INTRODUCTION:

Hepatocellular carcinoma (HCC) is the fifth most malignant tumor¹, where malignancy has been found to be upregulated by certain genes to produce an effect of transformation. It has also reported that former induce chronic inflammation and fibrosis development.^2,3. Primary liver cancer, hepatocellular carcinoma has less frequency of occurrence than colorectal and pancreatic carcinomas being the common sources of hepatic metastases⁴. There is close link between hepatitis B (HBV) and C virus (HCV) infections, particularly alcohol-induced liver injury, and consumption of aflatoxin proved byepidemiologic studies⁵.

The family of claudin (CLDN) consists of proteins which are required for the development of tight junctions (TJs) present epithelial as well as endothelial cells.

Tight junctions play a vital role in the control of paracellular transport and in the maintenance of cell polarity. Involvement of tight junction proteins results in signal transduction and regulatory pathways⁶. In formation of tight junctions, tissues express many CLDNs that interact with both homotypic and heterotypic. Overexpression of CLDNs in cancer which are less likely to be occurred are related in the formation of tight junctions. CLDNsacts as a membrane proteins which are involved in survival of cancer cells and helps in various therapeutics⁷. It has two extracellular loops, the first loop is more elongated than the second loop, and a short carboxyl intracellular tail. In tail, last amino acids which form PDZ binding motifs are highly conserved in the family. This implies that few of the CLDNsgenes were produced by gene multiplication, and that neighboring genes may be coordinately regulated. Function and localization of CLDNs are affected by post translational modifications like phosphorylation which has various roles in CLDN proteins like CLDN-1 which is involved by mitogen actuated protein kinase (MAPK)⁸, CLDN-5 promotes the barrier domain of tight junctions, and CLDN-16 increases Mg2+ transport in the tail region⁹. Change in transepithelial resistance has been found by overexpression in cell lines by the permeability properties of CLDNs. One of the features ofCLDNsare that it acts as barrier properties in CLDN-1, -4, -5, -6, -8, -9 and -19 and in CLDN -2, -10, -15 and -16, itcauses pores forming junction leak to ions. On the basis of mouse CLDN proteins that they are classified to classic and non-classic proteins¹⁰. The function of CLDN-10 in cancer is not clear. Researchers have found that members of CLDN family show cell invasion and migration. Overexpression of CLDN-10 (NM_006984, encrypts 228 amino acids) were reported in lung carcinoma as the common isoform in different tissue organs⁵. Survival rate was seen more in increased expression of CLDN-10 in hepatocellular carcinoma. During metastasis cells break down the adhesion barrier which shows changes in CLDNs¹¹. This expression of CLDN can vary between the primary and metastatic lesion of the same tumor¹². Differentiation of tumors of biliary tract and hepatocellular carcinoma were found during the overexpression of increased CLDN-4¹³. Expressions ofCLDNsplay a vital role in mammalian kidney growth of the pronephros, mesonephros and metanephros. Studies show that CLDN-10 was greatly expressed in pronephric tubules and in the mesonephric and metanephric kidneys¹⁰. The studies depict that only hepatocellular carcinoma recurrence and papillary thyroid carcinoma, show the interconnection of CLDN-10 expression signifying the functional involvement during cancer growth. In vitro assays with the CLDN-10 overexpression and small interfering RNA–mediated knockdown transfectants are two domains of CLDN-10 in two distinct hepatocellular carcinoma cell lines. Hepatacellular3B which lacks CLDN-10 expression, are also affected by the over expression of CLDN-10 impart malignant phenotypes to hepatocellular carcinoma.

