IJC International Journal of Cancer
A novel transcript, VNN1-AB, as a biomarker for colorectal cancer Marthe Lïvf1,2,3, Torfinn Nome1,2, Jarle Bruun1,2, Mette Eknæs1,2, Anne C. Bakken1,2,4, John P. Mpindi5, Sami Kilpinen5,6, Torleiv O. Rognum7,8, Arild Nesbakken2,9, Olli Kallioniemi5, Ragnhild A. Lothe1,2,3,4 and Rolf I. Skotheim1,2,4 1
Department of Cancer Prevention, Institute for Cancer Research, the Norwegian Radium Hospital, Oslo University Hospital, Oslo, Norway Centre for Cancer Biomedicine, Faculty of Medicine, University of Oslo, Oslo, Norway 3 Department of Biosciences, University of Oslo, Oslo, Norway 4 Cancer Stem Cell Innovation Center (CAST), Oslo University Hospital, Oslo, Norway 5 Institute of Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki, Finland 6 MediSapiens Ltd, Helsinki, Finland 7 University of Oslo, Oslo, Norway 8 Division for Forensic Medicine, Department of Forensic Pathology and Clinical Forensic Medicine, the Norwegian Institute of Public Health, Oslo, Norway 9 Department of Gastrointestinal Surgery, Aker University Hospital, Oslo University Hospital, Oslo, Norway
Colorectal cancer is a global health challenge with high incidence rate and mortality. The patients’ prognosis is strongly associated with disease stage and currently there is a need for improved prognostic and predictive biomarkers. In this study, novel colorectal cancer-specific transcript structures were nominated from whole transcriptome sequencing of seven colorectal cancer cell lines, two primary colorectal carcinomas with corresponding normal colonic mucosa and 16 normal tissues. The nominated transcripts were combined with gene level outlier expression analyses in a cohort of 505 colorectal cancers to identify biomarkers with capacity to stratify colorectal cancer subgroups. The transcriptome sequencing data and outlier expression analysis revealed 11 novel colorectal cancer-specific exon–exon junctions, of which 3 were located in the gene VNN1. The junctions within VNN1 were further characterized using rapid amplification of cDNA ends (RACE) and the prevalence of the subsequently characterized novel transcript, VNN1-AB, was investigated by real-time RT-PCR in 291 samples of miscellaneous origins. VNN1-AB was not present in any of the 43 normal colorectal tissue samples investigated, but in 5 of the 6 polyps, and 102 of the 136 (75%) colorectal cancers. We have identified a novel transcript of the VNN1 gene, with an organ-confined complete specificity for colorectal neoplasia.
Colorectal cancer (CRC) is a global health challenge with about 1.2 million new cases and 600,000 deaths recorded per year.1 The development of CRC follows a morphological change from a benign adenoma to a malign carcinoma and involves a stepwise accumulation of molecular changes. Survival rates of CRC greatly correspond to stage at primary diagnosis, with a 5-year survival rate of 93 and 8% for stage I and IV, respectively.2
Numerous cytogenetic, genetic and epigenetic alterations in CRC have been published since the Vogelstein model,3 but few are implemented in clinical use. Still, the golden standard for detection of precancerous lesions and carcinomas is colonoscopy. Development of non-invasive testing awaits robust biomarkers with validated high specificity and sensitivity. With the exception of microsatellite instability, only clinical variables are currently used to predict prognosis.4
Key words: alternative splicing, colorectal cancer, biomarker, transcriptome, VNN1 Abbreviations: ACTB: actin beta; cDNA: complementary deoxyribonucleic acid; CRC: colorectal cancer; CT: cycle threshold; ERBB2: verb-b2 erythroblastic leukemia viral oncogene homolog 2 neuro/glioblastoma derived oncogene homolog (avian); ERG: v-ets erythroblastosis virus E26 oncogene homolog (avian); GTI: gene tissue index; mRNA: messenger ribonucleic acid; RACE: rapid amplification of cDNA ends; RT-PCR: reverse transcription polymerase chain reaction; TMPRSS2: transmembrane protease serine 2; VNN1: vanin 1 Additional Supporting Information may be found in the online version of this article. Grant sponsor: Molecular Life Science, University of Oslo; Grant sponsor: Norwegian Cancer Society; Grant number: PR-2007-0166; Grant sponsor: Research Council of Norway through its Centres of Excellence Funding Scheme; Grant number: 179571; Grant sponsor: NorStore; Grant number: NS9013K DOI: 10.1002/ijc.28855 History: Received 21 Jan 2014; Accepted 6 Mar 2014; Online 19 Mar 2014 Correspondence to: Rolf I. Skotheim, Department of Cancer Prevention, Institute for Cancer Research, Oslo University Hospital HE— Norwegian Radium Hospital, P.O. Box 4953 Nydalen, NO-0424 Oslo, Norway, Fax: 147–2278-1745, E-mail:
[email protected]
C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
Cancer Genetics
2
2
VNN1-AB, as a biomarker for colorectal cancer
Cancer Genetics
What’s new? For colorectal cancer (CRC), prognosis is strongly associated with disease stage. Improved prognostic and predictive biomarkers would be extremely helpful. In this study, the authors used whole transcriptome sequencing of CRC and normal cells, followed by outlier-expression analysis from a large set of samples, to identify variant mRNA transcripts that might be useful as biomarkers of CRC. Real-time PCR showed that a novel transcript, VNN1-AB, had high sensitivity and 100% specificity for CRC. This transcript may thus be a valuable prognostic tool for stratifying CRC risk.
Molecules with high cancer specificity and sensitivity are ideal as biomarkers. Furthermore, because CRC is a heterogeneous disease, such molecules may not be present in all cancers. One cellular mechanism which may yield cancer-specific molecules is alternative splicing of the primary transcripts, which gives rise to multiple mRNA transcript variants per gene. Alterations of this process are common in cancer due to fundamental difference in expression patterns of known splicing-regulatory genes in cancerous compared to normal tissues.5 This can result in the production of mRNAs being exclusively present in malignant cells. Such cancer-specific transcripts can be ideal as biomarkers. Further variation in gene expression can be produced by the use of alternative core promoters. This enables diversification of transcriptional regulation within a single gene and thereby plays a significant role in the control of gene expression in various cell lineages, tissue types and developmental stages. The molecular mechanisms behind the selective use of multiple promoters are not completely understood, but the use of diverse core promoter structures, variable concentrations of cis-regulatory elements and regional epigenetic mechanisms are thought to be important factors (reviewed in Ref. 6). Several oncogenes and tumor suppressor genes have multiple promoters and the aberrant use of one promoter over another in some of these genes is directly linked to abnormal cell growth.7,8 Multiple studies have compared cancerous to noncancerous tissues to identify differentially expressed biomolecules. However, only a few of these studies have examined the heterogeneity that exists between individual cancers. Identification of genes with an overexpression in a subset of samples has proven successful in finding critical players in carcinogenesis. For example, overexpression of the well known cancer-critical gene ERBB2 is seen in about 30% of human breast cancers and is in clinical use as a companion diagnostic biomarker identifying those that are likely to respond to treatment with Herceptin.9,10 Also, genes with outlier expression pattern may reveal biologically different and diverse cancer subtypes. For example, searching for such outlier expression pattern was the basis for the identification of the TMPRSS2-ERG fusion gene, which is present in about half of prostate cancers.11 Several statistical methods have been developed to identify genes with outlier expression pattern.11–14 In this study, whole transcriptome sequencing followed by outlier expression analysis from a large set of samples was used to identify novel transcript structures with biomarker
potential. Further analyses revealed a novel transcript, VNN1AB, with high sensitivity and specificity for neoplasia in the large bowel. We have also investigated its usefulness as a biomarker for CRC.
