High-level IGF1R expression is required for leukemia-initiating cell activity in T-ALL and is supported by Notch signaling

Abstract

T cell acute lymphoblastic leukemia (T-ALL) is an aggressive cancer of immature T cells that often shows aberrant activation of Notch1 and PI3K-Akt pathways. Although mutations that activate PI3K-Akt signaling have previously been identified, the relative contribution of growth factor-dependent activation is unclear. We show here that pharmacologic inhibition or genetic deletion of insulin-like growth factor 1 receptor (IGF1R) blocks the growth and viability of T-ALL cells, whereas moderate diminution of IGF1R signaling compromises leukemia-initiating cell (LIC) activity as defined by transplantability in syngeneic/congenic secondary recipients. Furthermore, IGF1R is a Notch1 target, and Notch1 signaling is required to maintain IGF1R expression at high levels in T-ALL cells. These findings suggest effects of Notch on LIC activity may be mediated in part by enhancing the responsiveness of T-ALL cells to ambient growth factors, and provide strong rationale for use of IGF1R inhibitors to improve initial response to therapy and to achieve long-term cure of patients with T-ALL.

Figure 1.

IGF1R is expressed broadly in human and mouse T-ALL. (A and B) Western blot and (C and D) flow cytometric analysis of total and surface IGF1R protein expression, respectively, from human cell lines (A and C), primary mouse leukemias (B) derived by retroviral transduction/transplantation of bone marrow with an activated form of Notch1 termed ΔE, and xenograft-expanded primary human samples (D). Western blot controls in (B) are mouse embryonic fibroblasts derived from IGF1R^null mouse embryos (R−) and the same cells stably transfected with an IGF1R cDNA expression construct (R+). At least 20,000 events were collected within each gate for all flow cytometry assays. Data depicted are representative of at least two independent experiments.

Figure 2.

Pharmacologic inhibition of IGF1R blocks growth of T-ALL cells. Flow cytometric analysis of cell proliferation by BrdU incorporation after treatment with a small molecule IGF1R inhibitor (BMS-536924) versus DMSO vehicle (mock), or IGF1R blocking antibody (αIR3), for 48–72 h in vitro. (A) Three representative independent primary mouse Notch1(ΔE) leukemias (#324, #327, #329). (B) Four independent xenograft-expanded primary human T-ALL samples (D115, K419, K424, and M69). (C) Two human T-ALL cell lines (ALLSIL and HPBALL). Error bars indicate standard deviation for assays performed in triplicate. Data depicted are representative of at least three independent experiments.

Figure 3.

Primary mouse T cell leukemias are generated efficiently by activated Notch1 despite reduced IGF1R expression. (A) Schematic of the IGF1R^neo allele. The retained neo cassette within the second intron results in reduced expression of full-length IGF1R protein in IGF1R^neo/neo mice (Holzenberger et al., 2000). (B) Survival of mice transplanted with retroviral Notch1(ΔE)-transduced bone marrow from either WT (IGF1R^+/+; n = 6) or IGF1R hypomorph (IGF1R^neo/neo; n = 14) donor animals. ***, P < 0.0001 (Log-rank test). (C) Spleen and liver organ weights at necropsy of individual morbid mice transplanted with Notch1(ΔE)-transduced bone marrow from IGF1R^+/+ (n = 6) and IGF1R^neo/neo (n = 11) backgrounds. Error bars indicate standard deviation. (D) Immunophenotypic analysis of representative primary mouse ΔE leukemias derived on the IGF1R^neo/neo background. (E) Western blot analysis of total IGF1R protein expression in representative primary mouse ΔE leukemias on IGF1R^+/+, IGF1R^neo/+, and IGF1R^neo/neo backgrounds. R+ and R− mouse embryonic fibroblast staining controls and Erk2 loading control are indicated. Data depicted in D and E are representative of at least seven independent samples.

Figure 4.

Mouse T cell leukemias with reduced IGF1R expression are defective in serial transplantation. Survival of mice transplanted with primary mouse ΔE leukemia cells from either IGF1R^+/+ or IGF1R^neo/neo backgrounds by i.v. (A) intrafemoral (IF; B) injection route. Each numbered sample represents a different primary leukemia from mice in Fig. 3 B injected into secondary recipients. Raw survival data, including numbers of animals in each cohort, are provided in Table S1. Data depicted are collated from four independent transplantation experiments.

Figure 5.

Transplantable clones with reduced IGF1R expression show compensatory PI3K–Akt activation. (A) Western blot and (B) flow cytometric analysis of PTEN protein expression in primary and transplantable secondary ΔE-IGF1R^neo leukemias. (C and D) Akt activation as measured by intracellular phospho-Akt(Ser473) flow cytometry in response to stimulation with FBS (C) or (D) recombinant IGF-1 × (D) 10 min after 5-h serum starvation. (E) Steady-state level of Akt activation as measured by pAkt(Ser473) in transplantable ΔE-IGF1R^+/+ and ΔE-IGF1R^neo (#3105) leukemias (n = 3 independent mice for each cohort). For C and D, the five ΔE-IGF1R^neo transplantable clones shown are (4 x #3115, 1 x #3112). Error bars indicate standard deviation. *, P < 0.05; **, P < 0.01 (Student’s t test). Data depicted in A and B are representative of at least three independent experiments. Data depicted in C–E are representative of two independent experiments and include at least three independent mice per cohort per experiment.

