Development of UPLC-ESI-MS/MS Method

In order to optimize the extraction conditions, the effect of different solvents and variation of time were investigated. It was found that pure methanol was more efficient solvent for extraction among the tested different concentrations of methanol and ethanol. In addition, the efficiency of extraction by ultrasonicator (Bandelin SONOREX, Berlin) were measured in different time intervals i.e., 20, 30, 40, and 50 min at room temperature. It was observed that the components extracted completely within 30 min. Finally, the sample solutions were prepared with 50 ml 100% methanol by ultrasonic extraction for 30 min.

Optimization of chromatographic and MS/MS conditions

In this study, mobile phase was optimized using different compositions of solvents and adjusting their gradient elution for separation of all the compounds. Compared to methanol, acetonitrilepossesses stronger elution ability, which can shorten the elution time and thus selected for this method. On the basis of the polarity of compounds an Acquity UPLC BEH C₁₈ (50 mm × 2.1 mm id, 1.7 µm; Waters, Milford, MA) column was selected for their separation, which was more applicable for acidic mobile phase with smoother baseline in the separation as compared to other relative to other columns. Compared with acetic acid, formic acid was more effective for ionization of those detected under the positive and negative ESI modes. Thus, different concentration strengths (0.05%, 0.1% and 0.2%) of formic acid were investigated. Final gradient elution with 0.1% formic acid in water and acetonitrile at a flow rate of 0.3 mL/min with the column temperature of 25°C resulted in separation of the 28 compounds in 10 min chromatographic runtime was selected for analysis.

All the compounds dependent MS parameters (precursor ion, product ion, declustering potential (DP) and collision energy (CE) were carefully optimized for each target compound individually, in both positive and negative ion modes, which was performed by flow injection analysis (FIA) of the individual standard solution into the mass spectrometer. The chemical structures of 28 components were characterized based on their retention behaviour and MS information such as, quasimolecular ions [M+H]⁺, [M+Na]⁺ fragment ions, [M-H-COO]^–, [M-H-COO-CH3]^–, [M- CO-H2O] comparison with standards and literatures [31-37]. MRM parameters: DP, EP, CE and CXP were optimized to achieve the most abundant, specific and stable MRM transition for each compound as shown in Table S3 and the MS/MS spectra were shown in Fig. S1 (a) and Fig. S1 (b). The type of polarity switching used in the LC–MS/MS system was continuous switching between the positive and negative ionization mode. A positive/negative switching time of 100 ms was optimized for continuous switching; the software algorithm builds an acquisition method that aligns both polarities at appropriate MRM transitions. Apart from that, the fast switching of positive-negative ionization mode has allowed the simultaneous determination of both positively and negatively ionized bioactive compounds in shorter analysis time, while refrain from additional injections into the LC-MS/MS system. MRM extracted ion chromatogram of analytes and both internal standards are shown in Fig. 2.

Analytical Method Validation

The proposed UPLC-MRM method for quantitative analysis was validated according to the guidelines of international conference on harmonization (ICH, Q2R1) by determining linearity, lower limit of detection (LOD), lower limit of quantification (LOQ), precision, solution stability and recovery.

• Linearity and detection limit

The quantitative method was assessed by linearity and sensitivity. Calibration curves were prepared by plotting the peak area of marker compounds against the corresponding concentrations. The regression lines are linear in the concentration range studied and the corresponding coefficients of correlation are showed in Table 1. Good linear relationships with r² in the range of (0.9984-0.9999) are demonstrated for each analyte. Limits of detection (LOD) and quantification (LOQ) under the present chromatographic conditions were determined on the basis of response and slope of each regression equation at a signal noise ratio (S/N) of 3 and 10, respectively. The LOD for the twenty-eight compounds ranged from 0.10 to 2.24 ng/ml and LOQ ranged from 0.33 to 3.88 ng/ml.

• Precision, Stability and Recovery

The intra-day and inter-day variations, which were chosen to determine the precision of the developed method, were investigated by determining twenty-eight analytes with IS in six replicates during a single day and by duplicating the experiments on three consecutive days. Variations of the peak area were taken as the measures of precision and expressed as percentage relative standard deviations (RSD). The overall intra-day and inter-day precisions were not more than 1.66% and 2.11% respectively are showed in Table 1. Stability of sample solutions stored at room temperature was investigated by replicate injections of the sample solution at 0, 2, 4, 8, 12 and 24 h. RSD value of stability samples of the twenty-eight analytes is ≤ 1.89%. A recovery test was applied to evaluate the accuracy of this method. Three different concentration levels (high, middle and low) of the analytical standards were added into the samples in triplicate and average recoveries were determined by the following equation. The analytical method developed had good accuracy with overall recovery in the range from 91.27-105.92 % (RSD ≤ 3.74 %) for all analytes Table S2 (Supporting information).

