Protein profiling uncovers IGF-1R inhibition potential of 3-(2-furoyl)-indole scaffolds in hepatocellular carcinoma

ABSTRACT

Aim

This study investigates the anti-proliferative potential and possible molecular mechanisms of 3-(2-furoyl)-indole derivatives against HepG2.

Method

Identified hit compounds (4a, 4b, 4c) using MTT screening, were further investigated for their efficacy and mechanism of action through FACS studies, in-silico molecular docking, molecular dynamics (MD) simulations, and label-free quantitative proteome and ADMET prediction.

Results

Lead compound 4a, showed IC50 of 27 µM against HepG2 cells and a binding score of −8.077 kcal/mol against IGF-1 R (PDB ID: 5XFS) and formed a stable complex 100 ns. Proteomic study revealed significant downregulation of the IGF-1 R downstream signaling molecules and showed minimal toxicity and favorable drug-like properties.

Conclusion

These findings suggest that 4a is a promising IGF-1 R inhibitor and potential drug candidate against drug resistance hepatocellular carcinoma (HCC).

1. Introduction

According to the Cancer Agency of World Health Organisation (WHO), new cancer cases will increase to approximately 35 million by 2050, which is about 77% increase, and one in every five people is diagnosed with cancer [1]. Among patients diagnosed with cancer, the death ratio is 1:9 and 1:12 for men and women, respectively [2]. Hepatocellular carcinoma (HCC) is one of the major global health burdens, accounting for an incidence of over 1 million by 2025, which ranked as the fourth most frequent cause of cancer-related death worldwide [3,4]. Among various risk factors, hepatitis B and C virus infection and nonalcoholic fatty liver diseases (NAFLD) are the major risk factors of HCC. Currently, six approved therapies are available for the treatment of HCC: sorafenib, atezolizumab plus bevacizumab, regorafenib, lenvatinib, ramucirumab, and cabozantinib [5]. In addition, new therapies are under development which explore combinations of various inhibitors, such as checkpoint inhibitors, tyrosine kinase inhibitors, anti-VEGF agents, and immunotherapy regimens [6]. Despite newer therapies and control strategies, drug resistance is widespread in all types of cancer, including HCC [7]. Hence, there has been a continuous approach among researchers to develop new anticancer agents against drug resistance, especially for combination systemic therapy with a special emphasis on multi-targeting and novel mechanisms. In HCC, growth factor receptors (e.g., VEGFR, FGFR, TGFA, EGFR, IGFR), cytoplasmic intermediates (e.g., PI3K-AKT-mTOR, RAF/ERK/MAPK), and critical cell differentiation pathways (e.g., Wnt/β-catenin, JAK/STAT, Hippo, Hedgehog, Notch) are significantly dysregulated. Among growth factors, the IGF signaling pathway plays a key role in regulating cellular proliferation, differentiation, and apoptosis [8–10]. IGF-1 R overexpression has been reported in poor prognosis in HCC [11]. Furthermore, IGF-1 R expression and activation are correlated with drug resistance in tumors [12]. Therefore, silencing or inhibiting the expression of IGF-1 R might enhance the response of HCC to drugs like sorafenib. Contrarily, in drug-resistant breast cancer like Triple negative breast cancer (TNBC), it is correlated with low IGF-1 R expression, and the underlying mechanism is not clear. In TNBC, IGF-1 R targeting is an ineffective therapeutic target and has failed clinically [13]. However, IGF-1 R is considered a potential therapeutic target in HCC [14].

Indole is a crucial heterocyclic structure as it is incorporated into proteins through the amino acid tryptophan [15]. It is also found in the structure of vincristine and vinblastine, which are FDA-approved anticancer agents and well-known indole alkaloids obtained from the plant Catharanthusroseus [16]. Other examples of naturally occurring anticancer agents containing indole scaffold are harmine, dehydrocrenatidine, vinorelbine, (3’R)-hydroxytabernaelegantine C, eusynstyelamide B, chetomin, brucine, flavopereirine, jerantinine B [17–25]. Development of the indole-base synthetic compound has been an increasing interest in the field of medicinal chemistry, as it is a biologically accepted scaffold in many pharmaceutical compounds. It exerts a wide spectrum of pharmacological activity [26–28]. Indole is well known for its high binding affinity as well as its ability to bind multiple targets [15]. Thus, it enhances the efficacy of the treatment by blocking multiple pathways, especially in cancer [29]. Furthermore, more than 70 indole-containing drugs are globally marketed and have been proven to exhibit excellent clinical outcomes. Some examples of FDA-approved indole containing synthetic anticancer agents are panobinostat, osimertinib, rucaparib, Lurbinectedin, etc [30]. In particular, aryl/heteryl clubbed indole compounds have drawn special attention for anticancer drug development and there has been a well-established structure-activity relationship (SAR) [31]. Many natural compounds of 3-aryl/heteryl substituted indole compounds or 2–3 fused compounds demonstrate potent cytotoxic effects against HCC and breast cancer [32]. While reviewing the literature on synthetic indole scaffolds, we noted that 3-substituted indole derivatives are potential cytotoxic compounds and have been investigated against a variety of cancer cell lines [33]. Thus, we continued to investigate the furan-clubbed indole scaffolds on HCC. In view of the above and the continuation of the reported synthetic work of our laboratory on 3-(2-furoyl) indole scaffolds [34], herein, we aimed to investigate the anticancer activity of these compounds with reference to their target interaction in HCC. Considering the literature on the cancer-specific cytotoxic effects of furan-clubbed indole compounds on breast cancer and HCC, we tested these compounds in MCF-7 and HepG2 cell lines. Based on the potential cytotoxicity, molecular mechanism studies are designed for HCC.

Over a decay, several studies have demonstrated the synthesis of polycyclic indole derivative via metal-catalyzed reactions, N-heterocyclic carbene (NHC) catalysis, tertiary amine catalysis, acid-catalyzed, calcium phosphate catalysis etc. However, such catalysts are costly, toxic in nature, and low availability of active sites [35–38]. Therefore, a sustainable, low-cost, environmentally friendly, metal-catalyst-free synthesis was developed using graphene oxide (GO) to address these drawbacks. GO has a great advantage over the previously reported catalyst for synthesizing polycyclic indole derivatives because of its high surface area, a large number of active sites and exhibits excellent catalytic performance. Furthermore, the surface of the graphene oxide sheet is composed of sulfonic acid, epoxy, carboxyl, and hydroxyl group which gives to its mild acidic and oxidative properties. Thus, this mild acidic reagent is used instead of other hazardous acid catalysts because of its ability to tolerate a wide range of functional groups. In addition to this, GO can be easily recovered and recycled and reused five times, without considerable reduction in the catalytic activity of the reaction [39]. Therefore, in this study, GO was utilized as a catalyst to synthesized new furan-indole derivatives.

In this study, we performed an in silico molecular docking and mass spectrometry-based proteome profiling to unveil the mechanism and interacting pathway of the active compounds in HepG2 cell lines. In addition, the induced apoptotic activities of active compounds were studied against HepG2 cells using FACS analysis. Furthermore, the pharmacokinetics and toxicity of all synthesis compounds were studied to predict the overall druggability of the active compounds.

