No abstract

A patient with primary gastric adenocarcinoma with extremely high serum alpha-fetoprotein (AFP) levels (12,000 ng/ml) is described. Histologically, foci strongly resembling hepatocellular carcinoma with hyaline globules were noted. Within tumor cells, AFP was identified with both light and electron microscopy, showing the production of AFP by tumor cells themselves. Furthermore, 88% of serum AFP combined with Concanavalin A (ConA), revealing that it was hepatic-type AFP and not germ-cell-type. Localization of alpha-1-antitrypsin within tumor cells was also noted. Ultrastructural study showed that there were two types of structures corresponding to periodic acid-Schiff (PAS)-positive globules, one of which, the proteinaceous material in intracytoplasmic lumina, was found to contain AFP. Among gastric adenocarcinomas with a high serum AFP level (several thousand or more ng/ml of AFP), foci resembling hepatocellular carcinomas have been reported by several investigators. Those gastric carcinomas, together with the current case, may constitute a clinicopathologic entity, hepatoid adenocarcinoma of the stomach.

Hepatoid adenocarcinomas of the stomach are gastric carcinomas with both adenocarcinomatous and hepatocellular differentiations. They usually produce large amounts of alpha-fetoprotein (AFP) with a Concanavalin A-binding property of hepatic type. In this study, these carcinomas occurred in older persons, with the antrum being a common site. Observed grossly, growth of the tumors was nodular and massive. Prognosis was poor because of frequent liver metastases. In the cytoplasms of tumor cells, various serum proteins were identified, including AFP, alpha-1 antitrypsin (AAT), alpha-1 antichymotrypsin (ACT), albumin, and prealbumin. Localizations of ferritin, prothrombin, and transferrin were demonstrated with less frequency. Adenocarcinomatous foci were composed of well-differentiated, intestinal-type epithelial cells and often contained carcinoembryonic antigen. These adenocarcinomatous and hepatoid areas were often intermingled with each other. There were extensive venous involvements by tumor cells. The poor prognosis of the tumors may be attributed to these involvements as well as to production of AFP and presence of AAT/ACT, which have immunosuppressive and protease-inhibitory properties, respectively.

Pancreatic cancer still remains a serious health problem with <5% 5-year survival rate for all stages. To develop an effective treatment, it is necessary to identify a target molecule that is crucially involved in pancreatic tumor growth. We previously observed that Pim-3, a member of the proto-oncogene Pim family that expresses serine/threonine kinase activity, was aberrantly expressed in human and mouse hepatomas but not in normal liver. Here, we show that Pim-3 is also expressed in malignant lesions of the pancreas but not in normal pancreatic tissue. Moreover, Pim-3 mRNA and protein were constitutively expressed in all human pancreatic cancer cell lines that we examined and colocalized with the proapoptotic protein Bad. The ablation of endogenous Pim-3 by small hairpin RNA transfection promoted apoptosis, as evidenced by increases in a proportion of cells in the sub-G 1 fraction of the cell cycle and in phosphatidyl serine externalization. A proapoptotic molecule, Bad, was phosphorylated constitutively at Ser 112 but not Ser 136 in human pancreatic cancer cell lines and this phosphorylation is presumed to represent its inactive form. Phosphorylation of Bad and the expression of an antiapoptotic molecule, Bcl-X L , were reduced by the ablation of endogenous Pim-3. Thus, we provide the first evidence that Pim-3 can inactivate Bad and maintain the expression of Bcl-X L and thus prevent apoptosis of human pancreatic cancer cells. This may contribute to the net increase in tumor volume or tumor growth in pancreatic cancer. (Cancer Res 2006; 66(13): 6741-7)

Five cases of ovarian carcinoma with hepatoid features, three of them primary and two probably primary, are described. The tumor cells were arranged predominantly in sheets and contained moderate to abundant amounts of eosinophilic cytoplasm; varying numbers of tumor cells stained immunohistochemically for alpha fetoprotein. In contrast to the much younger age range of patients with ovarian hepatoid yolk sac tumors, the ages of the five patients with hepatoid carcinomas ranged from 42 to 78 (average, 63 years), and none of them had gonadal dysgenesis or recognizable germ cell components within their tumors. All the tumors presented as adnexal masses; in four cases they were Stage III and in one case, Stage IIB; this last tumor spread to the upper abdomen within 4 years. In three cases the clinical course and pathology findings indicated that the tumor had originated in the ovary; in the remaining two cases the tumors were interpreted as only probably primary in the ovary. Hepatoid carcinomas must be distinguished from other ovarian neoplasms, especially from hepatoid yolk sac tumors.