Introduction

Thyroid cancer is the most common endocrine cancer, with an increasing overall incidence in recent decades, although it only accounts for 1% of all malignant tumors. It is divided into several types and histological subtypes according to the cells from which the tumor derives, with different characteristics and prognoses. Only 2–5% of cases of thyroid cancer are derived from parafollicular cells and this type is called medullary thyroid cancer (MTC) [1, 2] and occurs in both familial (ORPHA:99,361) and sporadic forms (ORPHA:1332). Approximately 20% of patients develop distant metastases, which makes MTC an aggressive and rare cancer incurable nowadays as this cancer does not respond to radiotherapy or chemotherapy (13.4% all thyroid cancer mortality) [3]. About 75% of all MTCs are believed to be sporadic (sMTC), whereas the remaining 25% correspond to inherited cancer syndromes known as Multiple Endocrine Neoplasia type 2 (MEN2). MEN2 includes 3 clinically differentiable types: MEN2A (ORPHA:247,698), MEN2B (ORPHA:247,709) and familial thyroid cancer (FMTC, ORPHA:99,361) [4, 5], which present MTC as a common feature and have been defined based on presence or absence of hyperparathyroidism, pheocromocytoma and other additional characteristic clinical features. RET proto-oncogene alterations play a crucial role for thyroid cancer development [3]. Different mechanisms of RET activation, gene rearrangements and point mutations characterize MTC, representing a very strong factor for its poor prognosis [6, 7]. More than 100 gain-of-function RET mutations have been reported in patients with MTC, including germline mutations (in patients MEN2 with hereditary disease) and somatic mutations (in patients with sporadic disease) [8]. More than 95% of MEN2 cases have germline mutations in exons 5, 7, 8, 10, 11, 13, 14, 15 and 16 of the RET proto-oncogene, which lead to a gain of function of the receptor. In the case of MEN2A, 98% of patients have mutations grouped within a hot-spot that corresponds to five cysteine codons present in the extracellular domain of the protein (codons 609, 611, 618, 620 and 634) [9]. Approximately 87% of MEN2A mutations affect to codon 634 of RET, and the p.Cys634Arg mutation has been found in more than 50% of cases [9]. In Spanish MEN2A patients, the p.Cys634Tyr mutation is more prevalent, which suggests a founder effect [10,11,12]. Biochemical studies on mutated proteins in cysteine codons indicate that these mutations lead to a constitutive activation of the metabolic pathways of RET signaling [13]. Regarding MEN2B, 95% of patients have a single mutation in exon 16 (p.Met918Thyr) that causes a conformational change in the intracellular tyrosine-kinase 2 binding pocket and allows for constitutive kinase activation in the absence of dimerization, as well as altered substrate binding. However, in about 5% of familial cases with “MEN2-like” presentation patients, the genetic cause of the disease is unknown.

Nowadays, MTC patients have limited treatment options for metastatic disease. Therefore, the exploration of new mechanisms implicated in the onset of thyroid cancer as well as new therapeutic targets is crucial to improve treatment. This perspective makes epigenetics an interesting area of research [14]. In this sense, many epigenetic processes have been implicated in thyroid tumorigenesis of MTC, such as ncRNA deregulation, especially microRNAs [15, 16], although it is difficult to identify a good biomarker of the disease [16]. Regarding long non coding RNA (lncRNA), there are two studies describing them in association with sMTC [17, 18]. However, no studies have been performed on MEN2 syndrome patients and lncRNAs. The broad term lncRNA indicates a class of non-coding RNA transcript of minimum 200 nucleotides in length, bigger than miRNAs with about 21–25 nucleotides in length. They have gained broad attention in recent years as new players in transcriptional, epigenetic, or post-transcriptional regulation of gene expression [19].

