A Robust Procedure for Determination of Immunosuppressants Cyclosporine A and Tacrolimus in Blood Samples with Detection of LC–MS/MS

The immunosuppressants monitoring is crucial in the treatment of transplant patients to avoid the risk of organ rejection. So it is vital to develop an accurate method to obtain precise and reliable results. Herein, a liquid chromatography tandem mass spectrometry (LC–MS/MS) based method was developed and validated, which could simultaneously realize the accurate quantification of cyclosporine A and tacrolimus with a total run time of 4.5 min. Unlike the previously described methods, the highlight of this assay is that there is no using zinc sulfate in the protein precipitation of the whole blood sample, which can effectively prolong the lifespan of mass spectrometry. In addition, the linearity, matrix effect, carryover, accuracy, intraday and interday precision of this method were investigated. The linear range of tacrolimus and cyclosporine were 1–40 ng/mL and 50–2000 ng/mL, respectively; The correlation coefficients (R2) were both greater than 0.995. The intraday and interday precision of cyclosporine and tacrolimus were ≤ 10%. The lower limit of quantitation (LLOQ) of tacrolimus and cyclosporine A were 1 ng/mL and 50 ng/mL, respectively. The main features of this method are good accuracy, sensitivity, specificity, rapidness, robustness and high throughput. It is suitable for therapeutic drug monitoring of cyclosporine A and tacrolimus in routine clinical diagnostics.


Introduction
In the field of solid organ transplantation, cyclosporine A and tacrolimus are most commonly used immunosuppressive agents, because they have great potential to prevent allograft rejection and decrease patient morbidity and mortality [1][2][3][4][5][6]. However, due to their narrow therapeutic window and high drug toxicity, excessive immunosuppressant can induce a series of side effects, while insufficient dosage of immunosuppressant or low drug concentration in blood may lead to organ rejection [1,[7][8][9][10]. The adverse effects of cyclosporine are mainly nephrotoxicity and hypertension [2,11,12]. For recipients with increased blood pressure after postoperative application of cyclosporine or hypertension, available treatments include taking oral antihypertensive drugs, combining with mycophenolate mofetil to reduce the dose of cyclosporine, or switching to tacrolimus. In addition, adverse effects caused by cyclosporine also include hepatotoxicity, neurotoxicity, hypercholesterolemia, hyperuricemia, hyperkalemia, tremor, gingival hyperplasia, diabetes and hypertrichosis. The main side effects of tacrolimus are nephrotoxicity, neurotoxicity and diabetes. Other side effects of tacrolimus include tremor, bacterial infection, cytomegalovirus infection and gastrointestinal reaction.
Particularly, immunosuppressive agents show a great degree of inter-individual and intra-individual pharmacokinetic and pharmacodynamic variability [8,[13][14][15][16][17]. Moreover, the age, weight, combined medication, food and fluid intake of a patient can alter the metabolization of the drugs [18,19]. Therefore, it is of great significance for transplant recipients to do therapeutic drug monitoring (TDM) regularly to maintain the concentration of immunosuppressant 1 3 within the therapeutic range and minimize the toxic effects and risk of acute rejection after organ transplanted. [15,20] Up to now, there are many available analytical methods in therapeutic drug monitoring, e.g., microparticle enzyme immunoassay (MEIA), enzyme multiplied immunoassay technique (EMIT), fluorescent polarization immunoassay (FPIA), high performance liquid chromatography (HPLC), liquid chromatography-tandem mass spectrometry (LC-MS/ MS). [8,16,21] Immunoassay results tend to show significant positive bias due to cross-reactivity with metabolites, similarstructure drug or heterophilic antibodies [16-18, 22, 23]. LC-MS/MS is considered as the gold standard method for monitoring the above-mentioned immunosuppressants [18,[24][25][26] because of its high specificity and sensitivity [5,27]. In addition, LC-MS/MS is able to simultaneously measure several drugs in one single analytical run. [28] However, many LC-MS/MS methods adopt the pretreatment methods of liquid-liquid extraction and solid-phase extraction, which is time-consuming and complex [3]. And other LC-MS/MS methods commonly used zinc sulfate (ZnSO 4 ) and methanol or acetonitrile as protein precipitant [29][30][31]. Zinc sulfate is a nonvolatile inorganic salt which is difficult to remove after entering the mass spectrum. The instrument is easy to get dirty, which increases the frequency of maintenance and shortens the lifespan of the instrument. Furthermore, zinc sulfate solution is the main factor causing ionization suppression. [10] Here, we exhibit a rapid and reliable LC-MS/MS method, which does not use zinc sulfate and has broad linearity range covering the full therapeutic ranges of the analyzed drugs. The pretreatment procedures are extremely simple and do not involve time-and moneyconsuming steps.

Chemicals and Stock Solutions
Tacrolimus, cyclosporin A were purchased from Sigma-Aldrich (USA). The internal standards (IS) tacrolimus-13C and cyclosporin-d12 were purchased from Cmass (China). HPLC grade acetonitrile and methanol were obtained from Merck (Germany). LC-MS grade formic acid (FA) was purchased from Fisher scientific (USA). LC-MS grade ammonium formate was purchased from Fisher Chemical (Germany). A millipore purification system was used to produce ultrapure water (Merck Millipore).

Calibration Standards and Quality Control Samples
Starting from a 1 mg/mL concentration for tacrolimus, first, 50 μg/mL sub-stock solution of tacrolimus was prepared by further diluting the original stock solution using acetonitrile. Then 40 μL sub-stock solution of tacrolimus and 100 μL cyclosporine A original stock solution (1 mg/mL) was added to 860 μL of 50% methanol to prepare the working calibration standards which containing 2000 ng/mL of tacrolimus (TAC) and 100,000 ng/mL of cyclosporine A (CsA). This solution was further diluted by 50% methanol to 7-point calibrators: 20.0, 40.0, 80.0, 160.0, 320.0, 600.0, 800.0 ng/ mL for TAC and 1000, 2000, 4000, 8000, 16,000, 30,000, 40,000 ng/mL for CsA. For the purpose of method validation only, quality control (QC) working solutions of TAC and CsA were prepared in the same way but from different stock solutions. Three concentration levels prepared were 60.0, 180.0, 640.0 ng/mL for TAC and 3000, 9000, 32,000 ng/mL for CsA, respectively.
Commercially available QCs (ClinChek ® Whole Blood Controls, Germany) are used for internal quality assurance in routine clinical examination. Concentrations are 3.49, 7.11, 14.4 ng/mL for tacrolimus and 51.0, 105, 204 ng/mL for cyclosporine A, respectively. QCs were implemented before and after sample batch every day.
All stock solutions and working solutions were stored at − 70 °C and light-protected until analysis. Batches were stable for at least 6 months. One batch of calibrators and QC samples were thawed 10 min before sample preparation.

Sample Preparation
Since immunosuppressants significantly accumulate in red blood cells, whole blood is the preferred specimen of choice for the determination of CsA and TAC [4,5,17]. First, for the analysis of EDTA-anticoagulated patient blood samples, whole blood samples were shaken absolutely. Consequently, 100 μL of the samples were transferred into 1.5 mL Eppendorf tubes. Followed by the addition of 400 μL acetonitrile containing internal standard for CsA and TAC to precipitate protein. Then, they were vortexed for 10 s, sonicated for 10 s and centrifuged for 10 min at 10,000 rpm. Sonication was critical to guarantee adequate recovery of the immunosuppressants by breaking down sample components. Finally, 1 3 200 μL supernatant was transferred to 600 μL water in a sample bottle. After vortexing for 4 s, it was placed in the autosampler and an aliquot of 6 μL was injected into the LC-MS system.

Instrumentation and LC-MS/MS Method
A 4500MD triple quadrupole mass spectrometer (AB SCIEX, for in vitro Diagnostic use) coupled with Jasper LC System and peak gas generation (Genius AE 32) was used to determine the concentration of immunosuppressants in the whole blood. Data processing was performed on Analyst MD 1.6.2 Software.
Mobile phase A consisted of 2 mM ammonium formate with 0.1% formic acid (v/v) in water, which was also used as a weak wash solvent. Mobile phase B consisted of 2 mM ammonium formate with 0.1% formic acid (v/v) in methanol, which was also used as a strong wash solvent. 50% Methanol (v/v) was used as a seal wash. The mobile phase flow rate was maintained at 0.3 mL min −1 . From zero to 0.4 min, isocratic conditions were run with 55% B. Solvent B was increased from 55 to 98% in the time range from 0.4 to 0.9 min. Then, solvent B was kept at 98% for 2 min. The mobile phase reverted to 55% B from 2.9 min to 2.95 min. Re-equilibration was performed from 2.95 to 4.5 min at 55% B. The total instrumental analysis time was 4.5 min.

