O6-methylguanine-DNA methyltransferase (MGMT) transfectants of a 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU)-sensitive colon cancer cell line selectively repopulate heterogenous MGMT+/MGMT- xenografts after BCNU and O6-benzylguanine plus BCNU.

To evaluate the role of O6-alkylguanine-DNA alkyltransferase (AGT) in colon tumor chloroethylnitrosourea (CENU) resistance, AGT-deficient VACO 8 cells were transfected with a vector containing or lacking the human O6-methylguanine-DNA methyltransferase (MGMT) cDNA. VACO 8MGMT (V8MGMT) sublines possessed high levels of AGT activity in cell culture and were > 10-fold resistant to the CENU 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU). V8MGMT cells, VACO 8neo cells, and mixtures of both were grown as xenografts in nude mice. MGMT expression in VACO 8 xenografts reflected the percentage of V8MGMT cells present in the tumor inoculum. Xenografts originally containing 0-10% V8MGMT cells were sensitive to BCNU, although partial resistance was observed as the percentage of V8MGMT cells increased. Tumors containing 30-100% V8MGMT cells were completely resistant to BCNU with no regressions and no growth delays. Pretreatment with O6-benzylguanine (BG) depleted tumor AGT activity for at least 6 h and sensitized xenografts containing 1 and 100% V8MGMT cells to BCNU. After BCNU or BG + BCNU, xenografts growing from inoculums containing as low as 0.1% V8MGMT cells had high AGT activities similar to that found in V8MGMT xenografts, with the majority of the cells expressing MGMT. These results provide evidence that MGMT expression influences both intrinsic and acquired colon tumor CENU resistance, that selective expansion of AGT+ colon tumor cells commonly occurs after CENU exposure, and that BG is effective in sensitizing colon tumors to CENUs, even when only a small fraction of the cells in a heterogeneous tumor express MGMT.

Quantitative immunohistochemical estimates of O6-alkylguanine-DNA alkyltransferase expression in normal and malignant human colon.

A major mechanism of resistance to nitrosoureas is O6-alkylguanine-DNA-alkyltransferase. The alkyltransferase biochemical assay measures mean tissue activity but requires availability of fresh tissue and cannot assess tumor heterogeneity, an important component of tumor resistance to alkylating agents. We assessed the levels of alkyltransferase in human colon carcinoma and normal colon by biochemical assay, Western blot, conventional immunohistochemistry, and quantitative immunohistochemistry (using 5H7 and mT3.1 monoclonal IgGs) to correlate whole tissue levels with cell-specific expression. Alkyltransferase activity was 18.0 +/- 4.6 fmol/microgram DNA in normal colon and 15.0 +/- 6.5 fmol/microgram DNA in tumors. By Western blot estimates, alkyltransferase in normal colon was 14.8 +/- 4.2 fmol/microgram DNA and in tumors was 16.2 +/- 7.8 fmol/microgram DNA. Alkyltransferase estimates by biochemical and Western blots were correlated strongly (P < 0.0001). Conventional immunohistochemistry demonstrated that alkyltransferase was predominantly nuclear and in normal colon was concentrated in glandular epithelial mucosal cells close to the lumen, whereas in tumors, expression was heterogenous but localized to malignant epithelial cells. Two parameters of quantitative immunohistochemistry, integrated gray and mean gray, were correlated strongly with each other (P < 0.002) and with biochemical and Western blot estimates (P = 0.004-0.04). Thus, quantitative immunohistochemical estimates of alkyltransferase in fixed tissues are a reasonable alternative to biochemical analysis and have an added advantage of identifying heterogeneity of alkyltransferase expression in tumors.

Role of DNA repair in resistance to drugs that alkylate O6 of guanine.

The mechanism of cytotoxicity of a number of chemotherapeutic agents involves alkylation at the O6 position of guanine, a site that strongly influences cytotoxicity. Repair of these lesions by the alkyltransferase protects from cytotoxicity and is a major mechanism of resistance to these agents. O6-benzylguanine inhibition of alkyltransferase sensitizes tumor cells, and clinical trials are underway to determine its efficacy. The use of gene therapy to enhance the expression of alkyltransferase in hematopoietic cells may prevent dose-limiting myelosuppression and may enhance the utility of this class of chemotherapeutic agents.

Oxygen desaturation on swallowing as a potential marker of aspiration in acute stroke.

