Potential conflict of interest: Nothing to report.
Hepatitis A virus (HAV), the causative agent of type A viral hepatitis, is an ancient human virus that was first identified almost 35 years ago. It has several characteristics that make it unique among the Picornaviridae, particularly in terms of its mechanisms of polyprotein processing and virion morphogenesis, and which likely contribute to its pathobiology. Although efficacious vaccines containing formalin-inactivated virus produced in cell culture have been licensed in multiple countries, their use has been limited by cost considerations. Changes in public health sanitation and generally increasing standards of living are leading to a decreasing incidence of acute hepatitis A worldwide, with the result that the prevalence of preexisting immunity among adults is declining in many regions. These changes in the epidemiology of HAV may paradoxically enhance the disease burden, as greater numbers of individuals become infected at older ages when disease is more likely to be clinically evident, thus providing greater incentives for vaccine utilization. (Hepatology 2006;43;S164–S172.)
More than 30 years have passed since the virus responsible for hepatitis A—hepatitis A virus (HAV)—was first visualized by immune electron microscopy in suspensions of fecal samples collected from infected human volunteers.1 The initial identification of the HAV particle followed years of earlier research that had defined many of the clinical features of the disease, documented the fecal–oral nature of HAV transmission, and distinguished the infection from “homologous serum jaundice,” otherwise known as hepatitis B. Six years later, just before publication of the first issue of Hepatology , cultured mammalian cells were shown to be permissive for HAV replication.2 Soon after, the RNA genome of the virus was reverse-transcribed and a cDNA copy of it was molecularly cloned and, in 1987, RNA transcripts derived from a genome-length cDNA clone were shown to be infectious when transfected into cultured cells.3 In the meantime, continued progress in the cell culture propagation of HAV, and the development of methods for the successful inactivation of the virus, led to the production and licensure of safe and very effective vaccines for prevention of hepatitis A by the early 1990s.4, 5
Since the licensure of the first hepatitis A vaccines, interest in the disease and the virus that causes it has waned. With a vaccine on the shelf, and the means to protect against the disease with nearly 100% efficacy, hepatitis A has been accorded a lower priority for research by national funding agencies. Research efforts in the field of hepatitis have focused increasingly on the more recently identified hepatitis C virus (HCV), because it causes frequent chronic infections and is perceived to represent a more pressing public health problem. HAV causes only acute hepatitis (Fig. 1) and is not associated with chronic liver disease. Although the incidence of hepatitis A has fallen dramatically since the introduction of vaccines, HAV remains more than a troublesome nuisance in the United States. Almost 6,000 cases were reported in 2004, representing an estimated 60,000 cases nationwide. Moreover, HAV causes occasional, dramatic disease outbreaks, as well as isolated severe cases of fulminant hepatitis with fatal outcomes in otherwise healthy adults.
HAV, hepatitis A virus; HCV, hepatitis C virus; dsRNA, double-stranded RNA; IRF, interferon regulatory factor.
Molecular Virology of HAV
The molecular cloning of the viral genome led to recognition that HAV should be classified within the Picornaviridae family, in a separate genus, the genus Hepatovirus. As such, our understanding of HAV has benefited from studies of the prototype picornavirus, poliovirus. Much of what we know of the HAV particle, the function of its genome, and its means of replication, have been gained indirectly. However, in recent years there has been growing appreciation of a number of features of the virus that are unique to hepatoviruses, including important details of the processing of the viral polyprotein, the morphogenesis of the virion, and the interactions of the virus with host cells (reviewed in6). These unique attributes of HAV are likely to be of particular relevance to the pathobiology of the virus.
