Abstract
Background/Aim: The proinflammatory protein S100A8, which is expressed in myeloid cells under physiological conditions, is strongly expressed in human cancer tissues. Its role in tumor cell differentiation and tumor progression is largely unclear and virtually unstudied in kidney cancer. In the present study, we investigated whether S100A8 could be a potential anticancer drug target and therapeutic biomarker for kidney cancer, and the underlying molecular mechanisms by exploiting its interaction profile with drugs. Materials and Methods: Microarray-based transcriptomics experiments using Affymetrix HuGene 1.0 ST arrays were applied to renal cell carcinoma specimens from Saudi patients for identification of significant genes associated with kidney cancer. In addition, we retrieved selected expression data from the National Center for Biotechnology Information Gene Expression Omnibus database for comparative analysis and confirmation of S100A8 expression. Ingenuity Pathway Analysis (IPA) was used to elucidate significant molecular networks and pathways associated with kidney cancer. The probable polar and non-polar interactions of possible S100A8 inhibitors (aspirin, celecoxib, dexamethasone and diclofenac) were examined by performing molecular docking and binding free energy calculations. Detailed analysis of bound structures and their binding free energies was carried out for S100A8, its known partner (S100A9), and S100A8–S100A9 complex (calprotectin). Results: In our microarray experiments, we identified 1,335 significantly differentially expressed genes, including S100A8, in kidney cancer using a cut-off of p<0.05 and fold-change of 2. Functional analysis of kidney cancer-associated genes showed overexpression of genes involved in cell-cycle progression, DNA repair, cell death, tumor morphology and tissue development. Pathway analysis showed significant disruption of pathways of atherosclerosis signaling, liver X receptor/retinoid X receptor (LXR/RXR) activation, notch signaling, and interleukin-12 (IL-12) signaling. We identified S100A8 as a prospective biomarker for kidney cancer and in silico analysis showed that aspirin, celecoxib, dexamethasone and diclofenac binds to S100A8 and may inhibit downstream signaling in kidney cancer. Conclusion: The present study provides an initial overview of differentially expressed genes in kidney cancer of Saudi Arabian patients using whole-transcript, high-density expression arrays. Our analysis suggests distinct transcriptomic signatures, with significantly high levels of S100A8, and underlying molecular mechanisms contributing to kidney cancer progression. Our docking-based findings shed insight into S100A8 protein as an attractive anticancer target for therapeutic intervention in kidney cancer. To our knowledge, this is the first structure-based docking study for the selected protein targets using the chosen ligands.
Cancer is a group of diseases caused by dysregulation in molecular signaling pathways due to over- and underexpression and loss/gain of function mutations of key proteins associated with altered cell growth and cell-cycle progression. It develops via a multi-step process of initiation, promotion, and progression (1, 2). Among all types of urological cancers, kidney cancer is the second leading cause of death in adults, mainly due to lack of promising biomarkers for effective target therapy. We conducted transcriptomic profiling and functional pathway analysis to study the role of the S100A8 protein in tumor cells and in cancer development. Docking study revealed the potential of S100A8 as a target of therapeutic importance.
S100A8 is a low-molecular-weight proinflammatory protein of 10 kDa, belonging to the S100 family of Ca2+-binding elongation factor (EF) hand-type proteins constitutively expressed by myeloid cells, such as neutrophils and activated monocytes, under physiological conditions (3-6). However, increased expression is seen in epithelial cells under pathological conditions, including inflammation and wound healing (7). S100A8 is an essential gene for life since S100A8 knock-out mice die during embryonic development (8). Enhanced expression of S100A8 is one of the hallmarks of chronic inflammation and epithelial cancer. Overexpression of S100A8 is found in various types of carcinomas, including breast (9-10), prostate (11-12), lung (13), gastric (14), hepatic (15), pancreatic (16) and colorectal (17) cancer. Nevertheless, little research has been carried out on its expression in different types of tumor cells or its correlation with cancer development; the differential expression pattern or role of S100A8 in progression of kidney cancer has not been reported as far as we are aware.
