Molecular Docking Analysis of Four Drugs (Phenytoin, Amoxicillin, Aceclofenac and Ciprofloxacin) and Their Association With Four Human Leukocyte Antigen (HLA) Alleles

Background Numerous reports have shown the role of human leukocyte antigen (HLA) alleles in the induction of cutaneous adverse drug reactions by moderating drug metabolism. We therefore aimed to investigate the docking patterns of four HLA alleles (HLA-B x 5101, HLA-B x 1501, HLA-A x 02:06 and HLA-B x 57:01) against four commercial drugs. Methodology Four drugs (phenytoin (PHT), amoxicillin (AMX), aceclofenac (ACE) and ciprofloxacin (CIP)) were investigated for their docking behavior against four HLA alleles (HLA-B x 5101, HLA-B x 1501, HLA-A x 02:06, and HLA-B x 57:01) using the SwissDock method. In addition, toxicity (Tox) and the search tool for interactions of chemicals (STITCH) (protein-drug interaction) analyses were also carried out using the predicating the small molecule pharmaco-kinetic (pk) properties using graph-based signature method (pkCSM) and STITCH free online servers, respectively. Results Toxicity analysis showed that two drugs (amoxicillin and ciprofloxacin) exhibit hepatotoxicity. The STITCH analysis of the drug amoxicillin revealed its interaction with two human proteins. The drug phenytoin exhibited the lowest binding energy (LBE) with all four HLA alleles (HLA-B x 5101, HLA-B x 1501, HLA-A x 02:06, and HLA-B x 57:01). Conclusions The present findings provide new knowledge about the four drugs (phenytoin (PHT), amoxicillin (AMX), aceclofenac (ACE) and ciprofloxacin (CIP)) and their binding affinities with HLA alleles, which may cause cutaneous adverse drug reactions.


Introduction
Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are two rare, life-threatening allergic reactions (rxns) induced mainly due to the usage of drugs that affect both the skin and mucous membranes [1].Three variants of epidermal necrolysis (EN) are (i) SJS, which is characterized by the involvement of less than 10% of the body surface area; (ii) SJS-TEN overlaps, which are characterized by the involvement of 10-30% of the body surface area, and (iii) TEN, which is characterized by the involvement of more than 30% of the body surface area [2,3].The reported major risk factors for SJS and TEN include (i) possessing a slow acetylator genotype, which has been associated with accelerated susceptibility to certain autoimmune diseases (ADs), (ii) the presence of immune-suppression, which describes a weak prognosis in the advancement of SJS/TEN, (iii) usage of anti-convulsant agents concurrent with radiotherapy and (iv) possessing specific human leukocyte antigen (HLA) alleles like HLA-B x 15:02, HLA-A x 31:01 and HLA-B x 58:01 [4].Moreover, the recent understanding of EN is that it is an immune-linked pathway mediated by granylysin, which in turn releases drug-specific cluster of differentiation 8 (CD8) cytotoxic T cells and natural killer cells (NKC) [5].Furthermore, the death of keratinocytes is stimulated by cytotoxic CD8 T cells and NKC via interacting with the HLA and drug antigens [6].Approximately 5-20% of SJS and TEN cases remain idiopathic.Moreover, SJS and TEN are mainly caused by a combination of immune predisposition and exogenous stimuli such as drug/medication or infection that results in the apoptosis of epithelial cells.Furthermore, drug/medication exposure is associated with 50-95% of the cases, depending on the population examined.Especially, people with certain HLA serotypes, T-cell receptor (TCR) subtypes or differences in their ability to absorb, distribute to tissues, metabolize or excrete drugs/medications have a higher probability of developing SJS and TEN.In a nutshell, the pathophysiology of SJS and TEN has been evidenced by the involvement of acute inflammatory vesiculo-bullous reaction of the (a) mucosa of the ocular surface, (b) oral cavity, (c) genitals and (d) the skin and immune-associated reactions from both innate and adaptive immune systems [7].

Toxicity (Tox) analysis
Toxicity (Tox) analysis of the four chosen drugs was performed using the predicating the small molecule pharmaco-kinetic (pk) properties using graph-based signature method (pkCSM) free web server [13].

STITCH (protein-drug interaction) analysis
The four chosen drugs (PHT, AMX, ACE and CIP) were assessed for their interaction with the target human proteins using the search tool for interactions of chemicals (STITCH) free web server [14].

