Follicular thyroid tumors are often aneuploid. It was advanced that chromosomal instability is closely associated to RAS mutations, but such association remains unproven. H-RAS can be alternatively spliced in two different proteins, p21 and p19, the former being the active protein. In order to investigate the relationship between RAS mutational status and ploidy in thyroid tumors, we analysed RAS genes in a series of 99 follicular lesions (14 nodular goiters, 70 follicular adenomas and 15 follicular carcinomas), eight thyroid carcinoma cell lines and a control group of 102 blood donors, correlating the presence of RAS mutations with the ploidy of the tumors and evaluating the two spliced forms of H-RAS. Overall, 20% of the follicular tumors harbored RAS mutations and 62% of the patients with follicular tumors (and 51% of blood donors) harbored the H-RAS 81T → C polymorphism. The presence of RAS mutations was not associated with aneuploidy. The H-RAS polymorphism did not seem to confer a higher propensity for neoplastic transformation as it was also found in hyperplastic lesions, but was strongly associated with aneuploidy (P<0.0001). The presence of the H-RAS 81T → C polymorphism was associated with significantly higher amounts of total H-RAS mRNA expression, higher amounts of p21 isoform and a higher fraction of neoplastic cells in S phase. Our results suggest that the H-RAS 81T → C polymorphism may induce aneuploidy through overexpression of the active p21 isoform of H-RAS.
The three mammalian RAS genes, H-RAS, K-RAS and N-RAS, encode small GTP-binding proteins, termed p21s, that function as switches in the cytoplasmatic relay of external growth and differentiation signals by cycling between ‘on’ (GTP-bound) and ‘off’ (GDP-bound) configurations. Oncogenic point mutations in the three RAS genes, namely the hotspot mutations in codons 12, 13 and 61, are frequently observed in many tumor types (reviewed by Bos, 1989). These alterations increase the proportion of GTP-bound p21 independently of external signals (Noonan et al., 1991; Liwo et al., 1994; Monaco et al., 1995).
It is accepted that mutations in RAS genes are an early event in thyroid tumorigenesis, being even detected in some hyperplastic lesions (goiters), and most likely occurring before malignant transformation (Esapa et al., 1999; Nikiforova et al., 2003; Soares et al., 2003). Benign and malignant follicular tumors of the thyroid and, less frequently, papillary carcinomas, were shown to harbor RAS activating mutations, more often N-RAS, but also H- and K-RAS (Hara et al., 1994; Esapa et al., 1999; Sugg et al., 1999; Kimura et al., 2003; Nikiforova et al., 2003; Soares et al., 2003).
Follicular tumors of the thyroid are often aneuploid (Johannessen et al., 1982; Hostetter et al., 1988; Joensuu and Klemi, 1988; Schelfhout et al., 1990; Cusick et al., 1991; Castro et al., 2001). Although the presence of aneuploidy is not by itself an indicator of malignancy (Johannessen et al., 1982; Joensuu and Klemi, 1988; Hostetter et al., 1988; Schelfhout et al., 1990; Cusick et al., 1991; Castro et al., 2001), aneuploid carcinomas tend to carry a worse prognosis than diploid (or near-diploid) carcinomas of the thyroid (Joensuu and Klemi, 1988; Schelfhout et al., 1990).
The relationship between the presence of RAS mutations and aneuploidy was first reported in colorectal adenomas by Giaretti et al. (1995). These authors suggested a functional role of K-RAS activation (by mutation) in the control of chromosomal stability (Giaretti et al., 1995, 1998, 2004). Other authors have shown that fibroblasts transfected with mutant H-RAS displayed genomic instability (Denko et al., 1994, 1995). The same result was found in a thyroid cell line transfected with oncogenic H-RAS (Saavedra et al., 2000). With regard to thyroid tumors, it was also advanced a putative relationship between RAS activation (mutation) and chromosomal instability (Fagin, 2002) although this relationship has not been demonstrated.
In 1989, Cohen et al. (1989) described the regulatory pathway of H-RAS by alternative splicing. The D intron of H-RAS contains an additional exon, named IDX (Figure 1). The H-RAS mRNA can be processed in, at least, two ways according to the presence or absence of the IDX exon (Figure 1). The presence of IDX exon would result in the introduction of a premature stopcodon and generate a smaller protein, p19. Although it was firstly postulated that p19 would be unstable owing to its premature stop codon (Cohen et al., 1989), more recent work showed that p19 mRNA is as stable as p21 mRNA, both being present in similar amounts in mammalian cells (Guil et al., 2003a). In contrast to the p21 protein, the alternative form of H-RAS (the p19 protein) is localized in the cytoplasm and nucleus rather than at the plasma membrane (Guil et al., 2003a), does not have transforming properties in vitro (Cohen et al., 1993) and does not interact with most p21 effectors. It was also advanced that the abolishment of H-RAS splicing contributed to oncogenesis by increasing the production of the p21 isoform (Huang and Cohen, 1997).
