J Gastrointest Canc DOI 10.1007/s12029-013-9556-9
REVIEW ARTICLE
The Role of Cirrhosis in the Etiology of Hepatocellular Carcinoma Michael C. Kew
# Springer Science+Business Media New York 2013
Abstract Abundant evidence supports the belief of a causal relationship between cirrhosis and hepatocellular carcinoma, but one that differs between high- and low-incidence regions of the tumor. In high-incidence regions, the cirrhosis is of the macronodular variety, is typically asymptomatic, and is caused predominantly by chronic hepatitis B virus infection, whereas in low-incidence regions, the cirrhosis, although usually macronodular, may be micronodular, is commonly symptomatic and of long-standing, and is caused by chronic hepatitis C virus infection, alcohol abuse over many years, the metabolic syndrome, or hereditary hemochromatosis. In a minority of patients, hepatocellular carcinoma develops in the absence of cirrhosis, supporting a direct hepatocarcinogenic effect of some of the causal agents. Cirrhosis is the major risk factor for tumor formation in patients with chronic hepatitis C virus infection. This virus does not integrate into cellular DNA, and malignant transformation results from increased liver cell turnover induced by recurring injury and regeneration of cells in the context of persisting inflammation, oxidative DNA damage, fibrosis, cirrhosis, and changes induced by the virus at a DNA level that have yet to be fully defined. Hepatitis B virus causes malignant transformation by both direct and indirect routes. The direct route results, in part, from integration of the viral DNA into host cellular DNA; transcriptional
M. C. Kew Department of Medicine, Groote Schuur Hospital and University of Cape Town, Cape Town, South Africa M. C. Kew Department of Medicine, Johannesburg Hospital and University of the Witwatersrand, Johannesburg, South Africa M. C. Kew (*) Department of Medicine, Old Main Hospital Building, Groote Schuur Hospital, Main Road, Observatory, Cape Town, South Africa e-mail: [email protected]
activation of host growth regulatory genes by hepatitis B virus-encoded proteins; and effects on apoptosis, cell signaling, and DNA repair. The direct route may share some similarities with that of hepatitis C virus infection. The metabolic syndrome may cause malignant transformation by production of oxidative stress and the induction of a variety of mutations, including some in the p53 gene. Keywords Hepatocellular carcinoma . Cirrhosis . Hepatitis C virus . Hepatitis B virus . Metabolic syndrome
Co-existence of Cirrhosis and Hepatocellular Carcinoma Cirrhosis and hepatocellular carcinoma (HCC) often co-exist, raising the possibility that cirrhosis may play a role in the etiology and pathogenesis of the tumor. The two pathologies do not, however, have parallel geographical distributions. In most regions, HCC is uncommon [1], and cirrhosis is far more prevalent, whereas in Southeast Asia and sub-Saharan Africa, HCC occurs commonly, and may exceed cirrhosis in incidence [1]. Nevertheless, considerable evidence supports the existence of a causal relationship between cirrhosis and HCC, albeit one that differs between low- and high-risk regions of the tumor [2]. In geographic regions with low incidences of HCC, such as Europe, Oceania and North America, the great majority of patients who develop the tumor do so against a background of long-standing symptomatic cirrhosis. The cirrhosis is usually macronodular and is most often the result of long-standing alcohol abuse or, increasingly in recent years, chronic hepatitis C virus (HCV) infection [3–5] or the metabolic syndrome [6]. By contrast, in the high incidence regions of HCC in Southeast Asia and sub-Saharan Africa, chronic hepatitis B virus (HBV) infection is the major cause, and the far less common cirrhosis is almost always macronodular in type (7–12). The
J Gastrointest Canc
symptoms of cirrhosis, if any, are overshadowed by those of the tumor, and the presence of cirrhosis is typically recognized only during the diagnostic work-up for HCC or at necropsy (8–11). Further evidence, albeit indirect, for the differences in the presentation of HCC between patients in low- and highincidence regions of the tumor is provided by the observation that in low-incidence regions, the patients with co-existing HCC and cirrhosis are between 6 and 20 years (average, 8– 10 years) older than those with HCC in a noncirrhotic liver, or in those with cirrhosis alone [5, 12–15]. By contrast, Chinese [16, 17] and sub-Saharan Black African [18–21] patients with cirrhosis and high incidences of HCC, and Japanese [22] patients with cirrhosis and moderate incidences of the tumor, are the same age as those with HCC without cirrhosis, or those with cirrhosis alone. Evidence that the underlying cirrhosis in individuals in populations at low risk of HCC is at an advanced stage is provided by the observation that patients with both the tumor and cirrhosis are more likely than those with the tumor alone to be jaundiced, to have ascites, and to bleed from esophageal varices [12]. In addition, laboratory tests in these patients are more likely to show a prolonged prothrombin time, a low serum albumin concentration, and raised serum bilirubin and aspartate aminotransferase levels [12]. These clinical findings are no more common in subSaharan Black African and Chinese patients with both HCC and cirrhosis than in those with HCC alone [8, 10, 11, 18–21], and tests of hepatic function and damage are similar in Black Africans with HCC with and without cirrhosis [2, 21]. In the presence of cirrhosis, HCC is far more common in men than women (5–11:1) [5, 12, 16, 17, 21]. But when the tumor occurs in a noncirrhotic liver in high-risk regions, male predominance is less striking (4:1) [19, 20]; in low-risk regions, the sex ratio is almost parity [5, 12, 21]. In countries with a low incidence of HCC, patients in whom cirrhosis co-exists with the tumor survive for a significantly shorter time than do those with the tumor alone [12], whereas in countries with a high incidence of the tumor, the survival times in all of the patients are so short that any difference would be biologically insignificant [19, 20]. The presence of co-existing cirrhosis in populations at low risk for HCC also has an effect on the size of the tumor. For example, in an analysis of North American patients, the mean weight of the tumorous liver at necropsy was 2,530 g in the presence of cirrhosis and 3, 810 g in its absence [22]. The presence or absence of cirrhosis has a lesser effect on tumor size in populations at high risk for HCC. Thus, in sub-Saharan Black Africans with HCC and cirrhosis the mean weight of the tumorous liver was 3,692 g, compared with 4,195 g in patients without cirrhosis [23].