Surveillance of cancer cell, invasiveness and motility has higher risk in malignant phenotype. CLDN-10 showed growth in mRNA transcription and protein expression of membrane type 1-MMP (MT1-MMP). In our previous report, MT1-MMP is a type of protease which promotes intrahepatic metastasis.In cancer cell, the expression of CLDN-10 may influence the expression levels of its own family members by CLDN-1, CLDN-2, and CLDN-4 which in retreat was upregulated in CLDN-10 overexpression transfectants. Small interfering RNA–based knockdown of CLDN-10 in Hepatocellular Carcinoma Cell Line destroy invasion and reduce the activation of MMPs and CLDN member’s expression. Hepatocellular Carcinoma Cell Line is an invasive cell line with elevated level of CLDN-10 expression¹⁴.

MATERIALS AND METHODS:

Homology Modeling:

There are various methods to construct and predict resolution model of proteins from the sequence and one of which is homology modeling which works based on determining 3D structure experimentally of same protein called template protein. Homology modeling comprises certain steps (i) it helps in finding out the sequence of template structure (ii) combining both the sequences of template structure and query sequence (iii) based on the information obtained from the template structure and query a model structure is developed (iv) evaluation of the obtained model¹⁵.The software used to develop the homology model is I-TASSER¹⁶. Homology modeling of the CLDN-10 of hepatocellular carcinoma was executed using I-TASSER software. Initially, the sequence of CLDN-10 was taken from UniProt¹⁷. It consists of 224 amino acids residues. The CLDN-10 was then exposed to a PSI-BLAST search¹⁸ in order to find the homologous proteins from the Protein Data Bank (PDB)¹⁹. Anappropriate template for CLDN-10 was screened. Subsequently, modeling was performed for the CLDN-10 of hepatocellular carcinoma against the selected template using I-TASSER²⁰. Based on the internal score, the output of the designed structures was positioned and those with thelessnumber ofranking scores were found and used for model validation.

Assessment of Homology Model:

The authentication of structure model obtained from I-TASSER was performed by inspecting the structure conformation of the modeled structure was calculated by resolving the phi (φ) and psi (ψ) torsion angles using PROCHECK, as decided by Ramachandran plot. The results were also confirmed using Structural Analysis and Verification Server.Confirm 3D will gives visual investigation of the quality of structure for a protein and dissects the similarity of an atomic model of the protein with its corresponding amino acids [16].

Preparation of Macromolecule:

The molecular volume of models of CLDN-10 was determined using Discovery Studio 3.1 (Molecular Attributes, AccelrysInc).

Ligand Selection:

Eighteen ligands from alfalfa plant were selected from the literature²¹. These ligands have been reported by the research studies for the interaction with the CLDN-10. All the structures were prepared. The ligands were checked for 3D structure, associated fragments accurate bond lengths and bond angles, covalent bonds, hydrogen, and protonation states and were prepared by AutoDock. The AutoDock process contains a series of steps that execute conversions, apply corrections to the structures, produce changes in the structures, exclude unwanted structures, and optimize the structures. To initiate the process, the ligand structures in thepdbqt format were used as input for the AutoDock module.The structures were prepared separately. We also selected the option to retain the original state of the input molecule. To ensure that other molecules such as water and counter ions are excluded from the ligand structure, we used the default option in AutoDock²².

Virtual Screening:

Virtual Screening was performed in the targeted molecule of CLDN-10 with the eighteen ligands from alfalfa plant using AutoDock and PyRx 0.8 (Virtual Screening Tools)²³. The grid for docking calculations was centered on thebinding domain of CLDN-10. Based on the scoring of binding energy four best ligand molecules were chosen for molecular docking²⁴.

Molecular Docking:

Receptor Grid Generation:

Once the ligands and protein were set in the form of docking, receptor grid files were obtained with the help of aprogram which generates grid-receptor. Using the grid box option the grids were centered on the macromolecule for the active sites residues²². Corresponding X Y Z values were noted. The parameter library file was set to AD4 type and was saved in grid parameter file (gpf).

