Abstract
Anaplastic thyroid cancer (ATC) is a highly malignant disease with a very short median survival time. Few studies have addressed the underlying somatic mutations, and the genomic landscape of ATC thus remains largely unknown. In the present study, we have ascertained copy number aberrations, gene fusions, gene expression patterns, and mutations in early-passage cells from ten newly established ATC cell lines using single nucleotide polymorphism (SNP) array analysis, RNA sequencing and whole exome sequencing. The ATC cell line genomes were highly complex and displayed signs of replicative stress and genomic instability, including massive aneuploidy and frequent breakpoints in the centromeric regions and in fragile sites. Loss of heterozygosity involving whole chromosomes was common, but there were no signs of previous near-haploidisation events or chromothripsis. A total of 21 fusion genes were detected, including six predicted in-frame fusions; none were recurrent. Global gene expression analysis showed 661 genes to be differentially expressed between ATC and papillary thyroid cancer cell lines, with pathway enrichment analyses showing downregulation of TP53 signalling as well as cell adhesion molecules in ATC. Besides previously known driver events, such as mutations in BRAF, NRAS, TP53 and the TERT promoter, we identified PTPRD and NEGR1 as putative novel target genes in ATC, based on deletions in six and four cell lines, respectively; the latter gene also carried a somatic mutation in one cell line. Taken together, our data provide novel insights into the tumourigenesis of ATC and may be used to identify new therapeutic targets.
Introduction
Anaplastic thyroid cancer (ATC) is a rare and lethal disease, accounting for only 1–5% of all thyroid malignancies and yet being responsible for approximately 50% of deaths attributed to thyroid cancer (Kebebew et al. 2005). Patients presenting with this extremely aggressive tumour have a very poor prognosis, with a median survival of just 4–12 months from the time of diagnosis, compared with the high cure rates and long-term survival rates for patients diagnosed with a well-differentiated thyroid cancer, such as follicular thyroid cancer (FTC) or papillary thyroid cancer (PTC) (Are & Shaha 2006). ATC is characterised by rapid tumour growth and frequent metastasis, with half of patients presenting with distant metastases (Ain 1999).
The genetic basis of PTC and FTC has been quite well studied, showing frequent mutations in RAS genes, including HRAS, KRAS and NRAS. PTC also frequently display BRAF mutations and fusions involving RET, whereas FTC commonly harbour PAX8–PPARG rearrangements (Kimura et al. 2003, Nikiforova et al. 2003, Soares et al. 2003, The Cancer Genome Atlas Research Network 2014). However, few investigations have so far focused on the genomic landscape of ATC. Array comparative genome hybridisation studies have shown a high degree of structural and numerical chromosomal aberrations, with gains in chromosomal copy number being observed more frequently than losses (Wreesmann et al. 2002, Miura et al. 2003, Lee et al. 2007). Mutations are commonly seen in TP53 (30–70% of cases), BRAF (30–45%), NRAS (15–20%), USH2A (18%) and EIF1AX (10–15%), as well as in the TERT promoter (50–70%) (Liu et al. 2013, Nikiforova et al. 2013, Kunstman et al. 2015, Jeon et al. 2016, Landa et al. 2016). To date, there have been few reports of fusion genes in ATC, with only three cases with RET fusions, two cases with ALK fusions and single cases with NUTM1/BRD4, SS18/SLC5A11, MKRN1/BRAF, and FGFR2/OGDH fusions reported in the literature (Liu et al. 2008, Kelly et al. 2014, Godbert et al. 2015, Kasaian et al. 2015, Landa et al. 2016), with no large-scale screening using RNA sequencing being reported.
In the present study, we investigated the genetic landscape of ATC cell lines using SNP array analysis, RNA sequencing and whole exome sequencing (WES). We found that ATC cell lines display highly complex genomes, with multiple breakpoints and large variation in copy numbers. We identified not previously implicated genes that are targeted by recurrent deletions and mutations. In addition, we found several novel in-frame gene fusions that could result in translated protein products affecting the development of ATC. Our results increase our understanding of the aetiology of this extremely aggressive disease and may be used to identify new therapeutic targets.
Materials and methods
Samples
A total of 13 newly established, low-passage thyroid cancer cell lines were included in the study (Wennerberg et al. 2014, Gretarsson et al. 2016). Ten cell lines established from patients diagnosed with ATC were ATC1 (LU-TC-1), ATC2 (LU-TC-2), ATC7 (LU-TC-7), ATC8 (LU-TC-8), ATC10 (LU-TC-10), ATC12 (LU-TC-12), ATC14 (LU-TC-14), ATC15 (LU-TC-15), ATC17 (LU-TC-17) and ATC18 (LU-TC-18), and three cell lines established from patients diagnosed with PTC were PTC4 (LU-TC-4), PTC5 (LU-TC-5) and PTC13 (LU-TC-13). For the establishment of the cell lines, patients referred to the Departments of Oncology, ENT/H&N Surgery or Endocrine Surgery at the University Hospital in Lund, Sweden, for treatment of previously untreated ATC or PTC were asked for participation in the study. Tissue sampling was performed with a conventional fine-needle aspiration technique using a 0.6–0.7 mm needle. The aspirates were directly transferred to RPMI 1640 medium with stable glutamine, supplemented with 1 mmol/L sodium pyruvate, 1× MEM non-essential amino acids, 20 µg/mL gentamicin and 10% foetal bovine serum (FBS) gold (GE Healthcare). The suspensions were immediately transferred to cell culture flasks and left to attach and grow at 37°C under a humidified atmosphere with 5% CO2. The cells were sequentially transferred to a new flask until visibly free from fibroblasts. Cytogenetic analysis of the cell lines showed polyploid and highly complex karyotypes in all cases. RNA and DNA were extracted from the cell lines using the RNeasy Mini Kit and Gentra Puregene Cell Kit (Qiagen), respectively, according to the manufacturer’s instructions. Two matched peripheral blood samples for ATC17 and ATC18 and three normal thyroid tissue samples, collected from patients undergoing a routine thyroidectomy, were also included in the study. DNA was extracted from the peripheral blood samples using the Gentra Puregene Blood Kit (Qiagen). For the tissue samples, sections of tissue measuring no more than 5 mm across one dimension were immediately transferred to a vial containing either RNAlater RNA Stabilization Solution or Allprotect Tissue Reagent (Qiagen) and stored overnight at 4°C, before removal from solution and long-term storage at −80°C. RNA was extracted from the tissue samples using the RNeasy Lipid Mini Kit (Qiagen). The study was approved by the Ethical Review Board of Lund University, reference number 522/2008, and informed consent was provided according to the Declaration of Helsinki.
