Correlation of RET somatic mutations with clinicopathological features in sporadic medullary thyroid carcinomas

Screening of REarranged during Transfection (RET) gene mutations has been carried out in different series of sporadic medullary thyroid carcinomas (MTC). RET-positive tumours seem to be associated to a worse clinical outcome. However, the correlation between the type of RET mutation and the patients' clinicopathological data has not been evaluated yet. We analysed RET exons 5, 8, 10–16 in fifty-one sporadic MTC, and found somatic mutations in thirty-three (64.7%) tumours. Among the RET-positive cases, exon 16 was the most frequently affected (60.6%). Two novel somatic mutations (Cys630Gly, c.1881del18) were identified. MTC patients were divided into three groups: group 1, with mutations in RET exons 15 and 16; group 2, with other RET mutations; group 3, having no RET mutations. Group 1 had higher prevalence (P=0.0051) and number of lymph node metastases (P=0.0017), and presented more often multifocal tumours (P=0.037) and persistent disease at last control (P=0.0242) than group 2. Detectable serum calcitonin levels at last screening (P=0.0119) and stage IV disease (P=0.0145) were more frequent in group 1, than in the other groups. Our results suggest that, among the sporadic MTC, cases with RET mutations in exons 15 and 16 are associated with the worst prognosis. Cases with other RET mutations have the most indolent course, and those with no RET mutations have an intermediate risk.

Medullary thyroid carcinoma (MTC) is a rare tumour that represents 5 -10% of all types of thyroid cancer, and accounts for a disproportionate number of thyroid cancer deaths (Hundahl et al, 1998). Except for surgery, therapy for MTC is generally ineffective. MTC may occur sporadically (in about 75% of cases) or as a part of the autosomal dominantly inherited cancer syndrome, known as multiple endocrine neoplasia type 2 (MEN 2) (Mulligan et al, 1993;Eng, 1999;Frank-Raue et al, 2007). MTC is the most common cause of death in patients with MEN 2 (Skinner et al, 2005). This familial type of thyroid carcinoma usually originates as multifocal C-cell hyperplasia, its progression to MTC is extremely variable, and may take several years (Carling, 2005). In sporadic cases, the mean age at presentation is 50 years, with a slight female predominance (Matias-Guiu et al, 2004).
Activating germline mutations in the REarranged during Transfection (RET) gene are detected in over 95% of MEN 2 cases (Mulligan et al, 1993;Marx, 2005). The oncogenic potential of different RET mutations seems to be dependent on the site of the amino acid change, and may account for the diverse phenotypes observed in MEN 2 patients (Asai et al, 1995).
The screening of RET mutations has been carried out in different series of sporadic MTC, however the observed frequencies are variable (12 -100%) (Hofstra et al, 1994;Zedenius et al, 1994;Jhiang et al, 1996;Marsh et al, 1996Marsh et al, , 2003Romei et al, 1996;Wohllk et al, 1996;Bugalho et al, 1997;Scurini et al, 1998;Shan et al, 1998;Uchino et al, 1998Uchino et al, , 1999Bockhorn et al, 1999;Dvorakova et al, 2008;Elisei et al, 2008). Met918Thr RET mutation is the most common somatic mutation in sporadic forms of MTC, and its detection rate varies greatly (5 -66%) in the published literature (Zedenius et al, 1994;Marsh et al, 1996;Romei et al, 1996;Wohllk et al, 1996;Bugalho et al, 1997;Scurini et al, 1998;Uchino et al, 1998Uchino et al, , 1999Dvorakova et al, 2008;Elisei et al, 2008). However, in some of these studies, the authors have screened sporadic MTC for only a few specific mutations, mostly in codon 918 (Hofstra et al, 1994;Shan et al, 1998;Bockhorn et al, 1999;Marsh et al, 2003). Therefore, the number of exons screened, as well as the sizes of the analysed series, may explain some of the reported differences in the prevalence of RET mutations in sporadic MTC. In addition, ethnic or environmental factors, differences in detection or in sampling methods may also account for the reported differences (Uchino et al, 1998;Dvorakova et al, 2008). In some cohorts, besides the Met918Thr mutation, other somatic mutations were also detected, at a lower frequency in exons 10, 11, 12, 13 and 15 (Bugalho et al, 1997;Scurini et al, 1998;Uchino et al, 1999).
The major somatic mutation (Met918Thr) localised in the tyrosine kinase domain in exon 16 of RET (Marini et al, 2006) has been associated to a worse clinical outcome in sporadic MTC when compared with tumours that did not harbour this mutation (Zedenius et al, 1994(Zedenius et al, , 1995Wohllk et al, 1996;Schilling et al, 2001).
Several reports have presented contradictory results concerning the ploidy pattern in MTC. Schröder et al (1988) found that most MTC have a diploid DNA pattern, and that a benign disease course was twice as frequent in patients with diploid tumours compared with aneuploid tumours. Conversely, the results presented by Lindsay (1970) seemed to be more consistent with aneuploidy in MTC.
In this study, we carried out a comprehensive analysis of exons 5, 8 and 10 -16 of RET to evaluate the prevalence of somatic mutations in a series of fifty-one sporadic MTC and to correlate with clinicopathological characteristics of the patients, including tumour ploidy pattern.

MATERIALS AND METHODS Patients
A total of fifty-two unrelated patients with MTC without family history of the disease were studied for RET mutations. A detailed personal history was obtained from all patients. All individuals were of Caucasian origin (34 females and 18 males). Each patient underwent total thyroidectomy, with the exception of two patients who were submitted to partial thyroidectomy. The diagnosis of MTC was confirmed by histopathology of the surgically removed tumours. The Tumour-Node-Metastases (TNM) classification of all tumour specimens was carried out after the criteria described in the WHO classification of thyroid tumours . Stage grouping was addressed according to the TNM classification (Sobin and Wittekind, 2002), namely, stage I (T1N0M0), stage II (T2N0M0), stage III (T3N0M0 or T1-T3N1aM0) and stage IV (T1-T3N1bM0, T4N0-N1M0 or T1-T4N0-N1M1).
The number of truly sporadic MTC patients was reduced to fiftyone, as a germline mutation was found in one case.
This study was carried out following guidelines approved by the local institution ethical board.

DNA flow cytometry
DNA flow cytometry analysis was carried out on paraffinembedded material, according to the method of Hedley et al (1983), with slight modifications (André et al, 2007).

RET variant analysis
DNA was extracted from tumour samples frozen in liquid nitrogen (n ¼ 47) following standard protocols. Otherwise, DNA was isolated from formalin-fixed paraffin-embedded tumour tissues (n ¼ 5), as described earlier (Imyanitov et al, 2001). Exons 5, 8 and 10 through 16 of RET were amplified by PCR. Sequences of the oligonucleotide primers and amplification conditions are available on request. Sequencing was carried out in both sense and antisense directions, using the same primers as for PCR amplification and the ABI PRISM BigDye Terminator version 1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA), in an automated DNA sequencer (ABI PRISM 310 Genetic Analyser, Applied Biosystems). All the mutations identified were confirmed by two independent experiments (restriction enzyme analysis, or repeated sequence analysis). To support somatic origin of the mutations, constitutional DNA from peripheral blood or non-tumourous tissue from the same patient was also analysed.

Statistical analysis
The statistical analysis was accomplished using GraphPad Prism version 4.0 statistical software (GraphPad Software Inc., San Diego, CA, USA) and SPSS 13.0 for Windows (SPSS Inc., Chicago, IL, USA). Values were expressed as mean ± s.e.m. The w 2 or Fisher's exact tests, and one-way analysis of variance or Kruskal -Wallis test were used according to the studied variables. Survival curves were analysed using the Kaplan -Meier method, and the statistical significance was assessed by the logrank test. Values of Po0.05 were considered statistically significant.

