Frequencies of 23 Functionally Significant Variant Alleles Related with Metabolism of Antineoplastic Drugs in the Chilean Population: Comparison with Caucasian and Asian Populations

Cancer is a leading cause of death worldwide. The cancer incidence rate in Chile is 133.7/100,000 inhabitants and it is the second cause of death, after cardiovascular diseases. Most of the antineoplastic drugs are metabolized to be detoxified, and some of them to be activated. Genetic polymorphisms of drug-metabolizing enzymes can induce deep changes in enzyme activity, leading to individual variability in drug efficacy and/or toxicity. The present research describes the presence of genetic polymorphisms in the Chilean population, which might be useful in public health programs for personalized treatment of cancer, and compares these frequencies with those reported for Asian and Caucasian populations, as a contribution to the evaluation of ethnic differences in the response to chemotherapy. We analyzed 23 polymorphisms in a group of 253 unrelated Chilean volunteers from the general population. The results showed that CYP2A6*2, CYP2A6*3, CYP2D6*3, CYP2C19*3, and CYP3A4*17 variant alleles are virtually absent in Chileans. CYP1A1*2A allele frequency (0.37) is similar to that of Caucasians and higher than that reported for Japanese people. Allele frequencies for CYP3A5*3(0.76) and CYP2C9*3(0.04) are similar to those observed in Japanese people. CYP1A1*2C(0.32), CYP1A2*1F(0.77), CYP3A4*1B(0.06), CYP2D6*2(0.41), and MTHFR T(0.52) allele frequencies are higher than the observed either in Caucasian or in Japanese populations. Conversely, CYP2C19*2 allelic frequency (0.12), and genotype frequencies for GSTT1 null (0.11) and GSTM1 null (0.36) are lower than those observed in both populations. Finally, allele frequencies for CYP2A6*4(0.04), CYP2C8*3(0.06), CYP2C9*2(0.06), CYP2D6*4(0.12), CYP2E1*5B(0.14), CYP2E1*6(0.19), and UGT2B7*2(0.40) are intermediate in relation to those described in Caucasian and in Japanese populations, as expected according to the ethnic origin of the Chilean population. In conclusion, our findings support the idea that ethnic variability must be considered in the pharmacogenomic assessment of cancer pharmacotherapy, especially in mixed populations and for drugs with a narrow safety range.