Identification and preparation of target proteins
The 3D structures of human leukocyte antigen (HLA-B x 5101; PDB ID: 1E27 with a resolution of 2.20 Aᵒ), human leukocyte antigen (HLA-B x 1501; PDB ID: 1XR9 with a resolution of 1.79 Aᵒ); human leukocyte antigen (HLA-A x 02:06; PDB ID: 3OXR with a resolution of 1.70 Aᵒ) and human leukocyte antigen (HLA-B x 57:01; PDB ID: 3VRI with a resolution of 1.60 Aᵒ) were retrieved from Protein Data Bank (PDB) database.A chain of these proteins was prepared individually by removing other chains, ligands and even crystallographically observed water molecules by applying UCSF Chimera software tool [12].

Docking study
A docking investigation was carried out for the four selected drugs using a SwissDock free web server.Finally, the binding site interaction analysis was performed by using PyMOL software tool [13].

Docking procedure validation
The validation of the docking procedure was done using abacavir (Aba) as a reference drug.Moreover, the root mean square deviation (RMSD) analysis of all the docked complexes (four drugs) was individually compared with that of abacavir docked complex (reference drug) for each selected protein through PyMOL using the "align" command [15].The RMSD values were utilized as a quantitative guide to know how close the prediction matches with that of the binding mode of abacavir (reference drug).The RMSD values along with visual analysis and docking scores were utilized to assess the reliability of each of the docking programs.The scores were also utilized to investigate the relationship between the binding scores and positions [15].

Ligand
The docking analysis showed that ciprofloxacin demonstrates a maximum binding energy (MBE) of -8.13 kcal per mol with the human leukocyte antigen (HLA-B x 5101).In contrast, phenytoin has the least binding energy (LBE) of -7.28 kcal per mol with HLA-B x 5101 (Table 2).Phenytoin and amoxicillin interact with the Tyr7 amino acid residue (AAR) of the human leukocyte antigen (HLA-B x 5101) (  The root mean square deviation (RMSD) analysis reveals that if RMSD is less than or equal to 2.0 Aᵒ, it is indicated as "excellent", whereas if RMSD greater than 2.0 Aᵒ and less than 3.0 Aᵒ, then it is considered as "fair".On the other hand, if RMSD is greater than or equal to 3.0 Aᵒ, it is indicated as "poor".

Discussion
In the present study, the toxicity (Tox) analysis of the two drugs amoxicillin (AMX) and ciprofloxacin (CIP) showed that they exhibit hepatotoxicity.The current finding correlates with the previous report, in which both the drugs AMX and CIP displayed hepatotoxicity in an in vivo (rat) experiment [16].The association of cytochrome P450 2C9 x 3 (CYP2C9 x 3) and human leukocyte antigen (HLA-B x 51:01) allele in phenytoin (PHT)-induced cutaneous adverse drug reactions was reported in South Indian Tamil patients [17].The association between the human leukocyte antigen (HLA-B x 15:02) allele and carbamazepine-induced SJS was observed in Han Chinese patients [5].Yang and co-workers (2009) noted that abacavir showed good binding affinities towards HLB-B x 57:01 using the molecular docking approach [18].The association between the human leukocyte antigen (HLA-B x 1502) allele and carbamazepine-induced SJS was noted in Indian patients [19].Similarly, the association between the human leukocyte antigen (HLA-B x 5801) allele and allopurinol-induced SJS and TEN was reviewed by Somkrua and colleagues in 2011 [20].Moreover, in silico risk assessment of human leukocyte antigen (HLA-A x 02:06) associated SJS and TEN caused by various ingredients present in cold medicine in Japan was reported by Isogai and colleagues in 2013 [21].The strong association of human leukocyte antigen (HLA-B x 15:21) allele and carbamazepine-induced SJS was reported in Asian patients [22].The association between the human leukocyte antigen (HLA-B x 15:02) allele and carbamazepine-induced SJS and TEN was shown in Malaysian patients [23].Ramsbottom and colleagues pkCSM, predicating the small molecule pharmaco-kinetic (pk) properties using graph-based signature method; N, No; Y, Yes; AT, AMES toxicity; hERG-1, human ether-a-go-go-related gene inhibitor-1; hERG-2, human ether-a-go-go-related gene inhibitor-2; HT, hepatotoxicity; SS, skin sensitization; ORAT, oral rat acute toxicity (lethal dose LD50 in mol per kg; MT, Minnow toxicity (log mM).

FIGURE 4 :
FIGURE 4: interaction of the drug ciprofloxacin with the human proteins using STITCH free web server.Image credit: Radhakrishnan Narayanaswamy.

TABLE 5 : SwissDock binding energy analysis of the four drugs with human leukocyte antigen (HLA-B x 57:01) using the SwissDock method.
Note: *RMSD, root mean square deviation; **Abacavir, reference drug.