ORIGINAL ARTICLE

CLIP2 as radiation biomarker in papillary thyroid carcinomaM Selmansberger1, A Feuchtinger2, L Zurnadzhy3, A Michna1, JC Kaiser4, M Abend5, A Brenner6, T Bogdanova3, A Walch2, K Unger1,7,H Zitzelsberger1,7 and J Hess1,7

A substantial increase in papillary thyroid carcinoma (PTC) among children exposed to the radioiodine fallout has been one of themain consequences of the Chernobyl reactor accident. Recently, the investigation of PTCs from a cohort of young patients exposedto the post-Chernobyl radioiodine fallout at very young age and a matched nonexposed control group revealed a radiation-specificDNA copy number gain on chromosomal band 7q11.23 and the radiation-associated mRNA overexpression of CLIP2. In this study,we investigated the potential role of CLIP2 as a radiation marker to be used for the individual classification of PTCs into CLIP2-positive and -negative cases—a prerequisite for the integration of CLIP2 into epidemiological modelling of the risk of radiation-induced PTC. We were able to validate the radiation-associated CLIP2 overexpression at the protein level by immunohistochemistry(IHC) followed by relative quantification using digital image analysis software (P= 0.0149). Furthermore, we developed astandardized workflow for the determination of CLIP2-positive and -negative cases that combines visual CLIP2 IHC scoring andCLIP2 genomic copy number status. In addition to the discovery cohort (n= 33), two independent validation cohorts of PTCs(n= 115) were investigated. High sensitivity and specificity rates for all three investigated cohorts were obtained, demonstratingrobustness of the developed workflow. To analyse the function of CLIP2 in radiation-associated PTC, the CLIP2 gene regulatorynetwork was reconstructed using global mRNA expression data from PTC patient samples. The genes comprising the firstneighbourhood of CLIP2 (BAG2, CHST3, KIF3C, NEURL1, PPIL3 and RGS4) suggest the involvement of CLIP2 in the fundamentalcarcinogenic processes including apoptosis, mitogen-activated protein kinase signalling and genomic instability. In our study,we successfully developed and independently validated a workflow for the typing of PTC clinical samples into CLIP2-positive andCLIP2-negative and provided first insights into the CLIP2 interactome in the context of radiation-associated PTC.

Oncogene advance online publication, 6 October 2014; doi:10.1038/onc.2014.311

INTRODUCTIONOne of the major consequences of the Chernobyl nuclear accidentin 1986 has been a significant increase in the incidence ofpapillary thyroid carcinomas (PTCs) among children exposed tothe radioiodine fallout, particularly to iodine-131.1 To date, PTChas developed in 44000 individuals who were children oradolescents at the time of exposure.2 Thus, young age at exposureis a significant risk factor for the development of radiation-induced PTC.3 In order to delineate radiation-associated effects, itis crucial that the tumour cohorts of exposed and nonexposedcases are matched on age and other factors as closely as possible.4

This approach was enabled by the Chernobyl Tissue Bank(CTB, www.chernobyltissuebank.com) that systematically collectsthyroid tumour tissue samples from residents who lived in thecontaminated regions of Ukraine and the Russian Federation atthe time of the accident. The CTB collection also includes asubstantial number of thyroid tumours from nonexposed patients.Several studies evidence that radiation exposure can induce copynumber alterations and deregulation of gene expressions with thepotential of triggering carcinogenic processes, both of which wasobserved even at low doses of radiation.5–7 Common genetic

alterations in PTC are point mutations of the BRAF gene (V600E)and various variants of RET gene rearrangements, all of which leadto a constitutive activation of the mitogen-activated proteinkinase (MAPK) pathway.8 The frequency of RET/PTC rearrange-ments (that is, RET/PTC1 and RET/PTC3) was associated withradiation exposure levels in a study of the atomic bomb survivorsand some but not all studies of post-Chernobyl PTC.9–11 However,RET/PTC3 rearrangements have also been observed with similarfrequencies in sporadic PTCs from young patients, indicating arelation with young age of PTC onset.11,12 A more recentpublication by Ricarte-Filho et al.13 reported a higher frequencyof fusion oncogenes in radiation-induced PTCs, including rare TRKand BRAF rearrangements and the recently discovered ETV6–NTRK3 kinase fusion oncogene. These promising findings requirefurther validation in independent cohorts. We recently reported aradiation-specific DNA copy number gain on the chromosomalband 7q11.23 and a radiation-associated mRNA overexpression ofthe gene CLIP2, located on chromosome 7q11.23.4 Based on thesefindings, we aimed to investigate CLIP2 as a radiation biomarker inPTC. However, one of the crucial requirements for a biomarker tobe used in epidemiological studies and risk modelling is its validityin terms of sensitivity, specificity, reproducibility and biological

1Research Unit Radiation Cytogenetics, Helmholtz Zentrum München, German Research Center for Environmental Health GmbH, Neuherberg, Germany; 2Research Unit AnalyticalPathology, Helmholtz Zentrum München, German Research Center for Environmental Health GmbH, Neuherberg, Germany; 3Institute of Endocrinology and Metabolism, NationalAcademy of Medical Sciences of the Ukraine, Kiev, Ukraine; 4Institute of Radiation Protection, Helmholtz Zentrum München, German Research Center for Environmental HealthGmbH, Neuherberg, Germany; 5Bundeswehr Institute of Radiobiology, Munich, Germany and 6Radiation Epidemiology Branch, Division of Cancer Epidemiology and Genetics,NCI, Bethesda, MD, USA. Correspondence: Dr J Hess, Research Unit Radiation Cytogenetics, Helmholtz Zentrum München, German Research Center for Environmental Health,Ingolstädter Landstr. 1, Neuherberg 85764, Germany.E-mail: julia.hess@helmholtz-muenchen.de7Clinical Cooperation Group ‘Personalized Radiotherapy of Head and Neck Cancer’, Helmholtz Zentrum München, German Research Center for Environmental Health GmbH,Neuherberg, Germany.Received 21 March 2014; revised 16 July 2014; accepted 9 August 2014

Oncogene (2014), 1–9© 2014 Macmillan Publishers Limited All rights reserved 0950-9232/14

www.nature.com/onc

plausibility. A further aspect for the integration into molecularepidemiology is the suitability of both a highly specific marker andan appropriate assay for its detection.14 In addition, a biomarkermay provide insights into the mechanisms of the disease it is asurrogate for. Therefore, this study also aimed to test whether aCLIP2 radiation-specific mRNA overexpression persists at theprotein level and to establish a reproducible workflow that allowsa classification of PTCs from patients with unknown radiationhistory. Moreover, we intended to reconstruct the CLIP2interactome from global transcriptome data that were derivedfrom radiation-induced PTCs in order to clarify the role of CLIP2 inradiation-associated PTC carcinogenesis.15

