N -Acetyltransferase 2 Polymorphisms and Susceptibility to Infant Leukemia with Maternal Exposure to Dipyrone during Pregnancy

Epidemiologic data suggest that maternal exposure during the prenatal period contributes to an increased risk of childhood leukemia because the time frame of infant exposure is very small in comparison with older children (1). Infant leukemia (IL) is a rare disease, and it has been suggested that chromosomal translocations involving the mixed-lineage leukemia (MLL) gene may occur in early precursor stem cells during fetal development (2, 3). Although the etiology of IL is not fully understood, several studies have addressed the issue of a mothers' exposure during pregnancy to a variety of carcinogenic agents (pesticides, alcohol, medicine, and topoisomerase II inhibitors in the diet) as potential risk factors. Infants and young children are more vulnerable to the effect of toxicant exposure due to their higher rate of cell growth and division. Additionally, infants can be indirectly affected by toxicants through maternal exposure and because of the physiologic immaturity of the xenobiotic system. Maternal consumption of dipyrone during pregnancy has been associated with an increased risk of IL, MLL rearrangements, and Wilms' tumor, an embryonal malignancy that affects very young children (4-6).

Many carcinogen-metabolizing enzymes are present during gestation. However, the physiologic immaturity and polymorphisms may be important in susceptibility to xenobiotic exposure of infants and young children (7). N-Acetyltransferase 2 (NAT2) is among the important phase II metabolic enzymes that catalyze the conjgation (N-acetylation) and activation (O-acetylation) of carcinogens and heterocyclic amine drugs, including metamizole. Dipyrone, as metamizole is known, is a nonsteroidal anti-inflammatory drug used as a powerful analgesic and antipyretic. In some countries including Brazil, Bulgaria, Egypt, India, Israel, Mexico, Poland, Russia, and Turkey, it is still freely available and plays an important role in self-medication. After p.o. administration, dipyrone is rapidly hydrolyzed by the gastric juice to 4-methyl-amino-antipyrine. 4-Methyl-amino-antipyrine is converted to a variety of metabolites, including 4-formyl-amino-antipyrine and 4-amino-antipyrine. Then, NAT2 acts in the detoxification phase of 4-aminoantipyrine (8).

Genetic polymorphisms of the NAT2 gene result in either rapid or slow acetylation of potentially toxic substances, and they have been shown to modify the risk of childhood acute lymphoblastic leukemia. The NAT2 slow acetylators, characterized by the genotypes NAT2*5C and *7B, were associated with increased acute lymphoblastic leukemia risk (9). Considering the widespread use of dipyrone in Brazil and the possibility that maternal exposure might play a role in the etiology of IL, we questioned whether NAT2 variants in mother-child settings would increase the risk of IL. The rationale for this hypothesis is that dysfunctional metabolism by NAT2 polymorphisms could increase the deleterious effect of dipyrone, and consequently enhance the risk of IL.

This genotyping study is composed of 132 IL cases and 131 control samples. Additionally, 129 samples of mothers (86 mothers of IL cases and 43 mothers of controls) with available DNA were also analyzed. Cases were eligible if they were diagnosed with acute lymphoblastic leukemia at age 21 months or below by immunophenotyping and molecular analysis. The analyses to characterize MLL status were either performed by conventional karyotype, reverse transcription-PCR assay, and/or fluorescence in situ hybridization. Details of the clinical, immunophenotyping, and molecular analyses of this study are described elsewhere (10).

Cases and controls were age frequency matched. They were selected from a previous epidemiologic study of IL evaluating the magnitude of association between IL and maternal exposure to environmental risk factors during pregnancy. We selected all individuals with available biological samples after the exclusion of patients with acute myeloid leukemia. Controls were selected among hospitalized or outpatients children in the same regional hospitals where IL cases came from. The pathologies yielding to their hospitalization were trauma, infectious diseases, sickle cell anemia, and asthma. Children with any pathology resulting from chromosomal abnormalities, such as Down syndrome, and other selected conditions, such as myelodysplastic syndrome, Fanconi anemia, Bloom's syndrome, ataxia telangiectasia, and neurofibromatosis, were excluded.

