Development of Ion pair Chromatography Method for Assay of Telaprevir by Response Surface Methodology

Response surface methodology approach has been utilized for the assay of Telaprevir in pure and formulation using ion-pair chromatography. In this, A risk assessment approach (Control-Noise-Experiment) has been used for identifying the risk factors, i.e. Percentage of Organic Modi ier (% Acetonitrile), Buffer’s pH and low rate of the method. The Central Composite design was applied to optimize the criticalmethod parameters (CMPs) and to ind out the Design space (DS) of the method. The coef icient of correlation (R),%CV and Lack of it are utilized for the evaluation of method responses (Retention time and Asymmetric factor Evaluation of model is justi ied by two diagnostic plots (normal probability plot of residuals and plot of residuals vs predicted values). The mobile phase is Acetate Buffer (20mM) pH 4.4: Acetonitrile (35:65) with 0.9 ml/min of Flow rate. The separation has taken placed in the Eclip Plus C-18 column (250× 4.6 mm, 5μm) at 268 nm. The retention time of Boceprevir was found to be 4.6 min. The validation of the optimized method has performed according to ICH guideline. The method has been successfully used for routine analysis of the Telaprevir throughout the life cycle of the product.

Telaprevir, Quality by Design, Method Development, Validation, Ultra Flow Liquid Chromatography A Response surface methodology approach has been utilized for the assay of Telaprevir in pure and formulation using ion-pair chromatography. In this, A risk assessment approach (Control-Noise-Experiment) has been used for identifying the risk factors, i.e. Percentage of Organic Modi ier (% Acetonitrile), Buffer's pH and low rate of the method. The Central Composite design was applied to optimize the critical method parameters (CMPs) and to ind out the Design space (DS) of the method. The coef icient of correlation (R 2 ),%CV and Lack of it are utilized for the evaluation of method responses (Retention time and Asymmetric factor Evaluation of model is justi ied by two diagnostic plots (normal probability plot of residuals and plot of residuals vs predicted values). The mobile phase is Acetate Buffer (20mM) pH 4.4: Acetonitrile (35:65) with 0.9 ml/min of Flow rate. The separation has taken placed in the Eclip Plus C-18 column (250 × 4.6 mm, 5µm) at 268 nm. The retention time of Boceprevir was found to be 4.6 min. The validation of the optimized method has performed according to ICH guideline. The method has been successfully used for routine analysis of the Telaprevir throughout the life cycle of the product.

