https://orcid.org/0000-0003-3401-4315
Figures
Abstract
Fructosamine-3-kinase (FN3K) is involved in the deglycation of Nrf2, a significant regulator of oxidative stress in cancer cells. However, the intricate functional aspects of FN3K and Nrf2 in breast cancers have not been explored vividly. The objectives of this study are to design the human FN3K protein using homology modeling followed by the screening of several anticancer molecules and examining their efficacy to modulate FN3K activity, Nrf2-mediated antioxidant signalling. Methods pertinent to homology modeling, virtual screening, molecular docking, molecular dynamics simulations, assessment of ADME properties, cytotoxicity assays for anticancer molecules of natural/synthetic origin in breast cancer cells (BT-474, T-47D), and Western blotting were used in this study. The screened anticancer molecules including kinase inhibitors of natural and synthetic origin interacted with the 3-dimensional structure of the catalytic domain in human FN3K protein designed through homology modeling by significant CDOCKER interaction energies. Subsequently, gefitinib, sorafenib, neratinib, tamoxifen citrate, and cyclosporine A enhanced the expression of FN3K in BT-474 cell lines with simultaneous alteration in Nrf2-driven antioxidant signalling. Oxaliplatin significantly downregulated FN3K expression and modulated Nrf2-driven antioxidant signalling when compared to cisplatin and other anticancer drugs. Hence, the study concluded the potential implications of existing anticancer drugs to modulate FN3K activity in breast cancers.
Citation: Beeraka NM, Zhang J, Mandal S, Vikram P. R. H, Liu J, B. M. N, et al. (2023) Screening fructosamine-3-kinase (FN3K) inhibitors, a deglycating enzyme of oncogenic Nrf2: Human FN3K homology modelling, docking and molecular dynamics simulations. PLoS ONE 18(11): e0283705. https://doi.org/10.1371/journal.pone.0283705
Editor: Arabinda Ghosh, Gauhati University, INDIA
Received: September 20, 2022; Accepted: March 14, 2023; Published: November 1, 2023
Copyright: © 2023 Beeraka et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This study was supported by JSS College of Pharmacy, Mysuru, India. This study was also supported by National Natural Science Foundation of China (No. 81700729). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Obesity, inheritance, lifestyle patterns, and pollution can cause the risk of breast cancer incidence as age increases in women [1]. In addition, severe tobacco usage, and postmenopausal hormone therapy can be other risk factors to induce breast cancer. Mortality depending on age-standardized breast cancer incidence is enhanced in recent years [1]. However, the lack of effective timely prognostic strategies leads to late detection or diagnosis and poor clinical outcomes due to several tumor-wise, node-wise, and metastasis-wise invasions. Additionally, the stemness-inducing factors across the tumor microenvironment are another hindrance to the overall survival in breast cancer patients [2]. Therefore, efficient therapeutic strategies are required to inhibit breast tumor growth and metastasis. Since existing therapies such as surgery and chemoradiotherapy are significant therapeutic modalities but the recurrence rate of breast cancer is still higher. Hence, novel combination therapies that specifically deliver drugs with sustained release properties are urgently needed to treat breast cancers. Current therapies for treating metastatic advanced stages of breast cancers are either less effective or target only a set of population [3]. In addition, the development of drug-resistant phenotypes worsens this situation as the majority of existing treatments are found to be ineffective against these cancers. Albeit several chemotherapies have been preferred to target breast cancer, the cancer cells in the metastatic stage have been acquiring chemoresistance due to stemness. Henceforth typically a higher requirement of chemotherapy dosage regimen is preferred to the breast cancer patients which recurrently induce dose-related toxicity and early mortality.
Activation of Nrf2 is higher in the dividing cancer cells due to mutations in the majority of tumors such as liver, lung, head, and neck [4], bladder, and pancreas [4]. The Nrf2 could be considered the crucial oncogene involved in driving oncogenesis [5]. Previous studies delineated that the less-frequent somatic mutations, which could affect Nrf2 glycation sites including R499W, R569C, and R569H in major cancers such as endometrial, colorectal carcinoma, and melanomas. FN3K is reported to be involved in deglycating the Nrf2 through phosphorylation of lysine, and arginine and enhancing Nrf2 activity in hepatocellular carcinoma, consequently protecting the cancer cells through antioxidant mechanisms [4]. Therefore, the development of small molecule inhibitors is a significant approach to selectively block FN3K in cancers. Previous study described the profound efficacy of Nrf2 inhibitors for fostering the retardation of breast cancer [6]. FN3K is a deglycating enzyme reported to induce deglycation of Nrf2 which results in the activation of antioxidant responses to regulate oxidative stress to offer cancer cell protection [4]. Deglycated Nrf2 predominantly induces antioxidant responses by binding to sMAF. It may be a highly significant targeting molecule to promote the development of therapeutic molecules of either synthetic or natural origin to control the activity of FN3K, subsequently to confer Nrf2 in the glycated state in cancer cells including breast cancers.
Rational drug design through virtual screening, molecular docking, and molecular dynamics could enhance the rapid screening process of existing anticancer molecules to modulate the several oncogenic proteins involved in oncogenesis [4, 7]. Thus, it is possible to develop more efficient drug candidates. For instance, the kinases can have a prominent function in modulating the oncogenesis in many cancers and several kinase modulators could be implicated as antineoplastic agents for the retardation of cancers [8]. Furthermore, the phytochemicals like brusatol, topotecan, and platinum derivatives such as cisplatin, and oxaliplatin could be used as combinatorial regimens along with kinase modulators, which require substantial studies. In this study, we have designed the human FN3K three-dimensional protein through homology modeling and subsequently screened several anticancer kinase inhibitors of synthetic and natural origin and analyzed the modulating efficacy of platinum derivatives, and other anticancer drugs onto FN3K protein expression patterns using in vitro models. The overall workflow of the study was depicted in the following diagram.
Methods
Ethical statement
This study does not involve animals or human subjects. Therefore, ethical permission is not required.
Chemicals and reagents
DMEM with high glucose (4.5 g/L) (cat#: AL111), trypsin-EDTA (0.25%) (cat#: T001) were procured from HiMedia Laboratories Pvt. Ltd. (Bengaluru, Karnataka, India). DMSO pertinent to cell culture (cat#: D2650), radioimmunoprecipitation assay buffer (cat#: R0278), protease inhibitor cocktail (cat#: S8820), bovine serum albumin (BSA) (cat#: 05479), and cisplatin (cat#:1134357), oxaliplatin (CAS: 61825-94-3), brusatol (cat#:SML 1868) were procured from Sigma Chemical Company (St. Louis, MO, USA). Primary antibodies for FN3K were procured (Catalog # PA5-28603) from Invitrogen whereas the primary antibodies labeling for Nrf2 (cat#: 12721, cell signalling), NQO1 (cat#: 62262), HO-1 (cat#: ab13248), Keap1 (cat#: ab227828) were procured from Abcam (Cambridge, MA, USA) and Cell Signalling Technologies (Danvers, MA, USA). Secondary antibodies (Rabbit cat#: SC2357 and Goat cat#: SC2020 were procured from Santa Cruz Biotechnology (Santa Cruz, CA, USA). FBS (cat#:10270106), Pen-Strep (cat#:150763) were procured from ThermoFisher Scientific (Waltham, MA, USA) Dilutions of all the primary antibodies and secondary antibodies were carried out as per the dilution factors given by the specification of the manufacturer. Kinase inhibitors such as sorafenib, tamoxifen citrate, neratinib, gefitinib, and other anti-cancer drugs including cyclosporine A, methotrexate, topotecan, and brusatol were obtained from the Bi Biotech, New Delhi, India.
3D-protein homology sequence modeling
As there is no availability of three-dimensional structure of Human FN3K (Homo sapiens) available in either Uniprot or Protein Data Bank (PDB) databases, the 3-dimensional homology modeled FN3K was constructed using the available ‘specific template’ of FN3K (Arabidopsis thaliana, 6OID) using ‘SWISS EXPASY Tools’ (SIB) [9].
Primarily, the amino acid FASTA sequences (Uniprot, 309 AA, accession number: Q9H479), of FN3K in Homo sapiens and A. thaliana (6OID, 317AA (PDB)) were collected from NCBI (http://www.ncbi.nlm.nih.gov). Subsequently, the phylogenetic similarities in the sequences of these two were ascertained with the aid of Clustal W tools (https://www.ebi.ac.uk/Tools /msa/clustalo/). A template for homology structural modeling was developed with the help of the SWISS-Model (https://swissmodel.expasy.org/) [9]. Subsequently, these models were subjected to refining using Galaxy Web (http://galaxy.seoklab.org/index.html), and verified using Ramachandran plots, and PROCHECK (https://servicesn.mbi.ucla.edu/PROCHECK/) analysis, global model quality estimation (GMQE) score and qualitative model energy analysis (QMEAN) values. All these procedures were executed as per Bharadwaja Vadloori et al [10]. As a control, the 3D structure of FN3K reported from A. thaliana was considered to be a reference to validate the designed 3D homology model of human FN3K using Ramachandran plots and PROCHECK. During the modeling process, the crystallized ligand was placed in its original position to rebuild the binding cavity. The resulting human FN3K model was subjected to PROCHECK validation [11].
Computer-aided drug design [Screening of anticancer molecules including protein kinase inhibitors for molecular docking]: Ligand preparation
For our study, we have selected anticancer molecules including synthetic protein kinase inhibitors, and natural protein kinase inhibitors to check their binding affinity with ATP binding loop of designed human FN3K 3D-homology sequence to optimize kinase inhibitors using CADD techniques. A report by Kiyoaki Yonesu et al (2011) reported that the lysates of red blood cells incubated with 1-deoxy-1-morpholinofructose (DMF), a competitive inhibitor of FN3K, could inhibit ∼10% of total FN3K activity [12]. Hence, the DMF structure file was obtained from Pubchem (CID, 220877) and subsequently copied to Chem 3D ultra 10.0 and generated a three-dimensional model; later, the model was subjected to energy minimization with the aid of molecular mechanics (MM2) until the root mean square gradient attained specific value lower than 0.001 Kcal/mol. These kinds of structures with energy minimization were generally required for efficient docking. MDL MOL files of these structures were generated and saved as /MDL MOL. The ligand preparation was executed as per the procedures described by Awwad Radwan et al 2020 [13].
3D homology model protein preparation
X-ray diffraction properties and resolution of the structure (2.0–2.5 Ǻ) as well as the presence of co-crystallized ligand are the most significant factors required to choose a three-dimensional protein preparation for an efficient docking process. However, we do not have the ‘existing 3D of FN3K of Homo sapiens’, henceforth designed using SWISS Expasy tools. Ramachandran plot statistics were considered as the crucial validation filter to perform protein selection in which none of the amino acid residues were in disallowed regions. Assignment of Kollmann charges was executed before docking. All the protein preparation procedures were executed according to the procedures given by Awwad Radwan et al 2020 [13].
Molecular docking
Molecular docking was executed onto the monomer of FN3K using Discovery studio. Primarily, the blind docking was executed and subsequently, the docking within the specified search space around the probable binding site of FN3K was performed. Docking conformations were chosen as per the binding affinity. Pymol and discovery studio visualizer was used to visualize 3D graphical representations (https://www.pymol.org/). Natural kinase inhibitors and FDA-approved synthetic kinase inhibitors were collected from the literature reports and databases for molecular docking to dock onto the designed 3-dimensional FN3K protein. Preliminary screening of >100 anticancer molecules was executed through LibDock Program, [14] a screen-based docking procedure to screen the number of hits from the database. Based on the LibDock score, we identified a total of ‘40’ molecules as hits. Among them, the potential lead molecules were identified and docked using a simulated annealing-based CDocker approach [14].
Specific algorithms were used in the Discovery studio docker tool to examine the intermolecular interactions of numerous ligands with macromolecules like 3-dimensional proteins. CDOCKER was one of the significant algorithms applied in a simulated annealing-based method to allow structural parameterization in CHARMm force fields [15–17]. Later, any structural inappropriacy was eliminated by performing adequate and suitable corrections of loop regions or side chains, or conformations.
On the other hand, the ligands were also subjected to processing to fix valences and charges in CHARMm force field. Specific ATP binding region was identified by ‘searching for potential receptor cavity [binding site specifications: volume: 1098 Å3, point count: 8340 0.5(X), 0.5(Y), and 0.5(Z) directions-equal gird spacing]’ subsequently executed the docking procedures [sphere site: 23. 28 (X), 21.01 (Y), and 2.16 (Z)] by ligand hits onto active binding site cavities using docking algorithm.
Pose-resolved interactions pertinent to the docked ligands were collected for subsequent analysis and MD simulation [14]. All the molecular docking procedures were executed as described by Awwad Radwan et al 2020 [13].
Molecular dynamics simulations [Gromacs 5.0]
Before MD simulations of 3D-modeled human FN3K, we initially performed the protein structure alignment of the homology FN3K model (cyan) with MEK1 PDB structure (yellow) (Fig 2A) to explore the binding and interaction efficacy; subsequently, the chosen ligands such as synthetic anticancer kinase inhibitors such as ‘neratinib (PubChem ID: 9915743), AZD1480 (PubChem ID: 16659841), Jadomycin B (PubChem ID: 164484), tamoxifen citrate (PubChem ID: 2733526), sorafenib (PubChem ID: 216239) ponatinib (PubChem ID: 24826799), gefitinib (PubChem ID: 123631), tricirbine (PubChem ID: 65399), and natural multikinase inhibitor, EGCG (PubChem ID: 65064)’, DMF (PubChem ID: 6454484), an FN3K inhibitor were subjected to the MD simulations in Gromacs 5.0 as per the procedures described by Vadloori et al 2018. Topological parameters pertinent to ligands were collected using the ATB server (https://atb.uq.edu.au/). Primarily, the protein-complex was placed in the cubix box having water with the aid of SPC/E water models. GROMOS54a7 force field was used to execute energy minimization and subjected to equilibration at 300K with the aid of V-rescale for the duration of 200 ps. NVT ensemble was used consequently and equilibration was performed at one atmospheric pressure with the aid of Parrinello-Rahman Algorithm. The NPT ensemble duration was 200 ps. The equilibrated conformation has been conducted for the duration of 20 nanoseconds simulation. Verlet was used for applying LINCS algorithm to perform bond constraints. Root mean square deviations (RMSD) and radius of gyration (Rg) at the time of simulation were calculated with the aid of gmx_rms tool procured through Gromacs 5.0. Primarily, all the hits from virtual screening were subjected to a 20 ns simulation run to estimate the binding efficiency, which successfully validated and adapted the molecular docking protocol. However, the confirmation of protein-ligand stability under a thermodynamically equilibrated environment was assessed by extending the production run for a 100 ns time frame of unbound protein and specific ligands (sorafenib and gefitinib) bound protein complexes. These ligands were associated with a significant binding affinity than the other molecules. The extended MD run was further employed to estimate the binding energies through MM-PBSA calculation [18–22].
HPLC characterization
of sorafenib, tamoxifen citrate, neratinib, gefitinib, cisplatin, oxaliplatin, cyclosporine A, topotecan, brusatol at specific chromatography conditions was given as supplementary attachment (S1 Table).
ADME
(absorption, distribution, metabolism, and excretion) properties of molecules screened (Table 2) deciphered using the computational SWISS ADMET Server. Descriptors with significant molecular properties pertinent to pharmaceutical compliance to Lipinski’s rule of five RO5 were calculated through the ‘SWISS ADMET Server’ as described by Ramesh Itteboina et al 2017 [23].
Cytotoxicity assays [sulforhodamine B proliferation (SRB) assay]
BT-474 (luminal B) and T47D (luminal A) ductal invasive carcinomas type of breast cancer cells were cultured into 96 well plates (10000 cells/well) and subsequently subjected to incubation for 24 and 48 hours respectively. Later, the cells were treated with drugs for 24 hours and 48 hours respectively:
BT-474/T47D
- Neratinib: 10 μM, 5 μM, 2.5 μM, 1.25 μM, 0.625 μM, 0.312 μM, 0.156 μM
- Sorafenib/tamoxifen citrate/topotecan: 1.5 μM, 3.125 μM, 6.25 μM, 12.5 μM, 25 μM, 50 μM, 100 μM
- Cisplatin/oxaliplatin: 15.6 μM, 31.25 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM, 1000 μM
BT-474
- Methotrexate: 7.8 μM, 15.6 μM, 31.25 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM
- Cyclosporine A: 3.9 μM, 7.8 μM, 15.6 μM, 31. 2 μM, 62. 5 μM, 125 μM, 250 μM
- Brusatol: 20 μM, 10 μM, 5 μM, 2.5 μM, 1.25 μM, 0.625 μM, 0.312 μM
T47D
- Gefitinib: 7.8 μM, 15.6 μM, 31. 2 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM
- Cyclosporine A: 7.8 μM, 15.6 μM, 31. 2 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM
- Methotrexate: 15.6 μM, 31.25 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM, 1000 μM
- Brusatol: 20 μM, 10 μM, 5 μM, 2.5 μM, 1.25 μM, 0.625 μM, 0.312 μM
DMF (known FN3K inhibitor for BT-474 and T-47D: 0. 15 mM, 0.31 mM, 0.625 mM, 1.25 mM, 2.5 mM, 5 mM treated for 48 hours.
Vehicle control (DMSO) was maintained at <1%; cisplatin (100 μM) was used as a positive control; finally, each well in the plates was fixed using 100 μl cold 25% trichloracetic acid. Later, the plates were washed using distilled water and subjected to drying and, consequently, 50 μl SRB dye (0.4% SRB dissolved in 1% glacial acetic acid) was added to every well. Later, the plates were subjected to incubation for thirty minutes at 37°C. 1% glacial acetic acid was used for washing consequently the plates were dried for 3 hours. Later, every well in the plates was added 100 μl of Tris-NaOH buffer (pH 10.5, 10 mM) and subjected to incubation for one hour, and the reading was taken using Perkin Elmer multimode plate reader (UK) for ascertaining optical density at 540 nm. IC-50 calculations were performed with the aid of GraphPad Prism 7 package.
Western blotting
Primarily, the cell lines such as MDA-MB-231, MDA-MB-468, MCF-7, BT-474, T-47D, A549, and Beas2B were screened for the expression patterns of FN3K, Nrf2 and Keap1. Later, the drug-treated lysates of BT-474 alone were analyzed for the FN3K, Nrf2, and their downstream antioxidant signaling. The total protein content of drug-treated BT-474 cells (sorafenib/tamoxifencitrate/neratininb/gefitnib/cisplatin/oxaliplatin/cyclosporine/methotrexate) was collected with the aid of lysis buffer subsequently protein content was determined through BCA method. Total of 50 μg of equal quantities were taken for SDS-PAGE running in 12% gel by applying 100 V for two hours. Later, the proteins were separated into PVDF/nitrocellulose membrane with the aid of semidry equipment. Skimmed milk of 5% was used to block the membranes for one hour in TBST buffer. Blots were subjected to TBST washing 3 times subsequently subjected to primary antibody loading and incubated overnight at 4°C. Later, the blots were washed using TBST thrice and subjected to incubation with ‘HRP-linked secondary antibody’ for two hours. ECL (1:1 ratio of reagent 1 and reagent 2) was added to the above samples and observed for banding pattern in Chemi-UV tech in the dark till the bands appeared. Additionally, FN3K expression alone was probed in T47D cell lines using Western blotting.
NQO1 activity assay
This procedure was executed as per the protocol described by Bovilla et al (2021) [6]. NQO1 activity for the drug-treated lysates of BT-474 cells was examined with the aid of a standard reaction of glucose-6-phosphate (G6P) when incubated with G6P dehydrogenase (G6Pdase). In this reaction, NADPH was generated and it can induce the reduction of menadione into menadiol. Subsequently, the formazan in blue color was generated and ascertained at 610 nm with the aid of a multimode plate reader. Furthermore, dicoumarol has been used to examine the background activity induced through reductases.
The total protein content of 10 μg obtained from the drug-treated cell lines was diluted to 40 μL with Millipore water and subjected to incubation using NQO1 cocktail (200 μL) in the presence/absence of dicoumarol [3 samples with inhibitor whereas 3 samples without dicoumarol; 6 wells/test]. Subsequently, absorbance was obtained at 610 nm for 60 minutes with one minute interval. The NQO1 activity was determined by performing subtraction of readings pertinent to ‘samples with inhibitor’ from readings pertinent to ‘samples without inhibitor’. OD values were deciphered per minute whereas the mole units were determined by multiplying the OD/minute/molar extinction coefficient of MTT with the protein concentration of each sample. Finally, the NQO1 activity was represented in ‘specific activity (μmol/min/mg protein)’.