In order to investigate the relationship between RAS mutational status and ploidy in thyroid tumors, we searched for RAS mutations in a series of 99 follicular lesions (14 nodular goiters, 70 follicular adenomas and 15 follicular carcinomas) and analysed the two splicing forms of H-RAS (p21 and p19) in a subset of the same series of tumors.
Ploidy and RAS mutations
The ploidy of the thyroid lesions is shown in Table 1. The DNA index (DI) of follicular carcinomas ranged from 0.97 to 1.50, and eight of the tumors (53.3%) were aneuploid. The DI of fetal adenomas (FA) ranged from 0.95 to 2.0, and 35 of the tumors (76.1%) were aneuploid. The DI of follicular adenomas other than FA ranged from 0.97 to 1.09, and five of the tumors (20.8%) were aneuploid. The DI of the 14 goiters ranged from 1.0 to 1.8, and 50% (7/14) of the lesions were aneuploid.
The frequency of RAS mutations in the 95 cases is shown in Table 1. Twenty percent of follicular carcinomas harbored RAS mutations (all in N-RAS codon 61); 11.6% of FA harbored RAS mutations (three in N-RAS codon 61 and two in H-RAS codon 61) and 26.1% of FA harbored RAS mutations (four in N-RAS codon 61, one in H-RAS codon 61 and one in K-RAS codon 61). None of the 14 goiters harbored RAS mutations.
There was no significant association between the ploidy status of the tumors and the presence of RAS mutations (P=0.11). Of the 14 tumors harboring RAS mutations, six were aneuploid and eight were diploid; the same holds true when the comparison was restricted to each group (adenomas and carcinomas) per se (data not shown).
There was also no significant association between the presence of RAS mutations and any of the clinicopathologic parameters analysed in this study (gender, age and size of the tumors) (data not shown).
Ploidy, S-phase fraction and H-RAS 81T → C
We detected a high prevalence of H-RAS 81T → C polymorphism: 53 of the 85 (62.3%) cases harbored the polymorphism (Table 2 and Figure 2). In the thyroid tissue of 20 patients from which normal tissue adjacent to the tumors was available, the H-RAS 81T → C status in the tumors did not differ from that in the respective normal tissue. There was no significant association between the presence of the polymorphism and the histotype of the tumors (Table 2).
The prevalence of the H-RAS 81T → C polymorphism was similar in a control group of 102 blood donors and in the cases: 51.0% in controls and 62.4% in cases (P=0.12) (Table 2). There was no association between the presence of the polymorphism and the mutational status of the RAS genes in the tumors (data not shown).
We found a significant association between the presence of the polymorphism and the ploidy of the tumors. From the 48 patients with aneuploid tumors, only seven did not have the H-RAS 81T → C polymorphism (P<0.0001) (Table 3). A significant association was also found when the comparison was made regarding the allelotype distribution (P<0.0001) (Table 3). The significant association between the presence of the polymorphism and aneuploidy is kept when the hyperplastic lesions (14 nodular goiters) were added to the series (P<0.0001). Six of the seven aneuploid goiters and two of the seven diploid goiters harbored the H-RAS 81T → C polymorphism.
In order to verify if the aforementioned associations were established with the H-RAS 81 polymorphism or with other gene/loci close to H-RAS 81, we studied two microsatellite markers, one in H-RAS (microsatellite 1) and the other outside the gene (microsatellite 2) (Figure 1). No significant differences were found between the control group and the patients regarding the frequency of the alleles of each marker (data not shown).
We confirmed that the two markers analysed were in complete linkage disequilibrium with the H-RAS 81T → C polymorphism as it was expected by their proximity (Figure 1).
The results obtained in the evaluation of the microsatellite markers and the association with the ploidy status confirmed the strong association of H-RAS 81T → C with patients with aneuploid tumors. The ‘Exact Test of Sample Differentiation’, to discriminate patients with diploid tumors from patients with aneuploid tumors, only gave strong statistically significant results when the polymorphism H-RAS 81T → C was considered (P<0.0001±0.00). When the other two microsatellite markers were considered, the associations were poorer (although still statistically significant), suggesting that the association with the ploidy status is weaker (M1, 0.04±0.004; M2, P=0.03±0.004) than with the H-RAS 81T → C polymorphism. Although the two markers were in linkage disequilibrium, none of them could discriminate the patients with diploid tumors from those with aneuploid tumors so well as the H-RAS 81T → C polymorphism alone. In accordance with this, the haplotype distribution in patients with diploid and aneuploid tumors showed a significantly higher prevalence of haplotypes associated with the C allele in patients with aneuploid tumors than in patients with diploid tumors (data not shown).
The H-RAS 81T → C polymorphism was also significantly associated with the S-phase fraction of the tumors: tumors from patients with the polymorphism had a significantly higher (P=0.037) S-phase fraction (4.01±3.92) than those from patients without the polymorphism (2.20±2.70). We found a similar association between the S-phase fraction and the allelotypes of the patients: patients with TT allelotype showed a lower S-phase fraction (2.10±2.70) than patients with CT allelotype (4.01±2.70) and patients with CC allelotype (4.81±4.90) (TT vs CT P=0.047; TT vs CC P=0.043; and CC vs CT P=0.51).
Relation between the presence of H-RAS 81T → C polymorphism and the expression of two spliced variants of H-RAS
The expression of H-RAS total mRNA and that of the two isoforms (p19 and p21) mRNA was assessed by semiquantitative reverse transcription–PCR (RT–PCR).
The expression of total mRNA of H-RAS was significantly higher (P=0.020) in tumors having the H-RAS 81T → C polymorphism than in tumors without the polymorphism (Table 4 and Figure 3). The difference in the level of expression seemed to be higher in tumors of patients homozygous to the polymorphism than in those who are heterozygous, although the sample is too small to allow a meaningful comparison (TT vs CT P=0.14; TT vs CC P=0.051; CC vs CT P=0.55) (Table 4).
The relative level of expression of the p21 isoform was significantly higher (P=0.005) in tumors from patients with the polymorphism than in the other patients (Table 4). Tumors from patients with H-RAS 81T → C polymorphism in homozygosity had the highest level of expression of p21 (Table 4) (TT vs CT P=0.058; TT vs CC P=0.028; CC vs CT P=0.41).
The analysis of eight thyroid cell lines showed that the total H-RAS mRNA expression was higher in the six cell lines harboring the polymorphism (TPC1, K1, FB2, E1, 8505C and HTH4) than in the two cell lines without the polymorphism (BCPAP and XTC1), 91.5±1.7 (mean±s.d.) vs 57.0±1.8, respectively, but the size of the samples is too small to reach the threshold of statistical significance (P=0.06). The relative levels of p19/p21 mRNA expression were significantly higher (P=0.039) in the six cell lines with the polymorphism than in the two without the polymorphism: 85.6±14.1 vs 62.5±24.7, respectively.
We analysed the adjacent thyroid tissue in four cases (three cases of tumors harboring the H-RAS 81T → C and one without the polymorphism) and, although the size of the sample is too small to allow a meaningful comparison, we observed that the relative mRNA expression of the p19/p21 isoforms was similar in each tumor and adjacent thyroid, but the total amount of H-RAS mRNA expression was higher in all the tumors than in the adjacent thyroid tissues.
Thyroid tumors are particularly interesting regarding DNA content for two main reasons: first, in contrast to papillary carcinomas, which are usually diploid, follicular adenomas and follicular carcinomas are frequently aneuploid, and second, aneuploidy is not an indicator of malignancy in follicular lesions, but it is associated with a poorer prognosis if the comparison is restricted to carcinomas (Joensuu and Klemi, 1988; Schelfhout et al., 1990).
The mechanisms underlying aneuploidy in follicular thyroid tumors are unknown. In the present study, we have demonstrated, for the first time, an association between the presence of a polymorphism in H-RAS and the occurrence of aneuploidy in such tumors.
Aneuploidy in thyroid adenomas is associated with a fetal-like morphotype (Castro et al., 2001). We confirmed this observation in the present series: FA are more frequently aneuploid than follicular adenomas other than fetal adenomas. Moreover, the latter, whenever aneuploid, are always near-diploid or hyper-diploid. The association between aneuploidy and the so-called fetal growth pattern of thyroid tumors remains unexplained (Castro et al., 2005). We confirmed the dissociation between aneuploidy and malignancy in thyroid as we found 50% of aneuploidy among nodular goiters and 57% of aneuploidy among adenomas.