RESULTSCLIP2 protein expression is elevated in PTCs from patientsexposed to radioiodine fallout compared with a nonexposedcontrol groupIn order to evaluate the expression of CLIP2 at the protein levelin exposed and nonexposed cases of the Genrisk-T cohort,immunohistochemical staining with an antibody against CLIP2was performed in a highly standardized manner, followed bydigital image analysis using the Definiens software (Definiens AG,Munich, Germany). The obtained average marker stainingintensities within the analysed tumour regions (regions of interest)are listed in Supplementary Table 1. Statistical testing revealedsignificantly increased staining intensities (Mann–Whitney test,P-value = 0.0149) in PTC tissues from the exposed group ascompared with the nonexposed group (Figure 1). Representativeimmunohistochemically stained formalin-fixed, paraffin-embedded(FFPE) tumour sections from nonexposed and exposed patientsare shown in Figure 1, illustrating a radiation-associated over-expression of the CLIP2 protein.

Establishment of a workflow for an individual classification of theCLIP2 expressionWe simplified and standardized the whole procedure of CLIP2protein expression measurement and established an approach forscoring and classification of individual cases by visual inspectionof digital immunohistochemistry (IHC) images. Subsequently, wecompared the results with the above-mentioned software-basedapproach. For this purpose, the automated stained FFPE tissue

sections were scanned and the generated images imported intothe image viewer software Panoramic Viewer (3DHISTECH,Budapest, Hungary). The staining intensities were visually scoredas follows: negative staining (score 0), weak staining (score 1),intermediate staining (score 2) and strong staining (score 3), asdemonstrated in Figure 2. Detailed scoring criteria are outlined inthe Materials and methods section. The entire classificationworkflow is illustrated in Figure 3. The resulting visual scores forthe 33 cases of the Genrisk-T cohort are shown in Table 1. Astatistically significant difference (two-sided Fisher’s exact test,P-value = 0.005) between the exposed and nonexposed group wasalso revealed using the visual scoring approach. Correlationanalysis of the data from both evaluation approaches showed astrong correlation between digital analysis and visual scoring(Spearman’s correlation coefficients: 0.83, P-value: 1.63 × 10− 6,Figure 4). In order to obtain an individual biomarker classificationfor each case, cases with a visual score of 0 or 1 were consideredas CLIP2 biomarker negative, whereas cases with visual score 3were classified as CLIP2 biomarker positive. For a classification ofan intermediate CLIP2 staining (score 2), the genomic copynumber status (gained or not gained) of chromosomal band7q11.23 (localization of CLIP2) was taken into account. Preferably,data from array comparative genomic hybridization (array CGH)were used. However, if array CGH data were not available,interphase fluorescence in situ hybridization analysis with a probespecific for 7q11.23 was performed. Cases with a visual score of 2and a DNA gain of chromosomal band 7q11.23 were finallyclassified as CLIP2-positive, and those showing normal copynumber of chromosomal band 7q11.23 were classified as CLIP2-negative (Figure 3).

Analysis of CLIP2 as a radiation biomarker in three independenttumour cohortsIn order to validate CLIP2 as a radiation biomarker in PTCs ofyoung patients, two independent tumour cohorts (Genrisk-T-PLUSand UkrAm) comprising in total 115 Ukrainian PTC cases wereanalysed in addition to the discovery set Genrisk-T. Genrisk-T-PLUS, which is an extension of the initial Genrisk-T cohort,included exposed and nonexposed cases, whereas the set of PTCcases from the UkrAm study included PTCs from exposed patientsonly.16 The investigated cohorts are summarized in Table 2. AllPTC cases were individually scored for CLIP2 expression using our

Figure 1. Immunohistochemical staining of FFPE tumor sections from patients not exposed and exposed to radioiodine fallout using anantibody against CLIP2. A significantly increased CLIP2 expression (P= 0.0149) in exposed (n= 16, green boxplot) compared with nonexposedcases (n= 17, red boxplot) is shown at the protein level by IHC (middle). The marker staining intensities were evaluated by relativequantification using the Definiens software. The P-value was calculated using the Mann–Whitney test. Representative immunohistochemicallystained FFPE sections from PTCs from nonexposed (left side; cases from top to bottom: UA0312, UA1328 and UA1208) and exposed (right side;cases from top to bottom: UA0905, UA0648 and UA0501) patients are shown.

CLIP2 as radiation marker in PTCM Selmansberger et al

2

Oncogene (2014), 1 – 9 © 2014 Macmillan Publishers Limited

standardized visual scoring approach. Figure 5 summarizes theresulting CLIP2 scores and the final classification into CLIP2-positive and -negative cases for all three independent tumourcohorts. The results of single cases are listed in SupplementaryTable 1. Gain of the chromosomal band 7q11.23 in score-2 caseswas present in 0/9 Genrisk-T cases (0%), 4/11 Genrisk-T PLUS cases(36%) and 7/18 UkrAm cases (39%). In Supplementary Figure 1,array CGH profiles representing chromosome 7 from six UkrAmcases are exemplarily shown. Following the inclusion of thegenomic copy number status of chromosomal band 7q11.23, weobtained sensitivity rates of 75%, 75% and 72.4% for the Genrisk-T,Genrisk-T-PLUS and UkrAm cohorts, respectively. The specificityrates were 82.4 and 57.1% for the Genrisk-T and Genrisk-T-PLUScohorts (Figures 5d–f).

Association of CLIP2 protein expression with patient andhistological dataThe CLIP2 protein expression (IHC CLIP2 visual scores 0–3) was notassociated with age at exposure, sex, histological dominantpattern, presence of RET/PTC rearrangements or BRAF V600Emutations. CLIP2 protein expression was associated with exposureto radioiodine (P= 0.0015). Testing was applied to the mergedclinical data on all three cohorts.

CLIP2 gene regulatory network reconstructionThe global gene regulatory network based on global geneexpression data from 31 UkrAm cases consisted of 4746 nodes(genes) and 29 842 edges (interactions between genes).15 The firstneighbourhood network of CLIP2, which was extracted from theglobal gene regulatory network, consisted of the seven genesBAG2, CHST3, GOLM1, KIF3C, NEURL1, PPIL3 and RGS4 (Figure 6).Furthermore, four of the CLIP2 first neighbourhood genes wereassociated with each other (KIF3C with BAG2 and RGS4, BAG2 withRGS4 and NEURL1 with RGS4). All but one (interaction of GOLM1

with CLIP2) association of the network could be validated byquantitative reverse transcriptase–PCR (qRT–PCR; SupplementaryTable 2). The second neighbourhood (including the first neigh-bourhood, Supplementary Figure 2 and Supplementary Table 3) ofCLIP2 contained 218 nodes and 1304 edges, including the geneLMO3 that was shown to be associated with radiation dose in thestudy by Abend et al.15 Pathway enrichment analysis of thesecond neighbourhood revealed the significantly enrichedpathways signalling by Nodal, transport of mature transcript tocytoplasm, Ras activation upon Ca2 influx through NMDA receptorand immune system.

DISCUSSIONWe continued and extended our previous study describing thegenomic radiation marker 7q11.23 that was exclusively detectedin PTCs from patients exposed to the Chernobyl radioiodinefallout and the radiation-associated CLIP2 overexpression at themRNA level.4 Here, we confirmed CLIP2 overexpression in PTCsfrom radiation-exposed patients at the protein level. Moreover, wedeveloped a generally applicable workflow for the classification ofPTCs into CLIP2-positive or -negative cases and proposed CLIP2 asa surrogate biomarker for radiation exposure in PTCs, allowing forsubsequent integration of CLIP2 into epidemiological studies andthyroid cancer risk models. Furthermore, we reconstructed theCLIP2 interactome from published global mRNA expression dataof the UkrAm cohort.15

We were able to confirm the previously demonstratedradiation-associated CLIP2 mRNA expression at the protein levelin the same discovery cohort of cases (Figure 1). Within thiscohort, the CLIP2 protein overexpression was present in 7q11.23gained cases. The observed variation of CLIP2 protein expressionin cases with normal copy number on chromosome 7q11.23 canbe explained by epigenetic regulations of CLIP2 expression, forexample, by miRNA miR-16-5p (http://mirtarbase.mbc.nctu.edu.

Figure 2. Visual scoring of immunohistochemically stained FFPE tumour sections using an antibody against CLIP2. The marker stainingintensities were evaluated by visual scoring 0–3. Two representative immunohistochemically stained papillary thyroid carcinoma cases withscores of 0, 1, 2 and 3, respectively, are shown from the left to the right. Image details of A-a and B-a (black frames) are shown below in A-band B-b, respectively.

CLIP2 as radiation marker in PTCM Selmansberger et al

3

© 2014 Macmillan Publishers Limited Oncogene (2014), 1 – 9

tw), or other so far unknown epigenetic mechanisms like DNAmethylation or histone modifications. In order to perform CLIP2typing of individual cases, we subsequently established asimplified and standardized procedure. The motivation behindthis effort was based on the fact that the continuous intensityvalues generated with a digital image analysis software stronglydepend on the used digital slide scanner, the applied quantifica-tion algorithm and the regions of interest defined within the

digital image. Digital IHC slide images as generated by differentslide scanner models might exhibit varying colour spectra becauseof different implementations in their hardware and software.17

Another variable is the manufacturer-dependent image proces-sing software that might use diverging colour recognitionalgorithms, resulting in different relative staining intensities andintensity ranges within one data set. Therefore, because of theinfluences of the above-mentioned parameters, the calculatedaverage staining intensities generated by one lab are hardlyreproducible by another. However, in order to use the CLIP2biomarker in molecular epidemiology, major requirements includeits reproducibility and the possibility to classify individual cases.14

For this purpose, we established a generally applicable approachfor CLIP2 typing without the use of a digital analysis software. Theproposed approach for classifying individual cases in one out offour categories is comparable to the procedures used in clinicalroutine diagnostics, for example, HER2neu typing in breast

Figure 3. Schematic workflow for CLIP2 biomarker classification.

Table 1. IHC CLIP2 visual scoring for PTC cases of the exposed andnonexposed groups (Genrisk-T cohort)

Score 1 Score 2 Score 3

Exposed (n= 16) 2 2 12Nonexposed (n= 17) 7 7 3

Abbreviations: IHC, immunohistochemistry; PTC, papillary thyroidcarcinoma.

Figure 4. Correlation between relative marker staining intensity(Definiens software) and visual scores. Good correlation (Spearman’scorrelation coefficient rho=0.83, P-value= 1.63× 10− 9) of computer-based analysis (Definiens software) and visual scores of IHC CLIP2stainings (scores 0–3; no case with score 0 within the Genrisk-Tcohort) indicates consistency of analysis methods and justifies thetransition from the analysis by image processing software to visualscoring evaluation.

Table 2. Number of radiation-exposed and nonexposed papillarythyroid carcinoma (PTC) cases in the three investigated cohorts

Number of cases in cohorts Genrisk-T Genrisk-T-PLUS UkrAm

Exposed 16 32 76Nonexposed 17 7 —

CLIP2 as radiation marker in PTCM Selmansberger et al

4

Oncogene (2014), 1 – 9 © 2014 Macmillan Publishers Limited

cancer.18 For CLIP2 typing we particularly aimed at a simplificationwhile ensuring crucial aspects of biomarker detection such asreproducibility, sensitivity and specificity. The visual scoring resultsalmost perfectly reflected the initial software-derived results,justifying the proposed simplification (Figure 4). Beyond thismethodological optimization, we validated a radiation-associatedCLIP2 overexpression in two additional independent tumourcohorts (Genrisk-T-PLUS and UkrAm). The very similar sensitivityrates between 72.4 and 75.0% in all three cohorts demonstrateboth the robustness of the approach and the ability to detect ahigh number of true positive cases (Figure 5). The specificity in theGenrisk-T cohort of 82.4% is particularly high and, in conjunction

