To optimize peak shape with appropriate retention time, various combinations of mobile phases were investigated. The chromatographic separations were initiated to achieve a short runtime, symmetric peak shapes, minimum matrix interference and solvent consumption. Previous studies for teriflunomide 15, 16, 17, 18, pioglitazone 19, 20 and prednisone 21-25] have reported different columns with 5µm particle size, 3–4 mm inner diameter and columns lengths (125–150 mm) with runtimes ≥15 min. Thus, in the present work chromatographic separation was tried on XBridge C18 (150 mm×4.6 mm, 5 µm), Waters X-Terra MS C18 (100 mm×3.9 mm, 5 µm) and Inertsil ODS-3 C18 (150 mm×4.6 mm, 5 µm) columns. To find the best eluting solvent system, various combinations of methanol/acetonitrile with additives like ammonium acetate and ammonium formate in different concentration and volume ratios were tested. Best results were obtained in terms of higher sensitivity, superior retention and better peak shapes on X-Terra MS C18 column. The best separation was achieved with 0.01M ammonium formate (A) and acetonitrile (B) for estimating TFM, PDN and PGZ along with ISs (IBU and IMP). Initial gradient conditions were set at 95% A and 5% B. From 2 to 8 min, %B increased to 95% and the same maintained till 10 min. Again %B was decreased linearly to 5% upto 12 min and held at the same till 15 min (total runtime is 15 min). Reproducibility and recovery data for all the analytes and IS supported protein precipitation to be used as the preferred extraction technique. All the chromatographic separations are achieved with single step protein precipitation technique using acetonitrile without any interference from one another. The reversed-phase XTerra MS C18 column (100 mm x 3.9 mm, 5 µm) was used with a flow rate of 1.0 mL/min. The retention times of IBU, TFM, PDN, PGZ, and IMP were found to be 5.6, 5.9, 6.5, 6.9 and 7.6 min, respectively. LC chromatogram showing proper separation of all the analytes is represented in Figure 2.

The standard solutions of 10 µg/mL with respect to TFM, PDN, PGZ in methanol were infused directly into the mass spectrometer. The electrospray ionization (ESI) of TFM was conducted in negative ionization mode as it has high electron affinity due to the presence of trifluoromethyl group. Internal standard, IBP has given higher response in negative mode. The source dependent and compound dependent parameters were suitably optimized to obtain a consistent and adequate response for all the analytes. The recovery in other solvent systems was between 50% and 80%, but was inconsistent with some ion suppression(greater than 15% CV).

The observed full scan mass spectra in positive mode showed prominent protonated molecules [M+H]⁺ of m/z 359 and 357 for PDN and PGZ respectively in positive ion mode, and prominent deprotonated molecules [M-H]^- of m/z 269 and 205 for TFM and IBU respectively, in negative ion mode. The [M+H]⁺ ions and [M-H]^- ions of respective analytes were subjected to collision-induced dissociation (CID) at average collision energy of 30%. The collision energies were optimized for each analyte to obtain the most intense fragment ions. The spectra were acquired using the following conditions: ion spray voltage, 5000 V; turbo gas temperature, 250°C; nebulizer gas (compressed air), 55 psi; curtain gas (N₂), 20 psi; declustering potential, 80 eV; focusing potential, 200 eV; entrance potential, 10 eV. The optimum values for compound dependent parameters like declustering potential (DP), collision energy (CE), entrance potential (EP), focusing potential (FP) and cell exit potential (CXP) set were −46, −30, −10, −283 and −13V in negative mode and 40, 25, 12, 230 and 13V in positive mode respectively. The molecules underwent fragmentation to yield the following fragment ions of m/z 160, 237 and 122 for TFM, PDN and PGZ respectively. The MS spectra of three analytes are presented in Figure 3.

Figure 3

Based on their mass spectra and tandem mass spectra, the following MRM transitions: m/z 269→160, m/z 359→237, m/z 357→122 were selected for analysis of TFM, PDN, and PGZ respectively. The extracted MRM ion chromatograms of teriflunomide, prednisone, pioglitazone, ibuprofen and imipramine are represented in Figure 4.