SNP array analysis
SNP array analysis of all ATC cell lines except ATC17 and ATC18 and all three PTC cell lines was performed using the Illumina HumanOmni5-Quad BeadChip platform, containing ~5 million markers (Illumina), according to the manufacturer’s instructions. SNP array analysis of ATC17 and ATC18 and their matched normal control samples was performed with the Affymetrix Genome-Wide SNP Array 6.0 (Affymetrix) according to the manufacturer’s instructions. Probe positions were extracted from the GRCh37 genome build and data was analysed using the Genome studio v2011.1 (Illumina) and Nexus Copy Number 7.5 (BioDiscovery, El Segundo, CA, USA) software. For ATC17 and ATC18, constitutional copy number variants were excluded based on comparison with the matched control sample. For the remaining cases, all copy number changes <1 Mb were compared with copy number polymorphisms listed in the Database of Genomic Variants (http://dgv.tcag.ca/dgv/app/home) or Nexus and excluded from further analysis if there was substantial overlap. Regions displaying loss of heterozygosity (LOH) were included as aberrant if they comprised at least 5 Mb or were part of another rearrangement.
RNA sequencing
RNA sequencing was performed for all ATC cell lines, except ATC17 and ATC18, and all PTC cell lines. Paired-end RNA libraries were constructed using the TruSeq RNA Sample Preparation Kit v2 (Illumina) and sequenced on the Illumina HiSeq2000 platform according the manufacturer’s instructions by BGI Tech Solutions (Hong Kong) (Supplementary Table 1, see section on supplementary data given at the end of this article). The data has been submitted to the Gene Expression Omnibus database (GEO; https://www.ncbi.nlm.nih.gov/geo) and is available under accession number GSE94465.
Fusion gene analysis
Potential fusion genes were identified from the RNA-sequencing data using SOAPfuse version 1.26 (http://soap.genomics.org.cn/soapfuse.html), Chimerascan version 0.4.5 (http://code.google.com/p/chimerascan) and TopHat version 2.0.7 (http://ccb.jhu.edu/software/tophat/fusion_index.html). The GRCh37 build was used as the human reference genome. Potential fusion genes for further investigation were identified using a filtering pipeline based on spanning/total read number and previous identification of SNP events. Briefly, the output list of fusion genes was first filtered to remove chimeras identified as read-through transcripts before filtering to generate two lists of potential fusion genes containing (1) chimeras with one or more spanning reads, and (2) chimeras with five or more total reads. The lists were then both filtered to remove pseudogenes, unannotated genes and fusions between gene family members. Finally, chimeras containing gene partners that were less than 100 kb apart in distance were discarded unless an imbalance was detected by SNP array analysis in one or both of the fusion partners. Also, the C15orf57–CBX3 fusion, seen in three samples (ATC2, ATC12 and PTC4), was excluded from the results as it has been shown to result from a retrotransposition insertion in a substantial proportion of individuals (Schrider et al. 2013). To validate potential fusion genes, RT-PCR was performed in the corresponding cell line and three normal thyroid tissue samples. Briefly, cDNA was generated from 2.5 µg of RNA, and primers (available on request) specific to potential fusion transcripts were designed using Primer 3 (http://primer3.ut.ee/). PCR was performed according to standard methods and amplified products were sequenced using the BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems). All cell lines were subsequently screened with RT-PCR for fusion genes that were validated in this way.
Gene expression analysis
Gene expression levels based on RNA sequencing for all ATC cell lines except ATC17 and ATC18, and all PTC cell lines, were estimated using the TCGA UNC V2 RNA-Seq Workflow (https://webshare.bioinf.unc.edu/public/mRNAseq_TCGA/UNC_mRNAseq_summary.pdf). The expected read counts in the ‘.genes.results’ files generated by RSEM were used as the input for DESeq2 (Love et al. 2014), limma–voom (Law et al. 2014) and edgeR (Robinson et al. 2010), respectively. Differentially expressed genes were defined as those with a fold change >2 and FDR <0.01. Functional enrichment of differentially expressed genes in the KEGG pathways was performed by KOBAS (version 2.0) software (Xie et al. 2011), and KEGG pathways with P value <0.01 were considered significantly enriched.
Whole exome sequencing
WES was performed on all ten ATC cell lines and matched normal blood for ATC17 and ATC18 to an average read depth of ~200× (Supplementary Table 2). DNA libraries were constructed using the V5 50M Exon Kit (Agilent Technologies) and sequenced on the Complete Genomics sequencing platform according the manufacturer’s instructions by BGI Tech Solutions (Hong Kong). Raw read data were aligned using Teramap and variants were identified using a pipeline designed and implemented by BGI Tech Solutions (Hong Kong). The eight ATC cell lines without a matched normal blood sample were filtered against genetic variations reported in the 1000 Genomes Project, 6500 Exome Project, the Single Nucleotide Polymorphism Database (dbSNP) Build 129, and RefSeq: NCBI Reference Sequence Database and then evaluated only for known gene mutations previously reported in ATC and PTC. Somatic gene mutations were identified in cell lines ATC17 and ATC18 using a pipeline designed and implemented by BGI Tech Solutions (Hong Kong). The mutational signatures in ATC17 and ATC18 were investigated using MutSigCV (Lawrence et al. 2013). The data has been submitted to the Sequence Read Archive (https://trace.ncbi.nlm.nih.gov/Traces/sra/sra.cgi?view=announcement) and is available under accession number SRP098778.
Analysis of TERT promoter mutations and validation of mutations
Mutations in the BRAF gene were validated using a standard PCR and primers designed for exon 15 as previously reported (Davidsson et al. 2008): BRAFEx15 For: TGCTTGCTCTGATAGGAAAATGAG and BRAFEx15 Rev: TCTCAGGGCCAAAAATTTAATCA. Mutations in the promoter region of the TERT gene were validated using a standard PCR as previously reported by Liu and coworkers (Liu et al. 2013) using the AmpliTaq Gold 360 Master Mix (Thermo Fisher Scientific). Validation of mutations detected by whole exome sequencing in the genes TP53, NRAS, PTEN, PIK3CA, NEGR1 and EIF1AX was done using Sanger sequencing according to standard methods (primers available upon request).
Fluorescence in situ hybridisation (FISH) analysis
In order to investigate whether the NEGR1 deletions were associated with chromosomal rearrangements, metaphase FISH analysis was performed according to standard methods in cell lines ATC7, ATC8, ATC10 and ATC15. Probes used were the bacterial artificial chromosomes RP11-115M14 (distal of NEGR1), RP11-82L20 (in 3′ part of NEGR1; includes exon 7), RP11-566G9 (in intron 1 of NEGR1) and RP11-292O17 (proximal of NEGR1), obtained from the BACPAC Resource Center (https://bacpacresources.org/).