with the sensitivity of 75.0%, points to a high quality of the CLIP2biomarker. The Genrisk-T-PLUS cohort was classified with thesame sensitivity of 75.0% but a lower specificity of 57.1%, probablybecause of the low number of nonexposed cases in this cohort.However, because of the very low rate of sporadically occurringPTCs diagnosed at young age, the number of nonexposed caseswithin this cohort cannot be increased within a reasonabletimeframe.19 The sensitivity rate of 72.4% within the UkrAm cohortof exposed cases is similar to the rates determined in the Genrisk-T and Genrisk-T-PLUS cohort. Moreover, the sensitivity rates forGenrisk-T, Genrisk-T-PLUS and UkrAm are in good accordancewith the estimated proportion of patients with radiation-inducedPTC in the exposed cohorts. Hess et al.4 estimated that ∼ 85% ofPTCs from the exposed group of the Genrisk-T cohort were indeedradiation induced. Compared with the Genrisk-T cohort (mean ageat exposure 2 years; mean age at operation 16 years), patientsfrom the UkrAm cohort were in average older at the time ofexposure (8 years) and at the time of operation (25 years). Theproportion of radiation-induced cases in the UkrAm cohort wasdetermined using the estimated average thyroid dose of 0.65 Gyand an excess relative risk of 5 Gy− 1 for the screening prevalencecohort and 1.91 Gy− 1 for the incidence cases (second to fourthscreening).20,21 Consequently, the estimated proportion ofpatients with radiation induced PTC in the UkrAm cohort was55–75%. The sensitivity rates for the CLIP2 marker of 75% and 72%for Genrisk-T and UkrAm, respectively, are likely to reflect theestimated frequencies of radiation-induced PTCs and suggest avery sensitive detection of radiation-induced cases.The radiation-specific overexpression of CLIP2 in PTCs on both

the mRNA and the protein levels strongly suggests an important

Figure 5. Stacked bar plots a, b, and c show the percentages of scores 0, 1, 2 and 3 of IHC staining intensities for the three different cohortsand their subgroups of exposed and nonexposed cases. Stacked bar plots d, e, and f show the classification into biomarker-positive andbiomarker-negative cases for the three different cohorts after inclusion of the genomic-level information (presence of gain on chromosomalband 7q11.23).

Figure 6. First neighbourhood network of CLIP2. De novo generegulatory network reconstruction was performed using theGeneNet method. All edges were independently validated byqRT–PCR (high correlation between the expressions of connectedgenes), except the association between GOLM1 and CLIP2.

CLIP2 as radiation marker in PTCM Selmansberger et al

5

© 2014 Macmillan Publishers Limited Oncogene (2014), 1 – 9

role of CLIP2 in radiation-induced carcinogenesis of PTC. In thepublished literature, CLIP2 is mainly known to be associatedwith the Williams–Beuren syndrome because of its genomiclocalization within the Williams–Beuren syndrome hemizygousdeletion.22,23 Amplifications of CLIP2 have been detected inglioblastomas and colorectal carcinomas.24,25 CLIPs are cytoplas-matic linker proteins binding to the ends of growing microtubulesand are thereby involved in the dynamics regulation of thecytoskeletal network including the localization of the dynein–dynactin complex.26,27 The latter, in turn, is involved in variousprocesses such as spindle organization, chromosome alignmentand chromosome segregation in cell division.28 Moreover, CLIP2shows high structural similarities to the protein CLIP1 and sharesconserved protein domains like cytoskeleton-associated proteinglycine rich and structural maintenance of chromosomes that islinked to chromosome segregation and cell division.29,30 CLIP1 isessential in the G2/M transition, facilitates the formation ofkinetochore–microtubule attachments during mitosis and inter-acts directly with the dynein–dynactin complex.31–34 For CLIP2 anindirect connection to the dynein–dynactin complex via BICD2 isproposed.35,36 Despite these published links, a detailed knowledgeabout the function of CLIP2 is quite limited.To gain knowledge about the functional role of CLIP2 in

radiation-associated PTC, we used the reconstructed generegulatory network inferred from microarray transcriptome dataand extracted the putative CLIP2 interaction partners from thisnetwork.15 The global mRNA expression data used for GRNreconstruction were previously published by Abend et al.15 andwere generated from PTC tissue samples from 31 patients ofthe UkrAm cohort.15 A subgroup (n= 15) of UkrAm cases of thetranscriptome data set was also subjected to CLIP2 typing in thisstudy. Thus, the reconstructed GRN was partly conducted on thesame cases that were used for CLIP2 biomarker typing. We assumethat this network represents the CLIP2 interactome, that is, itsdirect or indirect interaction partners. The identified six firstnetwork neighbours of CLIP2 (BAG2, CHST3, KIF3C, NEURL1, RGS4and PPIL3), which were validated by qRT–PCR on RNA from FFPEsections from the same cohort, are known to be involved infundamental carcinogenic processes, indicating a functional roleof CLIP2 in the carcinogenesis of radiation-associated PTC. Thepublished literature reveals that these genes are likely to belinked to the hallmarks of cancer resisting cell death, sustainingproliferative signalling and genome instability and mutation.37 BAG2and NEURL1 are involved in apoptosis and might therefore play arole in evading cell death by epithelial thyroid cancer cells.37–39

BAG2 as well as RGS4 are known to play a role in thyroid cancerand are also known to be involved in the MAPK signallingpathway.40,41 The importance of the MAPK pathway has been wellestablished in thyroid carcinogenesis. The MAPK pathway isfrequently constitutively activated in PTCs by genetic alterationssuch as rearrangements of the RET gene (RET/PTC), TRK or BRAF orby point mutations of the BRAF and RAS genes as well as by therecently discovered kinase fusion oncogenes ETV6–NTRK3.13,42,43

The aforementioned NEURL1 is known to be involved in the Notchpathway that in turn is activated by MAPK signalling in PTC.38,44

Furthermore, the subsequent pathway enrichment analysisincluding all first and second CLIP2 neighbourhood genesrevealed the significantly enriched pathways Ras activation uponCa2 influx through NMDA receptor and signalling by Nodal. Both arealso connected to the MAPK pathway, the latter of which by amolecular cross-talk between the aforementioned Notch pathwayand Nodal signalling.45–47 Activation of the MAPK pathway plays afundamental role in the regulation of cell proliferation, linking thefunction of the assumed CLIP2 interacting genes to the cancerhallmark sustaining proliferative signalling. Moreover, the CLIP2 firstneighbour KIF3C suggests a connection to the cancer hallmarkgenome instability & mutation. The motor protein KIF3C belongs tothe kinesin family and is part of a microtubule-associated

complex.48 Beside their function in microtubule-dependenttransport, kinesin motor proteins also contribute to the organiza-tion and movement of spindle poles and chromosomes duringmitosis.49–51 Therefore, KIF3C might represent an essential mitoticcomponent required for accurate cell division including accuratechromosome segregation—prerequisites for genomic stability. Aninteraction of KIF3C with CLIP2 would indicate a potential role ofCLIP2 in chromosome segregation and cell division as it hasalready been shown and discussed above for CLIP1. Genomicinstability is a well-established phenotype after irradiation andmight be a critical step in the radiation-associatedcarcinogenesis.52–55 An effect of CLIP2 on genomic instabilitycould therefore contribute to radiation-induced carcinogenesis ofPTC. Moreover, the identification of CLIP2 interacting genes opensthe possibility to investigate their role as additional biomarkers inradiation-induced thyroid carcinogenesis.In conclusion, this study provides the validation of the recently

published radiation marker CLIP2 at the protein level in radiation-associated PTCs from independent cohorts. We established astandardized procedure for CLIP2 typing, an essential step inintegrating a molecular marker into epidemiological studies.Finally, analysis of the CLIP2 interactome suggests the involve-ment of CLIP2 in the fundamental carcinogenic processesapoptosis, MAPK signalling and genomic instability, indicating afunctional role of CLIP2 in the carcinogenesis of radiation-associated PTC.

MATERIALS AND METHODSPatient data and tumour tissuesTissue samples from 124 PTCs that developed in young patients (0.1–17years of age at exposure, born before 26 April 1986) after exposure toradioiodine fallout as a consequence of the Chernobyl reactor accident aswell as 24 sporadic PTCs were obtained from the CTB. The patients wereresidents in one of the following oblasts of Ukraine: Cherkassy, Chernigov,Kiev (including Pripyat city), Rovno, Sumy or Zhytomyr. Sporadic PTC casesfrom patients born after 1 January 1987, and therefore not exposed toradioiodine fallout from the Chernobyl accident, were matched onresidency, age at operation and sex to the exposed PTC cases. Pathologicaldiagnosis was performed at the Laboratory of Morphology of EndocrineSystem (IEM, Kiev, Ukraine) by two pathologists (LZ/TB) and reviewed bythe CTB Pathology Panel.56 All tumours were diagnosed as PTCs. Thedominant histological patterns (follicular/papillary/solid) of the studiedsections were determined (by LZ/TB). An overview of the investigatedtumour cohorts is given in Table 2. The patients’ individual data are listedin Supplementary Table 1. We analysed a discovery cohort consisting of 33PTC cases (16 exposed and 17 nonexposed cases, the so-called ‘Genrisk-T’)and two validation cohorts consisting of 39 PTCs (32 exposed and 7nonexposed cases, so-called ‘Genrisk-T-PLUS’) and 76 PTCs (exposed casesonly, so-called ‘UkrAm’), respectively. FFPE tumour tissue sections of allcases and tumour DNA isolated from fresh-frozen tissue of the Genrisk-Tand Genrisk-T-PLUS cases were provided by the CTB. Tumour DNA fromthe UkrAm cases was isolated from FFPE tissue sections using the QiagenAllPrep DNA/RNA FFPE Kit (Qiagen, Hilden, Gemany). Tumour RNA wasisolated using the Qiagen RNeasy FFPE Kit. RET/PTC1 and RET/PTC3rearrangements, as well as BRAFV600E mutation status, were determinedas described previously.4

ImmunohistochemistryImmunohistochemical staining of FFPE tumour tissue sections wasperformed using a primary antibody against CLIP2 (HPA020430; SigmaPrestige Antibodies, St Louis, MO, USA). The antibody was selected fromthe ‘The Human Protein Atlas’ database that comprises informationabout the antibody specificity and staining patterns.57,58 Antibodyspecificity was validated by western blot analysis with protein lysatesfrom a CLIP2-expressing cell line and a CLIP2 small interfering RNA(Ambion, Carlsbad, CA, USA, Silencer Select ID: s14847) knockdown cell line(Supplementary Figure 3). Primary antibody was used in a dilution of 1:100in the automated staining instrument Discovery XT (Roche, Ventana,Tucson, AZ, USA) and Discovery-Universal (Ventana) was used as secondaryantibody. Signal detection was performed using peroxidase-DAB-

CLIP2 as radiation marker in PTCM Selmansberger et al

6

Oncogene (2014), 1 – 9 © 2014 Macmillan Publishers Limited

(diaminobenzidine)-MAP chemistry (Roche, Ventana). The stained tissuesections were fixed in an ethanol series and coated by a coverslip beforescanning at × 20 objective magnification with a digital slide scanningsystem (Mirax Desk, Carl Zeiss MicroImaging, Jena, Germany). The resultingstaining was confirmed by a pathologist (AW).

Image analysis of immunohistochemistryDigital image analysis. The marker staining intensities were evaluated byrelative quantification using digital image analysis platform DefiniensTis-sueStudio (Definiens AG). For this purpose, the digital slide images wereimported into the image analysis software using the tissue portal(DefiniensTissueStudio). In the first step, regions of interest, that is, tumourareas, were defined. In order to detect and quantify stained tissue areas,a continuous spectrum of brown staining intensity in relative units(0.00–3.00) was obtained using predefined algorithm and optimizedsettings. Finally, the developed image analysis was automatically appliedto all digital images in a batch process analysing the relevant regions ofinterest.

Visual scoring classification. The visual scoring classification was carriedout by two independent observers (JH/MS) in a blinded manner withrespect to the PTC exposure status. Only epithelial tumour cells wereevaluated. Tumour stroma and infiltrations, such as lymphocytes, were notconsidered. The tumour area with the most pronounced IHC staining wasused for scoring and classified into one of the four staining categories:negative staining (score 0), weak staining (score 1), intermediate staining(score 2) and strong staining (score 3). Each case was independentlyevaluated three times in a blinded scenario by each of the two observers.Thereby, six scores were obtained for each individual case. Subsequently,the modus (that is, the most frequent score) of all six scores was taken as aconsensus result. In case of a bimodal distribution (that is, two differentscores with the same frequencies), the higher value was taken. If the sixscores differed more than one scoring level, this particular case wasexcluded from further analysis.

Statistical analysis. The average marker staining intensities (obtained bythe Definiens software) of the FFPE PTC tissue sections from the exposedand nonexposed tumour groups were tested for statistical significantdifferences using the Mann–Whitney test (R base function wilxox.test with‘paired’ option set to false). Correlation of the continuous values (Definiens)with the visually assessed scores (0/1/2/3) was tested using Spearman’scorrelation method (R base function cor.test). Possible associations of theobtained CLIP2 visual scores with clinicopathological and patient datawere statistically tested using Fisher’s exact test (R base function fisher.test). P-values of o0.05 were considered statistically significant.

