The oncogenic mutation in the pleckstrin homology domain of AKT1 in endometrial carcinomas

Abstract

Background:

The phosphatidylinositol 3′-kinase (PI3K)–AKT pathway is activated in many human cancers and plays a key role in cell proliferation and survival. A mutation (E17K) in the pleckstrin homology domain of the AKT1 results in constitutive AKT1 activation by means of localisation to the plasma membrane. The AKT1 (E17K) mutation has been reported in some tumour types (breast, colorectal, ovarian and lung cancers), and it is of interest which tumour types other than those possess the E17K mutation.

Methods:

We analysed the presence of the AKT1 (E17K) mutation in 89 endometrial cancer tissue specimens and in 12 endometrial cancer cell lines by PCR and direct sequencing.

Results:

We detected two AKT1 (E17K) mutations in the tissue samples (2 out of 89) and no mutations in the cell lines. These two AKT1 mutant tumours do not possess any mutations in PIK3CA, PTEN and K-Ras.

Interpretation:

Our results and earlier reports suggest that AKT1 mutations might be mutually exclusive with other PI3K–AKT-activating alterations, although PIK3CA mutations frequently coexist with other alterations (such as HER2, K-Ras and PTEN) in several types of tumours.

The AKT serine/threonine kinases regulate diverse cellular processes, including cell survival, proliferation, invasion and metabolism (Vivanco and Sawyers, 2002). The phosphatidylinositol 3′-kinases (PI3Ks) are widely expressed lipid kinases that catalyse the production of the second messenger phosphatidylinositol 3,4,5-triphosphate (PIP3), which activates AKT by recruitment to the plasma membrane through direct contact of its pleckstrin homology (PH) domain (Stokoe et al, 1997; Lemmon and Ferguson, 2000). Constitutive PI3K–AKT pathway activation can result from various types of alterations in this pathway, including mutation or amplification of receptor tyrosine kinases (such as EGFR and HER2), mutation of Ras, mutation or amplification of PIK3CA (the p110α catalytic subunit of PI3K) and inactivation of the tumour suppressor gene, PTEN (Yuan and Cantley, 2008). In addition to amplifications in multiple AKT isoforms in pancreatic, ovarian and head and neck cancers (Engelman et al, 2006), a somatic missense mutation in the PH domain of AKT1 (E17K) was identified in breast, colorectal, ovarian and lung cancers and in melanoma (Carpten et al, 2007; Bleeker et al, 2008; Davies et al, 2008; Malanga et al, 2008). However, the AKT1 mutation has not been identified in hepatocellular, gastric and pancreatic cancers, leukemia, as well as in glioblastoma multiforme (Bleeker et al, 2008; Cao et al, 2008; Kim et al, 2008; Mahmoud et al, 2008; Mohamedali et al, 2008; Riener et al, 2008; Zenz et al, 2008). Further study is required to fully understand which tumour types take advantage of Akt1 (E17K) mutations to activate the PI3K–AKT pathway.

We reported earlier that PIK3CA mutations frequently coexist with other PI3K-activating alterations in breast (with HER2 and HER3) and endometrial cancers (with PTEN and K-Ras), and that mutant p110α combined with mutant Ras efficiently transformed immortalised human mammary epithelial cells (Oda et al, 2005, 2008). Frequent overlapping mutations of K-Ras and PIK3CA were also reported in colorectal cancer (Parsons et al, 2005). Although coexistent mutations of AKT1 and PIK3CA mutations are suggested to be infrequent in breast cancer (Carpten et al, 2007; Bleeker et al, 2008), it remains to be elucidated whether AKT1 mutations are mutually exclusive with all the other PI3K–AKT-activating alterations in various tumour types.

Endometrial cancer is one of the tumour types in which the PI3K–AKT pathway is frequently activated by alterations of various genes. The frequency of mutations for PTEN, PIK3CA and K-Ras in endometrial cancer is reported as 54, 28 and 11%, respectively (Yuan and Cantley, 2008). In this study, we screened 89 endometrial carcinoma specimens and 12 endometrial carcinoma cell lines for mutations in Akt1 (E17K) and analysed whether AKT1 mutations coexist with any mutations in PTEN, PIK3CA and K-Ras.

Materials and methods

Tumour samples and genomic DNA

Surgical samples were obtained from 89 patients with primary endometrial carcinomas who underwent resection of their tumours at the University of Tokyo Hospital. All patients provided informed consent for the research use of their samples and the collection, and the use of tissues for this study was approved by the appropriate institutional ethics committees. Genomic DNA was extracted by a standard SDS-proteinase K procedure. Patient characteristics (histology, tumour grade and stage) are available in Supplementary Table 1. A detailed distribution of the histological subtypes was as follows; 81 (90%) endometrioid adenocarcinomas, 3 adenosquamous carcinomas, 1 clear cell carcinoma, 1 squamous cell carcinoma and 3 mixed carcinomas.

PCR and sequencing

The primer sequences and PCR conditions of exon 4 of the AKT1 gene are forward: 5′-CACACCCAGTTCCTGCCT G-3′ and reverse: 5′-CCTGGTGGGCAAAGAGGGCT-3′. The PCR amplifications were with denaturation at 94°C for 5 min, followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, 72°C for 60 s and final extension at 72°C for 10 min. The PCR conditions and the PCR primers for PIK3CA (exons 9 and 20), PTEN (exons 1–9) and K-Ras (exons 1 and 2) were described earlier (Minaguchi et al, 2001; Samuels et al, 2004; Oda et al, 2008). The PCR products were sequenced using the BigDye (Applied Biosystems, Foster City, CA, USA) terminator method on an autosequencer.

Cell lines

In this study, AN3CA, KLE, HEC-1B and RL95-2 were obtained from the American Type Culture Collection (Manassas, VA, USA) and HHUA was obtained from the RIKEN CELL BANK (Tsukuba, Japan). Ishikawa3-H-12 was a generous gift from Dr Masato Nishida (Kasumigaura Medical Center, Ibaraki, Japan). HEC-6, HEC-50B, HEC-59, HEC-88, HEC-108 and HEC-116 cell lines were also analysed in this study. The culture condition of all these cell lines was described earlier (Oda et al, 2008).

DNA methylation analysis

Bisulphite treatment was performed using the EZ DNA methylation kit (Zymo Research, Orange, CA, USA). As described earlier (Ehrich et al, 2006), we used Sequenom's MassARRAY platform to perform quantitative methylation analysis of multiple CpG sites for PTEN in 53 endometrial tumour specimens (Sequenom, San Diego, CA, USA). Chromosomal localisation of CpG islands for PTEN and the primer sequences in this study are shown in Supplementary Figure 1.

Immunohistochemistry (IHC)

Immunohistochemistry for PTEN on 4-μm tissue sections was performed and evaluated as described earlier (Minaguchi et al, 2007). In this study, the anti-PTEN Rabbit monoclonal antibody (138G6) (Cell Signaling, Beverly, MA, USA) was applied at a dilution of 1 : 100.

Single nucleotide polymorphism (SNP) array

Single nucleotide polymorphism array was performed in the two AKT1 mutant tumours with tumour DNA. Experimental procedures for GeneChip were performed according to GeneChip Expression Analysis Technical Manual (Affymetrix, Santa Clara, CA, USA), using a Human mapping 50K Array Xba I (Affymetrix).

Results and discussions

The sequencing analysis for exon 4 of the AKT1 gene in 89 tumour tissue samples of endometrial carcinomas showed the point mutation of G to A at nucleotide 49 (E17K) in two tissue samples (2.2%) (Figure 1). Both of the tumours were well-differentiated endometrioid adenocarcinomas with positive oestrogen receptor and progesterone receptor, suggesting that these two tumours are oestrogen dependent (corresponded to type I endometrial cancer). No mutations were detected in the 12 endometrial cancer cell lines.

Figure 1.

The sequence traces of two tumours and a normal control for exon 4 of AKT1. The E17K mutation is caused by a missense mutation (G to A) indicated. In tumour-2, the level of the mutant band (A) is much higher than that of the wild-type band (G). It is possible that this weak band is derived from DNA of normal cells and that the tumour might lose one allele at this locus.