Type and Etiology of Cirrhosis Co-existing with Hepatocellular Carcinoma The cause and type of cirrhosis that co-exists with HCC differ appreciably between high- and low-risk regions for the tumor.
High-Risk Regions Between 60 and 85 % of Chinese and sub-Saharan Black African patients with HCC and co-existing cirrhosis show serological markers of current HBV infection, and antigenic markers for the virus can often be detected in the cirrhotic liver [20, 21, 24]. A far smaller percentage of patients with HCC in these populations is chronically infected with HCV [25]. No specific differences in the clinical manifestations of cirrhosis or HCC have been documented between patients chronically infected with HBV or HCV or with prolonged alcohol abuse. A history of alcohol abuse is uncommon in Chinese patients with HCC. In rural-dwelling sub-Saharan Black Africans, who have the highest incidences of HCC in the subcontinent, evidence of alcoholic liver disease is rarely present. A history of alcohol abuse may be obtained, but the beer consumed has a low alcohol content (3 %; the beer has a high iron content as a result of home brewing of the liquor in iron drums or pots, which leeches iron into the beer and causes iron overload of the liver, which in turn may cause HCC) [26]. In recent years a small, but increasing, number of urban Black African males with HCC have been shown to have evidence of conventional alcohol-induced liver disease [27]. Other causes of cirrhosis, such as hereditary hemochromatosis, Wilson's disease, and primary biliary cirrhosis are rarely, if ever, seen in sub-Saharan Black Africans or Chinese. The often young age of Black African [9, 11, 18] and Chinese [7, 10] patients with HCC, whether or not cirrhosis is present, suggests that the causative agent or agents of one or both diseases are operative at an early age. Certainly, HBV infection is known to be acquired at a very early age in both of these populations [28, 29].
Low-Risk Regions In populations with a low risk of HCC alcohol abuse of long standing remains an important cause of cirrhosis and HCC. In many of the remaining patients in these populations, and in increasing numbers during recent years, serological evidence of chronic HCV infection is present. Also increasing in frequency and importance as a cause of cirrhosis and HCC during this time in these populations is the metabolic syndrome [6]. Chronic HBV infection is present in most of the relatively few remaining patients.
J Gastrointest Canc
In contrast to the young ages at which exposure to environmental carcinogens occurs in regions with a high risk of HCC, the advanced ages of the patients with this tumor in regions at low risk of HCC favors exposure to the relevant etiologic agents during adult life. This is true of HCV infection and alcohol abuse, although the cause of the metabolic syndrome may be in place at a younger age. In patients with alcohol-induced HCC, the time elapsing between diagnosis of cirrhosis and the recognition of the tumor may be longer than that for the other etiologic forms of cirrhosis [14]. The frequency with which a proportion of cirrhotic patients either develop clinically recognizable HCC or are found at necropsy to have a tumor not detected during life, differs between low- and high-risk regions of the tumor, although the main determinants for this phenomenon are the type and etiology of the cirrhosis rather than the geographic location per se [2]. The risk is greatest with macronodular cirrhosis, irrespective of its etiology. In Japanese, Chinese, and Black Africans, the frequency of tumor development in a macronodular cirrhotic liver (which is almost always of viral origin) is of the order of 44–55 % [22, 30]; in low-incidence populations, it may be as high as 33 % [30]. Micronodular cirrhosis, most often as a result of alcohol abuse over many years, carries a lesser risk of tumor formation, with prevalences of 3–10% being usual [31, 32].
The Role of Cirrhosis in the Etiology and Pathogenesis of Hepatocellular Carcinoma In as many as 90 % of patients with HCC, the tumor develops in a cirrhotic liver, irrespective of the cause of the cirrhosis [33, 34, 36–45]. Although the close association between cirrhosis and HCC has long been recognized, the mechanisms by which cirrhosis contributes to malignant transformation have yet to be fully elucidated. The risk is greatest with macronodular cirrhosis. The proportion of patients with this form of cirrhosis who develop HCC ranges, in different parts of the world, from 15 to 55 %, although figures of between 40 and 55 % are usual [22, 24, 27, 33, 38, 40–42, 44]. In abusers of alcohol of long-standing, macronodular cirrhosis is likely to be present, and is complicated by HCC development in 15–24 % of patients, whereas the more common micronodular cirrhosis, present in abusers of shorter duration, is complicated in this way in only 3–10 % of patients [34–36]. The variable risk of developing HCC in the different types of cirrhosis is illustrated by studies in different population groups. For example, Becker and Chatgidakis [41] reported a rate of 44 % in southern African Black patients with HCC in whom the cirrhosis was almost exclusively macronodular, but a rate of only 6 % in Caucasians with predominantly micronodular cirrhosis. Thomson [44] reported corresponding figures of 54.5 and 2.6 % in the same population groups. The
severity of the cirrhosis also plays a role in the frequency of this complication—present in 50 % of patients with severe cirrhosis, compared with 13 % in those with moderate cirrhosis [46]. In patients in high-incidence regions of HCC in whom the symptoms of macronodular cirrhosis, if any, are overlooked, the presence of the cirrhosis is typically discovered coincidentally at the time the histological diagnosis of the tumor is made or at necropsy. One possible explanation for the close etiological association between cirrhosis and HCC is that cirrhosis itself is a premalignant condition, i.e., that hyperplasia of hepatocytes leads over time to neoplasia in the absence of additional causal factors, such as chemical or viral carcinogens. This sequence of events is supported by the observation in experimental animals that almost all forms of cirrhosis are eventually complicated by HCC [46]. But if this mechanism was universally operative, a parallel geographic distribution between cirrhosis and HCC would be expected. Although this may apply to macronodular cirrhosis, it is not true of micronodular cirrhosis. For example, in Central and Northern Europe, with high incidences of alcoholic micronodular cirrhosis, HCC