Genomic copy number typingIn order to detect copy number aberrations on chromosome 7q11,array CGH analysis was performed using Agilent (Santa Clara, CA, USA)60k or 180k (AMADID 252192/252206) CGH microarrays as describedpreviously.4 Alternatively, interphase fluorescence in situ hybridizationanalysis was performed on FFPE tissue sections as described previously.4

De novo reconstruction of the CLIP2 gene regulatory networkusing mRNA expression microarray dataWe used published global mRNA expression data from 31 PTC cases of theUkrAm cohort for the reconstruction of the CLIP2 GRN.15 Raw data import,filtering and normalization (preprocessing) of the data were carried outwithin the statistical programming framework R using the Bioconductorpackage limma.59–61 In order to solely consider genes that are likely to playa role within the GRN, only probes binding to transcripts with curatedRefSeq records were considered for the analysis. Replicated expressionswere summarized using the Limma function ‘summarize.probes’. Theresulting matrix with log2-normalized expression values (13 662 genes and31 samples) was subjected to de novo network reconstruction using theBioconductor package GeneNet.62 It is generally assumed that genes with astrong linear dependency in their expression patterns, and thus a highcorrelation between them, do either directly or indirectly interact witheach other. The GeneNet package allows the reconstruction of GRNs basedon pairwise partial correlation of gene expressions. Partial correlation is ameasure of the degree of association between two gene expressionvectors after removing the influence of all other genes in the data set and

thereby providing a much more accurate estimate of pair-wise correlationscompared with traditional correlation approaches. The approach imple-mented in GeneNet is particularly tailored for ‘large-p small-n’ situations inwhich the number of measurements is much higher than the number ofsamples analysed. For the definition of edges (that is, connections)between two genes we used a considerably stringent probability cutoff of0.988. From the resulting global network, the CLIP2-centred firstneighbourhood network (direct neighbours) and second neighbourhood(neighbours of first neighbours of CLIP2) were extracted. For pathwayenrichment analysis, Reactome pathway gene sets (674) were downloadedfrom GSEA.63 The gene list building the first and second neighbourhoodwas matched to the gene sets and tested for overrepresentation by two-sided Fisher’s exact test with an assumed number of total human genes of19 104 (all HUGO Gene Nomenclature Committee (HGNC) annotatedgenes).64 The P-values were corrected for multiple testing errors using theBenjamini–Hochberg correction.65 False discovery rates of o0.3 wereconsidered statistically significant.

Quantitative RT–PCRReverse transcription of RNA was performed using the VILO SuperScriptReverse Transcription Kit (Life Technologies, Carlsbad, CA, USA). TheqRT–PCR reactions (10 μl) were carried out in duplicates in a ViiA 7 RealTime PCR System in combination with the ViiA 7 Software v1.2.2 (LifeTechnologies). TaqMan gene expression assays (Life Technologies)detecting the following genes were used to validate the first neighbour-hood network (see previous section) of CLIP2: CLIP2 (Hs00185593_m1),BAG2 (Hs00188716_m1), CHST3 (Hs00427946_m1), RGS4 (Hs01111690_g1),NEURL1 (Hs00907830_m1), KIF3C (Hs01547426_m1), PPIL3 (Hs00368985_m1)and GOLM1 (Hs00895845_m1). Assays detecting the genes RPL30(Hs01066167_g1) and PGK1 (Hs99999906_m1) were used for the purposeof endogenous normalization. Relative expression levels were calculatedusing the ΔΔCt method.66 The expression levels of all eight genes(including CLIP2) from the CLIP2 first neighbourhood network weredetermined in five PTC samples of the UkrAm cohort. Correlations of thegenes were calculated based on the obtained expression values usingSpearman’s correlation method. Correlations were considered as ‘high’ andthus the association between the network genes was validated if theSpearman’s correlation coefficient was 40.6.

ABBREVIATIONSCGH, comparative genomic hybridization; CTB, Chernobyl tissuebank; FFPE, formalin-fixed, paraffin-embedded; GRN, generegulatory network; IHC, immunohistochemistry; PTC, papillarythyroid carcinoma.

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGEMENTSWe thank the International Pathology Panel of the Chernobyl Tissue Bank forconfirmation of diagnosis: Professors A Abrosimov, TI Bogdanova, G Fadda, G Hant,V LiVolsi, J Rosai and ED Williams; The Chernobyl Tissue Bank for collection of thyroidtissue samples; Professor G Thomas for establishing the matched Genrisk-T cohort;Dr Peter Jacob for discussion and determination of the proportion of radiation-induced tumours among the exposed cases in the UkrAm cohort; U Buchholz,C Innerlohinger, E Konhäuser, CM Pflüger and A Selmaier for technical support; andH Braselmann for mathematical/statistical support. This study was supported by theEuropean Commission, EpiRadBio project, FP7 Grant No. 269553 and in part by theEuropean Commission, DoReMi project, Grant No. 249689.

REFERENCES1 Cardis E, Hatch M. The Chernobyl accident-an epidemiological perspective. Clin

Oncol 2011; 23: 251–260.2 Tuttle RM, Vaisman F, Tronko MD. Clinical presentation and clinical outcomes in

Chernobyl-related paediatric thyroid cancers: what do we know now? What canwe expect in the future? Clin Oncol 2011; 23: 268–275.

CLIP2 as radiation marker in PTCM Selmansberger et al

7

© 2014 Macmillan Publishers Limited Oncogene (2014), 1 – 9

3 Cardis E, Howe G, Ron E, Bebeshko V, Bogdanova T, Bouville A et al. Cancerconsequences of the Chernobyl accident: 20 years on. J Radiol Prot 2006; 26:127–140.

4 Hess J, Thomas G, Braselmann H, Bauer V, Bogdanova T, Wienberg J et al. Gain ofchromosome band 7q11 in papillary thyroid carcinomas of young patients isassociated with exposure to low-dose irradiation. Proc Natl Acad Sci USA 2011;108: 9595–9600.

5 Iizuka D, Imaoka T, Takabatake T, Nishimura M, Kakinuma S, Nishimura Y et al.DNA copy number aberrations and disruption of the p16Ink4a/Rb pathway inradiation-induced and spontaneous rat mammary carcinomas. Radiat Res 2010;174: 206–215.

6 Ishida Y, Takabatake T, Kakinuma S, Doi K, Yamauchi K, Kaminishi M et al. Genomicand gene expression signatures of radiation in medulloblastomas after low-doseirradiation in Ptch1 heterozygous mice. Carcinogenesis 2010; 31: 1694–1701.

7 Mullenders L, Atkinson M, Paretzke H, Sabatier L, Bouffler S. Assessing cancer risksof low-dose radiation. Nat Rev Cancer 2009; 9: 596–604.

8 Klugbauer S, Lengfelder E, Demidchik EP, Rabes HM. High prevalence of RETrearrangement in thyroid tumors of children from Belarus after the Chernobylreactor accident. Oncogene 1995; 11: 2459–2467.