Genetic analysis
One out of the fifty-two (1.9%) cases of clinically apparently sporadic MTC carried a new germline mutation (Cys515Trp) located in exon 8 (manuscript under preparation). This case was excluded from further analysis.
In the remaining fifty-one sporadic MTC cases, thirty-three (64.7%) had mutations in RET exons 10, 11, 15 and 16 ( Table 2). The absence of these mutations in the constitutional DNA excluded a germline origin.
Among the RET-positive cases, exon 16 was the most frequently affected (60.6%) by the same specific Met918Thr mutation, followed by exon 11 (21.2%). RET mutations were also detected in exons 10 (9.1%) and 15 (9.1%). In the present series, two novel RET mutations (Cys630Gly and c.1881del18) located at exon 11 were identified. The novel Cys630Gly variant creates a restriction site for the enzyme BsrI, facilitating its independent confirmation (data not shown). The other novel variant (c.1881del18) is expected to lead to the replacement of seven amino acid residues by a glutamic acid residue. In this case, PCR amplification originated two fragments: the expected wild-type product (322 bp) and a smaller mutant product (304 bp), allowing the independent sequencing of both alleles (data not shown). Together with this mutation, and in the same allele, an unreported heterozygous nucleotide change in codon 634 (TGC to TGT) that does not predict an amino acid alteration (Cys to Cys) was found (data not shown).
In the MTC tissue of patient 51, with a Cys618Arg mutation (Table 1), the wild-type allele was not detected. The finding of allelic loss at flanking markers D10S141 and ZNF22 showed hemizygosity for this mutation (data not shown).
Forty-four (86.3%) tumours displayed a diploid DNA content and seven (13.7%) were aneuploid. No correlation between the presence or type of RET mutation and the ploidy pattern was observed (Table 3). Table 1 describes the clinical and pathological data of the fifty-one patients with sporadic MTC.

Clinical evaluation
In the thirty-three MTC patients (19 females and 14 males) carrying somatic RET mutations, the mean age at surgery and mean follow-up time were 52.9 years (median 55, range 26 -71) and 94.8 months (median 100, range 4 -303), respectively. Lymph node and distant metastases were present in 23/33 (69.7%) and 11/31 (35.5%) cases, respectively. According to the TNM classification, four patients (12.1%) had stage I disease, three (9.1%) had stage II, two (6.1%) had stage III and twenty-four (72.7%) had stage IV. At the time of the last clinical screening, nine patients (29.0%) were free of disease, and twenty-two (71.0%) were non-disease free (seven of them were deceased from MTC). The status at last control from two patients was not available. Among the twenty-two patients with persistent disease, ten (47.6%) showed a biochemical persistence of the disease with detectable levels of serum CT, but no evidence of distant metastases, whereas eleven patients (52.4%) were affected by metastatic disease. Clinical data from one case were not available.
As regard to the eighteen MTC patients (14 females and 4 males) without somatic RET mutation, the mean age at surgery and mean follow-up were 55.8 years (median 59.5, range 27 -82) and 90.8 months (median 73.5, range 0.6 -252), respectively. Lymph node and distant metastases were present in 11/18 (61.1%) and 3/18 For all these clinical and pathological data, there was no statistically significant difference between RET-positive and RETnegative patients. When the survival curves of RET-positive and RET-negative MTC patients were compared, a lower percentage of surviving patients was observed in the group with the somatic RET mutation, although not statistically significant (data not shown).
Group 1 cases (compared with group 2) had higher prevalence of lymph node metastases (P ¼ 0.0051), higher number of positive lymph nodes (P ¼ 0.0017), were more frequently multifocal (P ¼ 0.037) and had more often persistent disease at last control (P ¼ 0.0242). Moreover, group 1 was more frequently associated with detectable serum CT levels at the last screening (P ¼ 0.0119), as well as with stage IV (P ¼ 0.0145; vs stages I -III), than groups 2 and 3.
Although not statistically significant, there was a trend for a higher prevalence of distant metastases in group 1.
There was no statistical significant difference between patients of the different groups regarding the remaining clinicopathological characteristics (Table 3). Furthermore, when the survival curves of MTC patients from the three groups were evaluated, no significant differences were observed between each group (data not shown).

