Survivin, a new member of the family of apoptosis inhibitors, is expressed almost exclusively in proliferating cells, above all in cancers. Subcellular localisation and prognostic implications of the survivin protein have not yet been determined in oesophageal squamous cell carcinoma. The survival of 84 patients with oesophageal squamous cell carcinomas was correlated with the extent of immunohistochemical survivin expression in tumour cell nuclei. Tumours were scored positive when 45% cells stained positive. Patients were followed up for at least 5 years or until death. In normal oesophageal squamous cell epithelium, some cytoplasmic survivin expression was detected in the basal cells, whereas proliferating cells showed nuclear staining of survivin. Nuclear expression of survivin was also detected in 67 cancers (80%). The mean survival for patients of this group (28 months, range 20 -36) was significantly less than that for patients without survivin expression in the tumour cell nuclei (108 months, range 62 -154, P ¼ 0.003). Using univariate analysis, nuclear survivin expression (P ¼ 0.003), tumour depth (P ¼ 0.001), lymph node metastasis (P ¼ 0.003) and stage (Po0.001) were the best predictors of survival. In contrast, cytoplasmic survivin staining was noted in 53 (63%) tumours and had no prognostic relevance. In conclusion, the analysis of nuclear survivin expression identifies subgroups in oesophageal squamous cell cancer with favourable (survivin À ) or with poor prognosis (survivin + ). We suggest that the determination of nuclear survivin expression could be used to individualise therapeutic strategies in oesophageal squamous cell cancer in the future.

The frequency and prognostic significance of neuroendocrine marker expression in undifferentiated colorectal cancers has not yet been studied in great detail. Therefore, the survival of 20 patients with small cell undifferentiated colorectal cancers, treated at our institution between 1982 and 1997 (0.8% of all operated colorectal carcinomas), was correlated with the extent of neuroendocrine differentiation. Chromogranin A, synaptophysin, syntaxin1, VAMP2, SNAP25 and alpha/beta-SNAP were used as neuroendocrine markers. Based on the degree of immunoreactivity for these marker proteins, tumors were separated into group 0 (<2% cells stained positive for neuroendocrine markers) and group 1 (>2% cells stained positive). Patients were followed up for at least 5 years or until death. Of 20 (45%) undifferentiated colorectal tumors, 9 expressed neuroendocrine markers (group 1). Only one patient of this group survived for 2 years (11%), whereas the 2-year-survival rate was 45.4% in group 0. Of the 11 patients in group 0, 9 were diagnosed with UICC stages I-III, whereas 8 of 9 tumors with expression of neuroendocrine markers were diagnosed with UICC stage IV ( P=0.002). Our results show that neuroendocrine differentiation is often seen in small cell undifferentiated colorectal cancer. It correlates with a more aggressive course of the disease.

Telomerase activity (TA) has been shown to correlate with poor clinical outcome in various tumour entities, indicating that tumours expressing this enzyme may be more aggressive and that TA may be a useful prognostic marker. For breast cancer, however, TA is a controversial prognostic marker; whereas some studies suggest an association between TA and disease outcome, others do not find this association. This study used tissue microarrays (breast carcinoma prognosis arrays) containing 611 samples (each 0.6 mm in diameter) from the tumour centre of paraffin-embedded breast carcinomas to analyse the catalytic subunit of telomerase, human telomerase reverse-transcriptase (hTERT), and the internal RNA component (hTR), which are the core components of the telomerase holoenzyme complex. hTERT protein expression was obtained by immunohistochemistry (human anti-telomerase antibody Ab-2, Calbiochem), and hTR RNA was measured by radioactive in situ hybridization. hTERT and hTR expression were determined semi-quantitatively and graded (scores 1-4). Clinical data, such as histological subtype, pT stage, tumour diameter, pN stage, BRE grade, tumour-specific survival (in months), patient's age and others, were available for statistical analysis. A statistically significant correlation was found between tumour-specific survival (overall survival) and hTERT expression (p < 0.0001) or hTR expression (p = 0.00110). Tumours with higher scores (scores 3, 4) for hTR and/or hTERT were associated with a worse prognosis. In multivariate analysis, hTERT expression was an independent prognostic factor. Previous studies, focusing on analysis of TA in smaller numbers of fresh-frozen breast carcinomas by the TRAP assay, gave controversial results with respect to TA as a prognostic marker. Using tissue microarrays from 611 breast carcinomas, this study has demonstrated that increased expression levels of the telomerase core components, hTERT and hTR, are associated with lower overall survival. These findings suggest that TA should be included in future validation studies as a prognostic marker in breast cancer.

