The monoclonal antibody (MAb) Ki-67, which is directed against a proliferation-associated nuclear antigen, was used to measure tumor proliferation in 165 carcinomas of the esophagus, stomach, colon and rectum with an indirect immunoperoxidase technique. The percentage of Ki-67-positive tumor cells (Ki-67 index) was evaluated with the point-counting method. The Ki-67 index in gastric cancers (mean, 24.8%; standard deviation, 11.1%) was significantly lower than in tumors of the esophagus (35.7 +/- 12.6%), colon (37.6 +/- 15.2%), and rectum (34.3 +/- 16.4%). A wide range of the Ki-67 index (5.9-75.3%) could be observed within the various tumor types. In recurrent colorectal carcinomas, the Ki-67 index significantly increased to 51.9%. The Ki-67 index was independent of pathologic (e.g., TNM-stage, grading, tumor volume, tumor site) and clinical variables (age and gender of the patients). A marked heterogeneity of Ki-67 expression within different tumor stages was noted. Statistically significant regional variations in tumor proliferation existed between different areas within the same tumor.

Cell proliferation was measured by a three-step immunoperoxidase technique on cryostat sections of 61 resected colorectal adenocarcinomas using the monoclonal antibody Ki-67 that is directed against a proliferation-associated nuclear antigen. The percentage of Ki-67-positive cells was quantified with the point-counting method. The frequency distribution of the percentage of Ki-67-positive tumor cells in 52 unirradiated carcinomas was gaussian with a mean of 38.7% (range, 7.7% to 75.3%). Analysis of the unirradiated tumors showed no relation between Ki-67 staining and various clinicopathologic features including age, sex, tumor volume, tumor differentiation, location, and tumor stage. Although some tumors demonstrated intratumor heterogeneity of immunoreactivity, there was no significant difference in proliferative activity among peripheral, intermediate, and central tumor areas. Whereas the Ki-67 index increased to 47.1% in recurrent carcinomas, it decreased significantly to 24.7% in rectal carcinomas given radiotherapy before surgery. Therefore, Ki-67 immunostaining might be used as a tool to select and monitor patients with colorectal cancers who might benefit from radiotherapy. The marked differences of Ki-67 expression among different tumors may relate to heterogeneity in growth kinetics and may therefore carry prognostic implications.

Cancer testis antigens are encoded by germ line-associated genes that are present in normal germ cells of testis and ovary but not in differentiated tissues. Their expression in various human cancer types has been interpreted as 're-expression' or as intratumoral progenitor cell signature. Cancer testis antigen expression patterns have not yet been studied in germ cell tumorigenesis with specific emphasis on intratubular germ cell neoplasia unclassified as a precursor lesion for testicular germ cell tumors. Immunohistochemistry was used to study MAGEA3, MAGEA4, MAGEC1, GAGE1 and CTAG1B expression in 325 primary testicular germ cell tumors, including 94 mixed germ cell tumors. Seminomatous and non-seminomatous components were separately arranged and evaluated on tissue microarrays. Spermatogonia in the normal testis were positive, whereas intratubular germ cell neoplasia unclassified was negative for all five CT antigens. Cancer testis antigen expression was only found in 3% (CTAG1B), 10% (GAGE1, MAGEA4), 33% (MAGEA3) and 40% (MAGEC1) of classic seminoma but not in nonseminomatous testicular germ cell tumors. In contrast, all spermatocytic seminomas were positive for cancer testis antigens. These data are consistent with a different cell origin in spermatocytic seminoma compared with classic seminoma and support a progression model with loss of cancer testis antigens in early tumorigenesis of testicular germ cell tumors and later re-expression in a subset of seminomas.

Cellular differentiation in 22 surgically removed adenocarcinomas of the gallbladder was immunohistochemically studied with antibodies specific to mucins of gastric foveolar cells (M1), (pseudo)pyloric cells (M2) and intestinal goblet cells (M3), and also with antibodies against pepsinogen II and chromogranin A. More than 70% of tumours (16 of 22 cases) displayed gastric- and/or intestinal-type differentiation, most of which (12 of 16 cases) showed both types of differentiation. Two tumours showed an organoid growth pattern similar to the normal gastric mucosa. The presence of endocrine cells positive for chromogranin A was closely related to that of gastric- and/or intestinal-type cells. The present findings clearly indicate the multidirectional differentiation of gallbladder adenocarcinomas and suggest that most gallbladder adenocarcinomas develop and progress under induction of gastric and intestinal differentiation.

Summary The vascularisation of rectal and oesophageal carcinomas and of normal mucosa was studied using histochemical and immunohistochemical methods. Endothelial cells were stained for alkaline phosphatase (AP) using an azo-dye procedure. Histochemical results were compared with the immunohistochemical identification of endothelial cells using the monoclonal antibody BW200 recognising an epitope restricted to human endothelial cells. In Histochemical demonstration of alkaline phosphatase in the endothelium of the arterial part of the terminal vascular supply has been advocated for quantitative analysis of the vascularisation in human tumours (Mlynek et al., 1985). However, studies in a variety of different tumour types (Monis & Rutenburg, 1960) have shown that sometimes the simultaneous reactivity of stromal connective tissue may preclude the exact evaluation of blood vessels in human cancers.The capabilities of this histochemical method for the measurement of vascular density were therefore compared with an immunohistochemical technique using a monoclonal antibody directed against endothelium. This study was performed in rectal and oesophageal cancers because in these tumour types pre-and postoperative irradiation is often included in therapeutic regimens. Materials and methodsCryostat sections (5,um) of rectal (n=13) and oesophageal cancers (n= 17) and normal mucosa were air-dried, fixed in acetone and stained for alkaline phosphatase (AP) using an azo-dye method. Tissue sections were incubated in a solution containing diazotised triamino-tritolyl-methanechloride (New Fuchsine, CI no. 42520) and naphtol AS-BI phosphate as substrate (Stutte, 1967).For the immunohistochemical staining of vascular endothelium the monoclonal antibody BW 200 was used. This recognises an epitope on an antigen molecule restricted to human neoplastic and non-neoplastic endothelial cells (Alles & Bosslet, 1986 (Chalkley, 1943). In normal tissues, vascular parameters were evaluated in the mucosa and submucosa. It has been shown (Mlynek et al., 1985) that quantification of vascular parameters in normal colorectal mucosa was independent of the spatial orientation of vessels on histological sections. Therefore, only sections cut perpendicularly and not parallel to the luminal surface were used, allowing a simultaneous analysis of vessels in the mucosa and submucosa. Only this spatial orientation of histological sections rendered it possible to analyse the capillary plexus near the basal epithelial cells of normal oesophageal mucosa, which might have been missed on sections cut parallel to the luminal surface. In six patients, comparative microscopic analysis of two randomly oriented histological sections cut from a single tumour block did not reveal a preferential spatial orientation of vessels in tumours.Sections were examined through a graticule with a regular arrangement of 25 crosses inserted in a 10x eye-piece in combination with a 40 x objective. Sections were scanned by systematic sampling (Weibel, 1979 Because not all the vessel...