is much rarer than in sub-Saharan Africa or China, where this form of cirrhosis is very uncommon [22, 32]. Moreover, in lowincidence regions of HCC, the number of patients with cirrhosis far exceeds the number with the tumor, whereas in high-incidence regions, notably in sub-Saharan Black Africans, the number of cases of HCC equals or may exceed the number with cirrhosis. Moreover, if neoplasia followed inevitably on hyperplasia, patients with cirrhosis and HCC would be expected to be older than those with cirrhosis alone. This is true in Western countries [28, 46], but not in sub-Saharan Africa or the Far East [11]. Furthermore, for neoplasia to be an inevitable complication of hyperplasia, one would anticipate that the risk of HCC development would be similar in all types of cirrhosis. In clinical practice, the risk rate of HCC development ranges from approximately 50 % with macronodular cirrhosis in subSaharan Black Africans to less than 1 % for primary biliary cirrhosis and cirrhosis in Wilson's disease, which have degrees of hyperplasia of a similar order to that in alcoholic liver disease. Moreover, the risk of malignant transformation is not uniform within one type of cirrhosis. Thus, macronodular cirrhosis in sub-Saharan Africa has a risk rate of HCC development of 44–54 %, whereas in western countries it has a rate of 15–25 %. The most important determinants of malignant transformation in the cirrhotic liver in low-incidence populations are male sex and increasing age, the latter including, at least to some extent, the duration of the cirrhosis [5, 12, 45]. Another possible mechanism of tumorigenesis in the cirrhotic liver is that the presence of cirrhosis renders the individual susceptible to a variety of environmental carcinogens [2]. It might do so by virtue of the increased hepatocyte turnover rate in cirrhosis acting as a tumor promoter, cells in
J Gastrointest Canc
mitosis being more susceptible to DNA alterations by chemicals or other agents. In addition, rapid cell turnover rates interfere with DNA repair processes, resulting in DNA alterations being transmitted to daughter cells, and are thereby fixed in the progeny. These effects are presumably independent on the original cause of the cirrhosis. HBV and HCV are carcinogenic viruses and can cause HCC in the presence or absence of cirrhosis. Although the mechanisms involved in the malignant transformation of these forms of HCC have been more fully investigated than those of the other less common causes of the tumor arising in a cirrhotic liver—alcohol abuse of long standing, the metabolic syndrome [6, 32], hemochromatosis, and hepatic fibrosis [47]—the pathogenesis of all the causes of the tumor remains incompletely understood.
Chronic Hepatitis C Virus as a Cause of Cirrhosis-Associated Hepatocellular Carcinoma Approximately 170 million people (3 % of the global population) are chronically infected with HCV, of whom about 20 % will progress to cirrhosis [48]. The development of HCC in persons chronically infected with HCV is an increasing problem in many countries, but particularly in resourcerich countries. The virus is primarily transmitted parenterally. Partly as a result of its genetic diversity, HCV can evade the host's immune response. Although HCV-induced HCC may develop in a normal liver, it is more often preceded by a spectrum of hepatic pathologies ranging from nonspecific minimal inflammatory changes to cirrhosis [48–50]. In those countries with a low incidence of HCC and in Japan, with a moderate incidence, chronic HCV infection is a leading cause of progressive hepatic fibrosis, cirrhosis, and HCC. Cirrhosis is present in approximately 70 % of all liver cancers in Japan [51]. In countries with a high incidence of chronic HCV infection, cirrhosis is the main risk factor for malignant transformation, with an annual incidence rate of HCC of between 1 and 8 % [52, 53]. The global incidence of HCC is expected to increase over the next approximately 30 years [52].
Mechanisms of Hepatitis C Virus-Induced Hepatocellular Carcinoma The major risk factor for HCC development in chronic HCV infection is the presence of cirrhosis, with as many as 70–90 % of HCCs arising in this way [48, 52–54]. In the remaining patients, HCC develops in a noncirrhotic liver, supporting a direct carcinogenic effect of HCV and indicating that cirrhosis is not a prerequisite for malignant transformation [48, 52–54].
HCV does not integrate into the host genome, and the mechanisms responsible for its hepatocarcinogenicity are therefore likely to differ, in part, from those of chronic HBV infection, in which the viral genome is frequently integrated. Malignant transformation of hepatocytes induced by HCV occurs as a result of sequences of increased liver cell turnover induced by chronic liver injury and regeneration in the context of oxidative DNA damage and chronic inflammation [48–50, 53–55]. This microenvironment facilitates the occurrence of genetic and epigenetic alterations that, over decades, may lead to the development of cirrhosis and ultimately HCC. Steatosis is one of the main factors that induce reactive oxygen species, chronic hepatic inflammation, and liver fibrosis, that may ultimately progress to HCC [54, 55]. In HCV transgenic mice that develop HCC, carcinogenesis has been shown to be preceded by development of steatosis [56]. The presence of steatosis supports a direct cytopathic component in the pathogenesis of the malignant transformation [54]. The generation of oxidative stress is a key component of the development of both cirrhosis and HCC [55–67]. In animal models, chronic HCV infection produces oxidative DNA damage, chronic hepatic inflammation, and fibrosis [58–68]. Mitochondrial dysfunction, endoplasmic reticulum stress, immune cell-mediated oxidative outbursts, in addition to DNA damage, contributes to HCV-associated oxidative stress [63]. Oxidative stress may also activate stellate cells, thereby promoting fibrosis [64], and may lead to activation of cell signaling pathways that contribute to cellular transformation [65]. One way in which increased oxidative stress may be accounted for is by NAD(P)H oxidase, especially Nox-2 in peripheral mononuclear cells and Kupffer cells in the liver [65]. HCV also triggers mitochondrial permeability transition with production of reactive oxygen species, leading to activation of NF-ĸB and STAT-3 [59, 67]. Changes induced by the core, envelope, and nonstructural proteins are central to the hepatocarcinogenic effects of HCV [62–67]. These include disrupting cellular regulatory pathways associated with proliferation and apoptosis, and suppressing host immune responses. Some of these effects could also have a bearing on the role of cirrhosis in malignant transformation and thus have an indirect effect in the development of HCC, in addition to the direct carcinogenic effects. The lack of a small animal model of chronic HCV infection has hindered an understanding of these factors. HCV core protein, by itself, has a diverse range of functions, some of which may be relevant to the hepatocarcinogenic potential of the virus [60]. It binds to and inactivates a number of tumor suppressor proteins, including p53, p73, and pRB [68]. It can also modulate the expression of the cyclin-dependent inhibitor p21WAF1, which is a major target of p53 and regulates the activities of cyclin/cyclindependent kinase complexes involved in cell cycle control and tumor formation [69]. HBV core protein also
