Alan C. Paterson

Cytokeratin expression in hepatocellular carcinoma: an immunohistochemical study

Normal human hepatocytes express cytokeratins no. 8 and 18, whereas bile duct cells contain the same cytokeratins and, in addition, cytokeratins no. 7 and 19. This cytokeratin pattern is believed to be preserved during neoplastic transformation. Thirty-four cases of hepatocellular carcinoma (11 well differentiated, 16 moderately differentiated, 7 poorly differentiated) were studied on frozen sections using monoclonal antisera directed against individual cytokeratins no. 7, 8, 18, and 19 in an immunoperoxidase procedure. In 17 of 34 cases, tumor cells showed only reactivity with monoclonals anticytokeratin no. 8 and 18. However, 17 of 34 cases showed an aberrant pattern in that a variable number of tumor cells were stained with anticytokeratins no. 7 and/or 19 in addition to no. 8 and 18. Only three of 11 well-differentiated cases displayed an unexpected cytokeratin pattern, whereas an aberrant pattern was present in all seven of seven poorly differentiated cases. These results are in conflict with previously published data obtained by two-dimensional gel electrophoresis and immunohistochemistry. They indicate that the cytokeratin pattern might not always be preserved during neoplastic transformation. The implication of this finding for the differential diagnosis of metastatic gastrointestinal carcinomas is discussed.

Recombinant leucocyte interferon treatment of chronic hepatitis B: An analysis of two therapeutic trials

We have investigated the efficacy of recombinant alpha-interferon treatment of chronic hepatitis B virus (HBV) infection in two therapeutic trials. Forty-four patients positive for HBsAg, HBeAg, DNA polymerase and HBV-DNA were studied. Fourteen carriers were treated in the first trial with doses ranging from 18 to 50 million units (mu)/m2 3 times per week. Six of 14 treated carriers (43%) have a sustained loss of HBeAg, HBV-DNA and DNA polymerase. Four lost HBsAg (29%). Two of 11 (18%) untreated carriers lost HBeAg, but none lost HBsAg (P = 0.05). Nineteen patients were entered in a second trial to assess dose response. Fourteen were treated with doses ranging from 2.5 to 10 mu/m2. Five patients were untreated. Two treated patients seroconverted to anti-HBe, and a third cleared HBsAg and seroconverted to anti-HBs. None of the controls was anti-HBe-positive. Thus 9/28 (32%) carriers have lost replicating HBV versus 2/16 (13%) of untreated patients. Elevated pretreatment serum ALT concentrations and severe chronic active hepatitis were associated with inhibition of viral replication in treated patients suggesting that seroconversion may require an appropriate host response. The efficacy of recombinant interferon is restricted, but it may be of benefit in a proportion of carriers.

Transferrin receptor expression in human hepatocellular carcinoma: An immunohistochemical study of 34 cases

Using a panel of five monoclonal anti‐transferrin receptor antibodies, we investigated the transferrin receptor expression in 34 human hepatocellular carcinomas of Belgian (n=6), Italian (n=7) and South African (n=21) origin. For comparison the tumours were also stained with the monoclonal antibody BK 19.9, recognizing an antigen biochemically similar to the transferrin receptor, and with a monoclonal antibody against the epidermal growth factor receptor. Hepatocellular carcinomas express large amounts of transferrin receptors as demonstrated by the intense transferrin receptor immunostaining in 33/34 cases. Differences in staining pattern between and within the tumours were not related to the degree of tumour differentiation, nor to the origin or race of the patient. In 15 cases which included non‐tumoural tissue, the tumour was more intensely stained than the surrounding liver parenchyma. The BK 19.9 immunoreactivity was generally weaker and mainly involved stromal cells, except in three cases where an intense staining of the tumour cells was seen. The epidermal growth factor receptor staining was also weaker and only in four cases was the immunoreactivity of the tumour stronger than the surrounding parenchyma. Demonstration of the transferrin receptor may be useful for the detection of malignant foci in liver biopsies. This may be of particular interest in the histological investigation of minute hepatocellular carcinomas.

HLA expression in human hepatocellular carcinoma.

This study examines the expression of MHC class I and II antigens and their related invariant chains in 70 cases of human hepatocellular carcinoma (HCC), using monoclonal (Mabs) and polyclonal antibodies. In comparison with normal hepatocytes, the majority (94.3%) of HCCs show enhancement or acquisition of HLA-A, B, C in either a cytoplasmic or membranous distribution, with staining being uniformly distributed throughout the specimen. HLA-A, B, C was accompanied by beta 2-microglobulin expression in all but two cases. Although 44.9% of specimens showed HLA-DR expression, positively staining tumour cells were often sparse and heterogeneously distributed. By contrast, the invariant (I) chain, present in 47.1% of cases, was frequently intensively stained and extensive in distribution. HLA-DR staining was usually cytoplasmic although two cases showed faint membranous enhancement. In addition to HLA-DR and I-chain, two cases also showed HLA-DQ staining. Display of MHC antigens was not related to tumour differentiation or size of the lesion (resected vs. advanced tumours). It is possible that the acquisition of class I antigens by the majority of HCCs may influence tumour behaviour.

Non-transferrin-bound iron and hepatic dysfunction in African dietary iron overload.

Background: Circulating iron is normally bound to transferrin. Non‐transferrin‐bound iron (NTBI) has been described in most forms of iron overload, but has not been studied in African dietary iron overload. This abnormal iron fraction is probably toxic, but this has not been demonstrated.

