Immunohistochemical analysis of p53 protein overexpression in liver cell dysplasia and in hepatocellular carcinoma

[Epidemiology, natural history and pathogenesis of hepatocellular carcinoma]

Hepatocellular carcinoma (HCC) is the main type of primary liver cancers and the third most common cause of cancer mortality worldwide. In France, rising number between 5000 and 6000 cases are diagnosed each year. The major risk factor for hepatocellular carcinoma is chronic hepatitis: viral hepatitis B, viral hepatitis C, consumption of alcohol, hemochromatosis. Hepatocellular carcinoma is closely associated to liver cirrhosis, which is a true precancerous state. Because hepatocarcinogenesis is a long and heterogeneous process, there is still much to understand. Many genetic and epigenetic alterations are described leading to changes in cellular signalling cascades involved in regulation of growth, differentiation, apoptosis, motility. Hepatitis viruses play a direct oncogenic role through the interaction between viral and cellular proteins, which control cell homeostasis, or by the integration of hepatitis B virus genome into the host genome. Furthermore, hepatitis viruses play an indirect oncogenic role by causing chronic inflammation and hepatocyte regeneration related to viral hepatopathy. In expectation of a better understanding of hepatocarcinogenesis and new treatments, prevention from risk factors and ultrasonographic screening of patients with cirrhosis should increase prognosis.

Large liver cell change in hepatitis B virus-related liver cirrhosis

Large liver cell change (LLCC) refers to microscopic lesions often found in various chronic liver diseases; however, its nature is still controversial. Thirty-four formalin-fixed and 19 fresh frozen hepatitis B virus (HBV)-related cirrhosis samples were examined for the presence of LLCC, small liver cell change (SLCC), and hepatocellular carcinoma (HCC). The cell cycle checkpoint status (p21, p27, p16, Tp53), cell dynamics (proliferating cell nuclear antigen, Ki-67, terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling, M30), DNA damage (gamma-H2AX [H2A histone family, member X]), telomere lengths, chromosomal instability (micronuclei index), and senescence-associated beta-galactosidase (SA-beta-Gal) activity were evaluated using an in situ approach and compared to those in normal liver (n = 5) and liver with chronic cholestasis (34 cases of hepatolithiasis and three cases of primary biliary cirrhosis). In HBV-related cirrhosis, the p21, p27, and p16 cell cycle checkpoint markers were activated in normal-looking cirrhotic hepatocytes (NLCH), but diminished gradually from LLCC, SLCC, to HCC, with an increase in Tp53 expression. There was a general decrease in telomere length from NLCH, LLCC, SLCC, to HCC. Micronuclei, gamma-H2AX foci, and net cellular gain were significantly increased from normal hepatocytes, NLCH, LLCC, SLCC, to HCC. The SA-beta-Gal activity was weaker in LLCC compared to NLCH and absent in SLCC and HCC. In contrast, cholestatic LLCC showed retained expression of cell cycle checkpoint markers and decreased net cellular gain compared to adjacent normal-looking hepatocytes. HBV-related LLCC showed significantly higher Tp53 labeling index, gamma-H2AX labeling index, and micronuclei index; shorter telomere length; decreased SA-beta-Gal activity; and increased net cellular gain compared to cholestatic LLCC.

CONCLUSION

The nature of LLCC is rather heterogeneous depending on the biological setting. The characteristics of HBV-related LLCC are more consistent with dysplastic rather than merely reactive hepatocytes, whereas cholestatic LLCC more likely represents reactive change with more stringent cell cycle checkpoint control.

Large liver cell dysplasia: a controversial entity

Large cell change (LCC) is a noncommittal term used today to indicate liver cell dysplasia of the large cell type. Dysplasia was deleted from the original definition because not enough evidence has been collected over time to support premalignancy. LCC is a microscopically well-defined lesion, usually found in cirrhosis, whose origin, natural history, and fate are still debated. Different morphologic, phenotypic, molecular and clinical studies have been performed to address the issue of the dysplastic versus reactive nature of this lesion. The aim of this review is to critically evaluate the contributions to the topic and to underline that the heterogeneity of the lesion is an important issue to be taken into account for our biological understanding of it. While LCC has important morphologic analogies in experimental liver carcinogenesis, no comparable lesions are known in solid non-liver parenchymal human tissues that morphologically feature dysplasia, but in which it is uncertain whether the lesions are reactive or preneoplastic. The debate over the lesion may be useful in learning the actual limits of morphology and how additional information can be gained by looking inside the cells.