2. Experimental section

2.1. Chemistry

General procedure of the 3-(2-furoyl)-indole derivatives synthesis: 2.0 mmol of glyoxal 1, 2.0 mmol of 1,3 diketone 2 and 2.0 mmol of compound 3 were microwaved at 150W for 3 min X 2. Then, add 60 mg of GO and reaction mixture was heated at 70°C for 5 h. Then product was extracted from the surface of GO by adding 5 ml of ethanol followed by ultrasonication and decantation. The separated organic layer was evaporated to get the crude mass which was purified by column chromatography to obtain the desired product, which was further characterized by ¹H and ¹³C NMR (SI Figure 1).

Figure 1.

List of 3-(2-furoyl)-indole scaffolds used in this study and their synthetic route.

2.2. Molecular docking and MD simulations

In HCC, cell-signaling pathways are commonly altered, particularly through the involvement of growth factor receptors like VEGFR, TGFA, FGFR, EGFR, and IGFR [40]. Out of which IGF-1 R expression and activation are associated with drug resistance in tumors [12]. Previous studies showed promising IGFR inhibitory activity of indole base derivatives [41,42]. Therefore, based on the literature, these growth factor receptors were selected for docking studies. In order to evaluate the binding affinity and interactions of the synthesized compounds to the targeted HCC protein. The crystal structures of integrin (PDB ID: 1JV2), EGFR (PDB ID: 1M17), and IGF-1 R (PDB ID: 5FXS) were selected for docking studies due to their high resolution, biologically relevant active conformations, and the presence of co-crystallized ligands for validation. These structures provide accurate models for evaluating binding affinity and inhibitory potential of compounds against IGF-1 R and EGFR proteins. The PDB files were obtained from the Protein Data Bank (PDB) for further analysis. Initially, protein PDB ID was loaded into Schrodinger’s software, followed by protein preparation in the protein preparation wizard by assigning bond order, using CCD databases, replacing hydrogens, creating zero-order bonds to metals, and converting selenomethionines to methionines, deleting waters beyond hets 5.00 Å. Then, the protein minimization was done through OPLS4 force field. A grid box was generated as per the co-crystallized ligand of the respective protein. For ligand preparation, the structure of compounds was loaded into Schrodinger’s software in .mol file format and was prepared in the LigPrep module. Finally, docking was performed using glide. The resulting interaction between the ligand and protein amino acid residue was visualized through 2D and 3D interaction and binding energy, and docking score was observed to establish the affinity of the ligand toward the protein.

Molecular dynamics (MD) simulation was performed using Schrodinger for 100 ns to understand and validate the conformational changes of 5FXS and 5FXS −4a complex. The 5FXS −4a complex was placed within an orthorhombic box to simulate a realistic environment. The above complex was neutralized by adding Na+ and Cl−, and SPC water molecules were introduced for solvent environment. The complex was subjected to a simulation at a temperature of 300 K and a pressure of 1 bar. The Root Mean Square Deviation (RMSD) quantifies the average displacement of a set of atoms in a specific frame compared to a reference frame. In contrast, the Root Mean Square Fluctuation (RMSF) is helpful for assessing localized changes throughout the protein chain [43].

2.3. Pharmacophore modeling

Here, Schrodinger suite version 2021–2 was utilized to employ the phase model in constructing a pharmacophore hypothesis tailored for IGF-1 R inhibitors. Protein-ligand complexes were initially generated and subsequently imported into the workspace. Within this workspace, the receptor-ligand complexes were subjected to pharmacophore development procedures utilizing Auto (e-pharmacophore) functionalities. This approach allowed for the elucidation of crucial structural and chemical features essential for the interaction between IGF-1 R and its inhibitors, thereby aiding in the rational design of novel therapeutic agents targeting this receptor [44]. In the pharmacophore hypothesis development process, the entry ligand was carefully selected, and in the hypothesis settings, excluded volumes were delineated around the receptor with a shell thickness of 5Å. Subsequently, the pharmacophore model was generated, incorporating these excluded volumes. The generated hypothesis model was validated by loading both active compounds and decoys, where decoys were selected from the online database DUD A Directory of Useful Decoys. (https://dud.docking.org/r2/). The active compounds consisted of structurally similar molecules with IC₅₀ values spanning from 27 to 164 µM, corresponding to a range of pIC₅₀ values from −3.02 to −5.1. Meanwhile, decoy molecules were employed to assess the model’s discriminatory power. The enrichment factor of the decoy set was calculated to rank the compounds, providing insights into the model’s predictive capability and its ability to differentiate true ligands from non-binders or decoys.

2.4. In vitro cytotoxicity study

Cytotoxicity of the 13 compounds was screened against two cell lines HepG2 and MCF-7 by employing an MTT assay. IC₅₀ value of the screened compounds were determined. Initially, 100 µl of 1 × 10⁴ cells were seeded with 10% PBS in each well of 96-well and incubated for 24 hr. Cells were treated with 0–100 µM of all synthesized compounds and kept for 48 hr of incubation. 0.5 mg/mL of MTT dye was added to each well and incubated for 4 h. Remove the supernatant and add 100 µl of DMSO on each well and analyze the sample at 570 nm using Multimode Reader (Synergy Biotek).

2.5. Flow cytometry analysis

Out of all of the screened compounds, three potent compounds (4a, 4b and 4c) as per obtained IC₅₀ from MTT assay were selected to further investigate their anti-apoptotic activity against HepG2 cells. This assay was done by employing Annexin-V FITC/PI staining kit (BD Biosciences). Initially, 1 X 10⁶ cells were plated per dish of the 6-well plate using Dulbecco’s modified eagle medium media and incubated for 24 hr. Treated with the above three selected compounds and kept for incubation. After 12 h of incubation cells were harvested and transferred to sterile FACS tubes, and centrifugation at 2000 rpm for 5 min and gently removed the supernatants. Using cold PBS, cells were washed two times and re-suspended using 1X binding buffer (1 × 10^6 cells/ml). In a 5 ml culture tube, 100 µl of the solution (1 × 10^5 cells) was transferred. Subsequently, 5 µl of both PI and FITC Annexin V were added. The solution was mixed in vortex and kept in a dark room to incubate for 15 min at room temperature, followed by the addition of 400 µl of 1X binding buffer. Finally, it was analyzed by using flow cytometry (Becton Dickinson, CA, USA) [45].

2.6. Mass Spectrometric base of Proteome analysis

2.6.1. Sample preparation for quantitative proteomics

The sample preparation and work-flow was conducted as per our in house procedure and reported [46]. For proteomic analysis 20,000 hepG2 cells were seeded in each well of 6-well plates and was incubate for 12 hr. Followed by treatment with 4a, 4c, and 4b at concentrations of 30, 35, and 30 μM respectively, while DMSO was used as a vehicle control (non-treated), and incubated for an additional 24 h. For protein extraction, both treated and non-treated cells were collected and washed two times with 1×PBS buffer. The cells were lysed using RIPA buffer containing the protease inhibitor phenylmethylsulfonyl fluoride (PMSF). The mixture was vortexed and incubated at 37°C for 30 minutes, then centrifuged at 14,000 × g for 15 minutes to obtain the supernatant. Concentration of the lysate was determined using BCA assay. A protein volume corresponding to 200 μg was quantified and precipitated by adding four volumes of pre-chilled acetone. The mixture was incubated at −20°C for 10 hours. The precipitated protein was collected by centrifugation at 15,000 × g for 10 minutes at 4°C, and the remaining acetone was removed by gently air-drying the pellet. Further in-solution digestion was performed [45]. Protein pellet obtained after acetone precipitation was resuspended in 100 μL of denaturing buffer (7 M urea/50 mm ammonium bicarbonate). Samples were reduced with dithiothreitol (DTT) of 10 mm final concentration and incubated for 1 h at 37°C. Subsequently, alkylation was performed with iodoacetamide (IAA) of 40 mm final concentration, followed by incubation in the dark at 37°C for 1 h, further quenching of excess IAA with DTT of 10 mm final concentration at 37°C for 1 h. Finally, trypsin was added at a ratio of 1:50 (enzyme: protein) and incubated at 37°C for 16 h. Trypsin activity was quenched with formic acid (FA) at final concentration of 0.1%. Desalting of the tryptic peptides was done using C18 spin columns (Thermo Fisher Scientific). Samples were vacuum-dried using a speed vac system at 35°C and reconstituted with 0.1% FA before analysis by mass spectrometry.