Like HCV, HAV is a positive-strand RNA virus. However, unlike HCV, it lacks a lipid envelope. As a member of the Picornaviridae, its RNA genome is packaged within an icosahedral protein capsid composed of 60 copies of each of 3 major structural proteins, VP1, VP2, and VP3 (also known as 1D, 1B, and 1C). The genome itself is approximately 7,500 nucleotides in length and contains a single large open-reading frame encoding a polyprotein in which the major capsid proteins represent the amino-terminal third (P1 segment), with the remainder of the polyprotein comprising a series of nonstructural proteins required for HAV RNA replication: 2B, 2C, 3A, 3B (a small protein, also known as VPg, that is covalently linked to the 5′ end of the genomic RNA and that probably serves as the protein primer for RNA synthesis), 3Cpro (a cysteine protease responsible for most post-translational cleavage events within the polyprotein), and 3Dpol (the viral RNA-dependent, RNA polymerase) (see Fig. 2). Similar to HCV, an internal ribosome entry site located within the 5′ nontranslated RNA segment of the HAV genome directs the 5′ cap-independent translation of the polyprotein.
HAV has been adapted to replicate in many types of cultured mammalian cells, including cells of non-hepatic or non-primate origins.7 Several features of its replication cycle distinguish it from poliovirus and many other well-studied picornaviruses, including its slow and protracted time course, low virus yields, and a propensity to establish persistent infections in cell culture. The viral proteins are translated directly from the messenger-sense genomic RNA, which is delivered to the cytoplasm after uncoating of the viral particle. Proteolytic processing of the polyprotein occurs simultaneously with translation and is largely carried out by the 3Cpro protease (Fig. 2). Synthesis of the RNA follows the assembly of a large, macromolecular replicase complex containing the nonstructural viral proteins spanning the 2B-3Dpol segment of the polyprotein, and occurs on membranes that are usurped for this purpose from the cellular endoplasmic reticulum8 (Fig. 3). As with poliovirus, RNA transcription is most likely protein-primed, with uridylylated VPg (protein 3B) representing the primer for both a negative-sense RNA replication intermediate and subsequent positive-sense progeny RNA molecules.
Considerable controversy has surrounded 2 segments of the polyprotein that are likely to be involved in the assembly of virus particles: the extreme amino terminus, which comprises a short polypeptide sequence representing a putative fourth capsid protein, VP4 (also known as 1A), and the segment immediately downstream of the capsid proteins (2A segment), which represents a carboxy-terminal extension of VP1, the largest capsid protein (Fig. 2). The primary polyprotein cleavage event occurs at the 2A/2B junction, and differs from other well-studied picornaviruses in that it is mediated by the 3Cpro protease (Fig. 2).9 The resulting P1-2A structural precursor is further cleaved by the viral protease to generate 2 capsid protein precursors, VP0 (VP4-VP2) and VP1-2A (also known as pX),10 as well as the mature VP3 capsid protein (Fig. 2). VP1-2A is a critical structural intermediate in virion morphogenesis.11 Cleavage at the HAV VP1/2A junction occurs late in the process of virion morphogenesis and results from the action of an unknown cellular protease.12, 13 However, the mature 2A protein has never been identified directly in infected cells.
In the case of poliovirus, all of the non-structural proteins, including the 2A protease, as well as some precursors of the mature nonstructural proteins, are necessary for RNA replication. With HAV, however, this is not the case. Infectious virus can be recovered from recombinant HAV genomes containing exogenous protein-coding sequences inserted in-frame at the 2A/2B junction and flanked by consensus 3Cpro cleavage sites,14 indicating that the non-structural 2A polypeptide does not function in cis as a 2AB precursor. Viral RNA replication is not impaired when the C-terminal 60% of the 2A sequence is deleted.11 Furthermore, the complete deletion of 2A has no effect on the replication capacity of subgenomic RNA replicons lacking the P1 capsid protein region (Yi et al., unpublished observations). Thus, the non-structural 2A protein segment is not required for RNA synthesis.
