Małgorzata Oczko-Wojciechowska M.D., Department of Nuclear Medicine and Endocrine Oncology, Maria Sklodowska-Curie Memorial Cancer Centre and Institute of Oncology in Warsaw, Gliwice Branch, Wybrzeże Ak 15, 44–101 Gliwice, tel.: +48 32 278 97 20, e-mail: gosiaoczko@io.gliwice.pl

This research was supported by Polish National Science Centre grant no. NN401410639.

Introduction

Medullary thyroid cancer (MTC) arises from calcitonin secreting parafollicular C cells of thyroid and accounts for 3-5% of all thyroid cancers [1-3]. The disease in ultrasound imaging does not differ from other type of tumour [4]. Medullary thyroid cancer (MTC) occurs as both hereditary and sporadic form. The hereditary type of MTC is secondary to the RET (Rearranged during Transfection) proto-oncogene germline mutations that are associated with multiple endocrine neoplasia type 2 (MEN2a, MEN2b) and familial medullary thyroid cancer (FMTC).

The RET gene is located on chromosome l0q 11.2, contains 21 exons, and encodes a receptor tyrosine kinase, which is a transmembrane protein comprising of extracellular, transmembrane, and intracellular (cytoplasmic) domains. The RET gene plays an important role during morphogenesis and is normally activated by the ligand belonging to glial cell-derived neurotropic factor (GDNF) family and co-receptor GFRα. Mutations of the RET proto-oncogene lead to its autophosphorylation and to a gain of function resulting in constitutive activation of RET receptor [5-7].

The global risk of the detection of a germinal RET mutation in the Polish MTC population is about 10%, even if there is no positive family history and no features of MEN2a syndrome are present [8]. Simultaneously, the risk of MTC development in RET germline mutation carriers is nearly 100%. Therefore, molecular diagnostics is obligatory in all MTC subjects. These germline mutations concern the following RET gene exons 5, 8,10,11,13,14,15, and 16. A germline mutation with a higher transforming activity, located in RET exon 16 (M918T), results in MEN2b syndrome (medullary thyroid cancer, pheochromocytoma, and typical phenotype features). The most characteristic mutation for MEN2a syndrome, located in RET exon 11 (codon 634), also related with the highest probability of pheochromocytoma and parathyroid hyperplasia, is observed in up to 20% of MEN2a subjects [9, 10].

Somatic RET mutations are also detectable in two-thirds of sporadic MTC cases. Alteration in RET exon 16 (M918T) is the most frequent one, accounting for nearly 80% of all detected RET somatic mutations. It is associated with more aggressive course of the disease [11-13]. Recent studies suggest a high proportion of RAS somatic mutations inRET-negative tumours [14-16]. Thus,RAS gene mutations could be considered as an alternative genetic event in sporadic MTC tumorigenesis. These mutations are present in three RAS proteins: K-RAS, H-RAS, and N-RAS – small GTPases that play a role in cellular growth, differentiation, adhesion, and migration. Normally, RAS is activated by external signals from tyrosine kinases receptor (Fig. 1). Mutations of this gene lead to constitutive activation of the RAS protein. Most RAS mutations are limited to hot spot codons located in exons 2 and 3. A mutation in exon 3 of N-RAS (codon 61) is the most frequent in poorly differentiated thyroid cancers [17,18] (Fig. 1).

Figure 1. MAP Kinase signalling pathway. RTK – receptor tyrosine kinase
Rycina 1. Ścieżka sygnałowa MAPK. RTK – receptor kinazy tyrozynowej

The aim of our study was to evaluate the frequency of somatic RAS mutations in sporadic medullary thyroid cancer in the Polish population and to relate the obtained data to the somatic RET mutations.

Material and methods

Samples

Seventy-eight MTC samples were analysed (57 sporadic and 21 hereditary). Tumour samples were collected from the biobank of the Department of Nuclear Medicine and Endocrine Oncology in M. Sklodowska-Curie Memorial Cancer Centre and Institute of Oncology, Gliwice Branch. RET mutation status was evaluated in all 57 sporadic MTC tumour samples, whereas RAS mutation status was assessed in 49/57 sporadic MTC samples and in all 21 hereditary MTC tumours.