The aim of the present study is based on previous results from our group [18]. In that study, performed in tumoral and non-tumoral paired fresh frozen tissues from sMTC patients, and we have detected some lncRNAs already associated with thyroid cancer (GAS5, MALAT1, MEG3, PTCSC1, PTCSC3, H19), while some new were described (ZFAS1, RMST, SNHG16, FTX, IPW, ADAMTS9-AS2, PRNCR1 and RMRP). Thus, we aimed to amplify the cohort of patients, to validate such results. It is worthy to mention that ZFAS1 and SNHG16 were linked to papillary thyroid cancer after our publication [20,21,22] and that is the reason why we did not continue performing any additional study with them.

Therefore, unlike other previous assays, we have analyzed different lncRNA on both sMTC and MEN2 patients to find the link among their altered expression and their role on such manifestations, for the first time.

Materials and methods

Patients and tissue samples

From thirty-one MTC patients undergoing surgical resection, including both twenty-one MEN2 (twenty MEN2A and one MEN2B) and ten sMTC patients, medullary thyroid tumor tissues (FFPE, formalin fixed paraffin embedded tissues) and their corresponding adjacent non-tumor thyroid tissues were obtained. All the clinical data are compiled in Additional file 1: Supplementary Table 1.

Table 1 Aberrant lncRNAs in human sMTC and MEN2 FFPE tissues: data obtained by qRT-PCR (7500HT Taqman system) with RT2Sybr®Green RoxTMqPCR Mastermix using the cDNA obtained from the RNA isolated from the FFPE tissues of our cohort of patients
Full size table

RNA extraction from FFPE tissues and cDNA synthesis

RNA was extracted from all FFPE tissue specimens from MTC and MEN2 patients. Ten slides (ten µm each) were manipulated with the MasterPure™ Complete DNA and RNA Purification Kit (Lucigen) following manufacturer’s instructions. Briefly, we placed 30 mg of 10 µm thick paraffin sections in a tube with Proteinase K and Tissue and Cell Lysis Solution to mix. After 30 min at 65 °C (vortex and ice), we added MPC Protein Precipitation Reagent to pellet the debris by centrifugation (4 °C, 10 min, 10,000×g). The supernatant was mixed with isopropanol and rinsed twice with 70% ethanol to finally resuspend the total nucleic acids in TE Buffer. The contaminating DNA was removal using the DNAse I solution, MPC Protein Precipitation Reagent and different centrifugations. The resultant pellet with the purified RNA was resuspended with TE Buffer.

The RNA was quantified by Nanodrop (Invitrogen) and 1 μg of total RNA was reverse transcribed into cDNA using PrimeScript RT Reagent Kit (TaKaRa). Finally, RT2Sybr®Green RoxTMqPCR Mastermix (Qiagen) was used to determine lncRNA expression levels, using Beta Actin as reference gene.

qRT-PCR

All the reactions were carried out in triplicate. All the data were analyzed by Applied Biosystems software and the relative expression levels of lncRNAs were determined by the Equation  2−ΔΔCt. The qRT-PCR was performed at 7500 Fast Real Time PCR System (Applied Biosystems). Primers used for FTX, IPW and RMST were the same used into SYBR®Green qPCR assays (RT2 lncRNA PCR Array, Qiagen) in the previous study. For RMRP and Beta-Actin, the primers used were KiCqStart® SYBR® Green Primers. All primer sequences are available under request.

Annotation analysis was performed using DAVID Bioinformatics Resources v6.8 online tools (http://david.abcc.ncifcrf.gov/).

Statistics

Student’s t test was used to analyze lncRNA expression data collected from qRT-PCR.

Two-tailed t-test was used to analyze differences between tumor tissues and their corresponding adjacent non-tumor thyroid tissues. A P < 0.05 was considered as a statistically significant difference. Data are expressed as means ± standard error of the mean (SEM).