Ligand Docking:

Based on the genetic algorithm protocol ligand molecules were chosen and selected. The macromolecule was set to rigid docking which enables no change in the ligand and substrate and helps in ligand binding to the protein. Using the docking parameter the output was set in lamarican GA , the file was saved in docking parameter file (dpf).

Docking:

Using Autogrid and AutoDock option corresponding gpf and dpf files were launched and docked.

RESULTS AND DISCUSSIONS:

Homology Modelling:

Comparative modelling was used to model CLDN-10 macromolecule of Human. I-TASSER was used to develop the model of CLDN-10. Using UniProt (SwissProt) (P78369) FASTA sequence was downloaded and modeling was done. An appropriate template 3X29A and 4P79A for CLDN-10 was identified. Homology modeling was performed for the CLDN-10 of hepatocellular carcinoma against the chosen template using I-TASSER.

The outcome of the modeled structures wereclassified on the basis of an internal scoring function, and those with the less amount of internal scores were obtained and utilized for model validation.

The outcome of the modeled structures was classified on the basis of an internal scoring capacity, and those with the less number of internal scores were obtained and used for model validation.

The structure validation model was obtained from I- Tasserr by examining the backbone of the modeled structure calculated by analyzing the phi (φ) and psi (ψ) torsion angles, as found by Ramachandran plot. The results were also confirmed using Structural Analysis and Verification Server. (Figure 1).

Figure 1: CLAUDIN-10 structure

Virtual Screening:

Through virtual screening, the outcomes showed derivatives docked accurately with CLDN-10 for hepatocellular carcinoma. The good binding energy was seen in hit eriodictyol of -7.3 kcal/mol and lower binding energy in coumarin of -5.2 kcal/mol showed the changes in binding due to ligand atom interactions with CLDN-10. The best top four ligands - eriodictyol, apigenin, luteoloin and quercetin was selected as hits with binding energy of more than -6.6 kcal/mol confirmed the potential binding (Table 1). Hence, the virtual screening method was more useful in identifying the top hits from the ligand dataset.

Table 1 Virtual screening with alfalfa compounds binding energies with CLAUDIN-10.

S.No	Compound name	Binding energy
1	Apigenin	-6.9
2	Coumarin	-5.2
3	Chrysin	-6.3
4	Daidzein	-5.9
5	Eriodictyol	-7.3
6	Fisetin	-5.8
7	Flavanone	-6.4
8	Flavone	-6.6
9	Flavonol	-6.4
10	Genistein	-5.6
11	Hesperitin	-6.5
12	Kaempferol	-6.5
13	Luteolin	-6.8
14	Morin	-5.5
15	Myrecetin	-5.9
16	Naringenin	-5.9
17	Quercetin	-6.7
18	Umbelliferone	-5.6

Molecular Docking:

After hits identification, re-docking on the same were performed for identification of possible lead candidates with binding pose in AutoDock. The grid parameters were set to X = 126, Y = 90 and Z = 126, the offset value for X was 2.444 were set. With the help of docking results, conformations for each hit was generated and binding mode having low binding energy was selected as best conformation. The post docking analysis was evaluated by parameters like favorable energy, low inhibition constant and weak interactions involved between protein/ligand complexes. The energy terms include binding energy, final intermolecular energy, and electrostatic energy proved the energy contribution for the favorable docked complex. In clauin-10-ligand complex, interactions such as hydrogen bond and hydrophobic interactions plays crucial role in ligand recognition and protein stability after ligand binding.

From post docking analysis, four hits eriodictyol, apigenin, luteoloin and quercetin showed the convenient binding energy with range of -7.6 to -6.7 kcal/mol, this proved the effective docked complex. The binding energy was more reasonable for the transition induced and subsequent rearrangements in protein after ligand binding. Moreover, the 7effective interaction was further supported by energies like final intermolecular energy, electrostatic energy and hydrophobic. Besides energy terms, inhibition constant of four hits furnished the docking results and confirmed the significance of inhibition.

