Biological Effects of AL622, a Molecule Rationally Designed to Release an EGFR and a c-Src Kinase Inhibitor

Authors

  • Anne-Laure Larroque-Lombard,

    Search for more papers by this author
    • These authors contributed equally to the work.

  • Na Ning,

    Search for more papers by this author
    • These authors contributed equally to the work.

  • Suman Rao,

    1. Cancer Drug Research Laboratory, Department of Medicine, Division of Medical Oncology, McGill University Health Center/Royal Victoria Hospital, 687 Pine Avenue West Rm M-719, Montreal, Quebec, H3A 1A1 Canada
    Search for more papers by this author
  • Sylvia Lauwagie,

    1. Cancer Drug Research Laboratory, Department of Medicine, Division of Medical Oncology, McGill University Health Center/Royal Victoria Hospital, 687 Pine Avenue West Rm M-719, Montreal, Quebec, H3A 1A1 Canada
    Search for more papers by this author
  • Ruba Halaoui,

    1. Cancer Drug Research Laboratory, Department of Medicine, Division of Medical Oncology, McGill University Health Center/Royal Victoria Hospital, 687 Pine Avenue West Rm M-719, Montreal, Quebec, H3A 1A1 Canada
    Search for more papers by this author
  • Laëtitia Coudray,

    1. Cancer Drug Research Laboratory, Department of Medicine, Division of Medical Oncology, McGill University Health Center/Royal Victoria Hospital, 687 Pine Avenue West Rm M-719, Montreal, Quebec, H3A 1A1 Canada
    Search for more papers by this author
  • Ying Huang,

    1. Cancer Drug Research Laboratory, Department of Medicine, Division of Medical Oncology, McGill University Health Center/Royal Victoria Hospital, 687 Pine Avenue West Rm M-719, Montreal, Quebec, H3A 1A1 Canada
    Search for more papers by this author
  • Bertrand J. Jean-Claude

    Corresponding authorSearch for more papers by this author

Corresponding author: Bertrand J. Jean-Claude, bertrandj.jean-claude@mcgill.ca

Abstract

In breast cancer cells expressing c-Src and EGFR, a control of one of the two oncogenes over proliferation and invasion is observed, whereas in others, the synergistic interaction between them is required for tumor progression. With the purpose of developing molecules with the highest probability for blocking the adverse effects of these two oncogenes, we designed AL622, which contains a quinazoline head targeted to EGFR and a linker that bridges it to the PP2-like structure for targeting c-Src. In case the entire molecule would not be capable of blocking c-Src, we designed AL622 to hydrolyze to an intact c-Src-targeting PP2 molecule. After confirming its binary c-Src-EGFR targeting potency of AL622, we analyzed its potency in isogenic NIH3T3 cells transfected with EGFR and HER2 and human breast cancer cells known to be dominated by c-Src function. The results showed that in EGFR/HER-2-driven cells, it was more potent than PP2 and its activity was in the same range as the latter in more c-Src-driven cells. Its ability to block motility and invasion was comparable with that of PP2 and corresponding combinations, indicating that AL622 could be a better antitumor agent in cells where c-Src and/or EGFR play a role.

Tumor progression is characterized by a variety of signaling pathways often subverted by the mutation or overexpression of genes encoding key signaling proteins. Disordered expression of several growth factor receptors is not only associated with enhanced growth signaling, but also with tumor invasiveness. In the latter process, receptor tyrosine kinases can interact with non-receptor tyrosine kinases to promote motility and invasion (1,2). One such interaction has already been reported for the epidermal growth factor receptor (EGFR) and the non-receptor tyrosine kinase c-Src (3). Recent studies showed that activation of c-Src in some breast cancer cell lines led to phosphorylation of EGFR on Tyr845 and that the downstream effect is transduced through the STAT3/5 pathway (4,5). A significant body of work has now accumulated to suggest that EGFR and c-Src may contribute to an aggressive phenotype in some tumors (6).

The development of new drugs with dual or multiple targeting properties is now a new trend in anti-cancer drug discovery (7–10). Recently with the purpose of inducing a tandem and targeted blockade of EGFR and c-Src, we designed a series of agents containing a PP2-like warhead directed at c-Src and a gefitinib-like moiety targeted to EGFR. PP2 (see Figure 1) is an aminopurine inhibitor of c-Src (11), and the quinazoline moiety of gefitinib is known to anchor into the ATP site of EGFR (12). Studies on this series of molecules revealed that compounds such as SB163 (Figure 1; 13) have a strong EGFR inhibitory property but moderate c-Src targeting potential. The reduced c-Src targeting potency of SB163 was believed to be due to the bulkiness of the linker attached to the 6-position of the aminopurine PP2-like moiety (13). To circumvent this problem, we recently redesigned the molecule to release an intact PP2 molecule and a 7-substituted aminoquinazoline capable of blocking EGFR activation. Here, we study the cytokinetics of one such compound, AL622, wherein the hydrolysable linker is branched to the 6-aminopurine through an amide bond. It was expected that due to the electron deficiency of the pyridine linker, the amide bond could be readily hydrolyzed in the intracellular milieu.