2.6.2. Database search and protein detection parameters

Resulted samples both treated and non-treated were analyzed in triplicate (n = 3) using Orbitrap Exploris 240 mass spectrometer connected with Dionex Ultimate 3000 RSLC nano System (ThermoFisher Scientific). Peptides are separated using existing protocol with some modifications [47]. In summary, Peptides samples were injected on a C18 trap column (100 μm × 2 cm, 5 μm) at 5 μL/min using a 0.1% FA mobile phase. After loading on trap column, peptides were then passed into analytical C18 column (75 μm × 50 cm, 2 μ), equilibrated prior to elution with 95% solvent A (0.1% FA) and 5% solvent B (100% Acetonitrile, 0.1% FA), with a flow rate of 300nL/min, using a gradient elution program of 140-min (solvent B gradient as follows: 5% for 0–15 min; 5–50% for 15–75 min; 50–95% for 75–90 min; 95% for 90-120 min; 5% for 120–140 min) with injection volume of 1 µL. The analytical and trap column temperature was maintained at 40°C, whereas the auto-sampler temperature was set to 5°C. The mass spectrometer was used in a data-dependent acquisition mode. Survey scans were acquired in the Orbitrap at 350–2000 m/z range, with a resolution set to 60,000 and a data-dependent MS/MS setup that employed higher energy collisional dissociation (HCD) to induce fragmentation. Within the survey scans, the 15 most intense peaks were selected for fragmentation. The target intensity values for fragmentation were set at 5000 charged, with a resolution of 15,000. Protein quantification was performed using Proteome Discoverer 2.5 (Thermo Fisher Scientific), The protein database employed was a compilation of the reference proteome for Homo sapiens. (Fasta file consist of protein sequences retrieved from UniProtKB). MS/MS spectra were matched using SequestHT program. The carbamidomethylation and N-terminal acetylation of cysteine were set to a static modification. Methionine (oxidation) was designated as the dynamic modification and the lowest missed cleavage number was set at 2.

2.6.3. Gene Ontology (GO) analysis and KEGG pathway by STRING and proteome discoverer

Gene Ontology (GO) of all differentially expressed proteins treated with compounds 4a, 4b and 4c were mapped using Proteome Discoverer to their enriched GO with a p-value of ≤ 0.05. KEGG pathway analysis was performed using STRING to understand underlying protein-protein interactions which are involved in our proposed pathway of compound 4a.

2.6.4. In silico ADMET

The predicted pharmacokinetics properties and toxicity of our synthesized compound were studied using online software; Swiss ADME, admet SAR and proTox-II online. In this, Molar Refractivity, GI absorption, inhibition CYP1A2, Bioavailability Score and The Lipinski Rule of five (Molecular weight <500 Da, <5 hydrogen donors, <10 hydrogen acceptor, and partition coefficient < 5), Skin sensitization, Eye irritation had been taken into account to predict the drug-likeness and toxicity of the synthesized compounds [48].

3. Results and discussions

3.1. Chemistry

As part of our ongoing interest in green chemistry [39,49,50], we demonstrated a novel, highly efficient, economical and scalable synthesis, a for preparing furans with multiple substitutions from simple starting materials such as arylglyoxals and 1,3-diketones. This reaction was carried out as a one-pot reactions employing metal-free microwave-assisted GO-catalyzed as discussed in our previous work. Briefly, this reaction was carried out by reacting arylglyoxal, 1,3-dicarbonyl compounds, and indole was added to produce a 1,4-dicarbonyl intermediate under microwave conditions, which was then cyclized through a Paal-Knorr reaction in the presence of a Brønsted acid under heated conditions. Various types of arylglyoxals, indoles, and 1,3-diketones were reacted under the same conditions, with the results presented in Figure 1. This procedure is also suitable for 5-bromoindole, 2-methylfuran, and thiophenols bearing different substituents, such as methoxy (3c) and chloride (3d) moieties at the para position. Using the optimized condition both phenylglyoxal para-methylphenylglyoxal yielded 75% and 78% of 4b and 4a respectively. After achieving a good yield with electron-rich arylglyoxal, the scope of the protocol was further explored using a substrate containing electron-withdrawing groups, such as chlorine (Cl) and bromine (Br) at the para position of arylglyoxal. Notably, both arylglyoxals (1c and 1d) produced good yields. When ethyl acetoacetate was used in the reactions, it underwent annulations efficiently, producing the desired products 4e and 4f in 66% and 62% yields, respectively. Interestingly, 5-bromoindole (3b) was used in place of indole in the reaction, resulting in the desired product 4 g with a good yield. When thiocresol was used as the nucleophile in the reaction, the corresponding product 4i was obtained with a 79% yield. Next, 1,3-diketones were reacted with 2-methylfuran under microwave irradiation, followed by heating with GO for 3 hours, resulting in the formation of bifuryl derivatives. Initially, p-chlorophenylglyoxal was employed in this reaction, and the desired bifuryl compound 4j was obtained with a yield of 77%. Similarly, p-bromophenylglyoxal was examined, yielding the corresponding product 4k in a satisfactory yield.

3.2. In- silico studies

3.2.1. Molecular docking

Molecular docking study was performed to gain an understanding of binding affinity and intermolecular interactions of the synthesized compounds to the targeted proteins of HCC proteins. The target proteins were chosen based on previous literature knowledge on furan/heteryl clubbed indole compounds. Accordingly, the selected target proteins are integrin protein (PDB: 1JV2), EGFR (PDB:1M17), and IGF-1 R (PDB:5FXS). These proteins are known to play a key role in regulating apoptosis pathways in HCC and other cancers [51–53]. The docking analysis was conducted within RMSD value of 2Å, while the active site was linsitinib binding site, a reference ligand used in this study. The obtained docking results were comparatively evaluated for understanding and relating binding affinity and residue interaction of compounds against target proteins. Analysis of glide score and binding energy of the compounds (Figure 2), revealed that few compounds displayed comparable glide scores and binding energy to linsitinib. For instance, top-ranked pose compound 4b scored −8.421 against 5FXS, −6.901 against 1M17, and −4.776 against 1JV2. Similarly, compound 4c showed the second-highest docked scores of −7.852, −7.021 and −4.834 against 5FXS, 1M17, and 1JV2, respectively. Following, compound 4a displayed a docked score of −8.077, −6.786 and −4.34 against 5FXS, 1M17, and 1JV2, respectively. The comparative binding affinity of compounds indicates the promising binding affinity of compounds to IGF-1 R protein (5FXS) to other selected proteins. The low binding energy (Kcal/mol) recorded during docking analysis for 4a, 4b, 4c, 4 h and 4 g to IGF-1 R are impressive and similar to linsitinib. These results indicate that the synthesized compounds are posed with optimum binding toward IGF-1 R protein inhibition.