INTRODUCTION
Worldwide about 130,000,000 people have been affected by the Hepatitis C virus (HCV), with a burden of 366,000 deaths per year. In the Western world, HCV infection is a vital reason for liver disease (Gentile et al., 2009). Telaprevir is the irst direct-acting antiviral drug approved in 2011 (Rao et al., 2015). It acts as NS3/4A serine protease inhibitor (Clinical pharmacology and biopharmaceutics review, 2011). 3a,4,5,6,pyrrole-1-carboxamide.The molecular formula of Telaprevir is C 36 H 53 N 7 O 6 and its molecular weight is 679.85. It is white to an off-white powder with 0.0047 mg/mL water solubility (Clinical pharmacology and biopharmaceutics review, 2011).
The methods published so far for Telaprevir quanti ication in plasma and dosage formulations are based on liquid chromatography coupled to mass spectroscopy and UV detectors. For example, Reddy et al. has reported the HPLC method who separated Telaprevir on the Acquity UPLC-BEH C-18 column using a diode array detector (Reddy et al., 2015). Aouri et al. have developed an LCMS method where the separation was achieved on Hypercarb ® 148 3 µm, 2.1 mm x 149 100 mm column maintained at 80 • C using thermostated (Aouri et al., 2013). Panda et al. have published QbD method developed by HPLC on the Enable-C18G column (Panda et al., 2016). Chenet al. reported the LC-MS/MS method on a Waters XBridgeTM BEH Shield C18 column (Chen et al., 2014), Heinz et al. has separated Telaprevir on LUNA C18(2)-HST column and detected using a UV-VIS photodiode array detector. The column was protected using the 18 pre-column and maintained at 40 • C using a column thermostat (Heinz et al., 2015). Panda et al. performed method development on Phenomenex Luna ® C-18 column (Panda et al., 2017), and Penchala et al., used Accucore C18 column for chromatographic separation (Penchala et al., 2013). Tempestilli et al. developed the HPLC-UV method to quantify using X TerraVR RP18 column for chromatographic separation . Corrien et al. published an assay method for the determination of Telaprevir in dried blood using liquid chromatography and tandem mass detector using Phenomenex NX C18 column (CPWGM et al., 2015). In underdeveloped and developing countries like India, Bangladesh, and Nepal use of a dedicated column for a single compound is not cost-effective. This problem has received limited attention in the literature. So there is a demand for developing an effective and reliable technique using the C18 column (L1 packing) for cost-effective regular analysis.
Due to the poor peak-shape of Telaprevir chromatogram and inadaptability of the ion suppression technique because of the use of a mobile buffer phase of pH>8, Higher pH is not compatible with most of the standard reversed-phase columns (Jost and Hauck, 1983). In these situations, ion pair chromatography is the best option available. Here in this present work, we used 1-Octanesulphonic acid salt monohydrate as an Ion pair reagent in the mobile phase for HPLC based assay of Telaprevir.
This work aims to develop a simple isocratic Ion pair chromatography method using a common C18 column with L1 packing by applying the response surface methodology.
The present study is planned to be carried out in three phases, (i) Development of Ion pair chromatography method using C18(L1 packing) (ii) Optimization of the chromatographic condition by Central Composite Design (CCD) to ind out Method Operable Design Region (MODR) and inally to detect Design Space (DS) and (iii) Validation of the method according to ICH guideline.

Chemical and reagents
Telaprevir was procured from MNS Laboratories Pvt Ltd. Sodium Acetate Trihydrate (Analytical Grade) was purchased from Ranbaxy laboratories ltd, Mumbai, and HPLC grade Methanol was purchased from Thermo Fischer Scienti ic. Ultra-pure water (HPLC grade) was obtained from a Milli-Q Plus 185 water puri ication unit.

Instruments
All experiments for the method development and validation were performed on Shimadzu Ultra Flow Liquid Chromatography (UFLC) system equipped with LC-20AD pump with a PDA detector. The signal was processed and integrated using LC Real-time Analysis software.

Initial Chromatographic Condition
The initial separation was taken placed on Eclip plus C-18 column (250 × 4.6 mm, 5 µm) with 20 µl injection during 10 minutes run time 1.0ml/min low rate.
The mobile phase considered was 20mM Sodium acetate Buffer (pH 4.5): Methanol (33:67% v/v). The mobile phase was added with 5mM 1-Octanesulphonic acid salt monohydrate (IP reagent). The signals are monitored at 268nm in PDA.

Statistics
The obtained results were subjected to Central Composite Design (CCD) using Design-Expert ® 11 Software Trial Version.

Preparation of standard and sample solution
Preparation of primary Stock solution (1000µg/ml) 10mg of Telaprevir was dissolved in 10ml of acetonitrile to obtain the strength of 1000µg/ml

Preparation of working standard (20 µg/ml)
The stock solution was diluted accordingly to obtain the strength of 20µg/ml solution of Telaprevir.

Standard Curve construction
For calibration, standard Telaprevir solutions were made by diluting working standard with acetonitrile to get 5, 10, 15, 20, 25, 30, 35 & 40 µg/mL ( Figure 1). QC samples of the low, medium and high concentrations (10, 20 & 30 µg/mL) were prepared in the same solvent.   In the present studies, high-risk variables (Table 1) are identi ied and assessed with the help of Control-Noise-Experimentation (CNX) approach (Raman et al., 2015). Three critical method parameters viz. % of acetonitrile, low rate, and pH of Buffer were identi ied and imperilled to response surface methodology (Design of experiment) to ind a design space of the developed method.

