Abstract
Aim: To investigate polymorphisms that are probable indicators of response variability during cancer treatment with 5-fluorouracil (rs16430, rs2279198, rs1801159 and rs17878362). Materials and Methods: We investigated 1,038 individuals regarding allele distribution from different populations, out of which we genotyped 127 individuals from a Brazilian admixed population. Similarity analyses with parental populations were performed. Prevalence of potentially deleterious alleles was also evaluated. Results: Thirty-seven percent of the population had at least three potentially deleterious alleles and 38.6% had at least one potentially deleterious allele in homozygosis. Conclusion: Potentially deleterious alleles are present under diverse frequencies in different populations. Therefore, genotyping prior to 5-fluorouracil administration should be recommended.
Cancer is a complex and multi-factorial disease that is mainly based on disorganized cellular multiplication due to errors in genetic information, in previously healthy tissues (1). 5-Fluorouracil (5-FU) has been largely used and has been proven highly efficient in the treatment of many types of cancer, including gastric, colorectal, pulmonary and pancreatic cancer (2-8). However, this drug may be very toxic or debilitating. The main types of toxicities associated with 5-FU use are gastrointestinal and myelosuppression (4, 9-11), that may lead to treatment drop-out or even death of patients. Previous studies have proven that responses vary considerably among distinct parental populations (12). 5-FU is also the main compound of the S-1 chemotherapy scheme, that is very efficient among Asian patients (7), but not among European populations (5, 6).
Many genes participate in the metabolism of the 5-FU. Thus, several polymorphisms are associated with treatment response, as well as adverse reactions induced by the administration of the drug. Herein, we highlight four genes: DPYD, TYMS, OPRT and TP53.
The DPYD gene encodes the dihydropyrimidine dehydrogenase enzyme, acting as the first rate limiting reaction of the drug and is responsible for the almost immediate inactivation of near 90% of the administered 5-FU (13, 14). The DPYD gene is FDA (Food and Drug Administration)-approved as a pharmacogenetic marker for metabolizing 5-FU. Many studies have shown that, with partial or total deficiency in this enzyme, the levels of toxicity and death by 5-FU usage are elevated (9, 10, 14, 15). In fact, it has been suggested that patients with less than 70% DPD enzyme activity are considered at-risk of developing severe 5-FU-induced toxicity (16). Among the important polymorphisms for determining susceptibility to 5-FU toxicity, is rs1801159 (DPYD*5), a T>C SNP in exon 13 of DPYD gene. The mutant allele (C) is related to DPD deficiency, that leads patients with CC or TC genotypes to be more susceptible to present severe toxicity (17-19).
The TYMS gene encodes the thymidylate synthase enzyme, that catalyzes the methylation of deoxyuridine monophosphate (dUMP) to deoxythymidine monophosphate (dTMP), forming a complex that inhibits DNA synthesis and, subsequently, cellular proliferation (20, 21). A 6-bp deletion located 447 bp downstream of the TYMS transcription stop codon (rs16430) alters TYMS mRNA stability (22) and has been related to reduction of gene expression, drug resistance and poor outcome during therapy with 5-FU (23-25). However, genotypes with the 6-bp insertion (INS/INS or INS/DEL) are related to higher risk for development of many cancer types, including gastric and colorectal cancer (26, 27).
The TP53 gene leads to p53, which preserves genomic stability, responds to a variety of threats to the cell (28, 29). Among other roles, it orchestrates the cell death process, after DNA damage during the 5-FU metabolism. Mutations in TP53 may prevent it to properly perform apoptosis or DNA repair. Therefore, its tumor suppressor function will be deficient, generating a resistance to 5-FU treatment (30). Several TP53 polymorphisms, such as a 16 bp insertion/deletion (rs17878362), are associated with cancer susceptibility (31-34) and the predictive effect in oncological treatment with 5-FU (30, 35).
The OPRT gene (also known as UMPS) has a key role in the metabolic pathway of 5-FU. Therefore, it also should be considered an important susceptibility marker and an indicator of significant prognosis in patients (20, 36-39). In this metabolic pathway, OPRT is responsible for degrading 5-FU into FUMP. OPRT overexpression leads to reduction of drug activity and increase of cytotoxicity (40-42).
The rs2279198 polymorphism (A>G), despite the fact that it has not yet been specifically associated to the treatment with 5-FU, may influence the expression of the OPRT gene. Located four nucleotides before the gene transcription start site (3:124449209), inside the gene promoter and the E2F transcription factor 1 (E2F1) binding site, this polymorphism may lead to changes in the transcription complex stability and to binding difficulties of the E2F1 and RNA-Polymerase II (PolII). In comparison to the G allele, the A allele shows greater affinity to the E2F1's MOTIF (43), leading to a greater gene expression. As patients benefit from the lower activity of OPRT gene during the treatment with 5-FU (40-42), the A allele was considered as deleterious.