Iron-free neoplastic nodules and hepatocellular carcinoma without cirrhosis in Wistar rats fed a diet high in iron

Although excess hepatic iron in hereditary haemochromatosis and dietary iron overload in the African causes hepatocellular carcinoma, it usually does so in the presence of cirrhosis. A direct hepatocarcinogenic effect of iron has not been proved. Moreover, an animal model of hepatocellular carcinoma induced by iron overload has not been available. The aim of this study was to develop such a model and to use it to ascertain whether excess hepatic iron is directly hepatocarcinogenic. Sixty Wistar albino rats were fed a chow diet and 60 the same diet supplemented initially with 2% carbonyl iron for 12 months, followed by 0.5% ferrocene for 20 months. Five rats from each group were sacrificed every 4 months for 24 months for histological and biochemical monitoring. By 16 months, hepatocytes in all the rats receiving the iron‐supplemented diet showed grade 4 iron overload, comparable in degree with that seen in patients with advanced hereditary haemochromatosis and dietary iron overload. Altered hepatic foci and pre‐neoplastic nodules were first seen at 16 months. These increased in size and number with time, were iron‐free, stained positively with placental glutathione sulphydryl transferase, and showed the same histological features as the iron‐free foci described in patients with hepatocellular carcinoma complicating hereditary haemochromatosis. At 32 months the eight surviving rats in the iron overloaded group were sacrificed. The livers of five of these rats contained pre‐neoplastic nodules and one showed, in addition, an iron‐free, well‐differentiated hepatocellular carcinoma. The tumour stained positively for placental glutathione sulphydryl transferase. Neither cirrhosis nor portal fibrosis was present in this or any iron‐loaded animal. We conclude that hepatocellular carcinoma may complicate dietary hepatic iron overload in Wistar albino rats in the absence of fibrosis or cirrhosis, confirming an aetiological association between dietary iron overload and the tumour and suggesting that iron may be directly hepatocarcinogenic. Copyright

Smoking as a risk factor in hepatocellular carcinoma a case–control study in southern african black

Two hundred forty southern African black patients with hepatocellular carcinoma and control subjects matched for race, sex, age, and urban or rural background were questioned about their smoking habits. Patients with hepatocellular carcinoma were not more likely to smoke or to smoke heavily than the control subjects. This was also true of the subgroups: men and women, and urban and rural background. There was a slightly increased relative risk associated with smoking in all patients who showed no serum markers of current or past hepatitis‐B virus infection and in patients older than 50 years who did not have markers of current or past hepatitis‐B virus infection. However, this was not statistically significant, and was not supported by a linear trend, the risk in heavy smokers being less than 1. Rural black patients, who have a higher incidence of hepatocellular carcinoma than urban black patients, smoked less than their urban counterparts. We conclude that smoking is not an unqualified risk factor for hepatocellular carcinoma in southern African black patients. There may, however, be a trend toward smoking playing an etiologic role in patients without hepatitis‐B virus infection, especially in older patients.

Esophagitis in a population at risk for esophageal carcinoma

A previously unreported high incidence of endoscopically identifiable, diffuse, largely nonulcerative esophagitis, is described in a selected population of Southern African blacks at risk of cancer of the esophagus. This is comparable with findings in similar groups in Iran and China. The incidence of dysplasia and “early” cancer is still undefined in South Africa.

Susceptibility of chacma baboons (Papio ursinus orientalis) to infection by hepatitis B virus.

BACKGROUND Because baboons are being considered as a source of xenografts for human liver transplantation in patients with hepatitis B virus- (HBV) induced cirrhosis to forestall infection of the graft by the virus, we undertook a study to ascertain if baboons are resistant to HBV infection. METHODS Six chacma baboons were inoculated with serum containing HBV and were followed for 52 weeks to detect transmission of infection. RESULTS Anti-HBc was detected in the serum of four baboons 16 weeks after inoculation. Virions, small spherical particles, and tubular forms were seen at this time in the serum of the one baboon studied by transmission electron microscopy. HBV DNA was detected by polymerase chain reaction in the serum of the same four baboons throughout the period of follow-up, as well as in liver tissue obtained after 52 weeks. The specificity of the DNA was confirmed by Southern hybridization. Nucleotide sequences showed complete sequence identity between the HBV DNA in each of the baboon sera and one of the two HBV genotypes inoculated. Serum transaminase levels tested at 4-weekly intervals were always normal and histological examination of liver tissue after 52 weeks showed no evidence of chronic hepatitis. Examination of squash preparations of liver tissue by electron microscopy in one baboon revealed core-like particles. CONCLUSIONS Chacma baboons are susceptible to HBV infection and appear to develop a chronic carrier state. The use of xenografts from baboons should preferably be avoided, but if they are used again for HBV-infected patients it would be prudent to treat the patients as if they had received an organ from a human donor.

Colorectal cancer in South Africa: A heritable cause suspected in many young black patients

BACKGROUND Colorectal carcinoma (CRC) has a low incidence among the black African population. Largely unrecognised in the scientific literature is the fact that a disproportionately large number of young black patients (<50 years old) present with CRC. OBJECTIVES To analyse those tumours, which we propose may link them to morphological features associated with known genetic pathways. METHODS A retrospective review of South African patients histologically diagnosed as having CRC by the Division of Anatomical Pathology, National Health Laboratory Service (NHLS) and the University of the Witwatersrand (1732 patients from 1990 to 2003). The histology was fully reviewed in 609 patients (1997-2002), and all specimens from patients <50 years of age were subjected to immunohistochemistry tests for mismatch repair proteins, as well as APC and p53 proteins. RESULTS Most young patients (<50 years) were black (41% v. 10% white; p < or = 0.001). Blacks had predominantly proximal tumours and significantly more poorly differentiated and/or mucinous tumours (p = 0.006), and loss of mismatch repair protein expression was more evident than in whites. CONCLUSIONS It seems likely that CRC in young blacks develops through the accumulation of mutations, most probably via mismatch repair deficiency or promoter methylation, which in turn is linked to poor differentiation and a mucinous architecture.