Cell biological evaluation of liver cell carcinoma, dysplasia and adenoma by tissue micro-array analysis

The clinical and morphological definition of hepatocellular carcinoma (HCC), dysplasia and adenoma suffers from a lack of biological understanding. This is especially important in the histomorphological diagnosis of nodular liver lesions in needle biopsies. Therefore, we constructed a liver tissue micro-array (TMA) and evaluated 48 HCCs, 46 dysplasias, 8 adenomas, 20 cirrhotic specimens and 28 normal liver samples derived from 68 patients. Protein (over)expression by tumor suppressor genes p16, p53 and Rb1 was assessed by immunohistochemistry, the proliferative capacity was examined by immunostaining of Ki67. Further, DNA ploidy status (hyperdiploidy) was measured by fluorescent in situ hybridization (FISH) with a chromosome 1-specific repetitive DNA probe. An abnormal chromosome 1 number, i.e. the percentage of hyperdiploid cells, was 11.0, 13.7, 16.1, 23.7 and 31.3 for normal liver samples, adenomas, cirrhosis, dysplasias and HCCs, respectively. A significant difference was found for HCC versus cirrhosis (P = 0.024) or adenoma (P = 0.033), a trend (borderline significance) was seen for dysplasia versus cirrhosis (P = 0.094). Immunohistochemical protein localisation of p53 and Rb1, as well as Ki67 indicating proliferation, was clearly higher in HCC than in cirrhosis or dysplasia (all P < 0.001). Proliferation was also higher in HCC than in adenoma (P = 0.025), whereas a trend (borderline significance) was observed for Rb1 overexpression (P = 0.063). These data suggest that in the liver cell dysplasia-carcinoma pathway, changes in ploidy are followed by increased proliferation and cell biological perturbations involving p53 and Rb1. Adenomas can be distinguished from carcinomas, but not from dysplasias, based on ploidy and proliferation characteristics.

[Epidemiology, natural history and pathogenesis of hepatocellular carcinoma]

Hepatocellular carcinoma is one of the most prevalent tumors worldwide and its incidence is increasing due to hepatitis C virus infection. Other etiologic factors are hepatitis B virus infection, alcoholic liver disease and hemochromatosis. This tumor mainly develops in cirrhotic livers that are true precancerous states. Although mechanisms of hepatocarcinogenesis remain badly known, some signaling pathways are frequently deregulated: inactivation of the p53 tumor suppressor factor in 25% of HCC, activation of the Wnt signaling and the telomerase immortalization enzyme in most of tumors. Hepatitis viruses play a direct oncogenic role by interaction between viral proteins and cellular ones, which control cell homeostasis, or by integration of hepatitis B virus genome into the host genome. Furthermore, hepatitis viruses play an indirect oncogenic role by chronic inflammation and hepatocyte regeneration related to viral hepatopathy. In a near future, a better understanding of virus-specific oncogenic mechanisms should allow us to set up innovative preventive and curative therapeutic strategies.

Expression of bile duct-type cytokeratin in hepatocellular carcinoma in patients with hepatitis C virus and prior hepatitis B virus infection

The clinicopathologic findings in patients with hepatitis C virus (HCV)-related hepatocellular carcinoma (HCC) positive for biliary markers, those related to the hepatic progenitor cells, were investigated. Cytokeratin (CK) 19 was reactive for HCCs only in patients with prior hepatitis B virus (HBV) infection. The proportions of patients with prior HBV infection and poorly differentiated HCC were significantly higher among those with CK 19-positive HCC than among those with CK 19-negative HCC. Some HCCs develop from the hepatic progenitor cells in patients with HCV infection and prior HBV infection, which may affect the clinicopathologic findings of HCV-related HCCs.

