoptotic (Bcl-2/Bcl-xl) and depression of proapoptotic (caspase-3) proteins.
Another widely studied co-stimulatory pathway is the CD28/CD80/86 pathway. CD28 is constitutively expressed on T cells. e ligands for CD28 are CD80 and CD86 (B7-1, B7-2), both members of the immunoglobulin (Ig) superfamily, which are transiently expressed on activated APCs. Both CD80 and CD86 are increased in the liver aer I/R [26,27]. Ligation of CD28 by these molecules in conjunction with antigen recognition via the TCR complex leads to activation of the T cell. An additional feature of this pathway is the existence of an alternative receptor for CD80/86 called CD152 (CTLA-4), which unlike CD28, is upregulated aer T cell activation and results in suppressive T cell function. Indirect evidence for a critical role for T cells in kidney I/R came from blocking one of the costimulatory pathways necessary for T-cell activation. Blocking the B7-CD28 costimulation pathway by CTLA4 Ig, a recombinant fusion protein, containing the extracellular domain of human CTLA4 (a homologue of CD28), resulting in T cell anergy, ameliorated renal dysfunction and decreased mononuclear cell in�ltration in a model of renal cold ischemia [28]. It has yet to be elucidated whether such treatment during liver ischemia reperfusion would yield similar results.
e liver sinusoidal endothelial cell (LSEC) has been described as a new type of APC that resides in the liver [29,30]. LSEC is also believed to express the costimulatory moieties CD40, CD80, and CD86 and stimulate T cells through peptide presentation in the context of MHC class I and II molecules [22,31]. is would allow endothelial activation by T cells and vice versa, due to TCR-MHC and either CD40-CD154-or CD28-B7-dependent pathways. However, in a contrary report that compared LSEC and dendritic cells directly, it was found that LSEC expressed surface markers only re�ective of an endothelial phenotype. Further, highly puri�ed LSEC had undetectable levels of the co-stimulatory receptors CD40, CD80, and CD86 and only minimal MHC class II. is paper concluded that LSECs are poor stimulators of T cells, but other properties, such as their high capacity for antigen uptake and direct access to circulating lymphocytes, may enable them to contribute to the unique immunologic function of the liver [32]. e precise manner in which intrahepatic T cells interact with various APCs in the liver during the response to I/R has not been elucidated. However, the rapid recruitment of lymphocytes to the liver coincides with the induction of a robust hepatic in�ammatory response.

�. Initiation an� �ropa�ation of In�a��ation in the Liver by I/R
e hepatic in�ammatory response to I/R appears to begin with the elaboration of the cytokines, IL-12 and IL-23. Increased expression of these cytokines, at both the mRNA and protein levels, can be detected prior to hepatic reperfusion and this expression is short-lived, disappearing within 4-5 hours of hepatic reperfusion [18,33]. e importance of IL-12/23 in the initiation of the hepatic in�ammatory response to I/R was demonstrated in studies using both neutralizing antibodies as well as mice lacking the p40 subunit, which is common to both IL-12 and IL-23. In both experiments, a functional lack of IL-12 prevented increased expression of tumor necrosis factor-alpha (TNF ) and subsequent development of neutrophil-dependent liver injury. is prominent role of IL-12/23 was direct and not mediated by induction of interferon-gamma (IFN ) production, as mice nullizygous for IFN were indistinguishable from wild-type mice in their response to hepatic I/R [18]. e liver cell population responsible for producing IL-12/23 has not yet been identi�ed, but Kupffer cells and stellate cells are likely sources [34]. Similarly, the signalling mechanism through which IL-12/23 acts to illicit TNF production remains unde�ned. IL-12/23 is known to be a potent activator of the transcription factor, signal transducer and activator of transcription-4 (STAT4) [35,36]. However, STAT4-de�cient mice were not protected from liver I/R injury [37], suggesting that another signalling mechanism exists that is responsible for the in vivo gene induction of TNF by IL-12/23. e production of TNF and IL-1 by Kupffer cells aer I/R has long been thought to be the primary initiating events for propagation of the hepatic in�ammatory response [38][39][40][41]. We now know that expression of IL-12/23 is required for the full expression of TNF and that IL-1 plays only an accessory role in the hepatic in�ammatory response to I/R. As with many other acute in�ammatory responses, TNF is a central mediator in the hepatic response to I/R. Liver production of TNF does not occur during hepatic ischemia but begins to increase shortly aer reperfusion at a time when hepatic IL-12/23 levels are maximal [18]. e importance of TNF in the hepatic in�ammatory response has been well described and blockade of this mediator abolishes liver in�ammation and hepatocellular injury [38,42]. ese protective effects were subsequently found to be attributed to TNF induction of secondary in�ammatory mediators. Blockade of TNF prevents the expression of hepatic vascular adhesion molecules as well as the expression of CXC chemokines, which are chemotactic for neutrophils [43][44][45][46].
