High-throughput sequence analysis of the tyrosine kinome in acute myeloid leukemia

Abstract

To determine whether aberrantly activated tyrosine kinases other than FLT3 and c-KIT contribute to acute myeloid leukemia (AML) pathogenesis, we used high-throughput (HT) DNA sequence ana-lysis to screen exons encoding the activation loop and juxtamembrane domains of 85 tyrosine kinase genes in 188 AML patients without FLT3 or c-KIT mutations. The screen identified 30 nonsynonymous sequence variations in 22 different kinases not previously reported in single-nucleotide polymorphism (SNP) databases. These included a novel FLT3 activating allele and a previously described activating mutation in MET (METT1010I). The majority of novel sequence variants were stably expressed in factor-dependent Ba/F3 cells. Apart from one FLT3 allele, none of the novel variants showed constitutive phosphorylation by immunoblot analysis and none transformed Ba/F3 cells to factor-independent growth. These findings indicate the majority of these alleles are not potent tyrosine kinase activators in this cellular context and that a significant proportion of nonsynonymous sequence variants identified in HT DNA sequencing screens may not have functional significance. Although some sequence variants may represent SNPs, these data are consistent with recent reports that a significant fraction of such sequence variants are "passenger" rather than "driver" alleles and underscore the importance of functional assessment of candidate disease alleles.

Figure 1

Representative sequence trace. A representative reference (normal) and AML sequence trace comparison are shown for the BMX R540H mutation. The arrow notes the guanine to adenine substitution that results in an arginine to histidine substitution at codon 540. Protein alignment demonstrates this residue is highly conserved within the tyrosine kinase family.

Figure 2

Expression of nonsynonymous sequence variants does not alter cellular phosphotyrosine content. Evaluation of phosphotyrosine levels after expression of mutant DDR1, TEK, and FES. Ba/F3 cells were stably transfected. Cells were serum starved overnight and subjected to immunoblot analysis for phosphotyrosine (4G10) as well as total (A) DDR1, (B) TEK, or (C) FES.

Figure 3

Nonsynonymous sequence variations do not alter sensitivity to cognate ligands. Collagen stimulation of DDR1 mutants for evaluation of ligand hypersensitivity. Ba/F3 cells stably transfected with wild-type and mutant DDR1 as in Figure 2. Cells were serum starved overnight at which time they were collagen stimulated with (A) a time course or (B) a dose response for the indicated times and doses. Cell lysates were subjected to immunoblot analysis for phosphotyrosine (4G10) and total DDR1.

Figure 4

Evaluation of DDR1, TEK, and FES mutations in Ba/F3 WEHI independence assay. Ba/F3 cells were transfected with constructs expressing wild-type or mutant DDR1 (A), TEK (B), or FES (C). Following 4 weeks of neomycin selection, stably transfected cells were washed with WEHI-free media and plated in media containing 10-fold serial dilutions of IL-3, ranging from 1 ng/mL to 10⁻⁵ ng/mL. Cells were subjected to an MTS assay at day 3 after plating for determination of total viable cells. Values represent mean plus or minus SEM.

Figure 5

Evaluation of PDGFRB, FLT3, CSK, and EGFR mutations in IL-3 independence assay. Ba/F3 cells were transfected with constructs expressing mutant PDGFRB/FLT3 (A) or CSK/EGFR (B). Following 4 weeks of neomycin selection, stably transfected cells plated in media in the absence of exogenous cytokines. The number of viable cells was assessed by trypan blue exclusion. As a positive control, Ba/F3 cells expressing the constitutively active FLT3-D835Y allele were plated in the absence of IL-3 in panel A to demonstrate IL-3–independent growth. Values represent mean plus or minus SEM.