J Gastrointest Canc
downregulates E-cadherin expression at the transcriptional level, an effect which is strongly correlated with hypermethylation of CpG islands of the E-cadherin promoter [70]. This protein also upregulates the expression of transforming growth factor (TGF)-β [71, 72]. As HCV-infected livers progress from chronic hepatitis through cirrhosis to HCC, hepatocyte pSmad3L/PAI-1 increases and pSmad3C/p21WAF1 decreases with the fibrotic stage and necroinflammatory grade. This has the effect of accelerating fibrosis and the development of cirrhosis, and increasing the risk of malignant transformation [73]. HCV core protein can also activate the Raf/MAPK signal pathway [74], upregulate Wnt-1 [75] and effect transcription of cyclin-dependent kinase inhibitor 21 gene [76]. HCV envelope protein E2 exerts an inhibitory effect on natural killer cells through engagement of CD81 [77]. E2 also activates the MAP/ERK pathway, including the downstream transcription factor ATF-2, and maintains survival and growth of target cells [78]. Of the nonstructural proteins, NS3 protein inhibits the activity of the p21WAF1 promoter in a dose-dependent manner, and acts synergistically with the core protein and p53 in an NS3 sequence-dependent manner [79]. In addition, the expression of NS3 protein enhances cell growth, JNK activation, and DNA-binding activities of the transcription factors AP-1 and ATF-2 [80], and induces TNF-α production by activation of AP-1 and NF-ĸB [81]. NS5A protein is involved in a large number of cellular functions, including apoptosis, signal transduction, transcription, transformation, and production of reactive oxygen species [60]. It binds directly to p53, inhibiting transcriptional transactivation by the protein, and inhibits gene transcription of p21WAF1 [82, 83]. NS5A also interacts with and partially sequesters hTAF(II)32, a component of TFIID, and an essential co-activator of p53, suppressing p53-mediated transcriptional transactivation and apoptosis during HCV infection [83]. It forms complexes with TATA binding protein (TBP) and inhibits the binding of p53 and TBP to their DNA consensus binding sequences. NS5A protein also inhibits p53TBP and p53-excision repair cross-complementing factor 3 protein–protein complex formation [84]. It also interacts with Bax as a Bcl-2 homologue, and prevents apoptosis in a p53dependant manner [85]. The protein prevents TNF-αmediated apoptotic cell death by blocking the activation of caspase-3 and inhibiting proteolytic cleavage of poly(ADPribose) polymerase [86], and activates NF-ĸB leading to antiapoptotic activity [87]. NS5A protein can modulate TFG-β signaling through interaction with TGF-β receptor [88]. The protein also downregulates the expression of the mitotic spindle protein, ASPM, through the PKR-p38 signaling pathway, and induces aberrant mitoses, chromosomal instability, and ultimately HCC [89]. Recent evidence of the reversibililty of HCV-induced fibrosis and cirrhosis with effective antiviral therapy [90, 91]
gives promise of the future ability to prevent HCC caused by HCV and perhaps even HBV.
Chronic Hepatitis B Virus-Induced Cirrhosis as a Cause of Hepatocellular Carcinoma HBV infection is one of the most prevalent viral infections in humans, and is a serious public health problem in many countries. As much as 45 % of the world's population lives in regions of high HBV endemicity [92], and approximately 400 million people worldwide are chronically infected with the virus [93] and at risk of developing cirrhosis and HCC. As many as 25 % will develop the tumor [94]. HCC is the tenth leading cause of deaths worldwide, accounting for between 520,000 and 1.2 million deaths each year [95]. The geographic distributions of HBV-induced cirrhosis and HCC are remarkably similar. As many as 60 % of HBV-induced HCCs occur in cirrhotic livers in high-incidence regions of the tumor, and higher percentages may be present in low-incidence regions. A very close association exists between chronic HBV infection and HCC in sub-Saharan Black African and Chinese patients. Of these patients with HCC, 56–90 % show markers of current infection with the virus and most of the remainder markers of past infection [10, 16, 18, 19]. In these populations, HBV infection is more common in patients with both the tumor and cirrhosis than in those with cirrhosis alone: for example, in 83 and 62 %, respectively, and in 88 and 38 %, respectively, in two independent studies of Chinese patients living in Hong Kong [10, 16]. The difference is less striking in sub-Saharan Black Africans—in 60.5 % in those with the tumor and cirrhosis and 52.8 % in those with cirrhosis alone [19]. By contrast, in low-risk populations, the incidence of HBV infection is the same in patients with cirrhosis alone and those with cirrhosis and HCC [12]. In Chinese patients, the relative risk of a male carrier of HBV developing cirrhosis is 16:1 and of developing HCC when HBV infection is combined with cirrhosis 50:1 [16]. On the other hand, in populations with a low incidence of HCC, chronic HBV infection is relatively uncommon in patients with the tumor. For example, in British patients, the infection is present in 14 % of those without cirrhosis and 37 % of those with cirrhosis [5]; in Austrian patients, in 9.2 and 19 % [15], respectively; and in Greek patients, in 26 and 67 %, respectively [96]. Nevertheless, the lower prevalences of current HBV infection is appreciably higher than those in the general populations of these countries. Sub-Saharan Black Africans with HCC and cirrhosis survive for a shorter period from the time of diagnosis than patients with cirrhosis alone (p
REVIEW ARTICLE
The Role of Cirrhosis in the Etiology of Hepatocellular Carcinoma Michael C. Kew
