UPLC-MS/MS determination of 71 neuropsychotropic drugs in human serum

In this study, a UPLC-MS/MS method was developed for the rapid detection of 71 neuropsychotropic drugs in human serum for drug concentration monitoring and toxicity screening. The analytes were separated from the biological matrix by protein precipitation using a methanol-acetonitrile solvent mixture. The chromatographic separation was performed on a Kromasil ClassicShell C18 column (2.1*50 mm, 2.5 μ m) with gradient elution using acetonitrile-0.2 % acetic acid and 10 mM ammonium acetate as the mobile phases (flow rate 0.4 mL/min, column temperature 40 °C, injection volume 5 μL). An electrospray ion source in both positive and negative ion modes with multiple ion monitoring was used. The total run time was 6 min. All compounds were quantified using the isotope internal standard method. Totally, 71 drugs were detected within their linear ranges with correlation coefficients greater than 0.990. The intra- and inter-batch precision relative standard deviations (RSDs) for the low, medium, and high concentration points were less than 15 %, with an accuracy of 90%–110 %. All compounds except Moclobemide N-oxindole are stabilised within 7 days. The relative matrix effect results for each analyte were within ±20 % of the requirements. The method is validated according to Clinical and Laboratory Standards Institute guidelines, easy to use, and has a low cost.


Introduction
Mental illnesses account for a large proportion of clinically relevant diseases and are currently treated mainly by administering neuropsychotropic drugs that offer symptomatic control, but regular monitoring of blood levels is required due to interpatient differences, narrow therapeutic window for the drugs, high toxicity, and susceptibility to adverse reactions [1][2][3][4][5].
W. Jin et al.
Techniques for Quality Assurance of Therapeutic Drug Monitoring in 2021 [8].Several published guidelines recommend liquid chromatography-tandem mass spectrometry (LC-MS/MS) as an effective way to monitor drug concentrations in the blood.LC-MS/MS methods can detect neuropsychotropic drugs with high sensitivity and accuracy [9][10][11].These methods for detecting single or multiple classes of neuropsychotropic drugs by MS have been reported, but these methods have the following shortcomings: 1) Low throughput: most existing methods are designed for the simultaneous determination of multiple drugs of a single class [12] or separate determination of individual drugs [13][14][15][16][17]. 2) Low efficiency: various methodologies have to be used to determine drugs of distinct classes, necessitating frequent alteration of columns and mobile phases during the assay.This entails increased allocation of labor, resources, and finances.Furthermore, the capacity to concurrently detect multiple drugs remains restricted, thereby reducing assay efficiency [9,[18][19][20][21]. 3) Low sensitivity: in the simultaneous detection of multiple drugs, some drugs cannot be accurately detected due to low detection signals, which affects the accuracy of detection.In addition, the accurate quantification of drugs at lower concentrations is more challenging when also simultaneously determining drugs that are present at higher concentrations in the sample.
We developed an LC-MS/MS method to detect the concentrations of a total of 71 drugs in human serum by protein precipitation: 12 antiepileptic drugs and their 2 metabolites, 17 antipsychotics and their 5 metabolites, and 21 antidepressants and their 14 metabolites.The drugs were all present in the range of AGNP guidelines.The proposed method is simple, has good accuracy and reproducibility, and can provide technical support for the rational use of neuropsychotropic drugs.

Chemicals and reagents
The 71 drug standards and internal standards were obtained from Manhage (Shanghai, China) Biotechnology Co. Methanol and acetonitrile (Shanghai ANPEL Biotechnology Co., Ltd.; Shanghai, China), as well as ammonium acetate and acetic acid (Sigma-Aldrich Co., LLC.; MO, USA) were of chromatography grade.methanol and acetonitrile were from Shanghai ANPEL Biotechnology Co.,Ltd (Shanghai,China); ammonium acetate and acetic acid were from Sigma-Aldrich Co., LLC (MO, USA); Blank serum were from Real-BioGroup Co., Ltd.(Shenzhen, China).Ultrapure water (Millipore Express ultrapure water system; Millipore, Bedford, USA) was used in the analysis.
Mobile phase A was an aqueous solution of ammonium acetate (10 mM) containing 0.2 % acetic acid.Mobile phase B was acetonitrile.
The mobile phase gradient conditions used are given in Table 1.The column temperature was 40 • C, and the injection volume was 5 μL.
The MS conditions were as follows: electrospray ionization (ESI) source was used in the positive and negative ion, as well as multiple reaction monitoring (MRM) modes.The capillary voltage was 0.5 kV, desolvent gas temperature was 400 • C, desolvent gas flow was 1100 L/h, conical pore gas flow was 150 L/h, and nitrogen gas pressure was 7.0 bar.The parameters for MRM monitoring are shown in Table 2.

Preparation of stock and working solutions 2.3.1. Standard stock solution
The purity of each drug compound was different, and according to the purity stated on the certificate, the drug samples were divided into two drugs with a purity >95 % and drugs with a purity <95 %.When the purity of the drug was >95 % of the standard, 1 mg of the standard (default purity 100 %) was accurately weighed and dissolved in 1 mL of methanol.When the purity of the drug was <95 % of the standard, according to the actual purity of the quality of the conversion, and then dissolved in 1 mL of methanol.