In Brazil, as in many countries, 5-FU is widely used in the treatment of most frequent cancers in the population, such as breast, gastric and colorectal cancer. Despite the notorious importance of the above-described genetic markers to treatment success, no study has analyzed the frequencies of these markers in the Brazilian population. Therefore, the aim of the present study was to evaluate the distribution of four markers among an admixed Brazilian population in order to estimate the potential risk of toxicity development and adverse reaction in comparison to other continental populations (obtained from 1,000 Genomes database) and to improve patient outcomes.
Materials and Methods
Population. The studied population comprised of 127 cancer-free individuals from Belém, a city located in the Brazilian Amazon region. The participants signed an informed consent giving permissions to the usage of their samples, with the approval of the João de Barros Barreto Ethics Committee in Research (CAAE: 06649713.9.0000.0017). DNA was extracted from samples of peripheral blood according to the method developed by Sambrook et al. (1989) (44). Data pertaining to genetic ancestry were obtained from analysis with a panel of 48 autosomal Ancestry Informative Markers (AIM) using protocols that had been previously established (45).
Genotyping procedures. The markers were genotyped (allelic discrimination) using the following methodologies: (i) Real-Time PCR (for rs1801159); (ii) Fragment Analysis (for rs16430 and rs17878362); (iii) Direct Sequencing (for rs2279198). Real-Time PCR for allelic discrimination was conducted according to TaqMan protocol with C__1823316_20 (Life Technologies, Foster City, CA, USA). Multiplex PCR, genotyping both of the markers at the same reaction, was performed using an ABI Veriti Thermal Cycler (Life Technologies). This system used 5.0 μL from kit QIAGEN Multiplex PCR (QIAGEN, GmbH, Germany), 1.0 μl of Q-solution, 1.0 μL of Primer Mix, 2.0 μL of water and 1.0 μL of DNA (20 ng). Samples were incubated at 95°C for 15 min, then 35 cycles at 94°C for 45 sec, 60°C for 90 sec and 72°C for 60 sec, with a final extension of 70°C for 30 min. For capillary electrophoresis (Fragment Analysis), 1.0 μl from the PCR product was added to 8.5 μL of HI-FI deionized formamide and 0.5 μL of GeneScan 500 LIZ pattern size (Life Technologies). DNA fragments were separated by using the ABI PRISM 3130 Genetic Analyzer (Life Technologies) and analyzed with the program GeneMapper ID v.3.2 (Life Technologies). Forward and reverse primers for the TP53 gene marker were 5’GGGACTGACTTTCTGCTCTTGT3’ and 5’GGGACTGTAGATGGGTGAAAAG3’; for the TYMS gene marker, they were 5’ATCCAAACCAGAATACAGCACA3’ and 5’CTCAAATCTGAGGGAGCTGAGT3’.
Direct sequencing was performed after a regular PCR reaction. This amplification was achieved in an ABI Veriti Thermal Cycler with 18.6 μL of water, 1.8 μL of buffer, 1.0 μL of MgCl2, 1.3 μL of dNTP, 0.5 μL (5 mMol) for the forward primer (5’CATCCAGAAGAAACCAGGAGAA3’) and 0.5 μL (5 mMol) for the reverse primer (5’CGTACAGACCCGTCACCAAT3’), 0.3 μL of Taq polymerase enzyme and 1.0 μL of the DNA sample (20 ng). The used program in the thermal cycler was: 95°C for 5 min; 95°C for 80 sec, 60°C for 45 sec, 72°C for 90 sec, all three steps for 35 cycles; 72°C for 10 min; and 4°C for 5 min. Direct sequencing was developed in the ABI PRISM 3130 Genetic Analyzer with a total volume reaction of 15 μl, being 10.2 μl of water, 3.0 μl of Save Money buffer; 0.4 μl of primer; 0.4 μl of Big Dye™ kit (Terminator Cycle Sequence v. 3.0) and 1.0 μL of the PCR product. The used program was the same of the PCR step.
In silico analysis. To further improve the results (our findings on genotype distribution), our data were compared against the 1000 Genomes database (http://www.1000genomes.org/) (46) genotypes regarding the selected markers and three ancestral populations: European (populations from Finland, England, Scotland, Spain, Italy and population from Utah - USA, with Northern and Western Europe ancestry; N=379), African (populations from Kenya, Nigeria and African ancestry in the Southwest USA; N=246) and Asian (populations from Beijing, Southern China and Tokyo, Japan; n=286).
Frequencies of the alleles of the studied polymorphisms in different populations (European, African, Asian and Belém).
Statistical analysis. Statistical analyses were performed using two programs: SPSS v.14.0 (SPSS Ins. Chicago, IL, USA) and Arlequin v.3.5 (47). The p-value was considered significant if p≤0.05.
Results
General information and ancestry data. The studied sample of 127 subjects (71 females and 56 males) is representing the population of Belém city. In the present investigation, the used AIM panel showed a large genetic contribution from European populations (56%), followed by an Amerindian contribution (29%) and an African contribution (15%) in this sample cohort.
Allele frequencies comparison. Table I shows the allelic frequencies of the studied polymorphisms in TP53, TYMS, OPRT and DPYD genes. Frequencies for each marker were compared with the Fisher's Exact Test and the results are presented in Table II.
In the comparison between the continental populations, all markers showed significant difference for allelic frequency (p≤0.05), except for African and European populations regarding rs1801159 (DPYD gene). When our sample was compared to continental populations, the results varied. In the comparison with European populations, only rs1801159 (p=0.0089) showed a significant difference. The comparison with African populations presented contrary results: all markers indicated significant differences between the two populations, except rs1801159 (p=0.1581). In the comparison with Asian populations, two markers suggested no significant difference (rs2279198 and rs1801159), while the other two (rs17878362 and rs16430) showed that these populations are different from each other (p<0.00001, on both cases).
To improve data visualization, 10.000 Monte Carlo simulations were performed, by sampling randomly 100 individuals from each population and the frequency distribution of each sample is presented in Figure 1. This boxplot confirms our findings over the ethnic contribution in the population, by representing the proximity of our sample to the European for most markers.
p-Value comparison between parental and admixed populations (by Fisher's Exact Test), for each investigated polymorphism.
Counting of deleterious alleles in samples.
Analysis of deleterious alleles. Analysis of possibly deleterious alleles for the 5-FU treatment outcome in these polymorphisms for this population was also performed. On Table III, it is possible to see that 3% of the investigated sample present an elevated number of deleterious alleles (≥5) and that a representative part of the genotyped sample (25 individuals, 37.1% of the total) presents at least three alleles considered deleterious.
Allelic frequency distribution of the studied polymorphisms for each population.
We also analyzed data considering the number of homozygous individuals for deleterious alleles. By analyzing all four markers, it is possible to identify that two individuals of the sample (1.6%) are homozygous for two or three investigated markers and that 47 individuals have the homozygous genotype for one of the markers. Cumulative data indicate that 38.6% of the samples (47 individuals) has at least one deleterious allele in homozygosis. Another analysis of homozygous presence was performed, but only for the markers of genes directly related to the metabolism of 5-FU drug (TYMS, DPYD and OPRT). Data showed that 42 individuals (34.6%) are homozygous for one of the markers and two individuals (1.6%) are homozygous for two of these three markers. Therefore, the analyzed data allow estimating that at least 36.2% of the population (44 individuals) has a homozygous genotype for at least one of the alleles related to the poor outcome of 5-FU drug.
Discussion
The studied population is characterized by an admixture of European, African and Amerindian ethnicities. Ancestry data obtained in this investigation by AIM corroborates with the information that the Brazilian population is very heterogeneous (45).
Results from previous studies that did not investigate admixed populations do not correspond to drug response (and to cancer susceptibility, in the case of TP53 gene) in heterogeneous populations, such as Brazilian populations, and it would not be wise to extrapolate these results for such populations. This is reinforced by our results regarding the comparison of the allelic frequencies of the studied polymorphisms in the admixed population and in the parental populations (European, African and Asian).
Many studies have been performed in the Brazilian Amazon region regarding cancer susceptibility of different types of cancers, including gastric cancer (48-50). It is important to highlight that it presents one of the most elevated incidences of gastric cancer in the country. In that region, it is the second most frequent type of cancer in men (11.10/100.000 individuals) and third most frequent among women (5.91/100.000 individuals) (51). This type of cancer is usually treated with 5-FU in Brazilian hospitals. We focused our analyses for the studied polymorphisms in the presence of deleterious alleles that may be related to a higher susceptibility to gastric cancer (allele INS of rs17878362) as well as the toxicity development related to 5-FU administration (allele A of rs2279198, allele C of rs1801159 and allele DEL of rs16430).
Therefore, it is recommended to genotype each patient, prior to clinical application, in order to identify patients likely to benefit from 5-FU chemotherapy. Personal genotyping is especially recommended due to the admixture of this population, allowing a higher success rate of this chemotherapy. More in-depth studies should be conducted in other Brazilian populations to evaluate if other admixed populations show similar results.
It is noteworthy that current literature lacks studies on polymorphisms in genes responsible for metabolizing 5-FU in Brazilian populations. Thus, the present investigation is the first of its kind in this country.
Acknowledgements
This work is part of the Rede de Pesquisa em Genômica Populacional Humana (Biocomputacional – Protocol no. 3381/2013/CAPES). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. GIOVANNA CAVALCANTE is supported by PhD scholarship from CNPq/Brazil; ÂNDREA RIBEIRO-DOS-SANTOS is supported by CNPq/Produtividade; SIDNEY SANTOS is supported by CNPq/Produtividade; NEY SANTOS is supported by FAPESPA. We thank Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Fundação Amazônia de Amparo a Estudos e Pesquisas do Pará (FAPESPA) and Pró-Reitoria de Pesquisa e Pós Graduação/Universidade Federal do Pará-Fundação de Amparo e Desenvolvimento da Pesquisa (PROPESP/UFPA-FADESP) for the received grants. We also thank Pablo Diego do Carmo Pinto for great statistical and technical assistance.
- Received September 14, 2015.
- Revision received October 2, 2015.
- Accepted October 14, 2015.
- Copyright© 2015 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved