This study has focused on a broad collection, including 3235 FDA-approved compounds as well as those undergoing clinical evaluation. This collection was evaluated using a drug repurposing approach to identify potential drug candidates for the treatment of Merlin-negative tumors through the inhibition of the FAK protein. To achieve this, the protein and drug library were initially prepared for docking. Subsequently, a grid box was prepared on the protein (Fig. 1), and all ligands were docked into this grid box. This step, which involved comparative examination of docking scores in relation to reference compounds, served as an initial filter to identify molecules with the highest potential as drug candidates. Then, an attempt was made to reduce the number of candidates by performing MD/MMBG (1ns, 10ns, 100ns) analyses.
Molecules with docking scores lower than − 8.00 kcal/mol were selected for detailed analysis. The reference compound bound to the FAK protein with a docking score of -10.85 kcal/mol (Fig. 2). Given that this compound is the co-crystallized ligand in the PDB, obtaining the highest docking score in the initial stage was quite positive and expected. Out of the 3235 FDA-approved and clinically reviewed molecules, 71 had docking scores of -8.00 kcal/mol or lower, indicating better binding affinity. This threshold was strategically chosen to ensure that only the most promising candidates, demonstrating a strong potential for interaction with the FAK protein, were advanced to the subsequent stages of the study.
The docking scores are provided in Table 1. Among the molecules we studied, Amprenavir bound to the FAK protein with a docking score of -8.91 kcal/mol, followed closely by Flavin adenine dinucleotide with − 8,47 and Lactulose with − 8.23 kcal/mol. Bosutinib, Ferric derisomaltose, and Tafluprost demonstrated good binding affinities with scores ranging between − 8.12, -8.00, -8.47, and − 8.13, respectively. These molecules were selected for their high potential binding affinity. The binding conformation of Amprenavir, which has the best docking score, to the protein is provided (Fig. 3).
71 molecules with docking scores of -8.00 kcal/mol or better underwent simulation, followed by MM/GBSA calculations. The MM/GBSA score of the Reference Compound was found to be -69.56 based on the 1 ns simulations. Our tested molecules exhibited significantly better scores. For example, based on the MM/GBSA analyses conducted after the 1 ns MD simulations, Flavin adenine dinucleotide yielded a score of -89.15 kcal/mol, Ferric derisomaltose scored − 82.72 kcal/mol, Lactulose scored − 71.93 kcal/mol, Bosutinib scored − 75.79 kcal/mol, Amprenavir scored − 74.84 kcal/mol, and Tafluprost scored − 72.51 kcal/mol. Consequently, the top 10 molecules with the best scores out of the 71 were selected and subjected to a 10 ns simulation.
During the 10 ns simulation duration, the binding free energy scores remained relatively stable with no significant variations. Notably, Amprenavir, Bosutinib, Lactulose and Tafluprost mostly maintained their affinity with a score of -75.86, -74,38, -73,39 and − 73,13 kcal/mol respectively. Ferric derisomaltose, and Flavin adenine dinucleotide showed some changes in their binding affinities. The Reference Compound demonstrated almost stable with no significant variations, recording a score of -71.04 kcal/mol. In summary, the MM/GBSA scores after the 10 ns MD simulation were found to be nearly identical to the previous ones, indicating stability and consistency.
Subsequently, 9 molecules that exhibited better scores than the − 71.04 kcal/mol of the reference compound were selected for 100 ns simulations. Based on the results of the 100 ns MD, 9 molecules that demonstrated better scores than the − 69,20 kcal/mol of the reference compound were selected for 100 ns MD and MM/GBSA analyses. Upon examining the 100 ns MD-MM/GBSA results, it was found that Amprenavir exhibited a relatively strong binding affinity with a score of -72.81 kcal/mol. Bosutinib also demonstrated robust interaction with the protein, yielding a score of -71.84 kcal/mol. Ferric derisomaltose and Flavin adenine dinucleotide showed even stronger binding affinities, recording scores of -76.70 kcal/mol and − 69.09 kcal/mol, respectively. Conversely, the reference compound exhibited a binding affinity close to and even weaker than many of these molecules with a score of -69,20 kcal/mol.
Table 1
Docking scores and MM/GBSA analysis rely on 1ns, 10ns and 100ns molecular dynamics simulations.
Protein-Ligand Complex | Docking Score (kcal/mol) | 1 ns MD-MM/GBSA (kcal/mol) | 10 ns MD-MM/GBSA (kcal/mol) | 100 ns MD-MM/GBSA (kcal/mol) |
4GU6_Amprenavir | -8,91 | -74,84 | -75,86 | -72,81 |
4GU6_Bosutinib | -8,12 | -75,79 | -74,38 | -71,84 |
4GU6_Ferric derisomaltose | -8,00 | -82,72 | -73,52 | -76,70 |
4GU6_Flavin adenine dinucleotide | -8,47 | -89,15 | -82,99 | -69,09 |
4GU6_Lactulose | -8,23 | -71,93 | -73,39 | -74,86 |
4GU6_Tafluprost | -8,13 | -72,51 | -73,13 | -65,77 |
4GU6_Reference Compound | -10,85 | -69,56 | -71,04 | -69,20 |
At the specified time durations (1 ns, 10 ns, and 100 ns), the correlation between docking scores and MD-MM/GBSA results is evident. Notably, Amprenavir, Bosutinib, Ferric derisomaltose, and Lactulose have demonstrated significantly strong binding affinities. Flavin adenine dinucleotide and Tafluprost have also shown satisfactory results, closely following behind. In contrast, the reference compound exhibited weaker binding affinities compared to all other molecules in the 1 ns and 10 ns durations. After 100 ns of MD simulation, the MM/GBSA analysis indicated that Amprenavir, Bosutinib, Ferric derisomaltose, and Lactulose exhibited decreased binding affinities relative to the reference compound. The docked poses of the selected molecules and detailed docking and MM/GBSA score tables, were provided in the Supplementary Information. The alignment of docking and MD-MM/GBSA results highlights the reliability of computational approaches in predicting ligand binding affinities. Consequently, all six molecules identified in the study emerge as compelling candidates for further research, given their consistently low docking scores and stable binding interactions observed throughout the MD simulations. These findings significantly contribute to our understanding of the dynamic aspects of protein-ligand interactions and have substantial implications for the advancement of drug design and development.
The identification of common amino acid residues engaged in ligand-protein interactions provides valuable insights into potential binding sites and shared mechanisms among various molecules. This knowledge is crucial for understanding the consistent aspects of ligand binding and can guide future efforts in drug design by highlighting critical amino acid residues that consistently contribute to the stability and specificity of ligand-protein complexes. This comprehensive analysis enhances our understanding of the molecular basis of ligand-protein interactions and aids in the logical development of effective therapeutics. Therefore, this study analyzed the key amino acids involved in ligand-protein interactions, in accordance with the mentioned rationale. For the reference molecule, named 10N, interactions were observed with amino acids Cys502, Glu506, Arg550, Val552, Leu553, Asp564, and Leu567 during the 100ns MD simulation, confirming our findings. Importantly, amino acids Cys502 and Asp564 were previously highlighted in the literature as crucial, which supports our results [16, 26].
Notably, the Cys502 residue emerges as a common binding site, forming hydrogen bonds with both the reference molecule Compound 10N (Fig. 4), the candidate Ferric derimaltose and Bosutinib. This consistency in interaction highlights the significance of Cys502 in facilitating stable ligand binding and suggests its potential role as a key hotspot for ligand-protein interactions. Moreover, amino acid residues such as Thr503 and Asn51 exhibit recurrent involvement in the interactions. Thr503 is implicated in hydrogen bonding with Bosutinib, emphasizing its contribution to the binding affinity of the candidate drug. Similarly, Asn51 is involved in hydrogen bonding with Amprenavir, underlining its role as a critical residue in mediating ligand-protein interactions. These shared interactions across different molecules underscore the importance of specific amino acid residues in governing the binding specificity and affinity. Furthermore, the residues Asp564, Glu430, Glu506, and Cys427, identified in the interaction profile of Flavin adenine dinucleotide, appear to be significant across various ligands. Their recurrent presence indicates their pivotal role in the ligand-protein binding landscape. The interactions with water molecules also contribute to the stability of the complex, suggesting the dynamic nature of the ligand-protein interaction network involving these residues. Briefly, our in silico drug screening has demonstrated that candidate drugs consistently interact with the same amino acids, Cys502 and Asp564, as the reference molecule 10N found in the crystal structure during docking or molecular dynamics simulations. This consistent interaction pattern validates the accuracy and reliability of our in silico methodology. Considering all docking and MD simulation results, our compounds, especially Amprenavir, Bosutinib, Ferric derisomaltose, and Lactulose, outperformed the reference and are considered the most promising candidate drugs for FAK inhibition in cancer therapy. Furthermore, the other two molecules, Flavin adenine dinucleotide and Tafluprost, also showed similar potential to the reference molecule. In total, all six molecules have the potential to be promising candidates as FAK inhibitors for cancer treatment.
Zhan et. al.'s research [27] utilized molecular dynamics (MD) simulations and MM-GB/SA calculations to explore combinations, revealing an enthalpy-driven mechanism. Crucial residues, particularly Cys502 and Asp564, were identified for their essential roles in forming hydrogen bonds with inhibitors, consistent with experimental observations. Moreover, Glu500 was noted to establish non-classical hydrogen bonds with each inhibitor. Stronger electrostatic interactions with PHM16 and ligand3 were exhibited by Arg426. Hydrophobic interactions were facilitated by key residues such as Ile428, Val436, Ala452, Val484, Leu501, Glu505, Glu506, Leu553, Gly563, Leu567, and Ser568. These findings hold substantial significance for the advancement of FAK inhibitors, offering valuable insights for cancer research. Particularly noteworthy are the consistently low peaks of Cys502 in all three complexes, underscoring its pivotal role in hydrogen bond interactions. Our study echoes these observations, emphasizing the importance of residues like Cys502 and Asp564, validating the significance of these interactions in our computational drug screening approach. The identified residues, coupled with hydrophobic and electrostatic interactions, contribute significantly to the comprehension of FAK inhibition, supporting the design and development of potential cancer therapeutics [27]. The alignment of key amino acids identified in the study by Zhan et. al. with the significant interaction amino acids found in our study indicates a significant alignment and consistency between the two studies. This correspondence has increased the reliability of our findings.
Furthermore, our study supports the findings outlined in the research carried out by Mustafa et. al. [28] More specifically, they validate the pivotal functions of Cys502, positioned in the kinase hinge, and Asp564, found in the DFG motif within the ATP binding site of Focal Adhesion Kinase (FAK). This consistency emphasizes the importance of these residues in FAK and strengthens their significance in developing potential inhibitors for studying cancer. Our study reveals that the molecule Ferric derisomaltose interacts effectively with amino acids Ile428, Glu430, Glu500, Cys502, Glu506, Asn551, Asp564, and Leu567. The 2D interaction maps are presented in Fig. 5A, the types of interaction bonds in Fig. 5B, and the interaction analyses over a period of 100 ns in Fig. 5C. These findings are supportive of previous research and suggest that Ferric derisomaltose has the potential to be a highly effective FAK inhibitor, as indicated by previous studies conducted on this molecule.
Analysis of the biochemical and pharmacological properties of the leading ligands was conducted by using the MetaCore/MetaDrug platform. This tool enables the prediction of first-pass and second-pass metabolites and assesses various attributes such as reactivity, blood-brain barrier (BBB) permeability, protein binding, and water solubility. Additionally, MetaDrug employs Quantitative Structure-Activity Relationship (QSAR) models to forecast the potential toxic effects and therapeutic efficacy of the ligands under scrutiny. To determine the similarity between these ligands and those encompassed in the QSAR models, we utilized the Tanimoto Prioritization (TP) property. Further, we have previously comprehensively discussed the precision of the QSAR models in our prior studies [29, 30]. Bosutinib, Lactulose, Flavin adenine dinucleotide, and Ferric derisomaltose compounds were predicted to exhibit anti-cancer properties. Furthermore, Amprenavir and Tafluprost showed potential for improvement.
Amprenavir, a protease inhibitor approved for the therapeutic management of HIV infection by the United States Food and Drug Administration on April 15, 1999, has demonstrated efficacy with a dosing regimen of twice daily, marking a significant advancement over prior treatments necessitating administration every eight hours [31]. Beyond its primary application in HIV therapy, amprenavir has been explored both in vitro and in silico for potential utility in treating a range of diseases, including SARS-CoV-2 and various cancers [32, 33]. Subsequent investigations have revealed that amprenavir possesses the capability to inhibit tumor cell proliferation across a diverse spectrum of cancer types, as documented [34]. Additionally, research conducted by Jiang et. al. [33], has demonstrated amprenavir's ability to induce apoptosis in MCF-7 cells in vitro and in vivo through the inhibition of ERK2 kinase activity. Esposito et. al. [35], further substantiated amprenavir's anti-cancer potential by evidencing its capacity to impede migration and proliferation in human hepatocarcinoma cell lines. Considering the cumulative findings from prior studies alongside the outcomes of current research, it is posited that amprenavir may be effectively repurposed as a Focal Adhesion Kinase (FAK) inhibitor in oncological treatments.
Bosutinib, a tyrosine kinase inhibitor (TKI), was granted approval by the United States Food and Drug Administration (FDA) in September 2012 and by the European Medicines Agency (EMA) in March 2013. It is recognized for its application in the treatment of Philadelphia chromosome-positive chronic myeloid leukemia in the chronic phase (CML-CP), achieving its initial pediatric approval for this indication [36]. In a comparative study conducted among CML patients, Cortes et. al. [37] reported that patients treated with Bosutinib exhibited superior responses compared to those receiving Imatinib. Moreover, its efficacy has been investigated in various other cancer types. Singh et. al. [38] demonstrated in their study on MCF-7 cells that liposomal formulations of Bosutinib induced apoptosis in estrogen-positive cell lines. Segrelles et. al. [39] have stated that Bosutinib inhibits the growth of Head and Neck Squamous Cell Carcinoma (HNSCC) particularly by inhibiting the activity of the Epidermal Growth Factor Receptor (EGFR). Yu et. al. [40] concluded in their study with the HeLa cell line that Bosutinib could serve as a significant therapeutic agent against cervical cancer by decreasing the activity of the Src/NF-κB/survivin signaling pathway. Conversely, Watanabe et. al. [41] have reported potential adverse effects of Bosutinib, including diarrhea, hepatic toxicity, and severe lung injury. Given its more recent approval compared to Amprenavir, and notwithstanding the reported side effects, Bosutinib holds potential for repurposing as a Focal Adhesion Kinase (FAK) inhibitor, warranting cautious consideration of its application in this capacity.
Ferric derisomaltose, an iron carbohydrate complex composed of ferric hydroxide and the carbohydrate derisomaltose, received regulatory approval for the treatment of iron deficiency anemia from the European Medicines Agency (EMA) in 2009 and the United States Food and Drug Administration (FDA) in January 2020. A study conducted by Kassianides et. al. [42] demonstrated that the use of ferric derisomaltose in patients with anemia is more cost-effective and efficacious compared to previous formulations, while also exhibiting minimal side effects. In a comprehensive study, Kalra et. al. [43] investigated the application of ferric derisomaltose in patients with heart failure and iron deficiency, revealing a correlation between iron supplementation and a reduced risk of cardiovascular mortality in this patient group, thereby underscoring the safety of ferric derisomaltose use. Auerbach et. al. [44] further validated the rapid amelioration of iron deficiency and the reliability of ferric derisomaltose application in their study. Concerning the potential application of ferric derisomaltose within oncological treatment paradigms, the body of research remains limited. Nevertheless, Dickson et. al. [45] have administered ferric derisomaltose to oncology patients witnessing reductions in hematological parameters, notably hemoglobin concentrations, and have documented favorable outcomes. Review of the existing literature posits ferric derisomaltose as a promising candidate for investigation as a Focal Adhesion Kinase (FAK) inhibitor. Additionally, its deployment in the context of radiotherapy merits consideration for leveraging the radiosensitizing attributes of iron, thereby highlighting its prospects as an efficacious, non-toxic, and therapeutic agent.
Lactulose, synthesized through the isomerization of lactose, initially garnered approval from the United States Food and Drug Administration (FDA) in 1977 and has been recognized as one of the World Health Organization's (WHO) Essential Medicines [46]. In recent years, it has become one of the most frequently prescribed medications in the USA, primarily for the treatment of hepatic encephalopathy [47, 48]. Moreover, emerging evidence suggests that lactulose may play a role as a pharmacotherapeutic agent in the management and prevention of type 2 diabetes through its effects on gut microbiota [49]. Research conducted by Kishor et. al. [50], has indicated that lactulose could serve as an effective Galectin inhibitor, potentially useful in targeted cancer therapy and demonstrating anticancer agent capabilities. Furthermore, Fernández et. al. [51] have shown that galacto-oligosaccharides derived from lactulose significantly reduced the incidence of Colorectal Cancer (CRC) in in vivo models. Based on these findings, we conclude that lactulose can be reliably considered for use as a Focal Adhesion Kinase (FAK) inhibitor.
Tafluprost, a prostaglandin analog, received approval from the United States Food and Drug Administration (FDA) on February 13, 2013, for the treatment of ocular hypertension and glaucoma [52, 53]. The study conducted by Papadia et. al. [54], highlights not only the efficacy of Tafluprost, which is the first prostaglandin analog without preservatives, but also its safety profile and the minimal side effects associated with its use. In recent research, Wu et. al. [55] demonstrated that Tafluprost facilitates axon regeneration through the modulation of the Zn2+-mTOR pathway. While there has been no direct investigation into the applicability of Tafluprost in cancer treatment, studies involving prostaglandins have linkedwith the initiation, progression, and metastasis of cancer [4, 56]. Given this association, it has been concluded that the use of Tafluprost as a Focal Adhesion Kinase (FAK) inhibitor may not be appropriate.
The outcomes derived from this integrated computational methodology bear considerable importance, especially concerning the advancement of therapies for Merlin-negative tumors. The discovery of potent FAK inhibitors using this approach underscores the therapeutic promise of these compounds. Furthermore, the utilization of a diverse compound library encompassing both FDA-approved medications and substances undergoing clinical scrutiny provides opportunities for drug repurposing. Repurposing established drugs presents a potentially swifter and economically efficient approach to drug innovation, particularly in oncology, where the need for rapid treatment solutions is frequently critical.