RT-PCR for NEGR1
RT-PCR to detect NEGR1 transcripts were done using primers in exons 1a and 7 in all ten ATC cell lines, the three PTC cell lines, three normal thyroid tissue samples, and the Human Total RNA Master Panel II (Clontech Laboratories), containing RNA from twenty normal tissues. Complementary DNA was produced from 2.5 µg RNA using M-MLV reverse transcriptase (Invitrogen) and random hexamers (Invitrogen). The primers (R_exon 1: GATGGTCAGAAAAGGGGACA, F_exon 7: TCTCCTGCTGGTACCTTGTG, Life Technologies, Invitrogen) for NEGR1 were designed using Primer3 software (http://bioinfo.ut.ee/primer3/). PCR was done according to standard methods and amplified fragments were cut out from the gel and sequenced using the BigDye v1.1 Cycle Sequencing Kit (Applied Biosystems) on an ABI-3130 Genetic Analyzer (Applied Biosystems). Chromas Lite 2.1.1, free online software (http://technelysium.com.au/), was used for the analysis of the NEGR1 sequence data. In addition, RNA-Seq data was screened for junctions between exons and compared with the RT-PCR data.
Results
ATC cell lines have complex genomes and display overt signs of chromosomal instability
The SNP array analysis of ten ATC cell lines revealed highly complex genomes with multiple breakpoints and large variations in copy number within each case (Fig. 1A and B). The median number of breakpoints, defined as a change in copy number state, was 49.5 (range 9–167) (Table 1; Supplementary Table 3). There was no evidence of chromothripsis. Many of the identified breakpoints (36/601; 6.0%) were located in the centromeric regions. Breaks were most common in the centromeric regions of chromosomes 8 (5/10 cell lines; 50%), 3, 4 and 9 (4/10; 40% each), and 1, 2 and 17 (3/10; 30% each). ATC10 had the highest number of centromere breaks (n = 9), while ATC 1, 2 and 7 only had one each.
Another common feature of the ATC cell lines was LOH involving whole chromosomes (wcLOH), with 5/10 cell lines having one or more wcLOH. Chromosomes 13 (4/10 cases; 40%), 18 (4/10 cases; 40%) and 17 (3/10 cases; 30%) were most commonly affected (Fig. 1C; Supplementary Table 4). However, wcLOH was only seen for a minority of chromosomes in any given case; thus, no case displayed evidence of duplication of a near-haploid stem line (Fig. 1C).
Many breakpoints and massive aneuploidy in ATC cell lines. (A) Breakpoint map of ATC cell lines. Breakpoints were defined as a change in copy number. A large number of breakpoints in centromeric regions were seen. (B) Heatmap of copy number changes in ATC cell lines. All cases displayed massive aneuploidy. (C) Overview of regions displaying loss of heterozygosity (LOH). Note that cell lines ATC2, ATC12 and ATC15 were established from male patients; hence the LOH.
Citation: Endocrine-Related Cancer 24, 5; 10.1530/ERC-16-0522
Summary of breakpoints in ten anaplastic thyroid cancer cell lines.
| Cell line | Number of centromeric breakpoints* (%) | Total number of breakpoints* |
|---|---|---|
| ATC1 | 1 (1.0) | 105 |
| ATC2 | 1 (5.6) | 18 |
| ATC7 | 1 (11) | 9 |
| ATC8 | 6 (7.6) | 79 |
| ATC10 | 9 (5.4) | 167 |
| ATC12 | 4 (9.1) | 44 |
| ATC14 | 3 (8.1) | 37 |
| ATC15 | 7 (13) | 55 |
| ATC17 | 2 (6.7) | 30 |
| ATC18 | 2 (3.5) | 57 |
As determined by a change in copy number.
There were a total of 157 focal (<1 Mb) regions of copy number gain and loss among the ten cell lines (median 12.5; range 2–40), with the most common type of alteration being hemizygous loss (n = 123), followed by homozygous loss (n = 14) (Supplementary Table 4). Regions of focal copy number gains were rare (n = 20), with most of the gained regions affecting more than one gene; none of these regions were recurrent (Supplementary Table 4). There were no recurrent amplifications or regions of high copy number gain.
ATC cell lines harbour multiple out-of-frame fusion genes
A total of 21 fusion genes were found and validated among six ATC cell lines (Table 2; Supplementary Fig. 1). Six (29%) of these were expected to produce an in-frame transcript resulting in a protein product (Table 2). In addition, three fusion genes, including two in-frame chimeras, were seen in the three PTC cell lines (Table 2). Of the validated fusion genes, ten (42%) were identified by only one of the fusion transcript identifier programmes, whereas five (21%) were identified by all three programmes (Table 2; Supplementary Fig. 2).
Validated fusion genes identified in the ATC and PTC cell lines.
