1133O - Molecular profiling of small intestinal neuroendocrine tumours

Date 27 September 2014
Event ESMO 2014
Session Neuroendocrine tumours
Topics Neuroendocrine Cancers
Pathology/Molecular Biology
Basic Scientific Principles
Presenter Anna Karpathakis
Citation Annals of Oncology (2014) 25 (suppl_4): iv394-iv405. 10.1093/annonc/mdu345
Authors A. Karpathakis1, A. Feber2, T. Morris2, H. Dibra3, C. Pipinikas2, D. Oukrife2, J. Francis4, D. Mandair5, C. Toumpanakis5, T. Meyer2, T.V. Luong5, M. Caplin5, M. Meyerson4, S. Beck2, C. Thirlwell2
  • 1Cancer Institute, University College London, WC1E6BT - London/GB
  • 2Cancer Institute, University College London, London/GB
  • 3Cancer Institute, University College London, WC1 - EBT/GB
  • 4The Broad Institute, The Broad Institute, Boston/US
  • 5Neuroendocrine Tumour Unit, Royal Free Hospital, London/GB



Aberrant DNA methylation is known to play an important role in the pathogenesis of many human cancers, however little is known about its role in small intestinal neuroendocrine tumour (SI NET) development. We report the first unbiased genome wide DNA methylation analysis of a large cohort of SI NET, aiming to identify key epigenetic changes specific to SI NET which may contribute to tumorigenesis.


Illumina Infinium HumanMethylation450 Array (interrogating 99% of RefSeq genes) analysis was performed on DNA extracted from macrodissected SI NET primary tumours (n = 49) and normal small intestine (SI)(n = 21). Publicly available methylation data on >600 samples from The Cancer Genome Atlas was also assessed for comparison (colorectal, pancreatic and gastric adenocarcinoma and healthy tissue). Gene expression was determined using the Illumina DASL array on RNA extracted from SI NET primary tumours (n = 32) and normal SI (n = 6). Analysis was performed using ChAMP and limma R packages. A Bonferroni adjusted significance threshold value of p < 0.05 was used throughout.


Comparison of SI NET with normal SI identified a total of 130,083 significant methylation variable positions, including 1841 sites hypermethylated by >30% in tumour compared to normal tissue. 626 genes were found to have significant >3fold differential expression between SI NET and normal SI. Integrated analysis identified a group of 11 candidate genes where altered methylation and expression was significant and concordant (downregulated: CDX1, FBP1, C20orf54, GATA5; upregulated: PTPRN, PCSK1, PRLHR, CELSR3, GIPR, LMX1B, SCGN). Hypermethylation of GIPR (gastic inhibitory polypeptide receptor) was most significant (SI NET median methylation 0.67 vs normal SI 0.29 p < 2.2e−16), and 92% of SI NET were hypermethylated. Hypermethyation of GIPR was sensitive for the detection of NET compared to other GI malignancis with an AUC of 0.991 (95% CI .0.991-0.999).


This study is the first comprehensive analysis of the epigenetic profile of SI NET and identifies hypermethylation of GIPR as a potential novel biomarker. Novel radioligands targeting GIPR have been developed for use as imaging tools and it is a promising target for novel therapeutic agents.


All authors have declared no conflicts of interest.