The mothers of cases and controls were interviewed in person in the hospital with the aid of a well-structured questionnaire and asked to donate a biological sample (peripheral blood or buccal cells) at the time that the children were admitted to the hospital. Questions about demographics including family income, maternal age and education level, maternal history of disease, and reason for use of medication before and during pregnancy were asked and the results were published. The exposure assessment to dipyrone was determined by a qualitative analysis (yes/no) during the index pregnancy (5).

All Brazilian collaborating institutions approved the study, and written consent was obtained from the parents for both the interview and additional samples after the diagnostic procedures (CEP 005/06).

We used PCR followed by RFLP assay, as previously described by Cascorbi et al. (11), to determine the NAT2 variants (191G>A, 282C>T, 341T>C, 481C>T, 590G>A, 803A>G, and 857G>A), as simplified in Supplementary Fig. S1. Briefly, the first PCR amplifies the coding region of the NAT2 gene, generating a 1,211-bp (bp) fragment. This fragment was digested with KpnI (481C>T) and Bam HI (857G>A). Subsequently, 0.2 μL of the first PCR product was used to amplify a 442-bp fragment in a second PCR. The second PCR product was digested with Msp I (191G>A), FokI (282C>T), and DdeI (341T>C). Finally, the third PCR was done using 0.2 μL of the first PCR product to generate a fragment of 421 bp, which was then digested with TaqI (590G>A) and DdeI (803A>G).

The phenotype rapid or slow responders were defined according to the combination of allele variants that was first defined by Vastis et al. (12). Rapid responders were characterized by the presence of at least one copy of the NAT2*4, NAT2*11, NAT2*12, or NAT2*13 variant; slow responders were characterized by the presence of two alleles with slow acetylation capacity according to an international nomenclature committee (13).

The expected gene polymorphism frequency was calculated using the Hardy-Weinberg law based on the allele frequency in the control group. The χ² test was used to examine the difference in allele frequency between NAT2 variants in IL cases and controls. The polymorphism frequency in children and mothers was compared with that observed in the control group by χ² tests. The level at which results were considered significant (P ≤ 0.05) was calculated by χ² test and Fisher's exact test. Unconditional logistic regression was used to examine the association between maternal exposure to dipyrone during pregnancy, IL, and NAT2 phenotypes. In consideration of selected confounders such as hormonal intake during pregnancy and pesticide exposure, crude and adjusted odds ratios (OR) are given with a 95% confidence interval (95% CI). The Mantel-Haenszel ORs on the association between NAT2 maternal phenotypes and IL according to dipyrone maternal exposure were ascertained. The statistical package IBM SPSS Statistics 14.0 version was used.

The genotypes of seven polymorphism sites in the NAT2gene were successfully obtained in 132 IL cases, 131 healthy controls, and 110 samples from mothers. The genotypic variant distributions are in Hardy-Weinberg equilibrium in both the case and control groups, as well as in both the mothers of IL cases and mothers of controls (Supplementary Table S1). Table 1 lists the demographic characteristics and laboratory data. Maternal dipyrone use during pregnancy showed a borderline association with IL risk (OR, 1.78, 95% CI 0.98-3.23). For the frequency of NAT2 alleles of IL cases and controls, the population was divided into two subgroups (slow and rapid) according to the combination of NAT2 variants with the respective phenotype profile. There was a difference in the allele frequency distribution between cases and controls. The *4 wild-type haplotype was less frequent in cases than among controls. The allele variant NAT*5 (cluster 341T>C) was more frequent in the whole population. The sum of rapid-acetylation alleles was more frequent among controls, whereas the slow-acetylation alleles were more frequent among IL patients. The NAT2 alleles *5, *6, and *7 showed a 2-fold statistically significant association with IL risk (Table 1).

The distribution of NAT2 diplotypes is shown in Table 2. The combination of alleles that is a predictor of a NAT2 slow-acetylation genotype (two copies of any slow allele *5, *6, *7, or *14) was present in 67.4% of the IL group, whereas the combination was underrepresented in controls (34.3%). The distribution of NAT2 genotypes suggests that the slow acetylation activity of NAT2 is associated with an increased risk of IL (OR, 8.90; 95% CI, 1.71-86.7).

The frequencies of the NAT2 alleles as well as the genotype distribution in mothers of cases and mothers of controls were also analyzed. Allele frequency among the mothers of cases differed from that in the mothers of controls. Mothers of IL cases had higher frequencies of NAT2*5A (15.8%) and NAT2*5D (6.7%). The sums of allele variants that confer slow phenotypes were more frequent in the mothers of IL cases (82.5%), whereas allele variants that confer the NAT2 rapid phenotype were more frequent among the mothers of controls (50.8%).