RET and RAS analysis method

DNA for RET germline mutations was extracted from peripheral blood by the desalting method and Genomic Maxi AX Kit (A&A Biotechnology). Mutation screening was performed according to a standard algorithm approved by the American and European MTC Management Guidelines [9, 10], which assumes analysis of exons 10, 11, 13, 14, 15, and 16. These exons were sequenced directly using Big Dye 1.1 reagent and a 3130x1 Genome Analyser (Life Technologies). For RET and RAS somatic mutations, DNA was extracted from snap-frozen tissue samples by the DNeasy Blood and Tissue Kit (Qiagen). Twenty-one exons of the RET gene and exon 2 and 3 of H-RAS, K-RAS, and N-RAS (codons 12, 13, and 61) were sequenced directly as described above.

Deep sequencing

Next generation sequencing of RNA was performed in three randomly selected RET-negative MTC samples using TruSeq Sample Preparation Kit (Illumina) and TruSeq SR Cluster Kit (Illumina) according to the manufacturer’s protocol. Single-end 76 bp sequencing on Genome Analyzer IIx (Illumina) was carried out. Low quality reads were filtered out with FASTX_Toolkit 0.0.13. Next, good quality reads were aligned to the human genome hgl9 using TopHat 2.0.11 [19]. Duplicate reads were removed with PicardTools 1-56. Variants (single nucleotide variants, small insertions, and deletions) were identified using VarScan 2.3.6 [20]. The annotation of variants was performed with Annovar [21].

Results

Somatic RET mutational screening in sporadic MTC

Among 57 sporadic MTC samples somatic RET point mutations were found in 33 cases (57.8%). The most frequent RET mutation was localised in codon M918T, 16/33 (48.5%). The other RET mutations were detected in codon C634, 6/33 (18%), codon C 630, 2/33 (6%), and one in codon C618S (3%), codon A883F (3%), and H568N (3%). Deletions of 6 to 9 nucleotides in RET exon 11 were found in 4 sporadic MTC tumours and one deletion with insertion in exon 11. Detailed results are presented in Table I.

Somatic RAS mutational screening in sporadic MTC

The status of three RAS genes was investigated in 28 somatic RET-positive MTC samples and 16 somatic RET-negative samples. A RAS mutation was found in 2/33 of sporadic MTC samples with RET somatic mutation (6%) and in 11/16 (68.8%) of RET negative MTC tumours. Among sporadic RET-positive samples, one tumour harboured H-RAS codon 61 mutation, whereas K-RAS codon 13 mutation was found in the second sample. Among RET-negative samples, 8 H-RAS mutations were detected and 3 K-RAS. The details are reported in Table II.

None of 21 hereditary MTC samples showed any RAS mutations. Moreover, none of all analysed samples (21 germline RET-positive, 28 somatic RET-positive, and 16 somatic and germline RET-negative samples) showed N-RAS mutation.

Next generation sequencing

The NGS was performed in three randomly selected RET-negative MTC samples. In RNA-seq, 22 million good quality reads were obtained for MTC10, 24 million for MTC20, and 28 million for MTC22 to detect single nucleotide variants, insertions, and deletions. Among identified variants, two mutations were present in HRAS exon 3 in two distinct MTC samples. They were p.Q61K (c.C181A) HRAS mutation in MTC20 and p.Q61R (c.A182G) HRAS mutation in MTC22. No other putative driver mutations were found.

Discussion

Somatic mutations of three RAS genes have been reported with reference to several tumours [22], among them differentiated thyroid cancer [23-25]. From the clinical point of view, screening for RAS mutations is important in patients with colorectal carcinoma where the presence of K-RAS mutation is associated with poor response to anti-EGFR therapies [26-28]. However, in differentiated thyroid cancer, the detection of RAS mutation has not thus far determined any treatment decision. Interestingly, all mutations that play role in thyroid cancer despite its origin (follicular or parafollicular cells) are involved in MAPK kinase pathway.

The prevalence of RET somatic mutations (57,8%) in our analysed group is consistent with previous reports [15, 29-31], where the frequency of somatic RET mutation ranges between 38% [29] and 71% [15]. Moreover, similarly to other data [29, 30], the most frequent somatic RET mutation in our group was M918T, which constituted nearly 50% of all RET somatic mutations. However, other somatic mutations in exons 10-11 were also found in our material.

