Balkan Journal of Medical Genetics

Our results demonstrate that particular variant alleles of the biotransformation genes are associated with a higher risk of relapse in childhood acute leukemia. Most of the regularities we observed agree with the published data [16-20]. For instance, the association of polymorphic variant CYP1A1*2A with poor therapeutic prognosis has been shown in children with ALL (OR = 3.08, 95% CI = 1.62-5.88, p = 0.0011) [16]. An increase in frequency of the CYP1A1 genotype *1/*2A in children with ALL relapse can be explained by synthetic glucocorticoids (such as dexamethasone and prednisolone) being an important component of the ALL treatment protocol and in induction therapy. They cause lymphopenia and involution of lymphoid tissue that lead to immunosuppression. Published data demonstrate that they can potentiate the activity of CYP1A1 by action through the glucocorticoid-responsive elements [16,26]. The presence of a polymorphism in the CYP1A1 gene increases the enzymatic activity and, leads to increased concentration of intermediate genotoxic metabolites and of total mutagenic activity [14,19]. Since formation of additional mutations may cause resistance of cancer cells towards therapy, it is likely that the CYP1A1 *1/*2A genotype may decrease the efficacy of therapy and promote development of relapse [16].

We also observed that patients with the GSTT1 and/or GSTM1 null genotype have a lower risk of recurrence when compared with those with one or two functional alleles [17,18,20]. The GSTs are involved in metabolism of many anti tumor drugs, catalyzing conjugation of intermediate metabolites with reduced glutathione. In addition, they can determine resistance to chemotherapeutic agents [27]. Thus, ALL children carrying the GSTT1 null genotype have been observed to respond well to induction therapy with prednisone, while carriers of at least one functional GSTT1 allele displayed glucocorticoid resistance, a poor response to therapy, and a higher relapse rate [20]. The total 5-year survival in high-risk ALL has been shown to be substantially higher in children carrying the GSTM1 null genotype than in carriers of the non null genotype (p = 0.03) [17]. Our results suggest that, in carriers of the GSTT1 and/or GSTM1 null genotype, lack of GSTT1 and/ or GSTM1 enzymes leads to an accumulation of cytotoxic drugs, with enhancement of their efficacy and longer relapse-free survival.

The ALL-MB-02 treatment protocol does not include oxazaphosphorines. The cytostatics cyclophosphamide and ifosfamide are widely used for treating ALL according to the ALL-BFM protocol and exert an immunosuppressive effect by inhibiting proliferation of lymphocyte clones involved in the immune response. The GSTs catalyze conjugation of cyclophosphamide with glutathione [28]. Also among children with steroid-dependent nephropathic syndrome, carriers of the GSTM1 null genotype have a far greater relapse-free survival when compared with carriers of one or two functional GSTM1 alleles [29]. Thus, the GSTM1 null genotype improves the efficacy of cyclophosphamide: lack of GSTM1 allows accumulation of the cytotoxic and immonosuppressing metabolite 4-OH-cyclophosphamide which is a substrate of the enzyme GSTM1 [30].

An important component of induction therapy for ALL is the anthracyclines (e.g., daunorubicin and doxorubicin), which exert an immunosuppressive effect. It has been shown [31] that daunorubicin in vitro causes induction of GSTs, which may account for the resistance to cytotoxic effects. However by increasing the efficacy of the anthracyclines, the GSTs null genotypes may increase the risk of certain adverse reactions. Moreover, among AML patients with the GSTT1 null genotype, those receiving low doses of anthracyclines had a better prognosis than those who received high-dose therapy [32].

We found that the presence of both the CYP1A1 genotype *1/*2A and the GSTT1 and/or GSTM1 non null genotype has an additive effect on the risk of acute leukemia relapse. The polymorphic variant CYP1A1*2A has been associated with a higher risk of acute leukemia relapse, whereas the GSTT1 and/or GSTM1 null genotype exerts a protective effect [16-20]. Our results suggest that therapy is far more effective in patients carrying no CYP1A1*2A allele, increasing CYP1A1 activity, neither GSTT1 and/or GSTM1 functional alleles, where the lack of function favors the accumulation of cytotoxic metabolites and strengthens the therapeutic effect.

