Investigation of INDEL variants in apoptosis: the relevance to gastric cancer

Background Apoptosis is a type of cell death involved in different pathways inherent to the cell and the evasion from this mechanism has been related to cancer, although this process remains not very well comprehended. Gastric cancer (GC) is one of the most incident and aggressive types of cancer worldwide. In this study, we analyzed the distribution of INDEL variants in GC patients (Case) and individuals from the general population (Control) from the Amazon region, in which GC is remarkably frequent. Methods A panel of nine INDEL markers in apoptosis-related genes (BCL2 rs11269260, CASP3 rs4647655, CASP8 rs3834129 and rs59308963, CASP9 rs4645982 and rs61079693, FADD rs4197, FAS rs10562972 and TP53 rs17880560) was developed and genotyped by multiplex PCR in both groups. Results In our analyses, only marker rs4197 (FADD gene) was associated to GC development as follows: INS/DEL genotype of rs4197 increasing in about 2-fold the chances of developing this type of cancer (P = 0.046; OR = 1.940; 95%CI = 1.011–3.725). Conclusion Our results suggest that rs4197 (FADD gene) might play a role in gastric carcinogenesis in the investigated population. More studies are needed to clarify this relation. Here, we highlight the importance of investigating INDEL variants in genes involved in apoptosis.


Background
Apoptosis is a type of programmed cell death (PCD) that may occur as a response to intra-or extracellular stimuli, leading to intrinsic pathway or extrinsic pathway, respectively. These are independent pathways that converge to the executor phase, the final stage of this death process [1][2][3]. Several genes are involved in apoptosis, out of which we highlight: BCL2, CASP9 in the intrinsic pathway; FAS, FADD, CASP8 in the extrinsic pathway; TP53 in both pathways; and CASP3 in the executor phase. All of these genes are important to the apoptotic mechanism, so that variants affecting their functions might dysregulate cell homeostasis and contribute to tumor development [4]. Although the apoptosis process is still not completely comprehended, this type of cell death mechanism has been related to cancer.
Gastric cancer (GC) is one of the most frequent and aggressive types of cancer worldwide, usually presenting an unfavorable prognosis [5,6]. In Brazil, GC is the sixth most incident and the fifth most mortal type of cancer [6]. Brazil is a country with a highly admixed population, composed mainly of European, African, Amerindian and Asian ancestry. Many studies in this country have highlighted differences between these ancestries and the Brazilian regions [7][8][9][10][11], including aspects of cancer development [12]. This is also notable by the higher incidence of GC in northern Brazil in comparison to the general rates of this country and the world [13].
Therefore, in this study, we aimed to analyze the distribution of nine Insertion/Deletion (INDEL) polymorphisms in apoptosis-related genes in gastric cancer patients and cancer-free individuals from the general population of the Amazon region.

Sampling
Case group was composed of 93 individuals with gastric cancer diagnosis from Pará state and control group was composed of 98 cancer-free individuals of the general population from the same state, located in the Amazon region of Brazil. From each individual, blood or buccal swab samples were collected. All participants signed an informed consent, with approval by the Committee for Research Ethics of Hospital João de Barros Barreto under Protocol Number: 64399617.7.0000.5634.

DNA extraction and quantification
DNA extraction was performed with phenol-chloroform method, based on [14]. Quantification of DNA was performed with NanoDrop 1000 and Qubit™ (Thermo Fisher Scientific, Wilmington, DE, USA).

Selection of INDEL markers
The markers were selected according to the main criteria: (i) they must be in genes involved in apoptosis pathways; (ii) they must be INDEL-type polymorphisms, hence presenting the potential to modify the function of the protein; (iii) they must have minor allele frequency (MAF) ≥ 10%. Table 1 shows the technical features of the investigated markers.

Genotyping
The panel of markers described above was genotyped by Multiplex PCR, allowing the amplification of all of these markers in a single reaction, followed by capillary electrophoresis and fragment analysis. Protocol preparation of PCR samples was: 5.0 μL of QIAGEN Multiplex PCR Master Mix, 1.0 μL of Q-solution, 2.0 μL of water, 1.0 μL of primer mix and 1.0 μL of DNA per sample. Amplification reaction was done using Veriti thermal cycler (Thermo Fisher Scientific) with the following protocol: 95°C for 15 min, followed by 35 cycles of 94°C for 45 s, 60°C for 90 s and 72°C for 1 min, with a final extension at 70°C for 30 min. The protocol for capillary electrophoresis was: 1.0 μL of PCR product, 8.5 μL of HI-DI deionized formamide and 0.5 μL of GeneScan 500 LIZ pattern size standard (reagents by Thermo Fisher Scientific). DNA fragments were separated and analyzed using ABI PRISM 3130 genetic analyzer and GeneMapper ID v.3.2 software (both by Thermo Fisher Scientific).
In addition to the panel of apoptosis markers, a panel of 61 ancestry informative markers (AIM) previously developed and expanded by our research group was also employed in this study, following the established protocols [15,16]. This was done considering that the Brazilian population is highly admixed in terms of genetic contributions from different parental populations, mainly European, African and Native American. In order Table 1 Technical characterization of the markers included in the panel to control the possible influence of these ancestries in the genotypic distribution, leading to misinterpretation, we have used this AIM panel, composed of ancestryspecific markers that are able to estimate individual and population genetic contribution.