Conversely, HAV RNA lacking sequence encoding the N-terminal 40% of 2A is not capable of producing infectious virus particles,11 indicating a role for the 2A segment in capsid assembly. Such a deletion prevents the assembly of the structural precursors into pentamers, important intermediates in capsid morphogenesis that contain 5 copies of each of the capsid protein precursors VP4-VP2 (VP0), VP3, and VP1-2A (Fig. 4). This aspect of HAV particle assembly appears to be unique to the hepatoviruses.
The role of the small VP4 polypeptide in virion morphogenesis is unknown. Most other picornaviruses have 4 polypeptides within their capsid, including a small VP4 protein located at the amino terminus of the polyprotein. Whereas the polyprotein of HAV appears to possess a very short VP4 polypeptide segment at its amino terminus, this putative VP4 moiety has never been demonstrated directly in purified virus preparations. Moreover, whereas N-terminal myristoylation of other picornavirus VP4 proteins is important for virion morphogenesis, the HAV VP4 sequence does not contain a similarly placed myristoylation signal.15 Some data suggest that VP4 is necessary for assembly of pentamers into empty capsids.16 However, we have found that empty capsids can be generated in the absence of VP4 (Bénichou et al., unpublished data). Why the assembly of HAV particles should differ from that of other picornaviruses, such as poliovirus, is not clear, but it is tempting to speculate that it might relate specifically to the unique intrahepatic lifestyle of HAV.
Questions abound concerning the structure of the infectious HAV particle. Attempts to determine the atomic structure of HAV by x-ray crystallography have not been successful, although such studies have provided high-resolution images of virus particles from each of the other major picornaviral genera. Recently, however, medium-resolution images of the HAV particle have been obtained by cryo-electron microscopy (R.H. Cheng, unpublished data) (Fig. 5). Although this work remains in progress, this exciting new view of the HAV particle suggests significant differences in its structure compared with other picornaviruses. In particular, no well-defined “canyon” surrounding the particle's 5-fold-axes, a prominent feature and the site of cellular receptor binding in other picornaviruses, is present.
A putative receptor for HAV, an integral membrane mucin-like glycoprotein of unknown natural function, has been identified in African green monkey kidney cells. A human homolog of this protein also has been identified as a receptor binding HAV.17 However, this protein is not expressed selectively in the liver, and other work suggests that the hepatocellular asialoglycoprotein receptor may play a role in viral entry by mediating the uptake of immunoglobulin A (IgA)–virus complexes.18
The growth of wild-type virus is generally poor in cultured cells, and the virus must undergo a process of adaptation before becoming capable of efficient replication. This is reminiscent of the need for HCV replicon RNAs to acquire adaptive mutations before they are capable of efficient replication in human hepatoma cells,19 although the molecular basis of cell culture adaptation is likely to be quite different. Adaptive mutations that permit HAV to replicate efficiently in cell culture have been extensively characterized. Mutations within the internal ribosome entry site enhance cap-independent viral translation in a cell-type–specific fashion, whereas mutations in 2B promote viral RNA replication in all types of cultured cells.20–22
A few highly cell culture–adapted, rapidly replicating strains of HAV have been isolated.23, 24 These viruses are cytopathic and appear to cause cell death by inducing apoptosis.8, 24 However, in most HAV-infected cells, both in cell culture and probably also in vivo, there is no cytopathic effect. The virus apparently down-regulates its replication in cells commonly used for its propagation (FRhK-4 and MRC-5 cells).25 These observations may be of relevance to the mechanisms underlying establishment of persistent infections in vitro.
Although a large proportion of the newly replicated virus remain cell-associated, extensive release of progeny virus into cell culture supernatant fluids occurs by an unknown mechanism. In polarized, human colonic epithelial cell cultures, release of virus occurs almost exclusively into apical supernatant fluids, mimicking the secretion of HAV across the apical canalicular membrane of the hepatocyte into the biliary system (see below).26 This process is largely blocked by an inhibitor of the cellular secretory pathway, suggesting that virus release may involve vectorial cellular vesicular transport mechanisms.
