Advanced Computational Pipeline for FAK Inhibitor Discovery: Combining Multiple Docking Methods with MD and QSAR for Cancer Therapy

Abstract

Synthetic lethality, involving the simultaneous deactivation of two genes, disrupts cellular functions or induces cell death. This study examines its role in cancer, focusing on focal adhesion kinase and Neurofibromin 2. Inhibiting FAK, crucial for synthetic lethality with NF2/Merlin, offers significant cancer treatment potential. No FAK inhibitor has been clinically approved, underscoring the need for new, effective inhibitors. The small-molecule FAK inhibitors identified in this study show promise, with SP docking, IFD, QPLD, and MD simulations revealing intricate interactions. Based on the comprehensive analysis, the MM/GBSA scores from SP docking for amprenavir, bosutinib, ferric derisomaltose, flavin adenine dinucleotide, lactulose, and tafluprost were determined as −72.81, −71.84, −76.70, −69.09, −74.86, and −65.77 kcal/mol, respectively. The MMGBSA results following IFD docking MD identified the top-performing compounds with scores of −84.0518, −75.2591, −71.8943, −84.2638, −56.3019, and −75.3873 kcal/mol, respectively. The MMGBSA results from QPLD docking MD identified the leading compounds with scores of −77.8486, −69.5773, −71.9171, N/A, −62.5716, and −66.8067 kcal/mol, respectively. In conclusion, based on the MMGBSA scores obtained using the three docking methods and the 100 ns MD simulations, and considering the combined evaluation of these methods, amprenavir, ferric derisomaltose, and bosutinib are proposed as the most promising candidates.

1. Introduction

Cancer ranks as the second leading cause of global mortality, surpassed only by cardiovascular diseases. The financial resources allocated to cancer treatment and pharmaceutical research impose substantial burdens on national budgets and are anticipated to increasingly constitute a significant proportion of health expenditures in the years ahead [1]. Researchers are diligently exploring diverse modalities for cancer treatment, producing a multitude of publications based on in vitro and in vivo investigations. Notably, the contemporary emphasis on drug discovery has shifted towards molecular modeling studies [2]. Molecular screening facilitates the elucidation of molecular–target interactions within an in silico environment, thus enabling a reduction in the pool of drug candidate molecules prior to transitioning to in vitro and in vivo settings.

Focal Adhesion Kinase (FAK) emerges as a vital protein kinase governing cell–matrix interactions, particularly influencing cell adhesion and migration [3]. Its activation is instigated through integrin signaling, playing a crucial role in cells adhering to the extracellular matrix [4]. The autophosphorylation of FAK and its subsequent conformational activation intricately regulate fundamental cellular processes including growth, proliferation, and survival through the modulation of signal transduction pathways. Excessive FAK activation has been implicated in various diseases, notably cancer. Consequently, FAK inhibitors are under rigorous investigation as potential therapeutic agents in cancer management, with a specific focus on metastasis prevention [5].

The NF2 gene, responsible for encoding Merlin protein (Moesin–Ezrin–Radixin-Like), functions as a tumor suppressor protein [6]. Merlin’s normal role involves the regulation of cell growth and division, the control of cell adhesion, and the modulation of cellular interactions with the microenvironment. The frequent inactivation of NF2 is observed in diverse cancer types such as schwannoma, meningioma, and malignant mesothelioma (MM), the latter being a highly aggressive cancer with limited treatment options [7]. NF2 gene mutations disrupt the normative functionality of the Merlin protein, resulting in the manifestation of NF2 disease. Neurofibromatosis Type 2 (NF2) is characterized by a genetic predisposition causing an aberrant proliferation of nervous system cells, particularly manifesting as tumor formation between Schwann cells and other cellular entities, predominantly within the brain and spinal cord. The NF2 gene mutations induce aberrations in Merlin protein regulation, instigating uncontrolled cellular proliferation and tumor development [8].

Synthetic lethality constitutes a crucial concept in contemporary cellular biology and cancer research, particularly in the development of therapeutic strategies aimed at targeting cancer cells bearing specific genetic mutations [9]. A synthetic lethal interaction arises when the disruption of either gene is individually viable, but the simultaneous disruption of both genes leads to cell death. Central to leveraging synthetic lethality in cancer therapy is the discovery and detailed understanding of robust synthetic lethal genetic interactions [10]. Traditional approaches involve the suppression or inhibition of specific genes within tumors, a methodology effective in targeting cancer cells but often associated with collateral damage to normal cells. The synthetic lethality strategy seeks to invert this paradigm by exploiting genetic interactions. By suppressing the function of a gene in normal cells without harm, a severe detriment can be induced in cancer cells harboring specific genetic mutations [11].

Emerging research underscores a synthetic lethal relationship between focal adhesion kinase (FAK) and NF2/Merlin. Notably, cancer types such as schwannoma, meningioma, and malignant mesothelioma (MM) and MPM exhibit mutations in the NF2 gene, leading to the impaired synthesis of the Merlin protein [12]. The underlying hypothesis posits that in instances where cancer cells recurrently experience a loss of NF2/Merlin expression due to various inactivation mechanisms and mutations, the FAK protein presents a promising target for cancer therapy [13,14,15]. This study aims to identify small-molecule drug candidates specifically targeting FAK in Merlin-negative tumors. Therefore, our objective is to develop a specific therapy that effectively targets cancer cells through synthetic lethality, while healthy cells, due to their positive Merlin status, remain unaffected by this treatment approach. This approach ensures the development of therapeutic interventions that selectively kill cancerous tissues, offering a more precise and effective strategy for cancer therapy.

The ongoing advancement of novel FAK inhibitors underscores the importance of this approach in the landscape of cancer therapeutics. Despite concerted research efforts, however, no FAK inhibitor has yet garnered clinical approval [16]. Consequently, there is an urgent and critical necessity to develop new FAK inhibitors with heightened antitumor activity to achieve more potent antitumor effects [17].

In NF2-deficient cells, the selective targeting of FAK inhibitors without affecting healthy cells is a critical aspect that requires further elaboration. To address this, the current study focuses on FDA-approved drugs, emphasizing their established safety profiles and potential for repurposing. These drugs have been extensively evaluated for their effects on healthy cells, which supports their suitability for selective targeting in NF2-deficient cancer cells. By leveraging the safety data of FDA-approved compounds, this study aims to minimize adverse effects on healthy cells while maximizing therapeutic efficacy against cancerous cells. The investigation involved the docking of all FDA-approved drugs and compounds under clinical investigation, totaling 3235 compounds, against the FAK protein target. The evaluation of hit molecules with the highest potential as drug candidates was facilitated through a comparative analysis of their docking scores against reference compounds. Molecules with docking scores lower than −8.00 kcal/mol were selected for further investigation. Selected top-performing molecules underwent molecular dynamics (MD) simulations of 1 nanosecond (ns), 10 ns, and 100 ns, culminating in the identification of optimal drug candidate molecules. Furthermore, each top-ranked molecule identified as the best in SP was subjected to IFD and QPLD processes, followed by 100 ns long MD simulations and MM-GBSA analyses for all, to conduct detailed binding analyses of the ligands to the target. Subsequently, machine learning-based QSAR models were employed to predict the therapeutic activity of these molecules (Figure 1). The findings contribute to advancing our understanding of potential treatments for Merlin-negative tumors through FAK inhibition.

2. Results and Discussion

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 (Figure 2), and all ligands were docked into this grid box. This step, which involved a 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. To better understand the interaction between the protein and ligand, standard precision (SP) docking—a method known for its rapid and efficient prediction of binding affinity—was utilized. To account for potential dynamic changes in the receptor’s binding site, molecules that showed promising results in SP docking were further analyzed using induced fit docking (IFD). Additionally, quantum polarized ligand docking (QPLD) was applied to all molecules that performed well in SP docking. This approach integrates quantum calculations with both quantum mechanics (QM) and molecular mechanics (MM) methods to achieve highly accurate binding affinity predictions.

Molecules with SP docking scores lower than −8.00 kcal/mol were selected for detailed analysis. The reference compound was bound to the FAK protein with a docking score of −10.85 kcal/mol (Figure 3). 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 [18], 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. SP docking scores and MM/GBSA analysis rely on 1 ns, 10 ns, and 100 ns 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 ± 4.39
4GU6_Bosutinib	−8.12	−75.79	−74.38	−71.84 ± 4.49
4GU6_Ferric derisomaltose	−8.00	−82.72	−73.52	−76.70 ± 5.23
4GU6_Flavin adenine dinucleotide	−8.47	−89.15	−82.99	−69.09 ± 10.8
4GU6_Lactulose	−8.23	−71.93	−73.39	−74.86 ± 4.52
4GU6_Tafluprost	−8.13	−72.51	−73.13	−65.77 ± 6.49
4GU6_Reference compound	−10.85	−69.56	−71.04	−69.20 ± 3.58

. 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 (Figure 4).

Seventy-one molecules with docking scores of −8.00 kcal/mol or better underwent simulation, followed by MD simulations and 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 was demonstrated to be 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, nine 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, nine 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.

