The therapeutic potential of γ-secretase inhibitors in the human acute leukemias has focused on T-ALL,1
where up to 50% of cases harbor activating Notch 1 mutations.26
The involvement of Notch in early B-cell malignancies was largely dismissed, based on Notch1 association with the T-cell lineage commitment and evidence that overexpression of ICNs or Hes1 induces apoptosis in pre-B cell lines.27
Nevertheless, there are previous reports of JAG1
expression in precursor-B ALL cells25
and Notch receptors, ligands and targets have been implicated at various stages of B-cell development.28
Notch dysregulation has been documented in mature B-cell neoplasms, including multiple forms of lymphoma, B-cell chronic lymphoid leukemia and multiple myeloma.5, 9, 29, 30, 31
In mice, retroviral-mediated gene disruption implicated Notch-related genes (Notch 2, presenilin 1, Dtx2, Numb and ADAM) in pro-B and pre-B leukemogenesis,32
while long-term GSI treatment led to altered maturation of peripheral B cells.33
Here, we definitively show that expression of both Notch ligands and receptors are common features of leukemic B-cell precursors. This conclusion is based on analysis of 11 institutional patients and 207 high-risk patients enrolled in COG P9906. We further show that GSI-I induces cell death in precursor-B ALL cells ex vivo
and in vivo
using cell survival assays and a NOD/SCID xenograft model.
There are clear differences between Notch gene expression patterns and mutational status in the two lymphoid forms of acute leukemia, precursor-B ALL and T-ALL. First, comparison of gene expression profiles reported of T-ALL cells treated with GSIs34
shows only a small percentage of overlap between the gene sets in these studies, most notably the Notch target Myc. Massive induction of HSP70 and other stress response genes was not seen in T-ALL cells treated with DAPT or compound E,11, 34
despite previous connections between presenilin function and the unfolded protein response.35
The unexpected induction of Hes1
by GSI treatment () was also reported by others for B-cell leukemia and acute myeloid leukemia cell lines.4 Hes1
expression is also linked to activation of the JNK stress response pathway, which is elevated after GSI-I treatment ().36
Another explanation may be cross-talk between the Notch and NFκB pathways,37
because IκBα can bind the Hes1
promoter and suppress Hes1
Active Notch may compete with IκBα to bind with NFκB, thereby enabling increased Hes1 expression. By 6 h treatment with GSI, several genes associated with the NFκB pathway have significantly altered expression levels in precursor-B ALL cells (Supplementary Tables 3A and B).
We found that transcription of most Notch target genes in precursor-B ALL is modest compared with that reported for T cells. As a result of low expression levels, confirmation by quantitative PCR methods is an important part of our study. We suspect that Notch signaling at this stage of B-cell development is held in check by inhibitory cis-interactions between receptors and ligands expressed in the same cell, as recently demonstrated in elegant studies by Sprinzak et al.39
We note also that γ-secretase may have other significant targets in precursor-B ALL leukemia cells. CD44 (ref. 8
) is one likely candidate that merits further study.
Others have raised the possibility that GSIs might inhibit other cellular proteases.3
Our study confirms that GSI-I also inhibits the proteosome, an important cancer therapeutic target.14, 15, 39
Importantly, both protease pathways are mechanistically linked to apoptosis. Caspase activation and oxidative stress was found to be synergistic in chronic lymphoid leukemia cells treated with the proteasome inhibitor NPI-0052.40
ROS production distinguishes the killing activity of GSI-I from bortezomib (), consistent with disruption of Notch-mediated anti-apoptotic pathways that protect mitochondrial function and prevent induction of pro-apoptotic Bcl2 family members like Bim41
The significant enhancement in ALL blast cell killing because of GSI-I's dual inhibitory properties follows the recent trend of combined, targeted therapeutic regimens. In T-ALL, recent studies in mice have suggested that the combination of dexamethasone with compound E could enhance toxicity while protecting against GSI-induced gut toxicity.40
Other promising studies in mice combined a 3-day on/4-day off GSI dosing schedule with an mammalian target of rapamycin inhibitor.43
Our animal studies also showed a significant protective effect on an intermittent schedule, supporting the concept that short-term or periodic treatment with inhibitors (like GSI-I) that target multiple proteases might sensitize precursor-B ALL leukemia cells to chemotherapeutic reagents with less severe toxicity to normal organs. With these findings, we wish to further explore the effects of GSI-I in additional pre-clinical studies.