Several lines of evidence point to vital functions of S100A8 during tumorigenesis and, although its exact role within the tumor environment is still not understood, different tumor-promoting effects have been proposed. S100A8 preferentially forms heterodimeric complexes with S100A9 (S100A8–S100A9) (18), which undergoes conformational changes upon Ca2+ binding and functions as a sensor of intracellular Ca2+ (19). Extracellularly, S100A8–S100A9 acts as ligand for different receptors, including the receptor for advanced glycation end products (RAGE) (10, 20), toll like receptor-4 (TLR4) (21), CD36 receptor (22) and NADPH oxidase (23). S100A8 exerts potent pro-inflammatory activity (24-26), attracts neutrophils (27), influences myeloid cell differentiation (28-29), affects transendothelial migration of phagocytes (30) and induces expression of pro-inflammatory mediators (31). Studies suggest S100A8 to be an important driver of the inflammatory environment, ultimately promoting cancer progression (10, 15). A recent study showed a role of S100A8 at low concentrations in cell growth-promoting activity by binding to RAGE (9); however, its direct role in tumor cells and tumor progression is ambiguous. It has been demonstrated that primary tumors secrete soluble factors, including vascular endothelial growth factor-A (VEGF-A), transforming growth factor-beta (TGF-β) and tumor necrosis factor alpha (TNFα), which induce expression of S100A8 and S100A9 in myeloid and endothelial cells prior to tumor metastasis (13). S100A8 also increases the motility of circulating cancer cells by p38 mitogen activated protein kinase (MAPK)-mediated activation of tumor cells (32). S100A8 many exist as monomer, homodimer, heterodimer or heterotetramer depending on the presence and concentration of free calcium molecules. S100A8 and S100A9 share strong sequence homology and can form heterodimers (without calcium) or heterotetramers (with calcium) (18). Three-dimensional crystallographic analysis reveals that calcium-bound C-terminal EF-hand loops are necessary for tetramerization (33). Overall, this indicates that the S100A8, S100A9 or S100A8–S100A9 complex could be targeted to prevent the tumor cell migration and growth.
Molecular docking gives an optimized conformation and relative orientation for both the protein and ligand molecule such that the free energy of the overall bound system is minimal. Non-selective non-steroidal anti-inflammatory drugs (NSAID) such as aspirin, diclofenac, indomethacin, ibuprofen and naproxen inhibit both cyclooxygenase, COX-1 and COX2 but can lead to drastic side-effects such as gastric ulceration. However, selective NSAIDs such as celecoxib (Celebrex®) are much safer and only inhibit COX2 found at sites of inflammation, more than that which is normally found in the stomach, blood platelets, and blood vessels (COX1).
Aspirin, a prototypical analgesic, is very commonly administered for the treatment of mild to moderate pain. It has anti-inflammatory and anti-pyretic properties and acts as an inhibitor of cyclooxygenase (both COX1 and COX2), which results in the inhibition of prostaglandin biosynthesis. It also inhibits platelet aggregation and is used in the prevention of arterial and venous thrombosis. Daily aspirin intake has been shown to be beneficial in treatment of cancer and lowering the risk of cancer development (34-37). The molecular mechanisms involved in anticancer action of aspirin have not yet been elucidated.
Dexamethasone, is a synthetic adrenocortical steroid (analog of glucocorticoids) primarily used for its anti-inflammatory effects in disorders of many organ systems, and in immunomodulation, as it modifies the body's immune responses to diverse stimuli. It is an NSAID with anti-pyretic and analgesic actions used primarily in the treatment of chronic arthritic conditions and certain soft tissue disorders associated with pain and inflammation. It acts by blocking the synthesis of prostaglandins by inhibiting COX, which converts arachidonic acid to cyclic endoperoxides, precursors of prostaglandins. Inhibition of prostaglandin synthesis accounts for its analgesic, antipyretic, and platelet-inhibitory actions; other mechanisms may also contribute to its anti-inflammatory effects.
Diclofenac is another very common NSAID used as an analgesic, anti-inflammatory and anti-pyretic agent. It is often applied to treat chronic pain linked with cancer, in particular inflammation-associated pain (38). Apart from primarily inhibiting COX, evidence indicates that it also inhibits phospholipase A2 and the lipoxygenase pathways, thus reducing formation of leukotrienes.
We carried out a series of molecular docking analyses using aspirin, celecoxib, dexamethasone and diclofenac as inhibitors for S100A8, S100A9, S100A8–S100A9, respectively.
