Mutational Analysis of Oncogenic AKT1 Gene Associated with Breast Cancer Risk in the High Altitude Ecuadorian Mestizo Population

Breast cancer is the leading cause of cancer-related death among women worldwide. AKT1 encodes the kinase B alpha protein. The rs121434592, rs12881616, rs11555432, rs11555431, rs2494732, and rs3803304 single nucleotide polymorphisms have been identified in the AKT1 kinase gene. Activated AKT1 phosphorylates downstream substrates regulating cell growth, metabolism, apoptosis, angiogenesis, and drug responses. It is essential to know how breast cancer risk is associated with histopathological and immunohistochemical characteristics and genotype polymorphisms in a high altitude Ecuadorian mestizo population. This is a retrospective case-control study. DNA was extracted from 185 healthy and 91 affected women who live 2,800 meters above sea level. Genotypes were determined by genomic sequencing. We found a possible association between the noncoding intronic variant rs3803304 and breast cancer risk development: GG (odds ratio [OR] = 5.2; 95% confidence interval [CI] = 1.3-20.9; P ≤ 0.05; Q > 0.05). Regarding pathologic characteristics, we found significant risk between estrogen receptor, progesterone receptor, and HER2 status and molecular subtypes (P ≤ 0.001; Q ≤ 0.05). On the other hand, we did not find risk between variants and histopathological characteristics. Despite the small sample size, we found that the intronic variant, AKT1 rs3803304, may act as a predictive biomarker in the risk of developing breast cancer in the high altitude Ecuadorian mestizo population.


Introduction
Breast cancer (BC) in women involves the progressive accumulation of genetic, hereditary, hormonal, and environmental factors representing a significant health problem worldwide [1]. BC is the leading cause of cancer-related death among women (521,541 cases) and the most commonly diagnosed cancer (1,679,076 cases) [2]. The areas with a higher incidence of BC per each 100,000 inhabitants are Western Europe (89.9), Oceania (85.5), and Northern Europe (76.7), while South America has a lower incidence (44.3) [3]. In Ecuador, the incidence rate of BC has reached up to 32.7 in 2012 [4].
The E17K rs121434592, E319G rs12881616, L357P rs11555432, P388T rs11555431, rs2494732, and rs3803304 single nucleotide polymorphisms (SNPs) have been identified in the AKT1 kinase gene [22]. The E17K variant generates a conformational change in the PH domain of the AKT1 protein, making it possible to join the protein and the cell membrane, causing the process of intracellular phosphorylation [23]. The E319G, L357P, and P388T variants are located in the kinase domain (KD), and the rs2494732 and rs3803304 variants are placed in intronic regions ( Figure 1). Consequently, dysregulation in cell proliferation, survival, and growth drives progressive transformation of normal cells towards a malignant phenotype [18,[24][25][26][27]. Aberrant AKT signaling is the underlying defect found in several pathologies such as esophageal cancer [28], head and neck cancer [29], and non-small cell lung cancer [30].
The objective was to determine the risk of breast cancer associated with histopathological and immunohistochemical characteristics and genetic polymorphisms in a high altitude Ecuadorian mestizo population.

DNA Extraction and
Purification. DNA extraction and purification of control and case individuals were performed using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI) and the PureLink Genomic DNA Kit (Invitrogen, Carlsbad, CA), respectively. The DNA of the healthy individuals was extracted from peripheral blood samples and presented an average concentration of 135 ng/ l. Meanwhile, the DNA of the affected individuals, which presented an average concentration of 84 ng/ l, was extracted from ten sections (5 m) of formalin-fixed paraffin-embedded breast tumor tissue previously cut with a microtome CUT 6062 (SLEE, Mainz, Germany). Both calculations were obtained using NanoDrop 2000 (Thermo Scientific, Waltham, MA).  Table 1.

Amplification and
The SNPs E17K rs121434592 (G>A) (198 bp), L357P rs11555432 (T>C) (142 bp), E319G rs12881616 (A>G) (186 bp), and P388T rs11555431 (C>A) (171 bp) presented in exonic regions and rs3803304 (C>G) (171 bp) and rs2494732 (C>T) (171 bp) presented in intronic regions were amplified through the polymerase chain reaction (PCR) technique. Supplementary Table 2 details the genetic variants analyzed in this study. The PCR program started with an initial denaturation stage lasting 5 minutes at 94 ∘ C, followed by 35   ( Table 1), 45 seconds at 70 ∘ C, and a final elongation for 3 minutes at 72 ∘ C. Each run was completed using a Sure Cycler 8800 thermocycler (Agilent, Santa Clara, CA). The amplified fragment was then separated by electrophoresis in 2% agarose gels stained with ethidium bromide and was observed in an ImageQuant 300 transilluminator (General Electric, Fairfield, SC). Genotyping was performed using DNA sequencing analysis through Genetic Analyzer 3130 (Applied Biosystems, Austin, TX). The final volume of the reaction was 12 l and contained 2.8 l of Milli-Q water, 2 l of 5X buffer, 1 l of primer FW (3.2 pmol), 1 l of BigDye Terminator v3.1 sequencing standard (Applied Biosystems, Austin, TX), and 3 l of PCR product (3 to 10 ng). Once the product was amplified, it was then purified using Agencourt CleanSEQ (Beckman Coulter, Miami, FL). The amplification program consisted of 3 minutes at 96 ∘ C, followed by 30 cycles of 10 seconds at 96 ∘ C, 5 seconds at 50 ∘ C, and 4 minutes at 60 ∘ C. Finally, sequence analysis was performed using Sequencing Analysis Software 5.3.1 (Applied Biosystems, Austin, TX), and the alignment with sequences from GenBank (AKT1 NC 000014.9) was performed using Seq-Scape Software v2.6 (Applied Biosystems, Austin, TX) [7].