Applicability of the proposed method

The developed method was successfully applied to the simultaneous determination of 28 compounds (bacoside A3, bacopaside II, bacopaside X and bacopasaponin C, withanolide-A, withaferin-A, asiaticoside, madecassoside, jatrorrhizine, palmatine, magnoflorine, curcumin, gallic acid, protocatechuic acid, ferulic acid, caffeic acid, ellagic acid, rosamarinic acid, ursolic acid, catechin, apigenin, luteolin, quercetin, rutin, kaempferol-3-O-rutinoside, corilagin, chrysin and chlorogenic acid) in twenty different batches of same pharmaceutical manufacturer obtained from local market. The method was proved to be effective and reliable. All the contents were summarized in Table 2 and represented graphically in Fig. 3.

The commercial polyherbal tablets MT1-20 contains Bacopa monnieri and Centella asiatica major therapeutically active ingredients, but they show considerable variations in the contents of different bacosides, asiaticoside and madecassoside.

The content of gallic acid was highest in batch MT4 (5090.03 μg/g) and among other batches, because it was present in most of the herbs mixed in MT formulation. Bacoside-A3 (170.99 μg/g) and withaferin-A (110.03 μg/g) was detected highest in MT3 batch, while bacoside-X (482.50 μg/g) and bacopasaponin-C (146.52 μg/g) detected highest in MT2 batch comparison to others. Major content of asiaticoside (265.0 μg/g) and madecassoside (467.0 μg/g) was detected in batch MT15. However, the contents of these major compounds varied among these different samples, which could result in quality and efficacy of various sample.

Curcumin, kaempferol-3-O-rutinoside, rutin, rosamarinic acid and ferulic acid were found either in very low content or below detection limit in all the batches. Some effective components such as withanolide-A, chrysin, ursolic acid, chlorogenic acid, palmatine, magnoflorine and ellagic acid were observed in trace amounts and simultaneously determined. Besides, batch MT4 was having highest total contents of compounds, followed by batch MT1, MT3, MT2, MT7, MT5, etc as shown in Table 2. As a whole, multiple bioactive compounds, including macro and micro, were detected simultaneously and are frequently considered to be responsible for the therapeutic effects, which may work ‘synergistically’ and reflect the characteristic of HMs. It is suggested that the quality control study of MT should be focused on the raw materials control as well as in the procedures of processing/manufacturing and parameter optimization.

Quality assessment of MT by HCA and PCA

To evaluate the quality and chemical variation of MT different batches, multivariate analysis approaches such as HCA and PCA were performed.

• Hierarchical clustering analysis (HCA)

HCA was performed based on the characteristics of the contents of 28 marker compounds. The marker contents formed a 15 × 17 matrix in 20 batches of samples. The dendogram of samples as shown in Fig. 3 (a), shows similarity between different batches. In the present study compounds gallic acid and ellagic acid is an outlier. The dendogram as showed in Fig. S4 (Supporting information) showed that depending upon the quantity, the 28 compounds under study has been clustered based on complete linkage euclidean distance.

• Principal component analysis (PCA)

PCA is an unsupervised pattern recognition method used for analysing, classifying and reducing the dimensionality of numerical datasets in a multivariate. PCA was carried out to evaluate the quality and categorise the samples of 20 batches based on the characteristics of the contents of 28 compounds. The first 5 peaks were able to account for 83.13% information of data matrix. The first PC explained 45.77% variation the contribution to which was mainly from protocatechuic acid, catechin, caffeic acid, corilagin, magnoflorine, rosamarinic acid, jatrorrhizine, palmatine, bacoside-X and chrysin while PC2 explained 26% variation due to madecassoside, asiaticoside, quercetin, bacoside-II, bacoside-A3 and bacopasaponin-C. The biplot of PC1 vs PC2 as shown in Fig. 4 (a & b) suggests that 15 out of 20 formulations are clustered into 4 subgroups while remaining 5 species viz.; MT13, MT15, MT9, MT17, and MT10 appears to have distinctive quantities of chemical pattern.

Concluding remarks

In the present study, we have successfully developed an UPLC-ESI-MS/MS method to simultaneously analyze 28 bioactive compounds in a single run. This method was fully validated with respect to linearity, precision, stability and recovery. The application of multiple reaction monitoring acquisition mode enhances the specificity for complex multi-compounds in herbal drugs and achieve better accuracy in simultaneous quantification. The quantitative analysis revealed that quality is the major concern in herbal preparations/formulations therefore it is an important concern to analyze significant variation of multiple bioactive compounds. UPLC-MS/MS system is more suitable for complex analytical determination of pharmaceutical preparations as it offers low solvent consumption, reduced analysis time and thus low analysis cost, which is a very important aspect in many QC laboratories where high sample throughput and fast analytical speed are required. Hence, our findings suggested that the developed method for quantification of multi-constituents in polyherbal formulation is an effective approach for batch-to-batch monitoring in MT and promises wide applications in quality control of raw materials and finished products.