The NATs catalyze the addition of the acetyl group to the terminal nitrogen of arylhydrazines and arylamine-containing drugs and carcinogens [33]. The NAT2 is the main acetylating enzyme, while NAT1 acetylates fewer arylamines. Three main slow alleles which cause a slow NAT2 acetylation phenotype: NAT2*5 (T341C, C481T), NAT2*6 (G590A), and NAT2*7 (G857A), and one fast (wild type) allele NAT2*4, form the basis for comparing genetic and biochemical data. A combination of a fast and a slow allele yields an intermediate acetylation phenotype [34,35]. According to published data, the NAT2 polymorphic variants that determine the slow acetylator phenotype decrease either the activity or the stability of the enzyme. For instance, substitution of C for T at position 341 causes the substitution Ile^{114Thr and reduces the N-acetylation rate, while the substitutions}Arg^{197Gln and Gly286Glu}result in a less stable enzyme [33-36].

The NAT2 alleles can modulate the risk of various tumors [37-39], including acute leukemia [11]. We observed two characteristic NAT2 genotypes: 341C/–, 481T/–, 590G/G, 857G/G for ALL relapse and 341T/T, 481C/C, 590A/– for AML relapse. Since the intermediate/slow acetylator phenotype occurred at a higher frequency in both cases, the difference in genotype could be determined by the structure of NAT2 substrates. Since methotrexate and cytarabine are aryl and heterocyclic amines they may most probably be substrates for NAT2 [11,40] (Figure 1). Both are employed in acute leukemia: methotrexate for ALL in the consolidation phase, and cytarabine throughout AML therapy. NAT2 may metabolize 6-mercaptopurine (Figure 1) and its analogs used in the ALL therapy, but these are metabolized mainly by thiopurine-S-methyltransferase [6,22]. Thus, we speculate that the intermediate/slow acylator phenotype promotes accumulation of these drugs and their reactive metabolites in childhood acute leukemia. These metabolites cause multiple chromosome aberrations in bone marrow cells [41,42], which, with other adverse factors, lead to relapse.

An interesting finding is that patients with ALL relapse showed a significant increase in the frequency of the NAT2 genotype 341C/–, 481T/–, 590G/G, 857G/G in combination with GSTT1 and/or GSTM1 non null genotype when compared with primary leukemia patients, while patients with AML relapse displayed a higher frequency of the NAT2 genotype 341T/T, 481C/C, 590A/– in combination with the GSTT1 and/or GSTM1 non null genotype. This leads us to suggest that for carriers of "non null" GSTT1 and/or GSTM1 in combination with NAT2 genotype 341C/–, 481T/–, 590G/G, 857G/G (in children with ALL) or 341T/T, 481C/C, 590A/– (in children with AML), therapy is less effective than for carriers of the other genotypes. The presence of functionally significant alleles GSTT1 and/or GSTM1 causes effective detoxification of anti cancer drugs by the enzymes GSTT1 and/or GSTM1, and their elimination with a decrease in their effectiveness, while "unfavorable" NAT2 genotype favors accumulation of genotoxic agents which can provoke chromosomal aberrations in bone marrow cells and thus lead to relapse.

The analysis of genotype frequency for CYP1A1, GSTs and NAT2 genes in subgroups of patients with acute leukemia, divided according to their gender (boys and girls, separately), showed the same regularities as in total group of patients. It is worth noting that the differences found in the group of boys were more pronounced that those in the group of girls. The results are in good agreement with the data that boys not only more frequently develop the disease, but usually have a poorer prognosis due to higher risk of relapse [43-45]. The reason is not clearly understood, but it is likely that due to sexual difference in the metabolic work of biotransformation enzymes the unoptimized anti cancer therapy becomes more critical for boys than for girls [43].

To summarize, we have shown that a wide range of CYP1A1 allele combinations can be identified using the biochip designed to assess the biotransformation genes. The CYP1A1, GSTT1, GSTM1 and NAT2 polymorphic alleles are prognostic for increased risk of acute leukemia relapse in children. Our method can be employed in a search for other variant alleles that affect the risk of leukemia relapse and in clinical diagnosis aimed at individualizing therapy.