9 Hamatani K, Eguchi H, Ito R, Mukai M, Takahashi K, Taga M et al. RET/PTCrearrangements preferentially occurred in papillary thyroid cancer amongatomic bomb survivors exposed to high radiation dose. Cancer Res 2008; 68:7176–7182.

10 Leeman-Neill RJ, Brenner AV, Little MP, Bogdanova TI, Hatch M, Zurnadzy LY et al.RET/PTC and PAX8/PPARgamma chromosomal rearrangements in post-Chernobylthyroid cancer and their association with iodine-131 radiation dose and othercharacteristics. Cancer 2013; 119: 1792–1799.

11 Tuttle RM, Lukes Y, Onstad L, Lushnikov E, Abrosimov A, Troshin V et al. ret/PTCactivation is not associated with individual radiation dose estimates in a pilotstudy of neoplastic thyroid nodules arising in Russian children and adults exposedto Chernobyl fallout. Thyroid 2008; 18: 839–846.

12 Nikiforov YE, Rowland JM, Bove KE, Monforte-Munoz H, Fagin JA. Distinct patternof ret oncogene rearrangements in morphological variants of radiation-inducedand sporadic thyroid papillary carcinomas in children. Cancer Res 1997; 57:1690–1694.

13 Ricarte-Filho JC, Li S, Garcia-Rendueles ME, Montero-Conde C, Voza F, Knauf JAet al. Identification of kinase fusion oncogenes in post-Chernobyl radiation-induced thyroid cancers. J Clin Invest 2013; 123: 4935–4944.

14 Pernot E, Hall J, Baatout S, Benotmane MA, Blanchardon E, Bouffler S et al. Ionizingradiation biomarkers for potential use in epidemiological studies. Mutat Res 2012;751: 258–286.

15 Abend M, Pfeiffer RM, Ruf C, Hatch M, Bogdanova TI, Tronko MD et al. Iodine-131dose dependent gene expression in thyroid cancers and corresponding normaltissues following the Chernobyl accident. PLoS ONE 2012; 7: e39103.

16 Tronko M, Mabuchi K, Bogdanova T, Hatch M, Likhtarev I, Bouville A et al. Thyroidcancer in Ukraine after the Chernobyl accident (in the framework of the Ukraine-US Thyroid Project). J Radiol Prot 2012; 32: N65–N69.

17 Rojo MG, Bueno G, Slodkowska J. Review of imaging solutions for integratedquantitative immunohistochemistry in the Pathology daily practice. Folia His-tochem Cytobiol 2009; 47: 349–354.

18 Varga Z, Noske A, Ramach C, Padberg B, Moch H. Assessment of HER2 status inbreast cancer: overall positivity rate and accuracy by fluorescence in situ hybri-dization and immunohistochemistry in a single institution over 12 years: a qualitycontrol study. BMC Cancer 2013; 13: 615.

19 Williams D. Cancer after nuclear fallout: lessons from the Chernobyl accident. NatRev Cancer 2002; 2: 543–549.

20 Tronko MD, Howe GR, Bogdanova TI, Bouville AC, Epstein OV, Brill AB et al.A cohort study of thyroid cancer and other thyroid diseases after the chornobylaccident: thyroid cancer in Ukraine detected during first screening. J Natl CancerInst 2006; 98: 897–903.

21 Brenner AV, Tronko MD, Hatch M, Bogdanova TI, Oliynik VA, Lubin JH et al. I-131dose response for incident thyroid cancers in Ukraine related to the Chornobylaccident. Environ Health Perspect 2011; 119: 933–939.

22 Hoogenraad CC, Koekkoek B, Akhmanova A, Krugers H, Dortland B, Miedema Met al. Targeted mutation of Cyln2 in the Williams syndrome critical region linksCLIP-115 haploinsufficiency to neurodevelopmental abnormalities in mice. NatGenet 2002; 32: 116–127.

23 Ferrero GB, Howald C, Micale L, Biamino E, Augello B, Fusco C et al. An atypical7q11.23 deletion in a normal IQ Williams-Beuren syndrome patient. Eur J HumGenet 2010; 18: 33–38.

24 Suzuki T, Maruno M, Wada K, Kagawa N, Fujimoto Y, Hashimoto N et al. Geneticanalysis of human glioblastomas using a genomic microarray system. Brain TumorPathol 2004; 21: 27–34.

25 Lassmann S, Weis R, Makowiec F, Roth J, Danciu M, Hopt U et al. Array CGHidentifies distinct DNA copy number profiles of oncogenes and tumor suppressor

genes in chromosomal- and microsatellite-unstable sporadic colorectal carcino-mas. J Mol Med 2007; 85: 293–304.

26 Hoogenraad CC, Akhmanova A, Grosveld F, De Zeeuw CI, Galjart N. Functionalanalysis of CLIP-115 and its binding to microtubules. J Cell Sci 2000; 113(Pt 12):2285–2297.

27 Galjart N. CLIPs and CLASPs and cellular dynamics. Nat Rev Mol Cell Biol 2005; 6:487–498.

28 Karki S, Holzbaur EL. Cytoplasmic dynein and dynactin in cell division and intra-cellular transport. Curr Opin Cell Biol 1999; 11: 45–53.

29 Hoogenraad CC, Akhmanova A, Galjart N, De Zeeuw CI. LIMK1 and CLIP-115:linking cytoskeletal defects to Williams syndrome. Bioessays 2004; 26: 141–150.

30 Marchler-Bauer A, Anderson JB, Chitsaz F, Derbyshire MK, DeWeese-Scott C,Fong JH et al. CDD: specific functional annotation with the Conserved DomainDatabase. Nucleic Acids Res 2009; 37: D205–D210.

31 Yang X, Li H, Liu XS, Deng A, Liu X. Cdc2-mediated phosphorylation of CLIP-170 isessential for its inhibition of centrosome reduplication. J Biol Chem 2009; 284:28775–28782.

32 Wieland G, Orthaus S, Ohndorf S, Diekmann S, Hemmerich P. Functionalcomplementation of human centromere protein A (CENP-A) by Cse4p fromSaccharomyces cerevisiae. Mol Cell Biol 2004; 24: 6620–6630.

33 Tanenbaum ME, Galjart N, van Vugt MA, Medema RH. CLIP-170 facilitates theformation of kinetochore-microtubule attachments. EMBO J 2006; 25: 45–57.

34 Lansbergen G, Komarova Y, Modesti M, Wyman C, Hoogenraad CC, Goodson HVet al. Conformational changes in CLIP-170 regulate its binding to microtubulesand dynactin localization. J Cell Biol 2004; 166: 1003–1014.