Statistical analysis
Inference of genetic ancestry based on the AIM panel was done with Structure software v.2.3.4 [17]. R language [18] was used to assess Hardy-Weinberg Equilibrium (HWE) of genotype distribution. JASP software v. 0.9.2.0 [19] was used to perform all other statistical analyses (Student's t test, chi-squared test, logistic regression). P-values ≤0.05 were considered statistically significant.

Results
Firstly, we assessed HWE for genotypic distribution and found that only two markers, rs4645982 (CASP9) and rs4197 (FADD), deviated from HWE in case group (P = 0.0035 and P = 0.0026, respectively). Then, we compared sex, age and controlled genetic ancestry of case and control groups in order to verify possible confounding factors. In addition, staging (I-IV) of patients was assessed. These data are described in Table 2.
In these analyses, only age presented a statistically significant difference and it was included for correction in the panel analysis. It is noteworthy that 53% of the patients included in case group presented diffuse-type adenocarcinoma, while 47% presented intestinal-type adenocarcinoma, and the observed mean age (mean age ± SE) for these subtypes was 39.20 ± 1.397 and 44.34 ± 1.526, respectively, so that the mean age for case group as a whole was lower than it is usually observed for patients with GC in the world.
Nevertheless, in the analysis performed for each genotype (carriers of a genotype vs. non-carriers of such genotype as a reference group), only INS/DEL genotype of rs4197 (P = 0.046; OR = 1.940; 95%CI = 1.011-3.725) presented significant association to risk of developing GC in the investigated population (Table 3).

Discussion
In the last decade, it has been established that tumors display different abilities in order to grow, survive and proliferate, of which we highlight here resistance to cell death, including apoptosis [20]. However, there is still a lot to discover about these mechanisms.
Apoptosis is a complex process that involves the protein action of several different genes, such as BCL2, CASP3, CASP8, CASP9, FADD, FAS and TP53. Dysfunction in these genes may lead to deregulation of cell death and, thus, tumor development. Therefore, in this study, we investigated whether nine INDEL variants in the mentioned genes might influence gastric carcinogenesis, by comparing their distribution in GC patients and cancer-free individuals.
Regarding the observed deviation from HWE in the distribution of rs4197 and rs4645982 in case group, no previous studies were found with these markers, but their genotype distribution varies greatly between different populations in 1000 Genomes Project database [21]. Curiously, their distribution also presented a deviation from HWE in most populations in that database, suggesting that there could be a selective advantage leading to this pattern and/or that the deviation observed here could be due to population substructure, especially considering the relatively recent admixture process in Brazil, as previously observed for other markers in this population [7,22]. As such, it is an expected process in admixed populations, and it could even highlight the potential of these markers. In fact, it has been suggested that HWE deviation only in patient's group could support possible locus-disease associations [23].
Then, we carried on to the analyses performed for each genotype (carriers of a genotype vs. non-carriers of such genotype as a reference group) in cases and controls. In these analyses, our results suggest that individuals carrying INS/DEL of rs4197 have about 2-fold chances of developing GC than non-carriers (P = 0.046; OR = 1.940; 95%CI = 1.011-3.725).
To the best of our knowledge, this is the first study investigating variant rs4197. It is a 3-bp INDEL in the 3′ UTR of FADD gene, a key adaptor molecule that transmits death signals from death receptors during extrinsic apoptosis, thus being crucial to a variety of processes [24]. In fact, it seems that it could affect protein features due to the importance of 3′-UTR [25]. Depletion of FADD protein action may lead to failure of apoptosis and, thus, to tumor development. Absence of FADD expression has been suggested as marker of tumor development in mice and cancer prognosis in humans, due to the involvement of this protein with cell apoptosis, survival and proliferation [26]. Therefore, investigation of FADD gene should help the understanding of cancer development, including gastric carcinogenesis.
As for the variants that did not present a statistical association here, this may suggest that they either do not play a role in gastric carcinogenesis or it is a populationspecific treat. Indeed, some of these variants have been associated cancer development in a few studies, but not in others. For instance, one of these variants (rs3834129 Table 3 Genotypic distribution of the investigated variants for patients with GC in comparison to control group. P-value, OR and 95%CI were obtained with logistic regression adjusted for age, which was done for the genotype as reference in each line ( [27], but have not found any association of this variant with GC or colorectal cancer [22], so that our study corroborates the findings in the latter. Thus, the association of such variants to GC and other types of cancer is still a matter of great discussion.

Conclusions
Despite a few limitations (sample number and average age of both groups), this study contributed to an increased knowledge on variants in apoptosis-related genes in regard to GC development. As future perspectives, we recommend studies in the same population with a greater sample number and in different populations for comparison, as well as functional studies focused on the proteins to possibly reinforce the involvement of FADD gene in gastric carcinogenesis. Although such future studies are recommended to strengthen our results, this work contributes to a better understanding of these genes and INDEL variants in regard to gastric cancer.