Simple Extraction Protocol for Analysis of Immunosuppressant Drugs in Whole Blood

The immunosuppressant drugs cyclosporine (cyclosporin A), tacrolimus, sirolimus, and everolimus are used to prevent organ rejection. Because commercial immunoassays are available only for cyclosporine and tacrolimus, HPLC (1)(2)(3), often in conjunction with mass spectrometry (MS or MS/MS) (4)(5)(6)(7), has been used to quantify these drugs. For MS/MS, sample preparation has often involved simple protein precipitation and solvent dissolution. Protocols have used zinc sulfate followed by acetonitrile (8), methanol (9), or acetone (10). Sometimes these reagents are premixed (11)(12)(13)(14). During in-house assay development we encountered several problems when zinc sulfate and acetonitrile were used. For example, many specimens clumped when blood was added to the zinc sulfate. Even after addition of acetonitrile and vortex-mixing, the clumps remained, necessitating manual dislodgement of the pellet. This occurred with lysed human blood specimens, lyophilized whole-blood controls (e.g., Bio-Rad Lyphochek), and College of American Pathologists proficiency testing specimens. We substituted methanol for acetonitrile with minor improvement. Another problem was that the recovery of sirolimus from whole blood using zinc sulfate followed by acetonitrile or methanol was <100% (9). In addition, the observed MS/MS responses for commercial calibrators and controls were often ∼30% higher than for human blood specimens. Here we present a new extraction protocol that improves absolute recoveries, provides excellent precision, and shows less ion suppression. Although the performance of this protocol for several immunosuppressants is described, we present example data only for sirolimus. We also describe a solid-phase extraction (SPE) that may be added to further enhance the cleanliness of extracts. For recovery and extraction experiments, we used cyclosporine from Novartis, tacrolimus from Fujisawa, and sirolimus from LC Laboratories. These same materials were used to prepare whole-blood calibrators, except for tacrolimus, for which Abbott Tacrolimus II calibrators were used. For precision studies we used Bio-Rad Lyphochek whole-blood controls. The original zinc sulfate precipitation that we used was similar to previous publications. We added 50 μL of 0.1 mol/L aqueous zinc sulfate to a 1.5-mL polypropylene microcentrifuge tube. To this we added 50 μL of whole blood, followed by immediate vortex-mixing. We then added 500 μL of methanol containing 8 μg/L 32-desmethoxyrapamycin for sirolimus, 8 μg/L ascomycin for tacrolimus, or 20 μg/L cyclosporin D for cyclosporin A. After vortex-mixing for 30 s and centrifugation, the supernatant was analyzed. In our new protocol we placed 50 μL of patient blood, calibrator, or control in a polypropylene microcentrifuge tube. To this we added, with no intermediate mixing, 250 μL of deionized water, followed by 250 μL of aqueous 0.1 mol/L zinc sulfate, and finally 500 μL methanol containing the appropriate internal standards. After addition of all components, the tube was vortex-mixed for 30 s. After 5–10 min the tubes were centrifuged for 4 min, and the colorless supernatant was analyzed. SPE was performed with a 25-mg, 1-mL Varian LMS crossed-linked styrene divinylbenzene column. The column was conditioned with 1 mL of methanol followed by 1 mL of water. The supernatant prepared above was passed slowly through the column (1–2 mL/min). The column was washed twice with 1 mL of water and air-dried under reduced pressure for ∼30 s. The drugs were eluted into injection vials with 750 μL of acetonitrile. Before injection, 300 μL of water was added to each vial to make the solvent composition compatible with the initial mobile phase. For HPLC-MS/MS analyses we used an Agilent 1100 binary system and a Waters Quattro Micro mass spectrometer. In the positive-ion mode the monitored multiple-reaction monitoring transitions (m/z) were: cyclosporin A, 1219.7→1202.7; cyclosporin D, 1233.8→1216.8; sirolimus 931.6→864.5; 32-desmethoxyrapamycin, 901.5→834.4; tacrolimus, 821.4→768.3; and ascomycin, 809.4→756.3. Separation was performed with a Phenomenex 4 × 3 mm (i.d.) C18 guard column maintained at 50 °C. The injection volume was 30 μL with a mobile phase flow rate of 0.4 mL/min. The mobile phases were as follows: (A), 2 mmol/L ammonium acetate and 1 mL/L formic acid in water; and (B), 2 mmol/L ammonium acetate and 1 mL/L formic acid in methanol. The gradient program was 50% B for 0.1 min, followed by an immediate change to 90% B at 0.11 min. At 1.8 min, the mobile phase was increased from 90% to 100% B to clean the column. At 2.3 min, the