Materials and Methods
Patients and samples. The study was performed on samples from patients of the Kingdom of Saudi Arabia, diagnosed with renal cell carcinoma. The samples were collected from the King Abdulaziz University Hospital, Bakhsh Hospital and King Faisal Specialist Hospital and Research Center, Jeddah, Sandi Arabia, during 2010-2012. Out of 18 received cases, only two passed the criteria to be used for array expression analysis. One patient was a 61-year-old male, diagnosed with clear cell renal cell carcinoma of nuclear grade II and tumor size 4.5×3×4 cm. The other patient was a 47-year-old female, diagnosed with chromophobic renal cell carcinoma of Fuhrman's grade II. For gene expression analysis, fresh tumor tissue specimens were obtained from surgical resections adjacent to the sites on which final histological diagnosis was performed. Fresh normal kidney specimens were derived from surgically-resected normal kidney tissues. All collected tissue specimens were immediately placed in RNALater (Invitrogen Life Technologies, Grand Island, NY, USA) or RPMI_1640 medium (GIBCO BRL, Grand Island, NY USA).
Ethical approval. All patients included in the study provided written informed consent. The study was reviewed and approved by the Center of Excellence in Genomic Medicine Research (CEGMR) Ethical Committee (08-CEGMR-02-ETH).
RNA extraction and array processing. Total RNA was extracted from freshly-preserved kidney tissue specimens with the Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany) including an on-column DNAse treatment according to the manufacturer's recommendations. Quality of the purified RNA was verified on an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). The mean RNA integrity number (RIN) for processed samples was 8.0 only for two samples. Hence, the remaining were excluded from further expression studies. RNA concentrations were determined using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Two hundred and fifty nanograms of each RNA sample were processed according to the manufacturer's recommendations (Life Technology). After fragmentation and labeling, the samples were hybridized at 45°C for 17 h to Human Gene 1.0 ST GeneChip arrays (Affymetrix, Santa Clara, CA, USA). These arrays are conceptually based on the human genome sequence assembly UCSC hg18, NCBI Build 36 and interrogated with a set of 764,885 probes, 28,869 annotated genes.
Gene expression analysis. We conducted expression profiling on a small group of samples (two renal cell carcinomas and four normal kidney tissues), which requires confirmation by other studies. Therefore, selected kidney cancer expression data (GSE781, GSE7023, and GSE6344) were retrieved from the gene expression omnibus database for comparative meta-analysis and our own experimental findings were verified with a larger independent dataset available in the public domain (Table I). Affymetrix. CEL files were imported to Partek Genomics Suite version 6.6 (Partek Inc., MO, USA). The data were normalized using random multiple access (RMA) algorithm normalization. Principal component analysis (PCA) was performed on all probes to visualize high-dimensional data with multiple clinicopathological properties. PCA was used to assess quality control, as well as overall variance in gene expression between the diseased states. Analysis of variance (ANOVA) was applied on the complete data set and the list of differentially expressed genes was then generated using a false-discovery rate (FDR) of 0.05 with 2-fold change cut-off. Unsupervised two-dimensional average linkage hierarchical clustering was performed using Spearman's correlation as a similarity matrix.
Functional and pathway analysis. To define biological networks, interaction and functional analysis among the differentially regulated genes in kidney cancer, pathway analyses were performed using Ingenuity Pathways Analysis software (IPA) (Ingenuity Systems, Redwood City, CA, USA). Statistically differentially expressed datasets containing 1,335 genes and their corresponding probesets ID, gene symbol, Entrez gene ID as clone identifier, p-value and fold change values were uploaded into IPA. The functional/pathway analysis of IPA identifies the biological functions, diseases and pathways that are most significantly altered for the differentially expressed gene set. The significance of the relation between the expression data and the canonical pathways were calculated by ratio or Fisher's exact test.
Molecular docking. To investigate the role of S100A8 and its partner, S100A9, as inflammatory mediators and to establish their involvement in inflammation-associated cancers at molecular level, we carried out docking studies. The molecular structure of aspirin, celecoxib, dexamethasone and diclofenac were retrieved from PubChem compound database with CID 2244, 2662, 5743 and 3033 respectively (Figure 1). Protein targets taken for docking studies were S100A8 homodimer, S100A9 homodimer and S100A8–S100A9 heterodimer complex. The crystal structure of identified cancer signaling target proteins were retrieved from Protein Data Bank (PDBID: 1MR8, 1IRJ, and 4GGF). Structure visualization was profound using PyMol (39) (Figure 2).
