mTORC1 inhibition with RAD001 induces Akt activation in primary AML samples by activation of the IGF-1/IGF1-R signaling pathway, dependent on an IGF-1 autocrine loop. (A) Bone marrow (BM) blast cells from 19 patients with AML were starved for 4 hours in cytokine and serum-free minimal essential medium (MEM), with or without the following kinase inhibitors: 25 μmol/L LY294002, 10 μmol/L IC87114, or 10 nmol/L RAD001, added during the last hour of starvation. Western blots (WBs) were performed with anti-phospho-Akt (ser 473), anti-phospho-p70S6K (thr 389), and anti-Akt antibodies. Quantification of phospho-Akt signal intensity was normalized to Akt signal intensity. Each histogram of the graph represents the phospho-Akt signal intensity in RAD001-treated blast cells, expressed as percentage of signal intensity in control cells (M, medium without inhibitors). (B) BM blast cells from patients G192 and G194 were collected after Ficoll-Hypaque density gradient separation, then washed once in PBS buffer. Blast cells (5 × 105/mL) from patient G192 were starved for 4 hours in cytokine and serum-free MEM, then incubated without or with 10 nmol/L RAD001 for 1 hour. In independent experiments, 5 × 105/mL blast cells from patients G192 and G194 were incubated without or with RAD001 for 24 hours, in α-MEM with 10% fetal calf serum (FCS). RAD was used at 10 nmol/L for sample G192, and at 10, 50, or 100 nmol/L for sample G194. WBs were performed with anti-phospho-Akt (ser 473), anti-phospho-p70S6 kinase (thr 389), anti-phospho-S6R (Ser 235/236), and anti-Akt antibodies. (C) RAD001 increases IGF-1-stimulated Akt phosphorylation in AML blast cells. BM blast cells from patients G179, G194, and G149 were starved for 4 hours in serum-free MEM, without or with 10 nmol/L RAD001, then stimulated or not with 50 ng/mL IGF-1 for 10 minutes. WBs were performed with anti-phospho-Akt (Ser 473) and anti-Akt antibodies. (D) AML blast cells express IGF-1 at the RNA and protein level. BM blast cells from 8 patients were highly purified by flow cytometry cell sorting according to CD45low expression and side scatter. The MOLM-14 AML cell line was used as negative control, and the OPM2 myeloma cell line was used as a positive control for IGF-1 expression. IGF-1 mRNA expression was quantified in the purified blast cells by quantitative RT-PCR, and their levels were expressed relative to HPRT (hypoxanthine phosphoribosyl transferase) mRNA levels. Similar results were obtained using another housekeeping gene, UBCV (C ubiquitin; data not shown). Immunofluorescence staining was performed on purified blasts for the same 8 patients mentioned above, and on MOLM-14 and OPM2 cell lines, using a mouse monoclonal anti-IGF-1 antibody and fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse antibody. Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI). Images obtained from the representative patient G102 and from MOLM-14 and OPM2 cell lines are presented. (E) The mTORC1-mediated positive feedback on PI3K/Akt activity involves the IGF-1 receptor. BM blast cells from patients G179, G194, G199, and G205 were starved for 4 hours in serum-free MEM. Cells were incubated in the following conditions: medium alone, 10 nmol/L RAD001 for 1 hour, 10 nmol/L RAD001 for 1 hour plus, and 5 μg/mL αIR3 (added 30 minutes before RAD001). WBs were performed with anti-phospho-Akt (Ser 473) and anti-Akt antibodies. αIR3 is a blocking mouse monoclonal antibody directed against the alpha subunit of the IGF-1 receptor and was obtained from Calbiochem. (F) mTORC1 inhibition by RAD001 increases the expression of the IRS2 adaptor. BM blast cells from patients G72, G99, G189, and G191 were starved for 4 hours in serum-free MEM then incubated with or without RAD001 (10 nmol/L) for 1 hour. WBs were performed with anti-IRS2 and anti-actin antibodies.
