EGF induces hepatocyte proliferation in response to liver injury, thereby facilitating liver regeneration . Its role in hepatocellular transformation has been studied both in vitro and in vivo, showing that EGF enhanced hepatocyte transformation and that EGF over-expression in the liver caused HCC [12, 13]. A functional polymorphism at position 61 has been described with G/G or A/G genotypes associated with significantly higher EGF production both in normal peripheral-blood mononuclear cell cultures  and in serum and liver tissues of HCC patients . To our knowledge, population-based, epidemiologic data examining the role of this polymorphism in HCC development were non-existent. The present study shows a positive association between EGF 61A > G polymorphism and risk of HCC among low-risk non-Asians in Los Angeles, California, especially among those who were heavy users of alcohol or who possessed a genetic Th1/Th2 profile linked to high HCC risk . In contrast, no association between EGF genotype and HCC risk was noted among the native population of Southern Guangxi, China, who exhibit one of the highest incidence of HCC in the world.
The difference in the EGF-HCC risk association between the Chinese in southern Guangxi and non-Asians in Los Angeles could be due to the different allele frequencies of EGF 61A > G polymorphism between the two study populations. Compared with non-Hispanic whites in Los Angeles, Chinese in Guangxi had a significantly lower frequency of the 61*G allele of the EGF gene (55% versus 30%).
We also conjecture that the huge difference in risk of HCC between the Chinese in Southern Guangxi (120/100,000 person-years in men) and the non-Asians of Los Angeles (4/100,000 person-years in men) may be one explanation for the seemingly disparate findings in the EGF-HCC association between the two populations. If the EGF genotype effect on HCC occurrence is independent of the combined effect from other risk factors, then the statistical power to detect a given EGF-HCC association should not be influenced by the population’s background level of HCC risk. In other words, a multiplicative model of interaction between EGF genotype and background risk factors means one would observe the same magnitude of EGF-HCC association between low-risk non-Asians and high-risk native Chinese whose background level of risk vary by a factor of 30 (120/100,000 versus 4/100,000). But suppose the interaction model is additive rather than multiplicative, then a two-fold risk between low- versus high-risk EGF profile in non-Asians (the result reported here) would translate to a relative risk of 1.07 (32/30) in Chinese of Southern Guangxi. In other words, an additive model of interaction between EGF genotype and background risk factors would predict a null finding among the high-risk Chinese.
The observed modifying effect of Th1/Th2 genotypes on the association between EGF and HCC is not surprising, given the well-established interplay of NF-κB and JAK-STAT pathways and cytokine/growth factor signaling in liver regeneration. It’s been proposed that cytokines (predominantly TNFα)-mediated NF-κB activation causes “priming” of hepatocytes to enhance their sensitivity to direct mitogens (such as EGF) in the process of liver regeneration . In addition, growth factors (including EGF) can act as alternative “rescue” activators of NF-κB in the absence of main cytokine signaling . Similarly, STAT3, an important member of the JAK-STAT pathway, is activated by both EGF and cytokines (IL-10, IL-6), and has been proposed to play a central role in viral-induced HCC , where IL-6 and EGF were shown to act in concert to promote expression of HBV viral genes . Thus, utilization of common signaling pathways (NF-κB, STAT3) by inflammatory and growth factors provides a framework for their collaboration in liver carcinogenesis .
The present study also suggests a role of the EGF 61A > G polymorphism in the development of HCC among low-risk non-Asians of Los Angeles who were heavy users of alcohol. EGF can modulate the effect of ethanol on cell proliferation and DNA synthesis [27, 29]. A recent study demonstrated that EGF-like growth factors can reduce apoptosis and enhance cell proliferation caused by exposure to alcohol . Given that reduced apoptosis and enhanced cell proliferation are hallmarks of carcinogenesis, these experimental results suggest a plausible biological mechanism for the modifying effect of the EGF genetic polymorphism on HCC development among heavy drinkers.