Configuration of standard curves and QC
2.3.2.1.Standard curves.Sulpiride, chlorpromazine, agomelatine, chlorprothixene, desmethyl sertraline, sertraline, fluphenazine, haloperidol, reduced haloperidol, and flupenthixol were used in mixed concentration group 1. Blank serum was spiked with the appropriate amount of each drug to form high concentration mix 1, and the remaining five mixes were prepared from high  concentration mix 1 and blank serum.The following ratios of high concentration mix 1 to blank serum were followed: 1:2, 1:5, 1:11, 1:24, and 1:99.For the rest of the drugs as mixed concentration group 2, the appropriate amount of different drugs and blank serum were mixed to prepare high concentration mixing standard 2. The remaining five mixing ratios of the high concentration mixing standard 2 to blank serum were as follows: 3:2, 1:1, 2:3, 3:7, and 1:29.Mixing standard 1 and mixing standard 2 were mixed in the ratio of 1:9, and finally a standard curve of STD1-STD6 was prepared.
2.3.2.2.QC.QC preparations were obtained by diluting them from a new stock solution.QC samples comprise three concentrations covering the lowest detection limit, the range within the recommended reference concentration for the drug, and the range above the recommended concentration.The method of preparation of the three QC samples (low, medium, and high concentrations) was the same as that of standards 1, 3, and 5.

Preparation of the internal standard precipitation solution
The appropriate amount of each drug's internal standard was precisely weighed and placed in a 1 ml volumetric flask and dissolved in methanol.The volume was made up to obtain a concentration of 1 mg/ml of internal standard reserve solution.The appropriate volume of each drug's internal standard reserve solution was then taken in a reagent bottle, and the mixed internal standard solution was prepared by mixing with methanol, and then diluted 100 times with methanol-acetonitrile reagent mixture (4:6) to obtain the internal standard precipitation solution.

Sample preparation
Serum sample (50 μL), QC sample(50 μL), and standard sample(50 μL) were taken in a centrifuge tube, to which was added 300 μL of internal standard precipitating solution and vortex mixed.The tube was centrifuged at a high speed at 16000g for 10 min at 4 • C. The supernatant (100 μL) was collected in a collection plate, to which was added 100 μL of 5 % methanol in water and mixed well.The mix was analyzed using the Waters Acquity UPLC I-Class XEVO TQS.

Method validation
Method validation was performed according to CLSI C62-A [22] guidelines and 9012 Guidelines for Validation of Methods for Quantitative Analysis of Biological Samples [23].

Selectivity and specificity
Separate samples of blank human serum and serum samples with the drug-mixed standard solution added were tested.The presence of interference peaks and any effect of endogenous substances were studied.

Linearity and sensitivity
The standard curve was plotted by linear regression with the mass concentration of the drug of interest in serum (X) as the horizontal coordinate and the peak area ratio of the drug of interest to the internal standard (Y) as the vertical coordinate.The lower limit of quantitation (LLOQ) was defined as the lowest concentration on the standard curve satisfying the following conditions S/N ≥ 10, and it was noted that the RSD of precision of 5 replicate standard curve samples was within 20 %, and the concentration deviation was within ±15 % of the theoretical value.

Intra-and inter-day validation and accuracy
High, medium, and low concentrations of QC samples were prepared by adding standards to blank human serum.One batch was made each day for three days to investigate the intra-batch and inter-batch precision and accuracy, respectively.

Matrix effect
For validating the matrix effect, the post sample treatment addition method was adopted.The matrix sample was taken, pretreated, and then the target analyte was added at a certain concentration.Thus, three samples of low, medium, and high concentrations were prepared, and the measurement was repeated thrice for each concentration.The sample and the same concentration of pure solvent as the matrix sample were analyzed, and then their respective response values and the ratio of the internal standard were    compared.The difference between the ratio of the matrix sample response value to the internal standard and the ratio of the solvent sample response value to the internal standard to the solvent sample response value was considered the relative matrix effect.

Evaluation of stability
Standards were added to the blank serum to obtain specimens for stability studies at low, medium, and high concentrations.The control group was stored at − 80 • C, and the other groups were placed at room temperature.They were kept at room temperature for 1, 2, 3, 5, and 7 days and then transferred to the − 80 • C refrigerator.After all samples were transferred, they were tested to compare the experimental group with the control group.The samples were considered to be stable if the deviation was no more than 15 %.

Carryover
After the high concentration sample, a blank sample was analyzed to estimate the residual carryover effect.A carryover effect was considered to be absent if residues in the blank after the high concentration sample did not exceed 20 % of the lower limit of quantification and 5 % of the internal standard.

Selectivity and specificity
The chromatographic peaks of each drug were well resolved and had a good peak shape without obvious interference from impurity peaks, and the endogenous substances in the serum did not interfere with the detection, indicating that the method is highly specific and selective (Figs.1-3).

