To establish well-characterized cellular reagents for the study of colon carcinoma, we have examined 19 human colorectal carcinoma cell lines with regard to morphology, ultrastructure, expression of tumor-associated antigens, proliferative capacity in vitro, anchorage-independent growth, oncogene expression, tumorigenicity and malignant potential. Cell lines examined were cultured under identical conditions, and in vitro and in vivo analyses were performed in parallel on replicate cultures. Three classes of colorectal cell lines were defined according to their tumorigenicity in nude mice. Class-1 lines formed rapidly progressing tumors in nearly all mice at an inoculum of 10(6) cells. Cell lines belonging to class-2 were less tumorigenic, producing tumors later and at a slower growth rate. Class-3 lines were non-tumorigenic under all experimental conditions tested. By Northern analysis, the oncogenes c-myc, H-ras, K-ras, N-ras, myb, fos and p53 were expressed in nearly all cell lines examined. In contrast, transcripts for abl, src and ros were not detected. The best in vitro predictor of tumorigenicity was colony formation in soft agar. There was no detectable correlation between tumorigenicity and metastatic potential, doubling time in vitro, production of tumor-associated markers, xenograft histology or expression of specific oncogenes.

To investigate the role of oncogene activation in the pathogenesis of malignant tumors, we have studied the tumorigenic and metastatic properties of NIH/3T3 secondary transfectants (designated A51) containing an activated c-Haras-) gene derived from the human T24 bladder carcinoma cell line and compared them Tumors are classically defined as benign or malignant (1). Benign tumors are noninvasive growths that do not spread to distant organs. Unless located in a functionally vital site (e.g., brain), they pose little threat to the patient and usually can be removed surgically. In contrast, malignant neoplasms are readily invasive, metastasize to organs throughout the body, and eventually kill their host (1). The biochemical events that distinguish malignant from benign neoplasms remain unidentified. Work in several laboratories has implicated oncogene activation in the expression of tumorigenicity but the role of oncogenes in the pathogenesis of malignant versus benign tumors has received little attention. The NIH/3T3 cell line, although immortalized in vitro, is reportedly nontumorigenic and is widely used as a recipient cell line for detecting transforming gene sequences isolated from tumorigenic cell lines or tissues (2-6). For example, upon transfection with the activated ras gene, NIH/3T3 cells form foci in tissue culture, exhibit anchorage independence, and, in those experiments in which in vivo studies have been conducted, produce tumors in nude mice (6-10). However, the behavior of the tumors formed by transfected NIH/3T3 cells has not been rigorously evaluated and it remains unclear whether oncogene activation is an event associated strictly with the pathogenesis of benign neoplasms or whether activation is also an essential feature for expression of metastatic properties. We report that transfection with the activated c-Ha-ras-l gene accelerates the tumorigenicity and enhances the metastatic potential of NIH/3T3 Tumorigenicity and Metastasis. Tumorigenicity and spontaneous metastatic potential were assayed by inoculating mice with different cell doses in the footpad (i.m.) or the supraclavicular region (s.c.). Tumor size was monitored at the supraclavicular site every 2-3 days by caliper measurement. For studies on experimental metastasis, different numbers of cells were injected into the tail vein of nude mice. At autopsy the major organs of all animals were examined both grossly and histologically for evidence of metastases. Single sections were prepared from each organ except the lung, in which case multiple sections were examined.Detection of Activated c-Ha-ras Oncogene and Human Alu Sequences. For preparation of DNA (12), cell monolayers established from primary tumors or metastatic foci were dispersed into phosphate-buffered saline (Pi/NaCl), pelleted, rinsed, resuspended in 10 mM Tris HCl, pH 8.0/0.35 M NaCl/1 mM EDTA, lysed in 0.5% NaDodSO4, and treated for 4-12 hr with Pronase (0.1 mg/ml) at 37°C. DNA was extracted with phenol, ethanol precipitated, and dissolved in 10 mM Tris-HCl, pH ...