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. In this step of the study, each top-ranked molecule identified as the best in standard precision (SP) docking and subjected to a 100 ns molecular dynamics (MD) simulation was further analyzed using induced fit docking (IFD) and quantum polarized ligand docking (QPLD). These analyses allowed for a detailed examination of the interactions between the ligands and the target protein. Following the IFD and QPLD processes, all molecules underwent an additional 100 ns MD simulation, along with MM-GBSA analyses to calculate binding free energies. The results of these analyses are presented in

Table 2. QPLD and IFD scores and MM/GBSA analysis rely on 100 ns molecular dynamics simulations.

Ligand	QPLD Glide Score	QPLD 100 ns MD-MMGBSA	IFD Docking Score	IFD 100 ns MD-MMGBSA
Amprenavir	−9.02	−77.85 (±4.6)	−11.60	−84.05 (±4.3)
Flavinadeninedinucleotide	N/A	N/A	−14.56	−84.26 (±9.9)
Lactulose	−8.42	−62.57 (±8.9)	−8.73	−56.30 (±10.6)
Tafluprost	−7.96	−66.81 (±6.1)	−11.02	−75.39 (±5.3)
Bosutinib	−8.40	−69.58 (±5.1)	−11.38	−75.26 (±4.4)
Ferricderisomaltose	−8.97	−71.92 (±6.5)	−9.78	−71.89 (±10.0)
10N_Reference_molecule	−11.53	−78.80 (±4.3)	−11.07	−75.76 (±5.8)

. The outcomes from all three methods were collectively evaluated to provide a comprehensive understanding of the binding affinity and dynamic interactions between the ligands and the target protein. 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.

According to the QPLD and IFD analyses, the top-performing compounds were identified based on their MM-GBSA scores following QPLD docking and MD simulations. The MMGBSA results from QPLD docking MD identified the leading compounds with scores of −77.8486, −69.5773, −71.9171, N/A, −62.5716, and −66.8067 kcal/mol, respectively. The MMGBSA results following IFD docking MD identified the top-performing compounds with scores of −84.0518, −75.2591, −71.8943, −84.2638, −56.3019, and −75.3873 kcal/mol, respectively.

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 10 N, interactions were observed with amino acids Cys502, Glu506, Arg550, Val552, Leu553, Asp564, and Leu567 during the 100 ns MD simulation, confirming our findings. Importantly, amino acids Cys502 and Asp564 were previously highlighted in the literature as crucial, which supports our results [16,19].

Notably, the Cys502 residue emerges as a common binding site, forming hydrogen bonds with both the reference molecule compound 10 N (Figure 5), 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 demonstrated that candidate drugs consistently interact with the same amino acids, Cys502 and Asp564, as the reference molecule 10 N 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. When all docking and MD simulation results are considered, 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 [20] 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. 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. [21] 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 the amino acids Ile428, Glu430, Glu500, Cys502, Glu506, Asn551, Asp564, and Leu567. The 2D interaction maps are presented in Figure 6A, the types of interaction bonds in Figure 6B, and the interaction analyses over a period of 100 ns in Figure 6C. These findings are supportive of previous research, and they suggest that ferric derisomaltose has the potential to be a highly effective FAK inhibitor, as indicated in previous studies conducted on this molecule.

Analysis of the biochemical and pharmacological properties of the leading ligands was conducted using the MetaCore/MetaDrug platform. This tool enables the prediction of first-pass and second-pass metabolites, and it 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 [22,23]. Bosutinib, lactulose, flavin adenine dinucleotide, and ferric derisomaltose compounds were predicted to exhibit anticancer properties. Furthermore, amprenavir and tafluprost showed potential for improvement (QSAR-based therapeutic activity prediction table provided in the Supplementary Information, Table S4).

Amprenavir, a protease inhibitor approved for the therapeutic management of HIV infection by the United States Food and Drug Administration on 15 April 1999, has demonstrated efficacy with a dosing regimen of twice daily, marking a significant advancement over prior treatments necessitating administration every eight hours [24]. 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 [25,26]. Subsequent investigations have revealed that amprenavir possesses the capability to inhibit tumor-cell proliferation across a diverse spectrum of cancer types, as documented [27]. Additionally, research conducted by Jiang et al. [26], 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. [28], further substantiated amprenavir’s anticancer potential by evidencing its capacity to impede migration and proliferation in human hepatocarcinoma cell lines. The most common side effects of amprenavir include skin rashes, which affect up to 18% of patients, and diarrhea [29]. 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 [30]. In a comparative study conducted among CML patients, Cortes et al. [31] 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. [32] demonstrated, in their study on MCF-7 cells, that liposomal formulations of bosutinib induced apoptosis in estrogen-positive cell lines. Segrelles et al. [33] 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. [34] 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. Bosutinib was generally well tolerated, exhibiting a distinct safety profile compared to dasatinib in a similar patient population, with gastrointestinal adverse effects being the most common [35].