Intra and inter-day validation and accuracy
For each of the drugs analyzed, the intra-batch RSDs for the three concentration levels, low, medium, and high, were 1.2 %-9.2% and inter-batch RSDs for the three concentration levels were 1.1 %-10.6 %, and the accuracy ranged from 89.59 % to 111.18 % (Table 4).

Matrix effect
The relative matrix effect results for each analyte were − 14.98 %-19.67 % of the required limit (Table 5).

Stability evaluation
All compounds except moclobemide N-oxindole left at room temperature for 7 days were − 13.25 %-14.91 %15 % deviation from the control (Table 6).

Carryover effect
The residues in the blank samples after the high concentration samples of each compound were analyzed.The residues of the LLOQ were 0.00 %-18.73 % and the residues of the internal standard were 0.00 %-0.77 %, meeting the requirements(Table 7).

Discussion
Regional surveys in recent years have shown that the monitoring of psychotropic drug concentrations in China is still in its infancy, and the low coverage of inter-room QC and the lack of a unified detection method are unmet needs [24,25].Kirchherr et al. established a method for the simultaneous detection of 48 psychiatric drugs, but their method employed only 3 isotopic internal standards, which did not correspond to all compounds, and the detection time of this method was relatively long [26].Two reported methods by Remane et al. and Montenarh et al. were able to detect more than 30 psychiatric drugs simultaneously, but pre-treatment for both methods required complex liquid-liquid extraction [27,28].We hereby propose a method that is better than the reported ones in terms of number of compounds detected and sample pretreatment.
In this study, a UPLC-MS/MS detection method was established for the simultaneous determination of 71 neuropsychotropic drugs.During optimization of the method, it was found that the negative ion mode detection signals of phenobarbital, phenytoin, and valproic acid were higher and stabler than those obtained in the positive ion mode.Therefore, ammonium acetate was added to the mobile phase to improve the negative ion response, while taking into account that the vast majority of the drugs need to be detected in the positive ion mode.Acetic acid was also added to the mobile phase, and the concentrations of acetic acid and ammonium acetate were optimized for better compatibility of the signals with the two detection modes.
Our proposed method avoided the solvent effect by using methanol-acetonitrile as the protein precipitant, performing centrifugation at low temperature and high speed to separate the drugs to be tested, and adding methanol-water before sampling.The method was found to be simple and easy to perform and could meet the requirements of high-throughput TDM.The linear range of each drug covered the therapeutic concentration range provided by the AGNP guidelines and the laboratory alert concentration.All compounds except moclobemide N-indole were stable for 7 days at room temperature, but degradation of olanzapine, oxcarbazepine, bupropion, and other drugs stored at − 80 • C for 30 days was noted.However, the degradation could be avoided by the addition of 5 % 200 mg/mL vitamin C (final concentration of vitamin C was 10 mg/mL) as an antioxidant.In this case, the stability result of bupropion was in accordance with the AGNP guideline, and the stability result of olanzapine was inconsistent with the reported data [29][30][31].The stability of 10-hydroxycarbazepine is inconsistent with some studies, possibly because the sample used in our study was spiked serum rather than patient serum.When using UPLC-MS/MS for quantitative analysis, the ideal internal standard is the isotopic marker of the component to be measured.Accordingly, we used the respective isotopic markers of the drugs as the internal standard, so that the internal standard peak coincide with the analyte peak to reduce the matrix effect on quantification [32].However, there is one shortcoming: the matrix effect, we only evaluated one serum and this method did not fully assess the matrix effect.Another advantage of proposed method is that it requires a small sample volume of only 50 μL, which opens up the possibility of detecting samples from neonatal and pediatric populations and also reduces reagent costs.

Conclusion
The UPLC-MS/MS method proposed could successfully allow the simultaneous determination of serum drug concentrations of 71                         clinical neuropsychotropic drugs with excellent sensitivity and specificity.These properties are essential for the accurate identification and quantification of a wide range of neuropsychotropic drugs in the serum, ensuring that the method can reliably discriminate between compounds without cross-reactivity.In addition, the precision, accuracy, and stability of the method met the requirements of the relevant guidelines.Methanol-acetonitrile was used to precipitate proteins in this study.This method is very simple and easy to implement in a laboratory setting.In addition, the method requires a low sample volume of only 50 μL of serum.One of the most notable features of the method is the ability to obtain results rapidly.Each sample could be analyzed in less than 6 min, making our method significantly faster than many previously reported methods.A rapid turnaround is critical for timely clinical decision making, especially in cases of drug intoxication or clinical drug adjustments [33][34][35].

Declaration of competing interest
We have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Table 2
Mass spectrometry detection parameters of 71 psychotropic substances.

Table 3
Linear range, RSD of the assay at the lower limit of quantification, and bias for 71 neuropsychotropic drugs.

Table 4
Precision and accuracy of methods for the determination of 71 neuropsychotropic drugs.

Table 4
(continued on next page) W. Jin et al.

Table 5
Matrix effect evaluations of 71 neuropsychotropic drugs.
(continued on next page) W. Jin et al.

Table 5
(continued ) (continued on next page) W. Jin et al.