Accumulation of O6-methylguanine in human DNA after therapeutic exposure to methylating agents and its relationship with biological effects.

O6-Methylguanine has been measured in peripheral blood leukocytes of 14 patients during one or more cycles of treatment with procarbazine (daily treatment for 10 days) and in 12 patients during one or more cycles of treatment with dacarbazine (single dose per cycle). Adduct formation at levels up to about 0.4 fmole/microgram DNA was detected in all procarbazine- and all but one dacarbazine-treated patients at some point after treatment. O6-Methylguanine accumulated during procarbazine treatment in a dose-related manner (mean rate of accumulation 2.8 x 10(-4) fmole/microgram DNA per mg/m2 dose) and appeared to approach a plateau by the end of the cycle (above 600 mg/m2 cumulative dose). The average rate of O6-methylguanine formation 2 hr after dacarbazine treatment was 11 +/- 8 x 10(-4) fmole/microgram DNA per mg/m2 dose. Individuals examined on more than one treatment cycle with either drug showed broadly similar methylation responses. The rate of adduct accumulation showed a nonsignificant, negative correlation with the pretreatment lymphocyte levels of the repair enzyme O6-alkylguanine-DNA alkyltransferase (AGT) in the case of procarbazine and no correlation in the case of dacarbazine. No consistent lymphocyte AGT depletion was noted as a result of treatment with either drug. No correlation between O6-methylguanine formation and hematological toxicity was observed. In eight patients showing full remission after treatment with dacarbazine, the value of O6-methylguanine (averaged over all the cycles) was 0.252 +/- 0.120 fmole/microgram DNA while in four patients showing partial or no response it was 0.087 +/- 0.110 fmole/microgram DNA (p < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
The measurement of DNA adducts formed in patients treated with alkylating drugs is of interest because such adducts may indicate the biological dose received at the level of the individual patient, which may be correlated with therapeutic response or long-term complications (mainly therapy-induced carcino-genesis). During the past 2 years we have focused our attention on two methylating agents used in the chemotherapy of Hodgkin's lymphoma, procarbazine (PCZ) [N-isopropyl-a-(2methylhydrazino)-p-toluamide] and dacarbazine (DCZ) [5-(3,3-dimethyl-1-triazeno)imidazole-4-carboxamide] (1,2). Both drugs are converted by metabolism to SN 1-type methylating intermediates and give rise to a range of DNA adducts, including 06-methylguanine (06-MeG) (3,4), an adduct believed to play an important role in carcinogenesis, mutagenesis, and cytotoxicity (5)(6)(7). One of the most widely used therapeutic protocols for Hodgkin's lymphoma, the MOPP protocol (using PCZ in combination with nitrogen mustard, vincristine, and prednisone), has been shown to be associated with an increased risk ofacute nonlymphocytic leukemia in treated individuals, giving rise to a cumulative risk of more than 10% KYRTOPOULOS ETAL. within 10 years oftreatment (8)(9)(10). An alternative protocol, the ABVD protocol (using dacarbazine in combination with adriamycin, bleomycin, and vincristine), although not so far shown to be leukemogenic, may not be completely free ofcarcinogenic risk (11).
Apart from considerations of therapy-induced carcinogenesis, an additional reason that makes the study of the methylation of human DNA by PCZ and DCZ of interest is the need to develop methodologies for assessing of the risks posed to the general population by exposure to environmental methylating agents. Such exposure is known to occur widely via contact with exogenous or endogenously formed chemicals (12)(13)(14)(15) Here we review the current status of our ongoing studies and discuss the observed relationships between exposure dose, adduct formation, DNA repair and biological effects and discuss their usefulness in the assessment ofcancer risk associated with exposure to PCZ and DCZ and to environmental methylating agents.

Materials and Methods
Procarbazine treatment of patients involved 50-mg doses, three times per day (about 80 mg/m2day) for 10 consecutive days, followed by a 3-week rest period, a treatment normally  repeated for six cycles. Other drugs employed in the same protocol included vincristine and cyclophosphamide or mechlorethamine hydrochloride. Each cycle of DCZ treatment involved a single injection of the drug (300 mg, about 160 mg/M2) followed 2 hr later by IV injections ofbleomycin, vinblastin, and epirubicin. This treatment was repeated every 15 days for a total of four to six cycles. Blood samples were collected from each PCZ-treated patient just before the start of a cycle and on indicated days during the cycle, 4 hr after the morning intake of a 50-mg PCZ tablet. For DCZ patients, blood samples were collected just before and 1 hr after DCZ administration as well as on the next 1-2 days (where possible). In some cases a duplicate portion of blood was collected and immediately used for the isolation of lymphocytes and measurement of 06-alkylguanine-DNA alkyltransferase (AGT). Details of the assays used for determining 06-MeG and AGT have been published (1,2,16). Procarbazine-treatment of female Sprague-Dawley rats (150 g) was carried out by daily gavage (10 mg/kg, 59 mg/M2) for 10 days. Groups offour animals were killed 24 hr after they received the indicated cumulative doses. In a separate series of experiments, rats were given single IP doses of PCZ or dimethylnitrosamine (DMN) at the doses indicated and killed after 2 or 6 hr, respectively.

Results and Discussion
Fourteen patients (twelve with Hodgkin's lymphoma and two with non-Hodgkin's lymphoma) on PCZ treatment and twelve Hodgkin's lymphoma patients on DCZ treatment were followed for up to four cycles of chemotherapy. In almost all cases, no 06-MeG could be detected prior to the start ofa treatment cycle, even in patients starting a second or subsequent cycle, but 06-MeG was detected at some point after treatment (Figs. 1 and 2). Adducts accumulated during PCZ treatment, usually reaching about 0.2-0.3 fmole/yg DNA (0.32-0.48 ,mole 06-MeG/mole guanine) at the end ofeach cycle. Slightly higher adduct levels were seen 2 hr after treatment with DCZ. Where the same individuals were followed for more than one cycle of treatment, broadly similar methylation responses were observed. In all cases where data were available for the 24 hr immediately after cessation of treatment, a decrease in adduct levels was observed amounting to 44 8 % (n = 3) for PCZ-treated individuals and 43 ± 19 % (n = 7) for DCZ-treated individuals. As indicated in Figure 3, the accumulation of 06-MeG in PCZtreated individuals as a group is linearly related to cumulative dose (with a slope of 2.8 x 10 4 fmole/jLg DNA per mg/M2 dose) and approaches a plateau above a cumulative dose of 600 mg/i2. In DCZ-treated individuals, the average level of 06-MeG found 2 hr after exposure was 0.20 ± 0.14 fmole/pg DNA, corresponding to a rate of formation of 11 ± 8 x i0 4 fmole/Ag DNA per mg/m2 dose.
As already reported, no consistent changes in lymphocyte AGT levels were observed during PCZ treatment (1) and the average value ofAGT remained constant throughout the cycle. A similar picture was obtained with patients on DCZ treatment, where a) no significant differences were seen between AGT values before and after treatment and b) AGT levels in the same patients prior to the start of repeated cycles were sometimes constant and sometimes changed in either direction (2). The overall picture that emerges from our studies is that no significant AGT depletion occurred in most cases as a result of suicide repair of 06-MeG adducts formed in the context ofthe chemotherapy protocols we examined. This can be explained by the large excess of lymphocyte AGT present in most individuals examined (usually 5-15 fmole4cg DNA) relative to the observed adduct levels and by the de novo synthesis of AGT (17). Significant depletion in lymphocyte AGT was recently reported by Lee et al. (17) in melanoma patients treated with high doses of DCZ (up to 800 mg/m2), about 5-fold higher than the doses used in our protocols. Examination of 06-MeG accumulation in different individuals has provided evidence that interindividual variability of AGT levels may be an important parameter influencing adduct accumulation during exposure to PCZ (1). Based on the currently available data, while an inverse trend between lymphocyte AGT and 06-MeG accumulation is consistently observed, this correlation does not reach statistical significance (Fig. 4). In the case of DCZ exposure, formation of 06-MeG does not show any correlation with lymphocyte AGT before exposure. It is possible that any underlying relationship between adduct formation and AGT repair activity is in this case confounded by differences in individual rates of DCZ metabolism. , makes the assessment of any relationship between the two parameters difficult. However, it is notable that the two individuals in whom therapy failed were the only ones exhibiting an abnormal methylation pattern: patient no. 2 accumulated high levels of adducts up to day 7 and subsequently showed an unexplained rapid loss of adducts, and patient no. 4 accumulated very low adduct levels throughout the treatment cycle.
Animal-to-Human Extrapolations for PCZ, DCZ, and Environmental Methylating Agents A question of fundamental importance in any attempt to use blood leukocyte adduct levels as indicators of tissue-specific therapeutic effectiveness or carcinogenic risk is their relationship to adduct levels in the target cells. Fong et al. (18) have shown that doses of PCZ and methylnitrosourea resulting in similar extents of mammary carcinogenesis in female rats yield different amounts of 06-MeG in total mammary DNA, implying that the cell-specific distribution of adducts may be important in PCZ carcinogenesis. In order to obtain an indication of the tissue distribution of 06-MeG formed by PCZ, rats were treated daily peros with PCZ for 10 days with a dose comparable to that used in the MOPP protocol. Accumulation of similar amounts of 4 . . 06-MeG in the liver and lung and about 2.5-fold lower levels in blood leukocytes, bone marrow and leucocytes occurred (Fig. 6), suggesting that adduct levels in human blood leukocytes may be a good indicator of levels in the bone marrow (target tissue for leukemogenesis) and the lymph nodes (target tissue for chemotherapy). Furthermore, comparison of the rates of accumulation of 06-MeG in rat and human blood leukocytes suggests that humans have 5-fold lower susceptibility than rats to leukocyte 06-MeG induction by PCZ. Figure 7 describes a first approach toward using the above data to extrapolate the expected extent of 06-MeG formation in human blood leukocytes after exposure to DMN. Based on the comparison of 06-MeG formation in rat and human blood leukocytes by PCZ and assuming similar relative species susceptibilities to DMN-induced methylation, current daily human exposure to DMN (0.1-10 tsg) (12,19) would be expected to give rise to 06-MeG levels well below 1 x 10 9 mole/mole guanine. This manuscript was presented at the Conference on Biomonitoring and Susceptibility Markers in Human Cancer: Applications in Molecular Epidemiology and Risk Assessment that was held in Kailua-Kona, Hawaii, 26 October-I November 1991.
The technical assistance of Stella Kaila and Margarita Bekyrou is gratefully acknowledged. This work was partly supported by grants (to S.A.K.) by the Commission ofthe European Communities (contract no. EV4V-0062) and the international Agency for Research on Cancer (Collaborative Research Agreement BRI/89/09).