Material and Methods Patient samples and cell lines
Altogether, 291 samples were analyzed in this study (Table 1). This included 136 primary CRC samples from two independent clinical series collected from patients treated surgically at hospitals in the Oslo region, Norway, including 10, 63, 52 and 11 stage I, II, III and IV samples, respectively. Twelve of the samples showed microsatellite instability and 123 were microsatellite stable (1 sample not determined).15,16 Additionally, normal colonic mucosa taken from disease free areas of the colon of 43 CRC patients was included. In 12 cases both the tumor and the normal sample were derived from the same patient. Finally, six adenomas from patients enrolled in a multi-hospital study of young age onset CRC were included. In addition, we analyzed 86 cell lines from altogether 16 types of cancer (Supporting Information Additional file 1), and a panel of 20 normal tissue samples from different organs (FirstR Human Normal Tissue Total RNA, each a pool of ChoiceV RNA from at least three individuals, with the exception of an individual sample from the stomach; Ambion, Applied Biosystems by Life Technologies, Carlsbad, CA). RNA isolation
From the adenomas, CRC samples, and cell lines, RNA was extracted using the Qiagen AllPrep DNA/RNA Mini Kit (Qiagen GmbH, Hilden, Germany). The Ambion RiboPureTM kit (Applied Biosystems) was used to obtain RNA from the normal colonic mucosa samples. Both procedures were performed according to the manufacturers’ protocols. In silico RNA-sequencing data
Data from 16 normal tissue samples (Table 1) from Illumina Human Body Map v2 (ArrayExpress accession id [E-MTAB-513] and European Nucleotide Archive study [EMBL:ERP000546]) and 505 CRC samples and 44 normal samples from the In Silico Transcriptomics database was used in the analyses.17 Whole transcriptome sequencing
The transcriptome of two CRC and normal mucosa tissue from the same patients as well as seven cell lines (HCT15, HCT116, HT29, LS1034, RKO, SW48 and SW480) (Cell line C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
3
Lïvf et al.
Table 1. Samples investigated in this study Total (samples)
Samples CRC tissue
Cell lines In silico transcriptomics
Quantitative RT-PCR (samples)
Stage I (series I)
10
10
Stage II (total)
63
63
Series I
10
10
Series II
53
53
Stage III (total)
Nonmalignant tissue
Whole transcriptome sequencing (samples)
52
52
Series I
9
9
Series II
43
Stage IV (series I)
11
Colorectal adenomas
6
Colonic mucosa
43
43 2
11 6
2
43
Normal tissues
36
16
202
From 16 different cancer types
86
7
86
Normal tissue
44
CRC tissue
505
1
1
Adrenal, adipose, brain, breast, colon, heart, kidney, liver, lung, lymph, ovary, prostate, skeletal muscle, testes, thyroid and white blood cells. Adipose, bladder, brain, cervix, heart, kidney, liver, lung, oesophagus, ovary, placenta, prostate, skeletal muscle, small intestine, spleen, stomach, testes, thymus, thyroid, and trachea.
dataset from Nome et al.18) were sequenced on the Illumina Genome Analyzer IIx (Illumina, San Diego, CA). Construction of the libraries, including poly-A mRNA isolation, fragmentation and gel-based size selection followed the standard Illumina mRNA library preparation (Illumina, icom.illumina. com, 2009). Shearing of the cDNA to an average fragment size of about 250 base pairs was achieved using the Covaris S2 focused-ultrasonicator (Covaris, Woburn, MA). Seventysix base pairs, from each side of the fragments were sequenced according to the paired-end RNA-sequencing protocols from Illumina for Solexa with paired-end module. Between 23 and 38 million clusters were generated for each of the eleven libraries. Gene level outlier expression analyses from the in silico transcriptomics database
