We examined loss of heterozygosity at 13 loci on 5 chromosomes in hepatocellular carcinomas (HCCs) from 56 patients. In 42 of these cases, regenerative nodules of liver cirrhosis were also analyzed. High frequencies of allelic losses were detected on chromosomes 13q (47%), 16q (40%) and 17p (64%), whereas losses on chromosome 4p and 11p were observed in less than 22% of cases in HCCs. In contrast, LOH was not detected on any loci in cirrhotic nodules. On chromosome 13q, the common region of allelic loss was mapped to the region including the retinoblastoma (RB) locus, by using 8 polymorphic probes. Furthermore, one case with 13q loss had an interstitial deletion of the RB gene, indicating the involvement of inactivation of the RB gene in hepatotumorigenesis. Losses were associated with portal-vein thrombosis or intrahepatic metastasis, increased tumor size, a poorly differentiated phenotype and clinical stage. Losses occurring together on 13q, 16q and 17p were significantly higher in patients in clinical stage IV or histologically poorly differentiated tumors, suggesting that the accumulation of allelic loss occurs in advanced tumors and that patients with multiple allelic losses may have a worse prognosis than those with a single loss.

Activity of the DNA repair protein O6-alkylguanine-DNA alkyltransferase (AGT) is an important determinant of responsiveness of tumor cells to chloroethylnitrosoureas (CENUs), representative chemotherapeutic agents for primary malignant gliomas. In order to assess the real states of this repair protein in human malignant gliomas, we assayed AGT activity in surgically extirpated 42 malignant glioma samples and studied the distribution of the activity under certain clinical conditions. There were wide variations in AGT activity between individuals. No significant difference in AGT activity on average was seen either between glioblastoma and anaplastic astrocytoma, nor between primary and recurrent tumors. Among 42 malignant gliomas, 7 samples (16.7%) had low AGT activity less than 0.1 pmoles/mg protein. In the case of glioblastoma, tumors possessing higher AGT activity tended to be less responsive to post-operation remission-induction therapy including CENUs. The result of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) chemosensitivity assay by using the corresponding surgical specimens suggested a close relationship between cellular resistance to CENUs and AGT activity. It was found to be unlikely that a short term administration of CENUs had a significant effect on AGT activity of brain tumors in human body. We could detect a bit of definite evidences of the relevance of AGT to resistance to CENUs and need to conduct further investigations for other resistance factors.

Previous studies have shown that in about one-fifth of human tumor cell strains, the activity of O6-methylguanine-DNA methyltransferase (MGMT), which can repair O6-alkylguanine in DNA produced by alkylating agents, is deficient. These strains are termed Mer- cells. To see if there is any human tumor lacking MGMT activity, we measured the MGMT activity in extracts from liver tumors of 21 patients, and compared it to the activity in normal peritumoral tissues derived from the same patients. The MGMT activity was assayed by measuring the 3H radioactivity transferred from the substrate DNA containing [methyl-3H]-labeled O6-methylguanine to an acid-insoluble protein fraction. There was considerable variation in MGMT activity among individual extracts; the interindividual variation was approximately 6-fold in normal liver tissue and much larger in liver tumors. Although in many cases similar high levels of MGMT activity were found both in liver tumors and in the normal counterpart, six tumors had greater than 3-fold less activity compared with the normal liver tissue from the same patient. Liver tumors from two patients did not have any detectable level of MGMT activity by the present method used, in spite of the fact that the corresponding normal liver samples demonstrated significant activities. We also measured in the same tissue extracts the activities of two common enzymes, glutamic pyruvic transaminase (GPT) and lactic dehydrogenase (LDH). The activities of GPT and LDH in the liver tumor samples that showed undetectable levels of MGMT activity were similar to those in the surrounding normal liver tissues. These results may suggest the existence of human Mer- tumors, deficient or very low MGMT activity.

We have examined O6-methylguanine-DNA methyltransferase (O6-MT) activity of rat brain tumour cell strains with reference to cellular resistance to antitumour nitrosoureas, 1-(4-amino-2-methyl-5-pyrimidinyl)methyl-3-(2-chloroethyl)-3-nitrosourea hydrochloride (nimustine, ACNU) and methyl-6-[3-(2-chloroethyl)-3-nitrosoureido]-6-deoxy-alpha-D-glucopyrano side (ramustine, MCNU). The values of O6-MT activity were 52 and 160 fmol/mg protein extract in 9L and C6 rat brain tumour cells, respectively; while HeLa S3 cells, as a methyl excision repair positive (Mer+) cell strain, revealed a rather high value of 488 fmol/mg. 9L cells indicative of a low O6-MT activity showed 13 microM for ACNU and 18 microM for MCNU at a 10% survival dose (SD10), determined by a clonogenic cell assay as an index of cellular resistance. In contrast to this, C6 cells revealed a SD10 value of 67 microM and 36 microM for ACNU and MCNU, respectively, indicating higher resistance than 9L cells. HeLa S3 cells showed the highest SD10 value as follows: 84 microM for ACNU and 73 microM for MCNU. The relationship between the O6-MT activity and the cellular resistance was almost linear, with relatively resistant cell lines exhibiting the higher levels of the O6-MT activity. This correlation between the O6-MT activity and the cellular resistance to nitrosoureas as ACNU and MCNU was not observed among other antitumour drugs, which included bleomycin (BLM), neocarzinostatin (NCS), cis-diamminedichloroplatinum (II) (CDDP), and etoposide (VP-16) in clinical use for brain tumour chemotherapy. This indicates that O6-MT activity can be an indicator of cellular resistance to antitumour nitrosoureas in the chemotherapy of brain tumours.

In order to study the dynamic relationship in glioma cells between O6-alkylguanine-DNA alkyltransferase (AGT) activity and resistance to the cytotoxic effect of chloroethylnitrosoureas (CENUs), we investigated the changes in sensitivity to 1-(4-amino-2-methyl-5-pyrimidinyl)methyl-3-(2-chloroethyl)-3-nitrosourea hydrochloride (ACNU) after modulation of AGT activity. In ACNU-resistant rat glioma cell lines (9LR1, 9LR3, and 9LR12) and a human glioma cell (HNG-1), O6-methylguanine enhanced cytotoxicity to ACNU following a depletion of AGT activity. But no enhancement of cytotoxicity was seen in an ACNU-sensitive rat glioma cell line (9L). In the 9L and 9LR12 cells, equivalently sublethal doses of ACNU similarly depleted AGT activity but the regeneration rates of this repair protein were different. In the case of a 7-day pretreatment with human recombinant interferon-beta (HuIFN-beta), although it could modulate AGT activity in HNG-1 cells, no definite influence on cellular sensitivity to CENUs was observed. However, a 50-day pretreatment with HuIFN-beta conferred resistance to CENUs on them despite its effect to reduce AGT activity. Thus, diversity was seen in the relation between AGT activity and resistance to CENUs when AGT activity was modulated by HuIFN-beta. The results of this study suggest that AGT activity is one of factors affecting cellular sensitivity to CENUs but that alternative mechanisms of tolerance may be induced depending upon some environmental effects.