# Springer Science+Business Media New York 2013
Abstract Abundant evidence supports the belief of a causal relationship between cirrhosis and hepatocellular carcinoma, but one that differs between high- and low-incidence regions of the tumor. In high-incidence regions, the cirrhosis is of the macronodular variety, is typically asymptomatic, and is caused predominantly by chronic hepatitis B virus infection, whereas in low-incidence regions, the cirrhosis, although usually macronodular, may be micronodular, is commonly symptomatic and of long-standing, and is caused by chronic hepatitis C virus infection, alcohol abuse over many years, the metabolic syndrome, or hereditary hemochromatosis. In a minority of patients, hepatocellular carcinoma develops in the absence of cirrhosis, supporting a direct hepatocarcinogenic effect of some of the causal agents. Cirrhosis is the major risk factor for tumor formation in patients with chronic hepatitis C virus infection. This virus does not integrate into cellular DNA, and malignant transformation results from increased liver cell turnover induced by recurring injury and regeneration of cells in the context of persisting inflammation, oxidative DNA damage, fibrosis, cirrhosis, and changes induced by the virus at a DNA level that have yet to be fully defined. Hepatitis B virus causes malignant transformation by both direct and indirect routes. The direct route results, in part, from integration of the viral DNA into host cellular DNA; transcriptional
M. C. Kew Department of Medicine, Groote Schuur Hospital and University of Cape Town, Cape Town, South Africa M. C. Kew Department of Medicine, Johannesburg Hospital and University of the Witwatersrand, Johannesburg, South Africa M. C. Kew (*) Department of Medicine, Old Main Hospital Building, Groote Schuur Hospital, Main Road, Observatory, Cape Town, South Africa e-mail: [email protected]
activation of host growth regulatory genes by hepatitis B virus-encoded proteins; and effects on apoptosis, cell signaling, and DNA repair. The direct route may share some similarities with that of hepatitis C virus infection. The metabolic syndrome may cause malignant transformation by production of oxidative stress and the induction of a variety of mutations, including some in the p53 gene. Keywords Hepatocellular carcinoma . Cirrhosis . Hepatitis C virus . Hepatitis B virus . Metabolic syndrome
Co-existence of Cirrhosis and Hepatocellular Carcinoma Cirrhosis and hepatocellular carcinoma (HCC) often co-exist, raising the possibility that cirrhosis may play a role in the etiology and pathogenesis of the tumor. The two pathologies do not, however, have parallel geographical distributions. In most regions, HCC is uncommon [1], and cirrhosis is far more prevalent, whereas in Southeast Asia and sub-Saharan Africa, HCC occurs commonly, and may exceed cirrhosis in incidence [1]. Nevertheless, considerable evidence supports the existence of a causal relationship between cirrhosis and HCC, albeit one that differs between low- and high-risk regions of the tumor [2]. In geographic regions with low incidences of HCC, such as Europe, Oceania and North America, the great majority of patients who develop the tumor do so against a background of long-standing symptomatic cirrhosis. The cirrhosis is usually macronodular and is most often the result of long-standing alcohol abuse or, increasingly in recent years, chronic hepatitis C virus (HCV) infection [3–5] or the metabolic syndrome [6]. By contrast, in the high incidence regions of HCC in Southeast Asia and sub-Saharan Africa, chronic hepatitis B virus (HBV) infection is the major cause, and the far less common cirrhosis is almost always macronodular in type (7–12). The
J Gastrointest Canc
symptoms of cirrhosis, if any, are overshadowed by those of the tumor, and the presence of cirrhosis is typically recognized only during the diagnostic work-up for HCC or at necropsy (8–11). Further evidence, albeit indirect, for the differences in the presentation of HCC between patients in low- and highincidence regions of the tumor is provided by the observation that in low-incidence regions, the patients with co-existing HCC and cirrhosis are between 6 and 20 years (average, 8– 10 years) older than those with HCC in a noncirrhotic liver, or in those with cirrhosis alone [5, 12–15]. By contrast, Chinese [16, 17] and sub-Saharan Black African [18–21] patients with cirrhosis and high incidences of HCC, and Japanese [22] patients with cirrhosis and moderate incidences of the tumor, are the same age as those with HCC without cirrhosis, or those with cirrhosis alone. Evidence that the underlying cirrhosis in individuals in populations at low risk of HCC is at an advanced stage is provided by the observation that patients with both the tumor and cirrhosis are more likely than those with the tumor alone to be jaundiced, to have ascites, and to bleed from esophageal varices [12]. In addition, laboratory tests in these patients are more likely to show a prolonged prothrombin time, a low serum albumin concentration, and raised serum bilirubin and aspartate aminotransferase levels [12]. These clinical findings are no more common in subSaharan Black African and Chinese patients with both HCC and cirrhosis than in those with HCC alone [8, 10, 11, 18–21], and tests of hepatic function and damage are similar in Black Africans with HCC with and without cirrhosis [2, 21]. In the presence of cirrhosis, HCC is far more common in men than women (5–11:1) [5, 12, 16, 17, 21]. But when the tumor occurs in a noncirrhotic liver in high-risk regions, male predominance is less striking (4:1) [19, 20]; in low-risk regions, the sex ratio is almost parity [5, 12, 21]. In countries with a low incidence of HCC, patients in whom cirrhosis co-exists with the tumor survive for a significantly shorter time than do those with the tumor alone [12], whereas in countries with a high incidence of the tumor, the survival times in all of the patients are so short that any difference would be biologically insignificant [19, 20]. The presence of co-existing cirrhosis in populations at low risk for HCC also has an effect on the size of the tumor. For example, in an analysis of North American patients, the mean weight of the tumorous liver at necropsy was 2,530 g in the presence of cirrhosis and 3, 810 g in its absence [22]. The presence or absence of cirrhosis has a lesser effect on tumor size in populations at high risk for HCC. Thus, in sub-Saharan Black Africans with HCC and cirrhosis the mean weight of the tumorous liver was 3,692 g, compared with 4,195 g in patients without cirrhosis [23].