IL-1 was presumed to share many of the same effects of TNF , based on an early study showing that prophylactic treatment with IL-1 receptor antagonist protected against hepatic I/R injury [41]. However, a more recent study indicates that IL-1 functions to augment neutrophil recruitment but does not play an essential role in the development of liver in�ammation aer I/R [47]. In this study, IL-1 receptorknockout mice experienced the same degree of hepatocellular injury as their wild-type counterparts, but had reduced neutrophil accumulation in the liver. is was found to be associated with attenuated activation of the transcription factor, NF-B, and reduced expression of CXC chemokines. It appears that IL-1 functions to augment the in�ammatory response at later time points through the induction of CXC chemokine expression, but the lack of IL-1 function does not signi�cantly affect the development of liver in�ammation.

Transcriptional Activation of �roin�a��atory �yto�ines
A common theme amongst each of the three "early response" cytokines discussed above is their gene regulation. e gene expression of IL-12/23, TNF , and IL-1 is controlled, at least in part, by the transcription factor, NF-B [48].
Other transcription factors, such as AP-1 and members of the signal transducer and activator of transcription (STAT) family, also regulate proin�ammatory mediator expression. However, their roles during hepatic I/R injury are less well studied. e term NF-B refers to proteins of the Rel family which share a homologous amino acid sequence in their amino termini called the Rel homology domain that is necessary for dimerization, DNA binding and I B (inhibitor of NF-B) binding, [49,50]. ese proteins bind to form homo-or heterodimers with different degrees of transcriptional activity [51]. e classical form of NF-B, and the dimer found most commonly in the liver, is a heterodimer composed of p50 and p65 [49]. In unstimulated cells, NF-B is sequestered in the cytoplasm by inhibitors of B (I B) proteins, of which there are currently at least 5 known isoforms. I Bs prevent nuclear localization of NF-B by masking its nuclear localization signal peptide and block NF-B from binding to DNA by allosteric inhibition [52].
Two modes of NF-B activation have been described to occur in the liver during I/R injury. e �rst mode is the classical pathway of NF-B activation in which cell stimulation results in the serine phosphorylation of I B by the I B kinase complex (IKK complex). is kinase complex consists of two catalytically active subunits, IKK and IKK , and a nonenzymatic regulatory scaffold protein I K (also known as NF-B essential modi�er, NE�O) [53,54]. Phosphorylated I B then becomes the target of ubiquitin ligase which polyubiquitinates the protein for subsequent proteasomal degradation [55,56]. In addition to this wellcharacterized pathway, there appears to be an alternative method of NF-B activation that does not involve the IKK complex, serine phosphorylation, and proteosome mediated degradation of I B. is alternate mechanism of NF-B activation was originally described in hypoxic T cells and involves the phosphorylation of I B on tyrosine residue 42 that leads to its dissociation from NF-B [57]. However, tyrosine-phosphorylated I B is not proteolytically degraded. Experimental data suggest that activation of NF-B via this mechanism occurs predominantly aer hypoxia, whereas the classical pathway occurs primarily aer cytokine stimulation [58][59][60][61]. For both mechanisms of activation, once NF-B is freed from I B it translocates to the nucleus where it initiates the transcription of target genes.
e activity of free NF-B is also tightly regulated. First, NF-B activates the transcription of its own inhibitor I B , leading to the termination of NF-B response which is rapid but transient [56,62]. Second, posttranslational modi�cations of NF-B subunits by phosphorylation and acetylation affect the transcriptional activity, stability, DNA binding affinity, and subcellular localization of NF-B [63][64][65]. In particular, the robust induction of NF-B requires the phosphorylation of p65 on its serine residue 276 which permits the subsequent recruitment of coactivator CBP/p300 [66][67][68]. Multiple other p65 phosphorylation sites involving multiple kinases have also been identi�ed [63]. Overall, the phosphorylation of p65 appears to be essential for the optimal activity of NF-B. Similar to phosphorylation, acetylation of p65 occurs at multiple sites and affects the function of NF-B. For example, acetylation of lysine 221 on p65 enhances DNA binding and hinders association of I B , whereas the acetylation of lysines 122 and 123 reduces its DNA binding affinity [69,70]. p65 is the primary transcriptional activating component of NF-B and it appears that other NF-B subunits, such as p50, are expendable for function during liver injury and recovery [71,72].
Lastly, the nuclear translocation and stability of the released NF-B appear to be regulated by the prolyl isomerase, Pin-1. Pin-1 is an isomerase that binds to the phosphorylated serine-or threonine-proline motif of its target proteins, and causes cis-trans isomerization about the peptidyl-prolyl bond [73]. Depending on the substrate, this can affect substrate's stability, cellular localization, activity level and its ability to interact with other proteins. Not surprisingly, Pin-1 has diverse roles in the cellular processes which are re�ected in the diverse pool of substrates that include transcription factors, mitotic proteins and cytoskeletal proteins. NF-B p65 is one of the substrates of Pin-1, which binds to the phosphorylated threonine 254-proline motif of the p65 [74]. e binding of Pin-1 greatly stabilizes the NF-B complex by blocking the ubiquitin ligasemediated degradation of p65. Furthermore, Pin-1 inhibits the binding of I B to p65. is enhanced stability and inhibition of I B binding to p65 result in increased nuclear localization and prolonged activity of NF-B. e impact of Pin-1 in the activity of NF-B is seen in p65 mutants that have a threonine to alanine substitution at amino acid position 254. Such mutation prevents Pin-1 from binding to p65 and results in rapid degradation as well as failure of nuclear localization of NF-B [74]. e function of Pin-1 in liver I/R injury was recently con�rmed in a mouse model in which it was shown that Pin-1 expression was requisite for adequate p65 stability [75]. In addition, this study showed that normothermic ischemia reduced Pin-1 expression in hepatocytes, but not �upffer cells, suggesting a cell-speci�c regulation of NF-B activation during I/R injury by Pin-1.

Hepatic Recruitment of Neutrophils
e propagation of in�ammation in the liver aer I/R by TNF and, to a lesser degree, by IL-1 is accomplished through induction of the expression of adhesion molecules on vascular endothelial cells and stimulation of the production and release of CXC chemokines. ree classes of vascular cell adhesion molecules contribute to the adhesion and transmigration of neutrophils from the blood vessel lumen into the interstitial spaces. Selectins are glycoproteins expressed on endothelial cells (E-and P-selectins), platelets (P-selectin), and neutrophils (L-selectin) [76]. Selectins are responsible for leukocyte capture and transient adhesion to the vascular endothelium and all three family members have roles in leukocyte adhesion during hepatic I/R injury [77][78][79][80]. e increased neutrophil-endothelium interactions mediated by selectins facilitate the engagement of the other two classes of adhesion molecules, integrins and the immunoglobulin-like adhesion molecules. Integrins are expressed on the neutrophil surface (i.e., CD11b/CD18) and bind to immunoglobulin-like adhesion molecules that are expressed on the vascular endothelium (i.e., intercellular adhesion molecule-1, ICAM-1) [81]. ese interactions mediate �rm adhesion and transmigration and are essential for neutrophil recruitment into the liver aer I/R [82,83].
TNF is clearly the primary stimulus for vascular cell adhesion molecule expression in the liver aer I/R. TNF is responsible for increasing hepatic vascular endothelial expression of P-selectin as well as ICAM-1 [43,84]. P-selectin expression is not only important for the adhesion of neutrophils, but also for the adhesion of platelets to the hepatic endothelium [85]. Increased platelet accumulation within the hepatic microcirculation may enhance the subsequent adhesion of neutrophils, as adherent platelets are known to augment neutrophil adhesion at sites of in�ammation [86]. More importantly, increased neutrophil accumulation has been attributed to sinusoidal endothelial cell injury, contributing to hepatic microvascular dysfunction [87]. Hepatic vascular expression of ICAM-1 is also increased by TNF , and blockade of TNF attenuates neutrophil accumulation and subsequent liver injury [38,43].
In conjunction with vascular cell adhesion molecules, chemokines are an integral component of the process of neutrophil recruitment. Chemokines are a group of small (8-10 kD), basic, heparin-binding proteins that are secreted by leukocytes as well as various tissue cells [88,89]. While mainly involved in leukocyte chemoattraction, chemokines have also been implicated in other cellular activities, including regulation of angiogenesis, �brosis, proliferation, cytotoxicity, and apoptosis [90][91][92][93]. e nomenclature for chemokines is based on the con�guration of a conserved amino-proximal cysteine-containing motif [94]. ere are currently four branches of the chemokine family, CXC, CC, CX 3 C, and C (where X is any amino acid). CC and CXC are the two major branches, whereas CX 3 C and C each have only one representative, consisting of fractalkine (CX 3 CL1) and lymphotactin (XCL1), respectively [95]. e CC family is the largest, primarily involved in attracting mononuclear cells to sites of chronic in�ammation, while members of the CXC family mediate the chemoattraction of neutrophils and monocytes to sites of acute in�ammation [91]. CXC chemokines can be further classi�ed by the presence or absence of a Glu-Leu-Arg (ELR) amino acid motif in the Scienti�ca 5 amino terminus of the peptide. e ELR motif confers receptor-binding speci�city [96,97].
CXC chemokines exert their effects through the CXC chemokine receptors (CXCR) 1-6 [95]. CXCR1 and CXCR2 bind speci�cally to CXC chemokines which contain the ELR motif [90,94]. In addition to their leukocytechemoattractant properties, ELR + CXC chemokines have been shown to have important roles in angiogenesis and cellular proliferation [46,92,98]. CXCR1 and CXCR2 are expressed by neutrophils, monocytes, CD8+ T cells, epithelial cells, and endothelial cells, as well as in hepatocytes [99][100][101]. TNF and IL-1 stimulate the production of a number of chemokines [94]. Chemokine production and/or presentation by endothelial cells activates neutrophils during their initial interactions with the vascular endothelium and promotes their subsequent �rm adhesion. Hepatic production of chemokines forms a chemotactic gradient which serves to direct the recruitment of neutrophils into the injured liver. ere appears to be selective differences in the expression of various CXC chemokines in the ischemic liver versus the nonischemic liver that causes the preferential recruitment of neutrophils into the ischemic liver [45]. CXC chemokines are also upregulated in remote organs, including lung, and play an important role in the development of remote organ injury aer liver I/R [44,46].

Neutrophil-Mediated Hepatocellular Injury
In liver, accumulation of activated neutrophils within the hepatic parenchyma causes hepatocyte damage through the release of oxidants and proteases. Effective killing of hepatocytes by neutrophils probably requires direct cell contact via CD11/CD18-and ICAM-1-dependent mechanisms [102,103]. e primary neutrophil oxidant-generating pathway involves NADPH oxidase. Under normal conditions, this enzyme exists as inactive subunits located both on the cell membrane and in the cytoplasm. Cell activation causes translocation of cytosolic subunits to the cell membrane, resulting in assembly of a multimeric complex that exhibits oxidase activity. e active enzyme oxidizes NADPH and the released electron reduces molecular oxygen, forming O 2 • , superoxide anion. In addition to the generation of oxidants, activated neutrophils release a number of mediators by granule exocytosis. e contents of neutrophil granules include large amounts of proteases (i.e., elastase, cathepsin G, heparanase, and collagenase) and hydrolytic enzymes that may be directly cytotoxic to hepatocytes [106,107]. Serine proteases, such as elastase and cathepsin G may directly damage membrane components of hepatocytes, while metalloproteinases primarily degrade basement membrane and matrix components.

Modes of Hepatocyte Death after I/R
Apoptosis and necrosis are two distinct forms of cell death that differ morphologically. Necrotic cells are characterized by the loss of plasma membrane integrity and cellular architecture, vacuolization, and mitochondria swelling. Apoptotic cells, on the other hand, have as hallmarks chromatin condensation and nuclear fragmentation, cell shrinkage, and formation of apoptotic bodies. Although these two forms of cell death appear very different, they have some important similarities. First, mitochondrial dysfunction is a critical component of both forms of cell death. Speci�cally, opening of the nonselective permeability transition pores on the inner mitochondrial membrane leads to uncoupling of oxidative phosphorylation, membrane depolarization, and leaching of factors involved in cell death [108]. Second, both forms of cell death can be triggered by the same stimuli. In fact, intensity of a given stimulus and the intracellular ATP level appear to be important factors that determine whether a cell undergoes apoptosis or necrosis. An apoptotic stimulus can induce necrosis at higher intensity/concentration [109]. Alternatively, apoptotic stimuli can also cause necrosis if the intracellular ATP is depleted [110,111]. ere has been considerable debate about the primary mode of liver cell death aer I/R. Some laboratories have reported substantial hepatocyte apoptosis [112], while others have shown that broad caspase inhibitors protect against I/R injury [113]. However, critical examination has shown that many of the parameters used to assess apoptosis in these studies also are positive in necrotic cells and that the �nal mode of death in the vast majority of hepatocytes aer I/R is necrosis [114].
Because cells that would die by apoptosis undergo necrosis if intracellular ATP is depleted (a condition induced by prolonged ischemia), it is likely that necrotic cells seen aer I/R may represent two populations: one population consisting of cells that incurred severe lethal damage and die by necrosis; a second population consisting of cells that initially were triggered to undergo apoptosis but, due to the lack of intracellular ATP, switched to necrotic cell death. Inhibition of apoptosis may provide protection against I/R injury by allowing the latter population of cells, that are injured but viable, a chance to survive and recover. is concept is supported by studies in which inhibition of apoptosis by overexpression of the antiapoptotic gene Bcl-2 reduced liver injury aer I/R [115,116].
Given the central role of TNF in the injury response to hepatic I/R, it could serve as a primary stimulus for apoptosis. In fact, blocking TNF production greatly reduces hepatic injury and apoptotic parameters aer I/R, whereas the inhibition of Fas signaling had no effect on injury or evidence of apoptosis in this setting [117]. Similarly, TNF receptor-1 (TNFR-1)-knockout mice demonstrate less hepatic insult and apoptosis aer I/R [117]. ese �ndings suggest that TNF may induce apoptotic signaling aer I/R which may contribute to overall hepatocellular dysfunction and death.
One of the best known functions of NF-B is its role in inhibiting apoptosis aer TNF exposure. e in�ammatory and apoptotic response of TNF is mediated through the signaling proteins that are recruited to TNFR-1 upon ligand binding. TNFR-1 associated death domain (TRADD) binds to the receptor and is involved in the transduction of both responses [118]. Binding of Fas-associated death domain protein (FADD) to TRADD leads to TNF-induced apoptosis that involves caspase 8 [119]. On the other hand, binding of TNFR-associated factor (TRAF)-2 and receptorinteracting protein (RIP) results in activation of NF-B and c-Jun NH 2 -terminal kinase (JNK). Although TNF can elicit a potent apoptotic response, cells usually do not undergo apoptosis aer exposure to TNF because NF-B opposes this response. Failure to activate NF-B aer TNF stimulation results in apoptosis. For example, p65 −/− cells undergo apoptosis aer TNF stimulation. Cells that express a nonphosphorylatable form of I B fail to activate NF-B upon TNF stimulation and also undergo apoptosis [120,121]. Some of the most compelling evidence regarding the anti-apoptotic effects of NF-B came from the generation of p65 knockout mice, which die in utero due to massive hepatic apoptosis [122]. NF-B prevents apoptosis by transcribing a number of anti-apoptotic genes. For example, NF-B induces transcription of TRAF1 and 2 and inhibitory apoptotic protein (IAP)-1 and 2 which when expressed together block apoptosis by inhibiting caspase 8 activation [119]. When protein synthesis is blocked with cycloheximide, cells that are normally resistant to TNF become susceptible, re�ecting the fact that de novo synthesis of proteins induced by NF-B is required for cell survival [123]. us, NF-B may have a protective role aer I/R by inhibiting apoptotic signaling and allowing injured but viable cells a chance to recover rather than succumb to secondary necrosis.
Like all homeostatic processes, regulatory mechanisms exist to help prevent a runaway in�ammatory train that may otherwise lead to an overwhelming response. e degree of the insult oen is directly proportional to the magnitude of the in�ammatory response and with the evolution of advanced trauma stabilization and surgical procedures, patients are now surviving insults that in years past would have been lethal. While this is good for the patient initially, it sets the stage for an in�ammatory response proportional to a lethal stimulus. In this setting, regulatory mechanisms are oen overwhelmed and cannot effectively control the proin�ammatory response. ese regulatory mechanisms are oen induced well aer the initial insult and therefore represent a potential avenue for therapeutic intervention/supplementation. Our current knowledge of the endogenous mediators that regulate the hepatic in�ammatory response is quite limited. For example, to date there have been few anti-in�ammatory mediators identi�ed that play substantial endogenous roles in control of the hepatic response to I/R. e cytokines IL-6, IL-10, and IL-13 have all been shown to be expressed during hepatic I/R injury [124][125][126]. However, only IL-6 and IL-13 appear to play important regulatory roles. IL-6 has been shown to limit hepatocellular injury and promote hepatocyte regeneration aer I/R [126]. ese effects of IL-6 were linked to its capacity to reduce the expression of TNF and elaboration of c-reactive protein. IL-10, while expressed by the liver aer I/R, does not appear to play a signi�cant regulatory role [125]. Exogenous administration of IL-10, however, is highly protective and appears to suppress proin�ammatory cytokine expression by inhibiting activation of the transcription factor, NF-B [127,128]. e function of IL-13 is more complex. Exogenous administration of IL-13 prevents I/R injury by activating the transcription factor, STAT6, leading to blockade of transcriptional activation of the genes for TNF and MIP-2 [129]. STAT6 was subsequently found to compete with NF-B for nuclear transcriptional coactivators and decreased NF-B transcription activation [130]. Studies of IL-13-knockout mice provide a different story to the role of this cytokine in the endogenous regulation of liver in�ammation. IL-13 nullizygous mice display far more hepatocellular injury, but this occurred without signi�cant alterations in NF-B activation, proin�ammatory mediator expression, and coincided with a decrease in neutrophil accumulation [125]. ese studies went on to show that endogenous IL-13 has multiple effects on the liver including a positive modulatory effect on expression of the adhesion molecule, vascular cell adhesion molecule-1 (VCAM-1). A decrease in hepatic VCAM-1 expression in IL-13-knockout mice was associated with a neutrophil transmigration defect leading to increased adhesion of neutrophils to the hepatic venular endothelium and increased hepatic endothelial cell injury. In addition, this study found IL-13 to directly protect cultured hepatocytes from oxidant-induced cytotoxicity [125].
e protease inhibitor, secretory leukocyte protease inhibitor (SLPI), has been shown to be a very potent endogenous regulator of the hepatic in�ammatory response to I/R [131]. is small protein mediator was originally described as a secretory product of phagocytes that inhibited neutrophil elastase. It has subsequently been shown to be produced by a variety of cell types in various organs and has more complex functions, including inhibition of the transcription factor, NF-B [131][132][133]. Hepatic production of SLPI occurs prior to reperfusion, making somewhat unique amongst anti-in�ammatory mediators, which are normally expressed aer the initial surge of proin�ammatory mediators. Endogenous SLPI appears to regulate liver in�ammation by targeting the transcription factor, NF-B, and attenuating proin�ammatory cytokine expression [131]. Additionally, its properties as a potent protease inhibitor may contribute to neutralization of destructive enzymes released by activated neutrophils [106,[133][134][135][136].
Nitric oxide (NO) has also been shown to be an important regulatory mediator in the liver. Its role in the injury response, however, has been somewhat controversial [137][138][139][140][141][142][143][144][145][146][147]. In the liver, NO is produced by both endothelial nitric oxide synthase (eNOS) as well as inducible nitric oxide synthase (iNOS), with the latter being expressed primarily in Kupffer cells and sinusoidal endothelial cells. Studies have shown that NO is produced by eNOS functions in a highly protective manner, with pharmacological inhibition strategies and eNOS-knockout mice showing greatly enhanced indices of hepatocellular injury, whereas production of NO from iNOS increases liver injury [148][149][150][151]. Currently, it is not clear whether eNOS-derived NO protects the liver by scavenging reactive oxygen species or by modulation of proin�ammatory mediator expression.

