Biology of gastrointestinal stromal tumour and mechanisms of imatinib resistance

Abstract

The introduction of imatinib therapy for GISTs has represented a major breakthrough of oncology in the last decade, improving dramatically the prognosis of GIST patients. The discovery of an oncogenic event (kinase mutations) having a major predictive value underlines the need for a paradigm of classification combining thorough pathological examination and molecular analysis. The response to therapy with imatinib is indeed determined by the type of mutation in kinase genes. Genetic analysis should be therefore performed in all cases deserving such a therapy. The study of the resistance has enabled us to discover other important oncogenetic events (e.g. dysregulation of downstream pathways) that are important in the genesis and therapy of other tumours as well. Genetic studies have also allowed a molecular classification of GISTs, thus identifying subsets of GISTs (e.g. SDH-negative, paediatric) that do not behave as the “classical”, RTK-mutated tumours, probably representing different “entities”.

Introduction

Gastrointestinal stromal tumours (GISTs) are the most frequent mesenchymal tumours of the gastrointestinal tract, with an incidence between 15 and 20 new cases/10⁶/year.¹ It is however possible that the true incidence be underestimated. Autopsy studies have indeed shown that many GISTs are simply incidental findings at the autopsy and do not give any symptomatology.2, 3 The most frequent localization is the stomach, followed by the small intestine, the colon-rectum and the oesophagus. The existence of true extragastrointestinal GISTs has been proposed, but it is questioned by many authors. Although the concept that mesenchymal tumours of the gastrointestinal tract with leiomyomatous morphology are ‘bizarre or blastomatous’ had been recognized since the 40s by Stout, the particular nature of these tumours was defined first in the 80s, when Mazur et al, based mainly on electron microscopy findings, proposed the non committed term of “Gastrointestinal stromal tumours” (Figure 1). This concept was further developed with the help of immunohistochemistry (positivity of CD34 in 70% of cases, while the expression of smooth muscle markers was found only in a minority of cases). At the end, in 1998, two groups4, 5 showed independently that most GISTs show constitutively activating mutations of the CKIT gene that encode an important receptor tyrosine kinase (RTK) type III. KIT (CD117) immunohistochemical expression was contemporarily reported in more than 95% of GIST cases, thus becoming an important tool for the diagnosis⁶ (Figure 2a–d). In 2003, it was shown that a substantial fraction of KIT wild-type GISTs show mutations in the gene coding for PDGFRA (platelet-derived growth factor alpha), another RTK type III (Figure 2e–h) and that these mutations are mutually exclusive with mutations in CKIT.7, 8 Oncogenesis of GISTs is therefore probably related mainly to early activation of RTKs. The importance of this discovery is also underlined by the fact that KIT and PDGFRA are very good targets for the tyrosine kinase inhibitor imatinib mesylate (Gleevec^®, Novartis Pharma AG, Basel, Switzerland). Imatinib is now the first-line drug in the treatment of inoperable or metastatic GISTs and may be used before and/or after surgical treatment.⁹ Imatinib is indeed approved in US and in Europe for adjuvant therapy and may be used also in a neoadjuvant setting to reduce the tumour mass.¹⁰ The effect of imatinib on sensitive GISTs is dramatic, as it was shown in the first communications of Joensuu et al.¹¹ However, primary and secondary resistance to targeted therapy remains, in spite of this great success, a problem to solve. Beneath a minority of GISTs that simply do not respond to the therapy with imatinib (primary resistance), it is now well known that about half the patients develop disease progression by 2 years of treatment with imatinib.¹² The main predictor of the response to therapy is represented by the type and localization of the mutation in the RTK genes.13, 14 Therefore, the genetic alterations in the RTK genes are not only important early events in the genesis of GISTs; they also define the response to targeted therapy. In the last years, alternative/additional oncogenetic events have been identified, that could have at least partially a prognostic/predictive meaning, such as mutations in the small G-proteins BRAF and KRAS and hyperexpression of the transcription factor ETV1.15, 16, 17

As outlined above, in the last decade GISTs have represented an important model for the targeted therapy of solid tumours. The understanding of their biology has been of paramount importance for the developing of RTK inhibitors. Moreover, GISTs represent the perfect example for the need of a change in the paradigm of classification of disease, integrating the “classical” clinicopathologic parameters with the molecular alterations, a process that has been always put forward for haematologic malignancies.

This review will therefore focus on the biology and molecular pathology of GISTs, and on the mechanisms of resistance to the targeted therapy.