No abstract

In a study of a diverse set of human tumor cell lines previously shown to all have a defect in methionine metabolism (Stern, P. H., Wallace, C.D. and Hoffman, R.M., J. Cellular Physiology 119, 29-34, 1984), we demonstrate in this report that all have enhanced overall rates of transmethylation compared to normal human fibroblasts. Transmethylation rates were measured by blocking S-adenosylhomocysteine hydrolase and measuring the AdoHcy which accumulates as a result of transmethylation. The enhanced transmethylation rates may be the basis of the above-mentioned defects in methionine metabolism previously reported in human tumor cells, including the basis of the inability of the majority of the tumor cells to grow when methionine is replaced by homocysteine. The excess and unbalanced tRNA methylation observed for the last 25 years in many types of cancer may be at least in part explained by our results of elevated rates of overall transmethylation in cancer cells. The alteration of such a fundamental process as transmethylation in cancer may be indicative of its importance in the oncogenic process.

Methionine (Met) dependence-i.e., the inability of cultured cells to grow when Met is replaced by its immediate precursor homocysteine (Met-Hcy' medium)-is a frequent component ofthe oncogenically transformed phenotype. Normal cells, on the other hand, grow in this medium. There have been reports

An in vitro test of cell sensitivity to drugs that indicates in vivo response is an important need in cancer therapy and cancer drug development. Toward this end, we previously developed a collagen gel-supported culture system for growth of human tumors. This three-dimensional culture system is general and grows tumors at high frequency directly from surgery or biopsy that maintain important in vivo properties in vitro, including tissue architecture. We report here that with autoradiographic techniques measuring cellular DNA synthesis the drug responses of individual cells within the tissue structure of in vitro-grown tumors can be determined. Twenty tumor classes, including all the major ones, have been measured in toto at >50% frequency. Quantitative and qualitative results show increasing cell kill with rising cytotoxic drug concentration, differential drug sensitivities of multiple cell types within individual cultured tumors, differential sensitivities of a series of tumors of the same histopathological classification to a single drug, differential sensitivities of individual tumors to a series of drugs, and sensitivity patterns of various tumor types similar to the sensitivities found in vivo.Therefore, the results indicate that potentially important therapeutic data can be obtained from tumor specimens growing in vitro for the individual cancer patient as well as for rational and relevant screening for new agents active against human solid tumors.A major clinical problem is that cancers that are classified as identical according to their histopathological characteristics are nonetheless highly individual in their drug sensitivities and there is currently no way to predict clinical outcome of chemotherapy for individual patients (1). A second major problem is that there is currently no relevant assay to screen for new human anticancer agents, especially for solid tumors (2). To overcome these problems, many attempts have been made to develop in vitro drug-sensitivity tests for individual cancer patients about to undergo chemotherapy and to screen for new anticancer agents. These attempts have suffered in one form or another from their inability to support growth of human tumors such that they reflect the in vivo situation. Plating of dissociated tumor cells in soft agar (3)(4)(5)(6)(7)(8)(9)(10)(11)(12) and monolayer cultures (13-15) does not in many instances allow the growth of tumor cells. Often, when tumor cells do grow under these conditions, other cell types present in the original tumor probably have not grown. It is critical that cell types present in the original tumor be present in the assay since it has been shown that interactions between cell types can alter their drug sensitivities (16). Multicellular spheroids that are three-dimensional have been used recently for drug sensitivity testing (17,18), but these also involve dissociation of cells from the tumor and reassociation into structures that do not resemble the original tissues. Short-term in vitro assays of drugs on noncultured n...

A metabolic defect that is prevalent in human cancer cell lines was exploited to selectively kill these cells without killing cocultured normal human fibroblasts. Methionine dependence, a metabolic defect seen only in cancer cells or immortalized cell lines in vitro, precludes the cells from growing in media in which methionine is replaced by its immediate precursor, homocysteine, a condition that allows the growth of all normal cell strains tested. The methionine-dependent cells become reversibly blocked in late S-G2 (i.e., late-S and G2 phases) under the above condition, a block that was exploited for selective chemotherapy against these cells. In cultures that were initiated with equal amounts of cancer cells and human diploid fibroblasts, substitution of homocysteine and doxorubicin for methionine in the culture medium followed by methionine repletion with vincristine was totally effective at selectively eliminating a methionine-dependent human sarcoma and 3 methionine-dependent human carcinomas. The above protocol was nearly totally effective against a partially methionine-independent revertant of the sarcoma. The chemotherapeutic procedure used was not lethal to normal cells growing alongside the tumor cells and was ineffective when conducted totally in methionine-containing medium. The optimal procedure was 10(-10) M doxorubicin in methionine-free, homocysteine-containing medium for 10 days followed by 2 x 10(-7) M vincristine in methionine-containing, homocysteine-free medium for 1 day, in turn followed by drug-free methionine-containing, homocysteine-free medium. These results demonstrate the potential for treatment of solid tumors with chemotherapy based on metabolic differences between normal and tumor cells.