Figure 2.

Comparative docking results of compounds on integrin protein (PDB: 1JV2), EGFR (PDB:1M17), and IGF-1 R (PDB:5FXS) (a) Compounds versus docked glide score (-). (b) Compounds versus binding energy (-) of compounds to the active site.

Indeed, active site interaction of compounds within the target is very crucial for their effect on the catalytic property of bound protein. Accordingly, the 2D and 3D view of protein-ligand interaction was generated and analyzed for atom-residue interaction at ligand-bound sites SI Figure 2. Viewing at interactions in integrin (IJV2) pocket, compounds interacted with different amino acid residues via, hydrogen bonding, pi-pi interactions and hydrophobic interactions. The orientation, hydrogen bonding and hydrophobic interaction of top-ranked pose compounds on Chain A of IGF-1 R. The residue interaction for the active compounds is discussed below:

3.2.1.1. Reference ligand (Linsitinib)

The hydroxyl group substitution on cyclobutane established a hydrogen bond with Glu1080, while the hydrogen atom of cyclobutene formed a hydrophobic interaction with Met1142. Additionally, the amino group at the 8th position formed a hydrogen bond with LEU1005. Furthermore, the hydrogen atom within the quinoline moiety interacted with Gln1007 via a hydrophobic interaction.

3.2.1.2. Compound 4a

In the benzene ring of indole, the hydrogen atom formed a hydrophobic interaction with Ile 1160 residue. Additionally, at the 5th position of furan substitution, the hydrophobic benzene moiety interacted with Glu1080 and Met 1082 residues via hydrophobic bonding. Furthermore, the nitrogen group of indole engaged in hydrogen bonding with Asp-1086.

3.2.1.3. Compound 4c

The amine group of indole engaged in hydrogen bonding with Asp-1086. In the benzene ring of indole, the hydrogen atom participated in a hydrophobic interaction with Ile 1160 residue. Subsequently, at the 2nd position of furan substitution (eg -CO-CH₃), the hydrogen of the benzene moiety interacted with Glu 1080 and Met 1082 residues, forming hydrophobic interactions. Moreover, the methyl substitution at the 2nd position of furan facilitated a hydrophobic interaction with Thr1083. Notably, electronegative atom (−NH) of indole ring and substitutions (eg -CO-CH₃) at furan moiety participated in producing stable conformation of compounds for strong interaction, and were essential to increase the affinity of compounds to IGF-1 R.

3.2.1.4. Compound 4i

For example, the docked top posed 4iIGF-1R complex showed different sets of polar side chain residue of ASP1086 and LEU1005 within 4A^o. Similarly, linsitinib showed different H-bond interactions with residues of LEU1005 and GLU1080 within 4A^o. The methyl group in tolyl-thio engaged in hydrophobic interactions with Glu-1080 and VAL1063 residues. Furthermore, the methyl group of methoxy benzene at the 5th position interacted with Thr1083, forming a hydrophobic interaction, while the hydrogen atom in benzene made a hydrophobic interaction with Asp1153. Additionally, the methyl group at the 2nd position facilitated a hydrophobic interaction with Gln1007.

3.2.1.5. Compound 4f

The nitrogen atom within the indole ring formed a hydrogen bond with Asp1086. Additionally, the methyl substitution at the 2nd position of furan facilitated a hydrophobic interaction with Thr1083. Moreover, the hydrogen atom of the benzene substitution at the 2nd position of furan interacted with Glu 1080 and Met 1082, resulting in a hydrophobic interaction.

For the IGF-1 R (5FXS) protein, the top-ranked pose compounds 4a, 4b and 4f showed H-bond with Asp1086 residues and hydrophobic interaction with Glu1080 and Met1082. It was noted that the reference ligand posed two H-bonds with LEU1005 and GLU1080 residues. The binding mode and residue interaction of compounds 4a, 4b, 4c and 4 h mimics the linsitinib binding mode. Further, the spotted interacting residues at the binding site are directly responsible for the catalytic action of IGF-1 R. Overall, the lowest binding affinity, low binding energy and favorable interactions via H-bonds and pi-pi interaction between the residues of the entrance of the pocket and backbone suggest the inhibition of IGF-1 R.

This study concurrence the interaction of one of the top-ranked posed compounds 4i. Another study, target interaction at residues VAL1013, LEU1005, VAL1013, MET1156 MET1156, LYS1033,THR1083, GLY1085, ARG1084, MET1082, LEU1005, and SER1089 with a binding affinity of −7.1–7.9 (Ref. ligand: belzutifan: LEU1005, VAL1013, ASP1086, MET1142, MET1156, and ASP1086; binding affinity 7.3) showed inhibition [54]. These diverse molecular interactions underscore the complex interplay between hydrogen bonding and hydrophobic forces in molecular recognition and biological function, highlighting the intricacies of molecular interactions in biological systems.

Overall, while comparing the atom-residue interaction pattern of the selected proteins, there is a substantial difference in the binding mode between IGF-1 R and the other two proteins. Further, it was ruled out that a few compounds, 4a, 4b, 4c and 4i showed similar posing and binding modes to reference ligands. These findings encouraged us to conduct a cytotoxicity evaluation of these compounds against cancer cells. Since compound 4a offered the highest docking score and appropriate binding energy among all synthesized compounds. Thus, it was selected for subsequent molecular dynamic simulation.

3.2.2. Molecular dynamics simulation

The complex of the top hit molecule 4a was utilized for the MD simulations to understand the affinity of the ligand to protein IGF-1 R (5FXS). MD simulation is to understand the protein-ligand complex behavior at a given time, in this study; the protein-ligand complex was subjected to 100 ns in the assembly phase for the ligand complexes. The MD simulation of RMSD fluctuations was analyzed separately for the protein and ligand, with a limit of 3.5°A. Hence the complex contemplated being stable and the RMSD plot is shown in SI Figure 3. It reveals the ligand-bound proteins are significantly stable throughout the simulations, with minor instability fluctuations appearing in 90 to 100 ns during MD simulation. Ligand properties RMSD are the evaluation of radius of Gyration (rGyr) maximum extendedness reached between 4.1°A to 4.3 °A. Molecular surface area (MoISA) calculation in the range of 318 °A −327 °A². The Solvent Accessible Surface Area (SASA) represents the ligands attainable by water molecules from 60–150°A². Polar Surface Area (PSA) ranges from 60 °A²- 72 °A².

Figure 3.

Screening results of in vitro cytotoxicity of compounds against HepG2 (Liver cancer) and MCF-7 (Breast cancer) cell lines.