The specificity of this method was confirmed by comparing chromatograms of blank plasma, spiked plasma with analytes at a concentration of 1 ng/mL. PDN, PGZ and IMP (IS) in positive ESI experiment and TFM and IBU (IS) in negative ESI experiment were well separated under the described chromatographic conditions. No interfering endogenous peaks were observed around their retention times. The calibration curves of the analytes showed a good linearity over the studied concentration range of 1–500 ng/mL for TFM, PDN and PGZ with correlation coefficients (r²) 0.999. The LODs for all studied analytes were found to be 0.5 ng/mL. The LLOQs for all analytes were 1 ng/mL with acceptable precision and accuracy. The intra- and inter-day precisions for TFM, PDN and PGZ were less than 8.2%. The obtained intra-day accuracies were in the range of 97.2–102.2% and inter-day accuracies were in the range of 97.3–101.8%. The validation parameters are depicted in Table 2.

Table 2

Analyte	Nominal concentration (ng/mL)	Intra-day		Inter-day
		% Recovery	% RSD	% Recovery	% RSD
PDN	1	100.6	6.811	100.8	1.692
	5	98.6	3.331	98.8	1.383
	100	99.5	3.272	99.3	0.579
	400	99.3	1.800	99.3	0.328
PGZ	1	97.3	8.123	101.8	5.111
	5	100.3	3.832	97.3	2.433
	100	102.2	6.263	100.5	1.611
	400	102.2	7.172	99.6	1.066
TFM	1	97.9	6.963	97.7	1.127
	5	99.6	4.621	100.0	1.147
	100	98.3	5.888	99.1	1.362
	400	97.2	4.718	97.7	0.926
	stability (24h RT)	eeze-thaw stability (4 cycles)	Longterm stability (Day-30)

The extraction recoveries of all drugs from rat plasma were in the range of 65.5–85.3% with relative standard deviations less than 8.1%, which indicates the sample preparation technique is suitable for extracting the studied drugs from rat plasma. The results are displayed in false.

The stability studies of these drugs were performed at three QC concentration (low, medium and high) levels in six replicates (n=6). The predicted concentrations for each analyte deviated within ±5.6% of nominal concentrations after storage of plasma samples at room temperature for 24 h, four freeze–thaw cycles at -20°C. The mean accuracies were found to be more than 95% with relative standard deviations less than 4.3%, which are summarized in Table 4. Long-term stability of all the samples after Day-30 were found to be within 4.5% of actual concentrations.

Table 4

Analyte	Mean accuracy ± RSD
	Short term stability (24h RT)	Freeze-thaw stability (4 cycles)	Freeze-thaw stability (4 cycles)
PDN	97.5 ± 2.6	99.3 ± 2.3	99.9 ± 1.9
	99.9 ± 3.2	98.4 ± 4.3	97.6 ± 3.8
PGZ	99.0 ± 3.2	100.6 ± 1.9	100.0 ± 2.0
	101.4 ± 5.6	99.4 ± 3.2	97.4 ± 4.5
TFM	101.7 ± 4.6	99.8 ± 4.7	99.6 ± 2.2
	98.1 ± 2.8	99.3 ± 3.1	98.3 ± 3.8

All the samples were analyzed by the developed method and the mean plasma concentrations vs time profile of prednisone, pioglitazone and teriflunomide are shown in Figure 3.

Figure 5

The plasma concentration–time profile of all the three analytes was analyzed by the non–compartmental method using WinNonlin Version 5.1. The estimated pharmacokinetic parameters are shown in Table 5.

Table 5

Parameter	TFM	PDN	PGZ
Cmax (ng/mL)	100.1 ± 13.6	93.6 ± 17.0	47.4 ± 9.7
Tmax (h)	5.0 ± 1.4	2.3 ± 1.5	1.5 ± 0.7
t1/2 (h)	11.7 ± 4.6	4.8 ± 4.7	4.6 ± 2.2
Kel (h-1)	0.059 ± 0.098	0.144 ± 0.081	0.114 ± 0.021

Cmax: maximum plasma concentration.