Statistics
Experiments were repeated in triplicates and the data was indicated in Mean ± SEM. The level of significance was calculated with the aid of GraphPad Prism version 6.0 (Graphpad Software, CA, USA). Other statistical tests such as one-way ANOVA and Tukey’s post hoc test were executed to compare different experimental groups. The “p” values <0.05, <0.01, <0.001 were considered statistically significant.
Results
Sequence alignment and homology modeling
According to a previous study by Roberto Sánchez et al (1999), if two proteins exhibit 30% of “segment identity,” then Cα atoms pertinent to these proteins are within 3.5 Å [24]. For instance, AtFN3K (Arabdosis thaliana FN3K) adopts a canonical protein kinase fold (PKL-fold) due to the presence of an N-terminal ATP binding lobe as well as due to a C-terminal substrate-binding lobe, and these features are making FN3K as specific for docking or for obtaining a specific template for homology modeling of human FN3K using SWISS EXPASY tools [25]. Hence, we developed a homology model of human FN3K using a sequence template of AtFN3K. The amino acid sequence of Human FN3K was blasted against the PDB-BLAST database to identify a suitable template for homology modeling. In our study, we found 50 templates and 3 best sequence homology models viz., model-1 with 40.67% (oligostate: homodimer), model-2 with 38.57% (oligostate: monomer), and model-3 with 26.2% (Oligostate: monomer) respectively. Among them, the sequence homology model-1 between the human FN3K and the selected template of FN3K of Arabdosis thaliana was found to be 40.67% as the highest (Fig 1A) and it could be a valid target for a drug after validation as minimum 30% of sequence similarity is required for CADD and MD simulations [13, 24].
A. Human FN3K homology modeling was performed using Swiss-Model EXPASY tools and 3 reliable models were obtained including 1 ‘homodimer’, and 2 monomers. Homodimer with 40.67% of sequence similarity was further considered for molecular docking. B. Phylogenetic similarity was depicted by cladogram of human FN3K with FN3K of Arabidopsis thaliana (AtFN3K: PDB-ID: 6OID) and the multiple sequence alignment between Human FN3K and AtFN3K [9, 13].
Validation
- Clustal W analysis of multiple sequence alignment between human FN3K and A thaliana was given in Fig 1B, and a cladogram for phylogenetic similarity was depicted in Fig 1B [13].
- Selected model-1 of FN3K has shown GMQE score of 0.70 and QMEAN score of −4.67 (Fig 2A–2C) [10].
- Before the refinement using Galaxyweb, selected model-1 showed 8% residues in the disallowed regions of Ramachandran plots. After refinement using Galaxyweb (5 models in S2 Fig), the model-1 in PROCHECK validation of Ramachandran plots described the presence of only 0.4% residues in the disallowed regions (Fig 3B and 3C).
- The complete FASTA sequence of A. thaliana (6OID) was taken as a ‘reference’. We have assessed the ‘3D model’ of this monomer and constructed Ramachandran plots (0% of residues in disallowed regions) for comparison (Fig 3A). In conclusion, ‘Modelled & Refined Chain A MODEL-1 of FN3K homodimer’ is considered the ideal model for molecular docking and MD simulation. We performed the molecular docking with DMF (a selective FN3K inhibitor), >50 synthetic and ≥50 natural kinase inhibitors onto the ‘Modelled & Refined Chain A MODEL-1 of FN3K homodimer’. [Note: we have obtained the best model of Chain A of FN3K modeled homodimer, attached as supplementary (S2 Fig)]. We performed molecular docking for the 3D model of AtFN3K as a reference standard attached as supporting information (S1 Fig) and also performed the docking of DMF onto the homology modeled FN3K (Fig 4).
(A). Human FN3K target alignment with the template of AtFN3K. (B). The Local quality estimates of chain A and chain B for the homology of ‘human FN3K homodimer’. (C). Quality estimates by Z score and QMEAN of the FN3K homology homodimer with 40.67% seq. similarity.
Chain-A ‘validation’ of the homology model of a 3D homodimer of FN3K using ProCHECK and Ramachandran plots, where the disallowed regions were completely minimized to 0.4% in Refined Chain A MODEL-1 of FN3K as depicted (C) compared to the non-refined model (B), and reference model of A thaliana (A). Refined Chain A Model-1 was chosen out of 5 models (S2 Fig.) produced from ‘The galaxyweb protein refinement tool’ and comparatively considered as the best model, which was validated by the Ramachandran plots via ProCHECK and this was further used for molecular docking and MD simulation [9].
These molecules interacted with active sites of chain A homodimer was shown by (-) CDOCKER interaction energy; and the interaction energies for the individual docked molecules were given along with Pubchem IDs. Coordination bonds and hydrogen bonds were depicted in dashed lines [8].
Molecular docking results
The anticancer molecules were screened after docking with the modeled human FN3K and CDocker score, and CDocker interaction energies were described for benzoic acid derivatives, cinnamic acid derivatives, synthetic & natural kinase inhibitors, and reference kinase inhibitor molecules. As a preliminary docking process, 40 molecules out of ≥ 100 molecules were screened according to the CDOCKER interaction energies (Fig 5) subsequently, molecular docking for 10 molecules was executed which resulted in identifying natural and synthetic anticancer small molecules as modulators for the activity of FN3K protein model (Fig 4) with most stable conformations in static mode of interactions of ligands in the binding pocket of the protein.
CDocker interaction energies
of synthetic anticancer kinase inhibitors and DMF, a competitive FN3K inhibitor were selected as they could bind to the active pockets of modeled human FN3K with substantial binding energies (Table 1). Additionally, we have attached S1 Fig in which the DMF interaction energies when interacted with active sites of reported three-dimensional FN3K protein (PDB ID: 6OID) of A. thaliana.
ADME predictions of the screened anticancer molecules
Drug likeliness was ascertained by examining physicochemical properties as well as by applying Lipinski’s rule of five (Table 2). As per this rule, the molecular weight for any ideal molecule should be below 500 Daltons, H-bond donor < 5, ad H-bond acceptors < 10, as well as the log P value should be less than 5. The aqueous solubility (QPlogS) was in the acceptable range for sorafenib, ZD1480, and tamoxifen citrate, whereas the partition coefficient QPlogPo/w was within the acceptable range for sorafenib, jadomycin B, AZD 1480, neratinib, gefitinib, tamoxifen citrate, and ponatinib. The majority of the screened anticancer molecules can cross the blood-brain barrier and all the values were described in Table 2.
Molecular dynamic simulations
Initially, we have examined the protein structure alignment of the ‘homology built FN3K model’ (cyan) with MEK1, a potent oncogenic kinase that has similar structural homology as FN3K (PDB structure (yellow) (PDB ID: 4U7Z) (Fig 6A). This alignment described the ATP binding site in modeled FN3K as the most significant region for the binding of the following molecules. The multikinase inhibitors which we have chosen above include neratinib: (-) 57.18 Kcal/mol (PubChem ID: 9915743), AZD1480: (-) 38.39 Kcal/mol (PubChem ID: 16659841), jadomycin B: (-) 51.03 Kcal/mol (PubChem ID: 164484), tamoxifen citrate: (-) 47.81 Kcal/mol (PubChem ID: 2733526), sorafenib: (-) 55.01 Kcal/mol (PubChem ID: 216239), ponatinib: (-) 57.79 Kcal/mol (PubChem ID: 24826799), gefitinib: (-) 52.61 Kcal/mol (PubChem ID: 123631), tricirbine: (-) 37.04 Kcal/mol (PubChem ID: 65399), and natural kinase inhibitor, EGCG (PubChem ID: 65064)’ exhibit a significant interacting ability with the human FN3K homology modeled protein.
(a). Protein structure alignment of homology built FN3K model (cyan) with MEK1 PDB structure (yellow) (PDB ID: 4U7Z). Analysis of RMSD, Rg, hydrogen bond, SASA, and RMSF of FN3K homology model (protein) with 3g (ligand) complexes at 20000 ps (20 ns). (b,h). Time evolution of backbone RMSD of the FN3K-model protein (chain A of homodimer) alone and with kinase inhibitors of natural (EGCG Pubchem ID: 65064, Jadomycin B Pubchem ID: 164484), synthetic (neratinib Pubchem ID: 9915743, gefitinib Pubchem ID: 123631, sorafenib Pubchem ID: 216239, tamoxifen citrate Pubchem ID: 2733526) and reference ligand (AZD1480, Pubchem ID: 16659841) complexes. (c, i). Radius of gyration (Rg) of the protein backbone in its free and complex form over the entire simulation time. The ordinate is Rg (nm) and the abscissa is the time (ps) interval. (d, j). Hydrogen bonds occurring over the time (ns) of simulation between model protein and ligand. (e, k) SASA is indicated, where the ordinate is SASA (nm) and the abscissa is time (ns). (f, l) and (g, m). Residue-wise average RMSF plot of protein in the innate and ligand-bound form [8, 13, 76, 77].
Preliminary MD simulation for a 20 ns time frame was found useful for all the above molecules to draw a correlation between molecular docking and dynamic stability of protein-ligand complexes; however overall understanding of the attainment of stable conformation from the dynamics nature was achieved by extending the production run for 100 ns time frame for the specific molecules such as sorafenib and gefitinib which have significant (-) CDocker interaction score with the homology modeled 3-dimensional human FN3K.
In molecular dynamic (MD) simulation studies, the trajectory analysis (Fig 6) of protein-ligand complexes could conclude specific conformational changes occurring in the complexes during MD simulation. Dynamic nature was produced by the solvated protein systems which consequently generate fluctuations from initial conformations; hence, the root mean square deviation (RMSD) maps out the fluctuations and delivers the plot in time scale and provided ‘specific estimations pertinent to equilibrium as well as protein stability’ (Fig 6B and 6H) of the backbone C-α atoms that the apoprotein and ligand-bound protein complex reached equilibrium approximately at 4.5 to 5.0 ns time and remaining were evident of stable trajectory with minimal deviation in ~0.10 to ~0.15 nm RMSD range; this was indicating the attainment of the structural rigidity of protein being bound to ligand than its free form. The unbound protein exhibited stable trajectories, whereas, ligand-bound proteins reached equilibrium only after initial fluctuations during the simulation.
