Simultaneous inhibition of PI3K and PAK in preclinical models of neurofibromatosis type 2-related schwannomatosis

Abstract

Neurofibromatosis Type 2 (NF2)-related schwannomatosis is a genetic disorder that causes development of multiple types of nervous system tumors. The primary and diagnostic tumor type is bilateral vestibular schwannoma. There is no cure or drug therapy for NF2. Recommended treatments include surgical resection and radiation, both of which can leave patients with severe neurological deficits or increase the risk of future malignant tumors. Results of our previous pilot high-throughput drug screen identified phosphoinositide 3-kinase (PI3K) inhibitors as strong candidates based on loss of viability of mouse merlin-deficient Schwann cells (MD-SCs). Here we used novel human schwannoma model cells to conduct combination drug screens. We identified a class I PI3K inhibitor, pictilisib and p21 activated kinase (PAK) inhibitor, PF-3758309 as the top combination due to high synergy in cell viability assays. Both single and combination therapies significantly reduced growth of mouse MD-SCs in an orthotopic allograft mouse model. The inhibitor combination promoted cell cycle arrest and apoptosis in mouse merlin-deficient Schwann (MD-SCs) cells and cell cycle arrest in human MD-SCs. This study identifies the PI3K and PAK pathways as potential targets for combination drug treatment of NF2-related schwannomatosis.

Fig. 1. Combinatorial PI3K/PAK inhibition synergistically decreases MD-SC viability.

Screening results from human MD-SCs demonstrate loss of HS01 and HS03 cells viability with single drug treatments (A and D, EC50 or ECmax indicated with dotted line), with combination drug treatment (B, E), and synergy scores for each pair of combination treatments after 72 h (C, F). Similar screening performed on mouse MD-SCs shows loss of MS01 viability with single drug treatments (G), with combination drug treatment (H), and Loewe synergy scores for each pair of combination treatments after 48 h (I). In (B, E, H) color scale indicates loss of cell viability (light = high viability, dark = low viability) and numbers indicate percent of control cell viability. In (C, F, E) color scale indicates synergy (blue = synergy, red = antagonism) and numbers indicate percent difference in loss of viability compared to expected values (assuming no synergy).

Fig. 2. Combinatorial PI3K/PAK inhibition decreases orthotopic MD-MSC allograft growth rate and Weight, reduces cell proliferation, increase apoptosis, and reduces Akt phosphorylation and YAP levels in vivo.

A Representative bioluminescence images of signals at days 7, 14, and 21 post-cell injection of 5,000 MD-MSCs (corresponds with days 0, 7, and 14 of treatment; Avg. Radiance [p/s/cm²/sr]). All luminescent images are shown at identical scales selected for day 21 dpi; thus, no luminescence is seen at day 7 but was present and measurable indicating successful implantation of MD-MSC. B Tumor weight by treatment, excised tumors were significantly smaller in weight with pictilisib and combination treatments. C IVIS imaging fold change in bioluminescent signal over treatment duration significantly decreases with both PF and combination treatments. D Representative 400x magnification images of graft samples processed for histology and stained for hematoxylin and eosin, Ki67, Cleaved Caspase-3 (CC3), pPAK, pAkt and YAP. E Quantification of positive IHC staining demonstrates that both PAK inhibition alone, and combination treatment reduces cell proliferation in grafts as measured by Ki67 level. Data shown as individual data points with mean ± SD. Black bars in (B, C) indicate post-hoc significance between treatment groups over specific study increments or at study endpoint (p < 0.05). Black bars in E indicate post-hoc significance between treatment groups and vehicle controls (p < 0.05). Scale bars represent 200 µm.

Fig. 3. Combinatorial PI3K/PAK inhibition arrests human MD-SC at the G1 phase of the cell cycle.

A Graph of the distribution of cell cycle phases (gated for the live population) of all experimental runs as mean ± SD. **p < 0.01, *** p < 0.001, ****p < 0.0001 (B). Representative plots of the distribution of FITC (EdU) and FxCycle (DNA content)-labeled cells of 0.001% DMSO vehicle control and drug-treated cells. C Live cell imaging results of 72 h treatment at EC50 (concentrations noted in legend), including cell confluence (top) and fluorescent cleaved caspase 3/7+ signal area normalized to percent confluency.

Fig. 4. Pictilisib treatment comparison in human and mouse MD-SCs.

A Confluency measurements in two human MD-SCs (HS02 and HS05) and one mouse MD-SCs (MS01) treated with control DMSO (negative control), CUDC907 (positive control) and Pictilisib. Pictilisib significantly inhibited cell growth in all cell lines. B Cleaved Caspase-3/7 signals for human and mouse MD-SCs treated with Pictilisib (red) and CUDC907 (black). Pictilisib significantly induced CC-3/7 signal in mouse cells but did not in human MD-SCs. C Graphs presented all measured protein on apoptosis protein arrays, human on the right, mouse on the left. D Human Protein Profiler Apoptotic Array representative image, numbers correspond to the proteins in the graphs. E Mouse Protein Profiler Apoptotic Array representative image, numbers correspond to the proteins in the graphs. F Graphs representing protein levels detected by protein arrays, numbers correspond to the spots on the arrays; *p < 0.05; **p < 0.01, ****p < 0.0001 by Šídák’s multiple comparisons test.