The frequency of RAS mutations detected in the present study is in accordance with those previously reported in thyroid tumors (Esapa et al., 1999; Nikiforova et al., 2003; Soares et al., 2003). At variance with the findings of Giaretti et al., 1995, 1998, 2004) in colorectal carcinomas, the presence of RAS mutations in thyroid tumors was not significantly associated with the ploidy of the tumors.
We detected the H-RAS 81T → C polymorphism in many patients (62.4%) as well as in healthy controls (51.0%). This observation indicates that, differently from the findings in bladder cancer (Johne et al., 2003), the presence of the C allele does not seem to confer a higher propensity for neoplastic transformation in the thyroid.
Despite the lack of a significant association between the H-RAS 81T → C polymorphism and the occurrence of thyroid tumors, we found a strong association of the polymorphism with the ploidy status of the tumors (P<0.0001) suggesting a possible role of H-RAS in their aneuploidization. The same holds true for the aneuploidization of nodular goiters, which was also associated in our series to the H-RAS 81T → C polymorphism.
The data we have obtained in the study of the two microsatellites show that, although both are in complete linkage disequilibrium with the H-RAS 81T → C, none of them is able to discriminate patients with diploid tumors from patients with aneuploid tumors better than the polymorphism per se. This observation suggests that the association with ploidy status is with the 81T → C polymorphism. However, we cannot exclude the hypothesis that the association with ploidy may be related with other unknown alterations within H-RAS gene, or even with alterations at a distance from H-RAS gene if one considers the H-RAS 81T → C a regulatory polymorphism (reviewed in Pastinen and Hudson, 2004). As the 81T → C polymorphism does not alter the RAS protein sequence, it is possible that this SNP is linked to other polymorphic sites in functional intronic regions of H-RAS (such as a region of variable tandem repeats, about 1 kb downstream exon 4), with possible transcriptional enhancer activity (Krontiris et al., 1993). Moreover, it was recently shown that there is an intron silencer in intron D2 that acts as a negative regulator of IDX inclusion (Guil et al., 2003b), leading to the possibility that the H-RAS 81 may serve as a marker of other polymorphisms in H-RAS, namely in intron D2, that would act as regulators of IDX inclusion. This possibility would explain the different p21/p19 ratios observed in our study.
We have also analysed eight thyroid carcinoma-derived cell lines: two of them were negative for the polymorphism and six harbored the polymorphism. The ratio of p21/p19 mRNA in the cell lines with and without the polymorphism was statistically different (P=0.039). The differences we observed in the cell lines fit with those found in the tumors despite the small size of the sample (specially owing to the very small number of cell lines negative for the polymorphism).
The link between the H-RAS 81T → C and aneuploidy may reside on the fact that the allele 81 C is associated with significantly higher amounts of total H-RAS mRNA and/or significantly higher relative amounts of p21/p19. These findings support the hypothesis that H-RAS 81T → C could induce aneuploidy through the overexpression of p21 isoform of H-RAS. It remains to be fully clarified if activating mutations of RAS genes may lead to a similar effect in thyroid tumors (see below). Further studies are needed, using quantitative methods, to clarify the relationship between ploidy and H-RAS (over)expression and the relative amounts of p19/p21.
It is known that RAS proto-oncogenes acquire the property of transforming cells in culture by single point mutations (Tabin et al., 1982; Wang et al., 1997). Besides this, the biological consequences of quantitative changes in the expression of normal p21 RAS protein have also been demonstrated (Pulciani et al., 1985). A recent study in AML patients showed that only a minority of RAS mutated AML samples (22.2%) had strong RAS activity, whereas 38.5% of the patients presented strong RAS activity in the absence of RAS mutations. The clinical outcome of these patients was correlated with RAS activity and not with RAS mutations (Illmer et al., 2005). Other studies showed that the RAS effects could be due to both mutations and overexpression (Mascaux et al., 2005); however, in such studies neither the presence of the polymorphism nor the expression level of the p21/p19 isoforms was evaluated (Illmer et al., 2005; Mascaux et al., 2005).
It has been shown that N-RAS mutations are more frequent in thyroid tumors than those of H-RAS (as we have confirmed in the present study). This observation does not invalidate the aforementioned hypothesis, as different mechanisms can play different roles in the neoplastic transformation. As Mammas et al. 2004 have shown, H-RAS overexpression (and not K-RAS overexpression) is involved in the pathogenesis of cervical cancer, although the mutations detected in this type of tumor are in the K-RAS gene. Other possibility would be to consider that N-RAS mutations are important in the initiation of thyroid tumors, whereas activated H-RAS (p21) would be determinant in their aneuploidization. The discussion of these issues rests beyond the scope of the present study.
