A gene expression database for the molecular pharmacology of cancer.

We used cDNA microarrays to assess gene expression profiles in 60 human cancer cell lines used in a drug discovery screen by the National Cancer Institute. Using these data, we linked bioinformatics and chemoinformatics by correlating gene expression and drug activity patterns in the NCI60 lines. Clustering the cell lines on the basis of gene expression yielded relationships very different from those obtained by clustering the cell lines on the basis of their response to drugs. Gene-drug relationships for the clinical agents 5-fluorouracil and L-asparaginase exemplify how variations in the transcript levels of particular genes relate to mechanisms of drug sensitivity and resistance. This is the first study to integrate large databases on gene expression and molecular pharmacology.

Thiourea reverses cross-links and restores biological activity in DNA treated with dichlorodiaminoplatinum (II).

Cis and trans dichlorodiaminoplatinum (II) compounds bind to DNA and form DNA cross-links, which are usually considered to be irreversible. Thiourea can reverse these cross-links without any apparent breakdown of the DNA. In addition, cis- and trans-Pt (II) treatment of lambda decreases its transfectivity. After suitable incubation with thiourea, full transfectivity of Pt(II)-treated lambda DNA can be restored.

Neural computing in cancer drug development: predicting mechanism of action.

Described here are neural networks capable of predicting a drug's mechanism of action from its pattern of activity against a panel of 60 malignant cell lines in the National Cancer Institute's drug screening program. Given six possible classes of mechanism, the network misses the correct category for only 12 out of 141 agents (8.5 percent), whereas linear discriminant analysis, a standard statistical technique, misses 20 out of 141 (14.2 percent). The success of the neural net indicates several things. (i) The cell line response patterns are rich in information about mechanism. (ii) Appropriately designed neural networks can make effective use of that information. (iii) Trained networks can be used to classify prospectively the more than 10,000 agents per year tested by the screening program. Related networks, in combination with classical statistical tools, will help in a variety of ways to move new anticancer agents through the pipeline from in vitro studies to clinical application.

An information-intensive approach to the molecular pharmacology of cancer.

Since 1990, the National Cancer Institute (NCI) has screened more than 60,000 compounds against a panel of 60 human cancer cell lines. The 50-percent growth-inhibitory concentration (GI50) for any single cell line is simply an index of cytotoxicity or cytostasis, but the patterns of 60 such GI50 values encode unexpectedly rich, detailed information on mechanisms of drug action and drug resistance. Each compound's pattern is like a fingerprint, essentially unique among the many billions of distinguishable possibilities. These activity patterns are being used in conjunction with molecular structural features of the tested agents to explore the NCI's database of more than 460,000 compounds, and they are providing insight into potential target molecules and modulators of activity in the 60 cell lines. For example, the information is being used to search for candidate anticancer drugs that are not dependent on intact p53 suppressor gene function for their activity. It remains to be seen how effective this information-intensive strategy will be at generating new clinically active agents.

Dose-response transition from cell cycle arrest to apoptosis with selective degradation of Mdm2 and p21WAF1/CIP1 in response to the novel anticancer agent, aminoflavone (NSC 686,288).

Aminoflavone (AF, NSC 686,288) is beginning clinical trials. It induces replication-mediated histone H2AX phosphorylation, DNA-protein crosslinks and activates p53. Here, we studied p21(CIP1/WAF1) and Mdm2 responses to AF. Although p53 stabilization and phosphorylation at serine 15 increased with dose and time of exposure, Mdm2 and p21(CIP1/WAF1) protein levels displayed a biphasic response, as they accumulated at submicromolar doses and then decreased with increasing AF. As both Mdm2 and p21(CIP1/WAF1) mRNA levels increased with AF concentration without reduction at higher concentrations, we measured the half-lives of Mdm2 and p21(CIP1/WAF1) proteins. Mdm2 and p21(CIP1/WAF1) half-lives were shortened with increasing AF concentrations. Proteasomal degradation appears responsible for the decrease of both Mdm2 and p21(CIP1/WAF1), as MG-132 prevented their degradation and revealed AF-induced Mdm2 polyubiquitylation. AF also induced protein kinase B (Akt) activation, which was reduced with increasing AF concentrations. Suppression of Akt by small interfering RNA was associated with downregulation of Mdm2 and p21(CIP1/WAF1) and with enhanced apoptosis. These results suggest that the cellular responses to AF are determined at least in part by Mdm2 and p21(CIP1/WAF1) protein levels, as well as by Akt activity, leading either to cell cycle arrest when Mdm2 and p21(CIP1/WAF1) are elevated, or to apoptosis when Mdm2 and p21(CIP1/WAF1) are degraded by the proteasome and Akt insufficiently activated to protect against apoptosis.