35 Hoogenraad CC, Wulf P, Schiefermeier N, Stepanova T, Galjart N, Small JV et al.Bicaudal D induces selective dynein-mediated microtubule minus end-directedtransport. EMBO J 2003; 22: 6004–6015.

36 Fumoto K, Hoogenraad CC, Kikuchi A. GSK-3beta-regulated interaction of BICDwith dynein is involved in microtubule anchorage at centrosome. EMBO J 2006;25: 5670–5682.

37 Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011;144: 646–674.

38 Teider N, Scott DK, Neiss A, Weeraratne SD, Amani VM, Wang Y et al. Neuralized1causes apoptosis and downregulates Notch target genes in medulloblastoma.Neuro Oncol 2010; 12: 1244–1256.

39 Wang HQ, Zhang HY, Hao FJ, Meng X, Guan Y, Du ZX. Induction of BAG2 proteinduring proteasome inhibitor-induced apoptosis in thyroid carcinoma cells. Br JPharmacol 2008; 155: 655–660.

40 Arndt V, Daniel C, Nastainczyk W, Alberti S, Hohfeld J. BAG-2 acts as an inhibitor ofthe chaperone-associated ubiquitin ligase CHIP. Mol Biol Cell 2005; 16: 5891–5900.

41 Druey KM, Blumer KJ, Kang VH, Kehrl JH. Inhibition of G-protein-mediated MAPkinase activation by a new mammalian gene family. Nature 1996; 379: 742–746.

42 Nikiforov YE. Thyroid carcinoma: molecular pathways and therapeutic targets.Mod Pathol 2008; 21(Suppl 2): S37–S43.

43 Xing M. Molecular pathogenesis and mechanisms of thyroid cancer. Nat RevCancer 2013; 13: 184–199.

44 Yamashita AS, Geraldo MV, Fuziwara CS, Kulcsar MA, Friguglietti CU, da Costa RBet al. Notch pathway is activated by MAPK signaling and influences papillarythyroid cancer proliferation. Transl Oncol 2013; 6: 197–205.

45 Strizzi L, Hardy KM, Seftor EA, Costa FF, Kirschmann DA, Seftor RE et al.Development and cancer: at the crossroads of Nodal and Notch signaling. CancerRes 2009; 69: 7131–7134.

46 Grewal T, Koese M, Tebar F, Enrich C. Differential regulation of RasGAPs in cancer.Genes Cancer 2011; 2: 288–297.

47 Xia Z, Dudek H, Miranti CK, Greenberg ME. Calcium influx via the NMDA receptorinduces immediate early gene transcription by a MAP kinase/ERK-dependentmechanism. J Neurosci 1996; 16: 5425–5436.

48 Muresan V, Abramson T, Lyass A, Winter D, Porro E, Hong F et al. KIF3C and KIF3Aform a novel neuronal heteromeric kinesin that associates with membranevesicles. Mol Biol Cell 1998; 9: 637–652.

49 Goldstein LS. Kinesin molecular motors: transport pathways, receptors, andhuman disease. Proc Natl Acad Sci USA 2001; 98: 6999–7003.

50 Mandelkow E, Mandelkow EM. Kinesin motors and disease. Trends Cell Biol 2002;12: 585–591.

51 De S, Cipriano R, Jackson MW, Stark GR. Overexpression of kinesins mediatesdocetaxel resistance in breast cancer cells. Cancer Res 2009; 69: 8035–8042.

52 Little MP, Heidenreich WF, Moolgavkar SH, Schollnberger H, Thomas DC. Systemsbiological and mechanistic modelling of radiation-induced cancer. Radiat EnvironBiophys 2008; 47: 39–47.

53 Huang L, Snyder AR, Morgan WF. Radiation-induced genomic instability and itsimplications for radiation carcinogenesis. Oncogene 2003; 22: 5848–5854.

54 Morgan WF. Non-targeted and delayed effects of exposure to ionizing radiation: I.Radiation-induced genomic instability and bystander effects in vitro. Radiat Res2003; 159: 567–580.

CLIP2 as radiation marker in PTCM Selmansberger et al

8

Oncogene (2014), 1 – 9 © 2014 Macmillan Publishers Limited

55 Morgan WF. Non-targeted and delayed effects of exposure to ionizing radiation:II. Radiation-induced genomic instability and bystander effects in vivo, clastogenicfactors and transgenerational effects. Radiat Res 2003; 159: 581–596.

56 Williams ED. Guest Editorial: two proposals regarding the terminology ofthyroid tumors. Int J Surg Pathol 2000; 8: 181–183.

57 Uhlen M, Bjorling E, Agaton C, Szigyarto CA, Amini B, Andersen E et al. A humanprotein atlas for normal and cancer tissues based on antibody proteomics. MolCell Proteom 2005; 4: 1920–1932.

58 Uhlen M, Oksvold P, Fagerberg L, Lundberg E, Jonasson K, Forsberg M et al.Towards a knowledge-based Human Protein Atlas. Nat Biotechnol 2010; 28:1248–1250.

59 Team RDC. R: A Language and Environment for Statistical Computing.R Foundation for Statistical Computing, 2013.

60 Smyth GK. Limma: linear models for microarray data. In: Gentleman R,Carey V, Dudoit S, Irizarry R, Huber W (eds) Bioinformatics and ComputationalBiology Solutions Using {R} and Bioconductor. Springer: New York, 2005, pp397–420.

61 Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S et al.Bioconductor: open software development for computational biology andbioinformatics. Genome Biol 2004; 5: R80.

62 Opgen-Rhein R, Strimmer K. From correlation to causation networks: a simpleapproximate learning algorithm and its application to high-dimensional plantgene expression data. BMC Syst Biol 2007; 1: 37.

63 Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al.Gene set enrichment analysis: a knowledge-based approach for interpretinggenome-wide expression profiles. Proc Natl Acad Sci USA 2005; 102: 15545–15550.

64 Gray KA, Daugherty LC, Gordon SM, Seal RL, Wright MW, Bruford EA.Genenames.org: the HGNC resources in 2013. Nucleic Acids Res 2013; 41:D545–D552.

65 Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical andpowerful approach to multiple testing. J R Stat So 1995; 57: 289–300.

66 Livak KJ, Schmittgen TD. Analysis of relative gene expression data usingreal-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001; 25:402–408.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)

CLIP2 as radiation marker in PTCM Selmansberger et al

9

© 2014 Macmillan Publishers Limited Oncogene (2014), 1 – 9