Polymorphisms in the glutathione S-transferase theta and mu genes and susceptibility to myeloid leukemia in Brazilian patients

The null genotype for glutathione S-transferase (GST, EC 2.5.1.18) gene polymorphisms is considered a risk factor for leukemia in different populations. In this work we investigated the GSTT1 and GSTM1 polymorphisms using multiplex PCR in 53 patients with chronic myeloid leukemia (CML), 23 with acute promyelocytic leukemia (APL) and 304 apparently healthy controls. In this association study we found that the GSTT1 genotype was more frequent in our group of APL patients than in the control group [OR = 2.75 (95% CI = 1.10-6.88)], providing evidence that a deletion in the GSTT1 gene could be a risk factor for this type of leukemia.

Leukemias are complex diseases determined by a combination of several factors.It has been shown that DNA damage in hematopoietic precursor cells is directly linked to the risk of acute leukemia in adults (Rollinson et al., 2000) and may results from an interaction between reactive species generated by environmental or endogenous metabolites (Rollinson et al., 2000;Dalo et al., 2004).Human cells possess metabolic systems to eliminate toxic agents and several enzymes are responsible for the degradation of these xenobiotics, one systems being the glutathione S-transferase (GST, EC 2.5.1.18)group of enzymes which detoxify environmental carcinogens by conjugation with glutathione (Crump et al., 2000).The GST group is known to be coded for by 16 genes in six GST subfamilies, known as alpha (GSTA), mu (GSTM), omega (GSTO), pi (GSTP), theta (GSTT) and zeta (GSTZ).Two widespread genetic polymorphisms that involve deletions in GSTT1 and GSTM1 have been reported to lead to loss of enzyme activity (Bolufer et al., 2007) and have been investigated in many different populations, including those from Japan (Naoe et al., 2000), Italy (D'alo et al., 2004) and Spain (Bolufer et al., 2007).Furthermore, several studies have proposed that susceptibility to acute and chronic myeloid leukemia (AML and CML respectively) could be related to GSTT1 and/or GSTM1 deletions (Rollinson et al., 2000;Mondal et al., 2005;Ye and Song, 2005.Brazil is the largest country in South America, with a highly heterogeneous population due to several waves of immigration which have resulted in cultural, socioeconomic and ethnic diversity.In the Northeastern Brazilian state of Bahia, which has a highly mixed population of mainly African descent, the largest city is Salvador (population 2.7 million; RIPSA, 2006), 86% of the local population being of African, or European and African, descent.A recent study by Barreto et al. (2006) reported that the Bahian population shows a high prevalence of pediatric acute promyelocytic leukemia (APL), which accounted for 21% of all AML patients (n = 105) evaluated between 1995 and 2004).Ribeiro and Rego (2006) reported that patients with a Latin American background (i.e.some Amerindian genetic input) were much more likely to have APL (18.2%) than were white (7.7%) or black (10.3%) patients without a Latin American background.Based on this, we decided to investigate the frequency of GSTT1 and GSTM1 polymorphisms among patients with APL and CML from the city of Salvador.
We investigated a total of 76 patients with myeloid leukemia, 53 (30 (57%) male, 23 (43%) female; mean age 41 y ± a standard deviation (SD) of 21 y) with CML and 23 (11 (48%) male, 12 (52%) female; mean age 14 y ± 6.8 y) with APL which were selected at diagnosis between 2000 and 2003 from several institutions that attend patients with malignant diseases.Most of the APL cases were from a pediatric oncological institute (Clínica de Oncologia, Salvador, Bahia).The control group consisted of 304 (131 (43.2%) male, 173 (56.8%) female ; mean age 29 y ± 9.5 y) was composed of apparently healthy individuals randomly chosen from the staff of a private clinical laboratory (Labcheckup, Pituba, Salvador, Bahia).Both patients and control individuals all came from Bahia.The diagnoses of leukemia were made according to clinical, morphological and molecular criteria and the study was approved by the Oswaldo Cruz Research Foundation's Human Research Ethics Committee.Bone marrow and peripheral blood samples were obtained only after an informed consent form was signed.