| Cell line | Fusion gene | Chromosome | Inframe/frame-shift | Partner gene previously reported in fusion gene* | Software identifying fusion |
|---|---|---|---|---|---|
| ATC1 | MYBL1/VCPIP1 | 8q13.1/8q13 | Frame-shift | MYBL1 - Diffuse large B-cell lymphoma - Astrocytoma, pilocytic/juvenile (brain) |
SoapFUSE, Chimerascan |
| ATC1 | PPP6R2/CACNA1A | 22q13.33/19p13 | Frame-shift | SoapFUSE, TopHat | |
| ATC1 | SPOP/TBX21 | 17q21.33/17q21.32 | Frame-shift | SoapFUSE, Chimerascan | |
| ATC1 | AP2A2/TALDO1 | 11p15.5/11p15.5-p15.4 | Frame-shift | AP2A2 - Acute myeloid leukaemia, NOS |
SoapFUSE |
| ATC1 | DDAH1/ZNHIT6 | 1p22/1p22.3 | Frame-shift | SoapFUSE, Chimerascan, TopHat | |
| ATC1 | PEX14/KIF1B | 1p36.22/1p36.2 | Frame-shift | Chimerascan | |
| ATC1 | PIAS3/XPR1 | 1q21/1q25.1 | Inframe | SoapFUSE, Chimerascan, TopHat | |
| ATC1 | PLAU/SEC24C | 10q22.2/10q22.2 | Frame-shift | Chimerascan | |
| ATC2 | KIF20B/PRKG1 | 10q23.31/10q11.2 | Frame-shift | PRKG1 - Adenocarcinoma (lung) |
SoapFUSE, Chimerascan, TopHat |
| ATC2 | ATAD1/SOX5 | 10q23.31/12p12.1 | Frame-shift | SoapFUSE, Chimerascan | |
| ATC7 | RGS17/HBS1L | 6q25.3/6q23.3 | Frame-shift | RGS17 - Adenocarcinoma (breast) |
SoapFUSE, Chimerascan, TopHat |
| ATC8 | TANGO6/CDH13 | 16q22.1/16q23.3 | Inframe | CDH13 - Adenocarcinoma (breast) - Adenocarcinoma (lung) |
Chimerascan, TopHat |
| ATC8 | ZNF76/PPARD | 6p21.31/6p21.2 | Frame-shift | SoapFUSE, Chimerascan | |
| ATC8 | TMCC1/PTPRG | 3q22.1/3p21-p14 | Inframe | TMCC1 - Squamous cell carcinoma (lung) -Adenocarcinoma (breast)PTPRG - Adenocarcinoma (breast) |
Chimerascan |
| ATC10 | TTYH3/BRAT1 | 7p22/7p22.3 | Frame-shift | SoapFUSE, Chimerascan | |
| ATC10 | MMS19/SLIT1 | 10q24-q25/10q23.3-q24 | Frame-shift | Chimerascan | |
| ATC14 | KDM4B/PLIN4 | 19p13.3/19p13.3 | Inframe | SoapFUSE, Chimerascan | |
| ATC14 | ZFP14/BIRC2 | 19q13.12/11q22 | Inframe | SoapFUSE | |
| ATC14 | SAFB/GPI | 19p13.3-p13.2/19q13.1 | Frame-shift | SoapFUSE | |
| ATC14 | TMEM241/RPS12 | 18q11.2/6q23.2 | Frame-shift | TMEM241 - Adenocarcinoma (breast) |
SoapFUSE |
| ATC14 | DYNC2H1/CASP12 | 11q21-q22.1/11q22.3 | Inframe | CASP12 - Squamous cell carcinoma (lung) |
Chimerascan |
| PTC4 | MERTK/TG | 2q14.1/8q24 | Inframe | TG - Adenocarcinoma (thyroid) |
SoapFUSE, Chimerascan, TopHat |
| PTC4 | TOX4/TERT | 14q11.2/5p15.33 | Frame-shift | TERT - Astrocytoma, grade III-IV (brain) - Clear cell carcinoma (kidney) - Chronic lymphocytic leukaemia - Acute lymphoblastic leukaemia/lymphoblastic lymphoma |
SoapFUSE, Chimerascan |
| PTC13 | STIM1/PGAP2 | 11p15.5/11p15.5 | Inframe | STIM1 - Adenocarcinoma (large intestine) - Acute myeloid leukaemia, NOS |
Chimerascan |
Mitelman Database of Chromosome Aberrations and Gene Fusions in Cancer (2016). Mitelman F, Johansson B and Mertens F (Eds.), http://cgap.nci.nih.gov/Chromosomes/Mitelman.
ATC, anaplastic thyroid cancer; PTC, papillary thyroid cancer.
Multiple target genes identified in ATC cell lines
The SNP array analysis revealed several target genes in ATC cell lines, based on small targeted deletions or breakpoints involving only one gene in more than one case. The deletions comprised PTPRD in 6/10 cell lines (60%); NEGR1, AUTS2 and FHIT in 4/10 (40%) each; CDKN2A and MACROD2 in 3/10 (30%) each; and CNTNAP2, DLG2, IMMP2L, LRP1B, METTL15 and NAALADL2 in 2/10 (20%) each (Fig. 2; Supplementary Table 4). Most of these deletions were hemizygous, but homozygous deletions were also observed (Supplementary Table 2). In addition, breakpoints were recurrently seen in the genes PRKG1 (3/10 cell lines; 30%) and CDH13, DDAH1, DIAPH2, DYNC2H1, EYS, LRRTM4, MAGI2, SPAG9 and WWOX (2/10; 20% each) (Fig. 2; Supplementary Table 3).
Genomic profile of the ten ATC cell lines. The panels display the genetic analysis performed on the ATC cell lines, the somatic gene mutations, gene deletions and breakpoints detected in the ATC cell lines, and fusion genes identified and validated by RT-PCR in the ATC cell lines.
Citation: Endocrine-Related Cancer 24, 5; 10.1530/ERC-16-0522
Mutations
TP53 mutations were seen in six (60%) ATC cell lines, BRAF p.V600E mutations in four (40%), NRAS mutations in two (20%), and an EIF1AX mutation in one (10%) cell line (Fig. 2; Supplementary Tables 5 and 6). Mutations in the promoter region of the TERT gene were identified in 6/10 (60%) ATC cell lines; all were C228T mutations. All mutations in the above-mentioned genes were validated by Sanger sequencing (data not shown). For ATC17 and ATC18, where constitutional blood samples were available for identification of somatic mutations, the whole exome was analysed, showing a total of 71 somatic mutations in coding regions in ATC17 and 137 in ATC18 (Supplementary Table 6). ATC17 did not show a clear mutational signature, whereas ATC18 displayed predominant mutations in TpCpN trinucleotides, indicating the involvement of the APOBEC family of cytidine deaminases (Supplementary Fig. 3).
TP53 signalling and cell adhesion molecules are downregulated in ATC cell lines
Gene expression analysis of the RNA-sequencing data identified 661 genes differentially expressed between the ATC cell lines and PTC cell lines (Q value <0.01); 421 genes were downregulated in ATC and 240 genes were upregulated (Supplementary Table 7). Pathway enrichment analysis of the differentially expressed genes identified five significant pathways (P < 0.01): the TP53 signalling (KEGG hsa04115), the axon guidance (KEGG hsa04360), the cell adhesion molecules (CAMs) (KEGG hsa04514) and the autoimmune thyroid disease (KEGG hsa05320) pathways were downregulated, whereas the haematopoietic cell lineage pathway (KEGG hsa04640) was upregulated.