The hydrogen bond interactions of CLDN-10 ligand complex were examined using PyMOL. The results showed hydrogen bonds between the protein and ligand atoms. The interacting residues include Asp223, Trp50, Ser22 and Ser23 involved in active CLDN-10 inhibition (Table 2). The binding poses of hit clearly implied the effective inhibition (Figure 2, Figure 3, and Figure 4).

Table 2 Binding energies of best compounds of alfalfa

S.no	Compound	Binding Energy (kcal/mol)	IC50	Hydrogen Bonds
1	Eriodictyol	-7.6	426µM	TRP50 SER22
2	Apigenin	-4.9	255 µM	ASP223
3	Quercetin	-3.67	2.05mM	SER23

Figure 2 CLAUDIN-10 with eriodityol hydrogen bond interactions

Figure 3: CLAUDIN-10 with apigenin-hydrogen bond interactions

Figure 4: CLAUDIN-10 with quercetin-hydrogen bond interactions

Docking Results:

CLDN-10 surface binding pocket with eriodictyol, apigenin and quercetin are shown in Figure 5, Figure 6, and Figure 7.

Figure 5: CLAUDIN-10 surface binding pocket with eriodictyol

Figure. 6: CLAUDIN-10 surface binding pocket with apigenin

Figure 7CLAUDIN-10 surface binding pocket with quercetin

CONCLUSION:

Using computational approaches, we were able to find the CLDN-10 inhibitor. Surface domain of CLDN-10 was selected as the target region for inhibitor. From the virtual screening results, the least binding affinity of inhibitor obtained from alfalfa plant characterized the strong binding with CLDN-10. The results obtained from AutoDock Vina were re-screened and lead candidates were identified based on binding energy. Binding mode with least binding energy, low inhibition constant, and more number of hydrogen bond interactions in surface domain were the parameters for inhibition mechanism of lead candidates. The hydrogen bond interaction pattern between ligand and protein confirmed the high affinity of lead candidates. In conclusion, inhibitors act as potent CLDN-10 inhibitors. Thus, our computational findings initiate the hepatocellular carcinoma therapeutics using inhibitor and further experimental studies can confirm the effective human CLDN-10 inhibition.

ACKNOWLEDGEMENTS:

The authors thank the management of VIT University for providing the facilities to carry out this work.

CONFLICTS OF INTEREST:

The authors declare no conflict of interest.

REFERENCES:

1. Lai ECH, Lau WY. The continuing challenge of hepatic cancer in Asia. Surgeon. 2005; 3:210-215.

2. Murakami Y, Yasuda T, Saigo K, Urashima T, Toyoda H, Okanoue T, Shimotohno K . Comprehensive analysis of microRNA expression patterns in hepatocellular carcinoma and non-tumorous tissues. Oncogene. 2005; 25:2537-45.

3. Zhang J, Yang Y, Yang T, Liu Y, Li A, Fu S, Wu M, Pan Z, Zhou W. MicroRNA-22, downregulated in hepatocellular carcinoma and correlated with prognosis, suppresses cell proliferation and tumourigenicity. British Journal of Cancer. 2010; 103:1215 – 1220.

4. Holczbauer A, Gyöngyösi B, Lotz G, Szijártó A, Kupcsulik P, Schaff Z, Kiss A. Distinct Claudin Expression Profiles of Hepatocellular Carcinoma and Metastatic Colorectal and Pancreatic Carcinomas. Journal of Histochemistry& Cytochemistry. 2013; 61:294–305.

5. Cheung ST, Leung KL, Ip YC, Chen X, Fong DY, Ng IO, Fan ST, So S. CLAUDIN-10 Expression Level is Associated with Recurrence of Primary Hepatocellular Carcinoma. Clinical Cancer Research. 2005; 11:551-6.

6. Balda MS, Matter K. Tight junctions and the regulation of gene expression. Biochimica et Biophysica Acta. 2009; 1788:761–7.

