cDNA microarray analysis of multidrug resistance: doxorubicin selection produces multiple defects in apoptosis signaling pathways.

Doxorubicin plays an important role in the treatment of leukemias, lymphomas, and a variety of carcinomas. Tumor cell resistance to doxorubicin is often associated with expression of the multidrug resistance gene MDR1, which codes for the drug efflux pump P-glycoprotein, and a multidrug-resistant phenotype. Evidence from multiple sources suggests, however, that additional genes besides MDR1 are involved in development of multidrug resistance. To identify genes involved in the multidrug resistance phenotype, we created a 5760-gene cDNA microarray to search for differentially expressed genes between the human multiple myeloma cell line RPMI 8226 and its doxorubicin-selected sublines 8226/Dox6 and 8226/Dox40, both of which express MDR1 and are multidrug-resistant. The cDNA microarray results identified a set of differentially expressed genes, which included MDR1 as expected. Thirty Northern analyses were used to confirm the results of the cDNA microarrays; comparison with the microarray results showed a 90% agreement between the two techniques. Within the set of differentially expressed genes identified by the cDNA microarrays, 29 were of particular interest as they can participate in apoptotic signaling, particularly as mediated by ceramide and the mitochondrial permeability transition. The functional importance of these changes in gene expression is supported by their explanation of the 8226/Dox cell lines' cross-resistance to substances that are not P-glycoprotein substrates, such as Fas/CD95 ligand and staurosporine. We conclude that doxorubicin selection led to changes in gene expression that reduce the apoptotic response to death-inducing stimuli and thus contribute to the multidrug resistance phenotype.

Laminin-5-mediated gene expression in human prostate carcinoma cells.

Interactions between extracellular matrix (ECM) proteins and prostate carcinoma cells provide a dynamic model of prostate tumor progression. Previous work in our laboratory showed that laminin-5, an important member of a family of ECM glycoproteins expressed in the basal lamina, is lost in prostate carcinoma. Moreover, we showed that the receptor for laminin-5, the alpha6beta4 integrin, is altered in prostate tumors. However, the genes that laminin-5 potentially regulates and the significance of its loss of expression in prostate cancer are not known. We selected cDNA microarray as a comprehensive and systematic method for surveying and examining gene expression induced by laminin-5. To establish a definitive role for laminin-5 in prostate tumor progression and understand the significance of its loss of expression, we used a cDNA microarray containing 5289 human genes to detect perturbations of gene expression when DU145 prostate carcinoma cells interacted with purified laminin-5 after 0.5, 6, and 24 h. Triplicate experiments showed modulations of four, 61, and 14 genes at 0.5, 6, and 24 h, respectively. Genes associated with signal transduction, cell adhesion, the cell cycle, and cell structure were identified and validated by northern blot analysis. Protein expression was further assessed by immunohistochemistry. Mol. Carcinog. 30:119-129, 2001.

Cell line-directed screening assay for inhibitors of thioredoxin reductase signaling as potential anti-cancer drugs.

We have used a cell line-directed screening approach (CDSA) to identify novel inhibitors of the thioredoxin reductase signaling pathway which contributes to the transformed phenotype of some human tumors. Two 2-imidazolyl disulfide compounds, previously identified as inhibitors of thioredoxin reductase, were screened for growth inhibitory activity in the National Cancer Institute (NCI) human cancer cell line panel. The COMPARE pattern recognition algorithm was used to identify similar compounds from > 60,000 compounds in the NCI investigational drug database. Of 47 nondiscreet compounds tested in a thioredoxin reductase/thioredoxin insulin reduction assay, 37 (77%) were inhibitors with IC50s < or = 10 micrograms/ml and 15 of those (32%) had IC50s < or = 1 microgram/ml. These compounds were all as selective or more selective for thioredoxin reductase than for glutathione reductase, while three compounds were inhibitors of thioredoxin. In comparison to CDSA, the number of compounds with IC50s < or = 1 microgram/ml identified by screening of 52 compounds from the database whose growth inhibiting activity was unrelated to the activity of the disulfide compounds was only 2%. Screening of 221 randomly selected natural products gave only 3% of compounds with IC50s < or = 1 microgram/ml. Thus, the CDSA using data from the NCI cancer cell panel and known inhibitors of the selected target as seed compounds can greatly increase hit rates, compared with random screening, for identifying novel inhibitors of a target, in this case thioredoxin signaling.

Inhibition of the epidermal growth factor receptor tyrosine kinase by PD153035 in human A431 tumors in athymic nude mice.

PD153035 is a potent (Ki = 6 pm) and specific inhibitor of the epidermal growth factor (EGF) receptor tyrosine kinase that suppresses tyrosine phosphorylation of the EGF receptor in A431 cells at nanomolar concentrations in cell culture. We have examined the pharmacokinetics of this compound and its ability to rapidly suppress phosphorylation of the EGF receptor in A431 human epidermoid tumors grown as xenografts in immunodeficient nude mice. Following a single i.p. dose of 80 mg/kg, the drug levels in the plasma and tumor rose to 50 and 22 microM within 15 minutes. While the plasma levels of PD153035 fell below 1 microM by 3 hours, in the tumors it remained at micromolar concentrations for at least 12 hours. The tyrosine phosphorylation of the EGF receptor was rapidly suppressed by 80-90% in the tumors. However receptor phosphorylation returned to control levels after 3 hours despite the continued presence of the drug at concentrations which, based on previous in vitro results, were predicted to maintain inhibition. EGF-stimulated tyrosine kinase activity in tumor extracts was decreased and recovered in parallel with the effects of PD153035 on receptor phosphorylation though the activity had reached only about half of the control activity after three hours. These results demonstrate the potential for using small molecule inhibitors to inhibit the EGF receptor tyrosine kinase in vivo, though a fair evaluation of their potential anti-cancer activity will have to wait for solutions to problems with sustained delivery which may allow us to maintain suppression of EGF receptor phosphorylation.

Okadaic acid induces both augmentation and inhibition of beta 2-adrenergic stimulation of cAMP accumulation in S49 lymphoma cells.

To address the role of protein phosphatases in regulating hormonal responses in mammalian cells, we investigated the effects of okadaic acid, a potent inhibitor of protein phosphatases 1 and 2A, on epinephrine and prostaglandin E1 stimulation of cAMP accumulation and adenylyl cyclase in S49 WT and kin- lymphoma cells. Depending on the dose and time of okadaic acid pretreatment of both cell lines, there were two distinguishable effects on cAMP accumulation, an augmentation and an inhibition. The augmentation occurred rapidly (t1/2 < 1 min), was maximal with 3 microM okadaic acid, and was observed with concentrations of okadaic acid as low as 0.3 microM. Prolonged (t1/2 of 5-15 min) pretreatment of cells with okadaic acid caused an inhibition of epinephrine-stimulated cAMP accumulation, which was characterized by a 2-3-fold increase in the EC50 for the response to epinephrine. The EC50 for the okadaic acid-mediated inhibition was similar to that for the augmentation. In assays of adenylyl cyclase in membrane fractions prepared from okadaic acid-pretreated cells the inhibitory, but not the stimulatory, effects of okadaic acid pretreatment were observed. The data demonstrate that protein phosphatases play an important role in regulating adenylyl cyclase and suggest that cAMP-dependent protein kinase is not involved in either of its actions.

Cell-free heterologous desensitization of adenylyl cyclase in S49 lymphoma cell membranes mediated by cAMP-dependent protein kinase.

We have examined the cell-free heterologous desensitization of adenylyl cyclase in plasma membrane preparations from S49 wild-type (WT) and kin- cells (which lack cAMP-dependent protein kinase) incubated with purified catalytic subunit of cAMP-dependent protein kinase (cA.PKc). cA.PKc caused a rapid (t1/2 = 40 s) decrease in the hormone responsiveness of adenylyl cyclase in the WT membrane preparations that mimicked the intact cell heterologous desensitization; that is, there was an increase in the Kact for both epinephrine and prostaglandin E1 (PGE1) stimulations of adenylyl cyclase induced at the receptor level because neither forskolin- nor NaF-stimulated activity was affected. The desensitization was independent of agonist occupancy of the receptor, and the effects were blocked both by the active fragment (amino acids 5-22) of the specific inhibitor of cA.PK and by p[NH]ppA. cA.PKc treatment of kin- membranes resulted in a heterologous desensitization that resembled the effects of WT adenylyl cyclase, with the exception that forskolin-stimulated activity was also reproducibly decreased by 24%. cA.PKc had no effect on WT membranes isolated from cells that had previously undergone maximal heterologous desensitization during treatment with 10 microM forskolin. In contrast, cA.PKc-induced heterologous desensitization of kin- membranes was additive with the epinephrine-induced homologous desensitization of intact cells. Cell-free desensitizations were reversed by incubation of membranes with cA.PKc and ADP, conditions that drive the kinase reaction backward. The similarities of our cell-free cA.PKc-mediated heterologous desensitization of adenylyl cyclase with the intact cell desensitization support our hypothesis that heterologous desensitization of the WT lymphoma cells is mediated by cA.PK via a mechanism independent of homologous desensitization.

Activation of cAMP-dependent protein kinase is required for heterologous desensitization of adenylyl cyclase in S49 wild-type lymphoma cells.

We report here that, contrary to previously reported findings, treatment of S49 wild-type (WT) lymphoma cells with 0-50 nM epinephrine resulted in a heterologous desensitization of adenylyl cyclase (EC 4.6.1.1)--that is, epinephrine and prostaglandin E1 (PGE1) stimulations of adenylyl cyclase were reduced. Observation of this heterologous desensitization required the assay of adenylyl cyclase with submillimolar concentrations of Mg2+ and low concentrations of epinephrine. Also, whereas previously there had been no evidence for any role of cAMP-dependent protein kinase in the desensitization of the WT beta-adrenergic receptor, our data comparing the characteristics of the desensitization in WT, kin-, and cyc- lymphoma cells [where kin- and cyc- refer to variants of S49 WT cells lacking cAMP-dependent protein kinase activity (kin-) and the alpha subunit of the stimulatory guanine nucleotide-binding regulatory protein (cyc-)] now suggest that cAMP-dependent protein kinase mediates the heterologous desensitization of adenylyl cyclase. Specifically, we found that only the WT cells exhibited epinephrine-induced heterologous desensitization. The kin- and cyc- cells exhibited only homologous desensitization, and much higher concentrations of epinephrine were required to elicit the homologous desensitization in the variants relative to the heterologous desensitization of the WT. Treatment of WT and cyc- cells with dibutyryl cAMP or treatment of WT with forskolin or PGE1 caused the heterologous desensitization of adenylyl cyclase, indicating that neither receptor occupancy nor activation of adenylyl cyclase was necessary for the heterologous desensitization.

Beta-adrenergic receptor desensitization of wild-type but not cyc lymphoma cells unmasked by submillimolar Mg2+.

Treatment with low physiological concentrations of epinephrine (5-50 nM) rapidly desensitizes beta-adrenergic stimulation of cAMP formation in S49 wild-type (WT) lymphoma cells. Previous attempts to detect this early phase of desensitization in cell-free assays of adenylate cyclase (EC 4.6.1.1) after intact cell treatment were unsuccessful. We have now found that reducing the Mg2+ concentrations in the adenylate cyclase assays to less than 1.0 mM unmasked this rapid phase of desensitization of the WT cells, and that high Mg2+ concentrations (5-10 mM) largely obscured the desensitization. Submillimolar Mg2+ conditions also revealed a two- to threefold decrease in the affinity of epinephrine binding to the beta-adrenergic receptor after desensitization with 20 nM epinephrine. Detection of 4 beta-phorbol 12-myristate 13-acetate (PMA) desensitization of the WT beta-adrenergic receptor was also dependent on low Mg2+ as measured either by the decrease in epinephrine stimulation of adenylate cyclase or by the reduction in the affinity of epinephrine binding. Unexpectedly, when cyc- cells were pretreated with 50 nM epinephrine, the beta-adrenergic stimulation of reconstituted adenylate cyclase was not desensitized. The characteristics of the Mg2+ effect on epinephrine- and PMA-induced desensitizations suggest a similar mechanism of action with the most likely events being phosphorylations of the beta-adrenergic receptors. Our data indicate that cAMP-dependent protein kinase (EC 2.7.1.37) may play a role in the desensitization caused by low epinephrine concentrations inasmuch as this phase of desensitization did not occur in the cyc-. For the PMA-induced desensitization, the phosphorylation may be mediated by protein kinase C (EC 2.7.1.37).

Identification of an autostimulatory growth factor in the extracts and conditioned medium of GH3/C14 rat pituitary cells in culture.

An acid- and heat-stable low molecular weight (Mr 2,500-6,200) autostimulatory growth factor has been identified in extracts of GH3/C14 rat pituitary tumors, in extracts of these cells in culture and in their serum-free conditioned medium. The concentrations of autostimulatory activity required to achieve ED50 in serum-free culture were 4.3, 4.0 and 85 micrograms/ml for extracts of tumors, the cultured cells, and from the conditioned medium of the cells in culture, respectively. To characterize this activity, we have used an isolation method that included 0.1 M acetic acid extraction, heating at 95 C, SP-Sephadex chromatography, and finally, Sephadex G-50 chromatography. The identification of this activity if rapidly growing, relatively autonomous GH3/C14 tumors suggests an autostimulatory role in the conversion from estrogen-responsive growth to autonomy.