Genetic alterations in Ki-ras and Haras genes in Juvenile Nasopharyngeal Angiofibromas and Head and Neck Cancer

Head and neck tumors are the sixth commonest tumors in the world 1 and account for approximately 10% of all malignant tumors. In Brazil, metropolitan areas such as São Paulo show one of the highest incidences of head and neck tumors in the world. 2 Only in India is the risk of oral cancer greater than in São Paulo. It has been estimated that in the state of São Paulo in 1990 there were 6800 new cases of tumors from oral cavity, pharynx and larynx. 3

ras, Ha-ras and N-ras) are structurally related and code for a protein (p21) known to play an important role in the regulation of normal signal transduction and cell growth.Activation of ras genes is due to point mutations within codons 12, 13, 59 and 61, which take part in the p21 active site.10   The frequency of ras mutations is different from one type of tumor to another, suggesting that point mutations might be carcinogenspecific. 11Mutations of the ras gene family are a common event in the development and progression of adenocarcinoma of the pancreas (90%), colon (50%), thyroid (50%), bladder (50%) and lung (30%) 10 .Regarding head and neck cancer, most of the authors have found ras mutations in less than 5% of the tumors from western world patients, [12][13][14][15] although Nunez et al. (1992) 16 and Anderson et al. (1994) 30 found ras mutations in 36,3% (8/22) and 22% (6/ 27) of the tumors, respectively.Similar frequencies (35%) of Ha-ras mutations have only been reported in oral squamous cell carcinomas from India. 17A high frequency of ras mutations has also been found in carcinogen-induced tumors in animal models. 18Whereas activating ras mutations appear to be an infrequent event in head and neck tumors from the western world, studies using immunohistological staining and RT-PCR (Reverse Transcriptase-Polymerase Chain Reaction) have demonstrated that overexpression is a common event in those tumors. 19,20However, the exact mechanism accounting for ras overexpression is unknown and its association with existing prognostic factors is not clear yet.
When it comes to JNA, little is known about the molecular basis of this disease and only recently Giardiello et  In the present study, we investigated a possible association between Ki-ras and Ha-ras mutations within codons 12, 13, 59 and 61 and the development of head and neck tumors and JNA, by using PCR-SSCP (Polymerase Chain Reaction-Single Strand Conformation Polymorphism) analysis.The relative levels of Haras gene mRNA transcripts were also examined in head and neck tumors by Northern blot analysis.

METHODS
Tissue samples.Tumor samples were obtained from 60 head and neck patients under surgery at Hospital A.C. Camargo and from 28 patients with Juvenile Nasophar yngeal Angiofibroma (JNA) submitted to surgery at Hospital das Clínicas, Faculty of Medicine, University of São Paulo, São Paulo, Brazil.Normal tissue was also obtained from 12 out of the 28 patients with JNA and from all the 60 patients with head and neck cancer.Tumors consisted of squamous cell carcinomas of the head and neck, 28 localized in the oral cavity, 10 in the oropharynx, 8 in the hypopharynx and 14 in the larynx.Ages of the head and neck tumors patients at the time of operation ranged from 27 to 80 years, median 58.The study included a total of 52 males and 8 females.All the JNA patients were males and the age at the time of operation ranged from 11 to 23 years.DNA samples from one colorectal tumor and one endometrial tumor that had already been analyzed for ras gene mutations were used as positive controls for the SSCP analysis.
DNA and RNA extraction.Tissue was ground to a powder using a Frozen Tissue Pulverizer (Termovac).For DNA extraction the powder was resuspended in 1ml of lysis buffer (10mM Tris-HCl, pH 7.6, 1mM EDTA (ethylenediaminetetracetic acid), 0.6% SDS (sodium dodecyl sulfate)) and 100 µg/ml proteinase K, and incubated at 37°C overnight.High molecular weight DNA was extracted with phenol-chloroform and precipitated with ethanol.For RNA extraction tissue powder was homogenized in a solution containing guanidine isothiocyanate (4M guanidine isothiocyanate, 25mM sodium citrate pH 7.0, 0.5% sarcosyl and 100mM β-mercaptoethanol) and extracted as described by Chomczynski and Sacchi (1987).22   Northern blot analysis.Ten micrograms of total RNA from tumors and normal samples were denatured with formaldehyde-formamide, separated by electrophoresis on a formaldehyde 1% agarose gel, and transferred to nylon filters.Northern blot filters were hybridized under stringent conditions with a 32 P-labeled Ha-ras 6.6Kb BamHI fragment (the probe was labeled using the random oligonucleotide priming technique) for 24 hours.Filters were washed twice at room temperature in 2x SSC (sodium chloride/sodium citrate)/0.1% SDS for 10 minutes and twice at 50°C for 30 minutes in 0.1x SSC/0.1% SDS and then exposed to Kodak X-Omat XAR film with an intensifying screen at -70°C for 2 or 5 days.
PCR-SSCP (Polymerase Chain Reaction Single-Strand Conformation Polymorphism) analysis.DNA sequences containing codons 12-13 and 59-61 of the Ha-ras and Ki-ras genes were amplified using oligonucleotide primers described by Ichikawa et al. (1994). 23PCR reactions were performed in 25µl volume using 50-100ng of genomic DNA template, 1µM of each primer, 1.5mM MgCl2, 200µM of each deoxynucleotide triphosphate, 0.1µCi of [α 32 P-dCTP] (Amersham, specific activity, 3000Ci/ mmol), 50mM KCl, 10mM Tris-HCl pH 8.0, and 0.5 unit of Taq DNA polymerase (Pharmacia, NJ, USA).The reactions were performed with an automated Thermal Cycler -Perkin Elmer 580 as follows: 35 cycles of denaturation for 1 minute at 94°C, annealing for 1 minute at 55°C and extension for 1 minute at 72°C.Amplified products (1µl) were diluted 10-fold in a buffer containing 95% formamide, 20mM EDTA, 0.05% bromophenol blue and 0.05% xylene cyanol, heated at 83°C for 10 minutes and applied (3µl/lane) on 6% polyacrylamide nondenaturing gels containing 2.5, 5.0 and 10% glycerol.Electrophoresis was performed at 13W for 4 -6 hours at room temperature with two cooling fans.Band shift mobility was detected by autoradiography of dried gels using Kodak X-Omat XAR film with an intensifying screen for 5 to 48 hours at -70°C.Direct DNA sequencing.DNA samples from 1 colorectal tumor with Ki-ras mutation detected by SSCP gels, one endometrial tumor with Ha-ras mutation, three JNA samples and three head and neck tumor samples were reamplified and direct DNA sequencing was performed.PCR products obtained were purified using Wizard PCR Preps Kit (Promega Corporation, Madison, USA) in accordance with the manufacturer's procedures.Purified DNA was submitted to a dideoxy chain termination reaction using a double strand DNA Cycle Sequencing Kit (Pharmacia, USA) for both sense and antisense primers.Sequencing reaction products were denatured and resolved on 6% denaturing urea/polyacrylamide gels.Gels were fixed for 15 minutes in a 10% methanol/10% acetic acid solution, dried and exposed to X-ray film overnight.
Statistical Methods.Analysis of statistical correlations between Ha-ras overexpression and the clinicopathological characteristics of the patients were performed by the χ 2 test and Fisher's Exact test for frequency data in contingency tables.

RESULTS
Tumor DNA from 60 patients with head and neck cancer and from 28 patients with Juvenile Nasopharyngeal Angiofibromas (JNA) were examined for the occurrence of point mutation in Ha-ras and Ki-ras genes using PCR-SSCP analysis.Representative autoradiographs from SSCP analysis are shown in Fig. 1.No mutations were found in both series of DNA samples analyzed for codons 12, 13, 59 and 61 of the Ha-ras and Ki-ras genes.
Some JNA and head and neck tumor samples were chosen at random and submitted to direct DNA sequencing in order to confirm the absence of mutation within the specific codons of Ha-ras and Ki-ras genes.One colorectal and one endometrial tumor sample were also submitted to direct DNA sequencing, as positive controls of Ki-ras and Ha-ras mutations, respectively.Direct sequencing of both strands of the four amplified PCR products did not reveal any point mutation in head and neck tumor or JNA samples analyzed.Representative results showing evidence of absence of mutation within codons 12-13 of Ki-ras and Ha-ras genes in JNA and head and neck tumor samples are shown in Figs. 2 and 3.The presence of point mutations within the positive control samples were confirmed on Ki-ras codon 13 (colorectal tumor; GGC to AGC; Gly to Ser) and on Ha-ras codon 12 (endometrial tumor; GGC to GTC; Gly to Val).
In 32 of the head and neck cases where total RNA from normal and tumor samples were available, the relative level of Ha-ras mRNA expression was examined by Northern blot analysis.Representative results of the Northern blot analysis are shown in Fig. 4. Using densitometric scans, Ha-ras mRNA transcripts were found to be overexpressed in 17/32 (53%) of tumor samples relative to the normal sample derived from the same patient.The relative level of Ha-ras overexpression in these tumors ranged from 2-to 15-fold.A probe for 18S ribosomal RNA was used to correct the differences between normal and tumor RNA loading.
To evaluate the contributions of Ha-ras overexpression to the development and/or progression of the head and neck tumors analyzed, clinicopathological characteristics of the cases with Ha-ras overexpression were compared with the characteristics of those cases that showed normal Ha-ras expression.No correlation was observed between Ha-ras overexpression and clinicopathological characteristics of the patients, such as age, histology, site of the tumor, TNM stage or lymph node status (Table 1).Ras overexpression was observed in tumors of patients in all clinical stage but with a trend to be more frequent in stage III tumors.In the present study we found a significant association between ras overexpression and clinical outcome.Recurrence or death due to the disease occurred more frequently in patients with tumors showing high levels of Ha-ras gene (8/17) than in patients with tumors without Ha-ras overexpression (1/15) (p = 0.01).

DISCUSSION
Our results show no evidence of mutations within codons 12, 13, 59 and 61 of Ki-ras and Ha-ras either in Juvenile Nasopharyngeal Angiofibroma (JNA) or in head and neck tumors from a group of Brazilian patients.No data is available about Ki-ras and Ha-ras mutations related to the development of JNA but when it comes to head and neck tumors, our data are in accordance with other studies that show low rates of ras mutations (less than 5%) in head and neck tumors from western populations.) and 22% (6/27) of the samples analyzed, respectively, although these discrepancies might be due to geographical differences.Interestingly, this frequency of oral squamous cell carcinomas harboring activating ras mutations is similar to that found in India 17 and Taiwan, 25 where betel quid chewing is thought to be the initiating agent. 17,19A high frequency of ras mutations has also been found in carcinogen- induced tumors in animal models. 18 Although ras mutations do not appear to play a major role in head and neck tumors from Caucasian patients, several studies including ours have revealed that ras overexpression is a frequent event in these tumors. 15,19,20The implication of this finding is still unclear.The same situation is observed in breast cancer in which ras gene mutations are rare 26 but overexpression has been reported in about 60% of the tumors analyzed. 27,28In both types of tumors, ras gene amplification is a rare event and the overexpression observed may be due to another activation mechanism of gene expression. 19,27,28 In the present study, no correlations were observed between Ha-ras overexpression and clinicopathological characteristics of the patients.However, Ha-ras overexpression was associated with poor prognosis, in accordance with the results reported by Azuma et al. (1987). 31On the other hand, using a different technical approach, Field et al. (1992) 19 and Kiaris et al. (1995) 20 also reported that overexpression is a frequent event in squamous cell carcinomas of the head and neck, but found an association with a favorable prognosis.These differences regarding the prognostic value of ras overexpression in head and neck cancer may be mainly due to the use of small series of patients with heterogeneous composition.Further studies examining larger series of patients are required to clarify whether there is an association between Haras overexpression and clinical outcome.

Figure 1 -
Figure 1 -Representative autoradiographs from PCR-SSCP analysis in head and neck tumors (HN) and Juvenile Nasopharyngeal Angiofibromas (JNA) for detection of Ki-ras (1A) and Ha-ras (1B) gene mutation within codons 12 and 13.N: DNA from normal tissue; T: DNA from tumoral tissue; C: colorectal tumor used as positive control for Ki-ras mutation (the arrow indicates the mobility shift).

Figure 2 -
Figure 2 -Representative autoradiographs of direct DNA sequence of 12 and 13 codons of Ki-ras gene in Juvenile Nasopharyngeal Angiofibromas (JNA).One mutation in codon 13 was detected in the colorectal tumor sample used as a positive control for PCR-SSCP analysis.The arrow indicates the alteration observed (GGC to AGC; Gly to Ser).N: DNA from normal tissue; T: DNA from tumoral tissue.

Figure 3 -
Figure 3 -Representative autoradiographs from PCR-SSCP analysis (A) and direct DNA sequence (B) of codons 12 and 13 of the Ha-ras gene in head and neck tumors (HN).One mutation in codon 12 was detected in the endometrial tumor sample used as a positive control for PCR-SSCP analysis.The arrow indicates the alteration observed (GGC to GTC; Gly to Val).N: DNA from normal tissue; T: DNA from tumoral tissue; C: positive control.

16 and
Yarbrough et al. (1994) 15 analyzed 51 samples from head and neck squamous cell carcinomas and found no evidence of Ki-ras, Ha-ras and N-ras mutations within codons 12, 13 and 61.Irish and Bernstein (1993) 24 obtained similar results, not finding Ki-ras mutations within codons 12 and 13.In addition, Kiaris et al. (1995) 20 found ras gene mutations in only 1.7% of the samples analyzed, when studying a larger panel of squamous cell carcinomas of the head and neck.On the other hand, Nunez et al. (1992) Anderson et al. (1994) 30 found ras mutations in 36.3% (8/22

Figure 4 -
Figure 4 -Representative autoradiographs from Northern blot analysis of Ha-ras transcripts in the tissue samples of five patients with head and neck carcinomas (HN).N: DNA from normal tissue; T: DNA from tumoral tissue.

Table 1 -Associations of Ha-ras overexpression and clinicopathological characteristics in patients with head and neck squamous cell carcinomas
1Negative, patients without Ha-ras overexpression; Positive, patients with tumors with Ha-ras overexpression.
30asaka et al. (1993)29proposed that Ha-ras overexpression in HPV transfected cell lines might be due to loss of tumor suppressor gene function or direct integration of HPV DNA sequences in close proximity to cellular oncogenes.In addition, Anderson et al. (1994)30found Ha-ras overexpression associated with HPV infection in 11% of oral squamous cell carcinomas suggesting that viral infection might be associated with ras overexpression.