The metastatic properties of tumor cell clones isolated from individual lesions of B16 melanoma metastatic to lung have been examined at different stages in the evolution of metastasis. Clonal analysis of metastatic lesions produced by B16 melanoma populations containing clones with identifiable, stable drug-resistance markers revealed that the majority (>80%) of experimental metastases produced by intravenous injection oftumor cells are of unicellular origin. During the early stages of their growth (<25 days after initial tumor cell arrest), the majority of metastatic lesions contain cells with indistinguishable metastatic phenotypes (intralesional clonal homogeneity) although different clonally homogeneous lesions from the same host contain tumor cells with different metastatic phenotypes (interlesional clonal heterogeneity). Progressive growth of metastatic lesions is accompanied by emergence, within originally clonally homogeneous lesions, of variant tumor cells with altered metastatic properties (intralesional clonal heterogeneity). By 40-45 days after initial arrest of injected tumor cells in the lung, 90% of the metastatic lesions are populated by cells with heterogeneous metastatic phenotypes.Studies in several laboratories have shown that malignant tumors contain subpopulations of cells that differ widely in their metastatic abilities (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14). This cellular heterogeneity is believed to result from the formation of variant subpopulations of tumor cells with altered metastatic properties during progressive tumor growth (15). The factors that influence the genesis and regulation of cellular diversity within malignant tumors are poorly understood. Recent studies (16)(17)(18)(19) suggest that the rate at which tumor cell variants with altered metastatic properties are generated in vitro is influenced by the extent of subpopulation diversity within the overall cell population. We have shown that the rate of formation of metastatic variants is significantly higher in populations containing a limited number of tumor cell subpopulations than in highly heterogeneous, polyclonal populations containing multiple cellular subpopulations (16,17). We have also shown that the majority oflung metastatic lesions produced by the murine B16 melanoma arise from the single cells (17). Formation of such lesions thus represents a situation in which subpopulation diversity is restricted. It was therefore considered of interest to determine whether this situation would stimulate rapid formation of new tumor cell variants to generate phenotypically diverse subpopulations oftumor cells within individual metastases. Cells. The origin and properties of the B16 melanoma B16-F10 subline have been described (14,20). Drug-resistant (Dr) variants were selected from the B16-FlO subline by treatment with the chemical mutagen N-methyl-N'-nitro-N-nitrosoguanidine as described (16,17). Variants were selected for resistance to trifluorothymidine (2 ,ug/ml; TFT%), diaminopurine (47 p.M; DAPr), vin...

The growing body of evidence showing that malignant tumors are heterogeneous and contain diverse subpopulations of tumor cells is reviewed, with particular emphasis being given to the presence of tumor-cell subpopulations with differing metastatic properties. The factors that may influence the evolution of cellular diversity at different stages in the progression of malignant neoplasms are discussed. Emphasis is given to the possibility that interactions occurring amongst the constituent subpopulations of a malignant tumor may influence the rate at which new variant subpopulations emerge. Metastatic heterogeneity poses significant problems for experimental efforts to identify features unique to metastatic cells and also for the therapy of metastatic disease.

Metastasis is a complex process whereby tumour cells from a primary neoplastic growth disseminate throughout the body and establish secondary tumour foci in distant organs. Biochemical traits associated with, or essential for, the expression of the metastatic phenotype have not yet been identified. In the course of examining stimulation of the B16 murine melanoma adenylate cyclase by melanocyte-stimulating hormone (MSH) and by the diterpene forskolin, we noted that tumour cell clones isolated from common parent cell populations differed widely in their responses to these agonists. We report here that the accumulation of cyclic AMP induced by MSH or forskolin shows a strong positive correlation with the ability of B16 melanoma clones to form pulmonary tumour colonies when injected intravenously (i.v.) into syngeneic mice ('experimental metastasis'). In parallel in vitro analyses of cyclic AMP metabolism and in vivo assays of experimental metastasis using replicate cell preparations, highly metastatic tumour cell clones consistently show greater than a 30-fold increase in cellular cyclic AMP when exposed to MSH or forskolin. By contrast, clones with limited metastatic abilities respond to the same agonists with only a two- to threefold increase in cellular cyclic AMP. These data suggest that cyclic AMP metabolism is linked with biochemical pathways that are responsible for the formation of experimental metastasis by the B16 melanoma.