Cell-based screening approach for antitumor drug leads which exploits sensitivity differences between normal and cancer cells: identification of two novel cell-cycle inhibitors.

A cell-based in vitro screening approach for identification of antitumor drug leads that exploits the differential sensitivity between normal and cancer cells was developed. It is a three-step, high-throughput screen for antiproliferative and/or cytotoxic activity measured by a 7 day MTT [3-(4,5-dimethylthiazol-2yl)-2,5-diphenyl tetrazolium bromidel assay using small panels of proliferating primary human cells and established cancer cell lines. Proof-of-concept experiments successfully identified 11 known cancer drugs randomly mixed with 5000 test compounds. Application of this screening approach to a library of 110000 compounds allowed for the identification of several novel chemical classes of compounds active against an expanded panel of cancer cell lines in vitro. Two of the compounds representing novel mitotic inhibitors with in vivo potency against established breast cancer xenografts (MDA-MB-435) are reported here.

Quantification of telomerase activity by direct scintillation counting.

An improved telomerase assay was developed that allows direct quantification of the enzyme activity by scintillation counting of the labeled telomerase product. The assay measures the incorporation of 32P-dGTP into telomeric repeats synthesized at the 3' end of a biotinylated primer. Telomerase reaction product is separated from the reaction mix by streptavidin-coated magnetic beads and counted. The assay can be used for quantitative studies of human telomerase and its inhibitors.

Myeloid differentiation and retinoblastoma phosphorylation changes in HL-60 cells induced by retinoic acid receptor- and retinoid X receptor-selective retinoic acid analogs.

The ability of subtypes of retinoic acid receptors (RARs) and retinoid X receptors (RXRs) singly and in combination to elicit myeloid differentiation, G1/0-specific growth arrest, and retinoblastoma (RB) tumor suppressor protein dephosphorylation was determined in the human myeloblastic leukemia cell line HL-60 using subtype-selective retinoic acid (RA) analogs. RA analogs that selectively bind only to RARs (Am580 and/or TTNPB) or to RXRs (Ro 25-6603, SR11237, and/or SR11234) did not elicit the above-mentioned three cellular responses. In contrast, simultaneous treatment with both an RAR-selective ligand (Am580 or TTNPB) and an RXR-selective ligand (Ro 25-6603, SR11237, or SR11234) induced all three cellular processes. An RAR alpha-selective ligand used with an RXR-selective ligand generated the same responses as did all-trans RA or 9-cis RA, which affect both families of receptors, suggesting an important role for RAR alpha among RAR subtypes in eliciting cellular response. Consistent with this finding, the RAR alpha antagonist, Ro 41-5253, reduced the level of the cellular responses elicited by treatment with an RAR alpha-selective ligand plus RXR-selective ligand. The coupling of the shift of RB to its hypophosphorylated form with G1/0 arrest and differentiation in response to ligands is consistent with a possible role of RB as a downstream target or effector of RAR alpha and RXR in combination.

Conformationally defined 6-s-trans-retinoic acid analogs. 2. Selective agonists for nuclear receptor binding and transcriptional activity.

We recently demonstrated in animal models that a new conformationally defined RA isomer (Vaezi et al. J. Med. Chem. 1994, 37, 4499-4507) was as effective as RA in the prevention of skin papillomas but was less toxic. In order to provide more details concerning this improved action, we report here the preparation of a homologous conformationally defined 6-s-trans-retinoid (1) and investigate its ability to interact with proteins and to activate gene expression. Four configurational isomers of 1 were evaluated in binding assays for cellular retinoic acid binding protein, CRABP (isolated from chick skin); CRABP-I and CRABP-II (cloned from mouse); nuclear retinoic acid receptors (RARs); and nuclear retinoid X receptors (RXRs). In each assay the all-E-isomer of this retinoid had an activity that was comparable to that of (all-E)-RA. However, the 9Z-isomer was at least 200-fold less active than (all-E)-RA in binding to different RARs, while it was only 6-20 times less active than (9Z)-RA in binding to different RXRs. In an in vivo transient transfection assay, the all-E-isomer activated a reporter gene containing a retinoic acid response element (RARE) with efficiency similar to (all-E)-RA when expression vectors for either RAR alpha, RAR beta, RAR gamma alone or RAR alpha together with RXR alpha were cotransfected. In contrast, the 9Z-isomer was much less active than (9Z)-RA in the same assay systems. However, (9Z)-1 efficiently enhanced the DNA binding and transactivational activity of RXR alpha homodimers. Taken together, these studies demonstrate that the all-E- and 9Z-isomers of this retinoid are selective and potent agonists of RAR and RXR binding and activation.

Binding of 9-cis-retinoic acid and all-trans-retinoic acid to retinoic acid receptors alpha, beta, and gamma. Retinoic acid receptor gamma binds all-trans-retinoic acid preferentially over 9-cis-retinoic acid.

