Summary An analysis of the activity of compounds tested in pre-clinical in vivo and in vitro assays by the National Cancer Institute's Developmental Therapeutics Program was performed. For 39 agents with both xenograft data and Phase II clinical trials results available, in vivo activity in a particular histology in a tumour model did not closely correlate with activity in the same human cancer histology, casting doubt on the correspondence of the pre-clinical models to clinical results. However, for compounds with in vivo activity in at least one-third of tested xenograft models, there was correlation with ultimate activity in at least some Phase II trials. Thus, an efficient means of predicting activity in vivo models remains desirable for compounds with anti-proliferative activity in vitro. For 564 compounds tested in the hollow fibre assay which were also tested against in vivo tumour models, the likelihood of finding xenograft activity in at least one-third of the in vivo models tested rose with increasing intraperitoneal hollow fibre activity, from 8% for all compounds tested to 20% in agents with evidence of response in more than 6 intraperitoneal fibres (P < 0.0001). Intraperitoneal hollow fibre activity was also found to be a better predictor of xenograft activity than either subcutaneous hollow fibre activity or intraperitoneal plus subcutaneous activity combined. Since hollow fibre activity was a useful indicator of potential in vivo response, correlates with hollow fibre activity were examined for 2304 compounds tested in both the NCI 60 cell line in vitro cancer drug screen and hollow fibre assay. A positive correlation was found for histologic selectivity between in vitro and hollow fibre responses. The most striking correlation was between potency in the 60 cell line screen and hollow fibre activity; 56% of compounds with mean 50% growth inhibition below 10 -7.5 M were active in more than 6 intraperitoneal fibres whereas only 4% of compounds with a potency of 10 -4 M achieved the same level of hollow fibre activity (P < 0.0001). Structural parameters of the drugs analysed included compound molecular weight and hydrogen-bonding factors, both of which were found to be predictive of hollow fibre activity.

A series of N-alkylamide analogues of the lavendustin A pharmacophore were synthesized and tested for inhibition of the epidermal growth factor receptor (EGFR) protein tyrosine kinase and the nonreceptor protein tyrosine kinase Syk. Although several compounds in the series were effective inhibitors of both kinases, it seemed questionable whether their inhibitory effects on these kinases were responsible for the cytotoxic properties observed in a variety of human cancer cell cultures. Accordingly, a COMPARE analysis of the cytotoxicity profile of the most cytotoxic member of the series was performed, and the results indicated that its cytotoxicity profile was similar to that of antitubulin agents. This mechanism of action was supported by demonstrating that most compounds in the series were moderately effective as inhibitors of tubulin polymerization. This suggests that the lavendustin A analogues reported here, as well as some of the previously reported lavendustin A analogues, may be acting as cytotoxic agents by a mechanism involving the inhibition of tubulin polymerization.

Two purified animal venom toxins, crotoxin and cardiotoxin, have been combined to produce a unique natural product (VRCTC-310) currently under investigation as an antitumor agent by the National Cancer Institute. In vitro, it has demonstrated cytotoxic disease specificity and a unique mechanism of action when submitted to COMPARE analysis. In vivo, tolerance was developed to the neurotoxic properties of crotoxin which allowed comparison of several schedules of fixed and escalating daily i.m. doses to mice bearing s.c. Lewis Lung carcinoma. An 83% inhibition of tumor growth was achieved using an escalating dose schedule starting at 1.8 mg/kg and reaching 6.3 mg/kg/day on day 20. Although some irritation around the sites of i.m. injection was noted, animal weight loss was negligible and there were no other signs of adverse toxicity. This natural product represents a new, membrane interactive anticancer agent which produces a unique spectrum of cytotoxicity in vitro and which has demonstrated interesting in vivo antitumor efficacy.

Altered cell cycle control has emerged as a recurring theme in neoplasia. Strategies that would return toward normal the altered cell cycle control present in tumor cells have appeal as novel approaches to cancer treatment. Cyclin‐dependent kinases (CDKs) control the progression through the cell cycle, operating at the transition from the G2 to M and G1 to S phases, and progression through S. CDKs are regulated by a complex set of mechanisms, including the presence of activating cyclins, regulatory phosphorylations, and endogenous CDK inhibitors at “checkpoints.” This overview focuses on progress in defining compounds that can antagonize directly the action of CDKs. These have emerged as various types of ATP site‐directed inhibitors, including flavopiridol, N‐substituted adenine derivatives, the natural product butyrolactone, staurosporine derivatives, and, more recently, the synthetic paullones. Paullones appear to be of interest in that one of the most active members of the class, 9‐nitropaullone (alsterpaullone), requires relatively brief periods of exposure to living cells in order to effect lasting effects on cellular and proliferative potential. Two of these compounds, flavopiridol and UCN‐01 (7‐hydroxy‐staurosporine), have entered early clinical trials and achieved concentrations that might potentially modulate CDK function. In the case of UCN‐01, unexpected human plasma protein binding might prevent direct inhibition of CDKs but allow drug concentrations to be achieved that indirectly affect CDKs by checkpoint abrogation. Further studies with CDK inhibitors should define the expected end point of CDK inhibition more clearly in preclinical models and clinical systems, including cytostasis, apoptosis, or differentiation.

A new approach to cancer drug discovery targets molecules important in cancer pathogenesis. This approach is thought to be of greater promise than the antiproliferative screens which discovered cytotoxic agents and dominated cancer drug discovery for 60 years. However, one cannot lose sight of the fact that these targets exist in the cellular environment consisting of many additional influences on target function, and that effective drug treatment will take into account drug uptake, metabolism and elimination at the level of the cell as well as the organism. A key goal is to define for the new millennium a path to cancer drug discovery and development which accounts for the cancer cell phenotype in its totality rather than as arising solely from single molecular targets. The US National Cancer Institute maintains a cell-based drug discovery screen which can define a context for drug action in the milieu of more than 300 molecular targets and thousands of gene expression patterns which have been measured in the 60 human tumour cell lines which comprise the screening panel. The challenge of the millennium will be addressed by molecules active against defined targets but with selectivity of action occurring in the milieu of deregulated cancer cell biology in all its aspects. © 2000 Cancer Research Campaign http://www.bjcancer.com

No abstract

A new strategy is presented here which integrates combinatorial library technology with the antitumor in vitro screening system at the National Cancer Institute in the search for novel antitumor agents. Mixture-based synthetic combinatorial libraries (SCLs) representing hundreds of thousands to millions of individual compounds were screened against the cell-based assay, which evaluates compounds for their ability to inhibit the growth of 60 different human tumor cell lines. Five different SCLs, composed of peptides, peptidomimetics, polyamines or small molecules were first tested against three cell lines to identify the most active SCLs. Two SCLs, namely the N-perbenzylated pentamine and the N-acylated permethylated triamine, were deconvoluted to yield individual compounds having significant activities against the 60 tumor cell lines. Active compounds were tested in mice to determine the maximum tolerated dose, followed by in vivo testing in a hollow fiber assay. Using this strategy, three different compounds identified directly from SCLs are currently being evaluated in human tumor xenografts. This study demonstrates for the first time the use of in vitro cell-based assays to identify antitumor lead compounds from mixture-based combinatorial libraries.