Docking calculations were carried-out using Molecular Docking Server (40). The Merck molecular force field 94 (MMFF94) (41) was used for energy minimization of ligand molecules: aspirin, celecoxib, dexamethasone and diclofenac. Gasteiger partial charges were added to the ligand atoms. Non-polar hydrogen atoms were merged, and rotatable bonds were defined. Molecular docking of each ligand was performed individually with (S100A8)2 homodimer (1MR8), (S100A9)2 homodimer (1IRJ, in the presence and absence of prior existing ligand 3-[(3-cholamidopropyl) dimethylammonio]-1-propane-sulfonate, CPS, also known as CHAPS) and (S100A8–S100A9)2 hetrotetramer (4GGF) protein models to predict the binding orientation and interaction. Essential hydrogen atoms, Kollman united atom type charges, and solvation parameters were added with the aid of AutoDock tools (42). Affinity (grid) maps of 20×20×20 Å grid points and 0.375 Å binding site grid generation spacing were generated using the Autogrid program (43). AutoDock parameter set- and distance-dependent dielectric functions were used in the calculation of the van der Waals' and the electrostatic terms, respectively. Docking simulations were performed using the Lamarckian genetic algorithm (LGA) and the Solis and Wets local search method (44). Initial positions, orientation, and torsions of the ligand molecules were set randomly. Each docking experiment was derived from 10 different runs that were set to terminate after a maximum of 250000 energy evaluations. The population size was set to 150. During the search, a translational step of 0.2 Å, and quaternion and torsion steps of 5 were applied in the current series of docking analysis.
Results
The main focus of the present study was to discover novel anticancer drug targets by transcriptomic profiling and to identify possible protein–drug interactions by molecular docking analysis. We identified S100A8 as an important protein in kidney cancer and attempted to demonstrate its potential as an anticancer drug target.
Identification of differentially expressed genes. We profiled fresh kidney tissue specimens and compared them with normal control samples. We performed PCA scatter plot analysis for visualizing the high-dimensional array data where each point represents a chip or sample. The results of PCA of transcriptomic data showed that the samples from the same tissue type clustered tightly together. Clear differences were also observed between tumors and normal tissues, revealing distinct expression profiles for each tissue type. Comparison of the genome-wide expression of kidney cancer revealed 1,335 differentially expressed genes: 852 up-regulated and 483 down-regulated, with a 2-fold or greater change and FDA of p<0.05 (Figure 3, Table II).
Pathways and networks underlying kidney cancer. To understand the mechanisms by which the genes alter a wide range of physiological processes, we examined biological functions, molecular network and pathways associated with kidney cancer. Interestingly, cellular movement was significantly over-represented as a process for both down-regulated and up-regulated genes, indicating that metastasis is probably linked to a different equilibrium of switching on and off i.e. cell division cascade is on while tumor suppressor cascade is off. Functional analysis of kidney cancer-associated genes showed an overexpression of genes involved in cell-cycle progression, DNA repair, cell death, tumor morphology and tissue development. Pathway analysis showed significant disruption in certain signaling pathways, including atherosclerosis signaling, liver X receptor/retinoid X receptor activation, Interleukin-12 signaling and production in macrophages, production of nitric oxide and reactive oxygen species in macrophages, notch signaling, and clathrin-mediated endocytosis signaling (Figure 4, Table III). Extensive pathway analysis of differentially regulated genes may provide novel hypotheses underlying tumor invasion and metastatic progression of kidney cancer.
Docking studies. Molecular docking studies predicted potential interactions of our proposed protein drug target with the selected drug molecules. As far as we are aware of, this is the first structural attempt to study possible binding for cancer therapeutics. To understand the molecular interaction between drugs and S100A8, a series of molecular docking analyses were performed using three-dimensional structures available (PDBID: 1IRJ, 1MR8, and 1XK4) with four known anti-inflammatory drugs namely aspirin, celecoxib, dexamethasone and diclofenac. Molecular docking revealed that all four drugs are able to bind in the ligand-binding domain of these S100 proteins. The ligand-binding site was a hinge region containing two EF-hand motifs. Based on their size, stereochemistry and structural differences, the ligands exhibited different intensities in binding with the protein target molecules. The predicted parameters of estimated binding free energy, inhibition constant (Ki), total energy of van der Waals', hydrogen bonding, desolvation, electrostatic energy, total intermolecular energy and interacting surface area were evaluated to estimate the favorable binding of ligand drug molecules to the target protein. Molecular visualization was performed using PyMol. Complete interaction profiles (hydrogen bonds, polar, hydrophobic, pi–pi, cation–pi and other contacts), and hydrogen bonding interactions (HB plot) were also studied (Figure 5 and 6, Table IV).