Both 9-cis-retinoic acid (RA) and all-trans-RA (t-RA) compete for [3H]9-cis-RA binding to RA receptors (RAR alpha, beta, and gamma) in nucleosol fractions from transiently transfected COS-1 cells with IC50 values of approximately 12 and 5 nM, respectively. Curiously, 9-cis-RA competes for [3H]t-RA binding to mouse RAR alpha, beta, and gamma with IC50 values of 31, 8, and 60 nM, respectively, while t-RA itself does not exhibit such differential competition (IC50 values for RARs, 5 nM). A similar pattern is observed with human retinoic acid receptors (RARs). Differential binding of 9-cis-RA to the RAR beta and gamma receptors is also found following in vitro transcription and translation of these receptors. Displacement assays demonstrate that t-RA exhibits similar off-rates for RAR alpha, beta, and gamma. However, 9-cis-RA is 6-fold more rapidly displaced from RAR gamma than from RAR beta. When RAR-transfected COS-1 cells are incubated with [3H]t-RA, [3H]-9-cis-RA or various mixtures of these two radioligands, high performance liquid chromatography analysis demonstrates that the ligands bound in nucleosol fractions from RAR beta-transfected cells reflect the isomer content of the media. However, in identical whole cell assays, nucleosol fractions from RAR gamma-transfected cells preferentially bind t-RA over 9-cis-RA, consistent with the in vitro data. These binding kinetics in vitro and in whole cells suggest that there could be differences in the interactions of the receptor subtypes with the endogenous retinoic acids under physiologic conditions.

Distinct binding determinants for 9-cis retinoic acid are located within AF-2 of retinoic acid receptor alpha.

Retinoids exert their physiological action by interacting with two families of nuclear receptors, the retinoic acid receptors (RARs) and the retinoid X receptors (RXRs), which regulate gene expression by forming transcriptionally active heterodimeric RAR/RXR or homodimeric RXR/RXR complexes on DNA. Retinoid receptor activity resides in several regions, including the DNA and ligand binding domains, a dimerization interface, and both a ligand-independent (AF-1) and a ligand-dependent (AF-2) transactivation function. While 9-cis retinoic acid (RA) alone is the cognate ligand for the RXRs, both 9-cis RA and all-trans RA (t-RA) compete for binding with high affinity to the RARs. This latter observation suggested to us that the two isomers may interact with a common binding site. Here we report that RAR alpha has two distinct but overlapping binding sites for 9-cis RA and t-RA. Truncation of a human RAR alpha to 419 amino acids yields a receptor that binds both t-RA and 9-cis RA with high affinity, but truncation to amino acid 404 yields a mutant receptor that binds only t-RA with high affinity. Remarkably, this region also defines a C-terminal boundary for AF-2, as addition of amino acids 405 to 419 restores receptor-mediated gene activity to a truncated human RAR alpha lacking this region. It is interesting to speculate that binding of retinoid stereoisomers to unique sites within an RAR may function with AF-2 to cause differential activation of retinoid-responsive gene pathways.

Retinoic acid receptors and retinoid X receptors: interactions with endogenous retinoic acids.

The binding of endogenous retinoids and stereoisomers of retinoic acid (RA) to the retinoid nuclear receptors, RA receptor (RARs) and retinoid X receptors (RXRs), was characterized using nucleosol preparations from transiently transfected COS-1 cells. Among several stereoisomers of RA tested, including 7-cis-, 9-cis-, 11-cis-, 13-cis-, and all-trans-RA, only 9-cis-RA effectively competes with 9-cis-[3H]RA binding to the RXRs. Additionally, the endogenous retinoid trans-didehydro-RA (t-ddRA) does not interact with RXRs, whereas the 9-cis form of ddRA competes effectively. RXRs (alpha, beta, and gamma) bind 9-cis-RA with dissociation constants (Kd) of 15.7, 18.3, and 14.1 nM, respectively. In contrast to the selectivity of RXRs for 9-cis-RA, RARs bind both t-RA and 9-cis-RA with high affinity, exhibiting Kd values in the 0.2-0.7 nM range for both ligands. Unlike RARs, the cellular RA binding proteins CRABPI or CRABPII bind t-RA but do not bind 9-cis-RA. Consistent with the binding data, 9-cis-RA and 9-cis-ddRA transcriptionally activate both GAL4-RXR and GAL4-RAR chimeric receptors with EC50 values of 3-20 nM for 9-cis-RA and 9-cis-ddRA, whereas t-RA and t-ddRA efficiently activate only GAL4-RAR chimeric receptors. Thus, 9-cis forms of endogenous retinoids can contribute to the pleiotropic effects of retinoids by interacting with both the RARs and RXRs.

9-cis retinoic acid stereoisomer binds and activates the nuclear receptor RXR alpha.

