Abstract 3367

Point mutations that trigger ligand-independent proteolysis of the Notch1 ectodomain occur frequently in human T-cell acute lymphoblastic leukemia (T-ALL) but are rare in murine T-ALL, suggesting that other mechanisms account for Notch1 activation in murine tumors. Using a panel of 13 murine T-ALL cell lines arising in a variety of T-ALL prone genetic backgrounds, we identified two types of acquired somatic deletions involving the 5' end of Notch1 in all cell lines screened. Type 1 deletions were found in 10 cell lines. These rearrangements remove exon 1 and the proximal promoter of Notch1, yet paradoxically activate transcription from a conserved initiator element positioned adjacent to a cryptic internal promoter in exon 25, roughly 40 kb downstream from the normal transcriptional start site. Expression of type 1 transcripts is associated with markedly increased histone H3 acetylation in the 3' end of Notch1, indicating that deletion of the proximal promoter and exon 1 leads to global remodeling of chromatin within the locus. Type 1 transcripts lack the normal Notch1 translation start site and initiate translation from a conserved methionine, M1727, which lies within the transmembrane domain of Notch1. A Notch1 polypeptide with M1727 for its N-terminus scored as a strong gain-of-function molecule in reporter gene assays and required gamma-secretase cleavage for activation. Cloning and sequencing of alleles with type 1 rearrangements showed that the responsible DNA breaks occurred immediately adjacent to germ line sequences resembling RAG recombinase signal sequences (RSSs), and that the DNA sequences at the point of joining bear all of the structural hallmarks of RAG-mediated rearrangements. Moreover, ChIP-Seq analyses showed that RAG2 associates with the proximal Notch1 promoter close to the ectopic RSSs in normal thymocytes in a distribution similar to that of the histone mark H3K4-me3, placing RAG at the “scene of the crime” in normal T cell progenitors. The conserved nature of the DNA breaks that produce type 1 deletions allowed for design of a PCR assay that detected this type of mutation in the majority of primary murine tumors, which also expressed high amounts of aberrant 3' Notch1 transcripts, as judged by a ratiometric q-RT-PCR assay capable of comparing the levels of 5' and 3' Notch1 transcripts. The remaining three cell lines had type 2 deletions, which removed sequences between exon 1 and exons 26–28 of Notch1, appeared to be created by non-homologous or microhomology mediated end joining, and were associated with the expression of transcripts in which exon 1 is spliced out-of-frame to 3' Notch1 exons. Translation of type 2 transcripts also initiated at M1727. Both types of deletions in murine tumors often occur concomitantly with mutations in the 3' end of Notch1 that remove a C-terminal PEST degron domain, and the combination of 5' deletions and 3' PEST mutations produced synergistic increases in Notch1 signal strength in reporter gene assays. The absence of intact Notch1 ectodomains in tumors with type 1 and type 2 deletions render murine T-ALLs resistant to agents such as inhibitory antibodies that target the extracellular domain of Notch1. Unlike murine T-ALL, 5' deletions were not detected by comparative genomic hybridization on SNP arrays in nearly 200 human T-ALLs, suggesting that activating deletions are rare in human disease. The absence of 5' Notch1 deletions in human T-ALL may be related to evolutionary divergence of the RSS sequences, which are not highly conserved in man. One human mutation that bears some similarities to the Notch1 deletions in murine T-ALL was found in the human T-ALL cell line CUTLL1. CUTLL1 contains a t(7;9) that fuses TCR-beta with the 3' end of Notch1. The fusion gene expresses a TCR J-beta 2.4-Notch1 fusion transcript that initiates translation from M1738, which is homologous to murine M1727. We conclude that like human T-ALL, murine T-ALL is often associated with acquired mutations that cause ligand-independent Notch1 activation. These findings further reinforce the central oncogenic role of aberrant Notch1 signaling in both human and murine T-ALL, and also point out some limitations in using murine T-ALL models for development of therapeutic agents that target Notch1.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution