Molecular and Cellular Basis of Liver Failure

Clinical Manifestations

ALF is a clinical syndrome resulting from rapid loss of hepatocyte function. Hepatic encephalopathy is, by definition, present to some degree in all patients with ALF. Cerebral edema is a cardinal feature and may produce uncal herniation, yielding brainstem compression and death. ALF requires a multidisciplinary, collaborative effort among hepatologists, transplant surgeons, intensive care physicians, nephrologists, and neurosurgeons. Patients should be rapidly evaluated for cause and severity of liver injury, and an urgent assessment should be made regarding suitability for liver transplantation. Hallmarks of presentation include coagulopathy as evidenced by an elevated international normalized ratio greater than or equal to 1.5, jaundice, and elevated serum aminotransferase levels. Other common clinical manifestations include loss of vascular tone with hypotension, renal failure, infection and/or sepsis, hypoglycemia, electrolyte abnormalities, cardiac dysfunction, acute lung injury, gastrointestinal bleeding, and disseminated intravascular coagulation. Portal hypertensive bleeding and severe fluid retention are distinctly unusual ( Table 3-1 ). The syndrome of ALF is associated with high mortality, with most patients dying from cerebral edema and sepsis.

Some causes of ALF are associated with a better prognosis than are other causes. In general, the more rapid onset forms of ALF have a higher incidence of cerebral edema but an overall better prognosis, probably reflecting the lack of liver architectural derangement and thus more favorable conditions for hepatic regeneration. ALF due to acetaminophen overdose, hepatitis A, shock liver, or pregnancy-related disease showed a 50% or more transplant-free survival. In contrast, ALF caused by idiosyncratic drug reactions, Wilson’s disease, and indeterminate causes tends to carry a particularly poor prognosis.

The determination of the prognosis for ALF has immense value. Irreversible ALF recognized early can be treated so that life-threatening complications can be prevented. In turn, patients with recoverable liver function would be spared unnecessary surgery. Several prognostic indices have been developed. King’s College Hospital criteria are the most widely used, and they include clinical and biochemical data routinely available in clinical practice. However, no prognostic model to date has proved reliable in determining the prognosis for ALF, and ALF remains an unpredictable disease with high morbidity and mortality.

Etiology

ALF results from the abrupt loss of liver function secondary to severe injury from a variety of causes that may be grouped into several general categories ( Table 3-2 ). The most common causes of ALF in the United States currently are drugs and toxins, in particular, acetaminophen (APAP; 46%), ALF of indeterminate cause (14%), hepatitis A and B viral infections (11%), autoimmune disorders (5%), ischemia (4%), Wilson’s disease (2%), and a cluster of other diverse metabolic, structural, and undetermined causes (14%). Most patients who develop an acute hepatitis episode recover spontaneously. Recovery rate is higher if the triggering event is related to certain causes (i.e., hepatitis A virus, transient hypoxia, paracetamol intoxication, and mushroom poisoning). Although acetaminophen hepatoxicity is the most common cause of ALF in the United States, viral causes are the predominant cause of ALF in developing countries. The incidence of ALF from viral hepatitis A and B in the United States appears to be decreasing, perhaps in part the result of an active vaccination program, and together they now account for less than 10% of ALF cases per year. These diverse metabolic, toxic, and inflammatory insults result in liver injury and disease. A common feature of these insults is activation of apoptotic cell death. The subsequent sections of this chapter will discuss the pathogenesis of liver failure, focusing on the experimental evidence for cytotoxic pathway activation and molecular mechanisms whereby insult is translated into damage, and ultimately hepatobiliary disease.

Pathogenesis

To develop more effective prognostic tools and treatments in ALF, it is necessary to elucidate the molecular pathways that dictate the pathological changes and ultimately influence outcome. Knowledge of the degree of ongoing hepatic regeneration would be a useful tool, given that recovery of patients with ALF is thought to reflect the capacity of the liver for regeneration. The partial hepatectomy (PH) model in rodents has been a mainstay in studying hepatocyte proliferation and the initiation of the downstream cascades resulting in liver regeneration. Our most current model based on PH defines distinct phases of regeneration, each involving cytokine pathway interactions (e.g., tumor necrosis factor-α [TNF-α], interleukin-6 [IL-6]) between hepatocytes and nonparenchymal cells in conjunction with growth factor stimulation (epidermal growth factor [EGF], heparin-binding EGF, transforming growth factor-α [TGF-α], and hepatic growth factor [HGF]), which together induce transcription of early genes in hepatocytes (e.g., c-fos , c-jun , c-myc ) and activate multiple signaling pathways (mitogen-activated protein kinase [MAPK], signal transducer and activator of transcription 3 [STAT3], phophatidylinositol-3 kinase [PI3K]/Akt, and extracellular signal-regulated kinase [ERK1/2]) promoting the G0/G1 transition and cell progression. Although the hepatocyte is often the focus of attention in ALF, all of the various liver cell types ( Table 3-3 ) undoubtedly play important roles. Indeed, recent studies have shown the importance of macrophages, hepatic stellate cells (HSCs), lymphocytes, natural killer T (NKT) cells, natural killer (NK) cells, and endothelial cells in liver regeneration. Moreover, liver regeneration after PH may provide molecular insights into the self-renewal of mature cells, a property often ascribed exclusively to stem cells.

