Combined Inhibition of NEDD8-Activating Enzyme and mTOR Suppresses NF2 Loss-Driven Tumorigenesis

Abstract

Inactivation of NF2/Merlin causes the autosomal-dominant cancer predisposition syndrome familial neurofibromatosis type 2 (NF2) and contributes to the development of malignant pleural mesothelioma (MPM). To develop a targeted therapy for NF2-mutant tumors, we have exploited the recent realization that Merlin loss drives tumorigenesis by activating the E3 ubiquitin ligase CRL4^DCAF1, thereby inhibiting the Hippo pathway component Lats. Here, we show that MLN4924, a NEDD8-activating enzyme (NAE) inhibitor, suppresses CRL4^DCAF1 and attenuates activation of YAP in NF2-mutant tumor cells. In addition, MLN4924 sensitizes MPM to traditional chemotherapy, presumably as a result of collateral inhibition of cullin-RING ubiquitin ligases (CRL) involved in DNA repair. However, even in combination with chemotherapy, MLN4924 does not exhibit significant preclinical activity. Further analysis revealed that depletion of DCAF1 or treatment with MLN4924 does not affect mTOR hyperactivation in NF2-mutant tumor cells, suggesting that loss of Merlin activates mTOR independently of CRL4^DCAF1 Intriguingly, combining MLN4924 with the mTOR/PI3K inhibitor GDC-0980 suppresses the growth of NF2-mutant tumor cells in vitro as well as in mouse and patient-derived xenografts. These results provide preclinical rationale for the use of NAE inhibitors in combination with mTOR/PI3K inhibitors in NF2-mutant tumors. Mol Cancer Ther; 16(8); 1693-704. ©2017 AACR.

Figure 1. MLN4924 attenuates CRL4^DCAF1 activity and reduces the proliferation of NF2-mutant schwannoma cells

(A) 293T cells transfected with Flag-HA-tagged DCAF1 (FH-DCAF1) were treated with the indicated concentrations of MLN4924 for 4 hours and then lysed in RIPA buffer. Input and FLAG immunoprecipitates were immunoblotted as indicated. (B) COS-7 cells were transfected with FH-DCAF1 and Myc-Ubiquitin as indicated and then treated with MLN4924 at the indicated concentrations for 4 hours. FLAG immunoprecipitates and input were immunoblotted as indicated. Red asterisk indicates putative auto-ubiquitylated DCAF1. Red arrowhead indicates putative CRL4^DCAF1 substrates. (C) In vivo Lats1 ubiquitylation assay. 293T cells were transfected as indicated. 24 hours following transfection, cells were treated with MLN4924 (8 hours as indicated) and the proteasome inhibitor MG132 (4 hours, all samples) prior to lysis in a denaturing buffer. Input and Ni-NTA precipitates were immunoblotted as indicated. (D) FC-1801 cells were treated with the indicated concentrations of MLN4924 for 8 hours, lysed, and immunoblotted as indicated. (E) The Nf2-null mouse schwannoma line FC-1801 and the Nf2-wild-type immortalized adult Schwann cell line FH-912 were treated with increasing doses of MLN4924 and subjected to MTT proliferation assay after 72 hours of treatment. Proliferation of MLN4924-treated cells was normalized to respective DMSO-treated controls. The concentrations of MLN4924 required to inhibit cell proliferation by 50% (GI₅₀) are indicated.

Figure 2. MLN4924 rescues activity of the Lats Hippo pathway kinases and blocks cell cycle progression

(A) The NF2-mutant mesothelioma line Meso-33 was treated with the indicated concentrations of MLN4924 for 24 hours and immunoblotted as indicated. (B) Meso-33 cells were transfected with control or DCAF1 siRNAs and subsequently treated for 16 hours with DMSO or the indicated concentrations of MLN4924 and lysed 72 hours post-transfection for immunoblotting as indicated. P-YAP^S127/YAP fold change based on densitometric analysis is indicated. (C) Meso-33 cells were transfected with indicated siRNAs and harvested after 80 hours (left) or treated with the indicated concentrations of MLN4924 for 36 hours (right) for immunoblotting as indicated. (D) Meso-33 cells were treated with 0.5 μM MLN4924 for the indicated times and immunoblotted as indicated. (E) NF2-positive MPM lines and immortalized but untransformed mesothelial lines (Met5a and LP9) were treated with increasing doses of MLN4924 and subjected to MTT proliferation assay after 72 hours of treatment. Proliferation of MLN4924-treated cells was normalized to respective DMSO-treated controls. Error bars are ±SEM (n=3). (F) NF2-mutant MPM lines were treated as in (E).

Figure 3. NAE inhibition cooperates with first-line MPM chemotherapy

(A) Meso-33 and VAMT were treated with vehicle or the indicated drugs at their respective GI₅₀ concentrations for 24 hours. Cell lysates were immunoblotted as indicated. (B) Meso-33 and VAMT cells were treated with vehicle or the indicated drugs at twice their respective GI₅₀ concentrations for 72 hours and then subjected to an Annexin-5/PI apoptosis assay. Error bars indicate ±SEM (VAMT, n=3; Meso-33, n=2). (C) 15 × 10⁶ VAMT cells and (D) 8 × 10⁶ Meso-10 cells were injected bilaterally in the rear flanks of NSG mice. After xenografts exceeded 100 mm³, mice were treated with vehicle or the indicated drugs. All cisplatin treatments were 3 mg/kg (intraperitoneally, once every 7 days), pemetrexed treatments 100 mg/kg (intraperitoneally, once daily, 5 days on, 2 days off), and MLN4924 treatments 90 mg/kg (subcutaneously, twice per day, three times per week M-W-F). A 7-day holiday was provided between chemotherapy regimens while the MLN4924 was administered continuously. Data are means ± SEM (n≥12). **P < 0.01, ***P < 0.001, unpaired T-test.

Figure 4. MLN4924 combined with mTOR/PI3K inhibition causes complete MPM tumor growth suppression and significantly reduces the growth of NF2-mutant schwannoma

(A) Merlin deficient Meso-33 and VAMT cells were transfected with control or DCAF1 siRNA and harvested after 72 hours for immunoblotting as indicated. (B) FC-1801 cells were transfected with Flag-HA-tagged (FH) Merlin or empty vector and control or DCAF1 siRNAs as indicated. Cells were also treated with the indicated concentration of MLN4924 or DMSO for 4 hours prior to lysis and immunoblotted as indicated. (C) The indicated Merlin-deficient MPM lines were treated with DMSO (left) or 5 nM rapamycin (right) for 4 hours and immunoblotted as indicated. (D) VAMT cells were treated with the indicated concentrations of GDC-0980 for three hours and immunoblotted as indicated. (E) 8–10 × 10⁶ VAMT cells were injected bilaterally in the rear flanks of NSG mice. After xenografts exceeded 100 mm³, mice were treated with vehicle or the indicated inhibitors. MLN4924 treatments were 120 mg/kg (subcutaneously, twice per day, three times per week M-W-F). GDC-0980 treatments were 5 mg/kg (oral, once per day, five times per week). Data are means ± SEM (n=10). ***P < 0.001, unpaired T-test. (F) Primary Schwann cells from normal individuals (NF2 WT, left) and schwannoma cells from NF2 patients (NF2^−/−, right) were treated with DMSO or 100 nM MLN4924 for 72 hours. Graph indicates the percentage of proliferating cells based on immunostaining for Ki-67 positivity. Data are means ± SEM (NF2^+/+ n=3; NF2^−/− n=4). ***P < 0.001, unpaired T-test. (G) Primary NF2^−/− human schwannoma cells were treated with DMSO or the indicated concentrations of MLN4924, GDC-0980, or a combination of the two inhibitors as indicated. Ki-67⁺ cells were counted as in (F) and the graph indicates Ki-67⁺ relative to DMSO controls. Data are means ± SEM (n=3). *P < 0.05, unpaired T-test. (H and I) On-target mTORC1 and NAE inhibition in MSK-LX19 PDX. NSG mice bearing 200 mm³ MSK-LX19 PDXs were treated once with vehicle or the same doses of the indicated inhibitors as in (E). Tumors were isolated 7 hours after drug administration. (H) Representative images of tumors immunostained for phospho-S6 (P-S6) and counterstained with hematoxylin (H). (I) Lysates from the same tumors as (H) were immunoblotted as indicated to detect neddylated cullins (n=2). (J) Serially transplanted MSK-LX19 tumors were injected unilaterally in the rear flanks of NSG mice. After tumors exceeded 100 mm³, mice were treated as in (E). Data are means ± SEM (n=7). *P < 0.05, unpaired T-test.

Figure 5. Signaling pathway and mechanistic effects of MLN4924 and GDC-0980 in NF2/Merlin-mutant tumors

In NF2/Merlin-mutant tumors, dysregulated CRL4^DCAF1 and mTORC1 are two predominant oncogenic signaling nodes sustaining tumor growth. MLN4924 inhibits YAP activity in multiple Merlin-mutant tumors and transcription of canonical pro-proliferative and anti-apoptotic YAP targets in mesothelioma. GDC-0980 inhibits mTOR and PI3K, thereby inhibiting downstream mechanisms supporting tumor growth and maintenance including proliferation, survival, protein synthesis, and cell growth. Tumor-promoting mechanisms are indicated in red, tumor-suppressive mechanisms in blue.