Computational drugs repositioning identifies inhibitors of oncogenic PI3K/AKT/P70S6K-dependent pathways among FDA-approved compounds

Abstract

The discovery of inhibitors for oncogenic signalling pathways remains a key focus in modern oncology, based on personalized and targeted therapeutics. Computational drug repurposing via the analysis of FDA-approved drug network is becoming a very effective approach to identify therapeutic opportunities in cancer and other human diseases. Given that gene expression signatures can be associated with specific oncogenic mutations, we tested whether a "reverse" oncogene-specific signature might assist in the computational repositioning of inhibitors of oncogenic pathways. As a proof of principle, we focused on oncogenic PI3K-dependent signalling, a molecular pathway frequently driving cancer progression as well as raising resistance to anticancer-targeted therapies. We show that implementation of "reverse" oncogenic PI3K-dependent transcriptional signatures combined with interrogation of drug networks identified inhibitors of PI3K-dependent signalling among FDA-approved compounds. This led to repositioning of Niclosamide (Niclo) and Pyrvinium Pamoate (PP), two anthelmintic drugs, as inhibitors of oncogenic PI3K-dependent signalling. Niclo inhibited phosphorylation of P70S6K, while PP inhibited phosphorylation of AKT and P70S6K, which are downstream targets of PI3K. Anthelmintics inhibited oncogenic PI3K-dependent gene expression and showed a cytostatic effect in vitro and in mouse mammary gland. Lastly, PP inhibited the growth of breast cancer cells harbouring PI3K mutations. Our data indicate that drug repositioning by network analysis of oncogene-specific transcriptional signatures is an efficient strategy for identifying oncogenic pathway inhibitors among FDA-approved compounds. We propose that PP and Niclo should be further investigated as potential therapeutics for the treatment of tumors or diseases carrying the constitutive activation of the PI3K/P70S6K signalling axis.

Keywords: FDA-approved drugs; PI3K-dependent pathways; drugs network; gene expression signatures; oncogenes.

Figure 1. Computational drug repositioning identifies possible inhibitors of the PI3K-dependent pathway

A. Schematic representation of the drug repositioning approach used in this study. Reverse (1 & 2) and forward (3) oncogenic PI3K-specific gene expression signatures derived from isogenic cell lines generated new nodes in MANTRA network (See text and Experimental Procedures). Drugs are organized in a network of nodes (drugs) and edges (similarities) highlighting “communities” of drugs sharing a similar MoA. Edges display drugs with significant computed distance. Different colors refer to different drug communities. B. Drugs common to networks 1, 2 and 3 were selected and a new node was identified, here referred to as a PI3K intersection (PI3K int).

Figure 2. Niclo and PP effectively control oncogenic PI3K-dependent gene expression

A. and B. Total RNA was harvested from wild type HME(WT) or an isogenic clone carrying a PIK3CA(E545K) mutation (KI) after treatment (Trts) with vehicle (DMSO (D)), Niclo or PP. The most differentially expressed genes were predicted and selected as described in M&M. The expression of genes down-regulated by oncogenic PIK3CA (hCCNG2, hHBP1, hIL8, hHEY1, hTGFA and hCSGALNACT1) A. or up- regulated by oncogenic PIK3CA (hCCNE2, hDSCC1, hENDOD1 and hPPP2R1B) B. and selected WNT target genes (hAXIN1 and hAXIN2) C. was analyzed by quantitative real time PCR. Data indicate absolute values ± SD and are the average of at least three independent experiments. Statistical significance was evaluated through t-test, performed for the following conditions: DMSO-treated WT Vs DMSO-treated KI or DMSO-treated KI Vs Drugs-treated KI. **p < 0.05, ***p < 0.01, ns: not significant. C.L.: Cell Line.

Figure 3. Niclo and PP inhibit the phosphorylation of PI3K-dependent molecular targets

A. Immunoblot analysis of wild type human mammary epithelia HME cells (PIK3CA-WT) or isogenic cells carrying the oncogenic PIK3CA(E545K) mutation (PIK3CA-E545K): cells were treated for two hours with DMSO (D) or Niclo (N) or PP. Cell extracts were analyzed with the indicated antibodies. B. Immunoblot analysis of wild type human mammary epithelia HME cells (PIK3CA-WT) or isogenic cells carrying oncogenic PIK3CA(E545K) mutations (PIK3CA­(E545K): cells were serum starved overnight and stimulated for two hours with 20% serum (S) prior to harvesting. Equal amounts of protein extracts were analyzed with the indicated antibodies. Before adding serum, cells were pre-treated for 30 minutes with DMSO (D), Niclo (N), PP or LY294002 (LY) inhibitors. Data are representative of at least three independent experiments C. and D. Total RNA was harvested from wild type PIK3CA HME cells (WT) or the isogenic clone carrying the PIK3CA(E545K) mutation (KI) after treatments with DMSO (D), methanol (M), Niclo, Rapamycin (Rap) or PF-05212384 (PF) as described in Materials and Methods. Methanol was the vehicle for Rapamycin. The expression of genes down-regulated by oncogenic PI3K (e.g., hCCNG2, hHBP1 and hIL8) (C) or updown-regulated by oncogenic PI3K (e.g., hCCNE2, hENDOD1 and hPPP2R1B) (D) was analyzed by quantitative real time PCR. Data indicate absolute values ± SD and are the average of at least three independent experiments. Statistical significance was evaluated through t-tests, performed under the following conditions: D-treated KI Vs Niclo-treated KI; M­ treated KI, Vs Rap-treated KI; DMSO-treated KI Vs PF-treated KI. **p < 0.05, ***p < 0.01, ns: not significant.