Type and Etiology of Cirrhosis Co-existing with Hepatocellular Carcinoma The cause and type of cirrhosis that co-exists with HCC differ appreciably between high- and low-risk regions for the tumor.
High-Risk Regions Between 60 and 85 % of Chinese and sub-Saharan Black African patients with HCC and co-existing cirrhosis show serological markers of current HBV infection, and antigenic markers for the virus can often be detected in the cirrhotic liver [20, 21, 24]. A far smaller percentage of patients with HCC in these populations is chronically infected with HCV [25]. No specific differences in the clinical manifestations of cirrhosis or HCC have been documented between patients chronically infected with HBV or HCV or with prolonged alcohol abuse. A history of alcohol abuse is uncommon in Chinese patients with HCC. In rural-dwelling sub-Saharan Black Africans, who have the highest incidences of HCC in the subcontinent, evidence of alcoholic liver disease is rarely present. A history of alcohol abuse may be obtained, but the beer consumed has a low alcohol content (3 %; the beer has a high iron content as a result of home brewing of the liquor in iron drums or pots, which leeches iron into the beer and causes iron overload of the liver, which in turn may cause HCC) [26]. In recent years a small, but increasing, number of urban Black African males with HCC have been shown to have evidence of conventional alcohol-induced liver disease [27]. Other causes of cirrhosis, such as hereditary hemochromatosis, Wilson's disease, and primary biliary cirrhosis are rarely, if ever, seen in sub-Saharan Black Africans or Chinese. The often young age of Black African [9, 11, 18] and Chinese [7, 10] patients with HCC, whether or not cirrhosis is present, suggests that the causative agent or agents of one or both diseases are operative at an early age. Certainly, HBV infection is known to be acquired at a very early age in both of these populations [28, 29].
Low-Risk Regions In populations with a low risk of HCC alcohol abuse of long standing remains an important cause of cirrhosis and HCC. In many of the remaining patients in these populations, and in increasing numbers during recent years, serological evidence of chronic HCV infection is present. Also increasing in frequency and importance as a cause of cirrhosis and HCC during this time in these populations is the metabolic syndrome [6]. Chronic HBV infection is present in most of the relatively few remaining patients.
J Gastrointest Canc
In contrast to the young ages at which exposure to environmental carcinogens occurs in regions with a high risk of HCC, the advanced ages of the patients with this tumor in regions at low risk of HCC favors exposure to the relevant etiologic agents during adult life. This is true of HCV infection and alcohol abuse, although the cause of the metabolic syndrome may be in place at a younger age. In patients with alcohol-induced HCC, the time elapsing between diagnosis of cirrhosis and the recognition of the tumor may be longer than that for the other etiologic forms of cirrhosis [14]. The frequency with which a proportion of cirrhotic patients either develop clinically recognizable HCC or are found at necropsy to have a tumor not detected during life, differs between low- and high-risk regions of the tumor, although the main determinants for this phenomenon are the type and etiology of the cirrhosis rather than the geographic location per se [2]. The risk is greatest with macronodular cirrhosis, irrespective of its etiology. In Japanese, Chinese, and Black Africans, the frequency of tumor development in a macronodular cirrhotic liver (which is almost always of viral origin) is of the order of 44–55 % [22, 30]; in low-incidence populations, it may be as high as 33 % [30]. Micronodular cirrhosis, most often as a result of alcohol abuse over many years, carries a lesser risk of tumor formation, with prevalences of 3–10% being usual [31, 32].
The Role of Cirrhosis in the Etiology and Pathogenesis of Hepatocellular Carcinoma In as many as 90 % of patients with HCC, the tumor develops in a cirrhotic liver, irrespective of the cause of the cirrhosis [33, 34, 36–45]. Although the close association between cirrhosis and HCC has long been recognized, the mechanisms by which cirrhosis contributes to malignant transformation have yet to be fully elucidated. The risk is greatest with macronodular cirrhosis. The proportion of patients with this form of cirrhosis who develop HCC ranges, in different parts of the world, from 15 to 55 %, although figures of between 40 and 55 % are usual [22, 24, 27, 33, 38, 40–42, 44]. In abusers of alcohol of long-standing, macronodular cirrhosis is likely to be present, and is complicated by HCC development in 15–24 % of patients, whereas the more common micronodular cirrhosis, present in abusers of shorter duration, is complicated in this way in only 3–10 % of patients [34–36]. The variable risk of developing HCC in the different types of cirrhosis is illustrated by studies in different population groups. For example, Becker and Chatgidakis [41] reported a rate of 44 % in southern African Black patients with HCC in whom the cirrhosis was almost exclusively macronodular, but a rate of only 6 % in Caucasians with predominantly micronodular cirrhosis. Thomson [44] reported corresponding figures of 54.5 and 2.6 % in the same population groups. The
severity of the cirrhosis also plays a role in the frequency of this complication—present in 50 % of patients with severe cirrhosis, compared with 13 % in those with moderate cirrhosis [46]. In patients in high-incidence regions of HCC in whom the symptoms of macronodular cirrhosis, if any, are overlooked, the presence of the cirrhosis is typically discovered coincidentally at the time the histological diagnosis of the tumor is made or at necropsy. One possible explanation for the close etiological association between cirrhosis and HCC is that cirrhosis itself is a premalignant condition, i.e., that hyperplasia of hepatocytes leads over time to neoplasia in the absence of additional causal factors, such as chemical or viral carcinogens. This sequence of events is supported by the observation in experimental animals that almost all forms of cirrhosis are eventually complicated by HCC [46]. But if this mechanism was universally operative, a parallel geographic distribution between cirrhosis and HCC would be expected. Although this may apply to macronodular cirrhosis, it is not true of micronodular cirrhosis. For example, in Central and Northern Europe, with high incidences of alcoholic micronodular cirrhosis, HCC is much rarer than in sub-Saharan Africa or China, where this form of cirrhosis is very uncommon [22, 32]. Moreover, in lowincidence regions of HCC, the number of patients with cirrhosis far exceeds the number with the tumor, whereas in high-incidence regions, notably in sub-Saharan Black Africans, the number of cases of HCC equals or may exceed the number with cirrhosis. Moreover, if neoplasia followed inevitably on hyperplasia, patients with cirrhosis and HCC would be expected to be older than those with cirrhosis alone. This is true in Western countries [28, 46], but not in sub-Saharan Africa or the Far East [11]. Furthermore, for neoplasia to be an inevitable complication of hyperplasia, one would anticipate that the risk of HCC development would be similar in all types of cirrhosis. In clinical practice, the risk rate of HCC development ranges from approximately 50 % with macronodular cirrhosis in subSaharan Black Africans to less than 1 % for primary biliary cirrhosis and cirrhosis in Wilson's disease, which have degrees of hyperplasia of a similar order to that in alcoholic liver disease. Moreover, the risk of malignant transformation is not uniform within one type of cirrhosis. Thus, macronodular cirrhosis in sub-Saharan Africa has a risk rate of HCC development of 44–54 %, whereas in western countries it has a rate of 15–25 %. The most important determinants of malignant transformation in the cirrhotic liver in low-incidence populations are male sex and increasing age, the latter including, at least to some extent, the duration of the cirrhosis [5, 12, 45]. Another possible mechanism of tumorigenesis in the cirrhotic liver is that the presence of cirrhosis renders the individual susceptible to a variety of environmental carcinogens [2]. It might do so by virtue of the increased hepatocyte turnover rate in cirrhosis acting as a tumor promoter, cells in
J Gastrointest Canc
mitosis being more susceptible to DNA alterations by chemicals or other agents. In addition, rapid cell turnover rates interfere with DNA repair processes, resulting in DNA alterations being transmitted to daughter cells, and are thereby fixed in the progeny. These effects are presumably independent on the original cause of the cirrhosis. HBV and HCV are carcinogenic viruses and can cause HCC in the presence or absence of cirrhosis. Although the mechanisms involved in the malignant transformation of these forms of HCC have been more fully investigated than those of the other less common causes of the tumor arising in a cirrhotic liver—alcohol abuse of long standing, the metabolic syndrome [6, 32], hemochromatosis, and hepatic fibrosis [47]—the pathogenesis of all the causes of the tumor remains incompletely understood.
Chronic Hepatitis C Virus as a Cause of Cirrhosis-Associated Hepatocellular Carcinoma Approximately 170 million people (3 % of the global population) are chronically infected with HCV, of whom about 20 % will progress to cirrhosis [48]. The development of HCC in persons chronically infected with HCV is an increasing problem in many countries, but particularly in resourcerich countries. The virus is primarily transmitted parenterally. Partly as a result of its genetic diversity, HCV can evade the host's immune response. Although HCV-induced HCC may develop in a normal liver, it is more often preceded by a spectrum of hepatic pathologies ranging from nonspecific minimal inflammatory changes to cirrhosis [48–50]. In those countries with a low incidence of HCC and in Japan, with a moderate incidence, chronic HCV infection is a leading cause of progressive hepatic fibrosis, cirrhosis, and HCC. Cirrhosis is present in approximately 70 % of all liver cancers in Japan [51]. In countries with a high incidence of chronic HCV infection, cirrhosis is the main risk factor for malignant transformation, with an annual incidence rate of HCC of between 1 and 8 % [52, 53]. The global incidence of HCC is expected to increase over the next approximately 30 years [52].
Mechanisms of Hepatitis C Virus-Induced Hepatocellular Carcinoma The major risk factor for HCC development in chronic HCV infection is the presence of cirrhosis, with as many as 70–90 % of HCCs arising in this way [48, 52–54]. In the remaining patients, HCC develops in a noncirrhotic liver, supporting a direct carcinogenic effect of HCV and indicating that cirrhosis is not a prerequisite for malignant transformation [48, 52–54].
HCV does not integrate into the host genome, and the mechanisms responsible for its hepatocarcinogenicity are therefore likely to differ, in part, from those of chronic HBV infection, in which the viral genome is frequently integrated. Malignant transformation of hepatocytes induced by HCV occurs as a result of sequences of increased liver cell turnover induced by chronic liver injury and regeneration in the context of oxidative DNA damage and chronic inflammation [48–50, 53–55]. This microenvironment facilitates the occurrence of genetic and epigenetic alterations that, over decades, may lead to the development of cirrhosis and ultimately HCC. Steatosis is one of the main factors that induce reactive oxygen species, chronic hepatic inflammation, and liver fibrosis, that may ultimately progress to HCC [54, 55]. In HCV transgenic mice that develop HCC, carcinogenesis has been shown to be preceded by development of steatosis [56]. The presence of steatosis supports a direct cytopathic component in the pathogenesis of the malignant transformation [54]. The generation of oxidative stress is a key component of the development of both cirrhosis and HCC [55–67]. In animal models, chronic HCV infection produces oxidative DNA damage, chronic hepatic inflammation, and fibrosis [58–68]. Mitochondrial dysfunction, endoplasmic reticulum stress, immune cell-mediated oxidative outbursts, in addition to DNA damage, contributes to HCV-associated oxidative stress [63]. Oxidative stress may also activate stellate cells, thereby promoting fibrosis [64], and may lead to activation of cell signaling pathways that contribute to cellular transformation [65]. One way in which increased oxidative stress may be accounted for is by NAD(P)H oxidase, especially Nox-2 in peripheral mononuclear cells and Kupffer cells in the liver [65]. HCV also triggers mitochondrial permeability transition with production of reactive oxygen species, leading to activation of NF-ĸB and STAT-3 [59, 67]. Changes induced by the core, envelope, and nonstructural proteins are central to the hepatocarcinogenic effects of HCV [62–67]. These include disrupting cellular regulatory pathways associated with proliferation and apoptosis, and suppressing host immune responses. Some of these effects could also have a bearing on the role of cirrhosis in malignant transformation and thus have an indirect effect in the development of HCC, in addition to the direct carcinogenic effects. The lack of a small animal model of chronic HCV infection has hindered an understanding of these factors. HCV core protein, by itself, has a diverse range of functions, some of which may be relevant to the hepatocarcinogenic potential of the virus [60]. It binds to and inactivates a number of tumor suppressor proteins, including p53, p73, and pRB [68]. It can also modulate the expression of the cyclin-dependent inhibitor p21WAF1, which is a major target of p53 and regulates the activities of cyclin/cyclindependent kinase complexes involved in cell cycle control and tumor formation [69]. HBV core protein also
