Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy
An intelligent method based on feed-forward artificial neural network and least square support vector machine for the simultaneous spectrophotometric estimation of anti hepatitis C virus drugs in pharmaceutical formulation and biological fluid
Graphical abstract
Introduction
Hepatitis C is known as a liver disorder, which is caused by the hepatitis C virus (HCV). It is a risk factor in increasing liver cancer, liver cirrhosis, liver failure, and liver transplantation [1]. Also, researchers have revealed that the HCV affects other parts of the body such as the immune system, the lymphatic, and digestive system [2]. The optimum type, period of therapy, and prediction of the possibility of sustained virologic response are determined by HCV genotyping. The direct-acting antiviral (DAA) drugs are prescribed to treat HCV [3], [4]. The first generation of direct-acting protease inhibitors were Telaprevir and boceprevir, which approved to treat the genotype 1 (G1). The administration of these drugs was associated with interferon and ribavirin. Hence, side effects and limitations in their effectiveness have been reported. Second-generation of DAA drugs such as sofosbuvir (SOF) and daclatasvir (DAC) were introduced. A high pangenotypic activity and minimum unpleasant side effects are the advantages of these drugs [5].
SOF (Fig. 1a) has approved in December 2013. It is a nucleotide analog inhibitor that is responsible for inhibiting the HCV NS5B polymerase. SOF is used to treat chronic hepatitis C [6], [7]. Metabolism of this drug to the active uridine analog triphosphate can be done. The combination of this metabolite with the ribonucleic acid (RNA) of HCV leads to the termination of the chain [8]. In order to treat the disease in a faster and shorter time, SOF can be used incorporation with other drugs [9]. SOF along with DAC has revealed sustained virologic response and better treatment results in patients, which possess HCV genotype 1 or 3 [10].
DAC (Fig. 1b) is prescribed to treat the HCV infection. This drug aims diverse phases of the HCV lifecycle. It acts as a selective inhibitor of HCV non-structural protein 5A (NS5A) used by the virus to reproduce [11], [12]. In most chronic HCV cure protocols, the combination of SOF and DAC is preferred to use due to the presence of many HCV genotypes [13].
Several techniques, including high performance liquid chromatography (HPLC) [14], reversed-phase high-performance liquid chromatography (RP-HPLC) [15], [16], [17], high-performance thin-layer chromatography (HPTLC) [18], [19], ultra-performance liquid chromatography (UPLC) [20], ultra-high performance liquid chromatography tandem mass spectrometry (UHPLC–MS/MS) [21], liquid chromatography-tandem mass spectrometry (LC-MS/MS) [22], [23] have been utilized for the simultaneous determination of SOF and DAC with each other or with other drugs. Costly, time-consuming, require sophisticated instruments, and excessive consumption of pollutant solvents are the limitations of these methods [24], [25]. On the other hand, the spectrophotometry method has been widely employed in the simultaneous estimation owing to the ease of the approach, accessibility of instruments, accuracy, precision, and cost-effectiveness. However, in order to solve the problem of spectral overlap, the spectral data should be improved [26]. Various chemometrics methods coupling with UV/Visible spectrophotometry such as continuous wavelet transform (CWT) [27], artificial neural network (ANN), least squares support vector machine (LS-SVM) [28], derivative spectrophotometry (DS) [29], and so on have been applied to resolve overlapped spectra. The ANN is based on the human biological neural system first model. It consists of neurons, which are multiple connected simple processing elements. Input signals of the environment are transferred to these neurons. By connecting weights and via a training procedure, the signal is converted. Feed-forward neural network (FF-NN) is the most extensive type of ANN. This network contains input, hidden, and output layers. The back-propagation (BP) learning technique is used in FF-NN [30]. The LS-SVM is a classical statistical learning method that the basis of this model is the support vector machine (SVM). Unlike the SVM, LS-SVM represents the sum of the squared-error term in the function of target related to the SVM. The selection of kernel function is important in both SVM and LS-SVM. The radial basis function (RBF) is known as the most common kernel function [31].
In this study, the spectrophotometry method along with FF-NN and LS-SVM models as easier, faster, sensitive, accurate, and precise way was introduced for the simultaneous determination of SOF and DAC in pharmaceutical tablet formulation and biological fluid. Comparison between the proposed method and HPLC technique was performed using one-way analysis of variance (ANOVA).