Experimental Design
In experimental design, the number of experimental runs is constructed to achieve or identi ied true optimum points. Therefore, the effective variable on the HPLC method's ef iciency was optimized by using a central composite design and a quadratic model was constructed between the attributes (Retention time & Asymmetric factor) and the independent (% organic modi ier, mobile phase pH & Flow Rate) variables (Box and Wilson, 1951;Hashemi et al., 2010).
For an experimental design to get a bias-free response, a 20µg/ml of Telaprevir was used for all runs. The developed model is used for the main and interactive study of variables. The analysis of variance (ANOVA) was found to be signi icant (p < 0.05) while framing the polynomial equation. The parameters like lack of it, coef icient of correlation (R 2 ) and %CV are used for model itting (Sivakumar et al., 2007;Khodadoust and Ghaedi, 2013). The itness of the model is investigated and justi ied by diagnostic plots, such as the residual plot & normal probability plot (Olivero et al., 1995;Stalikas et al., 2009). The interaction study was carried out using 2D and 3D plot to understand the interaction between critical method parameters and attributes (Choisnard et al., 2003;?).

Validation
Validation of the method has been performed as per ICH guideline Q2 (R1) (Procedures, 1996).

Optimization of Chromatographic Condition
The separation of Telaprevir is optimized by experimental factors such as buffer and organic modiier of the mobile phase, detection wavelength and low rate of the elution. The optimized separation was achieved with Sodium acetate buffer and acetonitrile among different applied solvent systems. The excipient compounds must not interfere with the analysis of the targeted analyte. Linearity The standard linearity for Boceprevir was generated from 5µg/mL to 40 µg/mL R2 was found to be 0.9970 with y = 4638.92x -655.79 And Standard error was found to be 2540.6741 The correlation coef icient for ive concentration levels will be ≥0.997 for the range of 80 to 120% of the target Concentration.
Range 10µg/mL to 30 µg/mL range was used for the Accuracy and Precision study. %RSD was less than 2.
The acceptable range will be de ined as the concentration interval over which linearity and accuracy are obtained per the above criteria, and in addition, that yield a precision of ≤3%  Ion ion-pair (IP) reagent 1-Octanesulphonic acid salt monohydrate is added to the mobile phase to achieve better separation.

Effect of Concentration of Ion-pair reagent
In this study, 0.5mM to 10mM 1-Octanesulphonic acid salt monohydrate concentration in the mobile phase was studied for optimisation. The best concentration for separation and retention was found to be 5 mM.

Design of Experiment
The Design of the experiment is constructed to explore a better understanding of dependent and independent factors to achieve the best separation using central composite design (CCD). [15,16] In CCD, all experiments are performed in randomized order to minimize the effects of uncontrolled variables, as shown in Table 2.
Using this design interaction study was evaluated to optimized quadratic effects. The experimental results of the CCD have been itted with coded expressions for Retention time. The effects, as well as interactions of the dependent and independent variables, are evaluated following ANOVA using Design expert 11 . The P-value (P<0.05) for R1 & R2 is observed to be 0.002 & 0.0006, indicating the signi icance of the factorial effect at a 95% con idence level. The Model Fvalues of 7.20 for R1 & 10.24 for R2, respectively, implies the model is signi icant (Hashemi et al., 2010;Sivakumar et al., 2007).The it of the polynomial model equation is expressed by the coef icient of determination R 2, as shown in the result with 0.8663 and 0.9021 for the corresponding values of Retention time and Asymmetric factor, whereas 0.8087 and 0.8140 represented the same for the adjusted R 2 values. When adjusted R 2 values are ≥ 0.80, the relation between the itted model and experimental data is found to be good (Hashemi et al., 2010;Sivakumar et al., 2007). "The adequate precision value is a measure of the signal (response) to noise (deviation) ratio". A ratio greater than four is desirable (Hashemi et al., 2010;Sivakumar et al., 2007). In this study, the ratio is found to be 10.8811 for R1 & 10.6190 for R2, indicating the model is signi icant for the separation process. The value of 3.03 in the case of the parameter of %CV for all models is in agreement with previous literature, hence indicate reasonable reproducibility (Hashemi et al., 2010;Sivakumar et al., 2007).
The Model Responses R1 and R2 are evaluated using diagnostic plots, (i) a normal probability plot of residuals and (ii) plot of residuals vs predicted values. The close observation of a normal probability plot of residuals in Figure 2 tells that the residuals are fall on a straight line. Hence it is concluded that the distribution of errors is normal and the model its the data adequately (Khodadoust and Ghaedi, 2013;Olivero et al., 1995). It is observed in the plot of residuals vs predicted values in Figure 3 that there is no obvious pattern in the residual versus predicted response. The plot also exposes an almost equal distribution of residues above and below the x-axis predicting the suitability of the model. Due to regularity and continual adjustment of the residuals, the itted model for the R1 and R2 may be accepted (Khodadoust and Ghaedi, 2013;Olivero et al., 1995).

