Bifunctional DNA alkylator 1,3-bis(2-chloroethyl)-1-nitrosourea activates the ATR-Chk1 pathway independently of the mismatch repair pathway.

The presence of DNA damage initiates signaling through the ataxia-telangiectasia mutated kinase (ATM) and the ATM- and the Rad3-related kinase (ATR), which phosphorylate, thus activating, the checkpoint kinases (Chk) 1 and 2, which leads to cell cycle arrest. The bifunctional DNA alkylator 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) is cytotoxic primarily by inducing DNA monoadducts and ultimately, interstrand cross-links, which block DNA replication. In this study, we investigated the activation of the ATR-Chk1 pathway in response to BCNU treatment and the dependence of this response on the DNA mismatch repair (MMR) capacity. Medulloblastoma cells were exposed to low and moderate doses of BCNU, and the effects on this DNA damage signaling pathway were examined. In response to BCNU, Chk1 was found to be phosphorylated at serine 345 and exhibited increased kinase activity. Caffeine and wortmannin, which are broad-spectrum inhibitors of ATM and ATR, reduced this phosphorylation. Cell cycle analysis further revealed an accumulation of cells in the S phase in response to BCNU, an effect that was attenuated by caffeine. Small interfering RNA knockdown of ATR also reduced Chk1 phosphorylation after exposure to BCNU. However, knockdown of ATM had no effect on the observed Chk1 phosphorylation, suggesting that ATR was primarily responsible for Chk1 activation. Analysis of Chk1 activation in cells deficient in MMR proteins MutLalpha or MutSalpha indicated that the DNA damage response induced by BCNU was independent of the MMR apparatus. This MMR-independent activation seems to be the result of DNA interstrand cross-link formation.

Quenching of DNA cross-link precursors of chloroethylnitrosoureas and attenuation of DNA interstrand cross-linking by glutathione.

Interstrand DNA cross-linking is essential for the antitumor activity of chloroethylnitrosoureas (CENUs). The critical cross-links have been proposed to involve a rapid O6-guanine chloroethylation on one DNA strand, followed by a rearrangement of the O6-(2-chloroethyl)guanine and slow alkylation of the second DNA strand. In view of the relative intracellular abundance of glutathione (GSH) and nucleophilicity of its thiolate ion, the ability of GSH to react with and to inactivate 2-chlorethylated DNA and the possibility that this interaction decreases net DNA cross-linking by CENUs were investigated. Chloroethylated calf thymus DNA was reacted with GSH, the DNA was precipitated and redissolved, and subsequent DNA interstrand cross-linking was determined. The DNA cross-link index was compared for both GSH-treated and 2-chloroethylated untreated DNA. Simultaneously, Col E1 plasmid DNA was chloroethylated and reacted with GSH, and the extent of DNA interstrand cross-linking was determined by agarose gel electrophoresis and compared with controls. The results show both a time- and GSH concentration-dependent quenching of chloroethylated DNA, with a corresponding decrease in the DNA cross-link index. Using [methyl-3H]GSH, it was also demonstrated that 56% of the total GSH was bound to quenched 2-chloroethylated Col E1 DNA and 25% to quenched 2-chloroethylated calf thymus DNA. GSH binding to cross-linked DNA and native DNA was insignificant. It is concluded that, in addition to direct inactivation of reactive cytotoxic CENU species, GSH may also modulate cellular response to CENUs by quenching chloroethylated DNA, thereby decreasing the formation of potentially lethal DNA cross-links.

Optimization and characterization of the capillary human tumor clonogenic cell assay.