RMSF plotted the flexibility of each residue and secondary structure protein in the ligand-protein complexes. The RMSF deviation was not more than 2.5 Å throughout the 100 ns. Results have proven significant hydrogen bond, hydrophobic, ionic and water bridge inter linkage between the ligand and the residues during the simulations. That H-bond was observed with Arg1003, Met1082, Asp1086 respectively. The hydrophobic interaction was observed in Leu1005, Val1013, Ala1031, Val1063, Met1079, Met1082, Met1142, Met1156, Ile 1160. An ionic bond appeared at Asp1086, along with this the ligand made 9 contacts in the entire 100 ns. The most frequent contact and regular communications were made by Arg1003, Glu1015, Arg1084, and Met1142 amino acids. The predominant hydrogen bonding was exhibited by Arg1139 with 42% and followed by Asp1086 with 36%.

3.2.3. Pharmacophore model

Pharmacophore modeling represents an effective component of computer-assisted drug design for establishing essential structural motifs of a target protein which is responsible for ligand binding affinity and pharmacological activity [55]. In this modeling process, there are performance metrics such as average rank of actives, enrichment factor (EF), area under the receiver operating characteristic curve (ROC), Boltzmann-enhanced discrimination of receiver operating characteristic (BEDROC), area under the accumulation curve (AUAC), and robust initial enhancement (RIE). These metrics collectively determine the overall robustness of the pharmacophore model hypothesis [56]. Influential groups prefer using the area under the ROC curve to assess virtual screening performance, primarily because of its advantageous statistical characteristics. A value of 1/2 indicates that the ranking method is under-performing than random selection. This metric can be construed as the probability of an active compound being ranked before an inactive one. Its range lies between 0 (representing the poorest attainable performance) and 1 (indicating optimal performance), emphasizing the importance of minimizing its value [56]. EF is quantified as the ratio between the probabilities of discovering an active compound within the top X% of the dataset [57]. A maximal enrichment score of 100 is achieved when all active compounds (A = 1) are identified within the top 1% of the decoys (D = 0.01), underscoring the imperative to minimize this value for scientific rigor. BEDROC is a modified version of ROC AUC that prioritizes early retrieval of true positives, typically applied in scenarios where prompt recognition of relevant instances is critical.

In accordance with the above, the pharmacophore validation metrics for our compounds revealed the metrics with EF1% score of 83.37%, BEDROC of 0.77, ROC value of 1.0, and AUAC of 0.99. These findings demonstrate the accuracy and reliability of the developed pharmacophore model. The consensus pharmacophore model outlines the essential features, crucial for ligand-receptor interactions. Notably, it includes a hydrophobic region corresponding to the methyl group (-CH₃) of the benzene substituent positioned at the second position of furan, with a matching tolerance of 2Å. Additionally, a hydrogen donor feature associated with the nitrogen atom (−NH) within the indole moiety is represented, also with a matching tolerance of 2Å. The presence of three aromatic rings – indole, furan, and substituted benzene (SI Figure 4) – further underscores their pivotal role in ligand binding within the developed pharmacophore, notably highlighting the importance of the hydrophobic and hydrogen bonding characteristics. The developed consensus pharmacophore model reveals critical hydrophobic and hydrogen donor regions, underscoring their pivotal role in ligand-receptor interactions. The identified pharmacophore positions were found to be directly associated with ligand activity, as evidenced by the similar group interactions observed in molecular docking investigations. This observation is further corroborated by molecular docking studies conducted utilizing the crystal structure of PDB 5FXS bound to ligand 4a. The congruence between the pharmacophore model predictions and the spatial arrangements observed in the receptor-ligand complex lends credibility to the identified pharmacophore features. Furthermore, analysis of compound 4a‘s interaction with the IGF-1 R receptor reveals notable hydrophobic and hydrogen bond interactions, further emphasizing the indispensability of hydrophobic and acceptor features for effective binding and inhibition of IGF-1 R. This convergence between pharmacophore-based predictions and docking results further underscores the significance of the identified pharmacophore features in guiding the design of potent and selective IGF-1 R inhibitors.

Figure 4.

(a) Flow cytometry analysis of apoptotic population of HepG2 cells treated with compounds 4a, 4b and 4c by FITC-Annexin V (AV) and Propidium Iodide (PI) staining. After 12 h treatment 10,000 cells from each sample of treated and non-treated (control) were analyzed. The percentage of cell apoptosis was reported inside the quadrants. Cells which fall on the Q1 (AV-/PI+): necrotic cells; Q2(AV+/PI+): late apoptotic cells, Q3 (AV-/PI-): Viable cells; Q4 (AV+/PI-): early apoptotic cells. (b) The percentage of the cell population displayed early apoptosis after treatment with test compounds in compared to control (untreated) samples.

3.3. In vitro cytotoxicity study

Compounds 4a-k are screened for their cytotoxicity against two HepG2 and MCF-7 cell lines (Table SI). All screened compounds showed different degree of cytotoxicity against both cell lines. Whereby compounds 4a, 4b and 4c showed prominent cytotoxicity against HepG2 cell line with IC₅₀ = 27.7 µM, 28.35 µM, and 33.83 µM respectively. Compounds 4d and 4 g showed medium cytotoxicity against HepG-2 with IC ₅₀ = 47.02 and 49.41 µM respectively. Similarly, compounds 4a, 4b, 4c, 4d, 4g, 4 h, 4 l also showed medium cytotoxicity against MCF-7 cells with IC ₅₀ = 40.78, 41.18, 39.64, 45.85, 34.73, 37.54, 44.01 µM. Compounds 4j, 4k, 4 h, 4 l, 4e showed low cytotoxicity against HepG2 with IC ₅₀ = 83.77, 88.41, 67.93, 60.78 µM. Similarly, in the case of MCF-7 also both compounds 4j and 4k showed low cytotoxicity with IC ₅₀ = 90.53 and 65.5 µM respectively. Compounds 4i, 4f and 4 m showed very low cytotoxicity against both tested cell lines with IC₅₀ of more than 100 µM (Figure 3). Above result showed that compounds containing indole substituent at 3-position of furan showed prominent anti-proliferative activity against both cell lines compared to furan, thiophenyl substituents. Moreover, 2,3’-bisfuran and 3-thiophenylfuran derivatives completely showed lost of activity. Insight into structure-activity relationship (SAR), furan (indole) at 3^rd position of indole (furan) enhanced the cytotoxic effect, further the compound potency does not significantly affect by nature of substituents at 5^th of Indole and 4^th position of furan. It was noted that the more active compounds found against HepG2 were the highest scored compounds in silico docking against the protein IGF-1 R as per discussion in the previous section. Out of which compounds 4a, 4b and 4c showed potent cytotoxic activity and were selected for further investigation.

Overlap analysis of docking result on (5FXS) versus in vitro cytotoxicity of compounds on HepG2 and MCF-7 cells (SI Figure 5) suggests that our compounds showed better affinity toward IGF-1 R than EGFR, where a regression coefficient of 0.486 was found for pIC₅₀ vs glide score against IGF-1 R. Although, indole derivative can be dual acting on both EGF1R and IGF-1 R, the structural feature with substitution at 6 or 7^th position and fused system of indole at 2, 3 position of indole would favor dual selectivity [58]. A recent study quoted that substitution at 3^rd position indole plays a crucial role in IGF-1 R and nature of substitution determines the potency [59]. Accordingly, we found the synthesized compounds are relatively moderately potent as compared to those compounds. This may be due to either compensatory role of EGF1R as an alternative to IGF-1 R or a less flexible structure.

Figure 5.

(a) HCC (KEGG PATHWAY: Hepatocellular carcinoma – Homo sapiens (human)). Represented dysregulated protein involves in our proposed pathway on which compounds 4a, 4b and 4c act. (b) Depicted integrin signaling as a compensatory cell survival mechanism of IGF Signalling.

3.4. Flow Cytometry analysis

Flow cytometry analysis was done to evaluate the mode of cell death after treatment with compounds 4a, 4b and 4c at a concentration of their respective IC₅₀ value for 12 h. The HepG2 cells treated with 4a showed 13.3% early apoptosis and 3.7% late apoptosis. On 4b treatment, 11.5% of cells displayed early apoptosis and 3.9% late apoptosis. Similarly, 4c demonstrated 14.5% early apoptosis and 3.9% late apoptosis (Figure 4). Approximately twofold to threefold increase in apoptosis was observed in the cells upon treatment with test compounds as compared to the control cells. Thus, this finding suggests that all three compounds induce HepG2 cell death through early apoptosis. Comparatively, compound 4a and 4c treated HepG2 cells showed a higher percentage of cell death which further supports our docking result and MTT assay result as explained above.

3.5. NanoLC-Q-Orbitrap MS analysis of protein expression

In continuation of the MTT assay and FACS analyses, untargeted label-free quantification of the proteome profile of HepG2 cells treated with compounds was carried out using mass spectrometry. In this procedure, proteins were isolated using cold acetone precipitation and subjected to in-solution trypsin digestion as per our in-house protocol [45,60]. The samples were analyzed using NanoLC-Q-Orbitrap mass spectrometer with a mass resolution of 60,000, in triplicate (n = 3). The raw data was processed with Proteome Discoverer 2.5.0 for protein identification and relative quantification. A unique peptide match of more than two was allowed for inclusion in proteome expression data. For protein expression profiling, a p-value of less than or equal to 0.05 and a Log2 fold change of ≥ 0.5 (up-regulation) or ≤ −0.5 (down-regulation) were considered in pathway analysis.

3.5.1. Analysis of dysregulated proteins

Based on selection criteria we found 3381 dysregulated proteins which are mapped in venn diagram SI figure 6A. Out of which 3140 proteins were commonly expressed among all compound-treated groups (4a, 4b, and 4c). Specifically, 100, 94, and 50 proteins were uniquely expressed in the 4a, 4c, and 4b treated samples, respectively. 185 proteins were commonly expressed in both 4a and 4c treated samples, 233 proteins were shared between 4b and 4c treated samples, and 79 proteins were common between 4a and 4b treated samples. This dysregulation implies that the compounds significantly interfere with the proteome and normal cellular functions.

Further, univariate analysis of protein expressions was performed using volcano plot, provided the information related to fold change in terms of relative quantification of a specific dysregulated protein between control versus treated groups. Accordingly, for 4a treated sample, 828 proteins were down-regulated with log2 fold change ≤ −0.5 and 520 proteins were up-regulated with log2 fold change ≥ 0.5 (SI figure 6B). Similarly, 4b treated sample (SI figure 6C), revealed 702 down-regulated and 495 up-regulated proteins. For the 4c treated sample (SI figure 6D). 540 proteins were down-regulated, and 383 proteins were up-regulated.

Overall, we noted a significant difference in the dysregulated proteins profile, and this dysregulation is certainly driven by the compounds and their possible interaction with biochemical pathways involved in normal cellular function. For instance, 4a significantly shifted the protein expression toward down-regulation than up-regulation, where down-regulation is about 4-fold higher than up-regulations. It indicated that 4a exhibits significant interactome on protein dysregulation in HepG2 cells. Hence, this is further analyzed using other bioinformatics tools.

3.5.2. Gene-Ontology (GO) analysis of the dysregulated proteins

All dysregulated proteins were mapped to their enriched GO and categorized into three classes i.e., biological process, molecular function, and cellular component [61]. GO enrichment analysis of the 4a treated group revealed that a total of 901 dysregulated proteins were involved in biological process, with 67, 11 and 17 proteins involved in cell proliferation, cell-cell signaling, and cell adhesion. Notably, 180 proteins were involved in cell organization and biogenesis, indicating that the compounds exhibit cell-damaging potential (SI figure 7). For proteins involved in the cellular component category, we found, a total of 136 proteins associated with the plasma membrane, 197 proteins with the cytosol, and 313 proteins with other membranes (SI figure 8). In terms of molecular function, 131 dysregulated proteins were found to be responsible for nucleic acid binding activity, 24 proteins for kinase activity, 40 proteins for enzyme regulator activity, 38 proteins involved in transport activity, 26 proteins were involved in signal transduction activity, and 543 proteins were involved in other molecular functions (SI figure 9).

For compound 4b, 686 dysregulated proteins are involved in biological processes. Among these, 46 proteins were involved in cell proliferation, 7 proteins in cell-cell signaling, 15 proteins in cell adhesion, 141 proteins in cell organization and biogenesis, 138 proteins in protein metabolism, 124 proteins in transport proteins, and 75 proteins related to stress response (SI figure 10). For cellular component, 188 dysregulated proteins were distributed in the nucleus, whereas 100 proteins in the plasma membrane, 143 proteins in the cytosol, and 258 proteins in other membranes (SI figure 11). In terms of molecular function, 108, 17, 38, 39, and 29 proteins were respectively involved in nucleic acid binding activity, kinase activity, enzyme regulator activity, transport activity, and signal transduction activity or receptor binding (SI figure 12).

Similarly, GO enrichment analysis for 4c revealed that 670 dysregulated proteins of biological process, 43 proteins were involved in cell cycle or cell proliferation while only 7 proteins traced for cell-cell signaling. Apart, 22 proteins for cell adhesion, 149 proteins for cell organization and biogenesis, 123 for protein metabolism, 129 for transport proteins, and 85 for stress proteins were noted (SI figure 13). For cellular components, 170, 116 and 141 proteins were distributed in the nucleus, plasma membrane, and cytosol (SI figure 14). In terms of molecular function, 97 dysregulated proteins were involved in nucleic acid binding activity, whereas 19 for kinase activity, 38 for enzyme regulator activity, 37 for transport activity, and 35 for signal transduction activity (SI figure 15).

In summary, GO enrichment analysis reveals that our active compound acts through multiple mechanisms that align with anticancer activity, particularly by inhibiting cell proliferation or Cell cycle signaling, disrupting cellular communication, inducing stress, and targeting key molecular functions essential for cancer cell survival and progression.

3.5.3. Pathway analysis and protein-protein interaction (PPI)

The biological significance of the enriched pathways for the annotated genes involved in HCC was identified through KEGG pathways. Pathway enrichment analysis of the dysregulated protein of 4a treated HepG2 cells using STRING revealed 20 enriched pathways (Table SII). Cell cycle signaling pathway is a cancer-related pathway which is indirectly linked to our target proteins (IGF-1 R) [62,63]. Furthermore, pathway analysis by proteome discoverer revealed some prominent proteins listed in Table 1 that are directly and indirectly involved in the proposed IGFR/Ras/Raf/MEK/ERK pathway, also called as MAPK pathway. This is one of the well-studied HCC pathways [64]. The interaction of these listed proteins was analyzed by STRING, revealing 17 nodes and 24 edges, with an average node degree of 2.82 and a local coefficient of 0.553 (SI figure 16B). A significant protein-protein interaction (PPI) network was observed with a p-value of less than 6.57 × 10⁻⁷, indicating a notably higher number of interactions among listed proteins.