The present study was undertaken to develop a rapid and simple method to determine the prednisone which is in Phase-4 clinical trials for multiple sclerosis disease. Along with teriflunomide and pioglitazone this method has been developed and validated which can be highly useful in drug interaction studies with prednisone and also other possible random combinations. The current method is proved for its ruggedness with sufficient recovery and low matrix effect in a simple protein precipitation method with acetonitrile. This method may be applicable to other preclinical samples as well as clinical sample analysis of these three drugs given to the patients suffering with multiple sclerosis disease. In drug interaction studies, the concomitant drugs can be analyzed in a single run with current developed methods. Separation of ibuprofen, imipramine and teriflunomide is based on their affinity and interaction with the reverse phase C-18 column and the mobile phase. In the present study, based on sensitivity, matrix effect and reproducibility requirements extraction techniques of LLE (Liquid-liquid extraction) and SPE (Solid phase extraction) were tried during method development for the combinations associating teriflunomide. Thus, three analytes were distinctly resolved from one another by the present LC method. It should be noted that interfering peaks from endogenous compounds were not observed. Selectivity of the method in rat K₂EDTA plasma was evaluated in twelve individual matrix lots along with one lipemic and one hemolytic lot. Peak responses in blank lots were compared against the response of spiked LLOQ and negligible interference was observed at the retention time of analytes and internal standards. Calibration for the quantification was performed using the ratio of peak area of the analyte to that of IS. The results of intra- and inter- day accuracy and precision listed in Table 2 indicated that the method was accurate with excellent accuracy range of 97.2–102.2%, and the CV was within 10%. Matrix effect is a special phenomenon associated with LC–MS determination of drugs from biological fluids such as plasma and other matrices. Endogenous components extracted from plasma may suppress or enhance ionization of the analytes in electrospray source if they co-elute with the analytes. It is for this reason that matrix effect was evaluated under the experimental conditions used in this study. The average ion suppression or enhancement at low QC and high QC levels was < 11%, suggesting that matrix effect on the analysis was negligible. The stability results (Table 4) showed that the mixture of three analytes spiked into rat plasma was stable for 24 h at ambient temperature, for 30 days at 4°C, and four freeze–thaw cycles at -70°C. Stability of the analyte in the sample is of crucial importance to the validity of the split sample program. Thus, proper storage of all samples is very important to obtaining reliable results and their interpretation in analysis. Application of the current validated method has resulted positive results for pharmacokinetic study in rats which shall be highly useful information in clinical trials. Fig-5 shows the plasma-time concentration profiles of in vivo study conducted in albino rats for PDN, TFM, and PGZ. The extraction recovery of analytes from K₂EDTA plasma was determined by comparing the peak responses of plasma samples (n = 6) spiked before extraction with that of plasma samples spiked after extraction. The mean recovery of PDN, TFM and PGZ was found to be 70.9%, 79.0% and 71.4% respectively with %CV across the three levels ranging between 3.0 and 10.6%, as shown in Table 3. Dilution integrity experiment was carried out at 2 times the ULOQ concentration for all the three analytes. After 1/2 and 1/4 dilution the mean back calculated concentration for dilution QC samples was within 85–115% of nominal value with a %CV of ≤ 8.1.

We have developed and validated a highly sensitive, specific, reproducible and high throughput LC–MS/MS assay to quantify TFM, PDN and PGZ simultaneously in rat plasma. Simple and single step protein precipitation was used to extract analytes from rat plasma. The major advantages of the assay are simple sample preparation with equal sensitivity for all the three analytes. The obtained LODs and LOQs of all the drugs were adequate and proved successful to perform the pharmacokinetic study in rat plasma. Based on the results, we can conclude that the present method is suitable for quantification of multiple analytes simultaneously without any interference and matrix effects. The concomitant drug analysis along with the target analyte is more advantageous than single compound analysis and also useful in drug interaction and toxicology studies. This method may be useful for clinical sample analysis of prednisone which is currently in Phase 4 trials of multiple sclerosis disease along with coadministered drugs.

The authors are thankful to SK Healthcare Pvt Ltd, Hyderabad, India for providing facility and support to carry out this work.