mobile phase reverted to 50% B. Sirolimus, tacrolimus, 32-desmethoxyrapamycin, and ascomycin eluted at ∼1.6 min, whereas cyclosporin A and D eluted at ∼1.8 min. Extraction efficiencies were determined by enriching donor blood samples, stored at 4 °C, with sirolimus (10 and 20 μg/L), cyclosporine (100 μg/L), and tacrolimus (4 and 20 μg/L). Peak areas of the extracted samples were compared with areas obtained for injections of pure drug in saline (corrected for volume or dilution). After complete extraction, including SPE, recoveries were 95% for sirolimus, 95% for tacrolimus, and 103% for cyclosporine. These values are comparable to those reported by Deters et al. (7) and better than those reported by several other groups (6)(9)(11)(12)(13)(14) for other procedures. The improvements in absolute recovery from deidentified patient specimens, stored at 4 °C, obtained with our extraction protocol are shown in Fig. 1 , panels A and B. Comparisons were performed within a single run to avoid day-to-day variation in instrument response. The increase in absolute signal for sirolimus, corrected for any dilution, averaged 36% with the new water hemolysis protocol. To evaluate whether the difference was attributable to ion suppression, specimens were also subjected to further SPE purification. After SPE, a similar increase of 39% in absolute recovery was observed, supporting the fact that, for sirolimus, the new protocol provided greater extraction efficiency. A similar increase in area response of 29% was observed for 32-desmethoxyrapamycin. A comparison of area response for cyclosporine and tacrolimus showed mean increases of 21% and 31%, respectively, with our water hemolysis extraction protocol. A comparison of the interrun CVs (Table 1 ) showed an improvement for sirolimus, tacrolimus, and cyclosporine. The limits of quantification (± 20%) were 0.5 μg/L for sirolimus, 0.2 μg/L for tacrolimus, and 5 μg/L for cyclosporine. Interrun CVs. Tacrolimus (n = 10), cyclosporine (n = 10), sirolimus (n = 12). Interrun CVs. Tacrolimus (n = 10), cyclosporine (n = 10), sirolimus (n = 12). Comparison of results obtained with the existing and proposed hemolysis methods. (A and B), comparison of sirolimus signal (m/z 931.6→864.5) for zinc sulfate/methanol extraction protocol and the proposed water hemolysis/zinc sulfate/methanol protocol. (C–E), ion suppression profiles for zinc sulfate/methanol extract (C), water hemolysis/zinc sulfate/methanol extract (D), and water hemolysis/zinc sulfate/methanol plus SPE (E). The arrows in panels C–E indicate the elution time window for sirolimus under the chromatographic conditions described in the text. We evaluated ion suppression, using a 3 μL/min postcolumn infusion of a methanolic solution of sirolimus that provided a signal similar to a specimen containing 3 μg/L drug (15). For a Lyphochek whole-blood control containing no sirolimus, the original extraction protocol showed ion suppression at several points in the chromatographic profile (Fig. 1C ), whereas the water hemolysis protocol yielded cleaner extracts (Fig. 1D ). Ion suppression was further reduced if SPE was performed (Fig. 1E ). During initial development, we became aware that some laboratories observed more consistent results for sirolimus if vortex-mixed tubes were allowed to sit before centrifugation. We included a 10-min incubation with our water hemolysis protocol, although we have not evaluated the necessity of this step. There are several advantages to this new extraction protocol. One advantage is that the water helps lyse and uniformly dissolve blood components. Another advantage is that the use of water hemolysis and methanol as the solvent provides a desirable “milkshake” consistency after vortex-mixing. In addition, the extraction can be carried out with addition of all reagents followed by a single vortex-mixing step. Finally, this extraction could be adapted to an autodiluter for specimen delivery because water allows the specimen to be flushed from the dispenser with no precipitation in the line. There are potential benefits to the addition of SPE. Ion suppression can be further eliminated. SPE also allows extraction of larger specimen volumes. One purported advantage of MS/MS is the ability to monitor mass transitions for multiple drugs within the same analytical run. However, as the number of compounds being monitored increases, instrument dwell time or number of data points acquired for each mass transition may decrease, which could negatively affect the limit of quantification or precision. The use of a larger sample size in conjunction with SPE can help restore the observed response for the desired mass transitions.