Genetic dissection of the oncogenic mTOR pathway reveals druggable addiction to translational control via 4EBP-eIF4E

Abstract

We genetically dissect the contribution of the most prominent downstream translational components of mTOR signaling toward Akt-driven lymphomagenesis. While phosphorylation of rpS6 is dispensable for cancer formation, 4EBP-eIF4E exerts significant control over cap-dependent translation, cell growth, cancer initiation, and progression. This effect is mediated at least in part through 4EBP-dependent control of Mcl-1 expression, a key antiapoptotic protein. By using an active site inhibitor of mTOR, PP242, we show a marked therapeutic response in rapamycin-resistant tumors. The therapeutic benefit of PP242 is mediated through inhibition of mTORC1-dependent 4EBP-eIF4E hyperactivation. Thus, the 4EBP-eIF4E axis downstream of mTOR is a druggable mediator of translational control and Akt-mediated tumorigenesis that has important implications for the treatment of human cancers.

Figure 1. The Ability of PI3K-Akt-mTOR Signaling to Augment Protein Synthesis Is Necessary for Its Oncogenic Potential

(A) Schematic representation of the PI3K-Akt-mTOR pathway and the genetic strategy to restore protein synthesis to normal levels using a mouse minute mutant, haploinsufficient for ribosomal protein L24. (B) Cell size analysis of thymocytes from the indicated genotypes (**p < 0.00001; *p < 0.0009, n > 5/genotype). (C) Protein synthesis levels in thymocytes measured by [³⁵S]-methionine incorporation and TCA precipitation. The graph represents percentage of increase over WT levels (n = 6/genotype). *p = 0.0006 and **p = 0.008. (D) Transgenic CMV-Cap-HCV-IRES animals harboring a translational dicistronic luciferase reporter. (E) CMV-Cap-HCV-IRES mice were crossed with AKT^T and cap-dependent translation (Renilla luciferase activity) and IRES-mediated translation (Firefly luciferase activity) were measured (n = 3/genotype, p = 0.004; n.s., no statistical significance). (F) Kaplan-Meier curve showing lymphoma-free survival in AKT^T (n = 19) and AKT^T;L24^+/− (n = 14) mice (p = 0.0006). Data are presented as the average ± SEM. See also Figure S1.

Figure 2. Phosphorylation of 4EBP1, but Not rpS6, Is Required for Increased Protein Synthesis and Cell Size Control Downstream of Oncogenic Akt Signaling

(A) Schematic representation of the PI3K-Akt-mTOR pathway and the genetic strategy to specifically inhibit rpS6 phosphorylation or eIF4E hyperactivation. (B) Representative western blot analysis showing the hyperphosphorylation status of pAkt, p4EBP1, and prpS6 in AKT^T thymocytes. Western blot analysis of total Akt, 4EBP1, and rpS6 in AKT^T thymocytes is included. (C) A representative cap-binding assay to analyze eIF4E-eIF4G complex formation in thymocyte cell lysates of WT, AKT^T, AKT^T;4EBP1^M, and 4EBP1^M mice and densitometry of eIF4G/eIF4E ratio (n = 3/genotype) (*p < 0.05; n.s., no statistical significance). The eIF4E complex bound to the cap analog m⁷GTP-sepharose was pulled down and western blotted with antibodies against eIF4G and 4EBP1 revealed the relative amounts of eIF4E-associated proteins. The anti-4EBP1 antibody recognizes both the endogenous 4EBP1 and the 4EBP1^M proteins. (D) Protein synthesis rates in thymocytes derived from AKT^T, AKT^T;4EBP1^M, and AKT^T;rpS6^P−/− mice were measured by [³⁵S]-methionine incorporation and TCA precipitation represented as percentage of increase over WT levels; *p = 0.0006; **p = 0.001, n > 6/genotype. (E) Cell size analysis of thymocytes from the indicated genotypes; **p < 0.00001, n = 8; *p < 0.0007, n = 6. Data are presented as the average ± SEM. See also Figure S2.

Figure 3. Hyperactivation of eIF4E Is Essential for T Cell Lymphomagenesis Downstream of Oncogenic PI3K-Akt-mTOR Signaling, while Phosphorylation of rpS6 Is Dispensable

Kaplan-Meier analysis of lymphoma-free survival in two independent cohorts of AKT^T (n = 26) and AKT^T;4EBP1^M (n = 25) mice (A), p = 0.0001, and AKT^T (n = 19) and AKT^T;rpS6^P−/− (n = 18) mice (B), p = 0.3319. See also Figure S3.

Figure 4. Oncogenic PI3K-Akt-mTOR Signaling Leads to Enrichment of the DN3-4 Thymic Subpopulation, which Is Dependent on eIF4E Hyperactivity

(A) Percentage of DN3-4 thymocytes (prior to tumor formation compared to total DN thymocytes) in the thymus, as assessed by flow cytometry (n > 5/ genotype) (*p < 0.003; **p < 0.002; and n.s., no statistical significance). (B) Percentage of DN cell subsets in thymic tumors from AKT^T mice (n = 5). (C) Flow cytometry analysis of all thymocyte subpopulations in thymic tumors from AKT^T mice (n = 5). Data are presented as the average ± SEM. See also Figure S4.

Figure 5. eIF4E Hyperactivation Downstream of Oncogenic PI3K-Akt-mTOR Signaling Impairs Programmed Cell Death of Specific T Cell Precursors, at Least in Part, through Control of Translation of Mcl-1

(A) Percentage of Annexin V⁺ DN3-4 cells prior to tumor formation (n > 3 mice/genotype) (*p < 0.001, **p < 0.02, and ***p < 0.003). (B and C) Analysis of Mcl-1 protein expression levels in total thymocytes. (Right) Representative western blot. (Left) Densitometry analysis (*p = 0.005). All representative western blots were loaded, resolved, and transferred from the same gel. (D) Percentage of increase of Mcl-1 expression in AKT^T and AKT^T;4EBP1^M in the DN3-4 compartment compared to WT controls (n = 4/genotype) (*p = 0.01). (E) Fold change of Annexin V⁺ in DN3-4 cells after knockdown of Mcl-1 (n > 3 mice/genotype) (*p < 0.01 and n.s., no statistical significance). Data are presented as the average ± SEM. See also Figure S5.

Figure 6. eIF4E Hyperactivation Is Important for AKT^T Tumor Maintenance

4EBP1^M was induced in AKT^T;4EBP1^M mice at 10 weeks of age. (A) Average tumor volumes (p < 0.03). Data are presented as the average ± SEM. (B) AKT^T and AKT^T;4EBP1^M mice that were imaged in a GE 3T MRI at 10 and 14 weeks. Images from the same AKT^T and AKT^T; 4EBP1^M mouse at 10 and 14 weeks.

Figure 7. PP242, but Not Rapamycin, Has a Significant In Vivo Anti-Tumor Effect on AKT^T Tumor Allografts

(A) AKT^T thymocytes were injected subcutaneously and allowed to form tumors for 20 days. 30 tumor-bearing mice were randomized into three groups of 10: vehicle, 5 mg/kg rapamycin, or 100 mg/kg PP242. Treatments were given 7 days a week for 20 days. (B) Representative mice from each cohort after 20 days of therapy. Dotted yellow lines, primary tumor; dotted green lines, multiple sites of lymph node metastasis; green arrows (inset), areas of significant cervical lymphadenopathy. (C) Tumor areas in square millimeters. Black line, vehicle; blue line, rapamycin; red line, PP242. (D) Mouse weights in grams during the course of treatment. Data are presented as the average ± SEM.

Figure 8. PP242 Shows a Better Therapeutic Effect Than Rapamycin In Vivo, at Least in Part, through Inhibition of the 4EBP-eIF4E Axis

(A) Pharmacodynamic analysis of mTOR targets through western blot analysis. Each lane represents one post-treatment tumor 20 days after initiation of therapy. (B) Protein synthesis levels in tumor thymocytes treated with DMSO, rapamycin (50 nM), and PP242 (2.5 μM) ex vivo for 2 hr measured by [³⁵S]-methionine incorporation and TCA precipitation (n = 6/genotype) (*p = 0.001 and **p = 0.001). (C) Fold decrease in cell survival compared to WT as determined by propidium iodide/Annexin exclusion (n = 10/cohort) (*p < 0.01 and **p < 0.05). (D) Cell-cycle analysis of post-treatment tumor samples (n = 3/cohort). (E) Fold change in live AKT^T and AKT^T;4EBP1^M tumor cells after 24 hr treatment with PP242 (2.5 μM) ex vivo (n = 3) (*p < 0.001; n.s., no statistical significance). Data are presented as the average ± SEM.