Figure 6.

Inhibition of Notch signaling with GSI down-regulates IGF1R expression in human T-ALL cells. (A) qRT-PCR analysis of IGF1R mRNA in human T-ALL cell lines treated in vitro with γ-secretase inhibitor (GSI) to block Notch signaling versus DMSO vehicle (mock) for 2–10 d. Error bars indicate standard deviation for assays performed in triplicate. (B and C) Flow cytometric analysis of surface IGF1R expression by human T-ALL cell lines treated with GSI versus DMSO vehicle for 6–8 d. Data are representative of multiple replicates. (D) Flow cytometric analysis of surface IGF1R expression by xenograft-expanded primary human T-ALL cells. Cells were cultured on MS5-DL1 feeders to stimulate Notch signaling versus MS5 control feeders and then treated with 1.0 µM compound E (GSI) for 2–4 d to block Notch signaling. Flow histograms for a representative case are depicted on the left, and results from six different patient samples are summarized on the right. Error bars indicate standard deviation. **, P < 0.01 (Student’s t test). At least 20,000 gated live events were collected for all flow cytometry assays.

Figure 7.

Inhibition of Notch signaling with dominant-negative MAML1 and activation by ICN1 or DL1 ligand confirm IGF1R regulation by Notch. Flow cytometric analysis of surface IGF1R expression. (A) Human T-ALL cell lines were transduced with dominant negative Mastermind-like-1 retrovirus (DN-MAML1-GFP) or empty virus control (MigRI). (B) Human T-ALL cell lines were transduced with ICN retrovirus (Mig ICN) and then treated with 1.0 µM compound E (GSI) for 4 d (HPBALL and TALL-1) or 8 d (PF382) to block endogenous Notch signaling or DMSO vehicle (mock). Retrovirally transduced cells in A and B were discriminated from nontransduced cells by gating for GFP. Data depicted in A and B are representative of multiple replicates. (C) Xenograft-expanded primary human T-ALL cells were cultured in vitro on MS5-DL1 feeders to stimulate Notch signaling versus control MS5 feeders. CD45 expression levels were also assessed simultaneously with IGF1R by flow cytometry. Flow histograms for a representative case are depicted on the left, and results from six different patient samples are summarized on the right. Error bars indicate standard deviation. *, P < 0.05 (Student’s t test). At least 20,000 live events were collected within each gate for all flow cytometry assays.

Figure 8.

ICN1/CSL binds dynamically to a site within intron 20 of human IGF1R. (A) Alignment of sequencing reads over intron 20 of the IGF1R locus from ChIP libraries prepared from the human T-ALL cell line, CUTLL1, with antibodies specific for Notch1 and CSL as compared with input cells. (B) Genomic DNA sequence from human IGF1R intron 20 (NT_010274.17: 14461410–14461610) with sequence-paired ICN1/CSL binding sites highlighted in bold and 17-bp spacer underlined. (C) ChIP/qPCR analysis of IGF1R intron 20 from three different human T-ALL cell lines using a Notch1-specific antibody as compared with preimmune antiserum. (D) ChIP/qPCR analysis of IGF1R intron 20 from CUTLL1 cells using antibodies specific for Notch1 and CSL compared with control rabbit IgG. Cells were treated with GSI for 3 d to block Notch signaling (GSI x 3d) versus DMSO vehicle (DMSO). GSI was washed out, and cells were harvested 4 h later (GSI x 3d, then wash 4 h). Quantitation of immunoprecipitated DNA is expressed relative to input DNA (% input). Error bars indicate standard deviation for qPCR assays performed in triplicate. Numbers above bars in D indicate relative enrichment over control. ChIP libraries were prepared in duplicate, and local ChIP/qPCR analyses were performed twice. Representative results are shown.

Figure 9.

Notch-induced IGF1R expression enhances PI3K–Akt signaling in response to IGF1. (A) Flow cytometric analysis of intracellular phospho-Akt levels. Cells were treated with GSI versus DMSO vehicle for 6–8 d, serum starved overnight, and then pulsed with recombinant IGF1 for 10 min before assay. (B) Western blot analysis of PTEN protein expression in cells treated with GSI versus mock for 4 d. (C and D) Flow cytometric analysis for intracellular phospho-Akt levels (left) and surface IGF1R levels (right). Cells were retrovirally transduced with Mig IGF1R, treated with GSI versus DMSO vehicle for 4 d, serum starved overnight, and then pulsed with recombinant IGF1 for 10 min before phospho-Akt assay. Transduced versus nontransduced cells within the same culture were distinguished by gating for GFP. Surface IGF1R expression level was also assessed by flow cytometry immediately before stimulation with IGF1. Filled grey histograms represent second antibody staining controls. At least 20,000 events were collected within each gate for all flow cytometry assays. Data depicted are representative of at least two independent experiments.