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. [36] 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. [37] 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. [38] 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. [39] have administered ferric derisomaltose to oncology patients witnessing reductions in hematological parameters, notably hemoglobin concentrations, and have documented favorable outcomes. A review of the existing literature posits ferric derisomaltose as a promising candidate for investigation as a focal adhesion kinase (FAK) inhibitor. Ferric derisomaltose has an established safety profile with no known significant adverse effects. However, based on the reports of Khan et al., (2021) [40], this drug can be recommended for iron therapy, particularly due to its suitability for single-dose administration. Nevertheless, further studies are necessary to compare its efficacy with other parenteral formulations, especially in dialysis-dependent CKD patients. 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 [41]. In recent years, it has become one of the most frequently prescribed medications in the USA, primarily for the treatment of hepatic encephalopathy [42,43]. 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 [44]. Research conducted by Kishor et al. [45], 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. [46] have shown that galacto-oligosaccharides derived from lactulose significantly reduced the incidence of colorectal cancer (CRC) in in vivo models. Although effective for its intended purposes, lactulose is commonly associated with adverse gastrointestinal reactions, such as bloating, cramping, and diarrhea, which may limit its tolerability in some patients [47]. 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 13 February 2013 for the treatment of ocular hypertension and glaucoma [48,49]. The study conducted by Papadia et al. [50] 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. [51] 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 linked with the initiation, progression, and metastasis of cancer [52]. In the study by Katsanos et al., (2022) [53], conjunctival hyperemia, or eye redness, is identified as the most commonly reported adverse effect associated with the use tafluprost. Given this association, it has been concluded that the use of tafluprost as a focal adhesion kinase (FAK) inhibitor may not be appropriate.

In their study on calpain-mediated androgen receptor cleavage, Libertini et al., (2007) [54] reported that the increased synthesis of FAK (focal adhesion kinase) was associated with androgen receptor activity, contributing to the progression of prostate cancer. However, they also noted that calpain inhibitors were able to halt this progression. Additionally, they reported that amprenavir significantly reduced tumor growth, potentially mimicking the effects of calpain inhibitors. This effect may be attributed to amprenavir’s potential inhibition of FAK activity. Nisha et al. (2024) [55] investigated folate-functionalized bosutinib cubosomes against hepatic cancer cells through in vitro, in silico, and in vivo pharmacokinetic approaches. They demonstrated that bosutinib effectively inhibits both Src (a non-receptor tyrosine kinase) and FAK (focal adhesion kinase), enhancing its potential as a therapeutic agent in hepatic cancer. The literature on the effects of ferric derisomaltose and tafluprost are limited, and there are very few studies related to lactulose. Rabinovich et al. (2006) [56] demonstrated that synthetic lactulose amines represent a novel class of anticancer agents capable of inducing tumor cell apoptosis. These compounds exert their effect by binding to galectins-1 and galectins-3, which are involved in the FAK signaling pathway, thereby disrupting critical processes in cancer cell survival.

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.

3. Materials and Methods

3.1. Preparation of Ligands for Docking

The ligand structures were created using the Maestro 2D Sketcher program, and they were optimized for molecular docking studies. The preparation process ensures compatibility with a physiological pH of 7.4. This involves employing the PROPKA1 algorithm [57], which predicts protonation states of ionizable groups at a given pH, and utilizing the OPLS3e2 force field [58], a set of parameters that describe the interactions between atoms in the molecular system. Furthermore, the LigPrep3 module [59] was employed to refine and preprocess the molecular structures. LigPrep3 is a tool that can generate different ionization states, tautomers, and conformers of a molecule, enhancing the accuracy and reliability of subsequent molecular docking simulations. The Epik module [60] of the Maestro molecular modeling package (Schrödinger Release 2018, Epik, Schrödinger), a program engineered for the rapid and accurate prediction of aqueous phase pKa values and protonation state distributions for complex, drug-like molecules, played a crucial role in this phase. It facilitated the refinement of ligands, which were systematically assessed as potential small-molecule drugs, and meticulously prepared for subsequent docking procedures.

3.2. Preparation of Protein for Docking

Protein preparation and docking were conducted following established protocols detailed in a prior publication [61]. The crystallographic structure of the FAK protein, identified with PDB accession number 4GU6 [62], was retrieved from the Protein Data Bank and prepared as in our previous work [61]. Utilizing the Prime module, missing side chains and regions were comprehensively completed. The protein target was elaborately prepared to undergo molecular docking and simulation studies in accordance with the physiological pH of 7.4.

3.3. Grid Box Generation and Molecular Docking Studies

A grid box was created using the grid center coordinates derived from the co-crystallized ligand 10 N. Ligands were carefully positioned within this specified grid area on the protein target. This process was carried out meticulously using Maestro’s Glide module [63], utilizing standard precision (SP) settings. As a control measure, reference molecules underwent the same procedures, following identical methodologies as the ligands and ensuring methodological consistency [64]. IFD is a distinctive approach to computer-aided drug discovery that enables the rapid and cost-effective prediction of protein–ligand binding interactions. In order to allow for receptor flexibility and take into account the changes in the active region of the receptor, a second docking process was performed to confirm our SP glide docking results using the IFD method, which is a very robust and accurate method that allows flexibility in both the protein receptor and the ligand. Via the creation of a gridbox centered on co-crystallized ligand coordinates, all drug candidates that were FDA-approved and in clinical investigation were targeted against this gridbox. IFD was performed using the Maestro’s IFD module. IFD scores were generated for each output pose. The binding affinities and modes of molecules to proteins were examined. The IFD pose predictions for docking were conducted using the OPLS3e force field and the B3LYP-D3 functional with a 6–31 G** basis set for DFT calculations. Gas-phase parameters were applied, and the induced fit docking score was considered in the analysis. For quantum polarized ligand docking (QPLD), the OPLS3e force field was utilized for molecular mechanics calculations. DFT calculations were performed using the B3LYP-D3 functional and the 6–31 G** basis set.

3.4. Molecular Dynamics Simulations and MM/GBSA Calculations

For the top docking poses, molecular dynamics (MD) simulations were conducted to replicate physiological conditions using Desmond. An orthorhombic box with explicit water models (TIP3P) was employed. To neutralize the system, 0.15 M of NaCl was added to the simulation box. The temperature (310 K) and pressure (1.01325 bar) were kept constant throughout the simulations using the Nose–Hoover thermostat and Martyna–Tobias–Klein barostat. MD simulations utilized the OPLS3e force field. One hundred trajectory frames were collected for each system during the simulations. The molecular mechanics generalized Born surface area (MM/GBSA) method was subsequently applied to these trajectories, and the average MM/GBSA score for each studied compound was computed. The MM/GBSA calculations utilized the VSGB 2.0 solvation model in the Prime module of Maestro [23].

4. Conclusions

In conclusion, the integrated analysis of molecular docking, molecular dynamics, and MM-GBSA calculations for ligands interacting with the protein target 4GU6 provides valuable insights into potential drug candidates. The consistent trends observed in both docking and MM-GBSA analyses enhance the reliability of computational methods in predicting ligand binding affinities. In this study, the number of molecules in the library was initially filtered using 1, 10, and 100 ns MD simulations with SP docking. Promising compounds identified through this process were further subjected to QPLD and IFD analyses, followed by additional long-term MD simulations and MM-GBSA calculations. The collaborative use of SP, IFD, and QPLD methods, along with the comprehensive evaluation of their results, ensured a robust and thorough assessment of ligand–protein interactions. According to these results, amprenavir, bosutinib, and ferric derisomaltose consistently emerged as the most promising drug candidates, exhibiting the lowest docking scores and the most negative MM-GBSA values from MD simulations. This consistent and strong binding affinity highlights the potential of amprenavir, bosutinib, and ferric derisomaltose as effective therapeutic agents for the targeted protein. In contrast, the reference compound consistently showed higher docking scores and MM-GBSA values in both docking and MM-GBSA assays, indicating weaker binding affinities. Our study illuminates new therapeutic pathways targeting cancer through focal adhesion kinase (FAK) inhibition, a critical area where no clinically approved FAK inhibitor currently exists and where there is an urgent need for the development of novel FAK inhibitors with antitumor properties. This study lays the groundwork for promising new therapeutics for various cancer treatments while broadening the impact in the field of drug design by providing a general approach that extends beyond FAK inhibitors. Additionally, by repurposing FDA-approved drugs, this approach facilitates the approval process for phase-stage anticancer drugs, leveraging compounds with well-established safety profiles and known pharmacodynamic and pharmacokinetic properties, thereby accelerating the development timeline.