The effects of children's NAT2 phenotype, children's age range, and maternal exposure to dipyrone during pregnancy in cases and controls are presented in Table 3. The NAT2 slow-acetylation phenotype exhibited an increased positive association with IL in children younger than 12 months [adjusted (pesticides and hormone intake) OR, 5.14; 95% CI, 2.42-10.9]. A quite similar association pattern between slow-acetylation phenotype and IL was observed regardless of maternal dipyrone exposure during pregnancy. Mantel-Haenszel OR on the association of NAT2 slow phenotype and IL according to maternal dipyrone intake during pregnancy was 3.77 (95% CI, 1.90-7.64).

To further explore the association of NAT2 phenotype and IL with MLL rearrangements, data were stratified according to age (≤12 and 13-21 months old) and MLL gene status (positive, negative). There was no significant association of the NAT2 acetylator profile and IL with MLLrearrangements (Supplementary Table S2).

The effect of the joint mother/child NAT2 phenotype was assessed to evaluate the risk association with IL (Table 4). Using the NAT2 acetylator phenotype of mother rapid/child rapid combination as reference, an OR of 9.0 (95% CI, 2.28-42.1) was observed for the mother slow/child rapid NAT2 acetylators combination, and an OR of 30.0 (95% CI, 5.87-279.7) for the mother slow/child slow NAT2 acetylators combination. The effect of joint mother/child NAT2 phenotypes on the risk of IL with MLL rearrangements was also ascertained, showing quite similar risk patterns according to MLL status (Supplementary Table S3).

Polymorphisms in genes encoding xenobiotic-metabolizing enzymes are considered to be risk factors for childhood leukemia (9, 14, 15). The current study is the first to suggest a significant association between NAT2 slow alleles and the risk of IL. There are substantial age-related differences in susceptibility to environmental metabolism, mainly during fetal and the early postnatal period, when metabolic systems are still very immature. Additionally, the environmental exposures are greatly modified by the maternal system and placental metabolism (7). The effect of the NAT2 acetylation phenotype on IL in children whose mothers used dipyrone during pregnancy was thus explored.

First, the NAT2*5 variant that characterizes slow acetylation was the most frequently observed allele in IL. Additionally, the slow-acetylation alleles NAT2*6 and *14 in Brazilian children were also associated with increased risk of IL. These alleles are common in the Asian (NAT2*6) and African (NAT2*14) populations (16). The fact that Brazil is characterized by a mixed population composed of different ethnicities (European, African, Asian, and Amerindian ancestry) could explain the observed frequency of such NAT2*6 and *14 allele variants. The race distribution for our sampled population was based on the self-definition of skin color during the interview, and no differences between cases and controls were observed. The high frequency of NAT2 slow phenotype found among the IL subjects suggests that a deficient detoxification function of genotoxic metabolites contributes to IL pathogenesis.

Second, as hypothesized, these results are suggestive that infants exposed to dipyrone during fetal life may have an increased risk of developing leukemia in early infancy. Despite the relatively low risk of IL associated to maternal dipyrone exposure, this substance still may present a high attributable risk, considering its widespread population use. The 4-fold increased risk of leukemia in infants with the NAT2 slow-acetylation profile was similar regardless of the reported maternal dipyrone use during pregnancy. As a whole, these results are suggestive that the slow acetylator profile seems to be one of the most important risk factors in the causal effect of the leukemogenesis pathway. The fetal and early postnatal periods differ from the more mature stages of development in that a number of pathways are able to affect the metabolism of carcinogenic agents. The largest differences generally occur in the fetal period and during first infancy and include basic physiologic properties such as immaturity of the hepatic and renal systems (17).

