Repurposing FDA Drug Compounds against Breast Cancer by Targeting EGFR/HER2

Repurposing studies have identified several FDA-approved compounds as potential inhibitors of the intracellular domain of epidermal growth factor receptor 1 (EGFR) and human epidermal receptor 2 (HER2). EGFR and HER2 represent important targets for the design of new drugs against different types of cancer, and recently, differences in affinity depending on active or inactive states of EGFR or HER2 have been identified. In this study, we first identified FDA-approved compounds with similar structures in the DrugBank to lapatinib and gefitinib, two known inhibitors of EGFR and HER2. The selected compounds were submitted to docking and molecular dynamics MD simulations with the molecular mechanics generalized Born surface area approach to discover the conformational and thermodynamic basis for the recognition of these compounds on EGFR and HER2. These theoretical studies showed that compounds reached the ligand-binding site of EGFR and HER2, and some of the repurposed compounds did not interact with residues involved in drug resistance. An in vitro assay performed on two different breast cancer cell lines, MCF-7, and MDA-MB-23, showed growth inhibitory activity for these repurposed compounds on tumorigenic cells at micromolar concentrations. These repurposed compounds open up the possibility of generating new anticancer treatments by targeting HER2 and EGFR.


Introduction
Human epidermal receptor type 2 (HER2) is part of a superfamily of human epidermal grown factor (EGF) receptors that include four receptors: ErbB1 (HER1 or EGFR), ErbB2 (HER2), ErbB3 (HER3), and ErbB4 [1]. Of these receptors, EGFR and HER2 are the most significant targets for anticancer therapy. These receptors are composed of a transmembrane region, extracellular domain, and receptor tyrosine kinase (RTK) domain, and the latter has been one of the major targets for anticancer drug design [2]. Activation of these receptors begins with the binding of endogenous growth factors at the extracellular region, which encourages homo and heterodimerization of RTK of both receptors [3,4]. This produces a structural change in the activation loop of the RTK domain and a change in the N-lobe α-C helix to create active conformation [5][6][7][8]. The transition from the inactive to the active state is a two-step process for EGFR [5][6][7], whereas inactive, intermediate, and active states have been reported for HER2 [9][10][11]. Active and inactive EGFR conformations are used in drug design, identifying high specificity ligands by targeting the inactive

Structural Analysis of EGFR-Ligand Complexes
Analysis of the complexes between EGFR and alfuzosin, amodiaquine, antrafenine, bopindolol, carvedilol, doxazosin, pindolol, prazosin, quinacrine, saprisartan, terazosin, topotecan, trimetrexate, udenafil, and vardenafil showed that these complexes were stabilized by between 13 and 24 residues (Table 2). From these protein-ligand complexes, L718, G719, S720, V726, A743, K745, T790, L792, M793, G796, C797, R841, and L844 appear in most complexes. The formation of hydrogen bonds was observed in some complexes. Alfuzosin forms hydrogen bonds with backbone atoms of D855 and side chains of T790 and D855 ( Figure 4A). Amodiaquine makes hydrogen bonds with backbone atoms of L792 and C797 and with side chain atoms of D800 ( Figure 4B). Bopindolol forms hydrogen bonds with side chain atoms of N842 and D855 ( Figure 4D). Carvedilol forms one hydrogen bond with side chain atoms of D837 and one Pi-Cation interaction with F723 ( Figure 4E). Doxazosin established two hydrogen bonds with backbone atoms of M793 and P794 and through side chain atoms of D804 ( Figure 5A). Pindolol makes four hydrogen bonds with side chain atoms of K745, N842, T854, and D855 ( Figure 5B). Prazosin forms one hydrogen bond with backbone atoms of M793 ( Figure 5C). Quinacrine forms one hydrogen bond with the backbone atoms of M793 and a salt bridge with D800 ( Figure 5D). Saprisartan forms two hydrogen bonds with side chain atoms of K745 and R841 ( Figure 5E). Terazosin forms one hydrogen bond with the backbone atoms of C797 ( Figure 6A). Topotecan forms one hydrogen bond with backbone atoms of M793 ( Figure 6B). Trimetrexate establishes two hydrogen bonds with side chain atoms of K745 and D855 ( Figure 6C). Udenafil forms two hydrogen bonds through the backbone and side chain atoms of M793 and D800, respectively ( Figure 6D). Vardenafil establishes one hydrogen bond through side chain atoms of R841 ( Figure 6E). Stabilization of ligands at the ligand-binding site of EGFR establishes interactions with T790, whose mutations are associated with EGFR drug resistance [38,39]. We also identified characteristic interactions with M793, which was previously observed for other inhibitors of EGFR [36,37].