Section snippets
Methodology
The improved form of SVM is known as LS-SVM, which makes simple the computation process and computational cost. Also, SVM training becomes easier. A set of training samples is assumed: . Here, and are input and output vector of LS-SVM model, respectively. In addition, the number of training samples is shown by n.
The LSSVM model in the feature space can be represented in Eq (1).where and indicate a nonlinear mapping function and a weight
Materials
SOF and DAC (purity > 99%) were obtained from Khorshid Kadous pharma company (Tehran, Iran) as gratis samples. Sovodak tablet (400 mg SOF, 60 mg DAC) was acquired, which were manufactured by Fan Avaran Rojan Mohaghegh Darou company (Tehran, Iran). The solvent of these materials was ethanol (Merck, Germany).
Apparatus and software
In order to record spectra and measuring absorbance, T90+ double beam UV–visible (PG Instruments Ltd) equipped with UVWin5 (version 5.2.0) software was used. Analysis of the Sovodak tablet
UV spectra
Two standard solutions of SOF and DAC with different concentrations were selected for showing the absorption spectra of these components (Fig. 2). Extensive overlapping over the range 200–400 nm related to the absorption spectra of SOF and DAC is observed. This overlapping causes the use of common UV spectrophotometry along with the chemometrics method (FF-NN and LS-SVM) for the simultaneous estimation of concentration value.
The ANN model
A series of topologies were investigated to assess the optimum number
Conclusion
In this study, the proposed UV spectrophotometric technique along with FF-ANN and LS-SVM were successfully developed and validated for the simultaneous analysis of SOF and DAC. The short time of analysis is the advantage of this method, which causes the evaluation of SOF and DAC in the commercial formulation during the routine analysis in quality control laboratories. Also, this method is simple, rapid, cost-effective, robust, with suitable accuracy and precision, which can be suggested as an
CRediT authorship contribution statement
Kiarash Keyvan: Writing, Software. Mahmoud Reza Sohrabi: Project administration, Writing - review & editing. Fereshteh Motiee: Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (34)
- et al.
Optimization of A Sensitive and Robust Strategy for Micellar Electrokinetic Chromatographic Analysis of Sofosbuvir in Combination with its Co-Formulated Hepatitis C Antiviral Drugs
J. Chromatogr. A
(2020) - et al.
Development and validation of a new HPLC-DAD method for quantification of sofosbuvir in human serum and its comparison with LC–MS/MS technique: Application to a bioequivalence study
J. Chromatogr. B
(2017) - et al.
Development and validation of sensitive and rapid UPLC-MS/MS method for quantitative determination of daclatasvir in human plasma: Application to a bioequivalence study
J. Pharm. Biomed. Anal.
(2016) - et al.
Optimization and modeling of a green dual detected RP-HPLC method by UV and fluorescence detectors using two level full factorial design for simultaneous determination of sofosbuvir and ledipasvir: Application to average content and uniformity of dosage unit testing
Microchem. J.
(2019) - et al.
New, simple and sensitive HPTLC method for simultaneous determination of anti-hepatitis C sofosbuvir and ledipasvir in rabbit plasma
J. Chromatogr. B
(2018) - et al.
Validated spectrophotometric andchromatographic methods for analysis of the recently approved hepatitis C antiviral combination ledipasvir and sofosbuvirMéthodes
French Pharmaceutical Annals.
(2018) - et al.
Spectrophotometric and robust UPLC methods for simultaneous determination of velpatasvir and sofosbuvir in their tablet
Microchem. J.
(2019) - et al.
D’Avolio, A UHPLC–MS/MS method for the quantification of direct antiviralagents simeprevir, daclatasvir, ledipasvir, sofosbuvir/GS-331007, dasabuvir, ombitasvir and paritaprevir, together with ritonavir, inhuman plasma
J. Pharm. Biomed. Anal.
(2016) - et al.
Rapid bioanalytical LC-MS/MS method for the simultaneous determination of sofosbuvir and velpatasvir in human plasmaapplication to a pharmacokinetic study in Egyptian volunteers
J. Chromatogr. B
(2018) - et al.
Development a validated highly sensitive LC–MS/MS method for simultaneous quantification of Ledipasvir, sofosbuvir and its major metabolite GS-331007 in human plasma: Application to a human pharmacokinetic study
J. Pharm. Biomed. Anal.
(2017)