Mass Spectrometric Conditions
The mass spectrometer was operated in positive electrospray source ionization mode (ESI+). Detection parameters applied to both whole blood samples were as follows: curtain gas 30.0 μL/min, collision gas 7 mL/min, the source temperature was set at 400 °C, the ion spray voltage at 5500 V, ion source gas 1: 50 psi and ion source gas 2: 50 psi. Entrance Potential (EP)10.0, Collision Cell Exit Potential (CXP) 7.0. For each immunosuppressant, two transitions were followed: one of them was used for quantification (the quantifier), and the other was monitored for identification (the qualifier). The quantifier to qualifier ratio was used for peak identification based on criteria set forth by CLSI C62-A guidelines.
Multiple reaction monitoring (MRM) transitions for each compound are summarized in Table 1.

Analytical Method Validation
The method validation of whole blood was based on the European Medicines Agency (EMA), U.S. Food and Drug Administration (FDA) guidelines on bioanalytical method validation and C62-A. EDTA-anticoagulated blood samples from healthy volunteers were used as blank whole blood. The following performance parameters were evaluated: linearity, selectivity, carryover, lower limit of quantitation, recovery, matrix effect, the intraday and interday precision, accuracy and stability. In addition, the experiment was validated using external quality assessment (EQA) program, and passed with a score of 100. Results: All performance parameters were within acceptance criteria, and hence, it was determined that the validated method is fit for clinical routine analysis.

Linearity
To establish the relationship between the concentration of the analytes and the detector response, 7-point calibrators were prepared for two immunosuppressants, covering the entire range of expected concentrations in patient blood samples.
Generally, trough concentration of immunosuppressants were detected for therapeutic drug monitoring. In recent years, many studies have found that peak levels of CsA correlate better with the clinical outcome as compared to the trough levels [5]. So the calibration curve of CsA ranged from 50 to 2000 ng/mL. It is much higher than the quantitative upper limit of immunoassay (500 ng/mL) [36], which can decrease the error involved from dilution. Linearity was evaluated by linear regression. Each calibrator was injected on three consecutive days. Response was defined as the peak area ratio of the analytes to the internal standards (y) and was plotted against the nominal concentration (x) of each drug. The weighting of 1/x 2 was chosen for both analytes, calibration curves, the trueness and precision of the lower limit of quantitation (LLOQ) were < 15% bias and CV < 20%, the trueness and imprecision of other calibrators should not exceed 15%. Table 2 shows collecting the values for linear regression results, linear range and R 2 of CsA and TAC in human blood, respectively.
During routine analysis, calibration curves were run within each batch of unknown samples. A typical analytical sequence consisted of a calibration curve, QC samples and followed by patient samples. After all patient samples detected, QC samples were detected repeatedly.

Specificity/Selectivity and Carry-Over
Specificity/selectivity means the capability of differentiating or selecting the targeted analyte from its complex matrix. As shown in Fig. 1, in daily clinical routine calibrating, Double Blank (no analyte, no IS) and Blank (no analyte, with IS) samples were injected before injecting seven calibrators. According to the EMA guideline, the absence of interfering components is accepted when the peak area response of interfering peak at the retention time of analyte (each immunosuppressant in our case) is less than 20% of the LOQ for the analyte and 5% for the IS.
To assess carryover, a Double Blank (no analyte, no IS) sample was injected immediately after the highest calibrator. Criteria for acceptability included that the blank sample area following the highest calibrator should not be > 20.0% of the LLOQ area for each immunosuppressant and > 5.0% for the internal standards. No carry-over was observed for any drugs as shown in Table 3.

Lower limit of Quantification (LLOQ)
The LLOQ defined as the lowest point of a calibration curve was the lowest concentration of analyte that gave precision and accuracy values within the limits of condition (CV ≤ 20%, bias < 15%), and a signal-to-noise ratio (S/N) of at least 10 [3,7,32,33]. It was then determined experimentally by analyzing six replicates samples in three batches during different days. Calculated the average value, bias and CV. According to concentration of the immunosuppressants in the literature, CsA and TAC were clinically found at 150-1000 ng mL −1 and 4-70 ng mL −1 in human blood of transplanted patients, respectively [34,35], the LLOQ of this method (TAC: 1 ng/mL, CsA: 50 ng/mL) can meet the clinical demand easily. In fact, if the lower limits of quantification of CsA and TAC were required, the lower concentration can be achieved, because the signal-to-noise ratio (S/N) as shown in Table 4 is far greater than 10.
As illustrated in Table 5, the LC-MS/MS method we developed provides a favorable sensitivity compared with the hitherto reported protocols for TAC and CsA detection. Furthermore, the present assay approach exhibits advantages of simultaneous detection, no use of zinc sulfate and

