This article is mentioned in:

Introduction

Medullary thyroid carcinoma (MTC), a neuroendocrine tumor originating from thyroid parafollicular cells, accounts for ~4% of thyroid cancer cases (1). The majority are sporadic cases, however, 20-25% occur as a hereditary syndrome termed multiple endocrine neoplasia type 2 (MEN 2A and MEN 2B) and as familial MTC, both of which are associated with germline mutations in the RET oncogene (2).

Mutations in the RET oncogene have previously been identified in the tumor tissue of up to 64% of sporadic MTC cases (3). In addition, RAS gene mutations are observed in 10% of RET-negative cases and are associated with a subset of tumors with less aggressive behavior (4). While certain studies identified that ~90% of sporadic MTCs exhibited mutually exclusive mutations in RET, HRAS and KRAS (4–8), Moura et al (3) reported the presence of the RAS mutation in one case with RET-positive sporadic MTC and Rapa et al (9) identified no RAS mutations in 49 examined cases. Nevertheless, the clinical phenotype of sporadic and inherited MTCs is heterogeneous even in the presence of the same mutation; however the molecular mechanisms underlying the pathology remain to be fully elucidated.

In addition, it remains unclear whether there is a modulatory role in MTC tumor progression for additional genes such as BRAF, CDKN2A and PI3KCA. These genes participate in the tumorigenesis of several types of human malignancies such as tumors derived from neural crest cells, including melanoma, pheochromocytoma and paraganglioma (10–12).

BRAF, like RET and RAS, is involved in the mitogen-activated protein kinase pathway and has a well-established role in the pathogenesis of malignancies such as melanoma and papillary thyroid cancer (13). Nevertheless, the contribution in the tumorigenesis of MTC remains controversial. A previous study reported a high prevalence of the p.Val600Glu BRAF mutation in sporadic MTC cases (14); however, subsequent studies did not confirm this observation (3,9,15,16).

An additional tumor suppressor gene, CDKN2A/p16INK4A, is involved in the G1/S transition in the cell cycle. Mutations and deletions have been identified in melanoma, and polymorphisms in its 3′ untranslated region (UTR) have been associated with earlier progression from primary to metastatic disease (17). By contrast, polymorphisms in another tumor suppressor gene, CDKN1B, which is in the same CDKN family, are associated with improved outcomes (18).

Additionally, PI3KCA is a gene that serves an important role in signaling pathways and cell growth, and contributes to tumorigenesis in several types of human malignancy (19,20). However, the role of this gene in the tumorigenesis of MTC remains to be fully understood.

Therefore, the current study aimed to verify the prevalence of somatic mutations in BRAF, CDKN2A and PI3KCA, which have already been described in other neural crest-derived tumors, and to determine the possible supporting role of these genes in the tumorigenesis of MTC.

Patients and methods

Patients and tissue samples

From 128 patients with MTC assessed at the Multiple Endocrine Neoplasia outpatient clinic at the Universidade Federal de Sao Paulo (Sao Paulo, Brazil) between February 2007 and June 2013, formalin-fixed paraffin-embedded (FFPE) tumor tissues were selected from 31 patients on the basis of the availability of tumor tissues, with no other selection criteria. DNA extraction was subsequently performed, using an in-house method as previously described (21). Subsequent to DNA extraction, 20 samples (from 13 males and 7 females; mean age, 40.55±16.74 years) provided the appropriate quantity and quality of DNA. The study was approved by the Ethics and Research Committee of the Universidade Federal de Sao Paulo (protocol number 1945/10), and all patients provided informed consent. Additionally, 1,092 genotypes of variant frequencies (single nucleotide polymorphisms; SNPs) were obtained from the 1000 Genomes database (http://www.1000genomes.org/) as a population genetics control.

