Akt1 genetic variants confer increased susceptibility to thyroid cancer

The PI3K-Akt-mTOR pathway plays a central role in the development of non-medullary thyroid carcinoma (NMTC). Although somatic mutations have been identified in these genes in NMTC patients, the role of germline variants has not been investigated. Here, we selected frequently occurring genetic variants in AKT1, AKT2, AKT3, PIK3CA and MTOR and have assessed their effect on NMTC susceptibility, progression and clinical outcome in a Dutch discovery cohort (154 patients, 188 controls) and a Romanian validation cohort (159 patients, 260 controls). Significant associations with NMTC susceptibility were observed for AKT1 polymorphisms rs3803304, rs2494732 and rs2498804 in the Dutch discovery cohort, of which the AKT1 rs3803304 association was confirmed in the Romanian validation cohort. No associations were observed between PI3K-Akt-mTOR polymorphisms and clinical parameters including histology, TNM staging, treatment response and clinical outcome. Functionally, cells bearing the associated AKT1 rs3803304 risk allele exhibit increased levels of phosphorylated Akt protein, potentially leading to elevated signaling activity of the oncogenic Akt pathway. All together, germline encoded polymorphisms in the PI3K-Akt-mTOR pathway could represent important risk factors in development of NMTC.


Background
In recent years, the incidence of non-medullary thyroid cancer (NMTC) has steadily increased (1,2,3,4). Although most NMTC patients have a favorable prognosis, 20-30% of patients with locally advanced or metastatic disease is confronted with long-term disease and increased risk of death as no curative treatment options are available (5,6,7).
The intracellular proteins PI3K, Akt and mTOR are part of a central signaling pathway in NMTC tumorigenesis by facilitating signal transduction to induce angiogenesis, metabolic reprogramming, proliferation and invasion diagnosis and histological classification was performed by experienced thyroid cancer pathologists. RAI (I-131) ablation of residual thyroid tissue was performed 4-6 weeks after surgery. Patients were repeatedly treated with RAI to reach remission, if indicated. Cured disease was defined according to institutional cut-off values of TSH stimulated thyroglobulin (Tg, <1 pmol/L in the Dutch patients and <0.04 ng/mL in the Romanian patients) in the absence of anti-Tg antibodies and no evidence of loco-regional disease or distant metastasis on the whole body iodine scans (WBS) and/or neck ultrasonographic examinations at 6-9 months after RAI ablation. Tumor recurrence was defined as new evidence of loco-regional disease or distant metastasis after successful primary therapy. Current disease status was defined as in remission in case of undetectable unstimulated Tg (according to the institutional cut-off) in the absence of anti-Tg antibodies and no evidence of loco-regional disease or distant metastases at the last follow-up visit. Persistent disease was defined as detectable Tg and/or evidence of loco-regional disease or distant metastases. Recurrent disease was defined as new evidence, biochemical (e.g. Tg becoming detectable after having been undetectable) and/or radiological, of loco-regional disease or distant metastases. Histological, clinical and follow-up data were retrieved from the patients' medical records and are shown in Table 1. In addition, 188 Dutch and 260 Romanian healthy, genetically unrelated individuals, having no evidence of NMTC or other malignancies were recruited as population-based control subjects.

Genotyping
Single nucleotide polymorphisms (SNP) were selected based on population frequency, previously published associations with human diseases and/or known functional effects on protein function or gene expression (27,28,29,30,31,32) (Table 2). After obtaining informed consent, blood was drawn from the cubital vein of participants into EDTA collection tubes and subjected to DNA extraction using the GeneJET™ Whole Blood Genomic DNA Purification Mini Kit (Fermentas, Thermo Fisher Scientific) according to the manufacturer's instructions. Until further analysis, DNA samples were stored at −20°C. TaqMan SNP Genotyping assays (Life Technologies) designed with two specific probes and primers for each variant were utilized for genotyping the SNPs in PIK3CA, AKT1, AKT2, AKT3 and MTOR (Table 2). Ten nanograms of genomic DNA were amplified by quantitative PCR (qPCR) in a 7300 Real-Time PCR System 9:11 (Life Technologies), under standard conditions. The realtime PCR included an initial denaturation step at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and then at 60°C for 1 min. Quality control was performed by duplicating samples within and across plates and by the incorporation of positive and negative control samples.

PBMC isolation and Western blotting
For isolation of peripheral blood mononuclear cells (PBMCs), venous blood was drawn from the cubital vein of healthy volunteers into 10 mL EDTA tubes (Monoject). The mononuclear cell fraction was obtained by density centrifugation of blood diluted 1:1 in pyrogen-free saline over Ficoll-Paque (Pharmacia Biotech). Cells were washed twice in saline and suspended in culture medium (RPMI, Invitrogen) supplemented with gentamicin 10 µg/mL, l-glutamine 10 mM and pyruvate 10 mM. Cells were counted in a Coulter counter (Coulter Electronics) and the number was adjusted to 5 x 10 6 cells/mL. For Western blotting, cells were incubated with either culture medium (negative control) or with E. coli lipopolysaccharide (LPS, 100 ng/mL, Sigma) for 30 min, a well established activator of Akt signaling (33,34). For western blotting of (phosphorylated) Akt protein, 5 x 10 6 cells were lysed in 40 µL of lysis buffer (50 mM Tris (pH 7.4), 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 10% glycerol, 1% Triton X-100, 40 mM β-glycerophosphate, 50mM sodium fluoride, 200mM sodium vanadate, 10 mg/mL leupeptin, 10 mg/mL aprotinin, 1 mM pepstatin A, and 1 mM phenylmethylsulfonyl fluoride). The homogenate was frozen and then thawed and centrifuged at 4°C for 3 min at 14,000 g, and the supernatant was mixed with a loading buffer containing dithiothreitol, incubated at  95°C for 15 min, and taken for western blot analysis. Equal amounts of protein were subjected to SDS-PAGE using 10% polyacrylamide gels. After SDS-PAGE, proteins were transferred to nitrocellulose membrane (0.2 mm). The membrane was blocked with 5% (wt/vol) milk powder in TBS/Tween 20 for 1 h at room temperature, followed by incubation overnight at 4°C with a pAkt S473 antibody (1:1000, Cell Signalling #9018) or total Akt antibody (1:1000, Cell Signalling #2938) in 5% BSA in TBS/Tween 20 or with a β-actin antibody (loading control, 1:1000, A2066; Sigma) in 5% milk powder in TBS/Tween 20. After overnight incubation, the blots were washed three times with TBS/Tween 20 and then incubated with horseradish peroxidase-conjugated swine anti-rabbit antibody at a dilution of 1:5000 in 5% (wt/vol) milk powder in TBS/ Tween 20 for 1 h at room temperature. After being washed three times with TBS/Tween 20, the blots were developed with ECL (GE Healthcare) according to the manufacturer's instructions.

Statistical analysis
Genotypes and allele frequencies were calculated and the Hardy-Weinberg equilibrium was assessed using a goodness-of-fit X 2 -test for biallelic markers. The odds ratios (ORs) and 95% CI of the association between genotype frequencies and NMTC susceptibility in addition to clinicopathological characteristics and treatment outcomes were analyzed using logistic regression models. In addition, χ 2 analysis and Fisher's exact test were applied to determine whether tumor size, cumulative RAI activity (subdivided as 30-100 mCi (1.1-3.8 GBq), 101 -200 mCi (3.8-7.4 GBq) or ≥200 mCi (>7.4 GBq)) and disease status after thyroidectomy plus radio-ablation were associated with the genotype of the analyzed genes. All statistical analyses were carried out with SPSS for statistical computing and graphics. Differences in protein amounts detected by western blot were analyzed using the Mann-Whitney U test. Overall, statistical tests were two-sided and a P-value below 0.05 was considered statistically significant.

PI3K-Akt-mTOR pathway SNPs and susceptibility to NMTC
To assess the effects of genetic variation in PI3K-Akt-mTOR genes on susceptibility to NMTC, several SNPs were selected based on previously published associations with human diseases and/or known functional effects on protein function or gene expression. The genotypes corresponding to these SNPs were determined in the Dutch discovery cohort (154 patients, 188 healthy controls) and in the Romanian validation cohort (159 patients, 260 healthy controls).  Tables 3 and 4, respectively. These results demonstrate the association of the rs3803304, rs2494732 and rs2498804 polymorphisms in AKT1 with NMTC susceptibility in the Dutch discovery cohort by applying different genetic association models. Importantly, in the Romanian validation cohort the AKT1 rs3803304 polymorphism was confirmed as genetic risk factor for NMTC in the dominant model. Of note, genotype frequencies in both NMTC patients and controls study populations were in accordance with that expected under the Hardy-Weinberg equilibrium.

PI3K-Akt-mTOR pathway SNPs and clinical outcome of NMTC
Within the NMTC study populations recruited in The Netherlands and Romania, the impact of PIK3CA, AKT1, AKT2, AKT3 and MTOR genotypes on the clinical postoperative treatment response and outcome of NMTC patients was investigated. These analyses revealed that none of the investigated polymorphisms in the PI3K-Akt-mTOR pathway were associated with worse clinical manifestation of NMTC in any of the cohorts regarding histology, TNM staging, RAI treatment response and clinical outcome (Supplementary Tables 1 and 2, see section on supplementary materials given at the end of this article).

Functional consequences of AKT1 rs3803304 polymorphism for pAkt and total Akt protein expression in PBMCs
The observed genetic associations of the AKT1 rs3803304 polymorphism with NMTC susceptibility in both cohorts and of the AKT1 rs2494732 and rs2498804 polymorphisms in only the Dutch cohort suggest that these polymorphisms could influence Akt expression 9:11 Table 3 Genetic distribution of genetic variants in PI3K, Akt and mTOR genes in the Dutch cohort of thyroid carcinoma patients (n = 154) and healthy controls (n = 188).  or function. To assess the potential functional effects of these polymorphisms, healthy individuals were stratified for AKT1 rs3803304, AKT1 rs2494732 or AKT1 rs2498804 genotypes and their PBMCs were tested for differential levels of phosphorylated and total Akt protein, either in the unstimulated condition or after treatment with LPS for 30 min. Since individuals homozygous for the AKT1 rs3803304 minor allele are rare, only subjects either WT or heterozygous for the AKT1 rs3803304 minor allele could be included. The results indicate that no significant differences were apparent in total Akt expression between the genotypes. Interestingly, however, the amount of phosphorylated Akt is elevated in the individuals heterozygous for the AKT1 rs3803304 risk allele in both the unstimulated and LPS-stimulated condition as compared to WT subjects. In contrast, no differences in phosphorylated Akt are apparent between WT and homozygous AKT1 rs2494732 or AKT1 rs2498804 genotypes in either unstimulated or LPS-stimulated conditions ( Fig. 1A and B).