This is not the first report showing that polymorphisms of H-RAS increase the amount of p21. Cohen et al. (1989) had already shown that an intronic mutation of H-RAS increased p21 production by preventing the recognition of IDX during pre-mRNA splicing. A similar mechanism could explain the different amounts of the two isoforms detected in our series, although we did not detect this particular polymorphism in any case of a limited series analysed (data not shown).
It is generally accepted that RAS GTPases signal through many pathways that influence cell cycle (Feramisco et al., 1984; Mulcahy et al., 1985), and that overexpression of 21, not only via oncogenic mutations, may have a major effect on the progression through the cell cycle, namely on the G1 checkpoint (reviewed by Coleman et al., 2004). The higher percentage of H-RAS total mRNA, as well as overexpression of the p21 isoform we observed in tumors of patients with the H-RAS 81 C, may lead to a more rapid cell-cycle and, secondarily, to aneuploidy. The level of activity of H-RAS seems to be related to the levels of p21, as p19 does not have the ability to bind to GTP and thus is unable to activate the downstream effectors of H-RAS (Guil et al., 2003a). The association of H-RAS 81T → C polymorphism with increased amount of the p21 isoform, higher S-phase fraction (proliferation) and presence of aneuploidy remains, however, to be fully understood.
At variance with the results obtained in our study, it has been reported that individuals harboring the homozygous 81 C genotype of the H-RAS proto-oncogene appear to be at an increased risk for bladder cancer (Johne et al., 2003). Curiously, no significant association between the polymorphism and the presence of aneuploidy, which is also an important prognostic factor in bladder cancer, has been observed by Johne et al. (2003). It would be interesting to investigate, in other tumor models, the putative oncogenic role played by the H-RAS alternative splicing and/or the aneuploidization described in the present study.
Materials and methods
Ninety-nine tumors were either consecutively collected or retrieved from the files of frozen material of the Department of Pathology of Hospital S João. All the local ethical guidelines were strictly followed in this study. The lesions were classified according to Rosai et al. (1993) and DeLellis et al. (2004) as fetal adenoma (n=46), follicular adenoma (other than fetal adenoma) (n=24), follicular carcinoma (n=15) and nodular goiter (n=14) in sections stained with hematoxylin and eosin. The following criteria were used in the subclassification of benign lesions:
Fetal adenoma: Encapsulated follicular adenomas were classified as fetal adenoma whenever more than 50% of the lesion was composed of microfollicles, usually dispersed in an abundant, edematous or hyalinized stroma. In many cases, the microfollicular growth merged with foci of trabecular/solid adenoma; for the sake of simplicity, all these cases were classified as fetal adenoma regardless of the size of the trabecular/solid areas.
Follicular adenoma (other than fetal adenoma): All the remaining encapsulated benign tumors were included in this category. Most of the tumors displayed a predominant or exclusive normofollicular pattern of growth.
Nodular goiter: The diagnosis of nodular goiter was made whenever there were several nodules or a single nodule without a well-defined capsule. In most cases, the nodules displayed normo- or slightly macrofollicular pattern of growth with or without papillary hyperplasia.
Both in adenomas and carcinomas, no classificative importance was attributed to cell oxyphilia or lymphoid infiltration.
The adjacent normal thyroid from 24 cases and eight thyroid cell lines – TPC1, BCPAP, K1, FB2, XTC1, E1, 850, HTH4 – were also studied.
Flow cytometry was performed in fresh or frozen samples of the same lesion used for histological diagnosis. The samples were processed according to a previously reported procedure (Deitch et al., 1982), but slightly modified by us. Briefly, samples were thawed and complete cell lysis was assured with NP-40 treatment (final concentration 0.5%) and the isolated nuclei were washed once with ice-cold phosphate-buffered saline. Nuclei of chicken erythrocytes (NCE) were added as an internal standard at a final concentration of 5–10% of total nuclei. Nuclei (1 × 106) were stained with a propidium iodide (PI) solution containing 50 μl/ml of PI, 10 mM of Tris, 5 mM of MgCl2 and 0.5 mg/ml RNase (DNase free), pH 7.0. Tubes were vortexed and allowed to incubate protected from light for at least 30 min. After incubation, the suspension was syringed through a 27 needle and sieved through a 55 nylon mesh immediately before analysis. Acquisition was performed on a EPICS C flow cytometer (Coulter Electronics Inc., Hialeah, FL, USA). Computer analysis of DNA histograms was performed using MPLUS (Phoenix Flow Systems, San Diego, CA, USA). This software has included MCYCLE AV, a multiple option cell cycle fitting that automatically determines the DI and cell cycle phase fractions in cell or nuclei populations. The same batch of NCE was used with all the thyroid samples. The DI of the standard peak (DISP) was calculated by MCYCLE AV, as the ratio of the mean channel number of the standard peak and the mean channel number of a peak labeled as ‘diploid’. The coefficient of variation of DISP measurement intra-sample was less than 1.0% (in most cases, less than 0.5%). The mean coefficient of DISP inter-sample was 2.5%. The median value was 0.341%.