Heavy alcohol consumption is an important risk factor for HCC in non-Asians in Los Angeles. Forty-six percent (n = 54) of HCC patients and 18% of control subjects consumed 3 or more drinks per day among non-Asians in Los Angeles. Of the 54 HCC patients consuming 3 or more alcoholic beverages per day, 22 (41%) were free of HBV or HCV serological markers, two primary risk factors for HCC, indicating that heavy alcohol consumption plays an important role in the HCC development in low-risk non-Asians in Los Angeles, California. In contrast, only 13% (n = 32) Chinese HCC patients and 4% Chinese controls in Guangxi consumed 3 or more drinks of alcoholic beverages per day. Among the 32 HCC patients who consumed 3 or more alcohol drinks per day, 26 (81%) also tested positive for HBsAg and/or anti-HCV. In addition, dietary exposure to aflatoxin has been identified as another major risk factor for HCC in this high-risk population . Therefore, heavy alcohol intake per se plays a relatively minor role in contributing to the burden of HCC in this Chinese population.
The association between EGF 61A > G and the risk of developing HCC was initially reported in two independent cohorts of cirrhotic patients; cirrhotic patients possessing the EGF 61 G/G genotype had 2- to 4-fold increased risk of HCC compared with cirrhotic patients with the EGF 61A/A genotype . These initial findings were replicated in the Hepatitis C Antiviral Long-term Treatment against Cirrhosis (HALT-C) trial cohort; hepatitis C patients with an Ishak fibrosis stage of 3 or higher and possessing the EGF 61 G/G genotype experienced a doubling HCC risk relative to their counterparts with the EGF 61A/A genotype . Although these studies provided a strong support for a role of EGF genetic polymorphism in the progression of liver disease from fibrosis/cirrhosis to malignant stage, these data provide no information on the relationship between EGF genetic polymorphism and risk of HCC in a general population as discussed by Galmozzi and Colombo . It should be noted that the average EGF G allele frequency among the HCC patients of the two studies described above was 58% [15, 31], which was virtually identical to the percentage noted in white HCC patients in the present study (57%).
There were two recent reports of meta-analysis on the association between the EGF 61A > G genetic polymorphisms and risk of HCC in Caucasians, Chinese or mixed races [33, 34]. The latter meta-analysis  was based on data from 8 case–control studies totaling 1304 HCC cases and 2613 controls, which included all six case–control studies examined in the first meta-analysis report . This latter meta-analysis yielded a summary OR of 1.79 (95% CI = 1.39–2.29) for the EGF 61 G/G versus the A/A genotype . Significantly, this positive association between EGF 61A > G polymorphism and HCC was seen only hospital-based studies utilizing hospital controls. No association was observed among studies utilizing population-based controls. In other words, the overall null association noted in the present study is consistent with published literature.
There are inherent limitations in the present study that have been previously described [22, 23]. Briefly, both studies were of relatively small sample size. Thus, we were unable to conduct separate analyses for non-Hispanic white Americans and Hispanics/blacks, in spite of their differences in allele frequencies of EGF 61A > G polymorphism. Although race was one of the matching factors for the original study design, the imbalance in the race/ethnic distribution between HCC patients and controls of the Los Angeles study was the results of a greater proportion of black and Hispanic controls who refused to donate blood samples . In addition, the majority of eligible patients died before we were able to approach them for study participation due to the rapidly fatal nature of HCC following clinical diagnosis. Nonetheless, demographic features of eligible patients who were excluded from the study were similar to those who participated in our study. For the study in Guangxi, China, we were unable to review medical charts or obtain liver tissue slides to confirm the diagnosis of HCC and to estimate the prevalence of cirrhosis among HCC patients. If cirrhosis might modulate the influence of EGF polymorphism on HCC risk, the difference in the underlying etiologies for HCC between Chinese population in China (e.g., hepatitis B and dietary aflatoxin exposure) and non-Asians in Los Angeles (e.g., hepatitis C and alcohol abuse) could explain partially the observed differences in the EGF-HCC risk association between the two study populations since some of HCC patients with hepatitis B as the underlying cause did not show cirrhosis in the liver.