Figure 1.

 Schematic representation of the cross-talk between the epidermal growth factor receptor (EGFR) and the non-receptor tyrosine kinase, c-Src.

Here, we analyzed the degradation, the growth inhibitory potency as well as the mixed EGFR-c-Src targeting properties of this novel type of combi-molecule. Furthermore, given the significant role of c-Src and EGFR in invasion and motility, we analyzed the ability of AL622 to block these processes in breast cancer cells using the wound healing and Boyden Chamber assays.

Methods and Materials

Chemistry

1H NMR spectra and 13C NMR spectra were recorded on a Varian 300, 400, or 500 MHz spectrometer. Chemical shifts are given as δ values in parts per million (ppm) and are referenced to the residual solvent proton or carbon peak. Mass spectrometry was performed by the McGill University Mass spectroscopy Center, and electrospray ionization (ESI) spectra were performed on a Finnigan LC QDUO spectrometer. Data are reported as m/z (intensity relative to base peak = 100). All chemicals were purchased from Sigma-Aldrich (Oakville, Ontario, Canada).

Compound 2

To a solution of PP2 (0.5 g, 1.66 mmol) in dry THF (10 mL) at 0 °C, NaH (60% oil dispersion; 80 mg, 1.2 eq) was added portionwise. After 30 min, a solution of methyl 6-(chlorocarbonyl)nicotinate (0.66 g, 2 eq.) in THF (5 mL) was added dropwise at 0 °C. The mixture was subsequently kept under argon for 24 h, after which it was evaporated to give a green-blue crude powder (1.2 g), which was purified by silica gel chromatography column (CH2Cl2/MeOH 9/1 to 85/15) to give 2 as pure white powder (0.4 g, 52%). 1H NMR (400 MHz, DMSO-d6) δ ppm 1.82 (s, 9H), 3.93 (s, 3H), 7.42 (d, J = 8.4 Hz, 2H), 7.71 (br s, 2H), 7.98 (d, J = 8 Hz, 1H), 8.46 (dd, J = 8 Hz, J = 2 Hz, 1H), 8.86 (s, 1H), 8.99 (s, 1H), 11.07 (br s, 1H).

Compound 3

Compound 2 (0.4 g, 0.87 mmol) was dissolved in THF/CH2Cl2 (1/1) mixture (10 mL) at room temperature and potassium trimethylsilanolate (0.56 g, 5 eq.) was added. A white precipitate appeared within a few minutes, and the mixture was further stirred at room temperature under argon for 2 h 30 min after which it was evaporated to dryness. The resulting solid was triturated in ethyl ether, collected by filtration, and redissolved in water. The pH of the solution was adjusted to 1 with HCl 1N, and the product extracted three times with ethyl acetate. The organic phase was then dried over MgSO4, filtered, and evaporated to give 3 as a pure white solid (0.28 mg, 71%). 1H NMR (400 MHz, DMSO-d6) δ ppm 1.82 (s, 9H), 7.43 (d, J = 8.2 Hz, 2H), 7.70 (d, J = 8.2 Hz, 2H), 7.97 (d, J = 8.2 Hz, 1H), 8.44 (d, J = 8.2 Hz, 1H), 8.87 (s, 1H), 8.97 (s, 1H), 11.06 (s, 1H).

Compound 5

6-Amino-4-[(3-chlorophenyl)amino]quinazoline 4 (7.4 mmol, 2.0 g) was dissolved in dry THF (40 mL) containing pyridine (0.9 mL, 1.5 eq) and triethylamine (1.6 mL, 1.5 eq.) at 0 °C. A solution of 4-chloromethylbenzoyl chloride (2.1 g, 1.5 eq.) in dry THF (5 mL) was added dropwise. After 2 h at room temperature, the reaction mixture was concentrated in vacuo. The resulting green solid was washed with aqueous HCl (1N) and aqueous K2CO3 (10%). Additional washings with water, CH2Cl2, and Et2O gave a green solid, which was subsequently dried in vacuo to afford compound 5 as a yellow-green solid (2.33 g, 74 % yield). 1H NMR (400 MHz, DMSO-d6): δ 5.86 (s, 2 H), 7.15 (dd, = 8.0 Hz, = 1.0 Hz, 1 H), 7.40 (t, = 8.0 Hz, 1 H), 7.63 (d, = 8.5 Hz, 2 H), 7.82 (d, = 8.0 Hz, 2 H), 8.00–8.08 (m, 4 H), 8.60 (s, 1 H), 8.91 (d, = 2.0 Hz, 1 H), 9.97 (s, 1 H), 10.67 (s, 1 H).