The capillary human tumor clonogenic cell assay (HTCA) has been shown to have important advantages over conventional HTCAs. In the present report, this promising novel HTCA was further optimized and characterized using 46 primary human tumor specimens, 6 human tumor cell lines (1 astrocytoma, 2 colon carcinomas, 1 melanoma), and 2 murine leukemias. Hydrocortisone, epidermal growth factor, heat-inactivated fetal calf serum, and horse serum were investigated for their ability to modulate tumor colony formation in the assay. Critical assay parameters that can affect tumor colony formation, namely, cell seeding density, agarose concentration, culture volume, capillary tube geometry, and capillary tube sealing, were also investigated. The results showed that serum (optimum concentration, 20%) was obligatory for tumor colony formation, and that both epidermal growth factor (50 ng/ml) and hydrocortisone (2.5 ng/ml), although supportive of colony growth, were not absolute requirements. Plating at 2.5-3 x 10(5) cells/ml in a culture volume of 50 microliters/capillary tube and an agarose concentration of 0.2% optimized colony formation (number, size, and distribution of colonies along the capillary tube) by primary human tumor cells. The cell lines generally formed colonies best at lower seeding densities and in lower culture volumes (30 microliters/tube). Colony formation was significantly better in unsealed than in sealed capillary tubes and growth was just as good, and in some cases, better in round capillary tubes than in square ones. Using ovarian carcinoma cells, the Cellscan prototype system was demonstrated as feasible for automated counting and evaluation of tumor colony growth in capillary tubes. A comparison of the capillary HTCA and the agar double-layer assay in Petri dishes produced a median plating efficiency of 0.18 for the capillary HTCA and 0.036 for the Petri dish method. The overall success rate was 77% for the former and 53% for the latter assay.

Decreased DNA interstrand cross-linking and cytotoxicity induced in human brain tumor cells by 1,3-bis(2-chloroethyl)-1-nitrosourea after in vitro reaction with glutathione.

Although both direct and glutathione S-transferase (GST)-catalyzed interactions between many electrophiles and GSH generally result in inactivation of the former, there are several reports of compounds whose electrophilic, alkylating, and cytotoxic activities are potentiated by GSH. This study investigates the effects of direct in vitro interaction between GSH and BCNU at physiological pH (7.2) and temperature (37 degrees C) and how this affects the cytotoxic and DNA cross-linking activity of 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) in target human malignant brain tumor cells. The kinetics and dose-response relationship of this interaction were determined by measuring residual GSH and residual BCNU-cytotoxicity in aGSH/BCNU mixture over a 45-min period and at varying BCNU concentrations. The results demonstrate that reaction of BCNU with four times its molar concentration of GSH for 45 min significantly inactivates BCNU, as expressed by a 32% decrease in induction of cellular DNA cross-linking, a 21% increase in DNA synthesis, and a 15% increase in clonogenic survival of human brain tumor cells compared to incubates of BCNU alone. Equine liver (EL)-GST increased the inactivation of BCNU only slightly (insignificant at p = 0.05). These results suggest that, in contrast to agents such as the alkyl-N-nitro-N'-nitrosoguanidines which become more potent alkylators after reacting with GSH, the 2-chloroethylnitrosoureas (CENUs) undergo inactivation by GSH. We propose that such interactions between GSH and the CENUs may constitute an important aspect of CENU metabolism and provide a potential means by which brain tumor cells can circumvent CENU toxicity and exhibit resistance to this class of agents.

Glutathione content and glutathione-S-transferase expression in 1,3-bis(2-chloroethyl)-1-nitrosourea-resistant human malignant astrocytoma cell lines.

gamma-L-glutamyl-L-cysteinylglycine (GSH) has been shown to inactivate 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) and quench DNA crosslink precursors of BCNU. Because of the central role of 2-chloroethyl-nitrosoureas in brain tumor chemotherapy, we investigated the intracellular GSH content and the expression of specific glutathione-S-transferases (GSTs) in three human malignant astrocytoma cell lines (UWR1, UWR2, and UWR3) of varying BCNU resistance to determine the interrelationship of these parameters with brain tumor BCNU resistance. GSH was assayed by ion-exchange high performance liquid chromatography after derivatization with 1-fluoro-2,4 dinitrobenzene. Both bulk and specific GST (acid, near-neutral, and basic) activities were examined using substrates that show high specificities to the different GSTs. Western blot analyses with antisera against GST-alpha, -mu, and -tau subunits were also performed on partially purified GST from the cells of each cell line. The results showed GSH content of 91, 46.5, and 28.3 nmol GSH/mg protein for UWR1, UWR2, and UWR3, respectively. Bulk GST activity (with 1-chloro-2,4-dinitrobenzene as substrate) also correlated with increasing BCNU resistance. Of the three GST classes examined by both substrate specificities and Western blotting, only the expression of the acidic form, GST-tau, correlated significantly with the rank order of BCNU resistance of the cell lines. GST-mu and -alpha were present in only trace amounts in all three cell lines.

Application of in vivo and in vitro pharmacokinetics for physiologically relevant drug exposure in a human tumor clonogenic cell assay.

The biological half-lives and decay rate constants under the conditions of a human brain tumor clonogenic cell assay were determined for six clinically used anticancer agents. The agents studied were: 1,3-bis(2-chloroethyl)-1-nitrosourea; 3-(2-chloroethyl-3-nitrosoureido-2-deoxy-D-glucopyranose; cis-diaminedichloroplatinum(II); 2,5-diaziridinyl-3,6-bis-(carboethoxyamino)-1,4-benzoquinone; 4-demethylepipodophylotoxin-D-thylidene glucoside; and 9-hydroxy-2-N-methylellipticine. In vitro decay of all six drugs was found to be according to first order kinetics. The half-lives of two drugs, namely, 1,3-bis(2-chloroethyl-1-nitrosourea and 3-(2-chloroethyl-3-nitrosoureido-2-deoxy-D-glucopyranose under the human tumor clonogenic cell assay (HTCA) conditions were found to be similar to their terminal in vivo half-lives in humans. For the other drugs, however, there was a very large difference between their in vitro and in vivo pharmacokinetics. In the case of 2,5-diaziridinyl-3,6-bis(carboethoxyamine)-1,4-benzoquinone, we observed about an 80-fold difference between its in vitro half-life of 40.76 h and its in vivo terminal half-life of 0.52 h. We describe the principles upon which these data can be used to design clinically more relevant in vitro drug exposure protocols in HTCAs. Since, generally, tumor cells are exposed to drugs in the HTCA either continuously or for a specified duration, e.g., 1 or 2 h, we computed the initial in vitro drug concentrations to which tumor cells should be exposed such that the resulting in vitro (c X t) after a 2-h or a continuous exposure will be within clinically achievable levels. The application of these in vivo and in vitro pharmacokinetic principles will provide for more physiological testing of patient tumor cell sensitivity to anticancer drugs in the HTCA, and is likely to result in lower rates of false positive responses in clinical trials using clonogenic cell assays.

Simultaneous amplification of four DNA repair genes and beta-actin in human lymphocytes by multiplex reverse transcriptase-PCR.

We describe here the development, optimization, and use of a non-radioactive, quantitative, multiplex reverse transcriptase-PCR technique to measure, in a single reaction, the relative levels of the transcripts of four DNA repair genes (XPCC, hMSH2, XRCC1, and ERCC1) and the beta-actin gene in lymphoblastoid cell lines and frozen peripheral blood lymphocytes. Expression of defective DNA repair genes was not detected in DNA repair-deficient human cell lines, whereas the intact genes were detected in repair-proficient cell lines and in lymphocytes from a normal donor. The assay was reproducible, and repeated determinations of the same samples generated highly consistent results for each target gene. This approach should facilitate molecular epidemiological studies that incorporate screening for germline alterations that may affect gene expression and for changes in the levels of gene expression.

Potential antiglioma activity of 9-hydroxy-2-N-methylellipticine as determined by pharmacological and human tumor clonogenic cell studies.

The antiglioma activity of elliptinium (HME) was investigated in a human glioma clonogenic cell assay. Early passage cells of three human glioma cell lines (SF126, SF375, and SF407) were exposed to HME at the clinically achievable dose of 3 microM for 3 h. At this HME concentration, clonogenic cell survival was reduced by more than 3 logs in SF126 and SF375, and by 0.8 logs in SF407. A study of the kinetics of cell kill showed that whereas at moderate (less than or equal to 1.5 microM) HME doses cell kill increased with treatment time up to a maximum at approximately 3 h, cytotoxicity was more dose than time dependent at higher doses. Flash treatment of SF375 cells with 3 microM HME resulted in more than 2 logs clonogenic cell kill. Using high-pressure liquid chromatography, we investigated the in vitro decay kinetics of HME under our in vitro drug treatment conditions and observed a very rapid, protein nondependent 40% drop in HME concentration which was dose dependent and was probably due to HME adsorption on the surface of tissue culture plasticware. Subsequent decay of the drug was very slow, with a decay rate constant of 0.022/h and a half-life of 298 h. In order to determine whether HME crosses the blood-brain barrier, we measured the rat brain capillary permeability coefficient, P, of [3H]HME and [14C]HME. The mean P value of 2.2 X 10(-6) cm/s +/- 16% (SD) suggests that HME crosses the blood-brain barrier (t 1/2 = 46 min) consistent with its molecular size and octanol-water partition coefficient.

