Search

Background: The generation of DNA interstrand cross-links is thought to be important in the cytotoxicity of nitrogen mustard alkylating agents, such as melphalan, which have antitumor activity. Cell lines with mutations in recombinational repair pathways are hypersensitive to nitrogen mustards. Thus, resistance to melphalan may require accelerated DNA repair by either recombinational repair mechanisms involving Rad51-related proteins (including x-ray repair cross-complementing proteins Xrcc2, Xrcc3, and Rad52) or by nonhomologous endjoining involving DNA-dependent protein kinase (DNA-PK) and Ku proteins. We investigated the role of DNA repair in melphalan resistance in epithelial tumor cell lines., Methods: Melphalan cytotoxicity was determined in 14 epithelial tumor cell lines by use of the sulforhodamine assay. Homologous recombinational repair involving Rad51-related proteins was investigated by determining the levels of Rad51, Rad52, and Xrcc3 proteins and the density of nuclear melphalan-induced Rad51 foci, which represent sites of homologous recombinational repair. Nonhomologous endjoining was investigated by determining the levels of Ku70 and Ku86 proteins and DNA-PK activity. Linear regression analysis was used to analyze correlations between the various protein levels, DNA-PK activity, or Rad51 foci formation and melphalan cytotoxicity. All statistical tests were two-sided., Results: Melphalan resistance was correlated with Xrcc3 levels (r =.587; P =.027) and the density of melphalan-induced Rad51 foci (r =.848; P =.008). We found no correlation between melphalan resistance and Rad51, Rad52, or Ku protein levels or DNA-PK activity., Conclusion: Correlations of melphalan resistance in epithelial tumor cell lines with Xrcc3 protein levels and melphalan-induced Rad51 foci density suggest that homologous recombinational repair is involved in resistance to this nitrogen mustard.

We previously found that 2-chloroethyl-3-sarcosin-amide-1-nitrosourea (SarCNU), a new chloroethylnitrosourea analogue presently in phase I clinical trials, is a selective cytotoxin that enters cells via the extraneuronal transporter for monoamine transmitters (EMT). In this study, we assessed whether EMT expression correlates with SarCNU cytotoxicity by determining EMT expression in 23 human tumor cell lines with reverse-transcription PCR. Western blot analysis was used to measure protein levels of the DNA repair genes, O6-methylguanine-DNA methyltransferase (MGMT), and excision repair cross-complementing rodent repair deficiency gene 2 (ERCC2). SarCNU cytotoxicity was determined by the sulforhodamine B colorimetric anti-cancer-drug screening assay and correlated with gene expression. Almost all of the cell lines screened were positive for EMT expression. However, seven cell lines (MGR-1, MGR-2, T98-G, SKI-1, SKNSH, 297, and GBM) expressed low levels of EMT. Although there was no linear correlation between SarCNU cytotoxicity and EMT expression, SarCNU cytotoxicity significantly correlated with ERCC2 protein levels, and MGMT-rich (Mer+) cell lines (MGMT protein level >0.1) were more resistant to SarCNU than MGMT-poor (Mer-) cell lines (MGMT protein level <0.1). Moreover, multiple regression analysis indicated that the best correlation with SarCNU cytotoxicity was attainable with EMT plus MGMT and ERCC2 expression. This study suggests that in human tumor cell lines both EMT and DNA repair factors, specifically, MGMT and ERCC2, are important determinants of SarCNU activity. Because EMT is expressed in a wide variety of human tumors, SarCNU should be a more widely effective alternative chemotherapeutic agent.

Our previous studies with B-cell chronic lymphocytic leukemia (B-CLL) have suggested that one of the mechanisms of nitrogen mustard (NM) drug resistance is increased repair of drug-induced damage. We have postulated that recombination may play a crucial role in this process. The human homologue of Rad51, (HsRad51), has homology to the RecA protein in Escherichia coli, which is implicated in recombination repair and induction of DNA repair enzymes. In this report, we have examined the expression and distribution of HsRad51 protein in lymphocytes from patients with B-CLL to see whether the expression of HsRad51 is associated with NM damage to the malignant B lymphocytes, specifically chlorambucil (CLB), which is the standard alkylating agent used to treat patients with B-CLL. We have analyzed the intracellular distribution of HsRad51 protein in these lymphocytes before and after treatment with CLB by immunofluorescence. In vitro CLB treatment induces Rad51 expression, as measured by increased immunopositive staining in all CLL samples. In the CLB-resistant CLL lymphocytes, there was a linear correlation between induction of Rad51 protein at 5.4 microM CLB and the in vitro LD50 dose of CLB. Surprisingly, although it has been reported that Rad51 is induced in S phase and only 10% of cells from cell lines expressed positive immunostaining for Rad51, our CLL lymphocytes, which were not subjected to in vitro drug exposure, were 90% positive for Rad51, despite their nonproliferative state, which suggests that there is chronic activation of the protein. Our results suggest that CLB activates HsRad51-directed recombination repair and that this process may be important in NM drug-induced cytotoxicity.