(S100A8)2: 1MR8 (chain A and B). Aspirin binds at the ligand binding site without forming any H-bonds but displayed two polar contacts with the residue Gln 69 (A) and several hydrophobic and van der Waals' interactions mainly with the hydrophobic residues Leu 72 (A), Ile 73 (A), Ile 76 (B) and also with Gln 69 (A). Predicted free energy of binding is −3.26 kcal/mol.
Celecoxib binds to this protein molecule in a much better way compared to aspirin, with Ki of 960.05 μM. It is able to form H-bonds with a critical amino acid residue in the ligand binding domain site of S100A8 i.e. Gln 69 (A), which also forms three more polar contacts with the chosen ligand. More than two dozen hydrophobic and other interactions are also predicted to be mediated by Gln 69 (A), Leu 72 (A), Ile 73 (A), Ile 76 (B) and also by Val 80 (B). Interestingly, the three fluorine residues (H-bond acceptors) of celecoxib orient and form halogen bonds with Gln 69 (A).
The binding ability of dexamethasone is limited as there are no H-bonds formed and the free energy of the bound structure is −2.02 kcal/mol. However, there are around 17 hydrophobic interactions with the aliphatic non-polar hydrophobic residues of chain A Ile 73, Ile 76 and Val 80. Other contacts are mediated by polar residues Lys 77 (A) and Gln 69 (B) and the remaining hydrophobic amino acids.
Diclofenac shows good binding characteristics with the S100A8 dimer but forms neither H-bonds nor polar contacts. Estimated free energy of binding is predicted to be −4.32 kcal/mol. It mainly shows hydrophobic interactions: two with Leu 72 (B), one with Ile 73 (A), 12 with Ile 76 (A) and three with Val 80 (A). The remaining protein residues involved in other contacts are Ile 76 (A), Lys 77 (A) and Val 80 (A). All of these residues are at a maximum distance 3.1-3.8Å.
(S100A9)2: 1IRJ (chain A and B). In the docked structure there were no H-bonds but there were two hydrophobic interactions with Ile 16 (B) amongst other attractive contacts. We had not removed the present CHAPS (CPS) molecule from the crystal structure taken. Aspirin owing to its small shape and size was able to fit into the binding groove. Interestingly, aspirin binds to the S100A9 dimer at a totally new site different from the cavity where CPS is present. The drug-binding cavity was lined mainly by hydrophobic residues: Ile 16, His 20 (B), Phe 76, Ala 84 (A) and Leu 86. Glu 77 also exhibited indirect interactions. Attractive stacking pi–pi interactions between aromatic protein residues present, i.e. His 20 and Phe 76, were also noted.
Docking study in the presence of CPS and the docked structure has a positive free energy of binding, implying that binding is not feasible as most of the decomposed interaction energies are positive. Enhanced docking properties can possibly be achieved by increasing the simulation box size and making the ligand-binding domain free of any pre-existing bound ligand such as CPS in this case.
Similarly to the above case, docking was carried out in the presence of CPS and the docked structure has a positive value for free energy of binding, indicating that binding is not feasible. Better docking results are expected perhaps by increasing the simulation box size and making the ligand-binding domain lined by amino acids Thr 68, Glu 77, Met 81 and Arg 85 free of any pre-existing bound ligand. Hence, Ki cannot be estimated in this case.
The binding of diclofenac molecule was stabilized with the help of three H-bonds, one with the residue Thr 68 (B) and two with Arg 85 (B), one polar contact with Thr 68 (B), one halogen bond, and few hydrophobic interactions (van der Waals' contacts) primarily with Met 81 (B) (distances upto 3.7Å) and Glu 77 (A). It seems that diclofenac would bind at a site adjacent to the aspirin-binding cavity (not shown).
Aspirin binds with a low estimated free energy of binding of −3.20 kcal/mol with chain B and there is no H-bond formation between the drug and target protein molecule. The drug-binding site seems to be different from that of the above case in the presence of CPS and aspirin binds at the binding site previously occupied by CPS. Five hydrophobic interactions can be seen with Ile 62 and one with Leu 82. In addition, around 12 other contacts are observed with Asp 65, His 61, Glu 52, Arg 85 and Val 58.
