Elsevier

European Journal of Cancer

Volume 45, Issue 3, February 2009, Pages 461-469
European Journal of Cancer

Genome-wide expression profile of sporadic gastric cancers with microsatellite instability

https://doi.org/10.1016/j.ejca.2008.10.032Get rights and content

Abstract

Gastric cancers with mismatch repair (MMR) inactivation are characterised by microsatellite instability (MSI). In this study, the transcriptional profile of 38 gastric cancers with and without MSI was analysed. Unsupervised analysis showed that the immune and apoptotic gene networks efficiently discriminated these two cancer types. Hierarchical clustering analysis revealed numerous gene expression changes associated with the MSI phenotype. Amongst these, the p53-responsive genes maspin and 14-3-3 sigma were significantly more expressed in tumours with than without MSI. A tight immunosurveillance coupled with a functional p53 gene response is consistent with the better prognosis of MSI cancers.

Frequent silencing of MLH1 and downregulation of MMR target genes, such as MRE11 and MBD4, characterised MSI tumours. The downregulation of SMUG1 was also a typical feature of these tumours. The DNA repair gene expression profile of gastric cancer with MSI is of relevance for therapy response.

Introduction

Gastric cancers (GC) with defective mismatch repair (MMR) comprise 10–25% of all GC. These tumours accumulate DNA replication errors at short-repeat sequences that are identified by the presence of microsatellite instability (MSI) (reviewed in [1] and references therein). MSI is mostly associated with hypermethylation of the MLH1 promoter CpG island leading to loss of MLH1 expression in the tumour and occasionally in the surrounding gastric mucosa.2 MMR inactivation leads to accumulation of mutations in repeat tracts located within the coding and non-coding regions of tumour suppressor, anti-apoptotic and DNA repair genes as reported in both colon (reviewed in [3]) and gastric (reviewed in [1]) cancers with MSI. Defects in DNA repair pathways may impact on the tumour response to therapy (reviewed in [4]).

Recent data on gene expression profiling of endometrial5 and colon6 cancers with and without MSI support the hypothesis that these cancer subsets arise through distinct pathways of aberrant signalling. These pathways are likely cancer specific as indicated by the lack of common signature between MSI endometrial and colon cancers with hypermethylation of MLH1 as the only consistently observed feature.

The main objective of this study was to determine whether the MSI phenotype in GC could be distinguished from the microsatellite stable (MSS) phenotype using microarrays. Here, we show that MSI GC can indeed be discriminated from MSS tumours by their specific gene expression signature that mainly involves the immune response and the p53 gene network. Moreover, the DNA repair gene expression profiling showed that besides MLH1 silencing that characterised the majority of MSI GC, a set of DNA repair genes were differentially expressed in these tumours too. This information should be used to design therapies tailored to this GC subgroup.

Section snippets

Patients and clinical specimens

All tissues were collected from a series of GC cases identified in an area around Florence (Italy) characterised by high GC risk. Following full institutional review board approval, patients were recruited in the period 2000–2005 from the main hospitals of this area. All patients were residing in the Florence district, and they signed an informed consent before providing a blood sample and a gastric tissue during surgery. Immediately after resection, tissues from the invasive region of cancer

Gene expression analysis discriminates gastric cancer specimens with and without MSI

We first analysed 131 primary GC of similar grades, stages and histologies for the MSI phenotype by using two highly sensitive mononucleotide markers, BAT26 and BAT25. Overall, 20/131 (15.3%) GC showed contractions in both BAT26 and BAT25, and were thus classified as MSI. 19 MSI tumour samples were selected and matched by age, sex and grade with 19 MSI-negative (MSS) cancers characterised by lack of instability at both the loci tested. Demographic, clinical and histopathological features of

Discussion

MSI characterises approximately 15% of sporadic colorectal and gastric cancers. This study is the first report on transcriptional profile of MMR defective GC conducted on numerous sets of MSI tumours matched with MSS tumours (by age, sex and grade). The pathways that better discriminated the MSI status were the immune and apoptotic responses. Besides genes involved in the proinflammatory response, several components of the antigen presentation and binding machinery were differentially expressed

Conflict of interest statement

None declared.

Acknowledgements

We would like to thank Ivano Iavarone for statistical analysis and for critically reading our manuscript. Grant support: Progetto Integrato Oncologia, Regione Toscana-Ministero della Salute, Associazione Italiana per la Ricerca sul Cancro (AIRC) to ED, MIUR/FIRB (RBNE01RNN7), Compagnia di San Paolo (Coordinator: G. Frosina), programma di collaborazione ISS/NIH to ED.

References (26)

  • L. Ottini et al.

    Mutations at coding mononucleotide repeats in gastric cancer with the microsatellite mutator phenotype

    Oncogene

    (1998)
  • R.A. Irizarry et al.

    Summaries of Affymetrix GeneChip probe level data

    Nucleic Acids Res

    (2003)
  • C. Li et al.

    Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection

    Proc Natl Acad Sci USA

    (2001)
  • Cited by (287)

    View all citing articles on Scopus
    e

    These authors equally contributed to this work.

    View full text