Table 1.

List of dysregulated proteins found to be involved in the proposed IGFR/Ras/Raf/mek/erk (MAPK) pathway.

			Log2 Fold Change
Sl.No.	Accession	Description	4a	4b	4c
1.	P29353	SHC-transforming protein 1(SHC1)	−6.64	−1.5	0.06
2.	A0A024R3U9	Ras-related protein Rab-4(RAB-4)	−6.64	−0.03	−0.18
3.	P61106	Ras-related protein Rab-14(RAB-14)	−0.19	−0.33	−0.27
4.	P61006	Ras-related protein Rab-8A(RAB-8A)	−1.94	−0.29	−0.63
5.	Q53EZ9	Mitogen-activated protein kinase kinase 3 isoform B variant (Fragment) (MAPK3B)	−1.64	−0.67	−0.11
6.	P45985	Dual specificity mitogen-activated protein kinase kinase 4 (MAPK4)	−0.92	0.1	0.1
7.	E9PQW4	Mitogen-activated protein kinase (MAPK)	−0.25	−0.27	−1.87
8.	A0A286YF97	Stress-activated protein kinase JNK (JNK)	−6.64	0.13	−0.36
9.	A0A6Q8PGN9	Phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K)	6.64	–	–
10.	A0A8V8TRG9	Serine/threonine-protein kinase Mtor	−0.77	0.18	−0.58
11.	Q6PEY6	Insulin-like growth factor-binding protein 1(IGFBP1)	3.3	3.7	3.13
12.	E7EUI6	Integrin beta(ITGB)	0.49	0.34	0.33
13.	P56199	Integrin alpha-1(ITGA)	0.93	1.06	1.07
14.	P17301	Integrin alpha-2(ITGB)	1.14	0.93	0.8
15.	P24941	Cyclin-dependent kinase 2 (CDK-2)	−6.64	0.24	−0.4
16.	Q9UJX4	Anaphase-promoting complex subunit 5(ANAPC5)	−6.64	0.31	−0.57
17.	A0A024R411	Origin recognition complex subunit 2(ORC2)	−1.03	−0.4	−0.01
18.	P06493	Cyclin-dependent kinase 1(CDK-1)	−0.65	−0.43	−0.4
19.	Q05655	Protein kinase C delta type (PKCδ)	−6.64	−0.06	0.04

Our proteomics workflow for the identification of key proteins which are directly or indirectly related to IGF-1 R pathways revealed a list of 19 significant proteins (Table 1). Among which, proteins namely, SHC1), RAB-4, RAB-14, RAB-8A, MAPK3B, MAP2K4, JNK, CDK2, CDK1, ANAPC5, ORC2, PKCδ are down-regulated and PI3K, IGFBP1, ITGB, ITGA1, ITGA2, were found to be up-regulated.

Upon IGFs binding, a conformational change in the IGF-1 R receptor occurs which leads to the activation of receptor tyrosine kinase activity, which in turn phosphorylates several substrates including insulin receptor substrates (IRSs) and Src homology collagen (SHC) [65]. These phosphorylated substrates then activate downstream signaling pathways such as the PI3K/Akt/mTOR and Ras/Raf/MEK/ERK (MAPK or JNK/P38 MAPK) pathway which are essential for inducing various bioactivities of IGFs, including cell differentiation, proliferation, and survival [66,67]. Giorgetti et.al reported that the expression of exogenous SHC proteins increases basal MEK (MAPK/ERK kinase) activity, indicating involvement of SHC proteins in the MAP kinase pathway. However, SHC proteins do not appear to affect the activity of phosphatidylinositol 3-kinase (PtdIns-3-kinase), which is another important signaling pathway of (IGF) signaling pathway [66]. Furthermore, Calmont et, al. reports the induction of hepatic gene expression by FGF via specific activation of the MAPK pathway and is independent of PI3K signaling [68]. This furthermore, supports that our molecule inhibits the IGF-1 R signaling via MAPK signaling pathway.

Ras protein is a key mediator of growth factor-dependent cell survival. It activates protein kinase Raf, MEK (MAPK Kinase) and was reported to promote cell survival and inhibit apoptosis [69]. The MAPK family acts as a central integration point for different biochemical signals and is essential for various cellular processes. They consist of a three-tiered signaling module made up of MAPKKKs, MAPKKs, and MAPKs [70]. Llovet et al. reported that the oncogenic signals from the MAPK/ERK pathway account about 50% of early-stage HCC patients and are found in most patients with advanced HCC [71]. Furthermore, Numerous studies have shown that MAPK/ERK signaling plays a decisive and central role in the development of HCC [64].

JNK is a subfamily of MAPKs that is significantly involved in the development of several types of cancer, including HCC [72]. Furthermore, JNK also participate in extracellular signal and stress-activated cell-cycle checkpoint control. JNK participates in the regulation of G1/S and S-phase checkpoints [73]. Our data also showed down-regulation of proteins involving cell cycle signaling pathways such as Anaphase-promoting complex subunit 5 (−6.64 fold), Cyclin-dependent kinase 2 (−6.64 fold), Origin recognition complex subunit 2 (−1.03 fold), Cyclin-dependent kinase 1 (−0.57 fold). Moreover, Involvement of CDKs in regulation of cell cycle in cancer have been well studied and overactivity of different CDKs leads to cell cycle progression and checkpoint dysregulation which is responsible for the development of tumors [74]. It has been reported that overexpression of origin recognition complex (ORC6) is poorer prognostic outcomes in many cancers. Furthermore, ORC6 plays a role in the cell cycle pathway, DNA replication, and mismatch repair pathways across various tumor types [75].

In addition, we also observed up-regulation of IGFBP1 (3.3 fold), IGα1 (0.93 fold), IGα2 (1.14 fold), IGβ (0.49 fold). Integrin acts as a compensatory mechanism of the cell for survival in response to the inhibition of IGFR signaling. Zhang and Yee showed that IGFBP-1 could have dual effects on cancer cell motility by disrupting both IGRF and integrin receptor systems [76]. Non-phosphorylated IGFBP-1 binds to α5/β1 integrin receptor and potentiates the IGF signaling pathway via IGF-independent effect [77]. Studies have shown that IGFBP-1 can still influence cellular processes even in the lack of IGFs, indicating that IGFBP-1 can directly activate other cell-surface receptors. In particular, the RGD motif found in the C-terminal domain of IGFBP-1 binds to α5/β1 integrin, enabling IGFBP-1 to improve adhesion or cell migration [77–80]. This shows activation of the Integrin pathway by binding of IGFBP-1 which acts as a compensatory mechanism of the cell for survival in response to the inhibition of IGFR signaling (Figure 5).

In summary, the proteomics study showed very pronounced downregulation of the downstream signaling molecules of IGF-1 R receptor after treatment with compound 4a, thereby indicating its possible involvement in the modulation of the IGF-1 R signaling pathway. This finding correlated with the molecular docking studies, in which compound 4a showed a higher docking score and binding affinity toward IGF-1 R, suggesting a strong interaction. The molecular dynamics (MD) simulations also demonstrated that compound 4a continued to be stable during a trajectory period of 100 ns, supporting the idea of slightly complex behavior as a stable IGF-1 R binder. With this converging evidence that compound 4a downregulates IGF-1 R downstream molecules, the proteomics-docking-MD simulation shows possible evidence that compound 4a has a direct targeting on IGF-1 R and inhibits cell differentiation, proliferation, and survival that highlights its potential anticancer activity.