Statistical Analysis.
The information from the clinical records of the patients was collected and stored in a database. Allelic and genotypic frequencies of the AKT1 SNPs were calculated; also Hardy-Weinberg equilibrium was determined by using a tool available on the Internet (http://www.oege.org/software/hwe-mr-calc.shtml) [31]. With the use of IBM SPSS Statistics 22 software (SPSS Inc., Chicago, IL), chi-square ( 2 ), and odds ratio (OR) (with a  95% confidence interval [CI] and 2 x 2 contingency table), tests were applied to determine the association between the risk of developing BC, the SNPs, histopathological characteristics, and immunohistochemical characteristics. P < 0.05 was considered statistically significant. Consequently, the false discovery rate (FDR) Benjamini/Hochberg correction was performed in order to obtain Q values.

Results
The distribution of baseline characteristics in all patients with BC is shown in Table 2  Frequency of mutations in the AKT1 gene in breast cancer subtypes is detailed in Table 2 and genotype distribution and allele frequencies of AKT1 mutations in cases and controls are detailed in Table 3. The E17K AA homozygous genotype had a frequency of 0.02 in cases and 0.00 in control and was present in 2/91 affected individuals (2.2%) and 6.25% of basal-like type (P > 0.05; Q > 0.05). The E319G GG, L357P CC, and P388T AA genotypes had a frequency of 0.00 in cases and controls and were present in 0/91 affected individuals (0%). The rs2494732 TT mutant genotype had a frequency of 0.14 in cases and 0.22 in controls and was present in 13/91 affected individuals (14.29%) and 36.4% of HER individuals (P > 0.05; Q > 0.05). The rs3803304 mutant homozygous genotype had a frequency of 0.08 in cases and 0.02 in controls and was present in 7/91 affected individuals (7.7%) and 18.2% of HER individuals (P > 0.05; Q > 0.05), whereas the rs3803304 CG heterozygous genotype was significantly different among the four subtypes (P ≤ 0.05; Q ≤ 0.05), where luminal B had the highest percentage (63.6%). Finally, the alleles of the rs2494732 and rs3803304 SNPs presented Hardy-Weinberg equilibrium.
The association between the rs3803304 and rs2494732 polymorphisms and the histopathological and immunohistochemical characteristics is detailed in Table 5. Regarding the affected breast, tumor stage, LN status, ER status, PR status, HER2 status, and surgical margins, no statistically significant differences (P > 0.05; Q > 0.05) were found in relation to the genetic polymorphisms. Additionally, data of control individuals is fully detailed in Supplementary  Table 3.