The extent to which hepatic stem cells mediate liver regeneration remains under intense study. The liver progenitor cell, or oval cell, is widely used to describe hepatic progenitors; however, there is no consensus on the phenotypic or molecular traits of these cells. The rapid reconstitution of liver mass following injury is usually fulfilled by resident hepatocytes. However, in circumstances that overwhelm hepatocyte regeneration, progenitor cells reconstitute hepatic parenchyma, as determined in transplantation studies, and contribute to the formation of bile ducts. Based on the available data, it appears that oval cell activation reflects the effects of inflammatory cytokines (IL-6, IL-18, interferon-γ, TNF) and intracellular signaling pathways (e.g., Janus kinase (JAK)/STAT, Sonic hedgehog), initiating a cascade of events culminating in differentiation into biliary cells and hepatocytes. Further studies to delineate the molecular mechanisms controlling differentiation of hepatic progenitor cells are ongoing.

Full recovery from ALF is possible, and this suggests that outcomes may be improved not only if hepatic regeneration is enhanced, but if cell death is curtailed. A common feature of liver injury is activation of apoptotic or necrotic cell death. Hepatocytes can undergo apoptosis via an extrinsic, death receptor–mediated pathway or alternatively the intracellular intrinsic pathway of apoptosis. The molecular pathways leading to cell death are highly regulated and overlapping. Key regulatory signals from innate immunity cells (Kupffer cells, NKT cells, and NK cells) responding to injury interact with hepatocytes to initiate a molecular cascade resulting in hepatocyte apoptosis. Cytokine release (TNF, interferon-γ, IL-6) results in activation of multiple transmembrane signaling pathways (FasL, TNF-related apoptosis-inducing ligand [TRAIL], c-jun N-terminal kinase [JNK]), which in turn activate transcription factors (NF-κВ, c-jun, c-fos), mitochondrial proteins (Bcl-2, Bid, Bim, Bax, and Mcl-1), and caspases. Ultimately these pathways converge on the mitochondria, causing mitochondrial dysfunction, which is a prerequisite for hepatocyte apoptosis.

Prognosis in ALF depends on the balance of liver cell death with liver repair and regeneration. Indeed, survival critically depends upon rapid and robust recovery of liver cell function before the life-threatening complications, such as cerebral edema and sepsis, of ALF supervene.

Liver Regeneration and Repair

Although many of the main molecular pathways involved in liver regeneration after PH have been deciphered, recent studies highlight new insights into mechanisms involved in this process.

The adequacy of liver repair and regeneration following acute liver injury appears to be as important as the extent of the injury in determining outcome. Hepatic regeneration represents the culmination of a complex interaction among liver cells, matrix, cytokines, and hormones and is characterized by the activation of more than 100 genes encoding cytokines, growth factors, transcription factors, and cellular constituents. HGF, EGF, TGF-β, TNF-α, and IL-6 appear to have particularly important roles in hepatic regeneration. The plasma concentration of HGF, produced primarily by stellate cells, increases dramatically within 1 hour of a PH, and it acts through its receptor, c-Met, which is highly expressed on hepatocytes. Studies of growth signals in cultured hepatocytes have shown a fivefold to tenfold increase in DNA synthesis of HGF, and receptors for the ligands EGF and TGF. TNF-α, released primarily from Kupffer cells, although not directly mitogenic itself, appears to play a critical role in the initiation of the transcriptional cascade contributing to hepatocyte replication and experimentally is shown to enhance the effects of HGF, EGF, and TGF. Furthermore, proliferation is strongly enhanced by combining HGF and EGF.

The molecular mechanisms underlying hepatic regeneration have been elucidated primarily in the PH rodent model, in which two thirds of the liver, including the left lateral and medial lobes, is removed intact. Under normal conditions, only a small fraction of hepatocytes (∼1/20,000) are in mitosis. When hepatocytes are injured and die, they are usually replaced by mature hepatocytes. This was demonstrated by a critical early experiment in rodents using radiolabeled nucleotides after 70% PH that showed that nearly all hepatocytes incorporate radioactive nucleotides during liver regeneration. This landmark observation established that resident hepatocytes actively divide to recover the original cell number and liver mass, and that hepatocytes undergo roughly one or two rounds of cell division after 70% PH. After PH the onset of liver cell replication is rapid, with the peak of hepatocyte DNA synthesis occurring within approximately 24 hours, and the peak of nonparenchymal cell DNA synthesis occurring approximately 24 hours later. Amazingly, normal liver mass is restored after only 7 to 10 days following 50% PH in rats. More recent studies using a genetic tracing method to directly assess cell division has shown that not all hepatocytes undergo cell division. Interestingly, in the 70% PH model no cell division was observed in more than 40% of hepatocytes, and in a 30% PH model no cell division occurred even though liver cell mass was recovered in a shorter time interval compared to the 70% PH model. These observations indicate that hepatocyte proliferation alone does not account for liver regeneration after PH. Therefore recovery of liver mass encompasses both hypertrophy and hyperplasia. Increased hepatocyte size occurs much earlier than entry into the cell cycle, suggesting that cell size increase is the first response of hepatocytes to the loss of liver mass. This very early stage of liver regeneration is known as the priming phase, in which hepatocytes dramatically change their gene expression pattern to prepare for regeneration.

Biochemical studies and gene targeting technology have revealed influences of several signaling molecules in activation of cell cycle–associated genes and key transcription factors (e.g., cyclin D1, STAT3, and NF-κВ). Three main phases of liver regeneration after PH have been used to illustrate the molecular pathways of hepatocyte repopulation ( Fig. 3-1 ). In the “streaming liver hypothesis,” during the initial priming phase of replication, normally quiescent hepatocytes enter the cell cycle—moving from the G0 to the G1 phase—and become receptive to growth factors and replication competent. This phase, which lasts 4 to 6 hours, requires the secretion of cytokines such as TNF-α and IL-6. Increased circulating levels of TNF-α and IL-6 lead to the activation of the STAT3 pathway within the hepatocytes. Activation of the STAT3 pathway induces transcription of early genes in hepatocytes, including the proto-oncogenes c-fos , c-jun , and c-myc . Activation of these genes ultimately leads to progression through the early to mid-G1 phase of the cell cycle. Epidermal growth factor receptor (EGFR) and c-Met are activated during the second phase and stimulate progression through the cell cycle (G1 through S phases). EGFR downregulation induces a delayed and reduced hepatocyte proliferation because of a defect in G1/S progression with a compensatory activation of other ErbB receptors and c-Met. c-Met receptor regulates G2/M progression through an ERK1/2 activation. Both EGFR and c-Met will then recruit scaffolding proteins and activate multiple intracellular intertwined networks, among which MAPK, STAT3, PI3K/Akt, and ERK1/2 are the most important for liver regeneration. The early activation of NF-κB by a rapid posttranscriptional mechanism activates expression of IL-6, which in turn activates STAT3 and other genes. When NF-κB activity is blocked after PH, the residual liver undergoes massive apoptosis. Genetically modified mice that lack IL-6 or the receptor for TNF-α have deficient liver regeneration and develop liver failure following PH that is ameliorated by recombinant IL-6 administration, strongly suggesting that IL-6 is acting downstream of TNF-α in the regeneration cascade. Much less is known about how liver regeneration is terminated once the appropriate liver mass is restored. Although this final phase of regeneration must occur, the factors involved remain elusive. The TGF-β superfamily is known to be involved in this step. However, in mice lacking TGF-β receptor, hepatic overgrowth is only transient. Studies in Drosophila wing mass development have yielded conserved nuclear receptor kinases in mammalian species that also control hepatocyte proliferation. This suggests that other regulatory factors are involved and collaborate to stop liver growth.

However, in the setting of severe ALF, hepatic regeneration is impaired despite high serum levels of IL-6, TNF-α, and HGF, suggesting another pathway of regeneration. Whether new hepatocytes in the regenerating liver are derived from adult hepatocytes, intrahepatic stem cells, or circulating stem cells remains unclear. Current studies favor the hypothesis of an expansion of a progenitor cell population during regeneration and normal liver homeostasis, or the so-called streaming liver hypothesis ( Fig. 3-2 ). In this model, differential gene expression by hepatocytes arises during the hepatocyte maturation process, which represents lineage progression. A population of small portal zone cells in the Canal of Hering with a high nuclear-to-cytoplasmic ratio known as oval cells proliferate extensively and, upon migration into the lobule, differentiate into hepatocytes. The oval cell is best described as a heterogeneous liver progenitor cell whose exact phenotypic markers are not clearly defined, although cellular markers in multiple species have been identified (c-kit, flt-3, CD34, leukemia inhibiting factor, Thy-1, Sca-1/CD34/CD45, and OV6). Numerous studies have shown that in massive liver injury, where the typical regenerative pathways are overwhelmed, regeneration is strongly dependent on oval cell proliferation. The reliance on hepatic oval cells/progenitor cells to repopulate the liver in massive hepatic injury, such as seen in ALF, makes them a key focus for targeted therapy. The molecular mechanisms regulating the activation, proliferation, and differentiation of these cells is being elucidated. Both in vitro and in vivo studies demonstrate the importance of interferon-γ, TGF-β, IL-6, and TNF in activation and proliferation of oval cells. The most well-defined pathway used transgenic mice expressing TNF-like weak inducer of apoptosis (TWEAK) and showed hepatocytes from these mice display an oval cell response and progenitor-specific signaling in the liver. Hedgehog has also been implicated in progenitor activation induced by alcoholic steatohepatitis. Hedgehog inhibitors can impair progenitor proliferation, and indirect evidence suggests that activation of hedgehog signaling might be downstream of TGF-β. The ability to generate, manipulate, and then potentially transplant these hepatic progenitor cells could prove immensely valuable as a therapeutic intervention to treat severe ALF.

Necrosis, Apoptosis, and Hepatic Cell Death

Like other cells, liver cells die from apoptosis and necrosis. These two pathways of cell death are morphologically distinct but interrelated, and that probably should be viewed as two ends of a cell-death continuum. The cell-death pathway taken, either apoptosis or necrosis, appears to be related to the nature and severity of the inciting insult, the cell type, its metabolic status, and the integrity of the cell-death machinery. Both types of cell death probably occur simultaneously in most forms of ALF, and the same stimulus can result in either pathway. Morphologically, necrosis results in cell swelling, loss of cell membrane integrity, and lysis, which invariably elicits a secondary immune response. Adenosine triphosphate depletion due to loss of mitochondrial oxidative phosphorylation is a biochemical hallmark of necrosis. Mitochondrial dysfunction in necrosis is characterized by the mitochondrial permeability transition (MPT), pores that collapse the ion gradient across the inner mitochondrial membrane. MPT results in failure of the ion gradient, which drives oxidative phosphorylation. Cyclophilin D knockout mice have been shown to inhibit MPT and limit ischemic tissue injury. Necrosis is a prominent feature of APAP-induced liver injury, and N -acetyl- p -benzoquinone imine detoxification is associated with oxidative stress. Studies show that mice deficient in cyclophilin D are protected from APAP-induced liver injury and DNA damage. Also, JNK kinase is activated by APAP and mediates liver injury, whereas inhibition of JNK in APAP-injected mice protects from APAP toxicity. Oxidative injury to mitochondria secondary to TNF-α will also result in opening of MPT pores. This leads to release of intramitochondrial cytochrome c and apoptosis-inducing factor and to initiation of the apoptosis cascade via caspase-9. In general, liver cell necrosis rather than apoptosis tends to predominate, with extensive oxidative damage to mitochondria because this depletes cellular adenosine triphosphate stores and also may inhibit caspase activity, both of which are necessary for the successful execution of the apoptosis pathway.

In contrast, apoptosis, or programmed cell death, is characterized by a more orderly process of nuclear and cytoplasmic shrinkage, condensation, and blebbing without loss of cell membrane integrity or release of intracellular contents; thus it allows cellular debris to be removed without intense secondary inflammation and marked perturbation of neighboring cells. Hepatocyte apoptosis can be considered to be a pivotal step in most forms of liver injury. Apoptosis is a highly conserved process essential to organogenesis and immune cell homeostasis that was first recognized pathologically in liver 3 decades ago as acidophilic (Councilman) bodies. However, fundamental insights into the molecular details of the apoptosis pathway are more recent, initially gleaned from experiments in the worm Caenorhabditis elegans and only later in mammalian cells.

Diverse factors trigger liver cell death, such as hypoxia (e.g., with ischemia-reperfusion), reactive oxygen species (e.g., generated during drug metabolism), viral infection, and autoimmune injury. Susceptible hepatocytes then undergo apoptosis via an extrinsic death receptor–mediated pathway, or an intracellular stress-mediated intrinsic pathway. In either case, participation of mitochondria appears to be essential for apoptosis in hepatocytes ( Fig. 3-3 ). The extrinsic pathway of apoptosis involves the sequential activation of death receptors (Fas or TRAIL), followed by activation of a series of cysteine proteases called caspases and subsequent mitochondrial permeabilization. An alternative pathway via TNF-α signaling leads to lysosomal activation and subsequent mitochondrial failure. The intrinsic pathway is triggered by a variety of insults signaling apoptosis via JNK activation, or cellular organelles, such as mitochondria, endoplasmic reticulum (ER), and lysosomes. These cascades converge with the activation of apoptotic proteins (Bax, Bim, Bad, Bid) and inhibition of antiapoptotic proteins (Bcl-2, Bcl-xL), which results in mitochondrial permeabilization.