To highlight genes with outlier expression patterns, i.e., in which a group of samples are expressed at increased levels, the Gene tissue index (GTI) was applied.12 Briefly, for a given gene and a given cancer type, GTI calculates the outlier expression profile on the basis of both the fraction of samples with expression above a certain threshold, and the average expression change within these samples. Three different cutoffs were applied (inter quartile range 1 75th percentile, 90th percentile and 95th percentile) in both a general and a tissue specific manner and combined to a final GTI score. The GTI algorithm was applied on data from the In Silico Transcriptomics database, and the applied version, IST4_pub, contained Affymetrix gene expression data from approximately 15,000 human samples.17 The gene expression values in the database have a common normalization across all samples, C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
enabling analyses across the whole sample set. In this study, gene expression data from 505 CRC samples and 44 normal colorectal mucosa samples were used as input to the GTI analysis, and the resulting 50 genes with lowest GTI rank were included for further analysis (Table 2). Filtering of RNA-sequencing data
The paired CRC patient samples (tumor and normal tissue), seven CRC cell lines and 16 normal tissues from Illumina Human Body Map v2 were analyzed with TopHat 2.0.6 and Bowtie 2.0.519,20 (with default parameters). Only the junctions.bed file, produced by TopHat, was included in the pipeline. All junctions present in the two CRC samples were pooled together before excluding junctions present in any of the 18 normal samples (16 from Illumina Human Body Map and two from the CRC patients), all previously known junctions (all exons downloaded from BioMart with Ensembl release 70), and junctions not involving genes with low GTI rank (Table 2). Three of the remaining 11 junctions were present in the same gene, VNN1, and were investigated further. The VNN1 junctions were also present in two of the seven CRC cell lines, HT29 and LS1034 (Supporting Information Additional file 2). Characterization of known and novel transcripts
To validate the presence of novel transcripts observed with whole transcriptome sequencing VNN1 was investigated further in the HT29 cell line using 50 -RACE (SMART RACE cDNA Amplification kit and Advantage 2 PCR kit; Clontech, Mountain View, CA). RACE primer (GGCTTCAGACTAAA CAAGCGTCCGTCA) and nested primer (CTGGGTTCCG
Cancer Genetics
2
4
VNN1-AB, as a biomarker for colorectal cancer
Table 2. Genes with outlier expression profiles in colorectal cancer
Cancer Genetics
Ensembl
Gene symbol
Ensembl
Gene symbol
ENSG00000000005
TNMD
ENSG00000159184
HOXB13
ENSG00000083782
EPYC
ENSG00000160181
TFF2
ENSG00000085741
WNT11
ENSG00000163499
CRYBA2
ENSG00000099953
MMP11
ENSG00000163735
CXCL5
ENSG00000101850
GPR143
ENSG00000166819
PLIN1
ENSG00000103355
PRSS33
ENSG00000169248
CXCL11
ENSG00000104524
PYCRL
ENSG00000170373
CST1
ENSG00000105664
COMP
ENSG00000170835
CEL
ENSG00000109255
NMU
ENSG00000174697
LEP
ENSG00000109511
ANXA10
ENSG00000175426
PCSK1
ENSG00000111432
FZD10
ENSG00000177984
LCN15
ENSG00000111700
SLCO1B3
ENSG00000183971
NPW
ENSG00000112299
VNN1
ENSG00000184774
RP11–480I12.4
ENSG00000117983
MUC5B
ENSG00000185269
NOTUM
ENSG00000124875
CXCL6
ENSG00000196188
CTSE
ENSG00000124882
EREG
ENSG00000198028
ZNF560
ENSG00000126856
PRDM7
ENSG00000198535
C2CD4A
ENSG00000127362
TAS2R3
ENSG00000204528
PSORS1C3
ENSG00000129451
KLK10
ENSG00000204866
IGFL2
ENSG00000132781
MUTYH
ENSG00000211937
IGHV2–5
ENSG00000134389
CFHR5
ENSG00000211976
IGHV3–73
ENSG00000137745
MMP13
ENSG00000178589
ENSG00000145214
DGKQ
ENSG00000211646
ENSG00000146678
IGFBP1
ENSG00000197460
ENSG00000147206
NXF3
ENSG00000183133
The table is listing the top 50 genes according to their GTI rank in CRC vs. normal colonic mucosa.