p21CDKN1A allows the repair of replication-mediated DNA double-strand breaks induced by topoisomerase I and is inactivated by the checkpoint kinase inhibitor 7-hydroxystaurosporine.

This study provides evidence for the importance of p21(CDKN1A) for the repair of replication-mediated DNA double-strand breaks (DSBs) induced by topoisomerase I. We report that defects of p21(CDKN1A) and p53 enhance camptothecin-induced histone H2AX phosphorylation (gammaH2AX), a marker for DNA DSBs. In human colon carcinoma HCT116 cells with wild-type (wt) p53, gammaH2AX reverses after camptothecin removal. By contrast, gammaH2AX increases after camptothecin removal in HCT116 cells deficient for p53 (p53-/-) or p21(CDKN1A) (p21-/-) as the cells reach the late-S and G2 phases. Since p21-/- cells exhibit similar S-phase arrest as wt cells in response to camptothecin and aphidicolin does not abrogate the enhanced gammaH2AX formation in p21-/- cells, we conclude that enhanced gammaH2AX formation in p21-/- cells is not due to re-replication. The cell cycle checkpoint abrogator and Chk1/Chk2 inhibitor 7-hydroxystaurosporine (UCN-01) also increases camptothecin-induced gammaH2AX formation and inhibits camptothecin-induced p21(CDKN1A) upregulation in HCT116 wt cells. TUNEL (terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end labeling) assays demonstrate that gammaH2AX formation in late S and G2 cells following CPT treatment corresponds to DNA breaks. However, these breaks are not related to apoptotic DNA fragmentation. We propose that p21(CDKN1A) prevents the collapse of replication forks damaged by stabilized topoisomerase I cleavage complexes.

The gadd and MyD genes define a novel set of mammalian genes encoding acidic proteins that synergistically suppress cell growth.

A remarkable overlap was observed between the gadd genes, a group of often coordinately expressed genes that are induced by genotoxic stress and certain other growth arrest signals, and the MyD genes, a set of myeloid differentiation primary response genes. The MyD116 gene was found to be the murine homolog of the hamster gadd34 gene, whereas MyD118 and gadd45 were found to represent two separate but closely related genes. Furthermore, gadd34/MyD116, gadd45, MyD118, and gadd153 encode acidic proteins with very similar and unusual charge characteristics; both this property and a similar pattern of induction are shared with mdm2, whic, like gadd45, has been shown previously to be regulated by the tumor suppressor p53. Expression analysis revealed that they are distinguished from other growth arrest genes in that they are DNA damage inducible and suggest a role for these genes in growth arrest and apoptosis either coupled with or uncoupled from terminal differentiation. Evidence is also presented for coordinate induction in vivo by stress. The use of a short-term transfection assay, in which expression vectors for one or a combination of these gadd/MyD genes were transfected with a selectable marker into several different human tumor cell lines, provided direct evidence for the growth-inhibitory functions of the products of these genes and their ability to synergistically suppress growth. Taken together, these observations indicate that these genes define a novel class of mammalian genes encoding acidic proteins involved in the control of cellular growth.

Molecular interaction map of the mammalian cell cycle control and DNA repair systems.

Eventually to understand the integrated function of the cell cycle regulatory network, we must organize the known interactions in the form of a diagram, map, and/or database. A diagram convention was designed capable of unambiguous representation of networks containing multiprotein complexes, protein modifications, and enzymes that are substrates of other enzymes. To facilitate linkage to a database, each molecular species is symbolically represented only once in each diagram. Molecular species can be located on the map by means of indexed grid coordinates. Each interaction is referenced to an annotation list where pertinent information and references can be found. Parts of the network are grouped into functional subsystems. The map shows how multiprotein complexes could assemble and function at gene promoter sites and at sites of DNA damage. It also portrays the richness of connections between the p53-Mdm2 subsystem and other parts of the network.

Effects of c-erbB2 overexpression on the drug sensitivities of normal human mammary epithelial cells.

BACKGROUND: Overexpression of the gene c-erbB2, which encodes a receptor tyrosine kinase, in breast tumors has been linked with either increased or decreased response of breast cancer patients to various therapies. In breast cancer cell lines, overexpression of exogenous c-erbB2 sometimes alters drug sensitivities but sometimes has no effect. To avoid the genetic complexities associated with established cancer cell lines, normal human mammary epithelial cells (HMECs) were studied to determine whether c-erbB2 overexpression by itself would alter chemosensitivity. METHODS: HMECs were designed to overexpress c-erbB2, and these cells were then evaluated for alterations in chemosensitivity. RESULTS: HMECs overexpressing c-erbB2 failed to show any alterations in chemosensitivity to a panel of chemotherapeutic agents, as indicated by 95% confidence intervals on growth curves of cells treated with or without the agent of interest. With the use of fluorescence-activated cell sorting to enrich for HMECs overexpressing c-erbB2 on their surface, an 85% pure population of cells was isolated and their chemosensitivity was evaluated. Again, the cells failed to display any alterations in chemosensitivity. CONCLUSIONS: These results suggest that overexpression of c-erbB2 is not sufficient by itself to induce changes in chemosensitivity. Cellular studies using normal human cells in which the complexity of the system can be carefully controlled by the addition of one, two, or even more genes associated with cancer development may provide valuable information about how the products of the genes interact with each other and which combinations are critical in regulating chemosensitivity.

Use of the Kohonen self-organizing map to study the mechanisms of action of chemotherapeutic agents.

BACKGROUND: Many natural and synthetic compounds might prove to be effective in cancer chemotherapy. To identify potentially useful agents, the National Cancer Institute screens over 10,000 compounds annually against a panel of 60 distinct human tumor cell lines in vitro. This screening program generates large amounts of data that are organized into relational databases. Important questions concern the information content of the data and ways to extract that information. Previously, statistical techniques have revealed that compounds with similar patterns of activity against the 60 cell lines are often similar in structure and mechanism of action. Feed-forward, back-propagation neural networks have been trained on this type of data to predict broadly defined mechanisms of action of chemotherapeutic agents. PURPOSE AND METHOD: In this report, we examine the information that can be extracted from the screening data by means of another type of neural network paradigm, the Kohonen self-organizing map. This is a topology-preserving function, obtained by unsupervised learning, that nonlinearly projects the high-dimensional activity patterns into two dimensions. Our dataset is almost identical to that used in the earlier neural network study. RESULTS: The self-organizing maps we constructed have several important characteristics. 1) They partition the two-dimensional array into distinct regions, each of which is principally occupied by agents having the same broadly defined mechanism of action. 2) These regions can be resolved into distinct subregions that conform to plausible submechanisms and chemically defined subgroups of submechanism. 3) These results (and exceptions to them) are consistent with those obtained with the use of such deterministic measures of similarity among activity patterns as the Euclidean distance or Pearson correlation coefficient. CONCLUSIONS: Our results indicate that the activity patterns obtained from the screen contain detailed information about mechanism of action and its basis in chemical structure. The self-organizing map can be used to suggest the mechanism of action of compounds identified by the screen as potentially useful chemotherapeutic agents and to probe the biology of the cell lines in the cancer screen. Kohonen self-organizing maps, unlike the previously applied neural networks, preserve and reveal the relationships among compounds acting by similar mechanisms and therefore have the potential to identify compounds that act by novel cytotoxic mechanisms.

Beyond DNA cross-linking: history and prospects of DNA-targeted cancer treatment--fifteenth Bruce F. Cain Memorial Award Lecture.

The origin of cancer chemotherapy can be traced to the wartime discovery of the lymphotoxic action of nitrogen mustards. These and other bifunctional agents were later found to produce various types of DNA cross-links, and some of these agents continue to be mainstays of current therapy. The cellular pharmacology of these drugs was studied extensively during the 1970s and 1980s by means of DNA filter elution methodology. In the course of these investigations, DNA topoisomerases were discovered to be targets of anthracyclines and several other classes of anticancer drugs. DNA cross-linkers and topoisomerase blockers have generally similar cytotoxic mechanisms, which depend on DNA damage detection, DNA repair, cell cycle arrest, and cell death by apoptosis. The molecular control of these processes, involving oncogenes and tumor suppressor genes, is being revealed by current research. Cancer cells often have defects within these control systems, and these defects may confer selective sensitivity to appropriately designed drug therapy. Panels of human tumor cell lines may serve to link the molecular defects with specific drug sensitivities. Such correlations could guide the selection of drugs for therapy based on molecular diagnosis of individual tumors.

Interstrand cross-linking of DNA by 1,3-bis(2-chloroethyl)-1-nitrosourea and other 1-(2-haloethyl)-1-nitrosoureas.

Bifunctional alkylating agents are known to cross-link DNA by simultaneously alkylating two guanine residues located on opposite strands. Despite this apparent requirement for bifunctionality, 1-(2-chloroethyl)-1-nitrosoureas bearing a single alkylating function were found to cross-link DNA in vitro. Cross-linking was demonstrated by showing inhibition of alkali-induced strand separation. Extensive cross-linking was observed in DNA treated with 1-(2-chloroethyl)-1-nitrosourea, 1,3-bis-(2-chloroethyl)-1-nitrosourea, and 1-(2-chloroethyl(-3-cyclohexyl-1-nitrosourea. The reaction occurs in two steps, an intital binding followed by a second step which can proceed after removal of unbound drug. It is suggested that the first step is chloroethylation of a nucleophilic site on one strand and that the second step involves displacement of Cl- by a nucleophilic site on the opposite strand, resulting in an ethyl bridge between the strands. Consistent with this possibility, 1-(2-fluoroethyl)-3-cyclohexyl-1-nitrosourea produced much less cross-linking, as expected from the known low activity of F-, compared with Cl-, as leaving group. 1-Methyl-1-nitrosourea, which is known to depurinate DNA, produced no detectable cross-linking.

Mechanisms of DNA strand breakage and interstrand cross-linking by diaziridinylbenzoquinone (diaziquone) in isolated nuclei from human cells.

AZQ had been found to produce DNA strand breaks and interstrand cross-links in intact cells; evidence had indicated that these two DNA lesions arise by different chemical mechanisms and vary independently in degree in different cell types. In the present work, the mechanisms of the production of DNA strand breaks and interstrand cross-links by AZQ were studied in isolated cell nuclei. This system avoided the problem of poor penetration of test substances into cells. The DNA lesions were measured by means of the alkaline elution technique. It was found that the production of DNA strand breaks by AZQ in isolated nuclei required the addition of a reducing agent such as NADPH and was almost completely prevented by superoxide dismutase. This indicates that the mechanism of DNA strand breakage involves transfer of an electron from a reduced form of AZQ to molecular oxygen. Unexpectedly, interstrand cross-linking also was enhanced greatly by previous reduction of AZQ by NADPH or NaBH4. However, this reaction was not inhibited by superoxide dismutase. General alkylating activity of AZQ also was stimulated by reduction; the pH-dependence of this reaction was determined. The mechanism of DNA interstrand cross-linking by AZQ was surmised to stem from alkylation reactions of the two aziridine groups. The findings suggest the possibility that AZQ or related compounds may function as bioreductive alkylating agents which might be selectively toxic to hypoxic tissues.

Effect of pH on the bleomycin-induced DNA single-strand scission in L1210 cells and the relation to cell survival.

The susceptibility of cultured L1210 cells to bleomycin was investigated as a function of pH of the medium, and was compared with DNA damage measured by alkaline elution. With increasing pH of the medium, both cytotoxicity and DNA damage increased. This was observed in the effects of bleomycin on cell proliferation, colony-forming ability, and DNA elutability. The reduction of colony-forming ability correlated with DNA single-strand scission, and this correlation was independent of pH.

DNA-protein cross-linking and DNA interstrand cross-linking by haloethylnitrosoureas in L1210 cells.

Bifunctional alkylating agents are known to produce cross-links between DNA and protein and between paired DNA strands. The possibility of discriminating these two classes of cross-links in L1210 cells treated with haloethylnitrosoureas or nitrogen mustard was explored with the alkaline elution technique. Two classes of cross-links were demonstrated, based on sensitivity to proteinase K; the proteinase-sensitive cross-links appear to be DNA-protein cross-links, and the proteinase-resistant class may include interstrand cross-links. Proteinase-sensitive cross-links form more rapidly than do proteinase-resistant cross-links in cells treated with chloroethylnitrosoureas, perhaps because these agents can chloroethylate protein sulfhydryl or amino groups followed by rapid reaction of these chloroethylated groups with DNA. Although both types of cross-links produced by nitrogen mustard disappeared or were repaired after 24 hr, the removal of cross-links produced by chloroethylnitrosoureas either did not occur or was incomplete in 24 hr. In addition to cross-links, cells treated with haloethylnitrosoureas exhibited DNA strand breaks; a method is suggested for estimating the apparent frequencies of strand breaks and cross-links in the DNA.

Transport and binding of mesoporphyrin IX by leukemia L1210 cells.

The combination of mesoporphyrin IX and light is cytotoxic to leukemia L1210 cells, resulting in membrane damage and loss of viability. In this study, mesoporphyrin transport and the cellular environment of accumulated drug were examined. The latter was characterized by measurements of absorbance and fluorescence spectra and of effects of irradiation on subsequent capacity of cells for transport of the nonmetabolized amino acid cycloleucine. We observed a rapid accumulation of drug at a relatively hydrophilic cellular environment (dielectric constant, 20) from which light-catalyzed inhibition of cycloleucine transport was clearly demonstrable. Washing at 37 degrees rapidly depleted this cellular region of drug. Longer incubations resulted in accumulation of porphyrin at more hydrophobic locci (dielectric constant, approximately 10) from which the drug was not readily washed and from which the efficiency of light-catalyzed damage to membrane transport was relatively low. These findings are consistent with the hydrophobic nature of mesoporphyrin (octanol:water partition ratio, 10).

DNA reactivity and in vitro cytotoxicity of the novel antitumor agent 1,5,2,4-dioxadithiepane-2,2,4,4-tetraoxide (NSC-348948) in human embryo cells.

1,5,2,4-Dioxadithiepane-2,2,4,4-tetraoxide (cyclic-SoSo) is structurally a novel synthetic compound but may functionally act as an alkylating agent. The effects of cyclic-SoSo on DNA were studied in IMR-90 and VA-13 human embryo cells by means of DNA alkaline elution analysis. In contrast to a number of bifunctional alkylating agents, cyclic-SoSo produced no DNA-DNA interstrand cross-links in either cell line, even at concentration which produced a greater than 3 log cell kill. At equimolar concentrations cyclic-SoSo induced DNA-protein cross-links in both cell lines to a similar extent. Frank DNA breaks and alkali-labile DNA lesions were detected in both cell lines. A greater quantity of strand breaks appeared in the IMR-90 than in the VA-13 cell line after exposure to cyclic-SoSo. However, cyclic-SoSo was more cytotoxic to the VA-13 cell line in vitro than to the IMR-90 cell line. Thus cyclic-SoSo may not be a typical bifunctional alkylating agent in that its mechanism of action does not appear to involve DNA interstrand cross-linking.

DNA damage and differential cytotoxicity produced in human cells by 2-chloroethyl (methylsulfonyl)methanesulfonate (NSC 338947), a new DNA-chloroethylating agent.

2-Chloroethyl (methylsulfonyl)methanesulfonate (CIEtSoSo) is of interest as a possible new chloroethylating agent because its simpler chemistry suggests that it will not generate the hydroxyethyl products which are produced by chloroethylnitrosoureas (CIEtNUs). The effects of CIEtSoSo on DNA were studied in IMR-90 and VA-13 human embryo cells by means of DNA alkaline elution analysis. IMR-90 are normal cells, whereas VA-13 cells are SV40 transformed and are deficient in DNA-guanine O6-alkyltransferase activity. The effects of CIEtSoSo were essentially the same as those of CIEtNUs in the following respects: DNA interstrand cross-links were produced in VA-13 cells but not in IMR-90 cells; the interstrand cross-links in VA-13 cells were formed after a delay of 6 to 12 h and appeared to undergo repair; DNA-protein cross-links were formed promptly in both cell types and appeared to be repaired; DNA strand breaks and alkali-labile lesions were produced and repaired, and differences were observed between the two cell types; and VA-13 cells were more sensitive than IMR-90 cells (dose modification factor, 5). The differential cytotoxicity against VA-13 cells was similar to that produced by noncarbamoylating CIEtNUs, and significantly larger than that produced by carbamoylating CIEtNUs. The results suggest that CIEtSoSo acts by chloroethylating guanine O6 positions in DNA.

Quenching of DNA:platinum(II) monoadducts as a possible mechanism of resistance to cis-diamminedichloroplatinum(II) in L1210 cells.

A line of mouse leukemia L1210 cells resistant to cis-diamminedichloroplatinum(II) (cis-DDP) was compared with its parent cell line in order to determine whether the sensitivity difference could be related to DNA interstrand cross-linking as measured by the alkaline elution technique. The study was stimulated by a previous finding that the magnitude of DNA interstrand cross-linking, although somewhat reduced in this resistant line, did not account for the relatively high degree of resistance. Therefore, the kinetics of cross-link formation and removal was studied. Cross-link removal rates were determined by the use of thiourea to stop the delayed formation of interstrand cross-links from cis-DDP:DNA monoadducts. There was no significant difference between the cross-link removal rates in the parent and resistant lines. Computer-analyzed kinetics was consistent with an enhanced cis-DDP:DNA monoadduct quenching mechanism in the resistant cells.

Effects of carbamoylation on cell survival and DNA repair in normal human embryo cells (IMR-90) treated with various 1-(2-chloroethyl)-1-nitrosoureas.

The possibility was examined that the carbamoylating activity of some chloroethylnitrosoureas could interfere with the activity of normal human cells to survive treatment with these drugs; 1-(2-chloroethyl)-3-(trans-4-hydroxycyclohexyl)-1-nitrosourea, which has strong carbamoylating activity, inhibited the rejoining of drug or X-ray-induced DNA strand breaks in IMR-90 cells, whereas the noncarbamoylating cis-2-hydroxy isomer had little or no effect; 1-(2-chloroethyl)-3-(trans-4-hydroxycyclohexyl)-1-nitrosourea was twice as potent as the cis-2-hydroxy isomer in reducing colony survival. The moderate or high carbamoylating drugs 1,3-bis(2-chloroethyl)-1-nitrosourea and 1-(2-chloroethyl)-3-(cyclohexyl)-1-nitrosourea had effects resembling those of 1-(2-chloroethyl)-3-(trans-4-hydroxycyclohexyl)-1-nitrosourea. The low carbamoylating drug 1-(2-chloroethyl)-3-(2,6-dioxo-1-piperidyl)-1-nitrosourea had effects resembling those of the cis-2-hydroxy isomer. 1-(2-chloroethyl)-1-nitrosourea, although a strong carbamoylator in chemical systems, behaved biologically as if it were a low carbamoylator. This can be rationalized on the basis of limited cellular uptake of cyanate ion. The results suggest that carbamoylation may inhibit the nucleotide excision repair of chloroethylnitrosourea-induced DNA damage that may be crucial to the ability of normal human cells to recover from the action of these drugs. Previous work has indicated that susceptible human tumor cells are sensitive to chloroethylnitrosoureas because of a lack of a DNA repair protein (guanine O6-alkyltransferase) that is not involved in nucleotide excision repair. On the basis of these findings and other evidence, further clinical trials of appropriate noncarbamoylating chloroethylnitrosoureas would be justified.