Interference study
3D-response surface plots are used to analyse for identifying the interaction(s) among the dependent and independent variables (Stalikas et al., 2009;Choisnard et al., 2003) and shown in Figure 4.
The 3D-response surface plot showed a linear increasing Retention time value with a decrease in buffer pH. At the higher pH of the buffer and lowest low rate, optimum Retention time is shown. It is also observed a linear decreasing retention Time with increasing the Organic Modi ier as well as low rate. The asymmetric factor showed a linear increasing value with an increase in the pH as well as a % organic modi ier. The 3D-response surface plot also showed a linear increasing asymmetric factor value with an increase of the low rate and buffer pH.

Design Space and Desirability Function
Design Space (DS) is created using the modelling software Design Expert Trial Version. Twodimensional charts are created by taking three factors (% of Acetonitrile, pH & Flow rate) and represented in Figure 5. The shaded blue and yellow region of the 2D contour plots depict the design space for retention time as well as the asymmetric factor, which de ines the robust region of the method where results are within designated criteria (? Hadjmohammadi and Shari i, 2012).
Optimizing the Retention time of symmetrical peak and minimization of asymmetric factors are the main objectives of Derringer's desirability function (D). It is a technique to optimize different parameters with multiple responses (Hadjmohammadi and Shari i, 2012;Panda et al., 2015). The value of D with zero indicates a desirable range of all responses and D close to 1 indicates optimum responses with a near target value. The maximum desirability function (D = 1) is pull out from the response surface curve (Figure 6), signifying the model is excellent (Hadjmohammadi and Shari i, 2012;Panda et al., 2015). The coordinates produce the maximum desirability value at Acetonitrile 65% v/v, pH 4.4, and a low rate of 0.9 mL min_1.
Hence, these critical method parameters have been optimized and strictly control during the development of the method. The inal robust UFLC-method´s condition for Telaprevir estimation is shown in Table 3, and Chromatogram in Figure 7.

Validation of the method
The method has been validated by applying the working point conditions according to ICH guideline Q2(R1) with respect to selectivity, linearity, range, accuracy, precision, the limit of detection, and quantitation (Procedures, 1996). The summary of the validation report is shown in Table 4 .

CONCLUSIONS
In this work, the QbD approach has been successfully implemented to develop a robust method for the estimation of Telaprevir in API and formulation. To initiate the method development by QbD approach, the physiochemical properties of Telaprevir was considered in the selection of input variables for the Design of Experiment using Central Composite Design. Because of the narrow concentration range of Telaprevir, the concentration was not considered as a quantitative variable in this design. So, mobile phase pH, % organic modi ier mobile phase & low rate were considered as qualitative variable and were controlled. Each step of the Analytical QbD process has been studied to ind out the Design Space. Response surface plots graphically illustrated the major effects of mobile phase pH, % organic modi ier mobile phase and low rate on the separation. Using the QbD approach, the robustness of the method is already available before going for validation. The method was also validated for accuracy and precision, and the result was satisfactory. The method has been found to be cost-effective, pre-cise, accurate, and linear at concentrations ranging from 5 µg/ml to 40 µg/mL for Telaprevir using Eclip plus C-18 column in 10 minutes runtime.