Compound 6

To compound 5 (1.0 g, 2.36 mmol) in DMF (10 mL) was added potassium iodide (0.43 g, 1.1 eq.) followed by 2-methylaminoethanol (0.57 mL, 3 eq.) at room temperature under argon. After 3 h at 70 °C under argon, the DMF was azeotroped with heptane to give a brown paste, which was triturated in water. The resulting pale orange solid was filtered and dried in vacuo to give compound 6 (1 g, 92 %). 1H NMR (400 MHz, DMSO-d6): δ 2.17 (s, 3H), 2.44 (t, = 6.4 Hz, 2H), 3.51 (q, = 9.2 Hz, = 5.6 Hz, 2H), 3.58 (s, 2H), 4.44 (t, = 5.6 Hz, 1H), 7.14 (dd, = 8.0 Hz, = 1.4 Hz, 1 H), 7.39 (t, = 8.0 Hz, 1 H), 7.49 (d, = 8.4 Hz, 2 H), 7.81 (d, = 9.4 Hz, 2 H), 7.99 (d, = 8.4 Hz, 2 H), 8.02–8.07 (m, 2 H), 8.60 (s, 1 H), 8.90 (d, = 2.4 Hz, 1 H), 9.95 (s, 1 H), 10.58 (s, 1 H).

Compound 7 (AL622)

To a solution of 3 (0.54 g, 1.2 mmol) and 6 [0.55 g, 1 eq in DMF (10 mL)] were added dicyclohexylcarbodiimide (DCC) (0.37 g, 1.5 eq.), hydroxybenzotriazole (HOBt) (0.24 g, 1.5 eq.), and dimethylaminopyridine (DMAP) (15 mg, 0.1 eq.). A white precipitate appeared within a few minutes, and the mixture was further stirred at room temperature for 48 h under argon. The precipitate was filtered through celite, and the DMF evaporated in vacuo. The resulting paste was redissolved in ethyl acetate and washed with water and brine. The organic layer was removed and dried over MgSO4, to give a crude brown solid (1.6 g), which was further purified by silica gel chromatography column (CH2Cl2/MeOH 95/5 to 9/1). Further purification by preparative TLC (silica plate, CH2Cl2/MeOH 9/1) gave 7 as a pure pale yellow solid (0.37 g, 35%). 1H NMR (400 MHz, DMSO-d6): δ 1.78 (s, 9H), 2.32 (s, 3H), 2.78 (t, = 5.0 Hz, 2H), 3.66 (s, 2H), 4.48 (t, = 5.0 Hz, 2H), 7.13 (dd, = 8.0 Hz, = 1.5 Hz, 1 H), 7.37 (t, = 8.5 Hz, 1 H), 7.40 (d, J = 8.4 Hz, 2H), 7.45 (d, = 8.4 Hz, 2 H), 7.67 (br s, 1H), 7.71 (d, J = 8.8 Hz, 2H), 7.78 (dd, = 8.5 Hz, = 1.5 Hz, 1 H), 7.90 (d, = 8.4 Hz, 2 H), 7.96 (dd, = 8.5 Hz, = 1.8 Hz, 1 H), 8.03–8.06 (m, 2 H), 8.44 (dd, = 8.5 Hz, = 1.8 Hz, 1 H), 8.56 (s, 1 H), 8.76 (br s, 1H), 8.85 (d, = 1.6 Hz, 1 H), 8.92 (s, 1H), 9.86 (s, 1 H), 10.51 (s, 1 H), 11.01 (br s, 1H); 13C NMR (100 MHz, DMSO-d6) δ 168.8, 165.9, 165.7, 164.2, 157.7 (2C), 154.6, 154.5, 153.6, 151.8, 149.0, 147.2, 143.8, 141.4, 141.3, 139.4, 137.3 (2C), 133.8, 133.3, 133.1 (2C), 131.9, 130.8, 130.4, 129.1(2C), 128.6 (2C), 128.1 (2C), 123.3, 123.1, 121.8, 120.7, 115.7, 113.8, 105.0 (2C), 63.7, 61.5, 61.1, 55.3, 42.9, 29.2 (3C); ESI m/z 916.23 (MNa+ with 35Cl, 35Cl).

