Introduction

Thyroid cancer (TC) is one of the most frequent cancers, and its prevalence is increasing in the last years [1]. TC may originate from follicular or parafollicular cells, referred as non-medullary thyroid carcinoma (NMTC) or medullary thyroid carcinoma (MTC), respectively. A majority of thyroid cancer cases (80–90%) originate from follicular cells, and 3–9% of them present first-degree relatives with NMTC [2, 3].

Familial non-medullary thyroid cancer (FNMTC) is defined by the presence of thyroid cancer originating from follicular cells in two or more first-degree relatives, in the absence of predisposing environmental factors. Five percent of all FNMTC cases are syndromic and the susceptibility genes involved in syndromic FNMTC are known: APC in familial adenomatous polyposis (MIM: 175100), PTEN in Cowden’s disease (MIM: 158350), PRKAR1A in Carney complex type 1 (MIM: 160980), WRN in Werner’s syndrome (MIM: 277700), and DICER1 in the DICER1 syndrome (MIM: 606241) [4]. However, most cases of FNMTC (95%) are non-syndromic and the genetic causes are still unknown [5]. Thus far, different strategies have been carried out in order to find out the genes involved in the development of FNMTC: linkage [6, 7], GWAS [8], and next-generation sequencing studies [9, 10]. Several candidate regions such as PRN (1q21), NMTC1 (2q21), FTEN (8p23), MNG1 (14q32), and TCO (19p13.2), as well as candidate genes like SRGAP1, NKX2, FOXE1, HABP2, or MAP2K5, have been suggested, but none clearly validated as causative of FNMTC [11,12,13].

It has long been known that somatic mutations in the TERT promotor gene are involved in the development of TC [14], suggesting that genes associated with, or taking part of the telomerase complex, could be involved in the development of non-syndromic FNMTC. Some studies have already been carried out in TC, e.g., TERC gene [15], showing no germline mutations in this gene.

POT1 (protection of telomeres 1) gene is located in chromosome 7 and it is a member of the shelterin complex, encoding a nuclear protein involved in telomere maintenance. Germline variants in POT1 have been described mainly not only in families with melanoma [16], but also in familial gliomas and Li-Fraumeni-like familial cancer syndromes [17, 18]. Recently, it has been suggested that germline mutations in POT1 gene could be implicated in TC, specifically the new variant p.(Lys90Glu) of POT1 identified by whole-exome sequencing (WES) in a family with multiple family members affected with melanoma as well as thyroid, kidney, and breast cancers [19]. In another family with melanoma, a case of TC has been recently reported in an individual with the p.(Ile78Thr) mutation [20]. On the other hand, it has been described that the spectrum of cancers caused by mutations in the POT1 gene is more diverse than it seemed at first [21]. Therefore, with the purpose of reinforcing or not if TC is part of the mutational POT1 phenotype spectrum, it would be interesting to know if mutations in these gene could lead to non-syndromic FNMTC. In order to elucidate this possibility, we analyzed the POT1 gene in seven selected families, each of them with three or more affected cases to minimize the possibility that thyroid cancer could have been due to chance in these families.

Methods

Patients

We designed a multicentric study in Spain to collect blood specimens (15 ml of whole blood in potassium EDTA tubes), and clinical data from families with at least three first-degree relatives affected with NMTC, confirmed by histology, without history of other malignancies, and without clinical characteristics suggestive of syndromic FNMTC. We obtained blood samples and clinical data from the affected individuals belonging to seven families with non-syndromic FNMTC, from seven hospitals in Spain (Fig. 1). Patients gave written informed consent before undergoing evaluation and testing.

Fig. 1
figure 1

Families with non-syndromic FNMTC recruited. Index cases are indicated with black arrows. Asterisk: Affected individuals where the WES was performed. Gray circles: Patients affected with thyroid cancer

Full size image

DNA Extraction and Gene Sequencing

Genomic DNA was extracted from peripheral blood samples using conventional salt-precipitation protocol. The whole coding region and 5′UTR regions of POT1 gene (NM_015450.2), as well as intron-exon boundaries, were studied.

PCR followed by Sanger sequencing was performed. PCR conditions were as follows: denaturation at 95 °C for 5 min, 10 cycles (95 °C for 1 min, 65–60 °C for 1 min, 72 °C for 1 min), followed by 25 cycles (95 °C for 1 min, 55 °C for 1 min, 72 °C for 1 min), and extension at 72 °C (10 min). Sanger sequencing was performed using universal M13 primers by GENEWIZ (Takeley, UK). Sequences were analyzed using the SeqPilot 4.0.1 software (JSI Medical Systems, Ettenheim, Germany). POT1 sequencing was performed in seven affected individuals (the index cases) from seven different pedigrees with FNMTC.

Exome Capture and Next-Generation Sequencing

Whole-exome sequencing (WES) was performed in 10 affected individuals (marked with an asterisk in Fig. 1) from the first four families we recruited (kindreds 1, 2, 3, and 4).

The DNA library was prepared using the SureSelect exon v5-post kit (Agilent Technologies, Santa Clara, CA, USA) that enables the capture of target sequence of exonic regions in the human genome. The libraries were sequenced using the Illumina HiSeq 2000 sequencer (Macrogene, Seoul, South Korea) with 101-base pair (bp) average read length.

Results

Families

Regarding the whole data (Table 1) for the 7 pedigrees (in total, 24 FNMTC-affected members), their median age at diagnosis was 41 years (range 23–61 years), 54% female. Twenty out of 24 (83%) patients presented classic papillary thyroid cancer (PTC), and 17% follicular variant PTC. Tumor multifocality was present in 46% of cases and not present in 37% (in the remaining 17%, we did not have this data). In total, 29% of tumors were bilateral (46% unilateral) and 37% of tumors presented gross extrathyroidal extension and local invasion. Lymph node metastasis was found in 10 cases (41.6%), and distant metastasis was present in 2 cases (8.3%). One of the patients with distant metastasis died of thyroid cancer. All patients received radioactive iodine therapy (range 30–450 mCi), and to date 29% presented recurrences and persistence of the disease.

Table 1 Clinical characteristics, pathological findings, and outcome of 7 families with FNMTC (members according to Fig. 1)
Full size table

POT1 Analysis

None of the seven index members from FNMTC was found to be a carrier of POT1 mutations that could be interpreted as causative of TC or other neoplasia. During the analysis, only the following well-known benign changes were found: rs6959712, rs6977407, rs7794637, rs3815221, and rs142780416. All of them are located in intronic regions.

Whole-Exome Sequencing

Confirming the previously described sequencing results, none of the 10 subjects studied by WES was found to be a carrier of germline POT1 mutations. In addition, we looked for mutations in other genes related to FNMTC. According to the literature, we examined genes responsible for syndromic FNMTC (APC, PTEN, WRN, DICER1, and PRKAR1A), rearranged genes (NTRK, PPARG), genes involved in DNA repair (XRCC1, XRCC3, ATM), or in the development of the thyroid gland (PAX8, JAG1, CDC42, GSTM1, GSTT1, SRGAP1, TERT, THRB, AKT1, SEC23B, ESR2, NKX2, TBL1X), genes previously described that may predispose to non-syndromic FNMTC (HABP2, TTF1, THADA, SEC23B, FOXE1, KLLN, MAP2K5, CHEK2, MYH9, SRRM2), proto-oncogenes (RET, MET, KIT, MERTK), and genes with somatic mutations in TC (BRAF, NRAS, TP53, CDKN2A, ALK, ATF4). We did not find any genetic variant that could be considered causative of thyroid cancer.

Discussion

Regarding the rising incidence of TC in general, and familial TC in particular, there is an unmet need to identify the genetic risk factors associated with this disease. Many studies have been carried out in order to identify susceptibility genes involved in non-syndromic FNMTC. Numerous reports have found private mutations in specific genes in isolated families, suggesting them to be causative of FNMTC [22], but none of them has been validated across different research articles [23].

TC and melanoma are in a way genetically connected cancers since they share some features: both can be triggered by environmental factors such as radiation or ultraviolet exposure, suggesting that DNA repair genes may be involved in their pathophysiology [24]. Furthermore, the elevation of TC risk in patients with melanoma has been described [25], and both are observed in a syndrome such as PTEN hamartomatous tumor syndrome (Cowden disease). Interestingly, like melanoma, TC often harbors somatic promotor mutations in the TERT gene [26] or the p.Val600Glu mutation in the BRAF gene, both being important events in TC progression [27]. Moreover, it has been suggested that genes involved in telomere maintenance that predispose to develop melanomas may also be implicated in TC, as it has been described that patients with FNMTC have shorter telomeres, compared with unaffected family members, sporadic cases, and healthy controls [28]. He et al. [29] investigated the gene copy number and mRNA expression of different genes involved in telomere maintenance (POT1 gene included) in 13 patients from six families with FNMTC observing no significant differences. Cantara et al. [15] studied POT1 and RAP1 genes by DHPLC in 66 patients from 38 families, implying that the majority of families had only two affected cases. Although they did not detect mutations in POT1, we consider it was not strong enough to rule out the implication of POT1 as a gene implied in hereditary TC because the analysis was done by DHPLC and the study was carried out mostly in families with only two cases. Given the high prevalence of TC in the general population, Charkes [30] estimated that about 62% of families with two cases affected by FNMTC can be phenocopies (two sporadic cases associated by chance); therefore, only 38% would be truly hereditary. However, if there are three affected cases, the probability of being hereditary rises to 96%. Consequently, we designed this multicentric study to analyze in selected FNMTC families (with 3 or more affected patients/family, and without any feature that made us suspect a well-known syndromic FNMTC) if the POT1 gene could be a putative susceptibility gene for non-syndromic FNMTC. Our enrolled families are representative of non-syndromic FNMTC, with a very high probability of being hereditary, not only because of the important number of affected members, but also because they present similar clinical characteristics, as described before for other families with FNMTC, especially those with 3 or more affected members [31]: early age at diagnosis, lower female predominance, more tumor multifocality, more lymph node metastasis, and tumor aggressiveness and recurrence.

We performed WES analysis in four families and Sanger sequencing in the seven families. WES does not cover all the regions of POT1 gene (e.g., 5′UTR regions or large intron-exon boundaries). On the other hand, WES analysis provided us complementary information, since it allowed us to rule out the presence of pathogenic mutations in other genes previously described as implicated in FNMTC.

In our study, no germline potentially pathogenic mutations were detected in POT1 gene in seven families with FNMTC.

We did not find putative causative mutations in genes previously described as candidate genes for FNMTC in the four families studied by WES.

A limitation of the study is the relatively small number of families. However, it is difficult to recruit families with more than two members affected with NMTC, even with a multicentric collaboration. In fact, most scientific articles on this topic include a limited number of families analyzed, unless they include families with two affected members [15, 29].

In the multiple studies searching for causative genes of FNMTC in the literature, no reproducible results have been found. We hypothesize that FNMTC may be mainly a polygenic hereditary entity. This scenario would make it difficult to find a common link among the different affected families. Indeed, part of the families included in this study were included in a previous paper [32], where we suggest that non-syndromic FNMTC may be due to multiple mutations acting as risk alleles and modifier locus for FNMTC, in contrast with most hereditary cancers, in which only 2 or 3 high-penetrance causative genes have been described.

In conclusion, the absence of germline POT1 mutations in our set of TC families studied here does not completely exclude the possibility that POT1 mutations may be associated with an increased TC risk as part of the wide spectrum of POT1 cancers. However, our results minimize the chances of POT1 being a TC predisposition gene that would manifest itself in non-syndromic TC families.