Close correlation of p53 mutation to microvascular invasion in hepatocellular carcinoma

Mutation of p53 is a poor prognostic indicator of hepatocellular carcinoma (HCC). Although poor histologic differentiation of HCC has been associated with p53 mutations, the exact reasons for unfavorable clinical outcomes in patients with HCC remain to be clarified. In this study, we evaluate the association between p53 mutation and histopathologic features of HCCs, as well as tumor recurrences and survival. We examined 20 HCCs and surrounding liver tissues from patients who underwent surgical resection, and we performed direct sequencing of p53 gene. p53 mutations were found in 9 of 20 HCCs; none were found in the surrounding liver tissue. p53 mutations were frequent in large, multinodular, and poorly differentiated HCCs. Five of 9 with p53 mutation (in contrast, none of 11 with wild-type mutation) showed microvascular invasions. Hepatocellular carcinoma recurred in 6 of 9 with p53 mutation, in contrast to only 2 of 11 with wild-type mutation. The 1-year survival rate with p53 mutation was significantly lower than that with wild-type. In conclusion, it is suggested that p53 mutations tend to be commonly associated with microvascular invasions, which may result in micrometastasis, followed by frequent recurrences.

Preliminary results of interferon-alpha therapy on woodchuck hepatitis virus-induced hepatocarcinogenesis: possible benefit in female transgenic mice

BACKGROUND

C-myc activation is a potent oncogenic event in hepatocarcinogenesis. The aim of this study was to test the preventive effect of interferon-alpha (IFN-alpha) on the development of dysplasia and subsequent hepatocellular carcinoma (HCC) in transgenic (Tg) mice overexpressing c-myc in the liver.

METHODS

The WHV/c-myc Tg mice recapitulating woodchuck hepatitis virus-induced hepatocarcinogenesis were treated with IFN-alpha, starting early in life until sacrifice at pre-neoplastic or neoplastic stages. Transgene expression was assessed by reverse transcription-polymerase chain reaction (RT-PCR), hepatocyte proliferation was assessed by bromodeoxyuridine incorporation and RT-PCR for proliferating cell nuclear antigen, and apoptosis was assessed by in situ nick-end-labeling of DNA.

RESULTS

C-myc expression and hepatocyte proliferation were significantly reduced in treated female mice, without modification of apoptosis, correlating with a lower severity of dysplasia in 9/12 treated animals at pre-neoplastic stages. At the neoplastic stage, 2/3 treated females neither exhibited carcinoma nor dysplasia, while all 6/6 untreated mice and 3/3 treated males developed carcinomas.

CONCLUSIONS

Inhibition of c-myc and hepatocyte proliferation by long-term administration of IFN-alpha was associated with a decrease, or a delay, of oncogenesis in the mouse Tg HCC model. Whether c-myc and hepatocyte proliferation down-regulation could be relevant parameters of IFN-alpha efficiency for hepatocarcinogenesis prevention in cirrhotic patients should be established.

The immunohistochemical phenotype of dysplastic foci in human liver: correlation with putative progenitor cells

BACKGROUND/AIMS

In previous studies we found strong evidence for the existence and activation in human liver of putative progenitor cells resembling oval cells in rat liver. In view of the known role of rat oval cells in regeneration and hepatocarcinogenesis, we investigated a possible correlation between human putative progenitor cells and different types of dysplastic foci.

METHODS

We determined the immunohistochemical phenotype of dysplastic foci found in 20 cirrhotic liver explants of various etiology, using specific antibodies against hepatocyte-type cytokeratin (CK) 8 and CK18, bile duct-type CK7 and CK19, chromogranin-A (chrom-A) and rat oval cell marker OV-6.