DNA extraction and genotyping

DNA from peripheral blood and somatic DNA from 10-µm sections of FFPE MTC tissues was extracted using an in-house method as previously described (21). Polymerase chain reaction (PCR) was performed to amplify DNA corresponding to hotspot exons 2, 3 and 4 of HRAS; 2, 3 and 4 of KRAS; 2 and 3 of NRAS; 15 of BRAF; 9 and 20 of PI3KCA; and exons 2, 3 and the 3′UTR of the CDKN2A gene. The sequences of the primers are listed in Table I. The reactions were performed using 10 pM of each specific primer, 2.5 µl PCR buffer, 200 µM dNTP, 1.5 µM MgCl2 and 0.2 units Taq DNA polymerase (Invitrogen; Thermo Fisher Scientific, Waltham, MA, USA) in a 25-µl total reaction volume. The cycling conditions were as follows: 5 min at 95°C, 38 cycles of 45 sec at 95°C, 45 sec for annealing and 1 min at 72°C, and a final elongation for 10 min at 72°C. The PCR products were purified using the Illustra GFX PCR DNA and Gel Purification kit (GE Healthcare Life Sciences, Chalfont, UK) and were subject to sequencing using the Sanger method, with the Big Dye™ Terminator Cycle Sequencing Ready Reaction kit and the ABI PRISM 3130xl Genetic Analyzer (Applied Biosystems; Thermo Fisher Scientific). Gel electrophoresis of the PCR products was performed to analyze product quality and yield using a 1.8% agarose gel and a DNA ladder.

Table I Primers used in the present study.

Table I

Primers used in the present study.

Gene	Forward primer	Reverse primer
BRAF exon 15	5′-AACTCAGCAGCATCTCAGGG-3′	5′-CTTCATAATGCTTGCTCTGATAG-3′
CDKN2A exon 1	5′-ACCCTGGCTCTGACCATTC-3′	5′-CAGGTCACGGGCAGAC-3′
CDKN2A exon 2	5′-GACCTCAGGTTTCTAACGCC-3′	5′-CATATATCTACGTTAAAAGGCAGGAC-3′
PI3KCA exon 9	5′-TGGCAGTCAAACCTTCTCTC-3′	5′-GAGAAAGTATCTACCTAAATCCACAGA-3′
PI3KCA exon 20	5′-AAATGTTTTGGTGTTCTTAATTTATTC-3′	5′-GCAGCCAGAACTCTTTATTTTG-3′
C-kit exon 9	5′-GCCAGGGCTTTTGTTTTCTT-3′	5′-AGCCTAAACATCCCCTTAAATTG-3′
C-kit exon 11	5′-AACCATTTATTTGTTCTCTCTCCA-3′	5′-CCACTGGAGTTCCTTAAAGTCA-3′
C-kit exon 17	5′-TGGTTTTCTTTTCTCCTCCAAC-3′	5′-GGACTGTCAAGCAGAGAATGG-3′
HRAS exon 2	5′-GGCAGGAGACCCTGTAGGAG-3′	5′-AGCTGCTGGCACCTGGAC-3′
HRAS exon 3	5′-GTCCCTGAGCCCTGTCCTC-3′	5′-CAGCCTCACGGGGTTCAC-3′
HRAS exon 4	5′-CTCTCGCTTTCCACCTCTCA-3′	5′-GGGTGGAGAGCTGCCTCA-3′
KRAS exon 2	5′-TTAACCTTATGTGTGACATGTTCTAA-3′	5′-GGTCCTGCACCAGTAATATGC-3′
KRAS exon 3	5′-AGACTGTGTTTCTCCCTTCTCA-3′	5′-TGGCATTAGCAAAGACTCAAA-3′
KRAS exon 4	5′-GATATTTGTGTTACTAATGACTGTGCT-3′	5′-TTATGATTTTGCAGAAAACAGATC-3′
NRAS exon 2	5′-TCGCCAATTAACCCTGATTAC-3′	5′-TCCGACAAGTGAGAGACAGG-3′
NRAS exon 3	5′-TGGGCTTGAATAGTTAGATGC-3′	5′-AGTGTGGTAACCTCATTTCCC-3′

In silico analysis of HRAS mutations and CDKN2A polymorphisms

Mutational analysis of HRAS was performed by the use of Project HOPE to obtain structural information from the analysis of PDB-file 1CTQ (22). The in silico analysis for the CDKN2A polymorphisms was performed using the Functional Single Nucleotide Polymorphism database (http://compbio.cs.queensu.ca/F-SNP/) as previously described (23). This database provides information regarding potential deleterious effects of SNPs with respect to splicing, transcription, translation and post-translation based on SNP functional significance (FS). The FS score for neutral SNPs is 0.1764, whereas the FS score for disease-associated SNPs is in the range of 0.5-1.