Drug treatment

AL622 was designed and synthesized in our laboratory. PP2 was also synthesized in our laboratory following the methods already described in the literature (14), and Iressa® (gefitinib, AstraZeneca) was purchased from the Royal Victoria Hospital pharmacy and extracted from pills in our laboratory. In all assays, drugs were dissolved in DMSO and the concentration of DMSO never exceeded 0.2% (v/v). Subsequently, compounds were diluted in sterile media (DMEM or RPMI-1640) containing 10% fetal bovine serum (FBS) prior to addition to cells.

Cell culture

4T1 murine breast cancer cells (generous gift from Dr. Thierry Muanza, Jewish General Hospital, Montreal, Canada) were maintained as a monolayer in RPMI-1640 with 4.5 g/L of glucose, 2 mm l-glutamine, 10 mm of sodium pyruvate, 10 mm HEPES, 10% FBS, and 100 μg/mL penicillin/streptomycin cell culture. The human breast carcinoma MDA-MB-231 cell line was obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). Mouse fibroblast cells NIH 3T3 wild type or NIH 3T3her14 (transfected with erbB1/EGFR gene) and NIH 3T3neu (transfected with erbB2 gene) were provided by Dr. Moulay Aloui-Jamali (Jewish General Hospital, Montreal, Canada). Cells were maintained in Dulbecco modified Eagle’s minimal essential medium (DMEM) supplemented with 10% FBS, 10 mm HEPES, 2 mm l-glutamine, and 100 μg/mL penicillin/streptomycin (all reagents purchased from Wisent Inc., St-Bruno, Canada). All cells were grown in a humidified incubator with 5% carbon dioxide at 37 °C.

Enzyme binding assay

The EGFR, c-Src, Abl, and c-Met kinase assays are similar to the one described by Brahimi et al. (15). Briefly, the kinase reaction was performed in 96-well plates (Nunc Maxisorp) coated with PGT (poly l-glutamic acid l-tyrosine, 4:1) and incubated at 37 °C for 48 h using 4.5 ng/well EGFR or c-Src (Biomol, Enzo Life Science, Inc., Farmingdale, NY, USA). PGT served as the substrate to be phosphorylated by EGFR and c-Src in the presence of ATP (50 μm) when stimulated by EGF (100 μg/mL). Following drug addition (range 0.0001–10 μm), phosphorylation of EGFR or c-Src was initiated by supplementing the reaction with ATP. The phosphorylated substrate was detected using an HRP-conjugated anti-phosphotyrosine antibody (Santa Cruz Biotechnology, CA). The signal was developed by the addition of 3, 3’, 5, 5’-tetramethylbenzidine peroxidase substrate (Kierkegaard and Perry Laboratories, Gaithersburg, MD, USA), and the colorimetric reaction was monitored at 450 nm using a microplate reader ELx808 (BioTek Instruments). The IC50 values were calculated using GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA).

In vitro growth inhibition studies

Growth inhibitory activities were evaluated using the sulforhodamine B (SRB) assay (16). 4T1 and MDA-MB-231 were plated at approximately 5000 cells per well in 96-well plates, and cells were allowed to attach for 24 h. Thereafter, cells were exposed to different drug concentrations for 5 days to determine basal growth inhibition. Subsequently, cells were fixed with 10% ice-cold trichloroacetic acid for 60 min at 4 °C, stained with SRB (0.4%) for 4 h at room temperature, rinsed with 1% acetic acid, and allowed to dry overnight. The resulting colored residue was dissolved in Tris base (10 mm, pH 10–10.5), and the optical density recorded at 492 nm using a microplate reader ELx808. The results were analyzed using GraphPad Prism (GraphPad Software Inc.), and the sigmoidal dose–response curve was used to determine IC50 values. Each point represents the average of at least three independent experiments run in triplicate.

Wound healing assay

Breast cancer cell lines 4T1 and MDA-MB-231 were plated in 6-well plates and incubated overnight (37 °C, 5% CO2). A cross scratch was made in the middle of the cell monolayer. The cells were then washed twice with PBS, and the media was changed to fresh RPMI-1640 or DMEM. Cells were incubated without or in the presence of the drug at 12.5 and 25 μm concentrations. The images were captured at time 0 and at regular intervals of 24, 48, and 96 h to monitor cell migration using a Leica DM IL inverted fluorescence microscope (10×).

Matrigel invasion assay

The invasive property of breast cancer cells was determined using the two-compartment Boyden chamber assay (17). Cells were added onto polycarbonate transwell filter (8 μm pore size), which is a Matrigel-coated membrane separating the top and bottom chambers (50 μg/filter). A total of 5 × 104 4T1 breast cancer cells resuspended in DMEM starvation medium were added to the upper chamber, and the insert was placed in a 24-well plate containing DMEM starvation media either without or with 10% FBS and 20 ng/mL TGF-α as a chemo-attractant. AL622, PP2, gefitinib, and equimolar doses of PP2 and gefitinib were added to both the upper and lower chambers, and the cells were incubated at 37 °C for 4 h. Medium was removed, cells on both side of the filter were fixed with 3.7% paraformaldehyde (Sigma-Aldrich Canada Ltd) for 1 h, thereafter stained with 0.1% crystal violet solution for 30 min, and the cells on the upper surface of the filters were carefully removed with a cotton-tipped applicator and two washes with PBS. Cells that passed across the Matrigel transwell filter toward the lower surface were counted in three randomly selected non-overlapping fields (4× objective). The average cells counted from three independent experiments were reported, and representative images were photographed.