We extracted RNA, using Trizol R (Gibco-BRL, Life Technologies, USA), and DNA, using Genomic Blood DNA Purification Kits (Amersham Pharmacia Biotech, USA), from bone marrow cells and peripheral blood leukocytes according to the guidelines of the manufacturer.The translocations t(9;22)(q34;q11) for CML and t(15;17)(q22; q12-21) for APL patients were investigated using the reverse transcriptase polymerase chain reaction PCR (RT-PCR) method according to the methodology of Biernaux et al. (1995) and Miller et al. (1993).The GST polymorphisms were assessed using the multiplex PCR method described by Arruda et al. (2001), using the β-globin gene as an internal control.
Descriptive analyses included genotype frequencies and the odds ratio (OR) as an estimate of relative risk, with 95% confidence intervals (CI).Chi-square test using the Yates correction or the Fisher's exact test were applied and differences were considered significant at the p level.All analyses were carried out using the EPI INFO software, version 6.04 (Centers for Disease Control and Prevention, Atlanta, GA, USA).
All 53 CML patients had the t(9;22)(q34;q11) translocation and all 23 APL patients had the t(15;17)(q22; q11-21) translocation.The frequencies of the GSTT1 and the GSTM1 genotypes in both patients and controls are presented in Table 1.Although we found similar patterns for the GST multiplex PCR it is possible that the loss of heterozygosity in peripheral blood or bone marrow cells may have increased the frequencies of the GSTM1 null and GSTT1 null genotypes.However, the frequency distributions of these polymorphisms were in agreement with the Hardy-Weinberg equilibrium.No association was found between the GSTT1 or GSTM1 deletions and CML risk in the group studied but there was an increased APL risk for the GSTT1 null /GSTM1 normal (OR = 2.75, 95% CI = 1.1 to 6.88) and GSTT1 null /GSTM1 null (OR = 3.61, 95% CI = 1.37 to 9.51) genotypes (Table 1).
Several studies associating the presence of the GSTM1 and GSTT1 polymorphisms with lymphoid and myeloid leukemias have been performed (Crump et al., 2000;Naoe et al., 2000;Dalo et al., 2004;Mondal et al., 2005;Bolufer et al., 2007).Our analyses showed that the GSTT1 gene deletion was significantly higher among Brazilian APL patients than controls but no association was observed between APL susceptibility and the isolated GSTM1 null genotype, although when the GSTM1 null /GSTT1 null genotype was considered there was a significant difference between the APL group and the control group.The APL type of disease is a French-American-British (FAB, Bennett et al., 1976) classification subtype leukemia, also named AML-M3, frequently presenting the t(15;17)(q22;q12-21) translocation.We suppose that the deletions may lead to lack of detoxification of electrophilic compounds and/or a higher DNA damage ratio, which could contribute to the development and proliferation of leukemia.Rollinson et al. (2000) observed similar results, but found a weaker association between the GSTM1 or GSTT1 deletions and increased risk to AML, the same FABsubtype leukemia, in British adults.The differences between these studies may be related to age variation or to the diverse genetic backgrounds of the patients.Arruda et al. (2001) found an association between the development AML and the GSTM1 null or GSTT1 null genotypes among Brazilian individuals, but they did not separate the FAB subgroups.Ye and Song (2005) performed a systematic review of several studies of GST gene polymorphisms in relation to the risk of acute leukemia, their results suggesting that the GSTM1 null and GSTT1 null genotypes are not associated with AML.Our results, however, surprisingly showed a differ-40 Souza et al. ent pattern of GSTT1 polymorphism frequency among the APL patient group.Since the GSTT1 enzyme is responsible for the detoxification of environmental xenobiotics these results may be associated with high rates of APL in the investigated population.Barragan et al. (2007) recently suggested the influence of GST deletions on treatment follow up after chemotherapy in adult non-promyelocytic patients.
We observed similar frequencies in CML patients and controls but Lourenco et al. (2005) found a lower frequency of the GSTT1 null genotype in CML Brazilian patients who were both in the blast crisis or in the chronic phase, while Mondal et al. (2005) observed an increase in the GSTT1 null genotype in CML patients from India.
Our population was composed of an admixture of Amerindian, African and European-derived subjects, and the GST polymorphisms is known to exhibit different frequencies according to ethnic group (Gattas et al., 2004).Our results were different from those for individuals of European-descent from the Southeastern Brazilian state of São Paulo reported by Gattas et al. (2004), who found frequencies of 22.3 for the GSTT1 null and 55.4% for the tGSTM1 null genotypes.This difference can be explained by the high percentage of African genes present in Salvador population.
In conclusion, our analyses suggest that the GSTT1 genetic background might be an important marker for APL risk, at least in Salvador.