In the present study, a mutational analysis of EZRAS, K-RAS, and N-RAS in a series of 65 MTC samples (49 sporadic and 21 hereditary), representative for the Polish population, was performed. RAS mutation was detected in 26.5% of all sporadic MTC tumours, but when only RET-negative samples were considered, the prevalence of RAS mutation was 68.7%. A higher proportion of H-RAS mutation that constituted 72% of all detected RAS mutations was observed, which is consistent with all reported results [15, 16, 32] and comparable with studies published by Moura et al. and Boichard et al. that reported RAS mutations present in 68% and 81% of RET-negative sporadic MTC samples, respectively [14, 16]. The results of different analyses of RAS mutations in MTC are controversial with reference to their different frequency reported by different studies. Ciampi et al. suggest that RAS mutations are extremely rare in MTC and stand for ~11% of all sporadic MTC cases and 17.6% when only RET-negative MTC samples were evaluated [33]. Schlumberger also reported such a low prevalence of RAS mutation in MTC, only in 8% of all sporadic MTC³⁴. Some studies did not show any RAS mutation in MTC [35] or confirmed its very low frequency, like in the Shulten et al. study [36], where the only one sample among 15 MTC tumours analysed demonstrated RAS mutation, while in our study RAS mutation was diagnosed in 26.5% of all cases.

The reason of variability in the prevalence of RAS mutations in MTC is unknown. One explanation could be related to differences between analysed populations. However, when we look at Italian results only, some differences may also be noticed. Ciampi et al. reported a very low prevalence of RAS mutation in RET-negative sporadic MTC (17.6%) [32] while in the Scarpa and Fugazzola group the frequency of RAS mutation was 57% [31]. This percentage was even higher in our group – 68.7 % of RET-negative tumours showed RAS mutation. However, of note, the Elisei group involved the largest series of analysed MTC samples from different Italian centres [32]. The second explanation could be the different methodologies used for mutational RAS screening and their sensitivity. This is also visible in the Italian results, where Ciampi et al. used PCR and direct sequencing [32], while Símbolo et al. applied next generation sequencing [31]. Nonetheless, these differences do not explain all discrepancies because PCR and direct sequencing were used in many studies. The third explanation is related to the material used for analysis. Some studies comparing the results achieved from formalin-fixed paraffin-embedded (FFPE) samples and frozen tissues demonstrated a higher rate of RAS mutations in snap-frozen tissues than in FFPE samples [37]. Another important issue related to FFPE samples is the poor quality of the material leading to false-negative results [38]. This could explain the low rate of RAS mutation in the Schlumberger et al. and Rapa et al. studies as both groups used FFPE samples for RAS mutational screening, contrary to data from Ciampi et al. performed on snap-frozen samples [32, 34, 35]. We underline that we also used snap-frozen MTC samples.

Interestingly, in our set of RET-positive sporadic MTC samples, RAS mutations were also found (6%). Until now RET and RAS mutations were reported as mutually exclusive [15, 16, 32], and only in one study was RAS mutation detected in RET-positive MTC [14]. The coexistence of two oncogenic mutations is rare. In some papillary thyroid cancer cases RET rearrangements coexist with B-RAF or RAS mutation [39-41] or BRAF and K-RAS mutations [42]. A similar situation is observed in lung adenocarcinoma, where PIK3CA mutations are accompanied by with K-RAS or EGFR mutations [43, 44]. These data may suggest that RAS could be a secondary event, not a driver mutation. However, to date there is no published transforming assays demonstrating that RAS mutations play an important role in MTC tumorigenesis.

Summary

The analysis of three RAS genes (H-RAS, K-RAS, N-RAS) revealed that 29% of sporadic MTCs harbour RAS mutations. This rate was much higher when only RET-negative sporadic MTC were analysed – 69%. Most of these mutations were located in H-RAS codon 61 (72%). 3% of RAS mutations were detected in RET-positive sporadic MTC. No N-RAS-positive samples were found.

The prevalence of RET somatic mutations in the Polish population is high and is observed in nearly 58% of all sporadic MTC tumours. Among them, M918T RET mutations constitute 50%.

Conclusions

RAS mutations constitute a frequent molecular event in RET-negative sporadic medullary thyroid carcinoma in Polish patients. However, their role in MTC tumorigenesis remains unclear.