NEGR1 is a putative tumour suppressor gene involved in a high proportion of ATC cell lines
SNP array analysis showed that interstitial deletions in the NEGR1 gene were present in 4/10 (40%) ATC cell lines. These deletions frequently occurred in a stepwise fashion (from a baseline of 3 to 5 copies of chromosome arm 1p), involving different parts of the gene, but included exons 2 and 3 in 3/4 cell lines (Fig. 3). The deletions were hemizygous (i.e., at least one unaffected allele was still present) in all cell lines except ATC10, in which exons 2 and 3 were homozygously deleted. No rearrangements involving NEGR1 were detected with metaphase FISH (Supplementary Table 8). In addition, ATC17 harboured a somatic p.V156I mutation in NEGR1. This mutation occurred in a position between two of the Ig-like domains, but we could not predict its effects since the structure of the NEGR1 protein has not been determined. There are three normal splice variants of NEGR1: NEGR1_001, including exons 1a–7; transcript NEGR1_002, including exons 1b–7; and NEGR1_201, including exons 1a–3 and 5–7. RNA-Seq did not show any overall difference in the expression level of NEGR1 between ATC and PTC, but it showed the complete absence of expression of alternative exon 1b in all cell lines, excluding the expression of NEGR1_002. To further delineate the expression of NEGR1, RT-PCR was performed across the gene. Apart from the full-length transcript NEGR1_001, which was present in all cell lines except ATC8 and ATC10, and splice variant NEGR1_201, which was present in ATC2, ATC12 and ATC15, this analysis showed multiple extra transcripts in all cell lines with deletions (Fig. 3C and Supplementary Fig. 4A) as well as in ATC1 and ATC2 that lacked deletions detectable by SNP array analysis. Sanger sequencing showed that all these extra transcripts were the results of aberrant splicing of full exons, resulting in both in-frame and out-of-frame transcripts; at least one out-of-frame transcript predicted to lead to a truncated protein was present in all cases with deletion, as well as in ATC1 and ATC2 (Fig. 3C). To check whether these extra transcripts could be normal splice variants of NEGR1, we performed RT-PCR using the same primers in the three PTC cell lines, three samples from normal thyroid glands, and a tissue panel containing mRNA from 20 different tissues. The full-length NEGR1_001 transcript was seen in all investigated tissues and splice variant NEGR1_002 in prostate and foetal brain; however, no other transcripts were detected (Supplementary Fig. 4B). Taken together, our data indicate that focal deletions result in aberrant splicing of NEGR1 in ATC, possibly leading to lower levels of normal NEGR1 protein.
Deletions at the NEGR1 locus at 1p31.1 in anaplastic thyroid cancer cell lines. (A) SNP array data from ATC8. LogR ratios are shown on the Y axis; each dot corresponds to one marker. Stepwise deletions are seen. (B) Deletions in the cell lines ATC8, ATC10, ATC15 and ATC18. Hemizygous deletions are shown in green; homozygous deletions in blue. (C) Splice variants of NEGR1 detected in ATC cell lines. Two normal splice variants (NEGR1-001 and NEGR1-201 in GRCh37.p13, www.ensembl.org) are shown in blue. Of these, NEGR1-001 contains exons 1–7 and is detected in all cell lines except ATC8 and ATC10, whereas NEGR1-201 contains exons 1–3 and 5–7, and is seen in cell lines ATC2, ATC12 and ATC15. Aberrant splice variants are shown in red (out-of-frame) and green (in-frame). Cell lines harbouring the splice variants as detected by RT-PCR and RNA sequencing are shown to the left.
Citation: Endocrine-Related Cancer 24, 5; 10.1530/ERC-16-0522
Discussion
ATC has a close to 100% short-time mortality and there is no curative treatment for patients with disseminated disease (Kebebew et al. 2005, Are & Shaha 2006). Thus, there is an urgent need to identify possible targets for treatment. Considering that ATC is one of the most malignant diseases, surprisingly little is known about its underlying genomic landscape. Most studies have included small cohorts and have been focused on only one type of genetic aberration, such as copy number changes or sequence mutations. In the present study, we performed a full genomic characterisation, comprising parallel SNP array analyses, RNA sequencing and WES, to delineate all driver events in early-passage cells from ten new ATC cell lines. A potential problem with working with cell lines is that the cells may have acquired changes during culture that were not present in the primary tumour. Although we cannot exclude that some of the aberrations detected here were acquired during culture, we think that the majority were present in the primary tumour, based on (1) the cells analysed here were from early passages (passages 3–5), (2) the frequencies of well-known mutations, such as NRAS, BRAF and TP53 mutations as well as TERT promoter mutations, in our cohort closely resemble reports of primary ATC, and (3) several of the potential new driver events identified here, in particular PTPRD and NEGR1 deletions, were seen in a high proportion of the investigated cell lines. Thus, the abnormalities reported here are likely to be largely representative of primary ATC.
All ten investigated cell lines were polyploid and showed complex copy number aberration patterns, with frequent stepwise deletions and gains (Figs 1 and 3) and massive aneuploidy. Deletions were frequently seen at known fragile sites, indicating replicative stress (Bignell et al. 2010). There were a high number of breakpoints (median 49.5; range 9–167) in all cases as ascertained by a change in copy number; these numbers are likely an underestimate since balanced rearrangements will not be detected. Notably, the breakpoints were frequently in pericentromeric regions (Fig. 1A), similar to what has been reported previously in, e.g., colorectal cancer and head and neck squamous cell carcinomas (Stewenius et al. 2005). Such breaks could either result from telomere shortening leading to breakage-fusion-bridge cycles, which has been suggested to preferentially induce pericentromeric breaks, or from mitotic spindle defects leading to centromere shearing; either way, they are associated with chromosomal instability (Stewenius et al. 2005, Guerrero et al. 2010). There is no data in the literature on telomere lengths in ATC, but considering that TERT promoter mutations, leading to telomerase expression and presumably preventing telomere shortening, are common (found in 60% of the cell lines investigated here) it is likely that the underlying cause of the pericentromeric breaks in ATC cell lines is a malfunctioning mitotic spindle. Corver and coworkers (Corver et al. 2012) reported that massive loss of chromosomes, i.e., near-haploidisation, was common in FTC. Although LOH across whole chromosomes and large regions was frequently seen in the ten ATC cell lines, it always involved a minority of the chromosomes and was hence not indicative of a previous near-haploid step; we can thus exclude that this mechanism is common in ATC. Thus, on the chromosomal scale, the overall pattern of copy number changes and breakpoints detected in this study suggest that ATC exhibits replicative stress as well as genomic instability resulting from mitotic spindle defects. Considering that PTC generally appear to be relatively genomically stable (The Cancer Genome Atlas Research Network 2014), the induction of genomic instability could be one of the major factors leading to the development of ATC from a prior, more differentiated tumour.
Six of the eight investigated ATC cell lines harboured at least one fusion gene, with ATC1 displaying eight different chimeric genes. None of the fusion genes were recurrent and none have previously been described in ATC, although twelve of the fusion partner genes identified have previously been reported as fused to other partner genes in various solid tumours and haematological malignancies (Mitelman Database of Chromosome Aberrations and Gene Fusions in Cancer (2016). Mitelman F, Johansson B and Mertens F (Eds.), http://cgap.nci.nih.gov/Chromosomes/Mitelman). No cell line harboured any RET fusions, ALK fusions, or NUTM1/BRD4, SS18/SLC5A11, MKRN1/BRAF or FGFR2/OGDH, which has previously been reported to occur at a low frequency in ATC (Liu et al. 2008, Kelly et al. 2014, Godbert et al. 2015, Kasaian et al. 2015, Landa et al. 2016). Thus, our data suggest that there is no common fusion gene that drives ATC development. The high number of fusion genes detected in this study could be due to the genomic instability that generates many chromosomal breaks and some of them could thus be passenger events, in particular the out-of-frame fusions. However, considering that in-frame fusion genes are frequently potent driver events, a subset of the fusions identified here are likely to be involved in tumourigenesis.
In contrast to the fusion genes, there were several regions of recurrent focal deletions involving only one gene (Fig. 2; Supplementary Table 4), including the well-known tumour suppressor genes PTPRD (60%), FHIT (40%) and CDKN2A (30%). Several reports have shown previously that FHIT and CDKN2A are involved in thyroid tumourigenesis (Chang et al. 1998, Elisei et al. 1998, Zou et al. 1999, Lee et al. 2008), whereas the involvement of PTPRD in ATC is a novel finding. PTPRD encodes a receptor protein tyrosine phosphatase that is deleted in a wide range of tumours, such as neuroblastoma and glioblastoma; these deletions are believed to lead to STAT3 hyperactivation and thereby promote tumourigenesis (Molenaar et al. 2012, Ortiz et al. 2014). Another commonly targeted gene was NEGR1 in 1p31.1, which displayed intragenic deletions in four cases and a non-synonymous mutation in an additional case; thus, in total 50% of the investigated cell lines had somatic aberrations in this gene. NEGR1, also known as KILON, encodes a glycosylphosphatidylinositol (GPI)-anchored membrane protein involved in cell adhesion (Funatsu et al. 1999, Kim et al. 2014). It functions in neuronal development (Pischedda et al. 2014, Sanz et al. 2015), and constitutional variants affecting the expression of this gene have consistently been associated with obesity in genome-wide association studies (Willer et al. 2009, Wheeler et al. 2013). However, NEGR1 is expressed in most tissues (www.genecards.org and Supplementary Fig. 4B). NEGR1 has previously been reported to be somatically deleted in neuroblastoma and frequently downregulated in cancer (Takita et al. 2011, Kim et al. 2014). Knockdown of this gene was shown to promote cell migration and invasion in an ovarian adenocarcinoma cell line, suggesting that it may function as a tumour suppressor gene (Kim et al. 2014). Although we could not detect an overall lower expression of NEGR1 in cell lines with deletions compared with cell lines lacking such deletions, two cell lines (ATC8 and ATC10) lacked the full-length transcript (Supplementary Fig. 4A). In addition, RT-PCR across the gene showed multiple abnormal transcripts in all cell lines with deletions as well as in ATC1 and ATC2, which did not have any visible NEGR1 deletion as ascertained by SNP array analysis (Fig. 1 and Supplementary Fig. 3). These transcripts arose from aberrant splicing of whole exons and were both in-frame and out-of-frame (Fig. 3C). Thus, although the overall expression of NEGR1 may not be lowered by the deletions, they may lead to lesser amounts of functional protein. In contrast, the three PTC cell lines without deletions and a panel of 20 normal tissues, including thyroid gland, only showed full-length NEGR1 transcripts (Supplementary Fig. 4). Taken together, our data suggest that deletion of NEGR1, leading to aberrant splicing, may be a frequent driver event in ATC. Considering that NEGR1 is believed to function in cell adhesion, these deletions may increase invasiveness and thereby promote metastasis.
Global gene expression analyses based on the RNA-sequencing data showed similar results to a recent investigation by Kasaian and coworkers (Kasaian et al. 2015), with downregulation of TP53-regulated genes and cell adhesion molecules in ATC cell lines. The former is likely associated with the high frequency of inactivating TP53 mutations in this cohort, whereas the latter agree well with the high metastatic potential of ATC.
Taken together, we present copy number, transcriptional and mutational data from ten newly established ATC cell lines, likely to closely resemble primary ATC in terms of the genomic landscape. We found a high degree of genomic complexity and overt signs of genomic instability that likely contribute to the malignant phenotype of these tumours. No common fusion gene was found. Besides confirming that mutations in TP53, NRAS, BRAF and the TERT promoter are common, our findings suggest that PTPRD and NEGR1 deletions frequently may drive tumourigenesis in ATC.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-16-0522.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by the Swedish Research Council (grant number 521-2012-864); the Swedish Cancer Fund (grant number 14 0546), Åke Wibergs Stiftelse (grant number M14-0066), the Royal Physiographic Society of Lund (grant number 151111 KP), the King Gustaf V Jubilee Fund, the Foundations of the University Hospital of Lund, the Gunnar Nilsson Cancer Foundation, the Berta Kamprad Foundation for Investigation, Control of Cancer Diseases, the Laryngology Fund, Region of Scania R&D funding, and governmental funding of clinical research within the NHS.
Author contribution statement
K P, J W and L E conceived the study; E L W, A B, N R, M Y, L E and J W analysed the data; E L W, A B and K P wrote the manuscript, which was approved by all co-authors.
References
Ain KB 1999 Anaplastic thyroid carcinoma: a therapeutic challenge. Seminars in Surgical Oncology 16 64–69. (doi:10.1002/(SICI)1098-2388(199901/02)16:1<64::AID-SSU10>3.0.CO;2-U)
Are C & Shaha AR 2006 Anaplastic thyroid carcinoma: biology, pathogenesis, prognostic factors, and treatment approaches. Annals of Surgical Oncology 13 453–464. (doi:10.1245/ASO.2006.05.042)
Bignell GR, Greenman CD, Davies H, Butler AP, Edkins S, Andrews JM, Buck G, Chen L, Beare D & Latimer C et al. 2010 Signatures of mutation and selection in the cancer genome. Nature 463 893–898. (doi:10.1038/nature08768)
Chang TJ, Tsai TC, Wu YL, Yang HM, Chi CW, Yang AH & Lee CH 1998 Abnormal transcripts of FHIT gene in thyroid cancer. Oncology Reports 5 245–247. (doi:10.3892/or.5.1.245)
Corver WE, Ruano D, Weijers K, den Hartog WC, van Nieuwenhuizen MP, de Miranda N, van Eijk R, Middeldorp A, Jordanova ES & Oosting J et al. 2012 Genome haploidisation with chromosome 7 retention in oncocytic follicular thyroid carcinoma. PLoS ONE 7 e38287. (doi:10.1371/journal.pone.0038287)
Davidsson J, Lilljebjorn H, Panagopoulos I, Fioretos T & Johansson B 2008 BRAF mutations are very rare in B- and T-cell pediatric acute lymphoblastic leukemias. Leukemia 22 1619–1621. (doi:10.1038/leu.2008.14)
Elisei R, Shiohara M, Koeffler HP & Fagin JA 1998 Genetic and epigenetic alterations of the cyclin-dependent kinase inhibitors p15INK4b and p16INK4a in human thyroid carcinoma cell lines and primary thyroid carcinomas. Cancer 83 2185–2193. (doi:10.1002/(SICI)1097-0142(19981115)83:10<2185::AID-CNCR18>3.0.CO;2-4)
Funatsu N, Miyata S, Kumanogoh H, Shigeta M, Hamada K, Endo Y, Sokawa Y & Maekawa S 1999 Characterization of a novel rat brain glycosylphosphatidylinositol-anchored protein (Kilon), a member of the IgLON cell adhesion molecule family. Journal of Biological Chemistry 274 8224–8230. (doi:10.1074/jbc.274.12.8224)
Godbert Y, Henriques de Figueiredo B, Bonichon F, Chibon F, Hostein I, Perot G, Dupin C, Daubech A, Belleannee G & Gros A et al. 2015 Remarkable response to crizotinib in woman with anaplastic lymphoma kinase-rearranged anaplastic thyroid carcinoma. Journal of Clinical Oncology 33 e84–e87. (doi:10.1200/JCO.2013.49.6596)
- ↑
Gretarsson S, Ekblad L, Kjellén E, Jin Y, Paulsson K, Borg Å, Brun E, Tennvall J, Bergenfelz A & Wennerberg J 2016 Substantial intrinsic variability in chemo- and radio sensitivity in eight newly established anaplastic thyroid cancer cell-lines. Poster P-09, presented at 25th Meeting of the Scandinavian Society for Head and Neck Oncology (SSHNO), Reykjavik, Iceland, 26 May 2016. Aarhus, Denmark: Scandiavian Society for Head and Neck Oncology. (available at: http://www.sshno2016.is/)
Guerrero AA, Gamero MC, Trachana V, Futterer A, Pacios-Bras C, Diaz-Concha NP, Cigudosa JC, Martinez AC & van Wely KH 2010 Centromere-localized breaks indicate the generation of DNA damage by the mitotic spindle. PNAS 107 4159–4164. (doi:10.1073/pnas.0912143106)
Jeon M, Chun SM, Kim D, Kwon H, Jang EK, Kim TY, Kim WB, Shong Y, Jang SJ & Song DE et al. 2016 Genomic alterations of anaplastic thyroid carcinoma detected by targeted massive parallel sequencing in a BRAFV600E mutation-prevalent area. Thyroid 26 683–690. (doi:10.1089/thy.2015.0506)
Kasaian K, Wiseman SM, Walker BA, Schein JE, Zhao Y, Hirst M, Moore RA, Mungall AJ, Marra MA & Jones SJ 2015 The genomic and transcriptomic landscape of anaplastic thyroid cancer: implications for therapy. BMC Cancer 15 984. (doi:10.1186/s12885-015-1955-9)
Kebebew E, Greenspan FS, Clark OH, Woeber KA & McMillan A 2005 Anaplastic thyroid carcinoma – treatment outcome and prognostic factors. Cancer 103 1330–1335. (doi:10.1002/cncr.20936)
Kelly LM, Barila G, Liu P, Evdokimova VN, Trivedi S, Panebianco F, Gandhi M, Carty SE, Hodak SP & Luo J et al. 2014 Identification of the transforming STRN-ALK fusion as a potential therapeutic target in the aggressive forms of thyroid cancer. PNAS 111 4233–4238. (doi:10.1073/pnas.1321937111)
Kim H, Hwang JS, Lee B, Hong J & Lee S 2014 Newly identified cancer-associated role of human neuronal growth regulator 1 (NEGR1). Journal of Cancer 5 598–608. (doi:10.7150/jca.8052)
Kimura ET, Nikiforova MN, Zhu Z, Knauf JA, Nikiforov YE & Fagin JA 2003 High prevalence of BRAF mutations in thyroid cancer: genetic evidence for constitutive activation of the RET/PTC-RAS-BRAF signaling pathway in papillary thyroid carcinoma. Cancer Research 63 1454–1457.
Kunstman JW, Juhlin CC, Goh G, Brown TC, Stenman A, Healy JM, Rubinstein JC, Choi M, Kiss N & Nelson-Williams C et al. 2015 Characterization of the mutational landscape of anaplastic thyroid cancer via whole-exome sequencing. Human Molecular Genetics 24 2318–2329. (doi:10.1093/hmg/ddu749)
Landa I, Ibrahimpasic T, Boucai L, Sinha R, Knauf JA, Shah RH, Dogan S, Ricarte-Filho JC, Krishnamoorthy GP & Xu B et al. 2016 Genomic and transcriptomic hallmarks of poorly differentiated and anaplastic thyroid cancers. Journal of Clinical Investigation 126 1052–1066. (doi:10.1172/JCI85271)
Law CW, Chen Y, Shi W & Smyth GK 2014 Voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biology 15 R29. (doi:10.1186/gb-2014-15-2-r29)
Lawrence MS, Stojanov P, Polak P, Kryukov GV, Cibulskis K, Sivachenko A, Carter SL, Stewart C, Mermel CH & Roberts SA et al. 2013 Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 499 214–218. (doi:10.1038/nature12213)
Lee JJ, Foukakis T, Hashemi J, Grimelius L, Heldin NE, Wallin G, Rudduck C, Lui WO, Hoog A & Larsson C 2007 Molecular cytogenetic profiles of novel and established human anaplastic thyroid carcinoma models. Thyroid 17 289–301. (doi:10.1089/thy.2006.0246)