We have assessed the measurement of oxygen saturation (SaO2) as a means of detecting aspiration in patients with stroke. For 10 weeks all acute stroke [AS] admissions were seen within 48 hours. Basal SaO2 was measured by pulse oximetry. Patients swallowed 10ml water while sitting up and SaO2 was noted for 2 minutes. Two control groups [young, fit (YF) and inpatient age- and sex-matched, non-neurological disease (IP)] underwent the same assessment. AS subjects underwent independent assessment of swallowing by a speech and language therapist (SLT). Exclusion criteria comprised impaired consciousness, other neurological disease and chest infection. Forty-nine AS subjects [20 men; aged 46-93 (mean 71) years], 55 YF [26 men; aged 18-55 (mean 32) years] and 65 IP [28 men; aged 53-96 (mean 71) years] were studied. Mean (SD) SaO2 fall in AS subjects [2.6 (2.9)%)] was significantly more than in YF [1.1 (0.8)%] or IP [1.1 (0.9)%]. The lower 95% confidence limit for variation in SaO2 did not differ between YF and IP (3.0% 'fall'); 19 (39%) AS subjects desaturated below this 95% lower confidence limit. Mean (SD) SaO2 fall was significantly more in SLT-graded 'aspirators' [4.6 (2.7)%] than 'nonaspirators' [1.4 (1.0)%]. We conclude that (1) a fall in SaO2 on swallowing fluid is common in patients with acute stroke; (2) the presence or absence of desaturation agrees statistically with SLT assessment of aspiration; (3) SaO2 measures may aid bedside assessment of swallowing.

The immature thymocyte is protected from N-methylnitrosourea-induced lymphoma by the human MGMT-CD2 transgene.

N-methylnitrosourea (MNU) induces thymic lymphoma in a high proportion of susceptible C57BL/6xSJL (C57/SJL) mice. Expression of the human DNA repair gene, MGMT cDNA, which encodes O6-methylguanine-DNA-methyltransferase, in transgenic mice effectively prevents MNU-induced thymic lymphomas. In this study, we determined the phenotype of thymocytes expressing the transgene and defined whether the target cell population for MNU induced lymphomas were actually those that expressed the transgene. Transgene expression was characterized by in situ hybridization for MGMT mRNA and immunohistochemistry for the human alkyltransferase protein and was compared to the phenotype of the MNU induced lymphomas. The MGMT transgene was expressed uniformly in immature cortical thymocytes that were CD4+CD8+J11d+ and to a lesser extent in the medullary thymocyte. Lymphomas were induced by single [50 or 80 mg/kg] or multiple doses [30 mg/kg x 5] of MNU to evaluate the dose response of tumor induction and protection by the MGMT-CD2 transgene. Forty-seven of the 108 treated mice developed lymphomas: 38 of 58 nontransgenic and 9 of 50 MGMT+ mice. The T-cell phenotype of thymic lymphomas was established by immunohistochemistry and FACS analysis. Most of the lymphomas were J11d+ (98%), 70% of the tumors were CD4+CD8+, 21% were CD4-CD8+, 9% were CD4-CD8-, and none were CD4-CD8-. All lymphomas in MGMT+ transgenic mice were CD4+CD8+. Since the main phenotype of MNU induced lymphomas in these mice, CD4+CD8+J11d+, is also the cell phenotype which expresses the MGMT-CD2 transgene at high levels, it appears that MGMT-induced protection has occurred in the cell target for MNU induced transformation.

Transgenic expression of human MGMT protects against azoxymethane-induced aberrant crypt foci and G to A mutations in the K-ras oncogene of mouse colon.

Transgenic mice over-expressing MGMT, which codes for the human protein O6-alkylguanine-DNA alkyltransferase, are protected from methylating agent-induced thymic lymphomas. In this study we evaluated the ability of transgenic overexpression of MGMT in the colon to protect mice from the development of azoxymethane(AOM)-induced aberrant crypt foci (ACF) and mutations in K-ras. Colonic alkyltransferase in MGMT+ transgenic mice was > 5-fold higher than in nontransgenics: 10.5 +/- 1.1 vs 2.2 +/- 1.1 fmol/micrograms DNA, P = < 0.0001. Mice received 20 mg AOM/kg i.p. at 6 weeks or 15 mg AOM/kg at 6 and 7 weeks of age, and 8 wks later colons were examined for ACF. A significant protective effect of MGMT was seen in mice given single dose of 20 mg AOM/kg. The incidence of ACF/colon was lower in MGMT+ mice (2.0 +/- 1.2) than in nontransgenic mice (3.9 +/- 1.8, P = 0.02). G to A mutations in codon 12 of K-ras were detected by PCR-RFLP in ACF and in random samples of normal appearing mucosa. The incidence of ACF with mutant K-ras in MGMT transgenic mice (0.6 +/- 0.7/colon) was significantly reduced compared to nontransgenic mice (2.3 +/- 1.7/colon, P = 0.02). We propose that AOM induces at least two overlapping but not identical premalignant lesions (aberrant crypt foci and K-ras mutations) which can be prevented by over-expression of MGMT. Thus, MGMT may protect colonic mucosa from carcinogenesis involving methylating agents such as AOM.