Results

Expression of lncRNAs analyzed

A total of seven lncRNAs were selected to be further studied in the current study: four of them were up-regulated (RMST or rhabdomyosarcoma 2 associated transcript), FTX or FTX transcript and XIST regulator), IPW or imprinted in Prader-Willi syndrome) while other three lncRNAs were down-regulated (PRNCR1 or Prostate Cancer Associated Non-Coding RNA 1, ADAMTS9-AS2 or Antisense RNA 2 and RMRP or RNA component of mitochondrial RNA processing endoribonuclease). Unfortunately, neither from PRNCR1 nor ADAMTS9-AS2 expression was detected in FFPE samples, and thus they were eliminated from the study. Then, the lncRNAs finally analyzed were RMST, FTX, IPW and RMRP.

We used thirty-one formalin-fixed paraffin embedded (FFPE) tissues due to the difficulty to obtain fresh frozen tissue from both sMTC and MEN2 patients. The specimens included twenty-one patients with develop MTC due to a MEN2 syndrome and ten with sMTC (Additional file1: Supplementary Table 1). After analyzing the selected lncRNAs by qRT-PCR, we obtained a significant fold change (or log2 ratio) for the up-regulated lncRNAs: RMST (log2 = 11,359), FTX (log2 = 6,281), IPW (log2 = 14,616) and for the down-regulated RMRP (log2 = 0,338) (Table 1 and Fig. 1), which fits with the results obtained in our previous study performed on fresh sMTC tissue.

The functional annotation analysis of these lncRNA was performed based on their gene ontology categories (GOs) (Table 2). In either case, there is still little information about these lncRNAs to date [20,21,22,23,24,25].

Figure 1
figure 1

The expression levels of IPW, RMRP, RMST and FTX both in MEN2 and sMTC: This assay was performed by qRT-PCR in medullary thyroid tumor tissues (formalin-fixed paraffin embedded, FFPE) and their corresponding adjacent non-tumor thyroid tissues. Data represent the mean ± SEM from three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. N= normal; sMTC=sporadic medullary thyroid cancer; MEN2= Multiple Endocrine Neoplasia type 2.

Full size image
Table 2 Gene ontology (GO) functional annotation analysis: annotation analysis was performed using DAVID Bioinformatics Resources v6.8 online tools (http://david.abcc.ncifcrf.gov/)
Full size table

Discussion

Upregulated and downregulated lncRNAs, as post-transcriptional regulators of genetic expression, in different histological types of thyroid cancer, have been described [3, 6, 26]. Particularly, in MTC very few studies have been performed in this emerging field, despite the relevance of using lncRNAs as biomarkers in thyroid cancer diagnosis and prognosis, as well as their potential association with common genetic changes associated with thyroid cancer [17, 18]. As we previously mentioned, one study associates MALAT1 as pro-oncogenic lncRNA in sMTC and the other one, published last year by our group, detected some dysregulated lncRNAs in sMTC patients (RMST, FTX, IPW, PRNCR1, ADAMTS9-AS2 and RMRP). In addition, any study has linked lncRNAs with MEN2 syndrome to date. Thus, to our knowledge, this is the first study analyzing different lncRNAs (RMST, FTX, IPW and RMRP) and its role in MTC development, either in sMTC and MEN2 syndrome patients.

Briefly, RMST has been recently linked to cancer by its modulation of DNA methylation [27]. In addition, it has been described in hepatocellular [28] and endometrial cancers [29]. It functions in neurogenesis by helping in the association of Sox2 transcription factor to its target promoters (provided by RefSeq, Dec 2017). Some members of the SOX family are overexpressed in thyroid cancer [30]. In addition, SOX2 has been linked to Central Nervous System tumours as glioblastoma [31]. Such kind of tumours have been detected in MEN2A patients [32].

In this study, we have detected a significant upregulation of 3.7 fold (MEN2) and 2.1 fold (sMTC) in comparison with the normal tissue of such patients. These outcomes would indicate that maybe RMST is more implicated in familial cases than in sporadic ones from MTC. Thus, based on the literature and in our own study, a potential role for RMST in MTC pathogenesis, especially in familial cases, should be further analyzed.

RMRP encodes the RNA component of mitochondrial RNA processing endoribonuclease, which cleaves mitochondrial RNA at a priming site of mitochondrial DNA replication (provided by RefSeq, Mar 2010). It has been described that miR-675 directly targets MAPK1, and inhibits the oncogenicity of thyroid cancer while it is sponged by RMRP [33]. Moreover, mutations in RMRP have been associated with a spectrum of autosomal recessive skeletal dysplasias classified as cartilage-hair hypoplasia (CHH, ORPHA:175), whose patients may develop adult-onset immunodeficiency or malignancy such as thyroid cancer [34]. In the current work, we have found a significant downregulation of RMRP of a 63.49% (MEN2) and 69.75% (sMTC), which give us an indication that this molecule could be implicated in MTC with a little more incidence in sporadic cases. Then, together with those previously mentioned studies, our results would complement its potential association into medullary thyroid cancer development.

LncRNA FTX was firstly identified in XIST gene locus and was dysregulated in many human cancers [15, 35, 36]. FTX is located upstream of XIST, within the X-inactivation center, and produces a spliced lncRNA which positively regulate the expression of XIST, which is essential for initiation and spread of X-inactivation (provided by RefSeq, May 2015). It has been described that XIST was significantly upregulated in thyroid cancer [37], through promotion of cell proliferation and invasion [23]. We found here that FTX is significantly upregulated (6.2 fold in MEN2 and 6.8 fold in sMTC), although there are little differences in the expression pattern of this lncRNA in both type of patients. Then, if FTX is upregulated, then XIST will be also upregulated, potentially promoting the onset or develop of this pathology.

IPW is a non-protein coding gene exclusively expressed from the paternal allele, maybe playing a role in the imprinting process. Mutations in this gene are associated with Prader-Willi syndrome (PWS; ORPHA:739) [provided by RefSeq, May 2010]. Its overexpression in the critical region of the PWS locus, which regulates the DLK1-DIO3 region, resulted in chromatin modifications, leading to a subset of PWS phenotypes [21]. Subjects with PWS need to regularly control their thyroid function, because they present metabolic and endocrine complications, such as hypothyroidism, which would reinforce the implication of IPW into thyroid cancer [38, 39]. It is important to highlight that in this study we have detected a significant upregulation of IPW either in MEN2 (13.5 fold) and sMTC (18.4 fold). Then, IPW seems to have a role into MTC and especially in sporadic cases.

One of the major limitations of our study is the use of FFPE tissues instead of fresh tissue, because rare tumours are rarely available. That was the reason why in our previous study we only can account with four fresh tissue of sMTC patients. Another limitation is that, at this point, we have only confirmed the group of lncRNAs through performing a qRT-PCR. Thus, different further studies should be made to validate the obtained results, such as in situ hybridization to detect differences of lncRNA expression among normal and tumoral adjacent part of the tumour, studies in cell lines of MTC, such as TT cells (invasion, proliferation, silencing/overexpression of lncRNA target, apoptosis, cell cycle assays …) [8, 40].

In summary and being conscious that further functional analyses are needed, we propose RMRP and RMST as potential candidates for a deeper knowledge in MEN2 patients and FTX and IPW in sMTC.

Considering our results and all the information previously published, this preliminary study has allowed us to identify four potential lncRNAs as suitable novel biomarkers for these rare diseases.

Conclusions

Very few studies have been performed in the rising field of lncRNAs and medullary thyroid cancer. Based on previous results, we have validated four lncRNAs as potential diagnostic biomarkers in medullary thyroid carcinoma. Further future molecular analyses will be required to deep into their exact role in the pathology, focusing on their clinical validation and viability for thyroid cancer therapy in those patients who fail conventional therapy.