3.6. In silico ADMET

The ability of an antagonist to inhibit the receptors and enzymes is not sufficient to consider it a Potential drug [81]. But also, the ADMET properties, including drug-likeness analysis are also important in the field of drug discovery and decide whether these inhibitors are suitable to be administered to the human body. In addition, poor ADME properties are one of the major factors which lead to the failure of drugs in clinical trials [82]. The pharmacokinetics properties of all compounds were predicted using swissADME. Where by all compounds studied were shown to obey Lipinski’s rule of 5 as shown in Table 2. Their molecular weights are < 500 Dalton, the number of hydrogen bond donors are < 5 and the number of hydrogen bond acceptors < 10. The lipophilicity of the compounds was represented as log p _(o/w), the partition coefficient of the compound between the octanol and water system. All screened compounds showed a reasonable lipophilicity as per Lipinski’s rule of five, i.e., logP less or equal to 5, except compounds 4 g, 4f, 4 m with log P of 5.52, 5.18 and 5.08. Therefore all the screened compounds showed high gastrointestinal absorption except compound 4 g. Cytochrome P450 enzymes family are of enzymes responsible for the drug metabolism which significantly reduce drug bioavailability [83]. All screened compounds showed to inhibit CYP1A2 enzyme which means all these compounds are considered to have good bioavailability with a bioavailability Score of 0.55. The compound 1–13 toxicity risk such as eye irritation, and skin sensitization was predicted as shown in Table 2 using Pro Tox – II and admet SAR. Compounds 4a, 4b and 4c show none of the mentioned toxicities. Compound 4i showed mild Eye Irritancy and compound 4 m showed mild skin sensitization. Since, all compounds exhibited cytotoxicity, eye irritation, skin sensitization. This finding would be useful for chemist for handling of these compounds.

Table 2.

Predicted ADMET properties and drug likeness score of the synthesized compounds.

		Lipinski’s rule of five						CYP	BAS			LD50
Sl.No	Cpd.	MW	HBA	HBD	Log p	GIA	P-gp	1A2(-)	(0–1)	EI	SS	mg/kg
1	4a	329.39	2	1	4.64	H	YES	YES	0.55	–	–	842
2	4b	315.37	2	1	4.28	H	YES	YES	0.55	–	–	842
3	4c	349.81	2	1	4.84	H	YES	YES	0.55	–	–	1230
4	4d	394.26	2	1	4.93	H	YES	YES	0.55	–	–	870
5	4e	359.42	3	1	4.91	H	YES	YES	0.55	–	–	1190
6	4f	424.29	3	1	5.18	H	YES	YES	0.55	–	–	210
7	4 g	473.16	2	1	5.52	L	YES	YES	0.55	–	–	870
8	4 h	345.39	3	1	4.28	H	YES	YES	0.55	–	–	138
9	4i	368.45	4	0	4.48	H	NO	YES	0.55	+	–	115
10	4j	314.76	3	0	4.39	H	NO	YES	0.55	–	–	1230
11	4k	359.21	3	0	4.45	H	NO	YES	0.55	–	–	592
12	4 l	253.3	2	1	3.3	H	YES	YES	0.55	–	–	650
13	4 m	342.84	2	0	5.08	H	NO	YES	0.55	–	+	1655

HBA: No of -bond acceptor, HBD: No of H-bond donor, GIA: GI absorption, BAS: Bioavailability Score, Cytochrome P450 1A2, H- high, L: low, I: inhibitor; P-gp: P-glycoprotein substrate; SS: Skin sensitization; EI: Eye irritation.

4. Conclusion

In conclusion, the high expression of IGF-1 R in HCC, promotes cell proliferation, migration, and anti-apoptosis. Therefore, inhibition of IGF-1 R is a promising strategy to treat HCC. In this study we design to investigate the anti-proliferative activity and underlying mechanism of new 3-(2-furoyl)-indole derivatives against HCC. The data from in silico docking study and in vitro MTT assay indicated compound 4a showed a promising inhibitory activity against IGF-1 R, showing docking and glide score of −8.077 binding energy of −67.771 Kcal/mol against IGF-1 R and IC₅₀ 27.7 µM against HepG2 cell lines. Flow cytometry assay showed that after 12 hr of compound 4a treatment HepG2 cells undergo apoptosis significantly as compared to control, which indicates that 4a is a potential anti-HCC candidate. Furthermore, Mass Spectrometric base Proteomic study, revealed that compound 4a inhibits HCC through IGF-1 R/MAPK pathway. Comprehensively, ADMET properties prediction showed reasonable lipophilicity and minimal toxicity. Thus, predicting the overall potential of the compounds as a drug candidate. Further experimental validation is needed to confirm these findings and assess the efficacy and safety of the compounds in vivo.

Funding Statement

The authors are grateful to the generous financial support from the National Institute of Pharmaceutical Education and Research (NIPER)−Hajipur, Department of Pharmaceuticals, Ministry of Chemicals and Fertilizers, Govt. of India. P.R. acknowledge generous financial support from the Indian Council of Medical Research, Govt. of India for financial support [ICMR grant no. IIRP-2023–5331/F1]. A.J. and A.S. acknowledge generous financial support from the Science and Engineering Research Board, Govt. of India for financial support [SERB Start-up Research Grant No. SRG/2023/001403 & SRG/2022/002154]. The authors are grateful to the generous financial support from the National Institute of Pharmaceutical Education and Research (NIPER)−Hajipur, Department of Pharmaceuticals, Ministry of Chemicals and Fertilizers, Govt. of India.

Article highlights

• New 3-(2-furoyl)-indole derivatives showed better cytotoxicity towards HepG2 cells than MCF-7 cells.

• Cytotoxicity and molecular docking studies revealed that compound 4a showed better activity with IC50 of 27.7 μM, a docking score of -8.077, and a glide energy of -67.8 kcal/mol.

• FACS study revealed that the active compound causes cell death via inducing early apoptosis.

• Gene ontology enrichment analysis reveals that our active compound acts through multiple mechanisms that align with anticancer activity, particularly by inhibiting cell proliferation, disrupting cellular communication, inducing stress, and targeting key molecular functions essential for cancer cell survival and progression.

• Mass spectrometry-based label-free quantitative proteome profiling revealed IGF-1R inhibitory potential of 4a.

• The anti-proliferative mechanism is in good agreement with in silico, FACS and proteome profiling.

Declaration of Interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

Efficiency Myrsing: Investigation, experimental studies. H. M. Chandra Mouli: In Silico Study. Computational Study. Pallaprolu Nikhil: Investigation and reviewing Mass Spectrometry-Based Proteomics study. Abhishek Sahu: In Vitro Study, writing original draft preparation, reviewing, editing. Deepali: In Vitro Study. Anupam Jana: Conceptualization, supervision, writing original draft preparation, reviewing, editing. P Ramalingam: Conceptualization, supervision, writing original draft preparation, reviewing, editing. A.S., A.J and P.R. prepared the manuscript and ESI with input from, E.M. and H.M.C.M.

Data availability statement

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [1] partner repository with the dataset identifier P×D054680and 10.6019/PXD054680

Supplemental material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/17568919.2025.2467616