Extraction optimization and validation of HPLC method for determination of miquelianin in Cuphea glutinosa leaves

Popularly known in Brazil as "sete-sangrias", Cuphea genus is used in folk medicine as diuretic, antipyretic, anti-inflammatory and anti-hypertensive[1]. The chemical composition of some Cuphea species has been reported, presenting triterpenes, steroids, fatty acids, carbohydrates, tannins and flavonoids, mainly quercetin derivatives[2]. The present work aimed to optimize the extractive method and to validate an HPLC-PDA method for the quantification of miquelianin in leaves of Cuphea glutinosa. The species was collected in Alegrete, RS, Brazil, and the optimization of extractive method was performed in two steps (extractive method selection and optimization by Central Composite), through the statistical software MiniTab 17®[3]. The extractive methods tested were static maceration, ultrasound, infusion and reflux. After screening, five parameters were selected for Fractional Factorial Design (FFD). In addition, to estimate the experimental error, the Dong algorithm was associated with factorial design. In the Central Composite Design (CCD) the parameters were evaluated in two levels (-1 and +1): number of extractions and solvent percentage[4]. In the validation method, the parameters selectivity, linearity, matrix effect, precision, accuracy and robustness were evaluated[5,6]. The results indicated ultrasound-assisted extraction a better condition. Through FFD, 75% Dong algorithm and CCD, the optimal extraction was obtained with these parameters: plant:solvent 1:60 (w/v), percentage solvent 38% ethanol, 60 min time, five extractions and particle size ≤ 180 μm. Selectivity was proved by the addition of a reference substance in the sample solution, which produced an increase only in the peak area correspondent to the compound. Linearity was determined with reference solutions in concentrations of 60 - 140 μg/mL, with r2= 0.9999 and residual plot showed that the data were homoscedastic. Matrix effect was estimated by comparing the slopes of standard curves and matrix-matched curves and the lines were parallel. Repeatability and reproducibility presented adequate results, with RSD of 2.1818 % and 1.3971%, respectively. The mean recovery level for accuracy was 90.6%. Robustness was evidenced by Youden experiment, all results showed RSD of less than 1%. The average content of miquelianin was 1.0318%. Thus, the optimization of extractive method in C. glutinosa leaves increased the concentration of miquelianin in crude extract and the performed method is in accordance with the validation requirements of the current legislation.