Infection with HAV usually occurs by the fecal–oral route of transmission and is associated with extensive shedding of the virus in feces during the 3- to 6-week incubation period and extending into the early days of the illness (Fig. 6). This explains the high prevalence of infection in regions where low standards of sanitation promote transmission of HAV. HAV is exceptionally stable at ambient temperatures and at low pH. These features of the virus explain its ability to survive in the environment and to be transmitted by contaminated foods and drinking water. Resistance to acid pH and detergents also accounts for its ability to transit through the stomach, and to exit the host via the biliary tract. These are important features that contribute significantly to the pathogenesis of hepatitis A.
Chimpanzees, as well as several species of New World monkeys, including marmosets, tamarins, owl monkeys, and Saimiri monkeys, are susceptible to HAV and may be infected by either oral or percutaneous challenge.27–30 Much has been learned from these nonhuman primate models of hepatitis A, although they do not recapitulate the disease perfectly. Liver injury is usually mild compared with symptomatic infections in adult humans, although the course of the infection is otherwise very similar. Large amounts of virus are present in feces from 1 to 4 weeks after either percutaneous or oral exposure.31 Most of this virus appears to be produced in the liver and to reach the intestinal contents by secretion from infected hepatocytes via the biliary system.32 Nonetheless, some data suggest that HAV may undergo initial replication within crypt cells of the small intestine before reaching the liver.31
Fecal shedding of the virus reaches its maximum just before the onset of hepatocellular injury, at which point the individual is most infectious. This is accompanied by an extended viremia which roughly parallels the shedding of virus in the feces, but at a lower magnitude (Fig. 6).33 Liver injury follows (Fig. 1), often with marked elevation of serum aminotransferase activities. Viral antigen can be detected within the cytoplasm of hepatocytes, as well as within germinal centers of the spleen and lymph nodes and along the glomerular basement membrane in some primates.30 Viral antigen typically continues to be shed for 2 to 3 weeks after the first elevation of enzymes, although sensitive reverse transcription polymerase chain reaction assays can detect continued shedding of viral RNA for many weeks. Prolonged shedding of the virus has only been documented in infected premature infants.34 Importantly, older epidemiologic studies have demonstrated the disappearance of HAV from closed populations with time, suggesting that long-term, persistent fecal shedding of virus does not occur.
The mechanisms responsible for hepatocellular injury in hepatitis A (Fig. 1) are poorly characterized. However, type A hepatitis appears to be attributable to an immunopathologic response to infection of the hepatocyte, rather than to a direct cytopathic effect of the virus. HLA-restricted, virus-specific, cytotoxic, CD8+ T cells have been recovered from the liver in acute hepatitis A.35 Such cells have been shown to secrete γ-interferon, which likely stimulates the recruitment of additional, nonspecific inflammatory cells to the site of virus replication within the liver.36
Host Innate and Adaptive Immune Responses
After infection via the gastrointestinal tract, HAV replicates quietly within the liver for several weeks or more during the incubation period of the disease. By the end of this period, high titers of virus are present within liver tissue, bile, stool, and, to a lesser extent, blood (Fig. 6). Despite this, there is little evidence of liver injury and often no disease. Not until the fourth or fifth week of the infection do clinical manifestations of the infection appear along with the first evidence of an immune response to the virus.
Recent studies suggest that this prolonged period of clinical quiescence in the face of mounting viral replication may reflect the ability of the virus to disrupt cellular mechanisms by which mammalian cells recognize virus infection and induce synthesis of interferon-β.25 In HAV-infected cells, double-stranded RNA (dsRNA)-mediated activation of interferon regulatory factor 3 (IRF-3) is blocked. IRF-3 is a key transcription factor that is constitutively expressed in the cytoplasm. It is phosphorylated after virus infection, leading to its nuclear translocation and subsequent induction of interferon-β synthesis. Within hepatocytes, this occurs as the result of signaling transduced through 2 distinct pathways, 1 initiated by dsRNA engagement of Toll-like receptor 3, and the other by the interaction of dsRNA with a novel pathogen-associated molecular pattern receptor, the RNA helicase, retinoic acid–inducible gene I.37
IRF-3 activation through the retinoic acid–inducible gene I pathway is effectively blocked in cells infected with HAV,38 although the specific HAV protein responsible for this evasion of the host response is not known. Signaling through the Toll-like receptor 3 pathway also may be impaired partially. Remarkably, HCV infection also blocks the activation of IRF-3 through both pathways,39 and this has been postulated to play a role in the persistence of HCV infections. However, HCV also blocks the virus-induced activation of nuclear factor-κB (NF-κB), another critical transcription factor that controls expression of apoptotic pathways. Importantly, this does not appear to be the case with HAV,38 indicating a fundamental difference in the ways in which these 2 hepatotropic viruses go about disrupting innate cellular antiviral defenses. Very limited studies in cell culture suggest that HAV is sensitive to type 1 interferons,40 but it is unclear whether HAV, like HCV,39 also expresses proteins that antagonize the specific anti-viral effector mechanisms induced by interferons.
Unlike the situation with HCV, the adaptive immune response to HAV is robust and extremely effective in eliminating the virus. Neutralizing antibodies to the virus (anti-HAV) generally appear in the serum concurrent with the earliest evidence of serum aminotransferase elevation and hepatocellular injury (Fig. 6). The early antibody response is composed largely of IgM, although IgG may also be present shortly after the onset of symptoms. Anti-HAV IgG persists for life and confers protection against reinfection. Neutralizing antibody is of primary importance in protection against HAV infection and disease, as passively transferred antibody has been known for decades to protect against hepatitis A even when administered post-exposure. Re-exposure of seropositive individuals may lead to increases in anti-HAV titer, but is not associated with liver disease. In part, this reflects the absence of significant antigenic differences among strains collected from different regions of the world. Both fecal and serum anti-HAV IgA have been described, but the role of secretory immunity in protection against HAV infection appears to be very limited.41 Virus-specific, HLA-restricted cytotoxic T-cells have been identified within the liver during acute HAV infection and probably play roles both in viral clearance and in the production of liver injury.35, 42
Prevention of Hepatitis A
There is still a role for immune globulin in the prevention of hepatitis A, particularly when individuals are recognized to have been exposed to the virus within the preceding 2 weeks.43 However, its use in pre-exposure prophylaxis has been largely supplanted by inactivated hepatitis A vaccines. These vaccines contain viral particles that are produced in cell culture, purified, inactivated with formalin, and adsorbed to an aluminum hydroxide adjuvant. Inactivated HAV vaccines are highly immunogenic and protect against both infection and disease.4, 5 This protection is likely primarily antibody based,44 and is broadly directed against all strains of HAV, consistent with the identification of a single serotype of HAV among human strains.
In the past in industrialized countries, hepatitis A vaccination has been recommended for persons at increased risk of acquiring hepatitis A, including travelers to regions of high hepatitis A endemicity, users of illicit drugs, homosexually active men, and patients with clotting factor disorders who receive factor concentrates.43 Immunization also has been recommended for persons who are at increased risk of developing fulminant disease should they become infected with HAV, such as persons with chronic HCV infection. In recent years, however, recommendations for broader immunization of children in regions of the United States having a higher overall incidence of hepatitis A have led to an 88% decline in reported cases of hepatitis A in those states.45 Such data, coupled with evidence that the vaccine is immunogenic in children over the age of 12 months, recently led the Advisory Committee on Immunization Practices of the U.S. Centers for Disease Control and Prevention to recommend universal childhood immunization against hepatitis A in all regions of the country.a
Similarly, in Israel, a country of intermediate HAV endemicity, a recent study concluded that a nationwide, universal infant HAV immunization policy is both medically and economically justified.46 The incidence of hepatitis A among those younger than 14 years of age decreased by more than 95% after the introduction of universal immunization, and even disease rates in the elderly (who are particularly prone to severe or fulminant hepatitis) declined substantially. The ability of universal immunization to favorably impact the burden of hepatitis A disease is likely to be particularly evident in countries where the average age of infection has increased in recent years as socioeconomic conditions have improved. Such changes have led to increases in the incidence of clinically apparent infections that have an economic cost and also carry a risk of severe disease.
In contrast, in Southern Europe, where several studies have also confirmed an increase in the hepatitis A attack rate among susceptible adults, universal hepatitis A immunization seems less economically attractive at present.47 Cost–benefit analyses conducted in countries of high endemicity, such as Thailand,48 also suggest that the benefits of immunization do not justify the expenses incurred, regardless of prior screening for HAV antibodies and regardless of the age group targeted. These conclusions largely reflect the relatively high cost of the vaccine, which relates in part to difficulties inherent in propagating HAV in cell culture, but also possibly to the current limited scale of manufacture. As mentioned, however, changes in the epidemiology of hepatitis A may alter the future perspective on immunization in countries where sanitation is undergoing rapid improvement.
Candidate live, attenuated HAV vaccines have been developed using viruses that have been adapted to growth in cell culture. There was considerable enthusiasm for such vaccines early in the development process, but vaccine candidates were poorly immunogenic.49 Attenuation seems to have been achieved with these viruses at the cost of their ability to replicate within the liver, not by adaptation to a site of replication within the body that is different from that at which the virus causes disease (as is the case with attenuated poliovirus vaccines). Nonetheless, a live, attenuated hepatitis A vaccine has received relatively wide use in China and appears capable of inducing protective levels of antibody.50 A study conducted in children suggested that a single dose of this vaccine was approximately 95% effective in preventing overt disease, but that it was less effective in preventing asymptomatic infection during an outbreak of hepatitis A.51 Whereas an attenuated vaccine might have some advantages, inactivated vaccines work well. Furthermore, the magnitude of the challenge of bringing an attenuated HAV vaccine to successful licensure within the United States or western Europe is likely to preclude any serious efforts on the part of the vaccine industry to develop such a vaccine.
Summary and Conclusions
Because chronic hepatitis is not associated with HAV infection, relatively little “fear factor” is associated with hepatitis A. This sense of complacency, particularly given the availability of safe and extremely effective inactivated HAV vaccines, has contributed to a declining interest in the molecular virology of HAV and the pathogenesis of hepatitis A in recent years. Despite this, hepatitis A remains a public health problem in many countries, with approximately 1.5 million clinical cases reported worldwide annually, a figure that greatly underestimates the true incidence of infection. Recent changes in the incidence of HAV have led to modifications of immunization policies in the United States, something that is likely to occur in other countries in the future. In the meantime, HAV and its interactions with the hepatocyte represent a fertile field for future investigation. In particular, exploring the nature of this interaction may provide hints as to why this virus does not persist in the infected host, whereas HCV does.
The authors thank their co-workers, Lisette Cohen, Michael Beard, Danièle Bénichou, and MinKyung Yi for numerous stimulating discussions and past collaborative efforts in unraveling the many mysteries of HAV. We also gratefully acknowledge R. Holland Cheng (University of California, Davis) for providing the cryo-EM image of the HAV particle (Fig. 5) before publication, and the laboratory of K. Bienz, University of Basel, for assistance with the electron microscopic image shown in Fig. 4.
On October, 27th, 2005, the ACIP adopted a resolution calling for immunization of all children in the United States with hepatitis A vaccine at 12 months of age; detailed recommendations are currently being drafted by the Centers for Disease Control and Prevention to replace those cited in reference (43). For details see: http://www.cdc.gov/nip/recs/provisional_recs/hepA_child.pdf.