AAAGTGCCACTGAGG) were designed by the Primer3 software21 with default settings. The RACE products were separated by agarose gel electrophoresis, cut and extracted from the gel (MinElute Gel Extraction Kit; Qiagen). Finally, each product were cloned (TOPO TA Cloning Kit; Invitrogen by Life Technologies, CA) and sequenced (AB3730 DNA analyzer; Applied Biosystems by Life Technologies, Foster City, CA). All procedures were performed according to the manufacturers’ protocols. Quantitative RT-PCR analyses of VNN1
For 291 samples (Table 1), reverse transcription of two mg total RNA was performed using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems by Life Technologies), with Multiscribe reverse transcriptase and random primers, according to the manufacturer’s protocol. Predesigned assays were used to investigate the expression of the 5’ and 3’ end of the reference VNN1 transcript (Hs01546808_m1 and Hs01546812_m1, respectively). A third assay was custom made using the Primer Express Software v.3.0 (Applied Biosystems by Life Technologies. Primers were
designed to align in exons A (CATGATAGAATGATCTAG CTGGACCTT) and B (CCGCTAACTGGTCTTATTGTTT CC) (Fig. 1). The probe (AGTGATTACTTTCCACCTGC) was labelled at the 50 end with the fluorescent dye 6-FAM and modified at the 30 end with a nonfluorescent quencher suppressing the fluorescence of the dye prior to primer extension. Also, a minor groove binder, raising the melting temperature of the probe, was added to ensure its hybridization to the target prior to primer annealing and extension. The custom made assay had an efficiency of 94.4% (data not shown). All three assays were run in triplicates with realtime detection on an ABI 7900HT Fast Real-Time PCR System (TaqMan; Applied Biosystems by Life Technologies). Expression levels were reported as the median cycle threshold (CT) of the triplicates. Human ACTB (Applied Biosystems by Life Technologies) was used as a control of cDNA input. All normal mucosa samples had CT > 34 for the custom made assay targeting VNN1-AB, and CT 5 34 was set as the threshold for all three assays targeting VNN1, where all samples with a CT < 34 are scored as having positive expression. C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
5
Lïvf et al.
Results Identification of a novel transcript, VNN1-AB, in colorectal cancer
Whole transcriptome sequencing of two CRC samples revealed 228,982 unique exon–exon junctions. To enrich for cancer-specificity, junctions also present in at least one of 18 normal samples of miscellaneous origins were excluded from further analysis (32,240 remaining). In addition, all previously known junctions were excluded (20,341 remaining). Finally, a bioinformatic genome-scale search for genes with outlier expression patterns within CRC was performed and all junctions not present in such outlier genes were excluded. Three of the 11 resulting novel junctions were located within intron five of one gene, VNN1, and were investigated further. VNN1 has one annotated transcript variant containing seven exons in the Ensembl genome database (ENST00000367928). Analysis of whole transcriptome sequencing data of seven colon cancer cell lines (dataset from Nome et al.18) revealed the presence of the nominated cancer-specific junctions in HT29 and LS1034, with highest coverage in HT29. VNN1 expression in the HT29 cell line revealed novel transcript variants transcribed from a novel promoter within intron five (Fig. 1). The transcripts included two novel exons with variable sizes resulting in three novel variants collectively called VNN1-AB (Supporting Information Additional file 3 and 4). Prevalence and expression levels of the VNN1-AB transcript
Individual quantitative RT-PCR assays for the novel VNN1AB and reference VNN1 transcripts (Fig. 1) were designed and applied to a total of 291 tissues samples and cell lines (Table 1). Quantitative RT-PCR showed that the 50 -end of the reference transcript was present in 58% (25/43) of normal colonic mucosa samples, 100% (6/6) of colorectal adenomas, 95% (129/136) of CRCs and 26% (5/19) of CRC cell lines (Fig. 2a). The assay targeting the 30 -end of the reference tranC 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
script revealed its presence in 86% (37/43) of normal colonic mucosa samples, 100% of colorectal adenomas and CRCs (6/ 6 and 136/136, respectively) and 58% (11/19) of CRC cell lines (Fig. 2b). The novel VNN1-AB was not present in any of the normal colonic mucosa samples (n 5 43), but in 83% (5/6) of colorectal adenomas, 75% (102/136) of CRCs and 47% (9/19) of CRC cell lines (Fig. 3a). For 92% (11/12) of the tumornormal pairs, the presence of VNN1-AB was restricted to the tumor sample. One pair was negative in both samples. All four stages of CRC were represented among the positive samples (80% in Stage I, 68% in Stage II, 81% in Stage III and 82% in Stage IV). Both microsatellite instable as well as stable tumors expressed the VNN1-AB transcript (83 and 75%, respectively). Expression of the novel transcripts was not associated with disease free survival for any of the two patient series investigated (log rank test, P 5 0.60 and P 5 0.47 for Series I and II, respectively; data not shown). Although we had complete neoplasia specificity within the large bowel, we also wanted to explore the expression of VNN1 and VNN1-AB in other tissues. Samples from various healthy organs were analyzed (n 5 20), as well as cancer cell lines from other types of cancers (n 5 67). The reference transcript was present in 100 % (20/20; both 50 and 30 -end) of normal tissues tested and in 21% (14/67, 50 -end) and 58% (39/67, 30 -end) of the cell lines (Figs. 2c and 2d). The VNN1AB transcript was present in 30% (6/20) of normal tissue samples from noncolonic organs (adipose, lung, esophagus, ovary, stomach and trachea) and in 43% (29/67) of the cell lines (Fig. 3b).
Discussion By combining exon–exon junction identification by pairedend RNA-sequencing with analysis of a large sample cohort for outlier gene expression patterns in CRC and normal
Cancer Genetics
Figure 1. Identification of a novel transcript variant, VNN1-AB. In the HT29 cell line, whole transcriptome analysis revealed sequence reads from exons six and seven in the reference transcript of VNN1, in addition to two locations in intron five. No sequence at all was detected from exons one to five. Exons in the reference transcript (ENST00000367928) are represented by blue boxes, whereas novel exons are depicted in green. A primer for 50 -RACE (black arrow) was designed to target exon six in the reference transcript. Three quantitative RT-PCR assays were applied to investigate the different transcript parts of VNN1 (red lines).
VNN1-AB, as a biomarker for colorectal cancer
Cancer Genetics
6
Figure 2. VNN1 is expressed in both normal and cancer samples. Expression results from quantitative RT-PCR are reported as the cycle threshold value (CT), with higher expression corresponding to lower CT-values. The transcript was considered absent for CT >34. (a) Expression of the 50 -end of the reference transcript is seen in 25 of 43 normal colonic mucosa, 6 of 6 colorectal adenomas, 129 of 136 CRCs and 5 of 19 colon cancer cell lines. (b) Expression of the 30 -end of the reference transcript is seen in 37 of 43 normal colonic mucosa and all colorectal adenomas and CRCs, and in 11 of 19 colon cancer cell lines. (c) and (d) The reference transcript is expressed in all normal tissues of other organs investigated and in 14 of 67 (50 -end; panel c) and 39 of 67 (30 -end; panel d) cancer cell lines from noncolonic organs.
colonic mucosa samples, we have identified VNN1-AB as a novel transcript with neoplasia specificity within the large bowel. An ideal diagnostic biomarker should have high sensitivity, being present in the large majority of tumors, and high specificity being absent from normal control samples. However, these criteria are rarely fulfilled for individual biomarkers, and combination in molecular panels is necessary to achieve robustness to the test.22,23 The novel VNN1-AB transcript identified in this article has a great specificity and
adequate sensitivity within the large bowel and could therefore be useful as a cancer biomarker. Combining VNN1-AB with other CRC-specific biomarkers, would improve its applicability in cancer detection. For example, we recently identified a particular splice variant of SLC39A14 to have a similar organ-confined cancer-specific expression within the large bowel.24 One great potential for a cancer-specific biomarker is the ability to detect cancer in a noninvasive manner. For CRC C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
Figure 3. Neoplasia-specific expression of VNN1-AB within colon and rectum. Expression results from quantitative RT-PCR are reported as the cycle threshold value (CT), with higher expression corresponding to lower CT-values. The VNN1-AB transcript was considered absent for CT > 34. (a) For samples within the colon and rectum, all 43 normal colonic mucosa samples were negative, and positive expression was seen for 5 of 6 adenomas, 102 of 136 carcinomas and 9 of 19 cancer cell lines. (b) VNN1-AB was negative for most normal tissues from other organs, except from trachea, stomach, lung, ovary, adipose, and esophagus (decreasing order of expression). Twenty-nine of 67 cancer cell lines from noncolonic organs had positive expression of VNN1-AB.