RAS mutational status
Genomic DNA was extracted from 10-μm paraffin-embedded sections of the 85 tumors and of 14 nodular goiters. Slides were microscopically examined and tumor areas were marked and carefully dissected under microscopic observation. Dissected material was deparaffinized in xylene, washed in ethanol and rehydrated. DNA extraction was performed using the Genomic DNA Purification Kit (Gentra System, Lisboa) according to the manufacturer's tissue protocol. Briefly, microdissected material was resuspended in 300 μl of cell lysis buffer with 30 μg of proteinase K and incubated at 56°C during 3 days, with daily addition of 30 μg of proteinase K. After proteinase K inactivation and protein removal, genomic DNA was precipitated with isopropanol and glycogen, and washed with ethanol. DNA was then dried and rehydrated with hydratation buffer.
Sequences of H-RAS (exons 1 and 2), K-RAS (exons 1 and 2), N-RAS (exon 2) and both microsatellites (M1 and M2) were amplified (using primer pairs listed in Supplementary Information). The PCR mixture (25 μl) contained 2.5 μl of 10 × Complete PCR buffer (Bioron, Porto, Portugal), 1 μl of dNTPs (5 mM each), 0.1 μg of each primer, 0.1–0.5 μg of genomic DNA and 0.2 U of Taq DNA Polymerase (Bioron). After 10 min of initial denaturation, the PCR mixtures were subjected to 35 cycles of denaturation for 30 s at 95°C, annealing for 45 s at variable temperature according to the amplicon (see Supplementary Information) and extention for 45 s at 72°C. A final extension period of 10 min at 72°C was performed to complete the reaction.
PCR products of the RAS genes were subjected to an SSCP electrophoresis in 0.8% MDE gel in order to analyse possible mutations. As SSCP analysis was not able to discriminate alterations in exon 2 from N-RAS, this gene was studied by direct sequencing only. Running conditions are discriminated in the table in Supplementary Information. Gels were silver stained and samples with suspicious bands were purified by enzymatic treatment. Purified samples were sequenced on both strands as follows: 3 μl of purified PCR product was added to 7 μl of sequencing reaction containing 2.7 μl of 2.5 × Big Dye Buffer (Applied Biosystems, Porto, Portugal), 30 ng of primer and 1 μl of Big Dye terminator v3.1 cycle (Applied Biosystems). Sequencing was performed in an ABI Prism® 3100 Genetic Analyzer (Applied Biosystem).
Mutated samples were re-extracted, re-amplified and re-sequenced to confirm the result. For the microsatellite detection, the fragment size determination of the amplified PCR products was carried out using an ABI Prism 310 Genetic Analyzer instrument (AB Applied Biosystems). After the PCR amplification, 1 μl of the PCR product was combined with the internal lane standard GS500 labeled with the dye TAMRA and diluted in formamide (14.4 μl deionized formamide and 0.6 μl TAMRA 500 internal size standard per sample). Samples were analysed by electrophoresis using Performance Optimized Polymer 4 (POP-4; AB Applied Biosystems) and filter set C, using the following conditions: Module ‘GS STR POP4 (1 ml) C’; Inj. seconds: 5; Inj kV: 15.0; run kV: 15.0; run °C: 60; run time: 24 minutes. Data were analysed using a matrix (Set C) generated using 6-TET, HEX and TAMRA matrix standards.
Fragment sizes were determined automatically using the GeneScan®Analysis Software v.3.1 (AB, Applied Biosystems). We also studied 20 samples of normal adjacent parenchyma (of the study tumors) and 102 controls from healthy blood donors for the H-RAS polymorphism and microsatellite markers.
The expression of H-RAS total mRNA and of the two isoforms (p19 and p21) mRNA was assessed in frozen sections by semiquantitative RT–PCR in 20 tumors, four samples of ‘normal thyroid tissue’ and in eight thyroid carcinoma-derived cell lines.
For each sample, 1.0 μg of RNA was reverse transcribed in a reaction volume of 20 μl in the presence of 4 mM dNTP, 1.0 U/μl RNase inhibitor, 2.5 μ M random primer P (dN)6 and 10 U/μl M-MLV reverse transcriptase. Reverse-transcribed cDNAs (0.25 μg) from tumor tissues were co-amplified with a set of primers for the housekeeping gene GAPDH (Supplementary Information) and a set of primers for H-RAS that allow us to differentite the two different isoforms (p19 and p21) (Supplementary Information). All the quantitations were performed in triplicate. The PCR products were separated in an agarose gel (2%), and stained with ethidium bromide. The intensity of the fluorescence was automatically measured and integrated using the Genescan Software ‘Image Master’ (Pharmacia, Carnaxide, Portugal).
Values of ploidy status and RAS mutations are expressed in absolute numbers or percentages. The fractions of total H-RAS mRNA and p21 mRNA are expressed in mean (%)±standard deviation (s.d.). The χ2 test with Yate's correction and the (unpaired) t-test were used in the statistical analysis of the results. Two values were considered significantly different when P<0.05.
Allele/haplotype frequency distributions were compared in different groups of thyroid tumors (follicular carcinomas vs FA vs follicular adenomas; diploid vs aneuploid tumors; patients with follicular tumors vs control group) by means of an exact test of sample differentiation.
Hardy–Weinberg equilibrium test and exact test of differentiation between pairs of samples were carried out using the ARLEQUIN software 2.000 (Schneider et al., 2000). In all cases, exact test was based on more than 100 000 shufflings, for standard error <0.001. Significant differences were considered positive for non-differentiation whenever P-values were lower than 0.05. ARLEQUIN software was also used to measure the degree of association between alleles in the form of likelihood ratio test of linkage disequilibrium.
Bos JL . (1989). Cancer Res 49: 4682–4689.
Castro P, Eknaes M, Teixeira MR, Danielsen HE, Soares P, Lothe RA et al. (2005). J Pathol 206: 305–311.
Castro P, Sansonetty F, Soares P, Dias A, Sobrinho-Simoes M . (2001). Virchows Arch 438: 336–342.
Codony C, Guil S, Caudevilla C, Serra D, Asins G, Graessmann A et al. (2001). Oncogene 20: 3683–3694.
Cohen JB, Broz SD, Levinson AD . (1989). Cell 58: 461–472.
Cohen JB, Broz SD, Levinson AD . (1993). Mol Cell Biol 13: 2666–2676.
Coleman ML, Marshall CJ, Olson MF . (2004). Nat Rev Mol Cell Biol 5: 355–366.
Cusick EL, Ewen SW, Krukowski ZH, Matheson NA . (1991). Br J Surg 78: 94–96.
Deitch AD, Law H, White RD . (1982). J Histochem Cytochem 30: 967–972.
DeLellis RA, Lloyd RV, Heitz PU, Eng C (eds) (2004). Pathology and Genetics of Tumours of Endocrine Organs. IARC Press: USA.
Denko N, Stringer J, Wani M, Stambrook P . (1995). Somat Cell Mol Genet 21: 241–253.
Denko NC, Giaccia AJ, Stringer JR, Stambrook PJ . (1994). Proc Natl Acad Sci USA 91: 5124–5128.
Esapa CT, Johnson SJ, Kendall-Taylor P, Lennard TW, Harris PE . (1999). Clin Endocrinol (Oxford) 50: 529–535.
Fagin JA . (2002). Mol Endocrinol 16: 903–911.
Feramisco JR, Gross M, Kamata T, Rosenberg M, Sweet RW . (1984). Cell 38: 109–117.
Giaretti W, Molinu S, Ceccarelli J, Prevosto C . (2004). Cell Oncol 26: 301–305.
Giaretti W, Pujic N, Rapallo A, Nigro S, Di Vinci A, Geido E et al. (1995). Gastroenterology 108: 1040–1047.
Giaretti W, Rapallo A, Geido E, Sciutto A, Merlo F, Risio M et al. (1998). Am J Pathol 153: 1201–1209.
Guil S, de La Iglesia N, Fernandez-Larrea J, Cifuentes D, Ferrer JC, Guinovart JJ et al. (2003a). Cancer Res 63: 5178–5187.
Guil S, Gattoni R, Carrascal M, Abian J, Stevenin J, Bach-Elias M . (2003b). Mol Cell Biol 23: 2927–2941.
Hara H, Fulton N, Yashiro T, Ito K, DeGroot LJ, Kaplan EL . (1994). Surgery 116: 1010–1016.
Hostetter AL, Hrafnkelsson J, Wingren SOW, Enestrom S, Nordenskjöld B . (1988). Am J Clin Pathol 89: 760–763.
Huang MY, Cohen JB . (1997). Oncol Res 9: 611–621.
Illmer T, Thiede C, Fredersdorf A, Stadler S, Neubauer A, Ehninger G et al. (2005). Clin Cancer Res 11: 3217–3224.
Joensuu H, Klemi PJ . (1988). Am J Clin Pathol 89: 35–40.
Johannessen JV, Sobrinho-Simões M, Tangen KO . (1982). Am J Clin Pathol 77: 20–25.
Johne A, Roots I, Brockmoller J . (2003). Epidemiol Biomarkers Prev 12: 68–70.
Kimura ET, Nikiforova MN, Zhu Z, Knauf JA, Nikiforov YE, Fagin JA . (2003). Cancer Res 63: 1454–1457.
Krontiris TG, Devlin B, Karp DD, Robert NJ, Risch N . (1993). N Engl J Med 329: 517–523.
Liwo A, Gibson KD, Scheraga HA, Brandt-Rauf PW, Monaco R, Pincus MR . (1994). J Protein Chem 13: 237–251.
Mammas IN, Zafiropoulos A, Koumantakis E, Sifakis S, Spandidos DA . (2004). Gynecol Oncol 92: 941–948.
Mascaux C, Iannino N, Martin B, Paesmans M, Berghmans T, Dusart M et al. (2005). Br J Cancer 92: 131–139.
Monaco R, Chen JM, Chung D, Brandt-Rauf P, Pincus MR . (1995). J Protein Chem 14: 457–466.
Mulcahy LS, Smith MR, Stacey DW . (1985). Nature 313: 241–243.
Nikiforova MN, Lynch RA, Biddinger PW, Alexander EK, Dorn II GW, Tallini G et al. (2003). J Clin Endocrinol Metab 88: 2318–2326.
Noonan T, Brown N, Dudycz L, Wright G . (1991). J Med Chem 34: 1302–1307.
Pastinen T, Hudson TJ . (2004). Science 306: 647–650.
Pulciani S, Santos E, Long LK, Sorrentino V, Barbacid M . (1985). Mol Cell Biol 5: 2836–2841.
Rosai J, Carcangiu ML, DeLellis RA . (1993). Tumors of the Thyroid Gland (Atlas of tumor pathology, 3rd series, fascicle 5) Armed forces Isntitute of Pathology: Washington, DC.
Saavedra HI, Knauf JA, Shirokawa JM, Wang J, Ouyang B, Elisei R et al. (2000). Oncogene 19: 3948–3954.
Schelfhout LJ, Cornelisse CJ, Goslings BM, Hamming JF, Kuipers-Dijkshoorn NJ, van de Velde CJ et al. (1990). Int J Cancer 45: 16–20.
Schneider S, Roessli D, Excoffier L . (2000). Arlequin: a software for population genetics data analysis. Genetics and Biometry Lab, Department of Anthropology, University of Geneva.
Soares P, Trovisco V, Rocha AS, Lima J, Castro P, Preto A et al. (2003). Oncogene 22: 4578–4580.
Sugg SL, Ezzat S, Zheng L, Freeman JL, Rosen IB, Asa SL . (1999). Surgery 125: 46–52.
Tabin CJ, Bradley SM, Bargmann CI, Weinberg RA, Papageorge AG, Scolnick EM et al. (1982). Nature 300: 143–149.
Wang B, Soule HD, Miller FR . (1997). Anticancer Res 17: 4387–4394.
This study was partially supported by a PhD grant (SFRH/BD/6816/2001 – PC) from the Portuguese Science and Technology Foundation (FCT) and with further funding from the same source (Project – ‘Programa Operacional Ciência Tecnologia e Inovação/Ciências Biomédicas e Oncológicas/338567/2001’). Cell lines were kindly provided by David Wynford-Thomas, Jacques E Dumont, Marc Mareel, Massimo Santoro and F Savagner.
About this article
Cite this article
Castro, P., Soares, P., Gusmão, L. et al. H-RAS 81 polymorphism is significantly associated with aneuploidy in follicular tumors of the thyroid. Oncogene 25, 4620–4627 (2006). https://doi.org/10.1038/sj.onc.1209491
- RAS mutations
- H-RAS polymorphism
- H-RAS splicing