Identification in rat stomach mucosae of a cell population characterized by a deficiency for the repair of O6-methyldeoxyguanosine from DNA.

Immunohistochemical studies have been used to show the time course for the cell-specific methylation of DNA in the upper gastrointestinal tract of rats treated with N-methyl-N'-nitro-N-nitrosoguanidine (MNNG). The doses used were 1, 5, 25 and 50 mg MNNG/kg (i.g.) and tissue samples were analysed at intervals of from 1 to 192 h after treatment. Relatively little reaction with nuclear DNA was observed in the forestomach and still less in the oesophagus. Reaction with DNA was most extensive in the corpus, pylorus and the duodenum, reaching a peak of staining intensity between 2 and 4 h and then declining progressively there-after. Staining for the presence of O6-methyldeoxyguanosine (O6-MedG) in DNA was highly selective and tended to occur in the nuclei of the basal cell of the oesophagus and forestomach and in the cells of the lumenal border of the glands and villi of the corpus, pylorus and duodenum. There were also areas, 5-15, glands apart where staining for O6-MedG in the corpus and pylorus extended as far down as the basal mucosa. From 12 h after MNNG treatment, in the corpus and pylorus, a band of strongly methylated cells became apparent about 3-6 cells deep from the lumen and remained identifiable up to 168 h after treatment with the higher doses. These cells, which apparently have a very low O6-MedG repair capacity, are stationary (i.e. not part of the escalator) and are located in the mesenchymal tissue elements as demonstrated by staining of serial sections with cytokeratin or vementin. The significance of this population of cells is unknown.

Alkyltransferase transgenic mice: probes of chemical carcinogenesis.

Transgenic mice expressing DNA-repair genes are an instructive model with which to study the protective role of DNA-repair pathways in both spontaneous and chemical carcinogenesis. Of particular interest in chemical carcinogenesis is the DNA-repair protein O6-alkylguanine-DNA alkyltransferase (alkyltransferase) which repairs O6-alkylguanine-DNA adducts. Transgenic mice carrying expression constructs for the alkyltransferase gene--either the human MGMT cDNA or the bacterial ada gene--express increased levels of alkyltransferase and have increased capacity to remove O6-methylguanine-DNA adducts. Protection from the DNA damaging effects of N-nitroso compounds occurs specifically in the cells and tissues in which the alkyltransferase transgene is expressed. For instance, mice carrying the PEPCKada construct have increased alkyltransferase in the liver and more rapid removal of O6methylguanine-DNA adducts. The protective effect is noted in hepatocytes, which express PEPCK-linked genes, not in nonparenchymal cells of the liver, which do not. Other tissues that express the transgene in the various models include the thymus, spleen, testes, muscle, stomach and brain. Mice expressing the human alkyltransferase in the thymus have a reduced incidence of thymic lymphomas following exposure to methyl nitrosourea (MNU), evidence of a role for this DNA-repair protein in protection from carcinogenesis due to N-nitroso compounds. Protection has also been observed in the induction of hepatic tumors by N-nitroso-dimethylamine (NDMA). These models will be used to identify whether overexpression of a single DNA-repair gene can block the carcinogenic process of N-nitroso compounds in many different tissues.

Tissue and cell specific methylation, repair and synthesis of DNA in the upper gastrointestinal tract of Wistar rats treated with single doses of N-methyl-N'-nitro-N-nitrosoguanidine.

Several potential cancer risk factors have been monitored concurrently in the upper gastrointestinal tract of young adult male Wistar rats given single (i.g.) doses of N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) which readily induces forestomach tumours under these conditions. Radioimmunoassay was used to determine the formation of O6-methyl-2'-deoxyguanosine (O6-MedG) in DNA after MNNG doses of 1, 5, 25 or 50 mg/kg and was found to be highest in the pylorus, with progressively lower levels in the corpus, forestomach, duodenum, oesophagus and jejunum. Immunohistochemical procedures showed that cells with nuclei containing O6-MedG were heterogeneously distributed in these tissues. O6-Alkylguanine-DNA alkyltransferase activity in untreated animals was highest in the mucosae of the corpus, lower and relatively similar in that of the pylorus, duodenum and jejunum and lowest in the tissues of oesophagus and forestomach. Estimates of DNA synthesis and cell proliferation indicated a 5-fold increase in the DNA labelling index in the forestomach whereas perturbations of DNA synthetic activity in the other tissues of the upper gastrointestinal tract were much less marked. As a result of these changes, cells with nuclei that contained O6-MedG and were also undergoing DNA synthesis (determined by sequential immunohistochemical analysis and autoradiography) were found most commonly in the forestomach and to a lesser extent in the pylorus. This distribution of replicating damaged cells corresponds with the relative tumour yields in these upper gastrointestinal tract tissues and such cells are the probable targets in this single dose carcinogenesis regime. Thus, whilst the highest concentration of O6-MedG did not correlate tumour incidence, the overall risk for tumour induction did correlate with a significant level of DNA damage, a lower capacity for DNA repair and a marked increase in DNA synthesis over the constitutive level in the target cells. Carcinogenic risk in this system is therefore more readily determined by studying several risk factors simultaneously.

Tissue and cell specific methylation, repair and synthesis of DNA in the upper gastrointestinal tract of Wistar rats treated with N-methyl-N'-nitro-N-nitrosoguanidine via the drinking water.

Several potential cancer risk factors have been monitored concurrently in the upper gastrointestinal tract of young male Wistar rats given N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) via the drinking water, a regimen that induces a high yield of tumours in the pylorus and to a lesser extent in the duodenum. Radioimmunoassay was used to determine the amounts of O6-methyl-2'-deoxyguanosine (O6-MedG) formed in the tissue DNA of rats given MNNG at doses of 40 or 80 micrograms/ml for periods of 3, 6 and 12 weeks. The highest adduct concentration was found in the pylorus with progressively lower concentrations in the corpus and duodenum, jejunum, forestomach and oesophagus. Between 3 and 12 weeks these adduct levels decreased in all tissues and there was no evidence of a dose dependent accumulation of O6-MedG. When analysed by immunohistochemistry the distribution of cells with nuclei containing O6-MedG was seen to be heterogeneous in the various tissues. O6-Alkylguanine-DNA alkyltransferase activity increased during the 12 weeks of MNNG treatment in oesophagus and forestomach, but decreased to approximately 50% of the initial value in the corpus, pylorus, duodenum and jejunum. The major changes in DNA synthesis and cell proliferation were the marked upward expansion (i.e. towards the lumen) of the zone of replicating cells in the glands of the pylorus and the greatly increased numbers of replicating damaged cells (i.e. cells that contained O6-MedG whilst undergoing DNA synthesis) as determined by sequential immunohistochemical analysis and autoradiography. Such cells are the probable target cells in this chronic dose carcinogenesis regime. Although similar changes also occurred in the glands of the corpus these were of lesser extent and the changes of labelling index in the oesophagus and forestomach were relatively minor. In the duodenum, MNNG treatment led to erosion of the upper part of the glands so that the zone of cells containing O6-MedG overlapped with the zone of proliferating cells resulting in the formation of many replicating damaged cells. Thus, as in the single dose study (see preceding paper) the distribution of replicating damaged cells coincides with the tumour yield in the tissues of the upper gastrointestinal tract. As in the case of single doses of MNNG the risk factors for carcinogenesis are, a significant level of DNA damage, a lower capacity for DNA repair and an increased DNA synthetic activity, again suggesting that carcinogenic risk cannot readily be determined by studying risk factors individually.

N-methyl-N'-nitro-N-nitrosoguanidine-induced carcinogenesis: differential pattern of upper gastrointestinal tract tumours in Wistar rats after single or chronic oral doses.

Male Wistar rats were treated with N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) either as a single dose of 50, 125 or 250 mg/kg given by gavage or via drinking water for 28 weeks at a concentration of 40, 80 or 160 micrograms/ml, in the case of the higher concentration reverting to 80 micrograms/ml after the first 8 weeks. The single dose regimen had no effect on water intake or body weight, but the chronic exposure led to a dose-dependent reduction in water intake that was paralleled by a slower weight gain, with the final body weights at approximately 90, 84 and 79% of the control weight values. The combined yield of benign and malignant tumours (79-100% of the animals treated) occurred in the forestomach in the case of the single doses, whereas the chronic exposure resulted in a maximum yield of tumours located in the pyloric region of the glandular stomach (64-100% of animals treated). The principal histological types of tumours induced were squamous cell papilloma and carcinoma in the forestomach and adenocarcinoma in the pylorus. There was a persistent, but low yield (25-30% of animals treated) of tumours in the jejunum, mainly adenocarcinoma, after administration via drinking water, whereas after single doses, multiple solitary cysts and cholangioma (30% and 25-70% respectively of the animals treated) were found in the liver. This report differs from earlier reports in that marked effects were noted on water consumption and body weight gain and that tumour induction can be achieved after much shorter periods of exposure than previously reported in the literature. These data confirm the tissue specificity of MNNG when given either as a single or chronic dose regimen and provide a suitable model for the investigation of the target cell specificity of tumour induction.