Topoisomerase II inhibition and altered kinetics of formation and repair of nitrosourea and cisplatin-induced DNA interstrand cross-links and cytotoxicity in human glioblastoma cells.

By altering the accessibility of DNA sequences for alkylation or platination, and/or for subsequent repair, topoisomerase II can potentially affect the level of DNA interstrand cross-links induced in cells by bifunctional agents. In this study, we investigated the extent to which inhibition of topoisomerase II activity in a human glioblastoma multiforme cell line alters the kinetics of both the formation and the repair of total genomic DNA interstrand cross-links, as well as the sensitivity of the tumor cells to cis-diamminedichloroplatinum II (cis-DDP) and 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU). Cells were incubated with and without 200 microM novobiocin, a known topoisomerase II inhibitor, for 24 h, followed by exposure to 50 microM BCNU and 25 microM cis-DDP. DNA interstrand cross-linking was determined at various time points over 72 h, using a modified ethidium bromide-DNA binding assay. Sensitivity of the cells to cis-DDP and BCNU was also determined with and without novobiocin pretreatment with 200 microM novobiocin. This concentration of novobiocin showed no significant direct cytotoxicity, although it inhibited topoisomerase II activity in tumor cell nuclear extracts by 73%. A significant decrease in the rate of repair of both cis-DDP and BCNU induced DNA interstrand cross-links, with a corresponding decrease in the clonogenic survival of the cells, was observed following novobiocin exposure. Although the peak cross-link indices of novobiocin-treated cells relative to controls were not significantly increased, residual DNA cross-linking in the cells after 72 h was increased by 1.4-fold for BCNU and 3-fold for cells treated with cis-DDP, thus, indicating a greater effect of topoisomerase II on cross-link repair than on cross-link formation. These data suggest that inhibition of topoisomerase II may provide a potentially effective clinical strategy for sensitizing human brain tumors, and possibly other tumors as well, to DNA cross-linking anticancer agents.

Modulation of in vitro invasion of human glioblastoma cells by urokinase-type plasminogen activator receptor antibody.

Four human glioblastoma cell lines (U251, UWR1, UWR2, and UWR3) were tested for the expression of the cell surface receptor for urokinase-type plasminogen activator (uPA). To our knowledge there have been no previous reports about the uPA receptors (uPARs) in glioblastoma cell lines. All four glioblastoma cell lines we tested were found to bind recombinant Pro-uPA saturably and reversibly. Scatchard analysis of radioligand binding with acid-pretreated cells showed the presence of a single population of high-affinity uPARs on glioblastoma cells. Northern blot analysis confirmed that glioblastoma cells like other human cell lines express a 1.4-kilobase uPAR mRNA and 2.4-kilobase uPA mRNA. The significance of the uPAR in the invasive potential of the cells was examined by incubating uPAR antibody in an in vitro invasion assay. The anti-uPAR monoclonal antibody blocked the invasion effectively in a Matrigel assay, in which inhibition of invasion ranged between 20 and 57% for the cells studied. These data suggest that the uPARs contribute significantly to the invasive capacity of the cells, possibly by facilitating uPA activity.

Protein phosphorylation is a regulatory mechanism for O6-alkylguanine-DNA alkyltransferase in human brain tumor cells.