Gastroenteropancreatic neuroendocrine tumors represent a heterogeneous tumor entity. The growth pattern ranges from very slowly to fast growing, aggressive types of tumors. Survivin, a member of the family of apoptosis inhibitors, is a bifunctional protein that suppresses apoptosis and regulates cell division. In this study we determined the prognostic value of survivin in this tumor entity. Tumor specimens from 104 patients (38 foregut, 53 midgut, 13 hindgut) were studied immunohistochemically for cytoplasmic and nuclear survivin expression as well as for ki-67 antigen expression. 5-year-follow-up was complete in 89 patients. 29 patients with localized, well-differentiated gastroenteropancreatic neuroendocrine tumors (WDET, WHO class 1) had been curatively treated by surgical or endoscopic tumor resection. 50 patients suffered from well-differentiated endocrine carcinomas (WDEC, WHO class 2), 10 patients were diagnosed with poorly differentiated neuroendocrine carcinomas (PDEC, WHO class 3). Survivin expression was correlated with survival for the 50 patients with metastatic WDEC disease. All 29 WDETs were negative for nuclear survivin, whereas all 10 PDECs stained positive for nuclear survivin. In the 50 patients with metastatic WDECs, 5/50 (10%) tumors were nuclear survivin positive. Those 5 patients had a statistically significant worse prognosis (survival of 41 vs. 103 months, p = 0.01). ki-67 was not a prognostic factor in this subgroup of patients. Nuclear survivin expression thus appears to be upregulated during progression of gastroenteropancreatic neuroendocrine tumors. The analysis of nuclear survivin expression identifies subgroups in metastatic disease (WHO class 2) with good (survivin^–) or with less favorable prognosis (survivin⁺). We propose that the determination of nuclear survivin expression could be used to individualize therapeutic strategies in this tumor entity in the future.

Our previous microarray-based studies identified secretagogin to be highly expressed in normal colon mucosa compared to basal expression in colon adenocarcinomas. The aim of this study was to analyze the differential expression of secretagogin in normal mucosa, adenocarcinomas, and neuroendocrine tumors. Western blotting, immunohistochemistry, immunofluorescence microscopy and ELISA were applied. Western blot analysis detected a 32-kDa secretagogin band in samples from normal mucosa. Immunohistochemical analyses on tissue specimens showed that secretagogin is exclusively expressed in neuroendocrine cells and nerve cells in normal mucosa of the digestive tract. Tissues adjacent to benign hyperplasic polyps and adenomas showed a decreased number of secretagogin-expressing neuroendocrine cells. Secretagogin co-localized with neuroendocrine markers (chromogranin A, neuron-specific enolase, synaptophysin) in neuroendocrine cells in crypts of normal mucosa, and in tumor cells of carcinoids. Secretagogin was strongly expressed in the cytosol and the nucleus of 19 well-differentiated neuroendocrine carcinoids and carcinoid metastases, as well as in neuroendocrine tumors from the lung, pancreas and adrenal gland. Secretagogin was detected in plasma from carcinoid patients with distant metastasis. Combined immunohistochemical analysis of secretagogin and FK506-binding protein 65, a protein de novo synthesized in adenocarcinomas, distinguished well-differentiated carcinoids, adenocarcinoids and undifferentiated carcinomas. We conclude that secretagogin is a novel marker for neuroendocrine differentiation.