Discussion
During the last decade, breast cancer genome-wide association studies (GWAS) have identified ∼80 loci with small-tomoderate effects on OR ranging from 1.05 to 1.53 [10,[32][33][34]. Studying SNPs is crucial to fully understand breast cancer biology, as well as for the development of novel therapeutics for cancer treatment and for providing methods for prevention and early diagnosis [35]. The most studied populations have been Asian, European, and African. However, Latin American populations have been poorly studied, making genetic characterization essential to better understand the development of BC [7,36]. Mounting evidence exists that activation of AKT proteins is important in cancer development [14]. Hyperactivation of the AKT pathway has been detected in up to 50% of all human tumors and is closely associated with chemoresistance [37]. Therefore, AKT has been an attractive target for anticancer drug discovery [38]. In particular, genetic alterations of the AKT genes have been demonstrated in many human tumors, including breast, colorectal, and ovarian cancers [39]. Concerning BC, the activating mutations of AKT1 gene have not been widely reported, the first being a study conducted in a female Latin American mestizo population from a high altitude (2,800 masl). It is noteworthy that this retrospective research presented a limited number of cases. However, it gives us relevant information about BC risk and its association with genotype polymorphisms of the AKT1 gene.
The E17K (rs121434592) point mutation in the pleckstrin homology domain of the AKT1 gene is the major point mutation that has been reported in the literature. In 2007, Carpten et al. evaluated the complete coding regions of AKT family members for mutations in genomic DNA from clinical tumour specimens representing breast ( = 61), colorectal ( = 51), and ovarian ( = 50) cancers [39]. Analysis of these samples revealed a unique mutation in the PH domain of AKT1 which results in a lysine substitution for glutamic acid at amino acid 17. E17K was identified in 5 of 51 (8%) breast, 3 of 51 (6%) colorectal, and 1 of 50 (2%) ovarian cancers. Although the sample size was insufficient to document statistical significance, the lack of coincidence of these mutations indicates that the AKT1 mutation (E17K) is sufficient for pathological activation of the PI3K/AKT They revealed that mutations such as E17S, E319G, L357P, and P388T were probably damaging and these mutations should be considered alongside E17K in the therapeutic development of AKT inhibitors to treat human cancer [22]. In the high altitude Ecuadorian mestizo population, the E17K mutation was found in 2 of 91 individuals (2.2%), where one of them was luminal A and the other one was basal-like (Table 2). In spite of the low percentage found in our population, it is known that the E17K (rs121434592) substitution decreases the sensitivity to an allosteric kinase inhibitor, so this mutation may have important clinical utility for AKT drug development [43].
According to the analysis of prioritization of Shanti et al., the E319G (rs12881616), L357P (rs11555432), and P388T (rs11555431) exonic variants are high-risk missense mutations of the AKT1 gene [22]. Nevertheless, these variants are 100% present in their normal homozygous state in this study due to a small sample size of the high altitude Ecuadorian mestizo population. On the other hand, the analysis of the rs2494732 and rs3803304 intronic variants was carried out. As for the rs2494732 variant, the CT genotype is found in 40.66% of the affected population with a higher percentage in the basal-like subtype (50%) and luminal A subtype (45.3%), whereas the TT genotype is found in 14.29% of the affected population with a higher percentage in the luminal A subtype (9.4%) and HER2-enriched subtype (36.4%). The OR statistical test determined that the rs2494732 variant did not present risk with the development of breast cancer in the CT heterozygous genotypes (OR = 0.6, 95% CI = 0.3-1.0; P > 0.05; Q > 0.05) or TT mutant homozygous genotypes (OR = 0.4, 95% CI = 0.2-0.9; P ≤ 0.05; Q > 0.05), just as found in the Chinese Han population with basocellular skin cancer (OR = 1.00, 95% CI = 0.5-2.2; P > 0.05) [41]. As a matter of fact, in this population, the T mutant allele presented a higher allele frequency (0.46) in controls than in cases (0.35), and the C normal allele had a higher allele frequency in cases (0.65) than in controls (0.54). These results prove that the presence of mutations in the rs2494732 variant is related to a protective factor in the development of breast cancer and is associated with better general survival with a hazard ratio (HR) of 0.59 (95% CI = 0.40-0.86; P ≤ 0.05) and progression-free survival with a HR of 0.74 (95% CI = 0.53-1.03; P > 0.05) in a Korean population with non-small cell lung cancer [30]. Concerning the rs3803304 variant, the CG genotype is found in 33.0% of the affected population with a higher percentage in the luminal B subtype (63.6%) and HER subtype (45.5%), whereas the GG genotype is found in 7.7% of the affected population with higher percentage in the HER2-enriched subtype (18.2%) and luminal A subtype (7.5%). The OR statistical test determined that the rs3803304 variant presented a possible risk with the development of breast cancer in individuals with the GG mutant homozygous genotype (OR = 5.2, 95% CI = 1.3-20.9; P ≤ 0.05; Q > 0.05). In this mestizo population, the C normal allele presented a higher allele frequency in controls (0.82) than in cases (0.76), and the G mutant allele had a higher allele frequency in cases (0.24) than in controls (0.18). There is a statistically significant difference between the rs3803304 CG heterozygous genotype and the four subtypes (P ≤ 0.05; Q ≤ 0.05), where the luminal B subtype was present in 63.6%. Nevertheless, it has been observed that US individuals with esophageal cancer and the rs3803304 CG genotype presented a better response to treatments with chemoradiotherapy (OR = 0.5, 95% CI = 0.25-0.99; P ≤ 0.05) [28]. The results suggest that the rs3803304 variant is capable of regulating DNA transcription mechanisms and, therefore, it is capable of causing risk in the development of breast cancer in the Ecuadorian mestizo population living at high altitudes.
Regarding the rs3803304 variant that presented risk associated with the development of breast cancer, the CG+GG combined genotypes were present in 11  significant association between these combined genotypes and the histopathological characteristics (P > 0.05; Q ≤ 0.05). According to our results, the noncoding region rs3803304 variant may act as a predictive biomarker in the risk of developing breast cancer in the high altitude Ecuadorian mestizo population. Nevertheless, sample size should be increased in future analysis to enrich statistical tests. In conclusion, this study as well as our previous genetic studies on MTHFR in breast and prostate cancer [7,44,45], EGFR in lung cancer [46], and GPX-1 in bladder cancer [47] is an important contribution in order to integrate pharmacogenetics in clinical practice in Ecuador and Latin America [48,49].

Data Availability
All data generated or analyzed during this study are included in this published article and its supplementary information files. Supplementary table 1 and table 3 detail clinical data of cases and controls, respectively. Clinical data is structured by genotype information, estrogen status, progesterone status, HER2/neu status, molecular subtypes, age at diagnosis, affected breast, surgical margins, lymph node status, and tumor stage. Supplementary table 2 details information about polymorphisms, nucleotide change, and amino acid change. All these data support the conclusions of the study.