Sequence survey of receptor tyrosine kinases reveals mutations in glioblastomas

Abstract

It is now clear that tyrosine kinases represent attractive targets for therapeutic intervention in cancer. Recent advances in DNA sequencing technology now provide the opportunity to survey mutational changes in cancer in a high-throughput and comprehensive manner. Here we report on the sequence analysis of members of the receptor tyrosine kinase (RTK) gene family in the genomes of glioblastoma brain tumors. Previous studies have identified a number of molecular alterations in glioblastoma, including amplification of the RTK epidermal growth factor receptor. We have identified mutations in two other RTKs: (i) fibroblast growth receptor 1, including the first mutations in the kinase domain in this gene observed in any cancer, and (ii) a frameshift mutation in the platelet-derived growth factor receptor-alpha gene. Fibroblast growth receptor 1, platelet-derived growth factor receptor-alpha, and epidermal growth factor receptor are all potential entry points to the phosphatidylinositol 3-kinase and mitogen-activated protein kinase intracellular signaling pathways already known to be important for neoplasia. Our results demonstrate the utility of applying DNA sequencing technology to systematically assess the coding sequence of genes within cancer genomes.

Fig. 1.

FGFR1 mutations in glioblastoma. (A) Schematic representation of the domain structure of FGFR1 showing the location of mutation we identified in this study. The numbers indicate the amino acid residue number at the approximate boundaries of each domain as described by Webster and Donoghue (26). The N and C termini are labeled N and C, respectively. The peptide regions showing locations of the mutations are shown below the domain structure, and the mutated residues are indicated in the amino acid sequence. SP, signal peptide; D1-D3, Ig-like domains; AB, acid box; TM, transmembrane domain; TK1 and TK2, tyrosine kinase domains; KI, kinase insert region. (B) Sequence data showing the two somatic DNA sequence alterations (indicated by vertical arrows); these are (from left to right) N546K (C/C→C/A) and R576W (C/C→C/T), which are located within the kinase domain in glioblastomas.

Fig. 2.

Consequence of mutations on the structure of FGFR1 kinase domain. Structure predictions show effects of FGFR1 kinase mutations on protein properties. (A) A three-dimensional structure of the kinase domain of the human wild-type FGFR1 showing the locations of mutations and functional regions. The residues in the wild-type protein are shown in a ball and stick representation colored in magenta. The other domains are colored accordingly: α-helices, red; β-sheets, light blue; catalytic domain, yellow; activation loop, green; nucleotide-binding loop, dark blue; 14-aa unresolved portion of crystal structure bearing the tyrosine residues (Y583 and Y585), black. The locations of somatic mutations at positions 546 and 576 are indicated. (B) Distribution of surface charges in the human wild-type and mutant FGFR1. The color scale for surface potential is shown. Red color denotes acidic properties of the surface (negative charge), and blue corresponds to basic surface (positive charge). The regions affected by the mutations are circled in white and black on the wild-type FGFR1 and in the corresponding color in the affected region on the relevant mutant FGFR1.

Fig. 3.

Evolutionary conservation of FGFR at the mutated residues in the kinase domain. (A) Sequence alignment of the region containing mutations within the kinase domains of human paralogs FGFR1-FGFR4. The mutated residues in FGFR1 are indicated with an arrow, and the residue number is given. The corresponding residues in FGFR2-FGFR4 are boxed in blue along with the residue on FGFR1. Sequence identities of aligned proteins to FGFR1 are shown with a hyphen. (B) Residue distribution within the FGFR1 family at positions 546 and 576. The height of each bar is proportional to the total number of kinases with the corresponding residue at this position in the alignment. The color of the fragments within each bar denotes the organism from which kinases with the given residue have originated. The residues at positions 546 and 576 in the wild-type FGFR1 are indicated with a black arrow, and the residue in the mutant FGFR1 is indicated with a red arrow.

Fig. 4.

Models of the wild-type (Upper) and mutant (Lower) PDGFRA to demonstrate the effects of a 2-bp deletion. Both forms of PDGFRA have five extracellular Ig domains (Ig-like domains), a transmembrane domain (TM), a juxtamembrane domain (JM), and a bipartite tyrosine kinase catalytic domain (TK1 and TK2) separated by a kinase insert region (KI) and a C-terminal tail (CT). The N-terminal is labeled N. We identified a 2-bp deletion (AG shown in red text) within the codon encoding a serine residue. The deletion occurs at position 1048 in the final exon (exon 23) of the C-terminal tail. This deletion results in a 2-bp frameshift that introduces two alternative amino acid residues (CH underlined) and a premature STOP codon (indicated by ^*). The wild-type PDGFRA encodes a 1,089-residue protein, whereas our mutant PDGFRA encodes a 1,049-residue protein.