Repair and Regeneration of Liver after I/R
Hepatocytes possess the unique ability to proliferate upon appropriate stimulation, normally maintaining themselves in a stage of quiescence, known as the G 0 phase. e molecular basis of liver regeneration is composed of three different phases, including a priming phase, a proliferative phase, and a termination phase [152]. ese phases have also been quali�ed as cytokine, growth factor, and metabolic pathways, respectively, as it pertains to the factors predominantly mediating a particular phase [153]. It is important to note that while it is conceptually easier to denote the sequence of events into "phases, " there is in fact a highly coordinated, synchronous schema of interactions between growth factors, cytokines, and other mediators that allow the process of liver regeneration to occur [154]. Cytokines are a key factor in stimulating quiescent hepatocytes from the G 0 phase into the G 1 phase. TNF and IL-6, along with the transcription factors, STAT3 and NF-B, are required for the initiation of liver regeneration [153,155]. rough activation of STAT3 and NF-B, target genes are transcribed which are important to hepatocyte proliferation. Growth factors, speci�cally hepatocyte growth factor (HGF) and epidermal growth factor (EGF), then drive the cell from G 1 into the S phase of DNA replication [153]. Arguably one of the most important mediators of liver regeneration is HGF-a 100 kDa protein that was originally identi�ed in 1984 [156,157]. A potent mitogen for hepatocyte growth, HGF, is locally released and upregulated during the initiation of the regenerative process, cleaved from its inactive single-chain form into its active twochain form by uPA [152].
Phospholipase C 1 (PLC 1), phospholipase C 1 (PLC 1), phospholipase D1 (PLD1), and phosphoinositide-3-kinase (PI3K) have been implicated in the mechanisms of hepatocyte proliferation immediately aer HGF or EGF binding [158][159][160][161]. PLC 1 and PLC 1 appear to play different roles in the regenerating liver, with PLC 1 having more in�uence on the G 2 /M phase transition, and PLC 1 seeming to trigger DNA replication [158]. PLD1 may play a role in the activation of c-Jun/c-Fos transcription factors, further contributing to DNA synthesis [162]. e HGF receptor, a c-met oncogene, has been shown to function through tyrosine kinase activity. However, Adachi et al. [163], showed that pertussis toxin-sensitive G proteins were also involved in mitogen activated protein kinase (MAPK) activation and arachidonic acid release, speci�cally demonstrating that PLD activation was diminished to baseline levels in the presence of G i receptor complex inhibition. More recently, signaling through PI3K has been shown to be critical for the induction of cyclin D and DNA replication following HGF binding [161]. Further downstream, MAPK-dependent production of arachidonic acid (AA) through PLA 2 results in production of prostaglandins, further stimulating DNA synthesis [160]. Prostaglandins, most signi�cantly PGE 2 and PGF 2 , are known to promote growth in hepatocytes [164]. Conversely, during conditions in which hepatocytes may be stressed, activation of PLA 2 and increased release of arachidonic acid may have a deleterious effect on hepatocytes [165]. In the setting of hypoxic injury to hepatocytes, diminished ATP production leads to acidosis, therefore preventing activation of PLA 2 until the return to physiologic pH during reperfusion, resulting in AA release and increased cell death [165,166]. In vivo studies have revealed that COX-2dependent conversion of arachidonic acid to prostaglandins is crucial to the induction of protective mechanisms within the liver, and that COX-2 inhibition contributed to greater hepatotoxicity in the setting of carbon tetrachloride (CCl 4 ) injury, perhaps indicating that the level of COX-2 following hepatic injury is important to recovery [167].

