Involvement of TSC genes and differential expression of other members of the mTOR signaling pathway in oral squamous cell carcinoma

Background Despite extensive research, the five-year survival rate of oral squamous cell carcinoma (OSCC) patients has not improved. Effective treatment of OSCC requires the identification of molecular targets and signaling pathways to design appropriate therapeutic strategies. Several genes from the mTOR signaling pathway are known to be dysregulated in a wide spectrum of cancers. However, not much is known about the involvement of this pathway in tumorigenesis of OSCC. We therefore investigated the role of the tumor suppressor genes, TSC1 and TSC2, and other members of this pathway in tumorigenesis of OSCC. Methods Expression of genes at the RNA and protein levels was examined by semi-quantitative RT-PCR and western blot analyses, respectively. Loss of heterozygosity was studied using matched blood and tumor DNA samples and microsatellite markers from the TSC1, TSC2 and PTEN candidate regions. The effect of promoter methylation on TSC gene expression was studied by treating cells with methyltransferase inhibitor 5-azacytidine. Methylation status of the TSC2 promoter in tissue samples was examined by combined bisulfite restriction analysis (COBRA). Results The semi-quantitative RT-PCR analysis showed downregulation of TSC1, TSC2, EIF4EBP1 and PTEN, and upregulation of PIK3C2A, AKT1, PDPK1, RHEB, FRAP1, RPS6KB1, EIF4E and RPS6 in tumors. A similar observation was made for AKT1 and RPS6KB1 expression in tumors at the protein level. Investigation of the mechanism of downregulation of TSC genes identified LOH in 36.96% and 39.13% of the tumors at the TSC1 and TSC2 loci, respectively. No mutation was found in TSC genes. A low LOH rate of 13% was observed at the PTEN locus. Treatment of an OSCC cell line with the methyltransferase inhibitor 5-azacytidine showed a significant increase in the expression of TSC genes, suggesting methylation of their promoters. However, the 5-azacytidine treatment of non-OSCC HeLa cells showed a significant increase in the expression of the TSC2 gene only. In order to confirm the results in patient tumor samples, the methylation status of the TSC2 gene promoter was examined by COBRA. The results suggested promoter hypermethylation as an important mechanism for its downregulation. No correlation was found between the presence or absence of LOH at the TSC1 and TSC2 loci in 50 primary tumors to their clinicopathological variables such as age, sex, T classification, stage, grade, histology, tobacco habits and lymph node metastasis. Conclusion Our study suggests the involvement of TSC genes and other members of the mTOR signaling pathway in the pathogenesis of OSCC. LOH and promoter methylation are two important mechanisms for downregulation of TSC genes. We suggest that known inhibitors of this pathway could be evaluated for the treatment of OSCC.


Background
Oral squamous cell carcinoma (OSCC) is the sixth most common cancer in the world [1]. In India, it is the leading cancer among males and the third most common malignancy in females [1]. The five-year survival rate for OSCC is the lowest among all major cancers [1]. The etiology of this cancer is multifactorial, with important risk factors being tobacco intake, alcohol consumption and human papilloma virus (HPV).
A thorough understanding of the genetic and epigenetic changes that result in the activation of signaling pathways and provide the cells with a growth advantage during oral tumorigenesis is essential for the development of novel therapeutic strategies. Agents that can inhibit or reverse these changes by targeting molecularly defined pathways should receive increased attention as novel candidates for oral cancer prevention and therapy [2,3]. The molecular interplay between phosphoinositide-3-kinase, catalytic, alpha polypeptide (PIK3CA) and FK506 binding protein 12-rapamycin associated protein 1 (FRAP1) of the mTOR (mammalian target of rapamycin) signaling pathway in the control of cell growth and proliferation has been the subject of much interest among cell biologists [4]. Tuberin, encoded by the tumor suppressor gene tuberous sclerosis 2 (TSC2), and its interacting partner hamartin, encoded by another tumor suppressor gene tuberous sclerosis 1 (TSC1), have been placed as a complex in the mTOR signaling pathway and negatively regulate the pathway to inhibit mTOR mediated downstream signaling [4]. Several components of the mTOR signaling pathway are known to be dysregulated in a wide spectrum of human cancers [5]. Although some components (PIK3C2A, AKT1, PTEN, RPS6 and EIF4E) of this pathway have been implicated in OSCC [6][7][8][9], a comprehensive analysis is lacking. Further, very little is known about the roles of TSC tumor suppressor genes in tumorigenesis of OSCC [10]. The main aim of this study was to assess the role of TSC genes and other members of this pathway in the tumorigenesis of OSCC. The results of our study are presented here.

Sample collection
A total of 52 OSCC (oral cancer) samples were ascertained at Bangalore Institute of Oncology, Bangalore. All tumor samples were from the tongue and cheek areas of the mouth. Lesions were situated at the anterior 2/3 of the tongue over the lateral borders, a common site in Indian patients. This study was performed with informed written consent from the patients and approval from the ethics committees of the Bangalore Institute of Oncology and Indian Institute of Science. The specimens were obtained as biopsy or surgical samples from oral cancerous lesions and adjacent normal mucosa (taken from the farthest margin of the surgical resection). The patients had not been treated at the time of biopsy/surgery. The clinicopathological data for 52 patients is given in Table 1. Tumors were classified according to TNM (Tumor, Node and Metastasis) criteria [11]. Peripheral blood samples were also collected in EDTA-Vacutainer (Beckton-Dickinson, Franklin Lakes, NJ) tubes from 52 patients.

Cell culture
Three oral cancer cell lines (SCC 131, SCC 104 and KB) and four other cell lines (HeLa, HepG2, A549 and HEK-293T) were used. SCC 131 and SCC 104 cell lines were a kind gift from Dr. Susanne M. Gollin (University of Pittsburgh, Pittsburgh, PA). Cell lines were maintained either in Minimum Essential Medium with Earle's salt and lglutamine or in Dulbecco's modified Eagle's medium (Sigma-Aldrich, St. Louis, MO).

Genomic DNA isolation
Genomic DNA was isolated from peripheral blood and tumor samples using a DNA isolation kit (Roche Diagnos-tics™, Mannheim, Germany).

Semi-quantitative RT-PCR
Total RNA was isolated from 16 paired normal and tumor samples using the TRI REAGENT™ (Sigma-Aldrich, St. Louis, MO). cDNA was synthesized from 1 μg total RNA from each sample using random hexamers and the Rever-taid™ H Minus First Strand cDNA Synthesis Kit (MBI Fermentas, Burlington, ON, Canada). For RT-PCR, forward and reverse primers were selected from two different exons of genes to rule out the possibility of amplification of contaminating genomic DNA. Primer sequences and PCR conditions are available from the authors upon request. For each gene, the PCR protocol was optimized in order to get the amplification in a linear phase. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was amplified as a normalizing control. Images of RT-PCR ethidium bromide stained agarose gels were acquired with a Kodak CCD camera and quantification of the bands was performed by densitometric analysis using the Kodak Digital Science Image Station Imaging Software version 3.6.1. Band intensity was expressed as relative absorbance units. Data was expressed in arbitrary units (relative expression) as a ratio of normal/GAPDH and tumor/GAPDH and plotted using the GraphPad Prism software version 4.00 (GraphPad Prism Software, San Diego, CA). The significance of difference in mRNA levels between normal and tumor samples for a gene was assessed by Student's t-test and the results are expressed as mean ± SEM [12]. A probability value of p < 0.05 was assumed to be significant. PCR amplification for each gene was repeated once. A gene was considered to be upregulated when its mean expression value across 16 tumor samples was significantly higher than the mean expression value across 16 normal tissue samples and vice versa [13]. We defined the cutoff value for determining the upregulation or downregulation of a gene in a tumor sample as ≥ 1.8 fold difference in its expression between normal and tumor samples as described by Arora et al. [3] for differentially expressed genes in oral squamous cell carcinoma.

Mutation analysis
Mutation screening of the entire coding regions of TSC genes was carried out using PCR-SSCP [14] and DNA sequencing techniques.

LOH analysis at TSC1, TSC2 and PTEN loci
For LOH studies, matched normal and tumor DNA samples from 50 patients were genotyped using following microsatellite markers: D9S179, D9S1830 and D9S915 for the TSC1 locus; D16S3024, D16S3395 and D16S475 for the TSC2 locus; and D10S215, D10S1765 and D10S541 for the PTEN locus. Microsatellite analysis was performed as described in Kumar et al. [15]. LOH was scored if there was a complete loss of one of the two heterozygous alleles in tumor DNA or a decrease of 50% intensity of one of the two alleles in tumor DNA as compared to the corresponding peripheral blood DNA (allelic imbalance).
For western blot analysis, whole cell lysates were prepared from matched normal and tumor samples as well as cell lines using a standard procedure. Equal amounts of protein (~100 μg/lane) from tumor, normal oral tissue or different cancer cell lines were resolved by SDSpolyacrylamide gel electrophoresis and transferred onto a PVDF membrane. Primary antibody was detected with either HRP-conjugated goat anti-rabbit or goat antimouse secondary antibodies (Bangalore Genei™, India). Immunoreactive bands were visualized using the Western Lightning Chemiluminescence Reagent kit (PerkinElmer Life Sciences, Boston, MA) and X-ray films. β-actin was used to see equal protein loading.

5-azacytidine treatment of cell lines
SCC 131 and HeLa cells were seeded at a density of 1 × 10 6 cells/90 mm dish. After 24 hr, freshly prepared 5-azacytidine (Sigma-Aldrich, St. Louis, MO) was added into the dish to a final concentration of 10 μM. Total RNA was isolated after 2 and 5 days from the start of the treatment. Untreated cells were used as controls. Semi-quantitative RT-PCR was used to assess the expression of TSC1 and TSC2. GAPDH was used as a normalizing control.

Combined bisulfite restriction analysis
Methylation status of the TSC2 gene promoter was examined using combined bisulfite restriction analysis (COBRA) as described by Xiong and Laird [16]. Sodium bisulfite treated DNA was used in PCR amplification using primers designed for the bisulfite treated DNA. Primers were designed using the MethPrimer program [17]. Promoter region of the TSC2 gene is reported by Kobayashi et al. [18]. Sodium bisulfite treated DNA was amplified with following TSC2 promoter primers: F-5'gggattttagtttgtagtttttattt-3' and R-5'-ccataacttaaaactaaaaaatact-3'. Primers were designed to exclude binding to any CpG dinucleotide to ensure amplification of both methylated and unmethylated forms of DNA. PCR conditions for primers were as follows: an initial denaturation at 95°C for 3 min was followed by 35 cycles of 94°C for 30 sec, 60°C for 45 sec and 68°C for 45 sec with a final extension at 68°C for 5 min. TSC2 primer set generated a 571 bp amplicon. A second PCR was carried out using the product of the first amplification as a template to get enough DNA for COBRA. Approximately 500-600 ng of pooled and gel purified PCR product was digested with Aci I at 37°C for 6 hr. Digests were resolved in a 2.5% agarose gel and visualized by ethidium bromide staining. There are 18 Aci I sites and 65 CpGs in the TSC2 promoter. The restriction enzyme Aci I recognizes the sequence 5'-GCGG-3'. The cleavage of this sequence will occur only when the C residue in the recognition sequence is methylated. Promoter of TSC1 has been reported by Ali et al. [19]. Primers were also designed for the TSC1 promoter. However, despite repeated efforts using DNA polymerases from several vendors, we were not successful in amplifying the TSC1 promoter after bisulfite treatment.

Downregulation of TSC genes
Expression levels of TSC genes were studied by semi-quantitative RT-PCR analysis in a panel of 16 matched normal and tumor samples. The mean expression level of TSC1 was significantly lower (1.31 ± 0.18 in normal vs. 0.79 ± 0.11 in tumor, p = 0.0208) in tumor samples (Figure 1a). The mean expression level of TSC2 was also significantly lower (1.56 ± 0.15 in normal vs. 0.88 ± 0.12 in tumor, p = 0.0013) in tumor samples (Figure 1a), suggesting the involvement of these genes in the etiology of oral cancer.

Relative expression
Relative expression TSC1 and TSC2 were also downregulated in eight matched normal and tumor samples at the protein level ( Figure 1b). We wanted further to determine the expression of both the TSC genes in cell lines. Interestingly, TSC2 did not show a detectable level of expression in an oral cancer cell line SCC 131, whereas it was expressed in two other oral cancer cell lines KB and SCC 104 as well as in A549, HEK-293T, HeLa and HepG2 at the protein level ( Figure 1c). However, the TSC2 expression in KB was lower than in other cell lines (Figure 1c). TSC1 was expressed in all the cell lines, albeit at different levels (Figure 1c).

Mechanisms of downregulation of TSC genes
We then sought to determine the mechanism of downregulation of TSC genes in tumor samples. Given the fact that both are tumor suppressors, we hypothesized that their downregulation could be due to inactivating somatic mutations, LOH and/or promoter methylation in tumors.  in patient 40 at D9S179 and patient 53 at D16S3395 and D16S475 (Figure 2b). No LOH was found in two ED (epithelial dysplasia) samples at either of the two TSC loci (data not shown).
In spite of the exciting link of TSC genes with the mTOR signaling pathway so widely linked with the onset and progression of tumorigenesis, only few studies have investigated the role of TSC genes in sporadic cancers. Hebert et al. [10] have detected a few mutations in both TSC genes in head and neck tumors from the U.S. patients and SCC cell lines. Although reduced expression of tuberin was found in sporadic astrocytomas, no intragenic mutations were detected in either TSC gene [20]. LOH was seen in astrocytomas, ependymomas, gangliogliomas, glioblastoma multiforme, oligodendrogliomas and pilocytic astrocytomas at both the TSC loci [20]. Parry  without somatic mutations has been found to be responsible for the downregulation of these genes. For example, LOH without somatic mutations in TSC genes was found to be an important mechanism for downregulation of these genes in sporadic glial and glioneuronal tumors [20]. Promoter methylation and LOH without somatic mutations are known to downregulate ATM and FHIT genes in breast cancer [26,27]. It is reasonable to assume that the LOH due to deletion of one of the two alleles or due to loss of an entire chromosome with TSC1 or TSC2 gene in tumors will leave only one allele, resulting in the downregulation of these genes.
Aberrant hypermethylation of promoter C P G islands has been found to be an important alternative mechanism to intragenic mutations for the inactivation of tumor suppressor genes [28]. Since we did not find any somatic mutation in TSC genes in oral tumors, we therefore ascribed the downregulation of TSC genes in oral cancer to an epigenetic alteration, resulting in the methylation of TSC1 and TSC2 promoters. To investigate if the downregulation of TSC genes in tumors is due to their promoters being methylated, we selected the oral cancer cell line SCC 131, which did not show expression of TSC2 and a low level of expression of TSC1 (Figure 1c), and treated it with methyltransferase inhibitor 5-azacytidine. RT-PCR data showed a significant increase in the expression of both the TSC genes in this cell line following the treatment ( Figure  3). HeLa cells also showed a significant increase in the expression of TSC2 after 5-azacytidine treatment ( Figure  3b). However, no significant difference was observed in the expression of TSC1 in HeLa cells following the drug treatment for 2 and 5 days (Figure 3b). No change in the expression of both the genes was seen in both cell lines grown for 2 and 5 days without the drug treatment ( Figure  3b). In order to see if the downregulation of TSC genes is due to their promoters being methylated in tumors from the patients, we examined the methylation status of the promoters of both the TSC genes in a panel of 16 oral tumors, three normal oral tissues, two peripheral blood DNA samples from normal individuals, and two cell lines HeLa and SCC 131 by COBRA. Our repeated efforts to amplify the TSC1 promoter using different Taq DNA polymerases failed. We believe that this could be due to the nature of the sodium bisulfite treated DNA. However, we were able to successfully amplify the 571 bp long TSC2 promoter region. As can be seen from the Figure 4a, a 571 bp PCR product from all the tumors and both cell lines showed digestion, whereas PCR products from peripheral blood samples from two normal individuals and three normal oral tissues did not show digestion. Digested bands of varying sizes and intensities were seen in different tumor samples and both cell lines, in addition to a prominent digested band of ~175 bp and the uncut band of 571 bp (Figure 4a). In order to see if the digestion by Aci I was specific, we treated the PCR products with the restriction enzyme buffer only. The absence of any digestion suggested that the Aci I digestion of the samples was specific (data not shown). Sequence analysis of the PCR products from sodium bisulfite treated normal and tumor DNA showed that the treatment was specific (Figure 4b).
The variation in the degree of digestion by Aci I in different tumor samples might be accounted for by the fact that all surgical samples are likely to contain a heterogenous Expression of TSC genes in cancer cell lines following treatment with 5-azacytidine for 2 (2 d) and 5 (5 d) days ples. An MSP (methylation-specific PCR) assay could not be designed in this study because the relevant methylated region was limited. Taken together, our 5-azacytidine and COBRA data suggested that TSC genes are targets of epigenetic inactivation in oral cancer (Figures 3 and 4).
Upregulation of PIK3C2A has been reported in several cancers such as cervical, colon, breast, liver, stomach and lung cancers [29]. Expression of this gene has been investigated in oral tongue carcinoma, and head and neck cancer cell lines where the mean expression level in tumor samples was found to be significantly higher than in normal samples [7]. Our results have also shown that PIK3C2A is upregulated in 10/16 oral tumors (Figure 5a, Table 2). AKT1 (AKT) is a downstream effecter of PIK3C2A (PI3K). It has emerged as a central player controlling several signal transduction pathways that are activated in response to growth factors or insulin. Activation of AKT1 has been shown to be a frequent event in breast, colorectal, ovarian, pancreatic, and head and neck cancers [30]. One of the best characterized regulators of the mTOR signaling pathway is PTEN. The lipid phosphatase activity of PTEN acts as a negative regulator for PIK3C2A induced signaling as it dephosphorylates PIP3 [31]. PIP3 is a potent second messenger that recruits certain kinases to the plasma membrane including the Protein kinase B/Akt family of kinases and PDPK1. On membrane localization, AKT1 is activated in part through phosphorylation by PDPK1 and elicits several downstream cellular functions. Genetic inactivation of PTEN leads to constitutive activation of the mTOR pathway [32]. Our study has shown that AKT1 and PDPK1 show upregulation in 10/16 and 6/16 tumors respectively (Figure 5a, Table 2). PTEN on the other hand showed downregulation in 7/16 tumors (Figure 5a, Table  2). Mavros et al.
[33] identified a low LOH rate of 12% in 50 samples of OSCC. They however did not find any mutation in the coding region of PTEN and concluded that the PTEN gene alterations do not play a key role in tumorigenesis of oral squamous cell cancers. We have also found a low LOH rate of 13% (6/46 informative cases) at the PTEN locus in the same panel of 50 paired blood and tumor DNA samples (data not shown). The frequency of LOH was 4.5% (2/44 informative cases) and 10% (3/30 informative cases) at D10S1765 and D10S541 respectively (data not shown). It is possible that the downregulation of this gene in oral tumors examined in this study is due to inactivating somatic mutations or its promoter methylation. However, these possibilities need to be investigated in the future.
FRAP1 has a central role in controlling cell cycle progression and cell growth. It has emerged as a major cancer therapeutic target [34]. FRAP1 exerts its effect by phosphorylating EIF4EBP1 (4E-BP1) which binds to and inactivates EIF4E, thus inhibiting 5'-cap-dependent mRNA translation. Phosphorylation of EIF4EBP1 releases EIF4E and allows initiation of translation. Regulation of EIF4E mediated translation is an important target for therapeutic intervention in light of the fact that EIF4E has been shown to be overexpressed in several cancers and that overexpression can cause malignant transformation of rodent fibroblasts [35]. In our study, EIF4EBP1 showed downregulation in 10/16 tumors (Figure 5a, Table 2). Upregulation of EIF4E in 10/16 tumors (Figure 5a, Table  2) potentiates its role in the increase of translation leading to overall cell growth and proliferation. FRAP1 also regulates translation via phosphorylation of a serine/threonine kinase p70S6K1 (RPS6KB1). Upon phosphorylation, p70S6K1 promotes translation of mRNAs containing a 5' terminal oligopyrimidine (5' TOP) by phosphorylating the ribosomal subunit S6. Since ribosomal proteins and translation elongation factors are encoded by 5' TOP mRNAs, signaling along the p70S6K1 pathway promotes ribosome biogenesis and overall protein biosynthetic capacity [4]. Our study provides the evidence that FRAP1 is upregulated in 8/16 tumors, as also are RPS6KB1 (9/16) and RPS6 (10/16) (Figure 5a, Table 2). Our western blot results also indicated that the AKT1/RPS6KB1 pathway is active in oral tumors, as phosphorylated forms of both proteins show increased levels in tumor samples ( Figure  5b).
A novel positive regulator of FRAP1 is the small GTPase RHEB [4]. Tuberin/hamartin complex acts as a negative regulator of this pathway by C-terminal GAP activity of tuberin towards RHEB. When stimulated by growth factors, AKT1 relieves this inhibition by phosphorylation of tuberin, which dissociates the tuberin/hamartin complex [4]. Phosphorylated tuberin binds to the 14-3-3 family of proteins which control various cellular functions [4]. We found upregulation of RHEB in 9/16 tumors (Figure 5a, Table 2), whereas we did not find any significant difference in the expressions of the tuberin interacting protein YWHAB (14-3-3β) and IRS1 across the samples analyzed (data not shown). Interestingly, 14-3-3zeta, another tuberin interacting protein [36], was recently found to be upregulated in OSCC [37]. Further, the expression of 14-3-3sigma, which also interacts with tuberin [36], was reduced or absent in OSCC [38]. This suggested that different isoforms of the 14-3-3 family behave differently in OSCC.

Clinicopathological characteristics of patients with LOH at TSC1 and TSC2 loci
We correlated the presence or absence of LOH at the TSC1 and TSC2 loci in 50 primary tumors to their clinicopathological variables such as age, sex, T classification, stage, grade, histology, tobacco habits and lymph node metastasis. Fisher's exact test (two-sided) was carried out and a p < 0.05 was considered to be significant. Using the above criteria, none of the parameters examined demonstrated a significant correlation with LOH at either of the TSC loci (data not shown).

Conclusion
Collectively, the detection of LOH in a proportion of OSCC samples coupled with reduced gene expression both at the RNA and protein levels indicates a loss of function of TSC genes, implicating their role as tumor suppressors in oral cancer for the first time. Loss of function of these genes may thus contribute to the constitutive activation of the mTOR signaling pathway leading to overall cell growth and proliferation. Our studies have also shown for the first time that several key members of this pathway show aberrant expression in oral cancer and can provide useful therapeutic targets. Several inhibitors of this pathway, such as rapamycin and its derivatives which inhibit mTOR (FRAP1) and the PI3K (PIK3C2A) inhibitor wortmannin, are in fact now being actively evaluated in clinical trials for other cancers [31,34]. Further, rapamycin and its derivative CCI-779 have been shown to reduce OSCC tumor size in nude mice [8,39]. Thus, these inhibitors could also be evaluated for the treatment of oral cancer.