Finally, we observed that the presence of just one NAT2 slow acetylator allele, either from the mother or the child, increases the risk of IL. This risk was also enhanced when a combination of the two NAT2 slow acetylator alleles was present. The maternal and placental system combination in the NAT2 acetylator phenotype is very important in completing the detoxification of carcinogen metabolites. Although available data on animal models are scarce, an effect of acetylator genotype on 3,2-dimethyl-4-aminobiphenyl in Syrian hamster congenic lines with NAT2*16Apolymorphism and colon cancer has been reported. One of the main findings in that study was that the same effect was found in two lines of hamsters that are congenic at the NAT2 acetylator locus (18). The NAT2 enzyme activity is first detectable in the placental-fetal unit, protecting the fetus from toxicant exposure (19); thus, we supposed that when the duo (mother-child) dysfunctional metabolism of NAT2 is present, the longer “low-acetylation” might increase the deleterious effect of dipyrone metabolites. A study on the effects of dipyrone on purified CD34⁺ marrow cells showed not only a drug-dependent suppression of the in vitro growth of myeloid progenitors but also a drug-dependent suppression of primitive multipotential progenitors (20). However, there is no experimental study with substantial evidence that dipyrone metabolites are associated with IL, thus far.

The most common genetic alterations in IL are rearrangements of the MLL gene with diverse gene partners that can lead to a common or different effect in leukemogenesis. Therefore, other factors that affect the propensity for drugs or their metabolites to reach a higher concentration and cause DNA-adducts in the fetus should be considered.

One of the limitations of the present study is related to the sample size, especially among controls, for which fewer molecular profiling in a mother-child setting were performed. The lack of information on other environmental exposures that could have biological effects depending on the acetylator variant genotypes might also have introduced confounding factors. On the other hand, some strengths of this investigation can be mentioned. It is the first one to show evidence on the association between NAT2 slow-acetylation profile and IL, analyzing biological samples from the mother and the child. Moreover, the study results highlight the quantitative contribution of NAT2 slow-acetylation alleles in IL development, showing a 9- to 12-fold increased risk whether such allele originated in the mother or the child, and a 30-fold risk increase when the allele is present in both. Such findings also show evidence that the in utero NAT2 detoxifying metabolism (rapid acetylation profile either at the mother or the child) is equally significant in preventing leukemogenesis. The associations observed in this study need to be further validated with larger samples and functional studies.

In conclusion, the present data suggest that the NAT2 acetylation profile is an important risk factor associated with IL but does not support the hypothesis that the dysfunctional metabolism by NAT2 polymorphisms increases the deleterious effect of dipyrone.

Jane Dobbin¹, Fernando de Almeida Werneck¹, Silvia Maia Farias de Carvalho¹, Claudia Julia¹, Cynthia Curvello Neves², Jozina Maria de Andrade Agareno², Lilian Maria Burlacchini de Carvalho³, Flávia Nogueira Serafim Araújo², Maurício de Souza Meira³, Nilma Pimentel de Brito³, Isis Q. Magalhães⁴, Jose Carlos Cordoba⁴, Flávia Pimenta⁵, Eloísa Cartaxo⁵, Gilberto Ramos⁶, Rosania Maria Basegio⁷, Atalla Mnayarji⁷, Alejandro Arenciba⁸, Renato Melaragno⁸, Eduardo Preto Serafini⁹

Affiliations: 1. Hospital Câncer 1-INCA, Hospital dos Servidores do Estado, and HEMORIO, Rio de Janeiro (n, 29); 2. Sociedade de Oncologia da Bahia, Salvador-Bahia (n, 26); 3. Hospital Martagão Gesteira, Salvador-Bahia (n, 24); 4. Hospital de Apoio Brasília, Unidade de Onco-Hematologia Pediátrica, Brasília, DF (n, 27); 5. Hospital Napoleão Laureano, João Pessoa, Paraíba (n, 28); 6. Departamento de Pediatria, Faculdade de Medicina, UFMG, Belo Horizonte, MG (n, 17); 7. Hospital Pedro Pedrossian, Campo Grande, MT (n, 28); 8. Hospital Santa Marcelina, São Paulo, SP (n, 23); 9. Hospital Geral de Caxias do Sul, RS (n, 26).

All authors disclose that no financial or personal relationships with other individuals or organizations have inappropriately influenced this study.

Brazilian National Research Council (CNPq), the Instituto Nacional de Câncer-Fundação Ary Frauzino, and the Swiss Bridge Foundation. M.S. Pombo-de-Oliveira and S. Koifman were supported by CNPq research scholarships (#309091/2007 and # 577598/2008-2, respectively). The project was funded by INCT-Controle do Cancer, CNPq #573806/2008-0, and FAPERJ E026/2008.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.