Figure 4. Anthelmintics inhibit oncogenic PI3K-dependent cellular phenotypes

A. MCF10A-PIK3CA(E545K) KI cells were incubated for 24 hours in low serum (0.5% Foetal Bovin Serum)-containing medium, without Insulin and EGF (Low serum). Cells were then stimulated with 5% Foetal Bovin Serum plus Insulin and EGF (S) for 3 hours without or with the indicated drugs. Cytoplasmic extracts were separated on linear sucrose gradients and the absorbance profile at 260 nm was recorded. Region of the gradient corresponding to active polysomes is indicated in each panel. Position of 80S, 60S and 40S peaks is also indicated. B.-C. Niclo and PP have a strong cytostatic effect. Growth curve of HME (B) or MCF10A cells (C) carrying wild type PIK3CA or mutated PIK3CA (E545K) treated with Vehicle (DMSO), Niclo or PP. At the indicated time, cell numbers (upper panel) and the percentage of viable cells were calculated compared with non-treated cells. The data are expressed as the mean ± S.D of three independent triplicate experiments.

Figure 5. Niclo and PP inhibit oncogenic PI3K-mediated cellular migration

A. In vitro wound healing assay of MCF10A-PIK3CA(E545K) mammary cells kept in EGF-free serum. Percentage of migration distance referred to total distance was calculated at 7 hrs and 14 hours after in vitro scratch, in presence of DMSO, Niclo or PP. Data indicate average ± SD. B. and C. A transwell assay measuring cell migration of MCF10A cells with or without PIK3CA(E545K) in the presence or absence of Niclo (10μM) (B) or PP (3.4 μM) (C). The number of migrated cells is shown as the mean and SEM of six biological replicates. P-values are calculated by t test: *p < 0.05, **p < 0.005, *** p < 0.0005, unmarked > 0.05.

Figure 6. PP controls mammary branching morphogenesis and PI3K-dependent signaling in mouse mammary gland tissue

A. Representative H&E stained sections of mammary glands from female mice treated with DMSO (D) (panels a-b) or PP (panels c-d). a and c: images at 2,5x magnification; b and d: images at 10X magnification. Images are representative of different fields of gland sections derived from three different mice (n = 3). B. Duct epithelial thickness (upper panel) and ducts diameter (lower panel) were analyzed in H&E stained sections of mammary gland from female mice treated with DMSO or PP. Means ± SEM are shown in histograms. P-values were calculated by t test: ** p < 0.05. C.-D. IHC analysis of mammary gland sections of female mice treated with DMSO (panels a-c) or PP (panels d-f) and stained with anti phospho-AKT (C) or anti phospho-S6 antibodies (D) Images are representative of different fields of gland sections derived from three different mice (n = 3). Images were captures at 40x magnification; images at 4x and 20x magnification were also analyzed (Figure 5S). Scale bars: 400 μM (a and c H&E images), 100 μM (b and d H&E images), 30μm (IHC images).

Figure 7. PP inhibited the growth of breast cancer cells carrying PIK3CA mutations

A. Breast cancer cell lines carrying the indicated PIK3CA mutation were seeded in 96-well plates. DMSO or Pyrvinium Pamoate (PP) was added at indicated concentrations and cell viability was evaluated after 72 hours by the CellTiter-Glo assay. The results are the mean ± SD of 3 independent experiments showing the percentage of cells viability over that of DMSO-treated cells. B. Upper panel: Representative graphs from Kaleidagraph software showing dose-response curves to Pyrvinium Pamoate (20 to 200 nM) in different cancer cell lines. Dose is presented on the x axis, while cell viability (% vs control) is presented on the y axis. Lower panel: Tables showing IC50 and slope values. C. and D. HER2 positive (HER2+) cancer cell lines (MDA-MB-361 and BT-474) carrying PIK3CA mutations were treated with PP at the IC50 concentration or with Trastuzumab (Tz) alone (10μg/ml) or in combination (Tz+PP). The Transtuzumab-responsive SKBR3 cell line (HER2+/PIK3CA Wild-Type) and Transtuzumab-resistant MCF7 (HER2-/PIK3CA(E545K)) were used as positive and negative controls, respectively. C. Cell viability after 72 hours was evaluated as in A C.. D. Soft agar colony forming assays were carried out with the indicated cancer cell lines. 1000 cells were seeded per well (2000 for BT-474) and treated with PP (at IC50 concentration) or Tz alone (10μg/ml) or in combination with PP. Tz was added once a week for 2 consecutive weeks. Colonies were stained with crystal violet at ∼4 weeks after initial seeding. Statistical analyses were performed using GraphPad Prism version 4 (www.graphpad.com). Statistical significance was determined by one-way ANOVA P value < 0.05*, P value < 0.01**, P < 0.001***.