Extraction Recovery
To evaluate the loss of sensitivity caused by sample pretreatment. According to the sample preparing procedure, three levels of QC samples in blank whole blood matrix were processed, n = 3 for each group, and the peak area A was acquired. Then, the blank whole blood without adding internal standards were processed in accordance with the sample preparing method. Consequently, a certain amount of IS and QC (LQC\MQC\HQC) was spiked into the supernatant, n = 3 for each group. Subsequently, the peak area B was acquired by injecting the samples into the instrument. The recovery can be defined as the ratio of the peak area response of adding standard before matrix extraction (Area A) to the peak area response of adding standard after matrix extraction (Area B) [12].
Also, the internal standard extraction recovery was calculated the same as the analytes.
As shown in Table 6, recovery after sample extraction resulted in a coefficient of variation (cv) of 2.5% for TAC, 3.2% for CsA. This indicated that analyte losses during sample pretreatment was consistent, the developed sample process procedure was reliable.

Matrix Effect
The matrix components of whole blood generally include salts, lipids, proteins, organic small molecules and so on. Salts and lipids (especially phospholipids) are the most important factors affecting the ionization efficiency. Most matrix interferences can be removed by sample pretreatment or chromatographic separation. Matrix effect was assessed by analyzing blank whole blood samples from six different healthy volunteers as six different matrix sources. First, according to the sample preparing method, samples were pretreated without adding internal standards, and then a certain amount of IS and (LQC/HQC) was spiked to the supernatant. And the peak area B was acquired by injecting the samples into the instrument. Determine same amount of (LQC/HQC) and IS in the pure solution, and the peak area C was acquired, n = 3 for each group, calculate the average value.
Quantitative measurement of the matrix effect can be defined as the ratio of the peak area response in the presence of the matrix (Area B) to the peak area response in pure solution of analyte (Area C). Calculate the matrix factor and the cv% of 6 matrixes.
Also, the IS-normalized matrix effect was calculated as same as the analytes.  According to the CLSI C62-A, the variability in matrix effect as measured by the coefficient of variation (cv) should be less than 15%, so as shown in Table 7, this method can solve matrix problems.

The Within-Day and Between-Day Precision
The within-day precision was measured by replicate analysis (n = 6) of the three levels of QC samples in whole blood matrix during each analytical run. The between-day precision was determined by evaluating the QC samples during three different analytical runs on three consecutive days. The acceptance criterion for the bias was 15%, except for the LLOQ, where it was 20%.
For the two immunosuppressive drugs, the within-day and between-day precision were both within the acceptable limits stated for bioanalytical method validation (Table 8). According to these results, the assay is accurate and precise enough for the studied concentration range.

Stability
The stability of the samples included the stability of the unprocessed whole blood samples and extracted samples in the autosampler. According to the actual situation of the laboratory, it demands a short intra-laboratory turnaround time. Patient samples were detected as soon as possible, except fault diagnosis and maintenance when

Proficiency Testing
To compare the performance against other laboratories using the similar method, the laboratory has participated in the external quality assessment (EQA) program (organized by National Center for Clinical Laboratory). Results of the proficiency testing program are presented in Tables 9 and 10. Using the above workflow, the external quality assessment (EQA) program was full score passed with a low bias, which proved that this method could be successfully used in a clinical laboratory performing routine TDM service.

Conclusion
In summary, by improving the pretreatment method of protein precipitation, a fast, simple and robust LC-MS/MS method for the simultaneous determination of TAC and CsA was developed and fully validated according to EMA guidelines. The linear ranges were 1-40 ng/mL for TAC and 50-2000 ng/mL for CsA. The LLOQ of TAC and CsA were 1 ng/mL and 50 ng/mL, respectively. Moreover, the high selectivity and specificity of LC-MS/MS method without cross-reactivity between the parent drug and its metabolites can prevent the result overestimation. This procedure has been successfully used to detect more than 20,000 patient samples in our clinical laboratory with high customer satisfaction. In the future, we will focus on improving the automation and standardization of the method.