Biochemical and Functional Analysis of Novel KRAS Insertions in MPN and Other Cancers

The somatic KRAS mutations found in ~25% of human cancers commonly encode amino acid substitutions at codons 12, 13, and 61. Each of these mutations lead to elevated levels of active Ras-GTP by reducing intrinsic Ras GTPase activity and conferring resistance to GTPase activating proteins (GAPs). Amino acids 12, 13, and 61 play key roles in binding the γ phosphate of GTP and facilitating GAP-mediated catalysis. The germline KRAS point mutations found in Noonan syndrome encode different amino acid substitutions that have less severe biochemical consequences. Together, these observations and the structural conservation of the Ras/GAP switch through evolution suggest that a limited spectrum of mutations can constitutively activate Ras and cause disease. Juvenile myelomonocytic leukemia (JMML) is an aggressive myeloproliferative neoplasm characterized by driver Ras pathway mutations in 85% of cases, including common somatic KRAS and NRAS substitutions. We unexpectedly identified a novel somatic KRAS insertion, 66GQEEYSA67, in the bone marrow of a patient who meets most, but not all diagnostic criteria, for JMML and lacks any other Ras pathway mutation. This seven amino acid insertion is located in the conserved Switch II domain of Ras, which binds to and activates PI3 kinase and other Ras effectors. We characterized the functional and biochemical consequences of this mutation, and of a similar five amino acid insertion identified in two lung cancers and one colon cancer archived in the COSMIC database (66EEYSA67). We find that unlike WT K-Ras, the expression of both insertion mutants supports cytokine independent growth of IL-3 dependent pro-B Ba/F3 cells similar to oncogenic K-Ras^G12D. Ba/F3 cells expressing K-Ras^66GQEEYSA67 or K-Ras^66EEYSA67 mutant proteins have markedly elevated levels of Ras-GTP that are insensitive to IL-3 stimulation, and result in constitutive ERK and Akt phosphorylation. Confirming that these Switch II insertion mutants have oncogenic potential in the hematopoietic system, fetal liver cells transduced with KRas^66GQEEYSA67, KRas^66EEYSA67, or KRas^G12D formed cytokine independent colony forming unit granulocyte macrophage (CFU-GM) colonies and also exhibited GM-CSF hypersensitivity, which is a cellular hallmark of JMML. Biochemical characterization of recombinant K-Ras^66GQEEYSA67 and K-Ras^66EEYSA67 proteins showed them to have reduced intrinsic GTP hydrolysis, and to accumulate in the GTP-bound conformation when compared to oncogenic K-Ras^G12D. Furthermore, whereas recombinant p120 GAP markedly accelerated the rate of GTP hydrolysis by WT K-Ras, K-Ras^66GQEEYSA67 and K-Ras^66EEYSA67 were completely resistant. The most likely structural explanation is that these insertions perturb critical contacts between glutamine 61 in the Switch II domain of Ras and GAPs. We conclude that KRAS^66GQEEYSA67 and KRAS^66EEYSA67 are oncogenic driver mutations in JMML and other cancers. These data illuminate a new class of oncogenic RAS mutations and are of immediate diagnostic relevance. Furthermore, these data reveal unanticipated structural flexibility in a domain of Ras that plays a critical role in effector binding, which has implications for potential mechanism of resistance to emerging therapeutic approaches.