RESULTS

All 12 foci of large cell dysplasia had a phenotype similar to that of surrounding parenchyma. Oncocytic foci showed a strong cytoplasmic staining for CK7. Three out of six of these foci contained "progenitor cells", which are small cells immunoreactive for CK18, CK7, CK19, OV-6, chrom-A and stained more intensely for CK8 than surrounding hepatocytes. Four out of eight glycogen-storing foci contained CK7-positive intermediate hepatocyte-like cells and "progenitor cells". Sixteen out of 29 small cell dysplastic foci consisted of "progenitor cells" and intermediate hepatocyte-like cells which were immunoreactive for CK7, CK18, OV-6, chrom-A and showed a stronger cytoplasmic positivity for CK8 than surrounding hepatocytes.

CONCLUSIONS

Foci of large cell dysplasia show no correlation with putative progenitor cells. Half of the oncocytic and glycogen-storing foci contain "progenitor cells", while more than half of the foci of small cell dysplasia consist of small cells with the same immunohistochemical phenotype as putative progenitor cells and intermediate hepatocyte-like cells, suggesting that differentiating putative progenitor cells can give rise to foci of small cell dysplasia.

Overexpression of p53 protein is not directly related to hepatitis B x protein expression and is associated with neoplastic progression in hepatocellular carcinomas rather than hepatic preneoplasia

p53 mutations and binding of p53 to hepatitis B virus (HBV) x protein (HBx) have been suggested as alternative mechanisms of development of hepatocellular carcinomas (HCCs) in man, both processes resulting in intracellular accumulation of the protein which is detectable by immunohistochemical approaches. We have examined p53 expression in 149 explanted human livers, including 39 cases infected with HBV and 35 bearing HCC. p53 was demonstrated immunohistochemically in 51% of HCC samples (18/35), localized mainly in fast growing poorly differentiated areas. Accumulation of mutant p53 was verified by immunoprecipitation in most of the positive HCC samples (14/15), implying occurrence of p53 mutations. No cells positive for p53 were found in 354 preneoplastic hepatocellular lesions examined. This indicates that p53 mutation is associated with progression, rather than early development, of HCC in the low-aflatoxin B(1)-exposed region. The intracellular distribution patterns of p53 and HBx were different, with the former within nuclei and the latter confined to cytoplasmic compartment. HBx did not coimmunoprecipitate with p53. These data indicate that p53-HBx binding is infrequent, if it really occurs, in HBV-infected human liver, and that it cannot be a common mechanism of HBV-associated hepatocarcinogenesis. In addition, p53 accumulation was also observed in some parenchymal and ductular (oval) cells in cirrhotic livers and, more frequently, in fulminant hepatitis, being independent of HBx expression, and seemingly associated with the damage and/or regeneration of liver parenchyma, perhaps merely reflecting a cellular stress response.

Consilience across evolving dysplasias affecting myeloid, cervical, esophageal, gastric and liver cells: common themes and emerging patterns

In the present paper, an attempt is made to identify common biologic themes across dysplastic states affecting the marrow, gastro intestinal tissue, the cervix and liver as well as unifying patterns during disease evolution. The following algorithm appears generally applicable, although individual variations must necessarily be anticipated. It appears that there is an initial transforming event which in all dysplasias except that affecting the marrow has been found to be infectious. Increased cellular proliferation-increased apoptosis, telomere shortening, appearance of telomerase expression and clonal expansion follow the initial insult. Abnormalities in the cytokine environment are universally described and it is likely that the quintessential monoclonality aspect of dysplasia predisposes to accumulation of genetic mutations, and microsatellite instability leading to the appearance of evolved sub-clones. The conversion of a dysplastic phenotype to a malignant one reflects the success of one such sub-clone in developing a survival advantage over a large population of prematurely apoptotic neighbors. This state is usually acquired by silencing tumor suppressor genes through hypermethylation or actual loss or dysfunction. Thus, excessive apoptosis of cells resulting from a persistent infectious process predisposes the organ towards developing a cancerous phenotype. Evidence for the shared pathology is presented at length with the hope that these parallels between dysplastic states will be helpful in both biologic and therapeutic research.