The biochemical regulation of human O6-alkylguanine-DNA alkyltransferase (AGT), which determines the susceptibility of normal tissues to methylating carcinogens and resistance of tumor cells to many alkylating agents, is poorly understood. We investigated the regulation of AGT by protein phosphorylation in a human medulloblastoma cell line. Incubation of cell extracts with [gamma-32P]ATP resulted in Mg(2+)-dependent phosphorylation of the endogenous AGT. Immunoprecipitation after exposure of the cells to 32P-labeled inorganic phosphate showed that AGT exists as a phosphoprotein under physiological conditions. Western analysis and chemical stability studies showed the AGT protein to be phosphorylated at tyrosine, threonine, and serine residues. Purified protein kinase A (PKA), casein kinase II (CK II), and protein kinase C (PKC) phosphorylated the recombinant AGT protein with a stoichiometry of 0.15, 0.28, and 0.44 (mol phosphate incorporated/mol protein), respectively. Residual phosphorylation of the endogenous AGT by the PKs present in cell homogenates and phosphorylation of the recombinant AGT by purified serine/threonine kinases, PKA, PKC, and CK II reduced AGT activity by 30-65%. Conversely, dephosphorylation of cell extracts by alkaline phosphatases stimulated AGT activity. We also identified consensus phosphorylation motifs for many cellular kinases, including PKA and CK II in the AGT protein. These data provide the first and conclusive evidence of AGT phosphorylation and suggest that reversible phosphorylation may control the activity of this therapeutically important DNA repair protein in human normal and cancer cells.

Chemical structure of carbamoylating groups and their relationship to bone marrow toxicity and antiglioma activity of bifunctionally alkylating and carbamoylating nitrosoureas.

Although the antitumor effects of chloroethylnitrosoureas have been shown to be due primarily to DNA-DNA cross-linking by the alkylating moieties of these agents, the basis of the often accompanying bone marrow toxicity has been more controversial. We report on the relative bone marrow toxicity of four model nitrosoureas with different alkylating and carbamoylating activities: 1,3-bis(2-chloroethyl)-1-nitrosourea; 1,3-bis(trans-4-hydroxycyclohexyl)-1-nitrosourea; chlorozotozin, (2-[3-(2-chloroethyl)-3 -nitrosoureido]-2-deoxy-D-glucopyranose); and -3-(beta-D-glucopyranosyl)-1-nitrosourea. Inhibitions of DNA, RNA, and protein synthesis in murine bone marrow cells and of colony growth of myeloid precursor cells (granulocyte-macrophage colony-forming units) were used as in vitro end points of myelotoxicity. Further, we determined the antiglioma activity of the four nitrosoureas on two human gliomas in a clonogenic tumor cell assay and studied the effect of the non-nitrosourea carbamoylators potassium cyanate, chloroethyl isocyanate, cyclohexyl isocyanate, ethyl isocyanate, and ethyl isothiocyanate on granulocyte-macrophage colony-forming units. The results show that, at equivalent drug exposures, clonogenic glioma cell kill was significant and comparative for 1,3-bis(2-chloroethyl)-1-nitrosourea, 1-(2-chloroethyl)-3-(beta-D-glucopyranosyl)-1-nitrosourea, and chlorozotocin; 1,3-bis(trans-4-hydroxycyclohexyl)-1-nitrosourea showed little activity. In contrast, granulocyte-macrophage colony-forming unit toxicity was low with chlorozotocin and 1-(2-chloroethyl)-3-(beta-D-glucopyranosyl)-1-nitrosourea and very high with 1,3-bis(2-chloroethyl)-1-nitrosourea and 1,3-bis(trans-4-hydroxycyclohexyl)-1-nitrosourea. Of the isocyanates, bone marrow toxicity was highest with chloroethyl isocyanate and cyclohexyl isocyanate, intermediate with ethyl isocyanate, and lowest with KOCN and ethyl isothiocyanate. Our results indicate that (a) bifunctional alkylation is essential for antiglioma activity of nitrosoureas and (b) myelosuppression is at least partly linked with carbamoylation but that structural entities in the carbamoylating isocyanate rather than a quantitative degree of carbamoylation determine the degree of potential myelotoxicity.

Elevated DNA polymerase alpha, DNA polymerase beta, and DNA topoisomerase II in a melphalan-resistant rhabdomyosarcoma xenograft that is cross-resistant to nitrosoureas and topotecan.