Contribution of CXC Chemokines to Liver Repair and Regeneration
As described above, CXC chemokines are known to be important mediators of the in�ammatory cascade following hepatic injury and also appear to have a dichotomous role in hepatocytes that may be related to their level of expression [168]. For example, induction of CXC chemokines at relatively low levels is associated with liver repair and regeneration, whereas high expression levels have been associated with hepatotoxicity [101,[169][170][171]. e impact of CXC chemokines on the regenerative capacity of the liver was �rst examined using an in vivo model of partial hepatectomy [169]. Partial hepatectomy represents a clinically relevant model of hepatic resection, a procedure oen performed due to trauma or malignancy. ELR + CXC chemokines are upregulated aer partial hepatectomy, and blockade of chemokines or CXCR2 results in diminished liver regeneration [169]. Subsequent in vitro experiments demonstrated that hepatocytes treated with ELR + CXC chemokines proliferated to a degree similar to that induced by HGF. ese studies [101,[169][170][171] provided evidence that ELR + CXC chemokines were important hepatocyte proliferative factors that functioned in vivo to promote liver regeneration aer hepatectomy. However, as previously mentioned, the remnant liver aer resection or hepatectomy, without Pringle maneuver, is comprised of unstressed hepatocytes. e role of CXC chemokines may be distinctly different in a setting in which hepatocytes are under signi�cant stress, such as I/R. Liver recovery and repair aer I/R injury in this model begins approximately 48 hours aer reperfusion and is associated with increased expression of stathmin and marked hepatocyte proliferation [172]. Liver repair and regeneration typically return the liver to its normal, homeostatic state 8 Scienti�ca within 5-7 days aer reperfusion, depending on the severity of the injury. It is during this reparative/regenerative phase that it appears that the function of CXC chemokines switches from a proin�ammatory role to direct impingement on hepatocyte proliferation or death. Knockout of CXCR2, the primary receptor for ELR + CXC chemokines in rodents, resulted in accelerated liver recovery associated with increased activation of NF-B and STAT3 transcription factors resulting in increased hepatocyte proliferation [170]. Antibody blockade of CXCR2 aer induction of I/R injury had the same effect [170]. ese studies suggest that during the reparative/regenerative phase of I/R injury, ELR + CXC chemokines have harmful effects which delay liver recovery.
ese apparent harmful effects of ELR + CXC chemokines are likely a result of speci�c signaling via CXCR2 in hepatocytes. While the presence and involvement of CXCR2 in murine models of hepatocyte injury and regeneration have been well characterized [169,170], murine CXCR1 has only been recently identi�ed [173,174]. Previous work has demonstrated that CXCR2 is constitutively expressed in hepatocytes [101] and may be upregulated in the presence of certain cytokines [175]. While CXCR2 and its ligands appear to play a key role in hepatocyte proliferation following partial hepatectomy, and hepatocyte toxicity following I/R injury or acetaminophen toxicity, the role of CXCR1 is less clear. CXCR1 is not constitutively expressed in the liver [173]. is �nding was recently con�rmed, but CXCR1 were found to be induced in hepatocytes aer I/R [100]. Blockade and knockout of CXCR1 was found to result in delayed liver repair aer I/R, although there were no observed changes in hepatocyte proliferation in vivo [100]. While the effects of CXCR1 blockade or knockout on liver repair were not as striking as those observed with CXCR2, the �ndings suggest that CXCR1 has a divergent function in hepatocytes, compared to CXCR2.
While the stress level of the hepatocyte may alter its response to CXC chemokines, so may the concentration of available ligand. Following 70% partial hepatectomy, expression of ELR + CXC chemokines increases approximately 5-fold [169], whereas aer I/R it increases hundreds-to thousandsfold [170]. In vitro, stimulation of primary hepatocytes with CXC chemokines has hepatoprotective effects at low concentrations and progressively cytotoxic effects at increasingly greater concentrations, effects which are speci�c to CXCR2 [100,170]. Adenoviral-mediated liver overexpression (>100-fold) of the CXC chemokine, keratinocytederived chemokine, has been shown to result in massive hepatocellular necrosis within 48 hours [71]. Collectively, these studies suggest that moderate increases in CXCR2 ligands, occuring aer partial hepatectomy may promote liver regeneration, whereas much larger increases in expression of CXCR2 ligands, occuring aer I/R injury, may be hepatotoxic and/or oppose hepatocyte proliferation and regeneration.

Conclusions
Hepatic I/R injury is a primary complication of liver resection and transplantation and is also a consideration during vascular and trauma surgery. e impact of this injury on patient morbidity and mortality is signi�cant. Experimental studies have identi�ed the primary mechanisms of this injury response, which begins as an oxidative stress and culminates in a robust in�ammatory response leading to neutrophilmediated injury to hepatocytes. is entire process is regulated largely by in�ammatory cytokines and is regulated by endogenous expression of anti-in�ammatory mediators that serve to resolve the response. An equally complex process for tissue repair and regeneration of lost functional mass includes participation of several proin�ammatory mediators, such as CXC chemokines, which also serve as secondary mitogens for hepatocytes. Collectively, these experimental �ndings have helped identify many new therapeutic targets that can help reduce the incidence of and mitigate the impact of I/R injury to the liver to improve patient care.