this means with the use of blood or fecal samples, instead of biopsies taken during colonoscopy. The novel VNN1-AB transcript identified in this study is neoplasia-specific within the large bowel, and therefore has potential as a diagnostic biomarker for CRC, but the presence of the transcript in normal tissue from other organs may reduce its applicability for such non-invasive testing. Use of high-throughput RNA-sequencing in search for cancer-specific biomarkers is advantageous over proteomics in that the analysis can easily be done on a genome-scale, and the method is unbiased in the sense that it enables straightforward identification of previously not annotated transcripts. However, for robust biomarker measurements in various biospecimens, it may again be beneficial to transfer the test to protein based analyses. The reference transcript of VNN1 code for a 513 amino acid long membrane-anchored amidohydrolase, with a pantetheinase domain located between amino acids 22 and 491. The VNN1-AB transcript identified in this article potentially encode a protein lacking the 397 first amino acid C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
residues, leaving a potentially malfunctioning pantetheinase domain. Additionally, the exclusion of a putative signal peptide located between amino acid one and 21 of the VNN1 transcript probably disrupts the cellular localization. To our knowledge, VNN1 has no known role in cancer as such, or CRC in particular. However, unrelated to its function, it has been included as one of seven genes in a biomarker panel for stratifying current risk for developing CRC.25,26 In this biomarker panel, it is the reference transcript of VNN1 which is measured. Thus, the specificity of such a panel could be further improved by use of the novel transcript identified in this study. VNN1 is also linked to another disease of the large bowel by being a biomarker for inflammatory bowel disease.27 In addition, evidence suggests that VNN1, together with MMP9, can be used to differentiate pancreatic cancer-associated diabetes mellitus from type 2 diabetes mellitus.28 VNN1-AB is present in 5 out of 6 pancreatic cell lines investigated in this study and might improve differentiation between the two diseases.
Cancer Genetics
7
Lïvf et al.
8
VNN1-AB, as a biomarker for colorectal cancer
By the use of high throughput RNA sequencing and outlier expression analysis, novel exons within the gene boundaries of VNN1 have been identified. The novel transcript, VNN1-AB, is expressed in 75% of the analyzed CRC samples, but not in any of the analyzed normal colonic mucosa samples. VNN1-AB constitutes a highly neoplasia-specific biomarker within the colon and rectum, although also being expressed at other anatomical sites.
Acknowledgements The adenoma material was kindly provided through collaboration with Dr. Espen Thiis-Evensen, Oslo University Hospital, Norway. Cell lines from cholangiocarcinoma, gall bladder cancer and leukaemia were kindly provided by Dr. Gregory Gores, Mayo Clinic, MN, USA, Dr. Alexander Knuth, University Hospital Zurich, Switzerland and Thoas Fioretos, Lund University Hospital, Sweden. The authors thank Zeremariam Johannes for help with establishing the cloning protocol.
References 1.
2.
3. 4.
5.
Cancer Genetics
6.
7.
8.
9.
Ferlay J, Shin HR, Bray F, et al. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127:2893–917. O’Connell JB, Maggard MA, Ko CY. Colon cancer survival rates with the new American Joint Committee on Cancer sixth edition staging. J Natl Cancer Inst 2004;96:1420–5. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61:759–67. Guastadisegni C, Colafranceschi M, Ottini L, et al. Microsatellite instability as a marker of prognosis and response to therapy: a metaanalysis of colorectal cancer survival data. Eur J Cancer 2010;46:2788–98. Skotheim RI, Nees M. Alternative splicing in cancer: noise, functional, or systematic? Int J Biochem Cell Biol 2007;39:1432–49. Davuluri RV, Suzuki Y, Sugano S, et al. The functional consequences of alternative promoter use in mammalian genomes. Trends Genet 2008; 24:167–77. Agarwal VR, Bulun SE, Leitch M, et al. Use of alternative promoters to express the aromatase cytochrome P450 (CYP19) gene in breast adipose tissues of cancer-free and breast cancer patients. J Clin Endocrinol Metab 1996;81: 3843–9. Nakanishi T, Bailey-Dell KJ, Hassel BA, et al. Novel 5’ untranslated region variants of BCRP mRNA are differentially expressed in drugselected cancer cells and in normal human tissues: implications for drug resistance, tissuespecific expression, and alternative promoter usage. Cancer Res 2006;66:5007–11. Slamon DJ, Clark GM, Wong SG, et al. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 1987;235:177–82.
10. Gianni L, Dafni U, Gelber RD, et al. Treatment with trastuzumab for 1 year after adjuvant chemotherapy in patients with HER2-positive early breast cancer: a 4-year follow-up of a randomised controlled trial. Lancet Oncol 2011; 12:236–44. 11. Tomlins SA, Rhodes DR, Perner S, et al. Recurrent fusion of TMPRSS2 and ETS transcription factor genes in prostate cancer. Science 2005;310: 644–8. 12. Mpindi JP, Sara H, Haapa-Paananen S, et al. GTI: a novel algorithm for identifying outlier gene expression profiles from integrated microarray datasets. PLoS One 2011;6:e17259. 13. Tibshirani R, Hastie T. Outlier sums for differential gene expression analysis. Biostatistics 2007;8: 2–8. 14. Wu B. Cancer outlier differential gene expression detection. Biostatistics 2007;8:566–75. 15. Berg M, Danielsen SA, Ahlquist T, et al. DNA sequence profiles of the colorectal cancer critical gene set KRAS-BRAF-PIK3CA-PTEN-TP53 related to age at disease onset. PLoS One 2010;5: e13978. 16. Thorstensen L, Diep CB, Meling GI, et al. WNT1 inducible signaling pathway protein 3, WISP-3, a novel target gene in colorectal carcinomas with microsatellite instability. Gastroenterology 2001; 121:1275–80. 17. Kilpinen S, Autio R, Ojala K, et al. Systematic bioinformatic analysis of expression levels of 17,330 human genes across 9,783 samples from 175 types of healthy and pathological tissues. Genome Biol 2008;19:R139. 18. Nome T, Thomassen GO, Bruun J, et al. Common fusion transcripts identified in colorectal cancer cell lines by high-throughput RNA sequencing. Transl Oncol 2013;6:546–53.
19. Trapnell C, Pachter L, Salzberg SL. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 2009;25:1105–11. 20. Langmead B, Trapnell C, Pop M, et al. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol 2009;10:R25. 21. Rozen S, Skaletsky H. Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol 2000;132:365–86. 22. Lind GE, Danielsen SA, Ahlquist T, et al. Identification of an epigenetic biomarker panel with high sensitivity and specificity for colorectal cancer and adenomas. Mol Cancer 2011;10:85. 23. Ransohoff DF. Cancer: developing molecular biomarkers for cancer. Science 2003;299:1679–80. 24. Sveen A, Bakken AC, Agesen TH, et al. The exon-level biomarker SLC39A14 has organconfined cancer-specificity in colorectal cancer. Int J Cancer 2012;131:1479–85. 25. Marshall KW, Mohr S, Khettabi FE, et al. A blood-based biomarker panel for stratifying current risk for colorectal cancer. Int J Cancer 2010; 126:1177–86. 26. Yip KT, Das PK, Suria D, et al. A case-controlled validation study of a blood-based seven-gene biomarker panel for colorectal cancer in Malaysia. J Exp Clin Cancer Res 2010;29:128. 27. Gensollen T, Bourges C, Rihet P, et al. Functional polymorphisms in the regulatory regions of the VNN1 gene are associated with susceptibility to inflammatory bowel diseases. Inflamm Bowel Dis 2013;19:2315–25. 28. Huang H, Dong X, Kang MX, et al. Novel blood biomarkers of pancreatic cancer-associated diabetes mellitus identified by peripheral blood-based gene expression profiles. Am J Gastroenterol 2010;105:1661–9.
C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