Thereafter, we attempted to figure out the exclusivity of AKT1 mutations and other PI3K–AKT-activating mutations (Supplementary Table 1). The genotypic pattern of the four genes (PTEN, PIK3CA, K-Ras and AKT1) in 97 endometrial carcinomas (85 tumour tissue samples and 12 cell lines) was shown in Table 1. Coexistence with other mutations is frequently observed in the PIK3CA mutant (28 of 34; 82%) and in the K-Ras mutant (13 out of 17; 76%) tumours, but the two AKT1 mutant tumours do not possess any mutations in PTEN, PIK3CA and K-Ras. As PI3K and PTEN are competitive for PIP3 production, the PIK3CA mutation might require another upstream input or PTEN loss itself to fully activate the PI3K–AKT pathway. As AKT1 (E17K) functions downstream of PTEN and shows constitutive localisation to the plasma membrane in the absence of serum stimulation (Carpten et al, 2007), mutant AKT1 (E17K) alone might be sufficient for complete activation of this pathway.

Table 1. PI3K–AKT activating mutations and their coexistence in 97 endometrial cancers.

	n (%)
Wild-type	24 (25)
AKT1 mutation alone	2 (2)
K-Ras mutation alone	4 (4)
PIK3CA mutation alone	6 (6)
PTEN mutation alone	30 (31)
Double mutations of K-Ras and PIK3CA (w/o PTEN mutation)	2 (2)
Double mutations of K-Ras and PTEN (w/o PIK3CA mutation)	3 (3)
Double mutations of PIK3CA and PTEN (w/o K-Ras mutation)	18 (19)
Triple mutations of K-Ras, PIK3CA and PTEN	8 (8)

PI3K, phosphatidylinositol 3′-kinase.

Wild-type, no mutations in PTEN, PIK3CA, K-Ras and AKT1.

We also analysed DNA methylation and protein expression of PTEN, as hypermethylation and loss of heterozygosity (LOH) are other mechanisms to inactivate PTEN (Teng et al, 1997; Blanco-Aparicio et al, 2007). Quantitative analysis of DNA methylation using Sequenom's MassARRAY platform did not find promoter hypermethylation of PTEN in all the 53 samples that were examined (Supplementary Figure 2 and Supplementary Table 2), including the two AKT1 mutant tumours. Although PTEN methylation had been reported in 18% of endometrial carcinomas (Salvesen et al, 2001), Zysman et al (2002) suggested that the pseudogene on chromosome 9 (Genbank accession number: AF040103), not PTEN, is predominantly methylated in endometrial carcinomas. In IHC, both tumours with the AKT1 mutation were stained positively for PTEN in the cytoplasm, whereas all the four tumours with multiple frameshift mutations in PTEN were stained negatively (Supplementary Figure 3). We evaluated the chromosomal imbalances in the two AKT1 mutant tumours, using SNP array (with more than 50 000 SNPs). Single nucleotide polymorphism array analysis showed that the two AKT1 mutant tumours do not show copy number changes in the locus of PTEN (10q23.1) (data not shown). These data also support the fact that AKT1 mutations are mutually exclusive with PTEN inactivation.

We found multiple PTEN mutations in 13 out of 85 clinical specimens and in 8 out of 12 endometrial cell lines (Supplementary Table 1), whereas LOH of PTEN was reported approximately at 30% in endometrial carcinomas (Toda et al, 2001). Thus, biallelic PTEN inactivation might be achieved through either biallelic mutations or monoallelic mutation with LOH in endometrial carcinomas. Considering the correlation between PTEN mutations and microsatellite instability (MSI) in endometrial carcinomas (Bilbao et al, 2006), it would be of interest to analyse whether AKT1 and the other mutations in the PI3K pathway genes are also associated with MSI.

To date, AKT1 (E17K) mutations have been reported in breast (25 out of 427; 5.9%), colorectal (4 out of 243; 1.6%), lung (4 out of 636; 0.6%) and ovarian cancers (1 out of 130; 0.8%) and in melanoma (1 out of 202; 0.5%). Breast, colorectal and endometrial cancers are the tumour types that frequently possess PIK3CA mutations (Campbell et al, 2004; Samuels et al, 2004; Oda et al, 2005). In lung cancer, the AKT1 mutation was detected only in squamous cell carcinomas and not in any adenocarcinomas, which is in agreement with the higher incidence of PIK3CA mutations or amplifications in squamous cell carcinomas than adenocarcinomas (Kawano et al, 2006, 2007; Malanga et al, 2008). These data suggest that the AKT1 mutation might occur in a tissue-specific manner and is more associated with the tumour types with frequent PIK3CA alterations.