Statistical analysis

The allele and genotype frequencies were compared between patients with MTC and the 1000 Genomes database controls using a χ2 test. The clinicopathological features of patients carrying each of the polymorphisms rs11515 and rs3088440 were compared with those of patients without such polymorphisms using the χ2 test or the Student's unpaired t-test as appropriate. P<0.05 was considered to indicate a statistically significant difference, and the Hardy-Weinberg equilibrium was evaluated. Statistical analyses were performed using SPSS, version 22.0 (IBM SPSS, Armonk, NY, USA) and GraphPad Prism, version 3.0 (GraphPad Software, Inc., La Jolla, CA, USA).

Results

Screening of the RET, HRAS, KRAS and NRAS genes

Mutational screening of the RET gene was performed on all 20 patients. A total of 10 cases were identified to be familial tumors as confirmed by the presence of a germline mutation. In total, 30% of the sporadic cases (3/10) presented with a RET somatic mutation. The clinicopathological features and molecular analysis, including tumor staging based on the American Joint Committee in Cancer staging system (24), are summarized in Table II.

Table II Summary of patient clinicopathological features and molecular analysis.

Table II

Summary of patient clinicopathological features and molecular analysis.

Patient	Gender	Age at diagnosis (y)	pTNMa	Germline RET allele	Somatic RET allele	Somatic H-, K-, NRAS allele	Somatic CDKN2A
1	M	28	T2N1bMx	WT	WT	HRAS_p.Asp33Asn	rs11515
2	F	25	T3N1bMx	WT	p.Met918Thr	-	WT
3	M	38	T1N1aMx	WT	WT	NA	rs11515/rs3088440
4	M	56	T3N1bMx	WT	WT	WT	WT
5	F	49	T2NxMx	WT	p.Gln681Stop	-	WT
6	M	69	T2N0Mx	WT	WT	HRAS_p.Gln61Arg	rs3088440
7	M	27	T4N1Mx	WT	WT	WT	WT
8	M	51	T3N1bMx	WT	p.Met918Thr	-	rs11515
9	F	56	T1N1bMx	WT	WT	HRAS_p.Asp33Asn	WT
10	M	41	T4N1bMx	WT	WT	HRAS_p.His94Tyr	WT
11	M	27	T1N1aMx	p.Cys634Arg	-	-	rs11515/rs3088440
12	F	21	T1N1aMx	p.Gly533Cys	-	-	rs11515
13	M	61	T1N1aMx	p.Gly533Cys	-	-	WT
14	F	22	T2N0Mx	p.Cys634Arg	-	-	rs11515/rs3088440
15	M	43	T2N0Mx	p.Cys634Arg	-	-	rs11515
16	M	72	T1N0Mx	p.Cys634Arg	-	-	WT
17	M	45	T1N1aMx	p.Cys634Arg	-	-	rs3088440
18	F	31	T1NxMx	p.Cys634Arg	-	-	rs3088440
19	F	15	T1N1aMx	p.Cys634Arg	-	-	rs3088440
20	M	40	T1N0Mx	p.Gly533Cys	-	-	WT

a TNM (Tumor, Node, Metastasis)/American Joint Committee on Cancer staging system. M, male; F, female; y, years; NA, not available; WT, wild-type.

To investigate exclusive causative mutations in cases of sporadic MTC other than RET mutations, HRAS, KRAS and NRAS were screened for somatic mutations in the hotspots. The majority of these patients had been previously analyzed for RET germline mutations as part of our routine evaluation, and for RET somatic mutations in a previous study (25) Two novel HRAS mutations, p.Asp33Asn and p.His94Tyr, were detected in RET-negative MTC tumors. Mutational analysis using Project HOPE suggests that the p.His94Tyr mutation is deleterious, and that the p.Asp33Asn mutation is likely to be damaging (Fig. 1). No differences in the clinical presentation or histological observations were noted between patients with MTC that had a mutation in the RAS gene (Table II).