DISCUSSION
Medullary thyroid carcinoma is clinically diagnosed as sporadic when the patient does not present other endocrine tumours, and when no other cases of MTC, pheochromocytoma or parathyroid disease are identified in the patient's family. However, only the exclusion of germline mutations in the RET proto-oncogene allows a definitive diagnosis of sporadic MTC.
The herein reported cohort is one of the largest single-country studies. Fifty-one sporadic MTC were analysed and somatic mutations were found in thirty-three (64.7%) cases. Two novel mutations were identified in exon 11 of the RET proto-oncogene, in two sporadic MTC cases: a heterozygous point mutation at codon 630 (Cys630Gly), and a 18 bp deletion at nucleotide c.1881 associated in the same allele with a silent nucleotide substitution at codon 634 (Cys634Cys). Both mutations are located in the cysteine-rich domain coding sequence, which, when mutated, has been shown to constitutively activate RET (Santoro et al, 1995(Santoro et al, , 2002. In this study, eight earlier described missense changes in RET were detected in exons 10, 11, 15 and 16. In accordance with other studies, the most common mutation was Met918Thr with a detection rate of 39.2%, which represents 60.6% of all detected mutations. In one case, loss of heterozygosity at RET flanking microsatellite markers was detected, showing hemizygosity for a sporadic missense mutation (Cys618Arg), a rare genetic event that has been reported earlier for other RET mutations (Jindrichova et al, 2003;Dvoráková et al, 2006).
As RET mutations other than Met918Thr are rare, most of the reported series did not compare the clinicopathological characteristics of Met918Thr vs other RET mutations (Dvorakova et al, 2008;Elisei et al, 2008). In our study, 13/51 (25.5%) cases had a RET mutation other than Met918Thr, which allowed such comparison. On the basis of the recent literature, RET mutations have been stratified into three risk levels, regarding the predisposition to originate MTC, as well as their in vitro transforming activity. Patients with germline mutations in RET codons 883 (exon 15) and 918 (exon 16), for which thyroidectomy is recommended at an early age, have the highest risk for the early development and the most aggressive MTC growth (Evans et al, 2007). Likewise, these two mutations (which are considered as level 3) have the highest in vitro transforming activity. Therefore, in this study, cases with Ala883Phe or Met918Thr mutations were analysed in the same group (group 1), and compared with cases with other somatic RET mutations (group 2) and cases with no RET mutation (group 3).
A statistically significant correlation was shown between group 1 and the presence of lymph node metastases, as well as the number of positive lymph nodes at the time of surgery, multifocality and a non-disease free status, compared with group 2. This correlation may account for the significantly higher frequency of patients from group 1 in stage IV and with detectable serum CT level at last control, compared with groups 2 and 3, and also for the significantly increased levels of serum CT at the last control in group 1 cases in comparison with group 2. Furthermore, such correlation supports the hypothesis that mutations in RET exons 15 and 16 are related to a more aggressive behaviour of MTC. This could be explained by an earlier dissemination of Met918Thr and Ala883Phe cases to lymph nodes (Table 3). Indeed, 26.1% of MTC cases came to clinical attention because of lymph nodes in group 1, whereas this  occurred in only 0.0% and 11.8% of the cases in groups 2 and 3, respectively. The results presented in Table 3 show a trend towards the stratification of the three groups of sporadic MTC patients into risk levels on the basis of the statistically significant clinicopathological characteristics. Group 1 patients are at the highest risk for aggressive MTC, followed by group 3 at intermediate risk, and group 2 patients, which present the lowest risk for a worse clinical outcome. Therefore, our study shows that RET mutations in exons 15 and 16 are associated with a more aggressive behaviour of sporadic MTC than other RET mutations, as it has been shown in vitro, as well as in the hereditary variants of MTC.
Taken together, these results suggest that the screening of RET somatic mutations may be helpful in the management of patients with MTC, according to the presence and type of RET somatic mutation.