Discussion
In recent years, the major oncogenic pathways have been elucidated that drive tumor initiation and progression in NMTC, which mainly comprise the RAS-RAF-MEK-ERK and PI3K-AKT-mTOR signaling pathways (6,35). Although the contribution of somatic mutations in the activation of oncogenic signaling through these pathways has been well-established, the influence of germline variants on these pathways is still poorly characterized. The present study was performed to assess the effect of germline variants in the oncogenes PIK3CA, AKT1, AKT2, AKT3 and MTOR on NMTC susceptibility and clinical outcome. For this, a Dutch discovery cohort and a Romanian validation cohort were gathered, consisting of NMTC patients and healthy unrelated controls, that allowed the assessment of potential genetic associations of selected germline polymorphisms with susceptibility to NMTC and with its clinical presentation, treatment response and patient outcome. Interestingly, by the present study polymorphisms in the AKT1 gene were shown to be significantly associated with NMTC susceptibility in the Dutch discovery cohort and the Romanian validation cohort. Whereas three AKT1 polymorphisms (rs3803304, rs2494732 and rs2498804) were identified as statistically significant in the Dutch cohort, one of these, the rs3803304 polymorphism, was confirmed in the Romanian cohort. As opposed to polymorphisms in genes encoding PI3K, Akt2, Akt3 and mTOR, these results suggest major consequences of AKT1 polymorphisms for Akt function, especially rs3803304, in modulating the activity of the PI3K-Akt-mTOR signaling pathway. Importantly, this genetic association was observed in both cohorts despite the statistically significant differences in clinical parameters between the Dutch and Romanian patient cohorts listed in Table 1. For the other AKT1 polymorphisms that were only significantly associated with NMTC susceptibility in the Dutch cohort, it cannot be excluded that the lack of association with NMTC susceptibility in the Romanian cohort could be clarified by the differential distribution of these clinical parameters.
Additional analyses were performed to assess whether the selected polymorphisms are associated with clinical parameters including histology, TNM staging, cumulative RAI activity and remission rates. No statistically significant differences were observed, suggesting that AKT1 polymorphisms are involved in tumor initiation rather than in processes of tumor progression, RAI therapy resistance and disease persistence.
Of note, with solid statistical significance the Romanian cohort received twice more low and medium RAI activities as compared to the Dutch patients, however is not associated with differences in remission rates. Again, this suggests that AKT1 polymorphisms are involved in tumor initiation rather than in processes of tumor progression, despite the I-131 activities used for treatment. Also, there are differences in radiation exposure; Romania was among the countries affected by Chernobyl fallout and considering the average age of the Romanian patients (52 ± s.d.) this would be interesting to be studied in relation to the AKT1 polymorphisms.
By functional assays it was demonstrated that the AKT1 rs3803304 polymorphism, in contrast to the AKT1 rs2494732 and AKT1 rs2498804 polymorphisms, has a major effect on Akt phosphorylation, both in the naïve state and upon activation of the PI3K-Akt-mTOR pathway; cells bearing the heterozygous GC genotype, with the C allele conferring increased NMTC risk, exhibited elevated levels of phosphorylated Akt as compared to the GG genotype. These major functional consequences of the AKT1 rs3803304 polymorphism provides mechanistic insights into the observed genetic association. Furthermore, these findings support the current evidence that the PI3K-Akt-mTOR signaling pathway plays a major role in NMTC and could represent a promising strategy for targeted treatment (5,19,25,36).
Previously, the AKT1 rs3803304 polymorphism has been demonstrated to also influence susceptibility, disease progression or clinical outcome of head and neck squamous cell carcinoma, lung carcinoma and esophageal carcinoma, indicating its major functional and clinical implications. In case susceptibility analyses were performed in these studies, the AKT1 rs3803304 minor allele was associated with increased cancer susceptibility in all, confirming the contribution of the minor C allele in the etiology of multiple cancer types (28,31,32,37,38). Additionally to these reports, the present study provides mechanistic insights into the genetic association by linking the AKT1 rs3803304 minor C allele with elevated Akt phosphorylation. The exact biological consequences of the polymorphism and whether it promotes Akt phosphorylation or inhibits Akt dephosphorylation remains to be determined.
Although the heterozygous GC genotype, and most likely also the homozygous CC genotype, are demonstrated to predispose to development of NMTC by inducing hyperactivation of the PI3K-Akt-mTOR pathway upon the encounter of activating stimuli, it should be emphasized that this germline genetic variant is not capable of evoking thyroid tumorigenesis by itself because of limited genetic penetrance, but rather represents a risk modifier. Within the context of NMTC, these activating stimuli could range from growth factors to metabolites (e.g. lactate) and inflammatory molecules (e.g. danger-associated molecular patterns and pro-inflammatory cytokines) produced by the tumor microenvironment, triggering either receptor tyrosine kinases, metabolic receptors, toll-like receptors or cytokine receptors expressed by follicular thyroid (tumor) cells (39,40,41).
In conclusion, the present study suggests that germline variants in the AKT1 gene are an important risk factor in the etiology of NMTC, reinforcing the clinical utility of kinase inhibitors targeting the PI3K-Akt-mTOR pathway to abrogate pathological signaling driving the NMTC malignant process.