Hepatocellular proliferation and development of hepatocellular carcinoma: a case-control study in chronic hepatitis C

Patients with hepatitis C have an increased risk of developing hepatocellular carcinoma (HCC). This is related to the stage of chronic liver disease, as characterized histologically by hepatic fibrosis and architectural distortion, but it is unclear whether histological markers can define the risk of developing HCC. We conducted a case-control immunohistochemical study of Ki-67, a marker for hepatocellular proliferation, in livers of 18 patients who had developed HCC more than 2 years after the biopsy specimen had been taken. Using conditional logistic regression analysis, the results were compared with 18 selected controls, who were age-matched patients with hepatitis C of similar histological stage who had not developed HCC. We also examined livers for cellular dysplasia, p53 mutations, and bcl-2 overexpression, and assessed whether the results could be correlated with demographic and disease-related variables, such as gender, region of birth, alcohol consumption, severity of liver disease, HCV genotype, and markers of hepatitis B virus (HBV) infection. Livers from patients who developed HCC were more often positive for Ki-67 (13 of 18 [72%] v 9 of 18 [50%]; P = .06) and tended to have higher mean Ki-67 scores (6 +/- 7.5 v 3 +/- 4.4; P = .10) compared with control cases. In the HCC-predisposed group, three livers showed large cell dysplasia, two were positive for p53 mutations, and two for bcl-2 overexpression. In contrast, in the non-HCC group, only one case had dysplasia, and none were positive for immunostaining for p53 or bcl-2 mutations. With the exception of one case, all livers with large cell dysplasia or p53 mutations and bcl-2 overexpression were also positive for Ki-67. Twelve (55%) of the 22 Ki-67-positive cases were anti-HBc-positive in the serum, in contrast to 2 of 14 (14%) patients in the Ki-67-negative group (P = .01). Patients with evidence of past infection with HBV were more often Ki-67 positive than those who had no evidence of past infection (85% [11 of 13] v 45% [10 of 22]; P = .02). There were no other associations between demographic or disease-related variables and Ki-67 expression. Increased hepatocellular proliferative activity, as assessed by Ki-67 expression, may be one factor indicative of an increased risk of developing HCC among patients with chronic hepatitis C. Furthermore, past infection with HBV appears to be an important correlate of increased hepatocellular proliferation in hepatitis C.

Prevalence of liver cell dysplasia and association with HCC in a series of 100 cirrhotic liver explants

BACKGROUND/AIMS

Liver cell dysplasia of large (LLCD) and small (SLCD) cell types may represent a premalignant change. We sought to evaluate their prevalence, relationship with the gross type of cirrhosis, aetiology of liver disease, and the presence of hepatocellular carcinoma in a series of cirrhotic livers.

METHODS

The presence and pattern of SLCD and LLCD were evaluated by careful histological analysis in 100 consecutive cirrhotic livers of viral (49%) or non viral (51%) aetiology, and with or without hepatocellular carcinoma. Prevalences were compared using Chi-square or Fisher's tests; relative risk for hepatocellular carcinoma was evaluated by the odds ratio.

RESULTS

Dysplasia was found in 82/100 of livers. Eighty-one had LLCD, with (n=49) or without (n=32) associated SLCD. SLCD alone was found in only one case. LLCD and SLCD tended to be more frequent and extensive in mixed or macronodular cirrhosis than in micronodular cirrhosis. LLCD was significantly more frequent and extensive in cirrhosis due to hepatitis B, as was SLCD in cirrhosis due to hepatitis B virus or biliary diseases, where it showed a different pattern (focal vs diffuse, respectively). LLCD and SLCD were both significantly associated with the presence of hepatocellular carcinoma, even of small size. Small foci of SLCD and widespread LLCD were the two conditions which showed the strongest association with hepatocellular carcinoma, with odds ratios of: 6.33 and 3.88, respectively. Widespread SLCD was not relevant for hepatocellular carcinoma in biliary diseases.