Vitamin A (retinol) and its natural derivatives are required for many physiological processes. The activity of retinoids is thought to be mediated by interactions with two subfamilies of nuclear retinoic acid receptors, RAR and RXR. The RARs bind all-trans retinoic acid (t-RA) with high affinity and alter gene expression as a consequence of this direct ligand interaction. RXR alpha is activated by t-RA, yet has little binding affinity for this ligand. t-RA may be converted to a more proximate ligand that directly binds and activates RXR alpha, and we have developed a method of nuclear receptor-dependent ligand trapping to test this hypothesis. Here we report the identification of a stereoisomer of retinoic acid, 9-cis retinoic acid, which directly binds and activates RXR alpha. These results suggest a new role for isomerization in the physiology of natural retinoids.

The effect of culture conditions and toxicity on the Ames Salmonella/microsome agar incorporation mutagenicity assay.

Various methods were used to reconstruct an environment similar to a chemically bactericidal or bacteriostatic effect in the Ames mutagenicity assay in order to determine whether a false-positive result could occur. Neither population reduction, sonication, nor increased histidine produced an increase of spurious revertant colonies.

Use of rat/hamster S-9 mixture in the Ames mutagenicity assay.

Based on the findings of Nagao et al [1978] that phenacetin is negative in the standard Ames test with Aroclor induced rat S-9 and positive with hamster S-9, the test was performed with a mixture of rat/hamster S-9. Phenacetin was mutagenic with the mixture. The activity of the mixture was compared to the rat S-9 alone with low concentrations of 2-aminoanthracene (a strong promutagen for Salmonella typhimurium TA 1535, TA 100, TA 1537, and TA 98), nitrosopiperidine (a weak promutagen), and 1,2 epoxybutane (a weak, direct-acting mutagen). Except for an increased mutagenic activation by the mixture with nitrosopiperidine the mixture was comparable to the rat S-9 alone, indicating that replacing rat S-9 with a rat/hamster S9 mixture in the standard Ames test could increase the sensitivity of the test without interfering with rat S-9 activity.

An analysis of weak mutagens in the Ames assay.

Nitrosopiperidine, sodium nitrite and 1,2 epoxybutane were tested in the Ames agar incorporation assay in an attempt to establish exact criteria for detecting the activity of these weak mutagens. As regards minimum concentrations it was determined that at 500 microgram per plate a statistical analysis and several concentrations of the activating system are required in order to manifest an indication of the mutagenicity of 1,2 epoxybutane. The mutagenicity of sodium nitrite was detectable at 80 microgram/plate while nitrosopiperidine produced a 2.5-3.2 increase over background at 250 microgram/plate. It is concluded that since weak mutagenicity cannot be predicted it is advisable to utilize the maximum concentrations possible within the limits of solubility and bacterial toxicity. At the higher concentrations 1,2 epoxybutane, nitrosopiperidine and sodium nitrite produced revertant colony counts equal to 7, 17 and 62 times above background, respectively.

Chromosomal effects of carcinogens and non-carcinogens on WI-38 after short term exposures with and without metabolic activation.

The human diploid fibroblast culture, WI-38 was analyzed for chromosomal damage after 24 h exposures to benzo(a)pyrene (BP), 3-methylcholanthrene (MCA), n-methyl-n'-nitrosoguanidine (MNNG), 4-nitroquinoline-1-oxide (4NQO), pyrene and caffeine. A low concentration of 4NQO (0.15 micron) and MNNG (1.9 micron) produced breakage and exchange figures. A relatively high concentration of caffeine (1300 micron) caused breakage. The other compounds (BP, MCA and pyrene) caused little or no increase in damage above the control levels. A 1-h pulse exposure of WI-38 cells to BP (40 micron) in the presence of a rat liver homogenate supernate (S-9) resulted in damage significantly greater than the untreated cells or cells treated with BP alone. 4NQO (0.25 micron) produced exchange figures after a similar 1-h exposure, but this effect was eliminated by the S-9. A much higher concentration of caffeine (10,300 micron) was required to cause breakage greater than control levels after a one hour exposure. The results indicate a possible short term in vitro human cell system for distinguishing carcinogens, procarcinogens, and noncarcinogens.

The effect of methylxanthines on chromosomes of human lyphocytes in culture.

The effect of caffeine (I,3,7-trimethylxanthine), theophylline (I,3-dimethylxanthine), theobromine (3,7-dimethylxanthine), paraxanthine (I,7-dimethylxanthine) I-methylxanthine, 3-methylxanthine, and 7-methylxanthine added at the 48th h on the chromosomes of human lymphocytes in 72-h cultures has been investigated. Caffeine and the dimethylxanthines cause breakage at 750 mug/ml, with caffeine the most, and paraxanthine the least clastogenic. I-Methylxanthine and dimethylxanthines with a methyl group in the I-position are the most effective in depressing mitotic indices. No chromosome damage was exhibited by the monomethylxanthines.