Exclusion of ABCB8 and ABCB10 as cancer candidate genes in acute myeloid leukemia

Abstract

Acute myeloid leukemia (AML) exhibits great heterogeneity in manifestation, sensitivity to therapy and genetic basis of pathology. Numerous studies have implied that AML arises from the sequential accumulation of mutations at the level of the hematopoietic stem/progenitor cell, resulting in a disturbed balance of proliferation, differentiation and apoptosis of immature myeloid progenitors.1 To gain more insight into somatic mutations that affect biological pathways in tumorigenesis, a systematic mutation analysis of virtually all annotated human protein-coding genes in eleven breast and eleven colorectal cancers has been performed,2 leading to the identification of 280 genes as candidate cancer (CAN) genes. Importantly, several of the identified CAN genes are known to be mutated in hematological malignancies, including CBFA2T3 (ETO2), FBXW7, KRAS, NF1, PTEN, RUNX1T1 and TP53.1, 2 Several genes newly identified to be mutated in breast and colorectal cancers belong to the class of ATP-binding cassette (ABC) transporters,2 which are known to affect the response to therapy of AML and other cancers by conferring chemoresistance. Two of the ABC transporters discovered to harbor novel somatic mutations in breast cancer, the mitochondrial transporters ABCB8 and ABCB10, are highly expressed in both normal and leukemic hematopoietic stem cells.3 ABCB8 and ABCB10 are structurally closely related and are evolutionary highly conserved, suggesting a fundamental function in cellular physiology. Although their functions have not been completely elucidated, some clues regarding their biological roles can be postulated on the basis of homologues of ABCB10 in other species. Firstly, multi-drug resistance like 1 (MDL1), the ABCB10 homologue from Saccharomyces cerevisiae, regulates cellular responses to oxidative stress, such as reactive oxygen species (ROS).4 Reactive Oxygen Species signals stimulate proliferation and induce genetic instability in malignant transformed cells, conferring increased basal oxidative stress in cancer cells.5 ABCB10 might represent an important factor in mediating the response to oxidative stress in cancer cells. Secondly, the murine ABCB10 homologue, ABC-mitochondria erythroid (ABC-me), is a direct target of the transcription factor GATA1 and overexpression of murine ABCB10 enhances hemoglobin synthesis.6 Furthermore, a recent study identified ABCB10 to be essential for proliferation of human mammary epithelial cells.7 Thus, mutations in ABCB10 may cause resistance to oxidative stress, thereby contributing to cancer development. In this context, we pursued to investigate whether ABCB8 and ABCB10 are CAN genes in AML. To this end, the coding regions of ABCB8 and ABCB10 were screened by heteroduplex analysis (HA) using genomic DNA isolated from diagnostic bone marrow or peripheral blood samples from a panel of 94 AML patients who were referred to the Department of Hematology at the Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands, between 1991 and 2004. To identify potential mutations that affect splicing, the first 15 intronic nucleotides surrounding the intron–exon junctions were analyzed as well. Amplicons covering relevant regions were generated by a two-step polymerase chain reaction (PCR) followed by HA (for details see Supplementary Methods and Supplementary Figure S1 and 2; for primer sequences see Supplementary Table S1). Fragments exhibiting altered peak patterns in HA were subsequently sequenced to identify the nucleotide changes. Exons with multiple known single nucleotide polymorphisms (SNPs) were subjected to direct sequencing. A limitation of the HA method is that homozygous mutations or SNPs will not be detected. As all reported somatic mutations in ABCB8 and ABCB10 are heterogeneous,2 we reasoned that most point mutations in the cohort would be detected by HA.