Fluorescence microscopy imaging for intracellular localization of combi-molecule

4T1 breast cancer cells were seeded in 6-well plates (1 × 106 cells/well) and grown in RPMI-1640 with 10% FBS for 24 h. Cells were then treated with three different concentrations (15, 25, and 50 μm) of AL622 or PP2. Cells were observed by fluorescence microscopy every hour for the appearance of blue fluorescence (excitation 294 nm, emission 451 nm) resulting from the release of free PP2 after AL622 degradation into its two moieties. Images were recorded at 5 and 24 h after incubation.

Autophosphorylation assay

MDA-MB-231 cells (up to 1 × 106 cells/well) were plated in 6-well plates and preincubated in DMEM with 10% FBS at 37 °C for 24 h. Following overnight starvation, cells were exposed for 2 h to a dose range of each drug, and subsequently, cells were washed and stimulated with EGF (50 ng/mL) for 15 min at 37 °C. Cells were washed, detached by scraping in ice-cold PBS, and collected by centrifugation for 2 min at 5000 g. Pelleted cells were resuspended in cold lysis buffer [50 mm Tris–HCl pH 7.5; 150 mm NaCl; 1% Nonidet P-40, 1 mm EDTA; 5 mm NaF; 1 mm Na3VO4; protease inhibitor tablet (Roche Biochemicals, Laval, Canada)]. Lysates were kept on ice for 30 min and collected by centrifugation at 10 000 rpm for 20 min at 4 °C. The concentration of protein was determined using the Bio-Rad protein assay kit (Bio-Rad laboratories, Hercules, CA, USA). Equal amounts of proteins were loaded, resolved on 10% SDS–PAGE, and thereafter transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA, USA). Membranes were blocked with 5% milk in TBST (20 mm Tris–HCl, 137 mm NaCl, 0.1% Tween 20) for 2 h at room temperature followed by incubation with anti-phosphotyrosine antibody (clone 4G10, Upstate; 1:1000) at 4 °C overnight. After three 10-min washes with TBST, blots were incubated with secondary HRP-conjugated goat anti-mouse antibody (Jackson ImmunoResearch Laboratories) for 1 h in TBST solution. The membranes were subsequently incubated with anti-phospho-c-Src, total c-Src antibody (Cell Signaling Technology), and total EGFR (sc-03, Santa Cruz, CA). Immunoblot bands were visualized using ECL kit and enhanced chemiluminescence system (Amersham Pharmacia Biotech, Buckinghamshire, UK).

Degradation analysis of the combi-molecule AL622 by high-performance liquid chromatography (HPLC)

4T1 breast cancer cells were seeded in six-well plates (1 × 106 cells/well) and grown in RPMI-1640 with 10% FBS for 24 h. Cells were then treated with different concentrations of AL622 or PP2 at 37 °C for 5 and 24 h. The cells were washed with ice-cold PBS to remove extracellular drug. Cells were scraped with 1 mL of methanol and the resulting mixture centrifuged at 10 000 rpm for 6 min. The supernatant was collected and evaporated to dryness. The resulting residue was reconstituted in a smaller volume of methanol (100 μL). The HPLC analyses were performed with a Cogent C18 series column (150 × 4.6 mm), and the elutions achieved with 88% aqueous methanol at a 1 mL/min flow rate. Analyses were performed using a Thermoquest P4000 equipped with a UV2000 detector and a AS300 autosampler.

Results

Chemical synthesis

The synthesis of the molecule AL622 (7) proceeded according to Schemes 1–3. Following activation with NaH, PP2 (1) obtained according to a method described in the literature (14) was treated with an excess of methyl 6-(chlorocarbonyl)nicotinate to give 2, whose methyl ester was removed in the presence of potassium silanoate base to give acid 3 (Scheme 1). Alcohol 6 (AL621) was synthesized as described in Scheme 2. Briefly, the anilinoaminoquinazoline 4 (18) was coupled with 4-(chloromethyl)benzoyl chloride in the presence of pyridine and triethylamine to give 5, which was further treated with an excess of (2-methyl)aminoethanol to provide compound 6. As depicted in Scheme 3, coupling of 3 with 6 in the presence of DCC, HOBt, and DMAP gave 7 (AL622).

Figure 
                 1:
              .

 Synthesis pathway of intermediate 3.

Figure 
                Scheme 2:
              .

 Synthesis pathway of intermediate 6.

Figure 
                Scheme 3:
              .

 Synthesis pathway of the combi-molecule AL622, 7.

Hydrolysis

The central strategy employed in the design of AL622 was based on the use of a linker capable of releasing both the EGFR and the c-Src inhibitors upon hydrolysis. As depicted in Figure 1, AL622 was designed to contain a cleavable p-dicarboxylic acid ester group that links it to the EGFR inhibitory moiety and a hydrolysable amide that carries the PP2 moiety. The synthesis of AL622 will be reported elsewhere.

Analysis of AL622 hydrolysis in 4T1 mouse mammary cells revealed that its two primary metabolites were AL621 and PP2. After a 5-h drug exposure, the release of PP2 from AL622 was analyzed by fluorescence microscopy (Figure 2). PP2 fluoresces in the blue. The results showed that within the first 5 h, blue fluorescence was detectable inside the cells reaching a significantly high intensity 24 h later. PP2 released from AL622 could be observed in the perinuclear region. In contrast, PP2 administered alone did not show a significant perinuclear distribution.

Figure 2.

 Fluorescence microscopy imaging of intracellular localization of combi-molecule and degradation analysis by HPLC. 4T1 breast cancer cells were treated with three different concentrations of AL622 and PP2 for various time periods (5 and 24 h). The intracellular distribution of PP2 was observed by fluorescence microscopy. The cells were extracted with methanol to observe the degradation of AL622 at the two different time-points.

EGFR-c-Src inhibitory potency

The ability of AL622 to block EGFR and c-Src tyrosine kinase activity was studied in an ELISA (Figure 3A,B). While gefitinib showed IC50 in the nm range (24 nm), despite its bulkiness, AL622 induced a dose-dependent inhibition of EGFR TK (IC50 of 0.6 μm). In addition to being a prodrug for the potent c-Src inhibitor PP2 (IC50, 0.188 μm), it was also able to induce c-Src TK inhibition on its own (IC50 values of 1.099 μm).

Figure 3.

 Enzyme inhibition assay of AL622 in comparison with gefitinib and PP2. (A) Inhibition of EGFR and (B) inhibition of c-Src. Ninety-six-well plates were coated with PGT (poly l-glutamic acid l-tyrosine), which served as the substrate to be phosphorylated by EGFR and c-Src in the presence of ATP (50 μm) when stimulated by EGF (100 μg/mL). Dose-dependent inhibition by the drugs was observed by adding a chemiluminescent agent and measuring absorbance at 450 nm. (C) EGFR and c-Src activity at various ATP concentrations and AL622 (1 μm) in an ELISA-based competitive ATP binding assay. Loss of inhibitory activity was observed for c-Src at lower concentrations of ATP than for EGFR. (D) Inhibition of receptor and non-receptor tyrosine kinases by AL622 in an ELISA-based competitive ATP binding assay.

EGFR-c-Src competitive binding at various ATP concentrations and kinase selectivity

From the EGFR and c-Src inhibition studies to determine the IC50 of AL622, it appeared that it was a ca. twofold stronger inhibitor of EGFR than c-Src. To further ascertain this difference in potency, we performed an activity study with our ELISA using OD (optical density) that reflects the level of substrate phosphorylation by EGFR and c-Src. Significant loss of inhibitory activity was observed for c-Src at much lower concentrations of ATP (e.g., 10 μm ATP) than for EGFR (e.g., 250 μm; Figure 3C) indicating that AL622 (intact structure) has a much stronger affinity for EGFR than for c-Src. This was further corroborated by comparison with other kinases (e.g., c-Met, Abl). As shown in Figure 3D, AL622, while retaining strong inhibitory potency against c-Src and Abl, was more selective for EGFR.

Whole cell EGFR-c-Src inhibitory potency

The ability of AL622 to block phosphorylation of EGFR and c-Src in whole cells was studied in the MDA-MB-231 cell line, which is known to express the latter two oncogenes. The results showed that AL622 was capable of inducing strong inhibition of c-Src phosphorylation in MDA-MB-231 at doses ranging from 3.125 to 25 μm (Figure 4). Analysis of EGFR phosphorylation revealed that AL622 was capable of inducing a dose-dependent inhibition of EGFR TK with almost 100% inhibition of phosphorylation at a concentration as low as 0.1 μm. By contrast, at three times higher concentration (0.3 μm), PP2 induced a moderate level of EGFR phosphorylation. It is worth noting that while the isolated enzyme assays showed noticeable difference between EGFR and c-Src inhibitory activities for intact AL622, in the 3.125–25 μm range at which inhibition of growth, motility, and invasion experiments were performed, the two kinases were maximally inhibited (Figure 4D).

Figure 4.

 Dose-dependent inhibition of EGFR and c-Src phosphorylation in MDA-MB-231 cells upon treatment with AL622 (A,C), gefitinib, PP2, and gefitinib + PP2 (B). Western blot analysis was performed using serum-starved cells. Cells were treated with the drugs for 2 h followed by EGF (50 ng/mL) stimulation for 20 min. In (A) and (B), phospho-EGFR was determined by probing with pY99 mouse monoclonal primary antibody and an HRP-conjugated goat anti-mouse secondary antibody. In (C), phospho-c-Src levels were determined using the pY416 c-Src (rabbit) primary antibody and the HRP-conjugated goat anti-rabbit secondary antibody. Total EGFR and c-Src levels were determined using anti-EGFR and c-Src antibodies to check for equal protein loading. (D) Levels of phosphorylation of EGFR and c-Src following treatment with AL622 were quantitated by band intensity ratio (kinase/loading control) resulting from densitometric analysis of (A) and (C).

Effect of AL622 on cell motility

The ability of AL622 to block cell motility was studied on 4T1 and MDA-MB-231 cells (Figure 5). The results showed that AL622 prevented repopulation of the wound in a dose-dependent manner with levels of inhibition slightly less strong than the combination of gefitinib plus PP2 or PP2 alone. The potency of the AL622 in blocking wound healing was equivalent to that of PP2.

Figure 5.

 Wound healing assay of untreated and treated 4T1 cells. Cells were treated with 12.5 and 25 μm of AL622, PP2, or PP2 + gefitinib and were observed at 0 and 24 h after treatment. The control cells showed almost complete wound closure.

Effect of AL622 on cell invasiveness

The ability of AL622 to block cell invasiveness was studied in 4T1 cells, which invaded the Matrigel in the presence of serum (Figure 6). The results showed that at the lower dose, AL622 induced more significant blockade of invasion than gefitinib (p < 0.001), PP2 (p < 0.001), and the combination of these two drugs (gefitinib + PP2; p < 0.001).

Figure 6.

 Invasive capacity of 4T1 breast cancer cells treated with AL622, gefitinib, and PP2. The cells were incubated with the drug at 37 °C for 4 h.

Oncogene-selective growth inhibitory potency

To evaluate the targeting contribution of the EGFR warhead of AL622, we evaluated its potency and selectivity on ErbB oncogene-transfected cells. Using NIH3T3 cells transfected with ErbB1 (Her14/EGFR) and ErbB2 (Neu) genes, we showed that AL622, gefitinib, and PP2 were capable of inducing selective inhibition of the oncogene-transfected cells (Figure 7). However, of all the drugs tested and corresponding combinations (e.g., gefitinib + PP2), AL622 was the most selective, with minimal growth inhibitory effect against the normal NIH3T3 wild-type cells.

Figure 7.

 Growth inhibition assay of different cell lines treated with AL622, PP2, gefitinib, and PP2 + gefitinib (equimolar concentration). Approximately 5000 cells were allowed to attach for 24 h and exposed to drug treatment for 5 days. Growth inhibition was determined using the sulforhodamine B assay. (A) NIH3T3 wild type and transfected, (B) 4T1 cells, and (C) MDA-MB 231 cells.

Discussion

The design of AL622 stemmed from previous work on the combi-targeting concept that sought to prepare multitargeted molecules to interfere with multiple signaling pathways in tumor cells (13,18). These molecules termed ‘combi-molecules’ are now classified into two major groups: type I and type II (19). Type I combi-molecules are designed to be hydrolyzed to generate their binary targeting functions (e.g., EGFR and DNA), while type II combi-molecules do not require hydrolysis to generate their EGFR-DNA targeting function (20,21). In our search for potent EGFR TK and c-Src TK cross-targeting compound, we recently designed and synthesized SB163, a type II combi-molecule in which a basic linker was flanked by a quinazoline ring on one hand and an aminopurine analogue of PP2 on the other (Figure 1). Studies on the latter molecule showed that it could significantly block EGFR, but possessed a weak c-Src inhibitory activity (IC50 = 3 μm). Thus, we redesigned the molecules to release an intact c-Src inhibitor and a quinazoline derivative that retained EGFR inhibitory potency.

The c-Src inhibitory potency of inhibitors of the same class as PP2 is extremely sensitive to structural alteration. Molecular modeling suggested the 9-position of the purine ring to be the best substitution position for retention of optimal c-Src TK activity. When we replaced the tert-butyl group of PP2 by an ethylamino linker as in SB163 (Figure 1), c-Src inhibitory potency decreased by threefold when compared with PP2. In this study, we rationalized that a type I combi-molecule strategy (22) would lead to a prodrug capable of releasing upon hydrolysis, intact PP2 and the 6-substituted anilinoquinazoline, which usually retains significant EGFR inhibitory activity. As described herein, this strategy has led to the synthesis of the first prodrug, which in addition to being EGFR and c-Src inhibitor on its own (EGFR inhibition IC50 = 0.610 μm and c-Src inhibition = 1.099 μm) is capable of releasing two different kinase inhibitors: PP2 and AL621. The results suggest that AL622 is clearly a rationally designed drug for ‘mix-targeting’ EGFR and c-Src. Here, we used the selective EGFR inhibitor (gefitinib) and the selective c-Src inhibitor PP2 as control drugs to assess the binary cytokinetic profile of AL622. We used the sensitivity of highly invasive breast cancer cells 4T1 and MDA-MB-231 to PP2 and to the highly EGFR-selective inhibitor gefitinib. PP2 was used as a reference to assess the c-Src dependence of the potency of AL622, and gefitinib as a reference for its EGFR dependence.

In growth inhibition assays, AL622 was equipotent with PP2 against the 4T1 cells whose growth and invasion strongly depend on c-Src. It was slightly more potent against MDA-MB-231 cells in which c-Src and EGFR are known to synergize to promote growth (23,24). Importantly, c-Src and EGFR were found to be strongly inhibited at concentrations markedly lower than IC50 for growth inhibition in these cells. In NIH3T3 mouse fibroblast cells transfected with EGFR or HER2 whose growth strongly depends on EGFR or HER2, AL622 was significantly more potent than PP2 but less potent than gefitinib. Interestingly, the dominance of the potency of AL622 over PP2, gefitinib, and the corresponding combination PP2 + gefitinib was seen only in the invasion assay. Recently, EGFR and c-Src were shown to cooperatively promote aberrant invasive behavior in immortalized breast cells (6). Perhaps, in this case, due to the requirement for both EGFR and c-Src to promote invasion, the dual mechanism of action of AL622 led to a more pronounced effect. The superior potency of AL622 when compared with equimolar addition of individual drugs (PP2 + gefitinib) may be due to a greater level of released bioactive species inside the cells when administered via a prodrug than as a combination of individual drugs. Indeed, fluorescence microscopy showed that higher intensity of blue fluorescence associated with PP2 was seen when AL622 was administered than when PP2 was given alone.

The results presented herein indicate that blockade of c-Src in the cells is sufficient to induce significant levels of growth inhibitory activity and reduction of motility. Tandem blockade of EGFR and c-Src with either AL622 or combination of individual molecules did not enhance these activities, suggesting that c-Src is the dominant oncogene controlling both proliferation and motility in these cells. It is now known that c-Src can phosphorylate EGFR at Y845, and this is associated with mitogenic signaling through a pathway in which only STAT5 has been identified (5,25). Previous work by Biscardi et al. (5) showed that transfection of CH101/2 cells with a Y845F mutant form of EGFR led to ablation of EGF-stimulated DNA synthesis in the latter cells, suggesting that c-Src activation plays an important role in the growth of EGFR-expressing cells. Additional blockade of EGFR by AL622 and the individual drug combination gefitinib + PP2 did not enhance inhibition of motility. The activity of the two types of combinations (AL622 and PP2 + gefitinib) was comparable with that of the c-Src inhibitor PP2 alone. This response may also be due to the fact that c-Src is the major player in regulating motility in these cells, a role that is now suspected to be exerted through the PI3K pathway.

One particular and important observation is the markedly selective potency of AL622 in NIH3T3 ErbB transfectants. PP2 that primarily targets c-Src was the least selective of the panel. Gefitinib also induced a dose-dependent growth inhibition in the wild type. The decreased potency of AL622 in these non-transformed cells may be due to decreased intracellular metabolism of AL622 when compared with their transformed counterpart. Further studies are ongoing to prove this point.

The results in toto suggest that AL622 is a potent combi-drug that releases an intact c-Src and an EGFR inhibitor inside the cells. Its significant anti-invasive properties warrant further evaluation in vivo, in a model wherein c-Src synergizes with EGFR to promote invasion.

Acknowledgments

We would like to thank the Canadian Institutes of Health Research (CIHR) Team Grant for financial support. Dr. Na Ning thanks the National Science Council of China for financial support. Dr. Anne-Laure Larroque-Lombard thanks the CIHR and the Drug Development Training Program (DDTP) for her postdoctoral fellowship training award. Suman Rao thanks the McGill-CIHR Drug Development Training Program for financial support. Dr. Ying Huang thanks the MUHC Research Institute for her postdoctoral fellowship training award.

Ancillary