Affinity of Compounds
The difference in binding affinity for all complexes was determined using the MMGBSA approach. All systems showed thermodynamically favorable binding free energy (ΔGbind) values, where nonpolar interactions established by van der Waals energy Figure 6. Interactions of the complexes between EGFR and terazosin, topotecan, trimetrexate, udenafil, and vardenafil. Binding complex for EGFR terazosin (A); EGFR topotecan (B); EGFR trimetrexate (C); EGFR udenafil (D) and EGFR vardenafil (E).

Antiproliferative Assays
HER2 and EGFR are rare important watchdogs for normal cellular activities, and their dysregulation has been linked to protein overexpression that promotes the progression of several kinds of cancer [40,41]. MDA-MB-231 and MCF-7 cell lines are estrogen receptor-negative and positive, respectively [42], and both expressed EGFR and HER2, although MDA-MB-231 cells expressed HER2 and EGFR at higher concentrations than MCF-7 cells [43]. The evaluation of growth inhibition by alfuzosin, quinacrine, terazosin, prazosin, and irinotecan was conducted using the MTT assay in selected MDA-MB-231 ( Figure 7) and MCF-7 ( Figure 8) cell lines and compared with that of gefitinib and lapatinib. Although vardenafil and antrafenine exhibited good theoretical affinities for EGFR or HER2, they were not included in this study because they were not commercially available at the time this research was performed.
As shown in Table 6, all evaluated compounds exhibited IC 50 values in the µM range. The compounds with the best antiproliferative activity in the MCF-7 cell line were irinotecan and quinacrine, which exhibited greater antiproliferative activity than gefitinib or lapatinib. In the MDA-MB-231 cell line, quinacrine exhibited better antiproliferative activity than lapatinib or gefitinib. Prazosin and irinotecan exhibited better antiproliferative activity than terazosin and alfuzosin but lower antiproliferative activity than lapatinib and gefitinib. Although the affinity tendency predicted by the MMGBSA approach was not in line with the antiproliferative activity, the results support the affinity observed through theoretical methods of alfuzosin, terazosin, prazosin, and irinotecan over HER2 and EGFR.    As shown in Table 6, all evaluated compounds exhibited IC50 values in the µM range. The compounds with the best antiproliferative activity in the MCF-7 cell line were irinotecan and quinacrine, which exhibited greater antiproliferative activity than gefitinib or lapatinib. In the MDA-MB-231 cell line, quinacrine exhibited better antiproliferative activity than lapatinib or gefitinib. Prazosin and irinotecan exhibited better antiproliferative activity than terazosin and alfuzosin but lower antiproliferative activity than lapatinib and gefitinib. Although the affinity tendency predicted by the MMGBSA approach was not in line with the antiproliferative activity, the results support the affinity observed through theoretical methods of alfuzosin, terazosin, prazosin, and irinotecan over HER2 and EGFR.

Docking Studies
The twenty-four FDA small compounds were docked with EGFR and HER2 with AutoDock Tools 1.5.6 and AutoDock 4.2 programs [46]. Hydrogen atoms were included in ligands and receptors, and partial charges were assigned to receptors (Kollman) and ligands (Gasteiger). A grid size of 70 × 70 × 70 Å and 0.370 Å spacing was built on the receptor. A Lamarckian genetic algorithm was selected to evaluate a global conformational examination with a maximum of 1 × 10 7 energy calculations and 200 separate populations. For each compound, 20 runs were calculated, and the best binding structures were chosen using the criteria of the lowest energetic ligand conformations at the binding site of the receptor.

Molecular Dynamics Simulations
Docking complexes were examined via MD simulation studies employing the AM-BER16 package [47] and the ff14SB force field [48]. Systems were simulated in a dodecahedric box of 12 Å and solvated with the TIP3P water model [49]. Systems were neutralized with NaCl (0.15 M) to establish physiological strength. Ligand forcefields were constructed considering AM1-BCC atomic charges with the general Amber force field [50]. Solvated and neutralized systems were minimized using steepest descent through 4000 steps and equilibrated through 1 nanosecond (ns). Minimized and equilibrated systems were run by MD simulations for 100 ns with an NPT ensemble at 310 K, with each MD simulation run in triplicate. The time step for simulations was set to 2.0 fs. The SHAKE algorithm [51] was selected to restrict bonds at their equilibrium values. The PME method [52] was employed to describe the electrostatic term, and a 10 Å cutoff was chosen for the van der Waals forces. Constant temperature and pressure of 310 K and 1 atm, respectively, were kept using a weak-coupling algorithm [53], with coupling constants τ T and τ P of 1.0 and 0.2 ps, respectively. The results of the MD simulations were evaluated with the cpptraj tool in Amber16 to determine the root mean squared deviation (RMSD), the radius of gyration (R G ), and clustering analysis. Figures were constructed using PyMOL [54] and Maestro Schrödinger version 10.5 [55].

Affinity Prediction and Per-Residue Decomposition
Binding free energy (∆G bind ) values and per residue energetic contribution were calculated using the MMGBSA approach [56][57][58][59]. Calculations were performed considering 500 protein-ligand complexes at intervals of 100 ps (last 50 ns of simulation) using implicit solvent models [60], a salt concentration of 0.10 M. ∆G bind, and per-residue contributions for each protein-ligand complex were determined as previously described [11] and correspond to the median result of triplicate trials.

Biological Assays on Cell Lines
Gefitinib, lapatinib, terazosin, alfuzosin, prazosin, irinotecan, and quinacrine were purchased from Sigma Chemical (St. Louis, MO, USA). The breast cancer cell lines used in this study, MCF-7 and MDA-MB-231, were obtained from the American Type Tissue Culture Collection (ATCC), Rockville, MD, USA. MCF-7 and MDA-MB-231 cells were grown in Dulbecco's modified Eagle's medium (DMEM) high glucose with phenol red, and the culture medium was supplemented with 10% fetal bovine serum (FBS) (BioWest, Riverside, MO, USA) and 1% penicillin/streptomycin as an antibiotic. The cells were incubated in culture flasks (75 cm 3 ) at 37 • C in a humidified atmosphere of 5% CO 2 and 95% air, all of which were carried out under sterile conditions using a laminar flow hood.

Antiproliferative Assays on Cell Cultures
Breast cancer cell lines were detached with trypsin-EDTA (1%) for 5 min at 37 • C. After trypsin-EDTA inactivation with DMEM, the cells were centrifuged and resuspended in 4 mL of medium. Afterwards, cells were counted with CytoSmart cell counting (CytoSMART Technologies, Eindhoven, The Netherlands). Each cell line was seeded at 5 × 10 3 cells in 96-well tissue culture plates and allowed to attach overnight (24 h) in a CO 2 incubator before the assays. Cells were treated with the compounds at different concentrations, all compounds were dissolved in DMSO (0.1% final concentration), and the control contained medium with DMSO (0.1%). The assays were performed in triplicate for each concentration and incubated at 37 • C under a humid atmosphere with 5% CO 2 for 48 h.

Cell Viability Assays
Further cell viability was determined using MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5diphenyl-2H-tetrazolium bromide (Sigma)]. After 48 h of incubation, the medium was removed and replaced by 100 µL of MTT (0.500 mg/mL), dissolved in PBS, and incubated for 3 h at 37 • C and 5% CO 2 . After that, the MTT/PBS was discarded, and 100 µL of DMSO was applied to each well to dissolve the dark-blue formazan crystals in intact cells. The resulting solution was measured by spectrophotometry with a microplate reader (Thermo Scientific, Multiskan TM Sky) at a wavelength of 550 nm, and the quantity of formazan produced was directly proportional to the number of living cells. The results are expressed as the percentage of viable cells in relation to the control, whose viability was designated as 100%. Each data point was determined with n = 8 in three independent experiments, and the results are reported as the mean absorption ± SD using the GraphPad Prism 8 software.

Conclusions
We performed repurposing studies by screening DrugBank in the search for new FDA-approved drugs with chemical structures similar to lapatinib and gefitinib. Docking and MD simulations coupled to the MMGBSA approach using the selected DrugBank compounds, and considering the inactive state of EGFR and HER2, allowed us to identify that quinacrine, alfuzosin, and antrafenine could act as dual inhibitors of HER2 and EGFR but with higher selectivity to HER2, whereas irinotecan exhibited high selectivity to HER2. Alfuzosin, terazosin, and prazosin could act as dual inhibitors of EGFR and HER2, whereas vardenafil exhibited selectivity exclusively to EGFR. Per-residue decomposition analysis identified the main residues stabilizing the protein-ligand complexes with HER2 and EGFR systems, showing that V726, V734, A751, K753, L785, C805, L852, and T862 were present in the stabilization of HER2 alfuzosin , HER2 antrafenine , HER2 irinotecan , and HER2 quinacrine. In the stabilization of these complexes, the characteristic interaction with M801 in HER2 alfuzosin and HER2 antrafenine complexes was observed, which has been observed for other HER2 inhibitors were also observed. In the case of EGFR, L718, V726, A743, K745, M793, G796, C797, and L844 were present in the stabilization of EGFR alfuzosin , EGFR prazosin , EGFR terazosin , and EGFR vardenafil . All complexes formed interactions with M793, a characteristic interaction that has been observed for other inhibitors of EGFR. Finally, MTT assays showed that repurposed compounds exhibited antiproliferative activity on breast cancer cell lines, with irinotecan and quinacrine exhibiting greater antiproliferative activity than lapatinib and gefitinib in the MCF-7 cell line, and quinacrine exhibiting greater antiproliferative activity than lapatinib and gefitinib in the MCF-7 and MDA-MB-231 cell lines. Finally, we could show that our structure-based screening approach identifies novel repositioning candidates for the cancer target EGFR/HER2. Not only did we identify candidates structurally related to gefitinib and lapatinib, but also showed that they show the desired inhibitory activity on the target receptors. Particularly, the FDA-approved drugs irinotecan and quinacrine, which appeared as the top hit from our screen and were later validated, demonstrates the potential of our approach for drug repositioning.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/ph14080791/s1, Scheme S1: FDA-drugs with structural similarity with lapatinib and gefitinib, Table S1: Docking studies of candidates to inactive EGFR/HER2 inhibitors selected in the DrugBank, Table S2: Interacting residues between inhibitors selected in the DrugBank with HER2, Table S3: Interacting residues between inhibitors selected in the DrugBank with EGFR, Table S4: Average RMSD and RG values (Å) during the last 50 ns of MD simulations, and Table S5: FDA-drugs with structural similarity with lapatinib and gefitinib.

Conflicts of Interest:
The authors declare no conflict of interest.