INTRODUCTION
Cancer is a leading cause of death worldwide and the total number of cases globally is increasing. The number of cancer deaths is projected to increase 45% from 2007 to 2030 (from 7.9 million to 11.5 million deaths), influenced in part by an increasing and aging global population. The estimated rise takes into account expected slight declines in death rates for some cancers in high resource countries. New cases of cancer in the same period are estimated to increase from 11.3 million in 2007 to 15.5 million in 2030 (WHO, 2011). In Chile cancer have a rate of 133.7 × 100,000 inhabitants, is the second cause of death after cardiovascular diseases with a sustained increase in the time both, in the rates and in proportion of deaths. Chile has a number of 30,000 new cases and 36,000 hospitalizations per year. The first cause of cancer death in Chilean population is gastric cancer for both genders, with a rate of 19.8 × 100,000 inhabitants, followed by lung cancer (15 × 100,000 inhabitants), prostate (10.4 × 100,000 inhabitants), and gallbladder (8.1 × 100,000 inhabitants; DEIS, 2011).
Though the principal achievements in the fight against cancer has been those related to preventive measures, the success in the treatment of an established cancer depends on the cancer stage when detected, on the type of cancer, and on its location. It is quite infrequent to find patients considered as "therapeutical success" because there is always a possibility that a tumor process may appear back (Arrastoa, 1998).
Chemotherapy for cancers has progressed from its introduction to clinical practice and constitutes the mainstay modality of therapy in these pathologies. Nevertheless, its use is limited by the inability to predict the response. In most cases the therapy choice is empirical. Nowadays there are more than 100 antineoplastic drugs, which are used either alone or combined. A combined therapy allows that drugs with different mechanism of action work together to destroy the larger possible amount of tumor cells, in order to reduce the possibility of resistance to a particular antineoplastic drug. The therapy to select, the dose, the method of administration and the frequency, and duration of the treatment, will depend on the type of cancer, its location, the rate of growth, how it is affecting the normal functions of the body, as well as on the patient's general health condition. Usually therapies can be managed by means of cycles that alternate drug administration with washout periods that allow healthy cells to recover from the effect of the medication.
The biotransformation of drugs, including antineoplastic agents, is done basically in two phases: phase I, catalyzed mainly by the cytochrome P450 system and phase II, by transferases that catalyze reactions of conjugation of xenobiotics with diverse molecules of endogenous nature as glucuronic acid, sulfates, acetate, glutathione, or some amino acids. The final outcome of both phases is to increase the hydrophilicity of chemicals facilitating the excretion through urine or bile (Rooney et al., 2004).
There is limited information available in human beings regarding the metabolism and pharmacokinetics interaction of antineoplastic drugs. Nevertheless, it is well known that clinical significant interactions exist between drugs that can render them less efficient if used simultaneously, and in some cases, produce unexpected effects.
The cytochrome P450 (CYP) system is the most important metabolic system responsible for the oxidation of numerous chemotherapeutic agents, and it is responsible to a great extent for the variability observed in response to drugs. For example, some enzymes of the CYP3A family play an important role in the metabolism of epidophilotoxin, ifosfamide, tamoxifen, taxol, and vinca alkaloids. Cytochrome P450 enzymes are also of great importance in the study of chemotherapy resistance. (Kivisto et al., 1995;Yao et al., 2000;Lin and Yu, 2001;García-Martín et al., 2006a;Quiñones et al., 2008).
Another relevant enzyme is methylene tetrahydrofolate reductase (MTHFR) which converts 5,10-methylene tetrahydrofolate to 5-methyl tetrahydrofolate. This reaction is required for the multistep process that converts the amino acid homocysteine to another amino acid, methionine. Polymorphisms of this enzyme predispose to serious bone marrow toxicity during treatment with agents that inhibit folate synthesis (e.g., methotrexate; Chiuslo et al., 2002;Yang et al., 2012).
Because of enzymes CYP1A1, CYP1A2, CYP2A6, CYP3A4/5, CYP2C8, CYP2C9, CYP2C19, CYP2E1, CYP2D6, GSTM1, GSTT1, UGT2B7, and MTHFR take part in the metabolism of oncological drugs (Table 1), the main goal of this study was to determine the allele frequencies of variants of these enzymes in a group representative of the Chilean population in order to describe genetic polymorphisms that might be useful in public health programs, and to compare these frequencies with other populations, as the first approximation to the evaluation of ethnic differences in the response to chemotherapies.

STUDY POPULATION
Blood samples were obtained from 253 unrelated volunteers living in Santiago of Chile (both sexes, 27-55 years old). The study group has a 37% Amerindian-Caucasian admixture, as determined by ABO blood group distribution (Valenzuela, 1988). The research was authorized by the Ethics Committee of the Faculty of Medicine of the University of Chile. All subjects signed an authorized Frontiers in Genetics | Pharmacogenetics and Pharmacogenomics informed consent. Table 1 shows general characteristics of the studied population.

DNA EXTRACTION
Extraction of genomic DNA was done by a standard procedure from whole blood using a commercial kit (High Pure PCR Template Preparation Kit, Roche Diagnostics®) and DNA samples were stored at -20˚C until further analysis.

RESULTS AND DISCUSSION
Pharmacogenetic research is directed to identify genes or gene products associated with diseases and, especially, allelic variants in enzymes of biotransformation that alter the individual response to drugs. These variants can modify the magnitude of the pharmacological effect, the threshold of toxicity, the efficacy of the drug, side effects, and drug-drug interactions. In this respect, it is particularly important at the time, to define pharmacogenetic profiles of patients with cancer to determine suitable dosages, to improve efficacy, to avoid adverse reactions of the traditionally used drugs, and to develop new drugs according to the genetic -metabolic profile of the patients (Wilkinson, 2000). Moreover, ethnicity plays an important role in pharmacokinetics and pharmacodynamics of drugs (Ling and Lee, 2011;Kurose et al., 2012) giving rise a more complex situation in "mestizo" populations as South American people. In this sense, Chilean population is a genetic admixture originated primarily between Caucasian (mainly Spaniards) and native-American (mainly Mapuches) from a first single migration of Asians from Siberia 15,000 years ago through Beringia (Reich et al., 2012) and secondly from immigration, mostly from Germany, Croatia, France and Italy. This is a restriction to extrapolate the dosage of drugs with clinical studies performed in other ethnic groups. Another restriction is the poor information about the biotransformation enzyme polymorphisms in Chilean population. According to this, we have studied genetic polymorphisms of several enzymes, in a sub-group of the Chilean population, which metabolize mainly antineoplastic drugs used for chemotherapy in health institutions of Chile (Table 1). General characteristics of the analyzed population are shown in Table 2. The mean age identifies a young adult population which has, in average, a normal mean weight, height, and body mass index and are a representative sample of the middle class Chilean population, which is supported by the Amerindian-Caucasian admixture (37%). The allele and genotype frequencies for metabolic enzymes included in this research are shown in the Table 3. Due to DNA shortage, not all DNA samples were analyzed for all polymorphisms. In Table 4 we compare allele frequencies found in this study in relation to Caucasian and Asian populations.
Our results show the absence of variant alleles CYP2A6 * 3, CYP2C19 * 3, and CYP3A4 * 17 such as it was also observed for Caucasians. In Japanese people CYP2A6 * 3 and CYP3A4 * 17 are also absent, but the CYP2C19 * 3 allele has a frequency of 0.11. CYP1A1 * 2A and CYP2A6 * 2 allele frequencies are similar to Caucasians but higher than the reported for Japanese people. CYP3A5 * 3 and CYP2C9 * 3 frequencies are similar to those in Japanese people, but different to the Caucasian people. CYP1A1 * 2C, CYP1A2 * 1F, CYP3A4 * 1B, and CYP2D6 * 2 allelic frequencies are higher than those observed either in Caucasian or Japanese populations. MTHFRT allele frequency is higher than the observed in Caucasian and Japanese population, but also than the frequency reported previously by Nitsche et al. (2003) in other group of the Chilean population. We suggest that this difference could be explained by different genetic composition of the previously studied group, which could be more similar to Caucasians. Unfortunately, Nitsche et al. no reported the Amerindian-Caucasian admixture percentage to evaluate this point.
On the other hand, CYP2C19 * 2, GSTT1 null and GSTM1 null frequencies are lower than those reported in Caucasian or Japanese population. Finally for CYP2C8 * 3, CYP2C9 * 2 and * 3, CYP2D6 * 3, CYP2D6 * 4, CYP2E1 * 5B and * 6, as well as for UGT2B7 * 2 the frequencies described for Chileans are intermediate in relation to those described for Caucasian and Japanese population (Sullivan-Klose et al., 1996;Nasu et al., 1997;Chen et al., 1998Chen et al., , 1999Chida et al., 1999;Paris et al., 1999;Sachse et al., 1999;Bhasker et al., 2000;Matsuo et al., 2001;Murata et al., 2001;Roy et al., 2005;Skarke et al., 2005;Nakajima et al., 2006; Krishnakumar www.frontiersin.org  Kurose et al., 2012), which is expected because Chilean population is considered a mixed ethnicity between both races. Some limitations in this study should be pointed out. Only some polymorphism of the many (>80) CYP2D6 known were genotyped. We select only CYP2D6 polymorphisms that have shown better reported relationship with plasma levels of antineoplastic drugs and those that have higher allele frequencies, based on a previous pilot study in Chilean subjects (Dr. Monica Acuña, Personal Communication). Similarly, same criteria were used for choosing the other studied polymorphisms, based on both literature and our own previous research. Nevertheless, some potentially relevant CYP variants are currently under investigation in our laboratory (CYP2D6 * 2xN, CYP2D6 * 5, CYP2D6 * 10, CYP2C8 * 3, and CYP2C8 * 4), to complement the results shown in this paper.
Additionally, we have no data for other relevant polymorphic enzymes, such as for example, CYP1B1 and CYP2B6, responsible for the metabolism of several antineoplastic drugs (Quiñones et al., 2008) and UDP Glucuronyl transferase 1A1, involved in metabolism of irinotecan (Dias et al., 2012). We did not analyze thiopurine S-methyltransferase (TPMT), a cytosolic phase II enzyme involved in the metabolism of azathioprine, thiopurine, and thioguanine (Zhou, 2006). However the frequency of four allelic variants of this gene ( * 2, * 3A, * 3B, and * 3C) were analyzed previously in Chilean population by Alvarez et al. (2009) showing that the presence of * 3A allele is the most prevalent, which is similar to Caucasians, giving a first approach to the use of this polymorphism in clinical practice in Chilean patients.
Another limitation of the present research is the use of Japanese population as proxy of the ancestral Asians of Chilean people. We use this population as reference due to two main facts: (a) there is very good and complete information about these polymorphisms in Japanese people and (b) recently, have been reported no great differences among Japanese and other Asian populations, particularly with respect to Chinese population (Kurose et al., 2012).
On the other hand, drug-drug pharmacological interactions, some epigenetic and environmental factors, and alternative metabolic routes should not be excluded to describe response to antineoplastic agents, which is per se a multifactorial event. Thus, the research in this area must identify these factors and potential gene -environment interactions that modulate response to these drugs.
Some polymorphisms have been studied in other South American countries (Isaza et al., 2000;Gaspar et al., 2002;Fernández et al., 2004;Gattas et al., 2004;Vianna-Jorge et al., 2004;Almeida et al., 2005;Lizcano Fernández, 2005;Rossini et al., 2006;Schlawicke et al., 2007;Canalle et al., 2008;Rodríguez et al., 2008;Castaño-Molina et al., 2009;Dorado et al., 2011), but the comparison with Chilean Mestizo population is very difficult because of the divergent origin of these populations. Chilean population is different from other South American countries, from Brazilian people for example, which have African and Portuguese contribution with great pharmacogenomic diversity, or from Argentina and Uruguay populations, which are multiethnic countries, with Amerindian-European admixture, but mostly with Italians (Lizcano Fernández, 2005). In this respect, in South America, the region one of the most diverse genetic background in the world, four main components have contributed to the present-day population: Amerindians (pre-Columbian inhabitants); Iberians (conquerors) who dominated the continent until the nineteenth century, Africans (imported as slaves by the colonizers); and postindependence immigrants from overseas (mostly Italy and Germany but also from France, South Asia, and Japan). Therefore, we suggest that studies of these pharmacogenes in Chileans should be used to develop pharmacogenetic tools for this specific population, rather than extrapolating results obtained to other populations.    The frequencies observed in metabolic polymorphism studied in Chilean population were distinct from paternal races. These results contribute to better understanding of the basis of ethnic variation in drug metabolism and response (Agúndez, 2004;García-Martín et al., 2006b;Yasuda et al., 2008), and suggest a complex genetic profile of this "mestizo" population, which should be considered in pharmacotherapy, especially for drugs with a narrow safety range, particularly in cancer chemotherapy (García-Martín, 2008;Quiñones et al., 2008). These established genotype frequencies may be used for studying the phenotype variation in further studies. Thus, it may be a good contribution for further studies on the clinical application of pharmacogenomics in Asian-Caucasian mixed races.

CONCLUSION
Profound variation in polymorphisms of metabolizing enzymes have been described in diverse populations, including enzymes that take part in the metabolism of chemotherapeutics drugs (50). In this sense, our results agree with these observations when we compare the analyzed sub-group with Asian and Caucasian populations ( Table 4).
As Chilean population represents a mixed ethnicity mainly between native-Americans and Caucasians (mostly Spaniards) the data obtained might help to understand inter ethnic differences not only for single polymorphisms, but also the function of simultaneous polymorphisms in metabolic genes in each subject, to explain differences in response to chemotherapy.

Frontiers in Genetics | Pharmacogenetics and Pharmacogenomics
This investigation contributes to have a first pattern of several relevant polymorphisms in metabolizing enzymes (CYP1A1,CYP1A2,CYP2A6,CYP2C8,CYP2C9,CYP2C19,CYP2D6,CYP2E1,CYP3A4/5,GSTM1,GSTT1,UGT2B7,and MTHFR) in Chilean people, which can give course for a genetic -population investigation of these polymorphisms helping to the understanding of susceptibility to drugs and pathologies in this population, which already has been suggested by our group of research for some specific genes (Quiñones et al., 1999(Quiñones et al., , 2001(Quiñones et al., , 2008Acevedo et al., 2003;Lee et al., 2006;Cordero et al., 2010).
The knowledge of genetic variants involved in the metabolism of the antineoplastic drugs in the Chilean population will help to the prediction of their clinical efficacy and/or toxicity, and therefore, will help us to the design personalized cancer treatments to improve therapeutic response and diminish the adverse effects improving cost-efficacy of treatments.
Finally, based on scientific literature and our experience, we believe that, at least in Chilean population, MTHFR/methotrexate, GST/cisplatin, and CYP2D6/tamoxifen are the potential more relevant gene/drug pairs which are closer for monitoring use in clinical practice.