Celecoxib binds at the CPS-binding site of (S100A9)2 as shown in the figure made by using PyMol (Figure 6). There are two H-bonds between Phe 48 and Arg 85 belonging to chain B of the protein molecule at a distance of 3.35Å and 2.57Å. In addition, polar contacts (Arg 85), hydrophobic interactions (Leu 49 and Leu 86), van der Waals' and other interactions (Phe 48, Lys 51, Glu 52 and Leu 86) are noted.
Our results clearly demonstrate the binding ability of dexamethasone with critical amino acid residues of the ligand-binding domain of chain B only, potentiating inhibitory ability. The interacting residues of the drug-binding site are Leu 49, Glu 52, Val 58, His 61, Ile 62 and Leu 86. However, there are no H-bonds. Stacking attractive pi–pi interactions with His 61 and two carbon atoms of the drug are also found. Several polar, hydrophobic, van der Waals' and other contacts are seen in the docked complex. The estimated free energy of binding is predicted to be −5.51kcal/mol.
The diclofenac molecule is predicted to have good affinity and binding with the S100A9 dimer but without any H-bond formation. Around 15 hydrophobic interactions are seen with the non-polar residues lining the binding groove, i.e. Leu 49, Val 58, Ile 62, Leu 82 and Leu 86. Stacking pi–pi interactions with His 61 and the C14 atom of the drug are present. Asp 65 forms a halogen bond with Cl1 of diclofenac, with the distance between them being 3.49Å. Remaining weak interactions are found for Glu 52, Val 58, His 61, Asp 65 and Arg 85. The estimated inhibition constant Ki is 214.31 μM.
(S100A8–S100A9)2 (4GGF). Aspirin interacts only with the C and L chains of S100A9, exhibiting only polar contacts and hydrophobic interactions and no direct H-bonds. In total, three polar contacts are predicted, one with Glu 92 (C) and two with Arg 85 (L). Leu 82 and 86 create hydrophobic contacts and are at a distance of 3.81Å and 3.89Å, respectively.
Docking results for celecoxib are promising, with estimated free energy of binding of −3.01 kcal/mol, and pairwise decomposition of energies of interacting residues are zero. Similar to the case of aspirin binding, the ligand celecoxib is found in the binding site made by chains C and L of S100A9 and interacts with different residues present. It forms a total of six H-bonds: three with Glu 52 (L), two with Glu 92 (C) and one with Gly 97 (C). Several hydrophobic interactions are seen with the hydrophobic amino acids of chain L: Leu 49, Ile 62, Leu 82, and Leu 86. Six halogen bonds with fluorine residue of the ligand were seen, most of which were water-mediated. More than 12 other contacts were noted, formed mainly with the polar residues such as Arg 85 (L), Trp 88 (C), His 91 (C) and Glu 96 (C).
Dexamethasone binds very well at the site lined mainly by chain L (S100A9) residues and partly by chain K (S100A8). The docked structure does not show any H-bonding nor polar contacts. More than 12 hydrophobic interactions are possible, mediated by amino acids His 20 (L), Val 24 (L), Pro 29 (L) and His 87 (K). Other noticeable interactions are with Gln 21 and Ser 23.
Diclofenac binds best with the selected chains of the calprotectin complex structure taken for docking analysis and all the interacting residues belong to chain L (S100A9). It does not form H-bonds but Val 24 and Pro 29 are predicted to have good hydrophobic interactions, with a maximum distance of 3.7 Å. Interestingly, His 20 displays six stacking ring interactions and also one halogen bond. Two other contacts are predicted with Asn 17, His 20, Gln 21, Ser 23 and Val 24.
Discussion
Kidney cancer includes heterogeneous tumors with diverse molecular and clinical characteristics that is reflected in their response to specific treatments. In the present work, we identified S100A8 as a potential biomarker for kidney cancer and in silico analysis shows that aspirin, celecoxib, dexamethasone and diclofenac are predicted to bind to S100A8, presumably inhibiting downstream signaling in kidney cancer. Our finding leads to the hypothesis that S100A8 is a promosing anticancer drug target and aspirin, celecoxib, dexamethasone and diclofenac are S100A8 inhibitors.