J Gastrointest Canc
downregulates E-cadherin expression at the transcriptional level, an effect which is strongly correlated with hypermethylation of CpG islands of the E-cadherin promoter [70]. This protein also upregulates the expression of transforming growth factor (TGF)-β [71, 72]. As HCV-infected livers progress from chronic hepatitis through cirrhosis to HCC, hepatocyte pSmad3L/PAI-1 increases and pSmad3C/p21WAF1 decreases with the fibrotic stage and necroinflammatory grade. This has the effect of accelerating fibrosis and the development of cirrhosis, and increasing the risk of malignant transformation [73]. HCV core protein can also activate the Raf/MAPK signal pathway [74], upregulate Wnt-1 [75] and effect transcription of cyclin-dependent kinase inhibitor 21 gene [76]. HCV envelope protein E2 exerts an inhibitory effect on natural killer cells through engagement of CD81 [77]. E2 also activates the MAP/ERK pathway, including the downstream transcription factor ATF-2, and maintains survival and growth of target cells [78]. Of the nonstructural proteins, NS3 protein inhibits the activity of the p21WAF1 promoter in a dose-dependent manner, and acts synergistically with the core protein and p53 in an NS3 sequence-dependent manner [79]. In addition, the expression of NS3 protein enhances cell growth, JNK activation, and DNA-binding activities of the transcription factors AP-1 and ATF-2 [80], and induces TNF-α production by activation of AP-1 and NF-ĸB [81]. NS5A protein is involved in a large number of cellular functions, including apoptosis, signal transduction, transcription, transformation, and production of reactive oxygen species [60]. It binds directly to p53, inhibiting transcriptional transactivation by the protein, and inhibits gene transcription of p21WAF1 [82, 83]. NS5A also interacts with and partially sequesters hTAF(II)32, a component of TFIID, and an essential co-activator of p53, suppressing p53-mediated transcriptional transactivation and apoptosis during HCV infection [83]. It forms complexes with TATA binding protein (TBP) and inhibits the binding of p53 and TBP to their DNA consensus binding sequences. NS5A protein also inhibits p53TBP and p53-excision repair cross-complementing factor 3 protein–protein complex formation [84]. It also interacts with Bax as a Bcl-2 homologue, and prevents apoptosis in a p53dependant manner [85]. The protein prevents TNF-αmediated apoptotic cell death by blocking the activation of caspase-3 and inhibiting proteolytic cleavage of poly(ADPribose) polymerase [86], and activates NF-ĸB leading to antiapoptotic activity [87]. NS5A protein can modulate TFG-β signaling through interaction with TGF-β receptor [88]. The protein also downregulates the expression of the mitotic spindle protein, ASPM, through the PKR-p38 signaling pathway, and induces aberrant mitoses, chromosomal instability, and ultimately HCC [89]. Recent evidence of the reversibililty of HCV-induced fibrosis and cirrhosis with effective antiviral therapy [90, 91]
gives promise of the future ability to prevent HCC caused by HCV and perhaps even HBV.
Chronic Hepatitis B Virus-Induced Cirrhosis as a Cause of Hepatocellular Carcinoma HBV infection is one of the most prevalent viral infections in humans, and is a serious public health problem in many countries. As much as 45 % of the world's population lives in regions of high HBV endemicity [92], and approximately 400 million people worldwide are chronically infected with the virus [93] and at risk of developing cirrhosis and HCC. As many as 25 % will develop the tumor [94]. HCC is the tenth leading cause of deaths worldwide, accounting for between 520,000 and 1.2 million deaths each year [95]. The geographic distributions of HBV-induced cirrhosis and HCC are remarkably similar. As many as 60 % of HBV-induced HCCs occur in cirrhotic livers in high-incidence regions of the tumor, and higher percentages may be present in low-incidence regions. A very close association exists between chronic HBV infection and HCC in sub-Saharan Black African and Chinese patients. Of these patients with HCC, 56–90 % show markers of current infection with the virus and most of the remainder markers of past infection [10, 16, 18, 19]. In these populations, HBV infection is more common in patients with both the tumor and cirrhosis than in those with cirrhosis alone: for example, in 83 and 62 %, respectively, and in 88 and 38 %, respectively, in two independent studies of Chinese patients living in Hong Kong [10, 16]. The difference is less striking in sub-Saharan Black Africans—in 60.5 % in those with the tumor and cirrhosis and 52.8 % in those with cirrhosis alone [19]. By contrast, in low-risk populations, the incidence of HBV infection is the same in patients with cirrhosis alone and those with cirrhosis and HCC [12]. In Chinese patients, the relative risk of a male carrier of HBV developing cirrhosis is 16:1 and of developing HCC when HBV infection is combined with cirrhosis 50:1 [16]. On the other hand, in populations with a low incidence of HCC, chronic HBV infection is relatively uncommon in patients with the tumor. For example, in British patients, the infection is present in 14 % of those without cirrhosis and 37 % of those with cirrhosis [5]; in Austrian patients, in 9.2 and 19 % [15], respectively; and in Greek patients, in 26 and 67 %, respectively [96]. Nevertheless, the lower prevalences of current HBV infection is appreciably higher than those in the general populations of these countries. Sub-Saharan Black Africans with HCC and cirrhosis survive for a shorter period from the time of diagnosis than patients with cirrhosis alone (p