7. Hewitt KJ, Agarwal R, Morin PJ. The claudin gene family: expression in normal and neoplastic tissues. Biomed Central Cancer. 2006;6:186.

8. Fujibe M, Chiba H, Kojima T, Soma T, Wada T, Yamashita T, Sawada N. Thr203 of claudin-1, a putative phosphorylation site for MAP kinase, is required to promote the barrier function of tight junctions. Experimental Cell Research. 2004;295:36–47.

9. Kwon MJ. Emerging Roles of Claudins in Human Cancer. International Journal ofMolecular Sciences. 2013;14:18148-80.

10. Collins MM, Baumholtz AI, Ryan AK.Claudin family members exhibit unique temporal and spatial expression boundaries in the chick embryo. Tissue Barriers. 2013; 1:e24517.

11. Martin TA, Jiang WG. Loss of tight junction barrier function and its role in cancer metastasis. Biochimica et Biophysica Acta. 2009; 1788:872–91.

12. Tabaries S, Dong Z, Annis MG, Omeroglu A, Pepin F, Ouellet V, Russo C, Hassanain M, Metrakos P, Diaz Z, Basik M, Bertos N, Park M, Guettier C, Adam R, Hallett M, Siegel PM. Claudin-2 is selectively enriched in and promotes the formation of breast cancer liver metastases through engagement of integrin complexes. Oncogene. 2011; 30:1318–28.

13. Lodi C, Szabo E, Holczbauer A, Batmunkh E, Szijarto A, Kupcsulik P, Kovalszky I, Paku S, Illyes G, Kiss A, Schaff Z. CLAUDIN-4 differentiates biliary tract cancers from hepatocellular carcinomas. Modern Pathology. 2006; 19:460–9.

14. Ip YC, Cheung ST, Lee YT, Ho JC, Fan ST. Inhibition of hepatocellular carcinoma invasion by suppression of claudin-10 in HLE cells. Molecular Cancer Therapeutics. 2007; 6:2858–67.

15. Gadhwal MK, Patil S, D’Mello P, Joshi U. Homology modeling of aryl hydrocarbon receptor and docking of agonists and antagonists. International Journal of Pharmacy and Pharmaceutical Sciences. 2013; 5:76-81.

16. Suganya PR, Sudevan K, Kalva S, Saleena LM. Homology modeling for human adam12 using prime, i-tasser and easymodeller. International Journal of Pharmacy and Pharmaceutical Sciences. 2014; 6:782-786.

17. McGinnis S, Madden TL. BLAST: at the core of a powerful and diverse set of sequence analysis tools. Nucleic Acids Research. 2004; 32:W20–5.

18. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research. 1997; 25:3389–402.

19. Berman H, Henrick K, Nakamura H. Announcing the worldwide Protein Data Bank. Nature Structural Biology. 2003; 10:980.

20. Sali A, Blundell TL. Comparative protein modelling by satisfaction of spatial restraints. Journal of Molecular Biology. 1993; 234:779–815.

21. Peters NK, Long SR. Alfalfa root exudates and compounds which promote or inhibit induction of rhizobium meliloti nodulation genes. Plant Physiology. 1988; 88:396-400.

22. Jayaraman A, Jamil K, Kakarala KK.Homology Modelling and Docking Studies of Human α2-Adrenergic Receptor Subtypes. Journal of Computer Science and System Biology. 2013; 6:3.

23. Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization and multithreading. Journal of Computational Chemistry. 2010; 31:455-461.

24. Utomo DH,Widodo N, Rifa'I M.Identifications small molecules inhibitor of p53-mortalin complex for cancer drug using virtual screening. Bioinformation. 2012; 8:426–429.

Received on 28.06.2017 Modified on 28.07.2017

Research J. Pharm. and Tech 2018; 11(8): 3529-3533.

DOI: 10.5958/0974-360X.2018.00652.2