S100 proteins participate in numerous functions including protein phosphorylation, enzymatic activation, calcium homeostasis, and interaction with cytoskeletal components (45). Most genes encoding S100 proteins are clustered on a region of human chromosome 1q21 that is prone to chromosomal re-arrangements, suggesting a link of S100A8 and S100A9 proteins with metastasis and tumor formation (45-46). Abnormal expression of S100 proteins, including S100A8 and S100A9, were observed in a variety of different cancer types, such as gastric, lung, breast, liver, pancreatic and squamous esophageal carcinomas (9-11, 14-16, 47-50). Despite elevated expression and the distinct role of S100A8 in different cancer types, less is known about the expression status or role of S100A8 in the progression of kidney cancer.
S100A8 has cell growth-promoting activity at low concentrations by binding to RAGE. In addition, RAGE binding to S100A8–S100A9 promotes phosphorylation of LIM (cell lineage protein 11 (Lin-11), islet-1 (ISL1), Mechanosensory protein 3 (MEC-3)), domain kinase. This phosphorylation is a critical step in cofilin recycling and actin polymerization. Interestingly, RAGE binding to S100A8–S100A9 enhanced cell mesenchymal properties and induced epithelial–mesenchymal transition. Moreover, RAGE binding to S100A8–S1009 played an important role in promoting invasion and metastasis in cancer (9, 16). These studies indicate the potential of S100A8 as an anticancer target.
In our docking study of the S100A8 dimer with the four selected drugs, the best overall binding was exhibited in terms of estimated free energy of binding and Ki value by diclofenac followed by celecoxib and aspirin, and the least by dexamethasone. Binding of aspirin with S100A9 dimer does not show much difference in terms of energy and Ki in the presence or absence of CPS. S100A9 dimer also seems to have multiple ligand-binding areas as exhibited by the different ligand-binding sites for aspirin and diclofenac in the presence, as well as absence, of already bound ligand, CPS. However the CPS-binding cavity is quite hydrophobic and is the most preferable binding site for compounds having compatible shape and stereochemistry. Aspirin does not form H-bonds in any of the docked structures but it displayed the best binding capability with calprotectin. The binding results for dexamethasone and diclofenac are similar with calprotectin complex and neither of them forms H-bonds with the protein target. Even though celecoxib is able to form six H-bonds with S100A8–S100A9 heterotetramer, its estimated free energy of binding is comparatively low.
This was a pioneering structure-based approach to study S100A8 protein interactions with the chosen anti-inflammatory drugs at the molecular level. The computational results provide valuable insights into the binding modes of the four tested inhibitors to the S100A8–S100A9 complex and the key factors affecting binding affinity. It has been demonstrated that the hydrophobic interactions and hydrogen bonding with S100A8 make pivotal contributions to the binding structures and binding free energies, although the van der Waals' and electrostatic interactions also significantly contribute to the stabilization of the binding structures. The calculated binding free energies are in good agreement with the available experiment activity data. The detailed structural insight, binding modes and the crucial factors affecting the binding free energies obtained from the present computational studies may provide valuable insights for future rational structure-based design of novel, more potent S100A8 inhibitors.
Conclusion
Our analysis suggests distinct transcriptomic signatures for kidney cancer, with significantly high levels of S100A8 expression pointing to one of the underlying molecular mechanisms contributing to progression of kidney cancer. Although further validation is needed to corroborate these findings, analysis of kidney cancer tissue is a promising tool for identifying biomarkers of interest. Protein–ligand interaction studies play a vital role in structure-based computational drug design. Our docking-based findings shed insight into S100A8 protein as a potential target for therapeutic intervention in kidney cancer. S100A8 has been identified as an attractive target for anticancer drug development due to its central role in mediating inflammatory pathways. Further investigations such as quantitative structure–activity relationship studies are required to determine more favorable interaction with S100A8 and its partners. Better binding ligands with more affinity and efficacy can be further designed and validated using combinatorial chemistry and co-crystallization approaches.
Acknowledgements
This work was supported by King Abdullah City for Science and Technology, Riyadh, Saudi Arabia (KACST, Strategic Project ID. 10-BIO1258-03 and 10-BIO1073-03). The Authors would like to thank Nuha Alansari, Alaa Albogmi, Manal Shaabad, Amal M. Noor and Manar Ata for sample collection, performing RNA extraction, bioanalyzer assays and microarray experiments. We thank the patients, physicians, nurses, and pathologists of the King Abdulaziz University Hospital and King Faisal Specialist Hospital and Research Center, Jeddah, Saudi Arabia.
Footnotes
-
Conflicts of Interest
The Authors declare that there are no conflicts of interest.
- Received January 14, 2014.
- Revision received February 17, 2014.
- Accepted February 18, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved