Unlucky number 13? Differential effects of KIT exon 13 mutation in gastrointestinal stromal tumors^,

Gastrointestinal stromal tumor (GIST) represents the majority of soft tissue sarcomas arising from the GI tract. Most GISTs encode activating mutations in the kit gene, an important genetic event in tumorigenesis. Imatinib mesylate (imatinib, Gleevec™; Novartis, Basel, Switzerland) has revolutionized the management of patients with advanced disease by targeting the aberrant kinase activity of Kit.

Mutation status of kit in GIST has emerged as a prognostic and predictive factor for imatinib response. At diagnosis, most kit mutations are within exon 11 or less commonly within exon 9, 13, or 17 (Chen et al., 2004). Regarding exon 13, early reports suggested that tumors harboring these mutations responded suboptimally to imatinib therapy. However, the efficacy of imatinib in this subpopulation has been difficult to analyze due to its rarity with an estimated incidence of approximately 1% (Desai et al., 2007).

Chen and colleagues showed that primary mutation of the Kit kinase domain 1 (exon 13, V654A) correlated to imatinib resistance and rapid progression (Heinrich et al., 2006). Desai and colleagues demonstrated that 80% of patients with imatinib‐resistant nodules after initial response to imatinib harbored secondary mutations in kit exons 13 and 17 (Lasota et al., 2008). Thus, kit exon 13 can be altered by primary or secondary mutation that can confer resistance to imatinib. However, the differential effect on imatinib binding and clinical response for each particular mutation within kit exon 13 is not fully understood.

We encountered a patient with GIST harboring a kit exon 13 mutation (K642E) who experienced a robust and prolonged response to imatinib. This 66‐year‐old Caucasian man presented to The University of Texas — M. D. Anderson Cancer Center in August 2001 with recurrent GIST measuring approximately 3.5cm at the gastroesophageal junction and several small intra‐ and extra‐peritoneal implants (Fig. 1A). Pathologic evaluation revealed a highly cellular, spindle cell tumor with brisk mitotic activity. The tumor was diffusely reactive for Kit, CD34, and focally reactive for smooth muscle actin while being negative for desmin, S‐100, and pan‐cytokeratin. He was treated with 600mg daily imatinib, and experienced a decrease in tumor burden by 1 year (Fig. 1B).

Figure 1.

Radiography, mutational analysis, and molecular modeling. (A) Contrast‐enhanced CT of pre‐imatinib tumor (arrow) at the gastroesophageal junction. (B) Contrast‐enhanced CT of the same tumor after 1 year of imatinib therapy demonstrating decreased tumor size. (C) Contrast‐enhanced CT at the time of multi‐focal progression of disease. (D) Mutational analysis of kit exon 13 at codons 642 and 654 of the imatinib‐naïve tumor. The black bar and arrow indicate the codon and exact site of mutation. Codon 654 is not mutated in this imatinib‐sensitive tumor. *Wild‐type sequence. (E) Mutational analysis of the surgically resected, imatinib‐resistant tumor indicating mutations of kit exon 13 at both codon 642 and 654. (F) Secondary structure of the double mutant Kit kinase by molecular modeling. Juxtamembrane domain=green; c‐helix=yellow; imatinib=gray (carbon), blue (nitrogen), red (oxygen); E642=pink; A654=red. Hydrogen atoms and water molecules are omitted for clarity. E642 confers aberrant activity of the kinase while A654 confers decreased imatinib binding.

However, after 45 months of imatinib therapy, he had multi‐focal progression within the abdominal cavity and a pleural implant (Fig. 1C). He underwent surgical resection to debulk his tumor burden. Characteristic histopathologic features of treatment response to imatinib were not present in these recurrent tumor areas.

Mutational analysis of his tumor at the time of imatinib initiation revealed a missense mutation of kit exon 13, K642E (Fig. 1D). This mutation is the most common codon affected in kit exon 13 (Desai et al., 2007). Upon resistance and progression of disease, mutational analysis revealed a secondary kit exon 13 mutation (V654A) along with the primary K642E mutation (Fig. 1E). This finding corroborates those seen by Chen and others regarding the V654A mutation and imatinib resistance (Chen et al., 2004; Singer et al., 2002).

Importantly, molecular modeling of the K642E mutant predicts that the negatively charged glutamic acid (in contrast to the positively charged lysine) results in altered interactions of the control c‐helix with the Kit juxtamembrane domain which, in turn, leads to an uncontrolled phosphorylation of the kinase, without effecting binding of imatinib (Fig. 1F). However, the 654 position of Kit lies within the ATP binding pocket and directly participates in imatinib binding. Thus, substitution of a valine for a smaller alanine in the 654 position leads to a “void” in the binding site and, hence, a less efficient drug binding.

The calculated free energy of binding (ΔG_bind) between wild‐type Kit and imatinib is −10.22kcal/mol (Tamborini et al., 2006). The ΔG_bind for the K642E mutant is marginally less at −10.03kcal/mol. The ΔG_bind for the Kit V654A mutant is −8.7kcal/mol or a one‐fold decrease in drug affinity. The ΔG_bind for the Kit K642E/V654A double mutant is −8.04kcal/mol indicating a 1–2 fold decrease in affinity for imatinib. Thus, the two mutations work together to enhance the resistance to imatinib as compared to each mutation alone.

In conclusion, this GIST patient harboring kit exon 13 mutation K642E benefited from imatinib. Treatment selected for the double mutation that conferred resistance and eventuated death. Despite the minimal 11 codons separating these loci, the resultant phenotype is imatinib‐sensitive for K642E and imatinib‐resistant for K642E/V654A. As GIST kit mutation reporting becomes more prevalent, documentation of the specific mutation is important to distinguish those patients with probable imatinib resistance from those that may benefit from imatinib for many years.

McAuliffe John C., Wang Wei-Lein, Pavan Giovanni M., Pricl Sabrina, Yang Dan, Chen Su S., Lazar Alexander J.F., Pollock Raphael E., Trent Jonathan C., (2008), Unlucky number 13? Differential effects of KIT exon 13 mutation in gastrointestinal stromal tumors, Molecular Oncology, 2, doi: 10.1016/j.molonc.2008.05.002.