3D Modeling of Novel Transforming JAK Mutations in T-Cell Acute Lymphoblastic Leukemia Reveals Altered Pseudokinase-Kinase Domain Interactions That Result in Constitutive JAK Kinase Activity

Background: 

Many pediatric T-cell acute lymphoblastic leukemia patients harbor mutations in IL7Ra or downstream molecules encoded by JAK1, JAK3, N-RAS, K-RAS, NF1, AKT, and PTEN. These mutated signaling molecules can contribute to leukemia by disturbing a multitude of cellular processes such as the cell cycle, epigenetics, apoptosis, or affecting other important signal transduction pathways.

Aims: 

We aimed to determine the overall incidence of JAK family mutations in a large cohort of T-ALL patients. We also aimed to generate a 3D JAK1 model including known and newly identified JAK mutations in order to better understand how these contribute to JAK kinase activity and the transformation of cells.

Methods: 

We screened 146 pediatric T-ALL patient samples for mutations in the FERM, pseudokinase and kinase domains of the Janus kinase gene family (JAK1-3, TYK2). To establish a 3D JAK1 model, we superimposed individual pseudokinase and kinase crystallographic structures on the homologous TYK2 pseudokinase-kinase structure. We visualized JAK mutations and their effects on the 3D structure. We modified the IL3-dependent Ba/F3 cell line to express JAK mutant or wild type genes upon induction by doxycycline. We tested these Ba/F3 derivative lines for transforming ability, signaling, and resistance to various inhibitors in the absence of IL3.

Results: 

JAK1 or JAK3 mutations were found in 10 patients; no mutations were found in JAK2 or TYK2. We found JAK1 and JAK3 mutations as previously reported, but also identified amino acid substitutions as a result of novel JAK1 mutations including V427M, L624YPILKV, E668Q, P815S, and T901G. Our novel 3D model of JAK1 places most mutations in one of two crucial pseudokinase-kinase interaction sites, which can weaken the interaction and facilitate constitutive kinase activity. One interaction is between the hinge region of the pseudokinase domain and the loop in the kinase domain, which is supported by four salt bridges. Mutations in T-ALL disrupting these salt bridges include E668Q, R724H and its JAK3 equivalent R657Q, and T901G. The second interaction with the kinase domain is formed by a helical domain in the pseudokinase domain, located just upstream of the conserved F575-F636-V658 triad. This F-F-V triad is predicted to act as a structural switch that controls the catalytic activity of JAK kinases. Various mutations occur in the direct vicinity and can affect the function of this switch. V658F in T-ALL and its JAK2 equivalent V617F in polycythemia vera patients are mutations in this triad. The frequent JAK3 mutation M511I in T-ALL flanks the F513 residue (equivalent of JAK1 F575) and also affects the F-F-V triad. The L624YPILKV insertion mutation is located in a loop near the helical domain, which may also subtly compromise the F-F-V triad structural switch leading to derepression of the kinase domain. Expression of mutant JAK genes—in contrast to the wild type genes—transforms Ba/F3 cells by supporting IL3-independent growth, and by activating downstream RAS-MEK-ERK and PI3K-AKT pathways. This pathway activation as a result of ligand-independent mutant JAK kinase activity was confirmed by measuring phospho-proteins including p-MEK, p-ERK, p-AKT, p-mTOR, and p-p70S6K, and can be blocked by JAK inhibitors. Notably, JAK3 mutants signal significantly weaker than JAK1 mutants, possibly due to different dependence on (endogenous) receptors that normally mediate wild type JAK signaling.

Summary/Conclusion: 

In a 3D model, we show that JAK mutations are located in critical interface regions between the pseudokinase and kinase domains, maintaining the kinase in an open, active confirmation. The inducible Ba/F3 model system confirms the transforming capacity of JAK mutations, reveals constitutive active downstream signaling, and is also suitable to test the effect of various inhibitors. The visualization of various JAK mutations in a 3D model and how these contribute to kinase activity provides insight in how mutant JAK could be inhibited, helping guide the development of new small molecule inhibitors of mutant JAKs.