The radius of gyration (Rg) was considered a significant parameter that depicts the altered masses calculated to root mean square distances while in motion around the central axis of rotation. Rg plot (Fig 6C and 6I) was indicative of capability, shape, and folding at the time of motion at each time step of the entire trajectory. The protein was in its ‘apo form’ and it was found with a specific reduction in gyration terms ~0.25 nm allying to the ligands bound protein showing stable trajectories only fluctuating in ~0.15 nm distances.
H-bonds are the significant characteristic for any protein-ligand interactions to evaluate in terms of stability. The biological functions were analyzed over the total simulation period. The intermolecular H-bond plot (Fig 6D and 6J) in time scale indicates that the number of H-bonds remains consistent with molecular docking, while simultaneously breaking and reforming of bonds being observed.
Solvent accessible surface area (SASA) measures the area around the hydrophobic core formed between protein-ligand complexes (Fig 6E and 6K). Consistent SASA values were observed typically fluctuating in between ~20–40 nm2 areas.
The residue-wise fluctuations can be evident from the root mean square fluctuation (RMSF) plot (Fig 6F, 6L, 6G and 6M), where amino acids are rigidified in the complex form as compared to the innate state of the protein. RMSD, Rg, hydrogen bond, SASA, and RMSF of FN3K homology model (protein) with gefitinib (Pubchem ID: 123631), sorafenib (Pubchem ID: 216239) complexes at 100000 ps (100 ns) were given in Fig 7A–7F.
(a) Time evolution of backbone RMSD pertinent to the FN3K model protein (chain A of homodimer) alone and FN3K-gefitinib, FN3K-sorafenib complexes. (b) The radius of gyration (Rg) of the protein backbone in its free and complex form over the entire simulation time. The ordinate is Rg (nm) and the abscissa is the time (ps) interval. (c) Hydrogen bonds occurring over the time (ns) of simulation between FN3K model protein and ligand. (d) SASA was indicated, where ordinate was SASA (nm) and abscissa was time (ns). (e, f) Residue-wise average RMSF plot of protein in the innate and ligand-bound form [8, 13, 76, 77].
Along with other MD analysis, SASA was employed to access the solvent interactions with the protein and its complexes, which ultimately outline the information about the protein’s 3D conformation. Total SASA provides us with the solvent accessibility of the complex as a whole, whereas the buried SASA indicates the stability of the interface as a measure of the binding surface between the protein and the ligand which is formed by the binding of the ligand to the protein. Table 3 described both the complexes such as FN3K-gefitinib and FN3K-sorafenib with stable trajectories in a similar trend with minor perturbations.
To further examine the binding efficiency of the synthetic hits (gefitinib and sorafenib) on FN3K protein, binding free energy calculation by MM/PBSA was carried out and the results were listed in Table 4. According to the MM/PBSA calculation, sorafenib exhibited higher binding affinity with -29.23 ± 5.42 KJ/mol to FN3K protein when compared to gefitinib having an affinity of—24.45 ± 7.21 KJ/mol. Among the other four terms, ΔGpolar and ΔGnonpolar have contributed similarly, while ΔEvdw and ΔEelectrostatic differed considerably in both cases because of the changes in Vander Waals and electrostatic interactions due to variation in the intermolecular distances. The final result of the simulation provides insights into the bioactive nature of all the ligands.
Therefore, MD simulation studies confirmed the presence of stable H-bonding interactions. Among the molecular dynamics of anticancer molecules such as ‘neratinib, AZD1480, Jadomycin B, tamoxifen citrate, sorafenib gefitinib, and natural multikinase inhibitor, EGCG ‘, the SASA values and MM-PBSA values were ascertained for FN3K-gefitinib and FN3K-sorafenib complexes respectively. Thus, the summed-up result of the simulation provides insights into the bioactive nature of ligands towards FN3K homology-built protein.
Before the in vitro assays, the identification, and characterization of the selected anticancer molecules viz., such as sorafenib, neratinib, tamoxifen citrate, gefitinib, methotrexate, cisplatin, oxaliplatin, cyclosporine A, topotecan, brusatol at specific optimized chromatographic conditions in HPLC method were depicted in Fig 8.
The chromatographic peaks were observed which denotes the characteristic identification of drug molecules such as (a) sorafenib, (b) neratinib, (c) tamoxifen citrate, (d) gefitinib, (e) methotrexate, (f) cisplatin, (g) oxaliplatin, (h) cyclosporine A, (i) topotecan, (j) brusatol respectively.
IC-50 activity determination for screened anticancer molecules
IC-50s for potent kinase inhibitors including sorafenib, tamoxifen citrate, neratinib, gefitinib, and other anticancer drugs including cisplatin, oxaliplatin, cyclosporine A, and methotrexate were identified for BT-474 and T-47D cell lines respectively (Table 5). Additionally, we have performed cytotoxicity for anticancer phytochemicals including topotecan, and brusatol on these breast cancer cell lines (Fig 9) subsequently the selected concentrations were used for treatment onto BT-474, T-47D cell lines for Western blotting to examine the expression patterns of FN3K, Nrf2 antioxidant downstream signaling proteins.
(A), (B). Cytotoxicity: BT-474 and T47D breast cancer cells were cultured and seeded into 96 well plates (10000 cells/well) subsequently subjected to incubation for 24 and 48 hours respectively. Later, the cells were treated for drugs (BT-474/T47D: neratinib: 10 μM, 5 μM, 2.5 μM, 1.25 μM, 0.625 μM, 0.312 μM, 0.156 μM; sorafenib/tamoxifen citrate/topotecan: 1.5 μM, 3.125 μM, 6.25 μM, 12.5 μM, 25 μM, 50 μM, 100 μM; cisplatin/oxaliplatin: 15.6 μM, 31.25 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM, 1000 μM); BT-474: gefitinib: 1.5 μM, 3.125 μM, 6.25 μM, 12.5 μM, 25 μM, 50 μM, 100 μM, methotrexate: 7.8 μM, 15.6 μM, 31.25 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM; cyclosporine A: 3.9 μM, 7.8 μM, 15.6 μM, 31. 2 μM, 62. 5 μM, 125 μM, 250 μM; Brusatol: 20 μM, 10 μM, 5 μM, 2.5 μM, 1.25 μM, 0.312 μM, 0.625 μM,; T47D: gefitinib: 7.8 μM, 15.6 μM, 31. 2 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM; cyclosporine A: 7.8 μM, 15.6 μM, 31. 2 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM; methotrexate: 15.6 μM, 31.25 μM, 62. 5 μM, 125 μM, 250 μM, 500 μM, 1000 μM; Brusatol: 20 μM, 10 μM, 5 μM, 2.5 μM, 1.25 μM, 0.625 μM, 0.312 μM; vehicle control <1%; cisplatin (100 μM) was used as a positive control for 24 hours and 48 hours respectively. DMF (known FN3K inhibitor): 0.15 mM, 0.31 mM, 0.625 mM, 1.25 mM, 2.5 mM, 5 mM for 48 hours. (*p< 0.05 compared to control, **p< 0.01, ***p<0.001 compared to control, NS: nonsignificant). [Note: drug precipitation occurred at higher concentrations of cyclosporine treatment in both the cell lines]. (B),(D). Morphological changes such as cell shrinking, cell rounding was observed in a dose-dependent fashion when the T47D, BT-474 cells were treated with different anticancer drug molecules. In case of DMF treatment, the cytostatic effect was observed in both the cell lines as there was no 50% cell death. Magnification: 4x, 10x.
Differential expression patterns of FN3K and Nrf2 antioxidant signalling
We specifically selected anticancer molecules including sorafenib, neratinib, tamoxifen citrate, gefitinib, methotrexate, cisplatin, oxaliplatin, and cyclosporine A to examine the expression patterns of FN3K, Nrf2-Keap1 and its downstream antioxidant signalling proteins in BT-474 cell lines. Expression of FN3K was significantly higher in ductal invasive carcinomas such as BT 474, T47D, and MCF-7.
Furthermore, the expression of FN3K is significantly upregulated when compared to control in all the drug-treated groups such as neratinib, sorafenib, gefitinib, tamoxifen citrate, cyclosporine A, and methotrexate. Consequently, the Nrf2 protein expression was downregulated in neratinib, sorafenib, tamoxifen citrate, and gefitinib, but no changes of expression were observed in the case of cyclosporine A or methotrexate. Keap1 expression patterns were significantly altered in all the drug-treated groups compared to the control. However, oxaliplatin completely downregulated FN3K expression in all the treated concentrations when compared to cisplatin. Nrf2 and Keap1 expression patterns also altered with these platinum derivatives in BT-474 cells. Other Nrf2 downstream antioxidant genes such as NQO1 and HO-1 were also mitigated in their expression patterns upon drug treatments (Fig 10A–10G).
A. Western blot protein expression study delineated the substantial expression of FN3K in ductal invasive carcinomas such as BT-474, and T-47D (Luminal subtype) compared to the other molecular subtypes including Triple-negative breast cancers (TNBCs). In addition, the expression of Nrf2 and Keap1 was probed in TNBCs (MDA-MB-231, MDA-MB-468), MCF-7, and luminal molecular subtype cancers. FN3K expression in HepG2 cells was considered a positive control. B. BT-474 cells with neratinib treatment: Expression patterns of FN3K protein were increased followed by the enhancement in Keap1 expression with slight expression changes in Nrf2, subsequently the NQO1 and HO-1 antioxidant protein expression were minimized. C, D, E. BT-474 cells with gefitinib/sorafenib/tamoxifen citrate treatment: Expression patterns of FN3K protein were increased followed by the decline in Nrf2, Keap1, NQO1, and HO-1 antioxidant protein expressions. F. BT-474 cells with cyclosporine A/methotrexate treatment: Expression patterns of FN3K protein were increased followed by the decline in Keap1 compared to control without significant alteration in the Nrf2 expression pattern, whereas NQO1 and HO-1 antioxidant protein expression was not changed compared to control; In case of methotrexate, the expression of Nrf2 slightly decreased and Keap1, NQO1, the HO-1 expression pattern was downregulated compared to control. G. BT-474 cells with cisplatin/oxaliplatin treatment: expression of FN3K, Nrf2, and Keap1 was not completely declined with cisplatin but NQO1, and HO-1 expression was mitigated at higher concentrations; interestingly, oxaliplatin downregulated the expression of FN3K and Nrf2 consequently mitigated expression of antioxidant proteins such as NQO1 and HO-1. 1-deoxy-1-morpholinofructose (DMF) (dissolved in 50% ethanol), a competitive inhibitor of FN3K, was used as a positive control which could only inhibit ∼10% of total FN3K activity but the FN3K expression was not decreased upon DMF treatment [12]. GAPDH was used as an internal control in all the treatments. [Note: all the original blots were provided as supporting information files and attached as original blots-1, original blots-2 respectively].
FN3K expression in T47D cells: Furthermore, we have examined the FN3K expression patterns alone in the T47D, another ductal invasive carcinoma cell line. Kinase inhibitors such as neratinib, sorafenib, gefitinib, tamoxifen citrate, and cyclosporine A enhanced the FN3K expression compared to control; topotecan, a camptothecin derivative of anticancer drug and methotrexate also upregulated the FN3K expression. Interestingly, both cisplatin and oxaliplatin treatment have downregulated FN3K expression (S3 Fig).
Nrf2 inhibitor: Brusatol treatment onto BT-474, T-47D upregulated the expression of FN3K (S3 Fig).
NQO1 activity measurements
NQO1 activity was executed as per the procedures described by Prochaska, H.J et al and Shen, J. et al [26, 27]. Drug-treated lysates were analyzed for NQO1 activity and the specific drugs including sorafenib, tamoxifen citrate, neratinib, DMF, gefitinib, cisplatin, oxaliplatin decreased the NQO1 activity. Although NQO1 activity enhanced with cyclosporine A but consequently this enhancement was observed in a dose-dependent manner in cyclosporine A treatment and the patterns of NQO1 activity were declined with methotrexate when compared to control (Fig 11).
The NQO1 was declined in sorafenib, tamoxifen citrate, neratinib, DMF, gefitinib, cisplatin, oxaliplatin, methotrexate and although its activity enhanced with cyclosporine A but consequently declined in a dose-dependent manner compared to control. (*p< 0.05 compared to control, **p< 0.01, ***p<0.001 compared to control, NS: nonsignificant), A549 cell lysates for NQO1 activity as standards were used during the assay. S: sorafenib, T: Tamoxifen citrate, N: Neratinib, G: gefitinib, CIS: cisplatin, OX: Oxaliplatin, Cys: cyclosporine A, Met: methotrexate; DMF (2.5 mM), an FN3K inhibitor. Control: BT-474 lysates without treatment.
Discussion
Breast carcinoma (BC) continues to exhibit resistance to the existing therapies due to dose-related toxicity and adverse effects on non-target tissues. Hence, identifying key protein(s) that play vital roles in the progression of breast carcinoma is highly significant. Protein glycation is minimally explored for the development of anti-cancer agents. A recent report by Sanghvi V., et al 2019 has demonstrated that the oncogenic activity of Nrf2 is regulated by deglycation by FN3K in hepatocellular carcinomas (HCCs) [4]. Thus, FN3K has a significant role in deglycating the Nrf2 in HCCs. Studies from our laboratory have demonstrated the upregulation of Nrf2 in breast cancers [6, 28]. In this study, we reported the FN3K modulators by screening several anticancer molecules subsequently docked onto the human FN3K protein homology model.
A report by Roskoski Jr R et al 2019 demonstrated the structure-dependent binding, three-dimensional catalytic domains of several protein kinases using homology modeling for ascertaining the ligand-binding efficacy of several FDA-approved kinase inhibitors [8]. Several protein kinase inhibitors are preferred for treating chronic diseases, cancers, and malignant and nonmalignant diseases [8]. The most significant targets of the kinase inhibitors include ‘ALK, B-Raf, BCR-Abl, epidermal growth factor receptor (EGFR), and vascular endothelial growth factor receptor (VEGFR) [8]. Human FN3K consists of an equivalent P-loop Cys, which is also redox sensitive [29]. Authors in this study reported a significant role for the “ATP binding site” in the redox regulation of FN3Ks. P-loop is stabilized in an extended conformation by “Cys-mediated disulfide bonds” connecting two chains to form a covalently linked dimer. The P-loop of FN3Ks is characterized by conserved Gly residues in addition to the redox-sensitive Cys residue; the presence of both these features might be necessary for the strand exchange dimer [8].
A study by Kornev et al described the active, conformations, and functional residues pertinent to several protein kinases with the aid of a local spatial alignment algorithm [30, 31]. According to their report, a catalytic spine is formed by the 8 hydrophobic amino acid residues whereas the regulatory spine is formed by four hydrophobic amino acid residues. These units are the stable and catalytic spine that could enable the positioning of ATP whereas the R-spine could interact with the protein substrate to mediate the catalytic reaction. A few studies by Liao and Van Liden et al segregated the amino acid residues between small and large lobes of protein kinases into the front cleft, gate area, and back cleft [32, 33]. The front cleft is composed of a glycine-rich amino acid loop and (HRD(x)4N) residues whereas the gate area is composed of an activation segment of DFG [8]. The back cleft region is formed by the αC-β4 back loop, β4- and β5-strands, and αE-helix located in the small lobe, and large respectively. Designing or docking therapeutic small molecule inhibitors to attain selectivity through drug interactions is a significant challenge in medicinal chemistry to mitigate disease processes including cancers [34–37]. Binding pockets located in the catalytic region of kinases could play a prominent role in predicting the efficacy of kinase inhibitors/activators or to design novel protein kinase inhibitors with utmost drug affinity. Previous studies described the ligand-binding affinity of several kinases like ‘cyclin-dependent kinase [38, 39], EGFR family of protein-tyrosine kinases [40–42], Janus kinase inhibitors [34], MEK1/2 dual specificity kinase inhibitors [43], Raf protein-serine threonine kinases [37], VEGFR1/2/3 protein-tyrosine kinases [8, 44].
In our study, FN3K is a kinase in which several kinase inhibitors and other anticancer molecules interacted through hydrogen bonding, hydrophobic interactions, and dipole interactions indicating a higher ligand-binding affinity. In the current study, we generated the homodimer of the FN3K homology 3D model with different sequence similarities and selected the chain A of the human FN3K 3D model with the highest sequence similarity (40.6%) for further validation using Ramachandran plots for the alignment of amino acids in beta helices and alpha helices. The template and target sequence of the human FN3K were validated phylogenetically using cladogram construction in Clustal W tools and Ramachandran plots. In the binding energies of the screened anticancer molecules through molecular docking onto the designed FN3K protein exhibited stable confirmations as indicated by the RMSD, Rg, and RMSF in the MD simulations. The kinase inhibitors are neratinib, sorafenib, gefitinib, tamoxifen citrate, AZD1480, Jadomycin B, and EGCG’ which exhibited a significant interacting ability with designed human FN3K modeled protein. Initially, molecular dynamics simulation run for 20 ns for anticancer molecules such as ‘neratinib, AZD1480, Jadomycin B, tamoxifen citrate, sorafenib gefitinib, and natural multikinase inhibitor, EGCG’ has attained a stable trajectory. However, we extended the simulation for 100 ns specifically for sorafenib and gefitinib with substantial binding efficacy with modeled FN3K and the stable trajectories were evident throughout the 100 ns duration. SASA values and MM-PBSA values were ascertained for FN3K-gefitinib and FN3K-sorafenib complexes respectively. The summed-up result of simulation related to the above molecules with modeled human FN3K provides insights into the bioactive nature of ligands towards FN3K homology-built protein. Nrf2 has a significant oncogenic role in several cancers due to its ability to get influenced by metabolic reprogramming to trigger antioxidant responses toward the chemotherapy [4]. The oncogenic role of Nrf2 is vividly responsible for tumor progression [5]. FN3K could enhance the deglycation of Nrf2 consequently increasing the antioxidant role of Nrf2 in HCC, and offering protection for cancer cells during chemotherapy [4]. Therefore, the development of small molecule inhibitors is a significant strategy to inhibit selectively FN3K in HCC, a strategy which may be used to target breast cancers. Kinase inhibitors such as sorafenib, gefitinib, neratinib, tamoxifen citrate, and platinum derivatives including cisplatin, oxaliplatin, other anticancer drugs such as cyclosporine A, & methotrexate, and brusatol, topotecan of natural origin were examined for the anti-breast cancer activity against luminal type breast cancer cell lines, BT-474, T-47D. Subsequently, we have screened and selected already proven anticancer molecules including neratinib, sorafenib, gefitinib, cyclosporine A, tamoxifen citrate, cisplatin, oxaliplatin due to their broad-spectrum activity and evaluated for their efficacy in modulating the FN3K expression in vitro. For instance, sorafenib is a multikinase inhibitor involved in impairing the autophosphorylation of receptor tyrosine kinases such as VEGFR1, 2, and 3, PDGFRβ, and RET as these receptor tyrosine kinases typically involved in angiogenesis and tumor progression even in breast cancers [45, 46]. A report by Sun, X. et al 2016 delineated that the blocking p62/Keap1/Nrf2 signaling resulted in enhanced sorafenib-induced HCC progression [47].
In our study, the expression of FN3K was substantially enhanced followed by the mitigation of Nrf2 expression and downstream target genes including HO-1, and NQO-1 with sorafenib. Neratinib is another multikinase inhibitor that can potentially block the activity of HER2, HER4, and EGFR and is reported to be implicated in breast cancers [48–51]. Gefitinib [52] is another EGFR-tyrosine kinase inhibitor [53] implicated against liver and lung cancers [52] and it has exhibited its efficacy in mitigating the breast cancer cell proliferation of MCF-7 and BT-474 [54, 55]. Tamoxifen citrate could inhibit the protein kinase C activity and is implicated as an anticancer drug as it has a significant role as a selective estrogen response modifier, and also an anti-angiogenic factor [56]. This drug can mitigate breast cancer progression [57–59]. However, the tamoxifen-resistance is one of the significant aspects in breast cancer cells and the resistant mechanisms are being maintained by the recurrent activation of Nrf2-mediated antioxidant response elements in breast cancer cells through ERK signaling [60]. Cyclosporine A is an enzyme pyruvate kinase inhibitor that substantially can impair breast cancer progression as pyruvate kinase enzyme expression is higher in the breast cancer cells like MCF-7, MDA-MB-231, and MDA-MB-435 [61, 62]. In the present study, treatment of these drugs against BT-474 at optimum dose ranges have mitigated the FN3K expression consequently reduced the Nrf2, HO-1, and NQO-1 expression patterns.