Lee JJ, Au AY, Foukakis T, Barbaro M, Kiss N, Clifton-Bligh R, Staaf J, Borg A, Delbridge L & Robinson BG et al. 2008 Array-CGH identifies cyclin D1 and UBCH10 amplicons in anaplastic thyroid carcinoma. Endocrine-Related Cancer 15 801–815. (doi:10.1677/ERC-08-0018)
Liu Z, Hou P, Ji M, Guan H, Studeman K, Jensen K, Vasko V, El-Naggar AK & Xing M 2008 Highly prevalent genetic alterations in receptor tyrosine kinases and phosphatidylinositol 3-kinase/akt and mitogen-activated protein kinase pathways in anaplastic and follicular thyroid cancers. Journal of Clinical Endocrinology and Metabolism 93 3106–3116. (doi:10.1210/jc.2008-0273)
Liu X, Bishop J, Shan Y, Pai S, Liu D, Murugan AK, Sun H, El-Naggar AK & Xing M 2013 Highly prevalent TERT promoter mutations in aggressive thyroid cancers. Endocrine-Related Cancer 20 603–610. (doi:10.1530/ERC-13-0210)
Love MI, Huber W & Anders S 2014 Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology 15 550. (doi:10.1186/s13059-014-0550-8)
Miura D, Wada N, Chin K, Magrane GG, Wong M, Duh QY & Clark OH 2003 Anaplastic thyroid cancer: cytogenetic patterns by comparative genomic hybridization. Thyroid 13 283–290. (doi:10.1089/105072503321582097)
Molenaar JJ, Koster J, Zwijnenburg DA, van Sluis P, Valentijn LJ, van der Ploeg I, Hamdi M, van Nes J, Westerman BA & van Arkel J et al. 2012 Sequencing of neuroblastoma identifies chromothripsis and defects in neuritogenesis genes. Nature 483 589–593. (doi:10.1038/nature10910)
Nikiforova MN, Kimura ET, Gandhi M, Biddinger PW, Knauf JA, Basolo F, Zhu Z, Giannini R, Salvatore G & Fusco A et al. 2003 BRAF mutations in thyroid tumors are restricted to papillary carcinomas and anaplastic or poorly differentiated carcinomas arising from papillary carcinomas. Journal of Clinical Endocrinology and Metabolism 88 5399–5404. (doi:10.1210/jc.2003-030838)
Nikiforova MN, Wald AI, Roy S, Durso MB & Nikiforov YE 2013 Targeted next-generation sequencing panel (ThyroSeq) for detection of mutations in thyroid cancer. Journal of Clinical Endocrinology and Metabolism 98 E1852–E1860. (doi:10.1210/jc.2013-2292)
Ortiz B, Fabius AW, Wu WH, Pedraza A, Brennan CW, Schultz N, Pitter KL, Bromberg JF, Huse JT & Holland EC et al. 2014 Loss of the tyrosine phosphatase PTPRD leads to aberrant STAT3 activation and promotes gliomagenesis. PNAS 111 8149–8154. (doi:10.1073/pnas.1401952111)
Pischedda F, Szczurkowska J, Cirnaru MD, Giesert F, Vezzoli E, Ueffing M, Sala C, Francolini M, Hauck SM & Cancedda L et al. 2014 A cell surface biotinylation assay to reveal membrane-associated neuronal cues: Negr1 regulates dendritic arborization. Molecular and Cellular Proteomics 13 733–748. (doi:10.1074/mcp.M113.031716)
Robinson MD, McCarthy DJ & Smyth GK 2010 edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26 139–140. (doi:10.1093/bioinformatics/btp616)
Sanz R, Ferraro GB & Fournier AE 2015 IgLON cell adhesion molecules are shed from the cell surface of cortical neurons to promote neuronal growth. Journal of Biological Chemistry 290 4330–4342. (doi:10.1074/jbc.M114.628438)
Schrider DR, Navarro FC, Galante PA, Parmigiani RB, Camargo AA, Hahn MW & de Souza SJ 2013 Gene copy-number polymorphism caused by retrotransposition in humans. PLoS Genetics 9 e1003242. (doi:10.1371/journal.pgen.1003242)
Soares P, Trovisco V, Rocha AS, Lima J, Castro P, Preto A, Maximo V, Botelho T, Seruca R & Sobrinho-Simoes M 2003 BRAF mutations and RET/PTC rearrangements are alternative events in the etiopathogenesis of PTC. Oncogene 22 4578–4580. (doi:10.1038/sj.onc.1206706)
Stewenius Y, Gorunova L, Jonson T, Larsson N, Hoglund M, Mandahl N, Mertens F, Mitelman F & Gisselsson D 2005 Structural and numerical chromosome changes in colon cancer develop through telomere-mediated anaphase bridges, not through mitotic multipolarity. PNAS 102 5541–5546. (doi:10.1073/pnas.0408454102)
Takita J, Chen Y, Okubo J, Sanada M, Adachi M, Ohki K, Nishimura R, Hanada R, Igarashi T & Hayashi Y et al. 2011 Aberrations of NEGR1 on 1p31 and MYEOV on 11q13 in neuroblastoma. Cancer Science 102 1645–1650. (doi:10.1111/j.1349-7006.2011.01995.x)
The Cancer Genome Atlas Research Network 2014 Integrated genomic characterization of papillary thyroid carcinoma. Cell 159 676–690. (doi:10.1016/j.cell.2014.09.050)
Wennerberg E, Pfefferle A, Ekblad L, Yoshimoto Y, Kremer V, Kaminskyy VO, Juhlin CC, Hoog A, Bodin I & Svjatoha V et al. 2014 Human anaplastic thyroid carcinoma cells are sensitive to NK cell-mediated lysis via ULBP2/5/6 and chemoattract NK cells. Clinical Cancer Research 20 5733–5744. (doi:10.1158/1078-0432.CCR-14-0291)
Wheeler E, Huang N, Bochukova EG, Keogh JM, Lindsay S, Garg S, Henning E, Blackburn H, Loos RJ & Wareham NJ et al. 2013 Genome-wide SNP and CNV analysis identifies common and low-frequency variants associated with severe early-onset obesity. Nature Genetics 45 513–517. (doi:10.1038/ng.2607)
Willer CJ, Speliotes EK, Loos RJ, Li S, Lindgren CM, Heid IM, Berndt SI, Elliott AL, Jackson AU & Lamina C et al. 2009 Six new loci associated with body mass index highlight a neuronal influence on body weight regulation. Nature Genetics 41 25–34. (doi:10.1038/ng.287)
Wreesmann VB, Ghossein RA, Patel SG, Harris CP, Schnaser EA, Shaha AR, Tuttle RM, Shah JP, Rao PH & Singh B 2002 Genome-wide appraisal of thyroid cancer progression. American Journal of Pathology 161 1549–1556. (doi:10.1016/S0002-9440(10)64433-1)
Xie C, Mao X, Huang J, Ding Y, Wu J, Dong S, Kong L, Gao G, Li CY & Wei L 2011 KOBAS 2.0: a web server for annotation and identification of enriched pathways and diseases. Nucleic Acids Research 39 W316–W322. (doi:10.1093/nar/gkr483)
Zou M, Shi Y, Farid NR, al-Sedairy ST & Paterson MC 1999 FHIT gene abnormalities in both benign and malignant thyroid tumours. European Journal of Cancer 35 467–472. (doi:10.1016/S0959-8049(98)00370-0)