CONCLUSIONS

Liver cell dysplasia may be considered an additional risk factor for hepatocellular carcinoma in patients with cirrhosis and should be looked for in biopsies. Widespread LLCD and SLCD with a focal pattern are particularly relevant for hepatocellular carcinoma, whereas widespread small cell changes found in biliary diseases seem to have a different biological significance.

p53 expression in hepatocellular carcinoma in Greece. Correlation with epidemiological and histopathological data

Localization of p53 oncoprotein was investigated in 60 hepatocellular carcinomas (HCCs) from patients resident in the Northwest and Central Greece. The streptavidin-biotin complex immunoperoxidase method was performed in archival formalin-fixed, paraffin-embedded material, using the monoclonal antibody DO-1. The aim of our study was to correlate p53 expression with histological and epidemiological data. p53 overexpression in patients with serological hepatitis B or C was greater (47%) as compared to that observed in patients without these markers (p < 0.01). Morphologically normal liver tissue (NLT) and liver cell dysplasia (LCD) was recognized adjacent to HCC developing on non-alcoholic cirrhotic livers in patients with "NonA, NonB hepatitis" from between 1975-1986. NLT and LCD and p53 oncoprotein was expressed in 10% of the cases. No relationship was observed between p53 expression and tumor histological grade, patients' age and sex. These results suggest that in Northwest and Central Greece, p53 oncosupressor gene may be involved in some HCCs; it may be associated with viral chronic infection disease (HBV or HCV), and as yet with uncharacterized viruses which remain to be determined.

Prognostic significance of p53 nuclear and cytoplasmic overexpression in right and left colorectal adenocarcinomas

The prognostic significance of nuclear and cytoplasmic p53 protein, detected immunocytochemically using CM1 and PAb 1801 antibodies, was evaluated in right-sided and left-sided colorectal adenocarcinomas from 293 patients. CM1 nuclear and cytoplasmic p53 accumulation occurred in 38 and 25% of cases, respectively. PAb 1801 nuclear staining occurred in 18%, with no cytoplasmic staining. CM1 expression either in the nucleus or in the cytoplasm was positively related to PAb 1801 expression (P < 0.001 and P = 0.009, respectively). The incidence of CM1 nuclear and cytoplasmic expression was more frequent in right-sided tumours (P = 0.023 and P = 0.034, respectively), while PAb 1801 nuclear staining was more common in left-sided tumours (P = 0.011). In survival analyses, CM1 nuclear overexpression in the right-sided tumours (P = 0.016) and CM1 cytoplasmic overexpression in left-sided tumours (P = 0.04) were prognostic indicators, independent of Dukes' stage, DNA ploidy, PAb 1801 expression and each other. Further analysis showed that the prognostic value of CM1 nuclear expression was greater in right-sided tumours than in left-sided tumours (P = 0.018). The nuclear and cytoplasmic p53 protein detected with CM1 and PAb 1801 may play different roles in tumour progression and provide prognostic indicators for right- and left-sided colorectal tumours.

Immunohistochemical expression of p53 protein and proliferating cell nuclear antigen in hepatocellular carcinoma

There is an increased prevalence of p53 mutations in hepatocellular carcinomas (HCCs). A total of 62 HCC samples with adjacent liver tissue were analyzed immunohistochemically for the presence of p53 by two different commercial sources of Pab 1801. Polyclonal antibodies anti-HbsAg and anti-HbcAg were employed for the detection of HBV in the adjacent tissue and PC-10 for the detection of proliferating cell nuclear antigen (PCNA). Positive staining for p53 was identified in 42% and 55% of the HCC cases using each monoclonal antibody. p53 was found in 42% of the low grade and 43% of the high grade HCC. In 32% of the HCC cases, p53 was found in the adjacent liver tissue. In 52.6% of the cases with evidence of HBV infection, p53 positive expression was observed. PCNA was detected in 56% of the HCC cases (69% low grade, 57% high grade HCC). Eighty-one percent of the p53 positive tumours expressed PCNA, mostly with a high index. p53 and PCNA were not related to histologic grade. A trend for positive correlation was observed between p53 expression and HBV infection. The detection of p53 in non neoplastic tissue and the absence of a significant correlation between p53 expression and degree of differentiation support the hypothesis that the p53 gene mutation is involved in early stages of hepatocellular carcinogenesis.