Other anticancer drugs like platinum derivatives including cisplatin and oxaliplatin were also examined against their ability for modulating the FN3K expression in breast cancers after determining their cytotoxicity [63–65]. A stable expression of Nrf2 in cancer cells could confer chemoresistance to cisplatin, etoposide, and doxorubicin [66]. Nrf2 activation through sulforaphane mitigated the cisplatin-invoked cancer cell death in Nrf2 proficient cells when compared to Nrf2-Cas9 cells [67]. FN3K expression and Nrf2 expression patterns were not mitigated with increasing doses of cisplatin in BT-474 cells. In the case of oxaliplatin [68], in the current study, the expression of Nrf2 and FN3K declined in a dose-dependent manner. Methotrexate is an immunosuppressant and is implicated to mitigate metastatic breast cancer [69] and inflammation as it can act as an antioxidant and scavenges superoxides, and it has been concluded that the methotrexate can inhibit Nrf2 activation in the cells treated with malondialdehyde and acetaldehyde during the inflammation in vitro models [70].
A previous study by Sanghvi et al (2019) reported that the knockdown of FN3K could promote the downregulation of Nrf2 target proteins in murine MYC/sgKeap1 HCC liver tumor isografts and similarly in three pairs of FN3K-proficient and FN3K-deficient human xenografts (Huh1, H460, and H3255) models [4]. Nrf2 inhibitors can induce the chemosensitization of chemotherapeutic molecules in HCC [71]. Brusatol is a potent Nrf2 inhibitor that can induce impairment of melanoma growth through the regulation of Akt-Nrf2 signaling when given along with UV irradiation [72]. Topotecan is a semisynthetic anticancer drug and is a derivative of camptothecin which can impair the topoisomerase I activity and induce double-strand DNA breaks. A recent study by Birandra K. Sinha et al 2020 reported the underlying mechanisms of topotecan-induced cancer cell death by examining the gene expression through microarray analysis. The study concluded that the topotecan treatment onto MCF-7 cells mitigated the expression of P53-regulating genes, estrogen receptor alpha (ERα/ESR1), and antiapoptotic genes (Bcl2) [73]. In addition, enhanced expression patterns of ‘ROS-related sensor genes [73] and proapoptotic genes in ZR-75-1 cell lines could be the most significant factors implicated in mediated topotecan-induced cancer cell death [73].
In this study, we have examined the activity of brusatol and topotecan cytotoxicity activity against breast cancers BT-474 and T47D invasive type breast cancers. We selected several natural, synthetic kinase inhibitors and screened the most potent multikinase inhibitors such as sorafenib, neratinib, tamoxifen citrate, and gefitinib potentially enhanced the deglycating enzymatic protein, FN3K consequently modulated the expression of Nrf2-downstream genes such as Keap1, HO-1, and NQO1. With sorafenib treatment, the expression of Keap1 and FN3K is upregulated whereas tamoxifen citrate also enhanced the expression of FN3K with concomitant modulation in the Nrf2 expression. It has been observed that the substantial expression of FN3K can result in the deglycation of Nrf2 resulting in the overexpression of AREs to enhance the cancer cell antioxidant potential [4].
Treatment of anticancer drugs such as cisplatin [74], oxaliplatin [75], and topotecan [75] can alter the expression of mRNA pertinent to FN3K [https://ctdbase.org/]. So, we have ascertained the IC-50 using cytotoxicity assays for these drugs and treated the invasive ductal breast carcinoma cells to examine the FN3K protein expression. In our study, FN3K expression is substantially suppressed with oxaliplatin during the treatment process in BT-474 cells. However, the expression of Nrf2 and consequently ARE-related gene expression was minimized in ductal invasive carcinomas such as BT-474. Cisplatin did not decrease the FN3K expression in BT-474. Furthermore, the expression patterns of FN3K were upregulated with kinase inhibitors, methotrexate but downregulated with both cisplatin/oxaliplatin in T47D cells. But, brusatol is an Nrf2 inhibitor of natural origin significantly upregulated the FN3K expression in both BT-474 and T4-7D cells (S3 Fig) indicating the combinatorial implications of both FN3K and Nrf2 inhibitors in breast cancers by examining Nrf2-glycation patterns.
It has been concluded that the Nrf2 inhibition could impair the cancer cell migration and consequently induce sensitization of cancer cells to chemotherapy [6]. Knockdown of Nrf2 genes could induce A549 sensitization to cisplatin and etoposide. In addition, blocking the activity of Nrf2 by trigonelline can induce susceptibility of pancreatic cancer cells to apoptosis. In our study, the NQO1 activity declined with the treatment of sorafenib, tamoxifen citrate, neratinib, DMF, gefitinib, cisplatin, oxaliplatin, and although its activity enhanced with cyclosporine A but consequently mitigated in a dose-dependent manner with the treatment of cyclosporine A. Patterns of NQO1 activity was declined with methotrexate treatment than control. Our study reported the expression of FN3K in ductal carcinomas, which are invasive types. A previous study depicted the knockdown of FN3K could induce a higher glycated Nrf2 in the cancer cells [4]. Therefore, the development of small molecule inhibitors is a significant strategy to inhibit selectively FN3K in breast cancers. Substantial studies are required to examine the activity of Nrf2 inhibitors/semisynthetic camptothecin derivatives like irinotecan, and topotecan along with FN3K inhibitors (oxaliplatin) in combinatorial regimens to target several cancers.
Conclusion
Screening of several kinase inhibitors, and anticancer drugs by molecular docking and molecular dynamic simulations was executed and subsequently examined the screened molecules for their ADME properties and the FN3K protein expression patterns. The current study demonstrated that oxaliplatin significantly downregulated the FN3K protein expression, a deglycating enzyme involved in Nrf2 glycation, a process implicated in tumor progression. In addition, the breast cancer cell death was substantially higher although the expression of FN3K enhanced with certain kinase inhibitors used for breast cancer treatment and the Nrf2 signalling was modulated consequently NQO1 and HO-1 antioxidant protein expression patterns also declined. It is crucial to examine the combinatorial regimen of natural anticancer molecules/Nrf2 blockers/kinase inhibitors with oxaliplatin to examine glycation patterns of Nrf2. Therefore, the antioxidant responses can be mitigated to alleviate breast cancer progression.
Summary
- FN3K expression is significantly upregulated when compared to control in all the drug-treated groups such as neratinib, sorafenib, gefitinib, tamoxifen citrate, cyclosporine A, and methotrexate. Consequently, the Nrf2 protein expression was also modulated in neratinib, sorafenib, tamoxifen citrate, and gefitinib, but no changes of expression were observed in the case of cyclosporine A or methotrexate. Keap1 expression patterns were significantly altered in all the drug-treated groups compared to the control.
- However, oxaliplatin completely downregulated FN3K expression in all the treated concentrations when compared to cisplatin in BT-474, and T-47D. Nrf2 and Keap1 expression patterns also altered with these platinum derivatives in BT-474 cells. Other Nrf2 downstream antioxidant genes such as NQO1 and HO-1 were also mitigated in their expression patterns upon drug treatments.
- In the case of T47D cells, both oxaliplatin and cisplatin both downregulated FN3K expression but brusatol (Nrf2 Inhibitor) enhanced the FN3K expression in BT-474, T47D.
- NQO1 activity was declined in sorafenib, tamoxifen citrate, neratinib, DMF, gefitinib, cisplatin, oxaliplatin, and methotrexate, and although its activity enhanced with cyclosporine A in a dose-dependent manner compared to control.
Note: Supplementary figures and tables were attached.
Supporting information
S1 Fig. Supplementary for DMF as a reference (a known FN3K inhibitor) for docking sites onto the AtFN3K protein.
https://doi.org/10.1371/journal.pone.0283705.s001
(TIF)
S2 Fig. Refined chain A MODEL-1 produced from Galaxyweb comparatively considered as the best model for docking and MD simulations.
https://doi.org/10.1371/journal.pone.0283705.s002
(TIF)
S3 Fig. Upregulated expression of FN3K was observed with the treatment of kinase inhibitors including gefitinib, neratinib, sorafenib, cyclosporine A, tamoxifen citrate, and with other anticancer molecules such as topotecan in T47D cell line.
Cisplatin/oxaliplatin treatment resulted in the mitigated expression of FN3K in T47D cells. Brusatol is an Nrf2 inhibitor that upregulated the FN3K expression in luminal-type breast cancer cells such as BT-474 and T47D. DMF, a competitive inhibitor of FN3K, was used as a positive control. GAPDH was used as an internal control.
https://doi.org/10.1371/journal.pone.0283705.s003
(TIF)
S1 Table. HPLC characterization of sorafenib, tamoxifen citrate, neratinib, gefitinib, cisplatin, oxaliplatin, cyclosporine A, methotrexate, topotecan, brusatol—chromatography conditions.
https://doi.org/10.1371/journal.pone.0283705.s005
(DOCX)
S1 File. Original blots-1: The original blots were provided as supporting information files and attached as original blots-1.
https://doi.org/10.1371/journal.pone.0283705.s006
(PDF)
S2 File. Original blots-2: The original blots were provided as supporting information files and attached as original blots-2.
https://doi.org/10.1371/journal.pone.0283705.s007
(PDF)
References
- 1. Ginsburg O, Yip CH, Brooks A, Cabanes A, Caleffi M, Dunstan Yataco JA, et al.: Breast cancer early detection: A phased approach to implementation. Cancer 2020, 126:2379–2393. pmid:32348566
- 2. Berumen AV, Moyao GJ, Rodriguez NM, Ilbawi AM, Migliore A, Shulman LN: Defining priority medical devices for cancer management: a WHO initiative. The Lancet Oncology 2018, 19(12):e709–e719. pmid:30507437
- 3. Brown TC, Sankpal NV, Gillanders WE: Functional implications of the dynamic regulation of EpCAM during epithelial-to-mesenchymal transition. Biomolecules 2021, 11(7):956. pmid:34209658
- 4. Sanghvi VR, Leibold J, Mina M, Mohan P, Berishaj M, Li Z, et al.: The oncogenic action of NRF2 depends on de-glycation by fructosamine-3-kinase. Cell 2019, 178(4):807–819. e821%@ 0092–8674. pmid:31398338
- 5. Bai X, Chen Y, Hou X, Huang M, Jin J: Emerging role of NRF2 in chemoresistance by regulating drug-metabolizing enzymes and efflux transporters. Drug metabolism reviews 2016, 48(4):541–567. pmid:27320238
- 6. Bovilla VR, Kuruburu MG, Bettada VG, Krishnamurthy J, Sukocheva OA, Thimmulappa RK, et al.: Targeted inhibition of anti-inflammatory regulator Nrf2 results in breast cancer retardation in vitro and in vivo. Biomedicines 2021, 9(9):1119. pmid:34572304
- 7. Gagic Z, Ruzic D, Djokovic N, Djikic T, Nikolic K: In silico methods for design of kinase inhibitors as anticancer drugs. Frontiers in chemistry 2020, 7:873%@ 2296–2646. pmid:31970149
- 8. Roskoski R Jr: Properties of FDA-approved small molecule protein kinase inhibitors. Pharmacological research 2019, 144:19–50. pmid:30877063
- 9. Owoloye Afolabi J., Ligali Funmilayo C., Enejoh Ojochenemi A., Musa Adesola Z., Aina Oluwagbemiga, Idowu Emmanuel T., and Oyebola Kolapo M. "Molecular docking, simulation and binding free energy analysis of small molecules as Pf HT1 inhibitors." Plos one 17, no. 8 (2022): e0268269. pmid:36026508
- 10. Vadloori B, Sharath AK, Prabhu NP, Maurya R: Homology modelling, molecular docking, and molecular dynamics simulations reveal the inhibition of Leishmania donovani dihydrofolate reductase-thymidylate synthase enzyme by Withaferin-A. BMC research notes 2018, 11(1):1–7%@ 1756–0500. pmid:29661206
- 11. Laskowski RA, Rullmann JAC, W Mam K: R, and Thornton JM (1996). J Biomol NMR, 8:477–486.
- 12. Yonesu K, Kubota K, Tamura M, Inaba S-i, Honda T, Yahara C, et al.: Purification and identification of activating enzymes of CS-0777, a selective sphingosine 1-phosphate receptor 1 modulator, in erythrocytes. Journal of Biological Chemistry 2011, 286(28):24765–24775. pmid:21613209
- 13. Radwan A, Mahrous GM: Docking studies and molecular dynamics simulations of the binding characteristics of waldiomycin and its methyl ester analog to Staphylococcus aureus histidine kinase. PloS one 2020, 15(6):e0234215%@ 0231932–0236203. pmid:32502195
- 14. Schaller D, Hagenow S, Stark H, Wolber G: Ligand-guided homology modeling drives identification of novel histamine H3 receptor ligands. PLoS One 2019, 14(6):e0218820%@ 0211932–0216203. pmid:31237914
- 15. Watanabe M, Kazuta Y, Hayashi H, Yamada S, Matsuda A, Shuto S: P-253 DEVELOPMENT OF POTENT HISTAMINE H_3 AND/OR H_4 RECEPTOR ANTAGONISTS BY THE STEREOCHEMICAL DIVERSITY-ORIENTED CONFORMATIONAL RESTRICTION STRATEGY. In: 2006: the Science Council of Japan under the Auspices of the International Union …; 2006: _P-253_ %@ 2433–2844.
- 16. Moguilevsky N, Varsalona F, Guillaume JP, Noyer M, Gillard M, Daliers J, et al.: Pharmacological and functional characterisation of the wild—type and site—directed mutants of the human H1 histamine receptor stably expressed in CHO cells. Journal of Receptors and Signal Transduction 1995, 15(1–4):91–102%@ 1079–9893. pmid:8903934
- 17. Bruysters M, Pertz HH, Teunissen A, Bakker RA, Gillard M, Chatelain P, et al.: Mutational analysis of the histamine H1-receptor binding pocket of histaprodifens. European journal of pharmacology 2004, 487(1–3):55–63%@ 0014–2999. pmid:15033376
- 18. Hess B, Kutzner C, Van Der Spoel D, Lindahl E: GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation. Journal of chemical theory and computation 2008, 4(3):435–447%@ 1549–9618.
- 19. Bjelkmar P, Larsson P, Cuendet MA, Hess B, Lindahl E: Implementation of the CHARMM force field in GROMACS: analysis of protein stability effects from correction maps, virtual interaction sites, and water models. Journal of chemical theory and computation 2010, 6(2):459–466%@ 1549–9618. pmid:26617301
- 20. Zoete V, Cuendet MA, Grosdidier A, Michielin O: SwissParam: a fast force field generation tool for small organic molecules. Journal of computational chemistry 2011, 32(11):2359–2368%@ 0192–8651. pmid:21541964
- 21. Mark P, Nilsson L: Structure and dynamics of the TIP3P, SPC, and SPC/E water models at 298 K. The Journal of Physical Chemistry A 2001, 105(43):9954–9960%@ 1089–5639.
- 22. Mandal SP, Garg A, Prabitha P, Wadhwani AD, Adhikary L, Kumar BR: Novel glitazones as PPARγ agonists: molecular design, synthesis, glucose uptake activity and 3D QSAR studies. Chemistry Central Journal 2018, 12(1):1–21%@ 1752–1153X.
- 23. Itteboina R, Ballu S, Sivan SK, Manga V: Molecular modeling-driven approach for identification of Janus kinase 1 inhibitors through 3D-QSAR, docking and molecular dynamics simulations. Journal of Receptors and Signal Transduction 2017, 37(5):453–469. pmid:28537140
- 24. Sánchez R, Šali A: Comparative protein structure modeling in genomics. Journal of Computational Physics 1999, 151(1):388–401%@ 0021–9991.
- 25. Bastidas AC, Deal MS, Steichen JM, Guo Y, Wu J, Taylor SS: Phosphoryl transfer by protein kinase A is captured in a crystal lattice. Journal of the American Chemical Society 2013, 135(12):4788–4798%@ 0002–7863.
- 26. Prochaska HJ, Santamaria AB: Direct measurement of NAD (P) H: quinone reductase from cells cultured in microtiter wells: a screening assay for anticarcinogenic enzyme inducers. Analytical biochemistry 1988, 169(2):328–336%@ 0003–2697.
- 27. Shen J, Rasmussen M, Dong Q-R, Tepel M, Scholze A: Expression of the NRF2 target gene NQO1 is enhanced in mononuclear cells in human chronic kidney disease. Oxidative medicine and cellular longevity 2017, 2017%@ 1942–0900. pmid:28785379
- 28. Beeraka NM, Bovilla VR, Doreswamy SH, Puttalingaiah S, Srinivasan A, Madhunapantula SV: The taming of nuclear factor erythroid-2-related factor-2 (Nrf2) deglycation by fructosamine-3-kinase (FN3K)-inhibitors-a novel strategy to combat cancers. Cancers 2021, 13(2):281. pmid:33466626
- 29. Shrestha S, Katiyar S, Sanz-Rodriguez CE, Kemppinen NR, Kim HW, Kadirvelraj R, et al.: A novel redox-active switch in Fructosamine-3-kinases expands the regulatory repertoire of the protein kinase superfamily. bioRxiv 2020.
- 30. Kornev AP, Haste NM, Taylor SS, Ten Eyck LF: Surface comparison of active and inactive protein kinases identifies a conserved activation mechanism. Proceedings of the National Academy of Sciences 2006, 103(47):17783–17788%@ 10027–18424. pmid:17095602
- 31. Kornev AP, Taylor SS, Ten Eyck LF: A helix scaffold for the assembly of active protein kinases. Proceedings of the National Academy of Sciences 2008, 105(38):14377–14382%@ 10027–18424. pmid:18787129
- 32. Van Linden OPJ, Kooistra AJ, Leurs R, De Esch IJP, De Graaf C: KLIFS: a knowledge-based structural database to navigate kinase–ligand interaction space. Journal of medicinal chemistry 2014, 57(2):249–277%@ 0022–2623. pmid:23941661
- 33. Liao JJ-L: Molecular recognition of protein kinase binding pockets for design of potent and selective kinase inhibitors. Journal of medicinal chemistry 2007, 50(3):409–424%@ 0022–2623. pmid:17266192
- 34.
Kooistra AJ, Volkamer A: Kinase-centric computational drug development. In: Annual Reports in Medicinal Chemistry. Volume 50, edn.: Elsevier; 2017: 197–236%@ 0065–7743.
- 35. Fischer PM: Approved and experimental small‐molecule oncology kinase inhibitor drugs: a mid‐2016 overview. Medicinal research reviews 2017, 37(2):314–367%@ 0198–6325. pmid:27775829
- 36. Bajusz D, G Ferenczy G, M Keseru G: Structure-based virtual screening approaches in kinase-directed drug discovery. Current topics in medicinal chemistry 2017, 17(20):2235–2259%@ 1568–0266. pmid:28240180
- 37. Wu P, Nielsen TE, Clausen MH: FDA-approved small-molecule kinase inhibitors. Trends in pharmacological sciences 2015, 36(7):422–439%@ 0165–6147.
- 38. Roskoski R Jr: Cyclin-dependent protein serine/threonine kinase inhibitors as anticancer drugs. Pharmacological research 2019, 139:471–488%@ 1043–6618. pmid:30508677
- 39. Roskoski R Jr: Cyclin-dependent protein kinase inhibitors including palbociclib as anticancer drugs. Pharmacological research 2016, 107:249–275%@ 1043–6618. pmid:26995305
- 40. Roskoski R Jr: The ErbB/HER family of protein-tyrosine kinases and cancer. Pharmacological research 2014, 79:34–74%@ 1043–6618. pmid:24269963
- 41. Roskoski R Jr: ErbB/HER protein-tyrosine kinases: Structures and small molecule inhibitors. Pharmacological research 2014, 87:42–59%@ 1043–6618. pmid:24928736
- 42. Roskoski R Jr: Small molecule inhibitors targeting the EGFR/ErbB family of protein-tyrosine kinases in human cancers. Pharmacological research 2019, 139:395–411%@ 1043–6618. pmid:30500458
- 43. Roskoski R Jr: Allosteric MEK1/2 inhibitors including cobimetanib and trametinib in the treatment of cutaneous melanomas. Pharmacological Research 2017, 117:20–31%@ 1043–6618. pmid:27956260
- 44. Roskoski R Jr: Vascular endothelial growth factor (VEGF) and VEGF receptor inhibitors in the treatment of renal cell carcinomas. Pharmacological Research 2017, 120:116–132%@ 1043–6618. pmid:28330784
- 45. Fumarola C, Caffarra C, La Monica S, Galetti M, Alfieri RR, Cavazzoni A, et al.: Effects of sorafenib on energy metabolism in breast cancer cells: role of AMPK–mTORC1 signaling. Breast cancer research and treatment 2013, 141(1):67–78%@ 1573–7217. pmid:23963659
- 46. Wilhelm S, Carter C, Lynch M, Lowinger T, Dumas J, Smith RA, et al.: Discovery and development of sorafenib: a multikinase inhibitor for treating cancer. Nature reviews Drug discovery 2006, 5(10):835–844%@ 1474–1784. pmid:17016424
- 47. Sun X, Ou Z, Chen R, Niu X, Chen D, Kang R, et al.: Activation of the p62‐Keap1‐NRF2 pathway protects against ferroptosis in hepatocellular carcinoma cells. Hepatology 2016, 63(1):173–184%@ 0270–9139. pmid:26403645
- 48. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, et al.: Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. New England journal of medicine 2001, 344(11):783–792%@ 0028–4793.
- 49. Arribas J, Baselga J, Pedersen K, Parra-Palau JL: p95HER2 and breast cancer. Cancer research 2011, 71(5):1515–1519%@ 0008–5472. pmid:21343397
- 50. Ryan Q, Ibrahim A, Cohen MH, Johnson J, Ko C-w, Sridhara R, et al.: FDA drug approval summary: lapatinib in combination with capecitabine for previously treated metastatic breast cancer that overexpresses HER-2. The oncologist 2008, 13(10):1114–1119%@ 1083–7159. pmid:18849320
- 51. Schwartzberg LS, Franco SX, Florance A, O’Rourke L, Maltzman J, Johnston S: Lapatinib plus letrozole as first-line therapy for HER-2+ hormone receptor–positive metastatic breast cancer. The oncologist 2010, 15(2):122–129%@ 1083–7159. pmid:20156908
- 52. Luo P, Yan H, Du J, Chen X, Shao J, Zhang Y, et al.: PLK1 (polo like kinase 1)-dependent autophagy facilitates gefitinib-induced hepatotoxicity by degrading COX6A1 (cytochrome c oxidase subunit 6A1). Autophagy 2021, 17(10):3221–3237%@ 1554–8627. pmid:33315519
- 53. Crunkhorn S: Kinase inhibitor combination combats liver cancer. Nature reviews Drug Discovery %@ 1474–1776 2021. pmid:34344998
- 54. Im JH, Kang KW, Kim SY, Kim YG, An YJ, Park S, et al.: GPR119 agonist enhances gefitinib responsiveness through lactate-mediated inhibition of autophagy. Journal of Experimental & Clinical Cancer Research 2018, 37(1):1–15%@ 1756–9966. pmid:30497501
- 55. Corkery B, Crown J, Clynes M, O’Donovan N: Epidermal growth factor receptor as a potential therapeutic target in triple-negative breast cancer. Annals of Oncology 2009, 20(5):862–867%@ 0923–7534. pmid:19150933
- 56. Vafeiadou V, Hany D, Picard D: Hyperactivation of MAPK Induces Tamoxifen Resistance in SPRED2-Deficient ERα-Positive Breast Cancer. Cancers 2022, 14(4):954%@ 2072–6694.
- 57. Tuğrul B, İşseven M: The effect of DEBIO 1143 usage alone or in combination with Tamoxifen on estrogen receptor positive breast cancer cell lines. Pamukkale Tıp Dergisi 2020, 13(1):9–18.
- 58. Fu Y, Wang Z, Luo C, Wang Y, Wang Y, Zhong X, et al.: Downregulation of CXXC finger protein 4 leads to a tamoxifen-resistant phenotype in breast cancer cells through activation of the Wnt/β-catenin pathway. Translational oncology 2020, 13(2):423–440%@ 1936–5233.
- 59. Pawlik A: Słomi nska-Wojewódzka M; Herman-Antosiewicz A. Sensitization of estrogen receptor-positive breast cancer cell lines to 4-hydroxytamoxifen by isothiocyanates present in cruciferous plants. Eur J Nutr 2016, 55(3):1165–1180. pmid:26014809
- 60. Kim SK, Yang JW, Kim MR, Roh SH, Kim HG, Lee KY, et al.: Increased expression of Nrf2/ARE-dependent anti-oxidant proteins in tamoxifen-resistant breast cancer cells. Free Radical Biology and Medicine 2008, 45(4):537–546%@ 0891–5849. pmid:18539158
- 61. Jiang K, He B, Lai L, Chen Q, Liu Y, Guo Q, et al.: Cyclosporine A inhibits breast cancer cell growth by downregulating the expression of pyruvate kinase subtype M2. International journal of molecular medicine 2012, 30(2):302–308%@ 1107–3756. pmid:22580449
- 62. Yao Y, Li X, Cheng L, Wu X, Wu B: Circular RNA FAT atypical cadherin 1 (circFAT1)/microRNA-525-5p/spindle and kinetochore-associated complex subunit 1 (SKA1) axis regulates oxaliplatin resistance in breast cancer by activating the notch and Wnt signaling pathway. Bioengineered 2021, 12(1):4032–4043%@ 2165–5979. pmid:34288822
- 63. Buccioni M, Dal Ben D, Lambertucci C, Maggi F, Papa F, Thomas A, et al.: Antiproliferative evaluation of isofuranodiene on breast and prostate cancer cell lines. The Scientific World Journal 2014, 2014%@ 2356–6140. pmid:24967427
- 64. Li Q, Shu Y: Pharmacological modulation of cytotoxicity and cellular uptake of anti-cancer drugs by PDE5 inhibitors in lung cancer cells. Pharmaceutical research 2014, 31(1):86–96%@ 1573–1904X. pmid:23884568
- 65. Serova M, Calvo F, Lokiec F, Koeppel F, Poindessous V, Larsen AK, et al.: Characterizations of irofulven cytotoxicity in combination with cisplatin and oxaliplatin in human colon, breast, and ovarian cancer cells. Cancer chemotherapy and pharmacology 2006, 57(4):491–499%@ 1432–0843. pmid:16075278
- 66. Wang X-J, Sun Z, Villeneuve NF, Zhang S, Zhao F, Li Y, et al.: Nrf2 enhances resistance of cancer cells to chemotherapeutic drugs, the dark side of Nrf2. Carcinogenesis 2008, 29(6):1235–1243%@ 1460–2180. pmid:18413364
- 67. Garufi A, Pistritto G, D’Orazi V, Cirone M, D’Orazi G: The Impact of NRF2 Inhibition on Drug-Induced Colon Cancer Cell Death and p53 Activity: A Pilot Study. Biomolecules 2022, 12(3):461%@ 2218-2273X. pmid:35327653
- 68. Payandeh Z, Tazehkand AP, Mansoori B, Khaze V, Asadi M, Baradaran B, et al.: The Impact of Nrf2 Silencing on Nrf2-PD-L1 Axis to Overcome Oxaliplatin Resistance and Migration in Colon Cancer Cells. Avicenna Journal of Medical Biotechnology 2021, 13(3):116. pmid:34484640
- 69. Colleoni M, Rocca A, Sandri MT, Zorzino L, Masci G, Nole F, et al.: Low-dose oral methotrexate and cyclophosphamide in metastatic breast cancer: antitumor activity and correlation with vascular endothelial growth factor levels. Annals of Oncology 2002, 13(1):73–80%@ 0923–7534. pmid:11863115
- 70. Zimmerman MC, Clemens DL, Duryee MJ, Sarmiento C, Chiou A, Hunter CD, et al.: Direct antioxidant properties of methotrexate: inhibition of malondialdehyde-acetaldehyde-protein adduct formation and superoxide scavenging. Redox Biol 13: 588–593. In.; 2017. pmid:28803127
- 71. Tian B, Lu Z-N, Guo X-L: Regulation and role of nuclear factor-E2-related factor 2 (Nrf2) in multidrug resistance of hepatocellular carcinoma. Chemico-Biological Interactions 2018, 280:70–76%@ 0009–2797. pmid:29223570
- 72. Wang M, Shi G, Bian C, Nisar MF, Guo Y, Wu Y, et al.: UVA irradiation enhances brusatol-mediated inhibition of melanoma growth by downregulation of the Nrf2-mediated antioxidant response. Oxidative medicine and cellular longevity 2018, 2018%@ 1942–0900. pmid:29670684
- 73. Sinha BK, Tokar EJ, Bushel PR: Elucidation of mechanisms of topotecan-induced cell death in human breast MCF-7 cancer cells by gene expression analysis. Frontiers in Genetics 2020:775%@ 1664–8021. pmid:32765594
- 74. Lu J, Chen J, Xu N, Wu J, Kang Y, Shen T, et al.: Activation of AIFM2 enhances apoptosis of human lung cancer cells undergoing toxicological stress. Toxicology letters 2016, 258:227–236. pmid:27392435
- 75. Davis M, Li J, Knight E, Eldridge SR, Daniels KK, Bushel PR: Toxicogenomics profiling of bone marrow from rats treated with topotecan in combination with oxaliplatin: a mechanistic strategy to inform combination toxicity. Frontiers in genetics 2015, 6:14. pmid:25729387
- 76. Barazorda-Ccahuana HL, Valencia DE, Aguilar-Pineda JA, Gómez B: Art v 4 protein structure as a representative template for allergen profilins: Homology modeling and molecular dynamics. ACS Omega 2018, 3(12):17254–17260.
- 77. Yuan M, Luo M, Song Y, Xu Q, Wang X, Cao Y, et al.: Identification of curcumin derivatives as human glyoxalase I inhibitors: A combination of biological evaluation, molecular docking, 3D-QSAR and molecular dynamics simulation studies. Bioorganic & medicinal chemistry 2011, 19(3):1189–1196. pmid:21237663