Specific blockade of Rictor-mTOR association inhibits mTORC2 activity and is cytotoxic in glioblastoma

Abstract

A small molecule which specifically blocks the interaction of Rictor and mTOR was identified utilizing a high-throughput yeast two-hybrid screen and evaluated as a potential inhibitor of mTORC2 activity in glioblastoma multiforme (GBM). In vitro, CID613034 inhibited mTORC2 kinase activity at submicromolar concentrations and in cellular assays specifically inhibited phosphorylation of mTORC2 substrates, including AKT (Ser-473), NDRG1 (Thr-346) and PKCα (Ser-657), while having no appreciable effects on the phosphorylation status of the mTORC1 substrate S6K (Thr-389) or mTORC1-dependent negative feedback loops. CID613034 demonstrated significant inhibitory effects on cell growth, motility and invasiveness in GBM cell lines and sensitivity correlated with relative Rictor or SIN1 expression. Structure-activity relationship analyses afforded an inhibitor, JR-AB2-011, with improved anti-GBM properties and blocked mTORC2 signaling and Rictor association with mTOR at lower effective concentrations. In GBM xenograft studies, JR-AB2-011 demonstrated significant anti-tumor properties. These data support mTORC2 as a viable therapeutic target in GBM and suggest that targeting protein-protein interactions critical for mTORC2 function is an effective strategy to achieve therapeutic responses.

Fig 1. Identification of compounds which inhibit mTORC2 activity in glioblastoma cells.

(A) Yeast two-hybrid assay configuration used to screen for inhibitors of the human Rictor-mTOR interaction. Full-length human Rictor fused to the Gal4-DNA-binding domain and mTOR fused to the Gal4-activation domain was expressed in yeast containing reporters harboring Gal4 upstream activating sequences (UAS). (B) Screening of compounds which block Rictor/mTOR association. Yeast expressing either p53-DBD and SV40 large T antigen-AD fusions or Rictor-DBD and mTOR-AD fusions were plated. Compounds were pinned onto the plate surfaces and examined for halo formation (red circles). Compounds which inhibited Rictor/mTOR-mediated growth on selective media while having little or no effects on p53/T antigen-mediated growth were considered specific. Compounds which blocked growth of both strains were considered nonspecific and exhibited general antifungal activity. (C) Structures of compounds which inhibit Rictor/mTOR association. (D) CID613034 inhibits mTORC2 in vitro kinase activity. mTORC2 kinase reactions were performed using GST-tagged AKT as a substrate with the indicated concentrations of inhibitor. Reactions were subsequently immunoblotted for phospho-S⁴⁷³-AKT and GST-tagged AKT. (E) CID613034 blocks binding of Rictor to mTOR in LN229 cells. Cells were treated with 5 mM Antimycin (control compound), 50 nM CID613034 or 20 nM rapamycin for 15 min and mTOR immunoprecipitated. Immunoprecipitations were then immunoblotted for the indicated proteins. (F) mTORC2 signaling is inhibited in GBM lines following 24 h exposure to CID613034. LN229 or U87 cells were treated with 100 nM of inhibitor as shown and lysates immunoblotted for the indicated proteins. Data shown are representative of experiments repeated two times.

Fig 2. Anti-GBM effects of CID613034 in cell lines.

(A) Inhibition of GBM cell line proliferation following culture with CID613034 (black, 0 μM; blue, 0.5 μM; red, 1 μM; green, 2 μM) for the indicated time points. (*, P< 0.05). Data represent mean ±S.D. of three independent experiments. (B) CID613034 exposure inhibits anchorage-independent growth. Cells were suspended in soft agar to evaluate anchorage independent growth in the presence of the indicated concentrations of inhibitor and colonies counted after 14 days of growth. Data represent mean +S.D. of three independent experiments. (C) Migration of U87 and LN229 cells in the presence of the indicated concentration of CID613034. Cells were seeded into Boyden chambers and allowed to migrate towards BSA (white bars), vitronectin (grey bars) or fibronectin (black bars) (*, P<0.05). Data represent mean +S.D. of three independent experiments. (D) Invasive potential of U87 or LN229 cells in the presence of the indicated concentrations of CID613034 migrating through matrigel. Data represent mean +S.D. of three independent experiments. (E) Cell-cycle phase distributions were determined on the indicated lines treated with CID613034 as shown. One of three experiments with similar results is shown. (F) Percent apoptotic cells as determined via annexin V-FITC staining. Data represent mean +S.D. of three independent experiments.

Fig 3. Sensitivity to CID613034 is dependent on Rictor or SIN1 expression.

(A) Steady-state Rictor expression levels in GBM lines. Lysates from the indicated lines were immunoblotted for Rictor and actin levels. (B) Quantification of Rictor expression levels from (A) as determined by densitometric analyses. Relative Rictor expression in U87 cells was arbitrarily assigned a value of 1. Rictor expression in all cell lines is shown as mean +S.D., n = 3. (C) XTT proliferation assays performed on eleven GBM lines with increasing concentration of CID613034 at 72 h. (D) Correlation between CID613034 sensitivity and IC₅₀ determined for all GBM cell lines treated with CID613034 for 72 h and shown as means of 3–5 individual experiments. Relative Rictor expression was obtained from (B) above. (E) Sensitivity of U87, U87shRictor and U87-Rictor GBM cells to CID613034 treated with the indicated concentrations at 72 h. Data represent mean ±S.D. of three independent experiments. *, P < 0.05 (F) Sensitivity of wt SIN1 and SIN1-null MEFs to increasing concentrations of CID613034 at 48 h. Data represent mean ±S.D. of three independent experiments. *, P < 0.05

Fig 4. Synthesis of JR-AB2-000 (CID613034) analogs.

Thirty analogs were synthesized with the indicated group modification shown and further detailed in Table 1. Analog identification numbers are shown below the structures corresponding to analogs where a particular functional group is modified. Detailed synthetic procedures and NMR spectra are described in supplementary Materials and methods.

Fig 5. JR-AB-000 and JR-AB-011 bind to Rictor and prevent Rictor-mTOR association.

(A) Surface plasmon resonance analysis of JR-AB-000 binding to immobilized Rictor, mSIN1 or mTOR as indicated (left panel). Binding sensorgrams of immobilized mTOR with Rictor over the indicated concentration range (right panel). The K_on, K_off and K_d were calculated by simultaneous non-linear regression using a 1:1 binding model and BIAevaluation 3.1 software. (B) Competitive binding curves of Rictor-mTOR association in the absence or presence of JR-AB2-011 or JR-AB2-000 as indicated (left panel). Analysis of selectivity of JR-AB2-011 (middle panel) or JR-AB2-000 (right panel) binding to Rictor, Raptor, mLST8 or Deptor as shown. Samples were preincubated with inhibitors and Rictor, Raptor, mLST8 or Deptor proteins as indicated and run over sensor chip containing immobilized mTOR. The IC₅₀ values were calculated using the response units at the dissociation phase. (C) mTOR-Flag coupled beads binding to myc-Rictor in the presence of increasing JR-AB2-011 (top panels) or JR-AB2-000 (bottom panels). myc-Rictor was incubated with inhibitor for 1 h followed by incubation with FLAG agarose beads coupled to mTOR-Flag (mTOR-Flag beads). Binding of myc-Rictor to mTOR-Flag beads (Rictor-mTOR-Flag beads) was detected by immunoblotting with an anti-myc mAb. The amount of protein bound to FLAG agarose beads was detected with an anti-Flag mAb (loading control). Immunoblots were quantified via densitometric analyses and graphs are shown to the right of the blots. Three independent experiments were performed and one representative result is shown.

Fig 6. Effects of JR-AB2-011 treatment on GBM tumor growth in mice.

(A) Tumor burden of SCID mice implanted with LN229 cells and treated with the indicated schedules of vehicle, JR-AB2-011 (4 mg/kg/d) and JR-AB2-011 (20 mg/kg/d) for ten consecutive days and tumor growth assessed every two days following initiation of treatment (start, day 0). *, P < 0.05, significantly different from vehicle, JR-AB2-011 (4 mg/kg/d) and JR-AB2-011 (20 mg/kg/d). (B) Overall survival of subcutaneous LN229 tumors receiving the indicated treatment schedules. (C) Ki-67 positive cells were identified via immunohistochemical staining of sections prepared from harvested tumors at day 12 following initiation of treatment regimens (upper left panel). Apoptotic cells were identified by TUNEL assays of sections prepared from harvested tumors at day 12 following initiation of treatment regimens (upper right panel). Data are expressed as the number of positive apoptotic bodies divided by high power field (hpf; 10–12 hpf/tumor). Values are means +S.D., *, P < 0.05. Phospho-S⁴⁷³-AKT/total AKT protein ratio levels in tumors (lower left panel). Values are means ±S.D., *, P < 0.05, significantly different from vehicle, JR-AB2-011 (4 mg/kg/d) and JR-AB2-011 (20 mg/kg/d). Phospho-T³⁸⁹-S6K/total S6K protein ratio levels in tumors (lower right panel). Values are means ±S.D., *, P < 0.05, significantly different from vehicle, JR-AB2-011 (4 mg/kg/d) and JR-AB2-011 (20 mg/kg/d). Protein levels were quantified by Western analyses of harvested tumors from mice with the corresponding treatments as indicated and described in Material and methods.