Adjuvant nitrosourea chemotherapy fails to prolong patient survival significantly as many tumors demonstrate resistance to these drugs. It has been documented in cell lines that O(6)-methylguanine-DNA methyltransferase (MGMT) plays an important role in chloroethylnitrosourea (CENU) drug resistance. The authors evaluated MGMT expression in 22 glioma specimens by using an immunofluorescence assay and compared the results with clinical response of the patients to CENU-based chemotherapy. The patients were treated with CENU after evidence of progressive disease following surgery and radiotherapy. Eight tumor samples had no detectable MGMT, whereas other samples had from 9989 to 982,401 molecules/nucleus. In one group (12 patients), the tumor decreased in size or was stable (effective group), whereas in the other group (10 patients), the tumor demonstrated continuous growth during chemotherapy (progressive group). The median time to progression (TTP) was 6.7 months with a median survival of 13 months. The Mer(-) patients (MGMT < 60,000 molecules/nucleus) appeared to have more chance of stable disease or response to CENU therapy than the Mer(+) patients (MGMT > 60,000 molecules/nucleus) (chi-square = 4.791, p = 0.0286). In patients with glioblastomas multiforme (GBMs), the TTP of Mer(+) patients was shorter than that of Mer(-) patients (t = 2.04, p = 0.049). As a corollary, the MGMT levels were significantly higher in GBM tumors from the progressive group than those from the effective group (t = -2.26, p = 0.029). The TTP and survival time in the effective GBM group were also longer than those in the progressive GBM group. However, there was no significant correlation between MGMT levels and either the survival time (r = 0.04, p = 0.8595) or TTP (r = 0.107, p = 0.6444). Results from this study suggested that MGMT positivity is indicative of more aggressive disease that progresses more rapidly when exposed to CENU therapy. However, MGMT-negative tumors are not always sensitive to CENU agents, suggesting that other factors may also be important.

Nitrosoureas are among the most widely used agents used in the treatment of malignant gliomas. Here, the activity of 2-chloroethyl-3-sarcosinamide-1-nitrosourea (SarCNU) was compared with that of 1,3-bis-(2-chloroethyl)-1-nitrosourea (BCNU), in vivo against s.c. implanted SF-295 and U-251 central nervous system (CNS) tumor xenografts. When given i.v., q4d for 3 doses, to athymic mice bearing s.c. SF-295 tumors, SarCNU, at an optimum of 167 mg/kg/dose, produced 9 tumor-free animals of 10 total animals, 1 regression, and no evidence of overt toxicity (> or =20% body weight loss). With a similar dosing schedule, BCNU produced no tumor-free animals, six regressions, and one drug-related death at its optimum of 30 mg/kg/dose. Furthermore, SarCNU retained high antitumor activity at two lower dose levels, 66 and 45% of the optimal dose, whereas BCNU demonstrated a progressive loss of antitumor activity at lower doses. Following p.o. administration, SarCNU similarly demonstrated antitumor activity that was superior to that of BCNU. In the U-251 CNS tumor model, SarCNU yielded six of six tumor-free animals at 80 mg/kg/dose with i.p. administration q.d. for 5 days, starting on day 14, whereas BCNU, at 9 mg/kg/dose, yielded three of six tumor-free mice and one drug-related death. Again, SarCNU resulted in tumor-free animals at 66 and 45% of its optimal dose and was relatively nontoxic, in contrast to BCNU. Results of testing to date indicate that SarCNU is clearly more effective than BCNU against the human CNS tumors SF-295 and U-251 in vivo. These results encourage the initiation of clinical trials for SarCNU, in an effort to improve therapeutic approaches to glioma, but clinical trials must determine whether superiority of SarCNU in preclinical models can be extrapolated to patients.

We have developed a method to quantitate ERCC-2 gene expression in tumor cell lines. A mutant ERCC-2 DNA fragment (1-bp mutation) is used as a competitive DNA template in a coamplification PCR reaction with cDNA obtained by reverse transcribing DNase-free total RNA from six human tumor cell lines. The PCR products are separated on agarose gel by virtue of their differential banding pattern upon restriction enzyme digestion. Densitometric readings of the PCR products from a negative film of the gel are used to establish a linear regression curve, which in turn is used to quantitate ERCC-2 levels. Beta-actin expression is similarly quantitated. Normalized ERCC-2 gene expression (either to beta-actin or to total RNA) correlates with cytotoxicity of 1,3-bis-(2-chloroethyl)-1-nitrosourea or (2-chloroethyl)-3-sarcosinamide-1-nitrosourea, suggesting that ERCC-2 may play an important role in drug resistance in these cell lines. This method is reliable and can be used to quantitate gene expression in clinical tumor specimens.

Resistance to (2-chloroethyl)-3-sarcosinamide-1-nitrosourea (SarCNU), an experimental anticancer compound, was investigated in the chloroethylnitrosourea-sensitive Mer- SK-MG-1 and -resistant Mer- SKI-1 human glioma cell lines. The transport of [3H]SarCNU was examined in suspension. The uptake of [3H]SarCNU was found to be temperature dependent in SK-MG-1 and SKI-1, but less so in SKI-1. At 37 degrees C, uptake of 50 microM [3H]SarCNU was linear up to 4 s in both cell lines, with uptake being significantly faster in SK-MG-1 than in SKI-1 under initial rate conditions. There was no significant difference in the rate of influx at 22 degrees C between both cell lines. Equilibrium was approached after 1 min at 22 and 37 degrees C. At 37 degrees C, steady state accumulation of SarCNU at 30 min was reduced significantly (35%) in SKI-1 cells compared with SK-MG-1 cells, although accumulation was similar at 22 degrees C. In SK-MG-1 cells, uptake of [3H]SarCNU at 37 degrees C was found to be saturable, but uptake in SKI-1 cells was not saturable over a 1000-fold range of concentrations. Analysis of efflux in cells preloaded with 50 microM [3H]SarCNU revealed that the rate of efflux was equivalent in both cell lines but that the efflux rate was more rapid at 37 degrees C compared with 22 degrees C. Metabolism of SarCNU at 37 degrees C was not different in either cell line after a 60-min incubation, as determined by thin layer chromatography. SKI-1 cells, compared with SK-MG-1 cells, were 3-fold more resistant to SarCNU at 37 degrees C but only 2-fold more resistant at 22 degrees C, a temperature at which SarCNU accumulation was similar in both cell lines. The 2-fold resistance at 22 degrees C was similar to that of 1,3-bis(2-chloroethyl)-1-nitrosourea at 37 and 22 degrees C. These findings indicate that increased cytotoxicity in SK-MG-1 cells is associated with a greater accumulation of SarCNU via an epinephrine-sensitive carrier that is not detectable in SKI-1 cells. However, part of the chloroethylnitrosourea resistance in SKI-1 cells is not secondary to decreased accumulation.

Previous studies of ERCC-1 gene expression levels in chronic lymphocytic leukemia and ovarian carcinoma tumor specimens indicated that increased gene expression may correlate with a lack of response to the alkylating agents, melphalan, and cis-diamminedichloroplatinum (II). In order to demonstrate direct involvement of the ERCC-1 protein in repair of melphalan lesions, the ERCC-1 defective Chinese hamster ovary cell line UV20 was transfected with the human ERCC-1 complementary DNA. Stably transfected UV20 cells demonstrated an increase in resistance to melphalan. Wild type Chinese hamster ovary AA8 cells were then stably transfected with the same complementary DNA. The result was an increase in sensitivity to melphalan. There was no effect on sensitivity to uv light, but the ERCC-1 transfected AA8 cells had an increased sensitivity to cis-diamminedichloroplatinum (II). These results suggest that overexpression of human ERCC-1 may inhibit a pathway specific to the repair of bifunctional DNA damaging agent lesions in AA8 cells. ERCC-1 transfected AA8 cells should be useful in determining the precise role of ERCC-1 in repair of DNA cross-links induced by melphalan.

The transport of the amino acid amide N-[3H]sarcosinamide (methyl glycinamide) was investigated in human glioma SK-MG-1 cells. Sarcosinamide uptake was found to be temperature dependent, sodium independent, and linear up to 1 min at 22 degrees C. Equilibrium was reached after 10 min at 22 degrees C with accumulation slightly above unity. Sarcosinamide was not metabolized in the cells as shown by thin layer chromatography. The uptake of sarcosinamide was significantly decreased when the extracellular pH was lowered from 7.5 to 6.0 and significantly enhanced at pH values above 7.5. The latter effect may be due mainly to increased cell permeability at high pH. The uptake of the labeled sarcosinamide was trans-stimulated by excess cold sarcosinamide. Sarcosinamide uptake over a 200-fold range of concentrations followed Michaelis-Menten kinetics with a Km of 0.284 +/- 0.041 mM and a Vmax of 0.154 +/- 0.024 nmol/10(6) cells/min. The uptake of sarcosinamide was significantly reduced by iodoacetate but not by the metabolic poisons NaF, ouabain, or dinitrophenyl, suggesting that the uptake is not dependent on energy, rather it proceeds by facilitated diffusion. Several naturally occurring substrates were unable to inhibit the uptake of sarcosinamide. Leucine significantly reduced the uptake of sarcosinamide, while sarcosinamide was a weak inhibitor of leucine transport. 2-Aminobicyclo[2,2,1]heptane-2-carboxylic acid a specific substrate for the sodium-independent, 2-aminobicyclo[2,2,1]heptane-2-carboxylic acid-sensitive amino acid system L failed to inhibit the uptake of sarcosinamide. Epinephrine reduced the uptake of sarcosinamide and sarcosinamide was equally potent as an inhibitor of epinephrine transport. Dixon plot analysis demonstrated that epinephrine (Km = 0.270 mM) inhibits the uptake of sarcosinamide competitively (Ki = 0.260 mM). These results indicate that sarcosinamide is a substrate for the catecholamine transporter. The alkylating agent, sarcosinamide chloroethylnitrosourea, was tested for its ability to inhibit the uptake of sarcosinamide. The results of Dixon plot analysis were consistent with competitive inhibition of sarcosinamide uptake and the inhibition constant Ki for SarCNU was found to be 3.26 +/- 0.57 mM. The steady-state intracellular concentration of SarCNU was found to be significantly higher (cell:medium ratio of 1.03 +/- 0.01) than that of BCNU cell:medium ratio of 0.52 +/- 0.12). These findings indicate that SarCNU and sarcosinamide share the same carrier for uptake in SK-MG-1 cells. This transport mechanism may be responsible for the increased accumulation of SarCNU as compared to BCNU, a nitrosourea which enters cells by passive diffusion.

Chlorozotocin is a chloroethyl nitrosourea with a glucose carrier that has curative activity for the murine L1210 leukemia, but is nonmyelosuppressive in mice. To determine the mechanism for this unique property of reduced bone marrow toxicity, comparative studies were conducted with chlorozotocin and CCNU, a myelotoxic chloroethyl nitrosourea. Suspensions of L1210 leukemia and murine bone marrow cells were incubated for 2 h with 0.1 mM [(14)C]-chloroethyl chlorozotocin or CCNU. Chlorozotocin demonstrated a fourfold increased covalent binding of the chloroethyl group to L1210 nuclei when compared to equimolar CCNU. Chlorozotocin alkylation of L1210 cells resulted in the binding of 57 pmol of [(14)C]ethyl group/mg of DNA, which represented a 2.3-fold increased alkylation when compared to CCNU. In marked contrast, the binding of the chloroethyl group to bone marrow nuclei was equivalent for both drugs. In addition, chlorozotocin alkylation of murine bone marrow DNA, 45 pmol of [(14)C]ethyl group/mg of DNA, was equivalent to that of CCNU. The ratio of L1210:bone marrow DNA alkylation was 1.3 for chlorozotocin compared to 0.6 for CCNU. The intracellular carbamoylation of L1210 and bone marrow protein by CCNU was 400- to 600-fold greater than that produced by chlorozotocin. After a 2-h exposure to 0.1, 0.05, or 0.01 mM drug, both chlorozotocin and CCNU produced a reduction in the cloning efficiency of L1210 cells that was dose dependent. However, chlorozotocin was significantly more cytotoxic than CCNU at all three molar concentrations (P < 0.01). Chlorozotocin, 0.1 mM, reduced L1210 DNA synthesis to 1% of control by 48 h, in contrast to 16% with equimolar CCNU (P < 0.01). In mice bearing 10(5) L1210 cells, chlorozotocin produced its optimal antitumor activity (332% increased life span [ILS]) at doses of 48-64 mumol/kg, with >50% indefinite survivors. In contrast, CCNU at the same molar doses resulted in only a 191% ILS; a CCNU dose of 128 mumol/kg was required for comparable optimal L1210 antitumor activity, 413% ILS. On a molar basis, the dose of chlorozotocin that produced optimal in vivo L1210 antitumor activity was one-third to one-half that of CCNU. Chlorozotocin, unlike CCNU, produced no murine bone marrow toxicity at its optimal therapeutic dose. This unique combination of antitumor activity without myelosuppression can be correlated with the advantageous ratio of L1210:bone marrow in vitro DNA alkylation by chlorozotocin (1.3) as compared to equimolar CCNU (0.6).

Sodium cyanate is a selective in vivo inhibitor of protein synthesis in a variety of mammalian tumor cells without a corresponding effect on the normal tissues of tumor-bearing animals. The in vivo decrease of protein synthesis observed 4 h post-NaOCN i.p. administration in the murine P388 leukemia cell cannot be explained by decreased amino acid pools in the mouse peritoneal cavity. In addition, the decrease in protein synthesis observed with NaOCN in isolated P388 cells was shown not to be secondary to (a) alterations in the kinetics of amino acid transport or (b) effects on total nucleotide pools. The incorporation of [14C]phenylalanine in P388 cell-free lysates from NaOCN-pretreated mice was significantly decreased to approximately 55% of control lysates in the presence of exogenous amino acids. The addition of exogenous calf liver tRNA to the lysates did not alter this result. However, no difference was observed in polyuridylic acid-directed [14C]phenylalanine incorporation into polypeptides in micrococcal nuclease-treated P388 lysates from NaOCN-pretreated or control mice. Quaternary initiation complex (48S) formation and mRNA synthesis were found to be significantly decreased by 35 and 38%, respectively, in P388 cells from NaOCN-pretreated mice. DNA synthesis was decreased by 66% of control at 1 h and 62% at 4 h post-NaOCN i.p. administration. No apparent effect with NaOCN was observed on total RNA synthesis in P388 cells. These results suggest that the decrease in P388 cell protein synthesis observed with NaOCN in vivo appears to be due to alterations manifested in the synthesis of cellular mRNA and protein synthesis initiation processes. NaOCN does not appear to affect the P388 cell ribosomal machinery, tRNA, or protein synthesis elongation processes.

1-(2-Chloroethyl)-3-(beta-D-glucopyranosyl)-1-nitrosourea (GANU), a water-soluble nitrosourea, differs from 2-[3-(2-chloroethyl)-3-nitrosoureido]-D-glucopyranose (chlorozotocin) by the placement of the cytotoxic group on C-1 of glucose. Its biological and biochemical properties are compared with those of chlorozotocin. At a 10% lethal dose (10 mg/kg i.p.), GANU demonstrates minimal myelosuppression. This dose failed to depress normal bone marrow DNA synthesis, in contrast to a 96% inhibition in L1210 DNA synthesis. In L1210 cell suspension, equimolar doses of GANU and chlorozotocin produced equivalent degrees of inhibition in DNA synthesis. GANU has significant L1210 activity in BALB/c X DBA/2 F1 mice treated on Day 2 of tumor growth. A 117% increased life-span and 15% 45-day survivors are atained with 15 mg/kg i.p., a 50% lethal dose. However, in concurrent studies using randomly selected littermate groups of mice, GANU proved less active than chlorozotocin which produced a 306% increased life-span (15 mg/kg i.p.). GANU and chlorozotocin have similar in vitro alkylating activity but the in vitro carbamoylating activity of GANU is sevenfold that of chlorozotocin. On a molar basis, the lethal toxicity of GANU is twice that of chlorozotocin. The significant carbamoylating activity of GANU may contribute to its greater toxicity and therefore limit the mumoles of alkylating agent that can be administered to the tumor. These structure-activity studies further confirm that the addition of a glucose carrier to a cytotoxic nitrosourea moiety can selectively reduce bone marrow toxicity while retaining antitumor activity.