Figure 1 Mutational analysis of the HRAS somatic mutations p.Asp33Asn and p.His94Tyr. Electropherogram of tumor tissues (A) 1 and (B) 10; (C and D) sequence alignment of human HRAS protein residues in which the position of the conserved amino acids are indicated (arrows); multiple sequence alignment was generated with Clustal Omega software (http://www.ebi.ac.uk/Tools/msa/clustalo/), *indicates that the residues in the column were identical in all sequences in the alignment; schematic structures of the (E) original and (F) mutant amino acids in the two HRAS mutations; (G and H) structure of the HRAS proteins in ribbon-presentation; gray, protein; magenta, side chain of the mutation (p.Asp33Asn and p.His94Tyr).

No somatic mutations were identified in exon 15 of BRAF or in exons 9 and 20 of PI3KCA. Patient 9 was not analyzed for somatic mutations in PI3KCA due to an insufficient number of tumor samples.

Despite not having identified somatic mutations in CDKN2A hotspots, two polymorphisms in the 3′UTR regulatory region, 500 C→G (rs11515) and 540 C→T (rs3088440), were identified in the patients observed. The heterozygotic pattern of the two SNPs was observed in the same proportion, 7/20 MTC (35%). The genotype distribution was identified to be in the Hardy-Weinberg equilibrium and was not identified to exhibit linkage disequilibrium. To investigate whether the observed polymorphisms were limited to a somatic event, they were further analyzed in the peripheral blood, which confirmed germline inheritance. The in silico analysis demonstrated that the CDKN2A polymorphisms rs11515 and rs3088440 are located in the transcriptional regulatory region and that the nucleotide alterations may affect the binding of transcription factors.

In seven cases, it was possible to detect the presence of these polymorphisms in the secondary tumors in the lymph nodes (tumor metastases), however no differences between the genotypes of the primary and secondary tumors were observed, indicating that there was no additional somatic event in CDKN2A involved in the metastatic process. This analysis was additionally performed for BRAF and PI3KCA in metastatic tissues.

No associations between the polymorphisms and the clinicopathological features observed were identified (Table III). In addition, the frequency of the SNPs was compared with a population genetics control, and there was no significant difference between the two populations (Table IV).

Table III Correlation between CDKN2A SNPs and clinicopathological features in the patient cohort.

Table III

Correlation between CDKN2A SNPs and clinicopathological features in the patient cohort.

Clinicopathological feature	rs11515 (n=20)			rs3088440 (n=20)
	CC (n=13)	CG (n=7)	P-value	CC (n=13)	CT (n=7)	P-value
Gender			0.526			0.474
Male (n=13)	8/13 (61.5%)	5/7 (71.4%)		9/13 (69.2%)	4/7 (57.1%)
Female (n=7)	5/13 (38.5%)	2/7 (28.6%)		4/13 (30.7%)	3/7 (42.9%)
Age at diagnosis			0.088a			0.272a
Mean ± SD (y)	45.41±17.49	31.53±11.36		44.062±17.49	31.53±11.36
Tumor type			0.500			0.175
Sporadic (n=10)	7/13 (53.8%)	3/7 (42.9%)		8/13 (61.5%)	2/7 (28.5%)
Familial (n=10)	6/13 (46.1%)	4/7 (57.1%)		5/13 (38.5%)	5/7 (62.5%)
T category			0.464			0.291
T1	7/13 (53.8%)	2/7(28.5%)		5/13 (38.5%)	4/7 (57.1%)
T2	2/13 (15.3%)	3/7 (42.9%)		3/13 (23.1%)	2/7 (28.5%)
T3	2/13 (15.3%)	2/7 (28.5%)		3/13 (23.1%)	1/7 (14.4%)
T4	2/13 (15.3%)	0/7 (0%)		2/13 (15.3%)	0/7 (0%)
Tumor size			0.421a			0.689a
Mean ± SD (cm)	1.954±1.11	2.34±1.03		2.315±1.22	1.671±0.59
<2	8/13 (61.5%)	2/7(28.6%)	0.378	7/13 (53.8%)	4/7 (57.1%)	0.339
≥2	5/13 (38.5%)	5/7 (71.4%)		6/13 (46.1%)	3/7 (42.9%)
Lymph node metastases			0.742			0.742
N0	4/13 (30.8%)	5/7 (71.4%)		3/13 (23.07%)	3/7 (42.9%)
N1	9/13 (69.2%)	2/7 (28.5%)		10/13 (76.9%)	4/7 (57.1%)
AJCC stage			0.742			0.742
I and II	4/13 (30.7%)	2/7 (28.5%)		4/13 (30.7%)	2/7 (28.5%)
III and IV	9/13 (69.2%)	5/7 (71.4%)		9/13 (69.2%)	5/7 (71.4%)

{ label (or @symbol) needed for fn[@id='tfn2-mmr-13-02-1653'] } P-values were obtained using the χ² test;

a continuous variables analyzed with Student's t-test. SNPs, single nucleotide polymorphisms; SD, standard deviation; y, years; AJCC, American Joint Committee on Cancer.

Table IV Comparative analysis of the frequency of the non-coding CDKN2A germ line single nucleotide polymorphisms in patients with MTC and the control.

Table IV

Comparative analysis of the frequency of the non-coding CDKN2A germ line single nucleotide polymorphisms in patients with MTC and the control.

A, rs11515
	Genotype frequency			Allele frequency
Population	CC	CG	GG	C (32)	G (8)	P-value
MTC	0.60	0.40	-	0.80	0.20	0.25
1,000 genomesa	0.79	0.19	0.02	0.88	0.12
B, rs3088440
	Genotype frequency			Allele frequency
Population	CC	CT	TT	C (31)	T (9)	P-value
MTC	0.55	0.45	-	0.78	0.22	0.65
1,000 genomesa	0.73	0.24	0.03	0.85	0.15

a Sequences obtained from the 1000 Genomes database used as a population control. MTC, medullary thyroid carcinoma. The numbers in parentheses represent the frequency of each allele type in this locus in the studied cohort.

Discussion

The adjuvant role of additional genes in the tumorigenesis of MTC was investigated in the current study through analysis of tumor tissues from 20 patients. Screening in hotspot regions of BRAF, CDKN2A and PI3KCA did not identify any somatic mutations in the coding region. In addition, the results of the current study were not in agreement with the BRAF mutation frequency of 68.2% observed by Goutas et al (14). This suggests that BRAF does not serve an important role in the tumorigenesis of MTC. The observations of the current study concerning MTC are consistent with a previous study that demonstrated that somatic mutations in genes other than RET and RAS are very rare or even absent (5). Notably, the present study identified two novel HRAS mutations.

Additionally, two common polymorphisms in the 3′-UTR non-coding region of the gene CDKN2A were identified, rs11515 and rs3088440 (26). It is known that protein synthesis can be modulated by regulatory elements located in the 5′-UTR and 3′-UTR regions. The 3′-UTR, the site of the polymorphisms identified in the current study, serves an important role in translation and mRNA stability. Alterations in this region may be associated with the onset or progression of disease (27).

These polymorphisms have been investigated in various tumor types including urinary bladder neoplasm (28), esophageal adenocarcinoma (29) and cervical cancer (30) as presented in Table V. The two identified polymorphisms have been previously associated with an earlier progression from primary to metastatic disease in the case of melanoma (17), and rs3088440 was associated with the mechanism of tumor invasion in bladder cancer (28). Controversially, this polymorphism has been previously associated with a sub-group with reduced vertical growth of melanoma and a favorable outcome (31). However, additional studies have not identified a clinical correlation with tumor behavior (30,32,33).

Table V Summary of the studies on CDKN2A polymorphisms in different tumor types.

Table V

Summary of the studies on CDKN2A polymorphisms in different tumor types.

Source, year (ref)	rs11515 (%)	rs3088440 (%)	Tumor type	n	Sample	Method used
Sauroja et al, 2000 (17)	16.67	16.67	Melanoma	48	Frozen/FFPE tissue	PCR-SSCP/sequencing
Kumar et al, 2001 (26)	25	27.27	Melanoma	229	FFPE tissue	PCR-SSCP
Sakano et al, 2003 (28)	18.1	12.9	Bladder	309	Blood	PCR-SSCP
Geddert et al, 2005 (29)	13.3	-	ADC	315	FFPE tissue	PCR-RFLP
Chansaenroj, et al 2013 (30)	7.1	17.9	Cervical	56	Cervical swab	Sequencing
Straume et al, 2002 (31)	25	23	Melanoma	185	FFPE tissue	PCR-SSCP/sequencing
Boonstra et al, 2011 (32)	22.07	-	EAC	214	FFPE tissue	Sequencing
Pinheiro et al, 2014 (33)	21.05	-	ESCC	97	FFPE tissue	Sequencing
	15.63	-	HNSCC	96	FFPE tissue	PCR-RFLP
Jin et al, 2012 (34)	-	16.7	SGC	156	Blood	PCR-RFLP
Polakova et al, 2008 (35)	25.98	13.07	Colorectal	612	Blood	PCR-RFLP
Royds et al, 2011 (36)	31.78	-	GBM	107	Blood	Sequencing
Thakur et al, 2012 (37)	13.64	-	Cervical	150	Fresh tissue	PCR-RFLP
Zhang et al, 2011 (38)	-	17.0	SCCHN	1,287	Blood	PCR-RFLP
Zhang et al, 2013 (39)	-	20.5	DTC	303	Blood	PCR-RFLP
	-	20.9	PTC	273	Blood	PCR-RFLP
De Giorgi et al, 2014 (4 0)	16.67	-	Melanoma	12	Blood	Sequencing
Song et al, 2014 (41)	-	33.88	SCCOP	552	Blood	PCR-RFLP
Nascimento et al, 2015a	35	35	MTC	20	FFPE tissue + blood	Sequencing

a Indicates the current study. ADC, gastric and esophageal adenocarcinomas; EAC, esophageal adenocarcinoma; ESCC, esophageal squamous cell carcinoma; GBM, glioblastoma multiforme; SCCHN, squamous cell carcinoma of the head and neck; SGC, salivary gland carcinoma; DTC, differentiated thyroid carcinoma; PTC, papillary thyroid cancer; HNSCC, head and neck squamous cell carcinoma; SCCOP, squamous cell carcinoma of the oropharynx; FFPE, formalin-fixed paraffin-embedded; PCR; polymerase chain reaction; SSCP, single-strand conformation polymorphism; RFLP, restriction fragment length polymorphism.

Using in silico analysis, the current study identified that the polymorphisms rs11515 and rs3088440 are located within a transcriptional regulatory region, and the alteration of nucleotides can affect the binding of potential transcriptional factors. For example, the presence of the C allele in rs3088440 favors the binding of the transcription factor c-Myb, which potentially results in the transcriptional repression of the CDKN2A gene, compromising its normal function in cell cycle control (42). However, no association was identified between this polymorphism and the clinicopathological parameters investigated in the cohort studied (Table III).

In conclusion, it is suggested that BRAF, CDKN2A and PI3KCA, listed as potential adjuvants in the tumorigenesis of MTC, do not participate through somatic mutations as modulators of oncogenesis. To the best of our knowledge, the current study is the first to investigate these two CDKN2A polymorphisms in the pathophysiology of MTC. Therefore, CDKN2A and its regulatory regions and the additional genes involved in tumorigenesis warrant further investigation in MTC.

Acknowledgments

The authors would like to thank the team of the Laboratory of Molecular and Translational Endocrinology, particularly Ms. Teresa Kasamatsu, Mr. Gilberto Furuzawa, Dr João Roberto Martins, Dr Ji Hoon Yang, Dr Fausto Germano Neto and Mr. Fernando Soares. The authors would additionally like to acknowledge Mr. Gilmar Miranda from Siratec Ltd. for graphic art design. The current study was supported grants from the Sao Paulo State Research Foundation (grant nos. 2012/11036-3, 2012/02465-8, 2012/01628-0, 2009/50575-4, 2012/00079-3 and 2011/20747-8).

References