G. Riethmüller

Since a dysregulated synthesis of tumor necrosis factor alpha (TNF-alpha) may be involved in the pathogenesis of autoimmune diseases, it was of interest to precisely locate the recently reported NcoI restriction fragment length polymorphism (RFLP) of the TNF-alpha region. However, by mapping of 56.8 kb of overlapping cosmid clones and direct sequencing, we could localize the polymorphic NcoI restriction site within the first intron of the TNF-beta gene and not in the TNF-alpha gene. To study whether regulatory mechanisms are affected by this polymorphism, we analyzed the TNF-alpha/TNF-beta production of phytohemagglutinin-stimulated peripheral blood mononuclear cells of individuals homozygous for the TNF-beta NcoI RFLP by ELISA and concomitant Northern blot analysis. On days 2-4 after stimulation with mitogen, the TNFB*1 allele corresponding to a 5.3-kb NcoI fragment presented with a significantly higher TNF-beta response. A mRNA analysis demonstrated that higher protein levels of TNF-beta correlate also with increased amounts of TNF-beta transcripts. No allelic association was found in respect to TNF-alpha production. To further investigate a possible allelic influence on transcription, we determined the DNA sequence of 2 kb of the 5' portion of our cloned TNFB*2 allele and compared it with the available TNF-beta sequences. By computer-aided recognition motif search of DNA binding factors, we report putative binding sites conserved between mouse and man in the 5' flanking region as well as in intron 1 of the TNF-beta gene, found also in other cytokine promoter sequences. In addition, by polymerase chain reaction amplification and sequencing of 740 bp of the 5' part of TNF-beta of individuals typed homozygously for the NcoI RFLP, we could show that amino acid position 26 is conserved as asparagine in the TNFB*1 and as threonine in the TNFB*2 sequence. A previously reported, EcoRI RFLP in the 3' untranslated region of TNF-beta does not segregate with either of the two alleles. Thus, four TNFB alleles can be defined at the DNA level.

PURPOSE: As previously shown, antibody treatment increased survival of patients with resected colorectal cancer of stage Dukes' C. Since the 5-year analysis was criticized because of the wide range (2.7 to 7.5 years) of follow-up time, we performed a 7-year analysis with only four of 189 patients monitored for less than 5 years. PATIENTS AND METHODS: A total of 189 patients with resected Dukes' C colorectal cancer were randomly allocated to infusions of a total of 900 mg 17-1A antibody, 500 mg postoperatively followed by 4 monthly doses of 100 mg (n=99), or to observation only (n=90). Primary end points were overall survival and disease-free interval. Patients were stratified by a dynamic randomization according to center, sex, location of tumor, number of affected lymph nodes, and preoperative carcinoembryonic antigen concentration. RESULTS: Randomization produced balanced distribution of risk factors. After 7 years of follow-up evaluation, treatment had reduced overall mortality by 32% (Cox's proportional hazard, P < .01; log-rank, P=.01) and decreased the recurrence rate by 23% (Cox's proportional hazard, P < .04; log-rank, P=.07). The intention-to-treat analysis gave a significant effect for overall survival (Cox's proportional hazard, P < .01; log-rank, P=.02) and disease-free survival (Cox's proportional hazard, P=.02; log-rank, P=.11 ). While distant metastases were significantly reduced (Cox's proportional hazard, P=.004; log-rank, P=.004), local relapses were not (Cox's proportional hazard, P=.65; log-rank, P=.52). This differential effect of 17-1A antibody on disseminated isolated tumor cells versus occult local satellites may explain the increased significance seen in the overall survival. CONCLUSION: The now-matured study shows that 17-1A antibody administered after surgery prevents the development of distant metastasis in approximately one third of patients. The therapeutic effect is maintained after 7 years of follow-up evaluation.

BACKGROUND: The development of monoclonal antibodies (MAbs) to cytokeratins, which are integral components of the epithelial cytoskeleton, has made possible immunocytochemical detection of epithelial tumor cells. Importantly, this technique allows the detection of epithelial tumor cells that have metastasized from primary adenocarcinomas to secondary sites such as the bone marrow. PURPOSE: The aim of the study was not only to detect micrometastatic cells in bone marrow, but also to assess the expression of nuclear proliferation markers (Ki-67 and p120) and the erbB2 oncogene (also known as ERBB2) in these cells and, thus, hopefully improve prognostic precision. METHODS: Bone marrow aspirates were obtained from both sides of the upper iliac crest of 532 patients having definitive diagnoses of either breast or gastrointestinal cancer. The presence of micrometastatic epithelial tumor cells in bone marrow was assayed using the MAb cytokeratin 2 (CK2) to cytokeratin component 18 (CK18), in combination with the alkaline phosphatase-anti-alkaline phosphatase immunostaining technique. After primary screening of all marrow samples with MAb CK2, representative subgroups of CK18+ samples were selected for co-labeling with MAbs either to ErbB (n = 16), ErbB2 (n = 121), Ki-67 (n = 33), or p120 (n = 36) protein. An alternative labeling protocol based on the combination of immunogold and immunoenzymatic techniques was utilized to confirm the results derived from immunoenzymatic double staining. RESULTS: In total, single CK18-positive tumor cells were detected in 180 (33.8%) of 532 bone marrow aspirates, with few differences among patients with breast or gastrointestinal cancer in TNM stage M0 (i.e., no distant metastasis). In patients with overt metastasis (stage M1), however, the incidence of metastatic cells in marrow increased to 73.7% in breast cancer, 52.5% in gastric cancer, and 39.0% in colon cancer. Whereas expression of Ki-67 or p120 on micrometastatic cells was observed only in 11 (15.9%) of 69 cancer patients analyzed, ErbB2+/CK18+ cells were found in 48 (67.6%) of 71 breast cancer patients and 14 (28.0%) of 50 patients with gastrointestinal cancer (P = .0001). The incidence of ErbB2+/CK18+ cells was positively correlated with the clinical stage of tumor progression. CONCLUSIONS: The high incidence of ErbB2 expression on micrometastatic breast cancer cells in the bone marrow suggests that these cells might have been positively selected during early stages of metastasis. The majority of these cells appear to be in a dormant state of cell growth. IMPLICATIONS: Although support from clinical follow-up is still needed, this study demonstrates that, beyond the mere presence of micrometastatic cells in bone marrow, useful prognostic information can be obtained by analysis of additional cell growth markers.