Previous investigations have revealed that the human TE-671 MR human rhabdomyosarcoma xenograft selected in vivo for melphalan resistance (M. C. Rosenberg, et al., Cancer Res., 49: 6917-6922, 1989) is cross-resistant to a wide variety of alkylating agents and to bleomycin, but is collaterally sensitive to etoposide. Although glutathione levels were noted to be elevated in TE-671 MR compared to the melphalan-sensitive parental TE-671 xenograft, treatment with buthionine sulfoximine to deplete glutathione levels did not fully restore melphalan sensitivity in the TE-671 MR xenograft. The present studies were undertaken to search for additional mechanisms of resistance in the TE-671 MR xenograft. Drug sensitivity testing performed at the dose of agents that was lethal to 10% of the animals revealed that the TE-671 MR xenograft maintained resistance to the bifunctional cross-linking agent 1,3-bis(2-chloroethyl)-1-nitrosourea and was cross-resistant to the topoisomerase I poison topotecan. Treatment with buthionine sulfoximine did not sensitize the TE-671 MR xenograft to 1,3-bis(2-chloroethyl)-1-nitrosourea. Further, even though O6-alkylguanine-DNA alkyltransferase levels were high in both the TE-671 and TE-671 MR xenografts, depletion of O6-alkylguanine-DNA alkyltransferase activity by treatment with O6-benzylguanine substantially sensitized the TE-671 xenografts but not the TE-671 MR xenografts, suggesting an additional mechanism of resistance. Measurement of additional enzyme activities that might be involved in DNA repair revealed significant elevations in DNA polymerase alpha (46 +/- 8 (SD) units/mg protein in TE-671, 69 +/- 6 units/mg protein in TE-671 MR, P < 0.05) and DNA polymerase beta (0.43 +/- 0.01 units/mg protein in TE-671, 0.78 +/- 0.12 units/mg protein in TE-671 MR, P < 0.05) but not DNA polymerase delta or total DNA ligase. Examination of topoisomerases by activity assays and Western blotting revealed a 2-fold increase in topoisomerase II and a 2-fold decrease in topoisomerase I in the TE-671 MR xenograft compared to the parental xenograft, apparently explaining the collateral sensitivity to etoposide and cross-resistance to topotecan. These results suggest that TE-671 MR xenografts contain multiple changes in activities of DNA repair-related proteins and other nuclear proteins that could contribute to alkylating agent resistance.

Selective suppression of matrix metalloproteinase-9 in human glioblastoma cells by antisense gene transfer impairs glioblastoma cell invasion.

Increased expression of matrix metalloproteinases (MMPs) has been associated with human glioblastoma tumor progression. In this study, we sought to down-regulate MMP-9 expression by stably transfecting a high-grade glioblastoma cell line with a plasmid vector capable of expressing an antisense transcript complementary to a 528-bp segment at the 5' end of human MMP-9 cDNA. Stable transfectants were obtained through selection with G418. Of the clones transfected with vector, sense, and antisense constructs, Northern blotting, Western blotting, and gelatin zymography showed that MMP-9 expression was significantly reduced only in the antisense-transfected cells. A Matrigel invasion assay revealed marked reductions in invasiveness for the antisense clones relative to the parental, vector, and sense clones. Cocultures of tumor spheroids and fetal rat brain aggregates showed that the antisense-transfected stable clones showed no invasion of the rat brain aggregates; in contrast, 90% of the parental, vector, and sense clones invaded the rat brain aggregates. Intracerebral injection of antisense stable transfectants in nude mice produced no tumors or very small tumors, but intracerebral injection of parental or vector clones did produce tumors. These results suggest that MMP-9 expression is essential for the invasiveness of glioblastoma cells.

Methylator resistance mediated by mismatch repair deficiency in a glioblastoma multiforme xenograft.

A methylator-resistant human glioblastoma multiforme xenograft, D-245 MG (PR), in athymic nude mice was established by serially treating the parent xenograft D-245 MG with procarbazine. D-245 MG xenografts were sensitive to procarbazine, temozolomide, N-methyl-N-nitrosourea, 1,3-bis(2-chloroethyl)-1-nitrosourea, 9-aminocamptothecin, topotecan, CPT-11, cyclophosphamide, and busulfan. D-245 MG (PR) xenografts were resistant to procarbazine, temozolomide, N-methyl-N-nitrosourea, and busulfan, but they were sensitive to the other agents. Both D-245 MG and D-245 MG (PR) xenografts displayed no O6-alkylguanine-DNA alkyltransferase activity, and their levels of glutathione and glutathione-S-transferase were similar. D-245 MG xenografts expressed the human mismatch repair proteins hMSH2 and hMLH1, whereas D-245 MG (PR) expressed hMLH1 but not hMSH2.

Deletions and rearrangements inactivate the p16INK4 gene in human glioma cells.

Structural alterations in the p16INK4 gene were examined in early passage human glioma cell lines and related to the expression of p16 transcripts and protein. Using the Southern blot approach, we observed both homozygous and hemizygous deletions, as well as rearrangements of the p16 and p15 genes in 5 of the 7 cell lines (71%). Two cell lines, MGR3 and HBT28, revealed hemizygous deletion of the p16 and p15 genes combined with indistinguishable rearrangements of the remaining p15-p16 locus that resulted in loss of exon 2 sequences for p15 and p16, but retention of p16 exon 1; neither of these cell lines expressed p16 mRNA. Data for a third cell line, MGR2, indicated a similar, but unique rearrangement involving the p15 and p16 genes. MGR2, which retained a single wild-type p15-p16 locus, showed expression of p16 transcript, but not of p16 protein as indicated by Western blot analysis. All the glioma cell lines expressed similar levels of the retinoblastoma protein and no amplification of the cyclin-dependent kinase 4 gene. These results demonstrate that human glioma cells contain p16 gene microdeletions and rearrangements that contribute to inactivation of the cell cycle regulatory protein.

Down-regulation of cathepsin B expression impairs the invasive and tumorigenic potential of human glioblastoma cells.

Increases in abundance of cathepsin B transcript and protein correlate with increases in tumor grade and alterations in subcellular localization and activity of cathepsin B. The enzyme is able to degrade the components of the extracellular matrix (ECM) and activate other proteases capable of degrading ECM. To investigate the role played by this protease in the invasion of brain tumor cells, we transfected SNB19 human glioblastoma cells with a plasmid containing cathepsin B cDNA in antisense orientation. Control cells were transfected with vector alone. Clones expressing antisense cathepsin B cDNA exhibited significant reductions in cathepsin B mRNA, enzyme activity and protein compared to controls. Matrigel Invasion assay showed that the antisense-transfected cells had a markedly diminished invasiveness compared with controls. When tumor spheroids containing antisense transfected SNB19 cells expressing reduced cathepsin B were co-cultured with fetal rat brain aggregates, invasion of fetal rat brain aggregates was significantly reduced. Green Fluorescent Protein (GFP) expressing parental cells and antisense transfectants were generated for detection in mouse brain tissue without any post-chemical treatment. Intracerebral injection of SNB19 stable antisense transfectants resulted in reduced tumor formation in nude mice. These results strongly support a role for cathepsin B in the invasiveness of human glioblastoma cells and suggest cathepsin B antisense may prove useful in cancer therapy.

Prognostic significance of glutathione S-transferase pi expression and subcellular localization in human gliomas.

The glutathione S-transferase (GST)-pi gene is overexpressed in many human cancers and preneoplastic lesions and is associated with failure of cancer chemotherapy and poor patient survival. Although GST-pi overexpression in tumors of the central nervous system has been observed, the prognostic and/or clinical relevance of this overexpression has, to date, not been investigated. In this study, we analyzed the level of GST-pi expression and its subcellular localization in 61 primary gliomas and correlated the results with tumor histology, patient age, and patient survival. We observed a strong positive correlation between the level of GST-pi expression and tumor grade and between the presence of GST-pi in glioma cell nuclei and patient age. Univariate and multivariate Cox regression analyses and Kaplan-Meier curves showed the level of GST-pi expression and its nuclear localization to be inversely correlated with patient survival. Relative risk for death of patients with high versus low tumor GST-pi expression was 3.2 (P = 0.0069) by univariate analysis and 2.6 (P = 0.036) by multivariate analysis. The relative risk of death associated with the presence of nuclear GST-pi in glioma cells was 3.9 (P = 0.0001) by univariate analysis and 4.4 (P < 0.0001) by multivariate analysis. These data indicate that high GST-pi expression in tumor cells and the presence of the GST-pi protein in tumor cell nuclei are associated with clinically more aggressive gliomas and are strong predictors of poor patient survival.

Enforced expression of wild-type p53 curtails the transcription of the O(6)-methylguanine-DNA methyltransferase gene in human tumor cells and enhances their sensitivity to alkylating agents.

We used isogenic human tumor cell lines to investigate the specific and direct effects of wild-type (wt) p53 on the expression of O(6)-methylguanine-DNA methyltransferase (MGMT), a DNA repair protein that confers tumor resistance to many anticancer alkylating agents. A p53-null, MGMT-proficient lung tumor cell line (H1299) was engineered to express wt p53 in a tetracycline-regulated system. High levels of p53 induction achieved by tetracycline withdrawal were accompanied by G(1) cell cycle arrest without significant apoptosis in this cell line. p53 accumulation resulted in a gradual and dramatic loss of MGMT mRNA, protein, and enzyme activity, whose levels were undetectable by day 3 of induction. The loss of MGMT protein was, however, not due to its degradation because the ubiquitin-promoted in vitro degradation of MGMT, which mediates the cellular disposal of the repair protein, was not altered by p53. Run-on transcription assays revealed a significant reduction in the rate of MGMT gene transcription. The negative regulation of MGMT expression by wt p53 was confirmed in two other human isogenic cell lines, namely, the GM47.23 glioblastoma, which contains a dexamethasone-inducible wt p53, and the H460 lung cancer cell line, in which wt p53 had been inactivated by the human papillomavirus E6 protein. Furthermore, a panel of four human tumor cell lines, including gliomas with wt p53 status, displayed markedly lower levels of MGMT gene transcripts than those having p53 mutations. Induction of wt p53 in these models led to a 3- and 2-fold increase in sensitivity to 1,3-bis(2-chloroethyl)-1-nitrosourea and temozolomide, respectively, which generate the MGMT-repairable O(6)-alkyl adducts in DNA. These results demonstrate that p53 is a negative regulator of MGMT gene expression and can create a MGMT-depleted state in human tumors similar to that achieved by O(6)-benzylguanine, a potent inhibitor of MGMT currently undergoing clinical trials. Thus, our study exposes an additional benefit associated with p53 gene therapy and provides a strong biochemical rationale for combining the MGMT-directed alkylators with p53 gene transfer to achieve improved antitumor efficacy.

Adenovirus-mediated antisense urokinase-type plasminogen activator receptor gene transfer reduces tumor cell invasion and metastasis in non-small cell lung cancer cell lines.

The urokinase-type plasminogen activator (uPA) and its receptor (uPAR) play an important role in the proteolytic cascade involved in the metastasis of lung and other cancers. We report that the reduction in uPAR levels produced by an antisense strategy using an adenovirus construct (Ad-uPAR) in H1299 cells, an invasive human lung cancer cell line that produces high levels of uPAR, resulted in a decrease of uPAR levels to 80-90% of those seen in cells infected with mock or adenovirus (Ad)-cytomegalovirus vector controls. In addition, increasing the multiplicity of infection from 25 to 200 caused a corresponding decrease in the level of uPAR protein within 5 days of treatment, as shown by Western blot analysis. Furthermore, the in vitro translation of total RNA levels of Ad-uPAR-infected H1299 cells in a rabbit reticulocyte lysate system caused a 50-70% decrease in uPAR immunoprecipitate in Ad-uPAR-infected cells relative to the levels in cells of mock and vector controls. The Matrigel invasion assay showed the invasion of H1299 cells and A549 cells infected with Ad-uPAR to be decreased by 70% relative to mock- and vector-infected controls. Infection of tumor cells with Ad-uPAR before implantation significantly reduced the incidence of lung metastasis by 85% as compared with the control virus-infected cells injected into nude mice through the tail vein. Our collective results show that the uPAR system is a potential target of treatment for lung cancers.