Reding,2 Jean Paul Squlfflet,3. Tatiana. Besse,3 and Pierre E. Wallemacqâ4. Tacrolimus is a relatively new immunosuppressant used in organ transplantation to ...
CLIN. CHEM. 41/9, 1292-1 296 (1995)
#{149} Drug
Monitoring
and
Toxicology
HPLC-Microparticle Enzyme Immunoassay Specific for Tacrolimus in Whole Blood of Hepatic and Renal Transplant Patients flham Besse,3
Firdaous,’ Alexandre Hassoun,’ and Pierre E. Wallemacq”4
Jean Bernard
Otte,2 Raymond
Tacrolimus is a relatively new immunosuppressant used in organ transplantation to prevent graft rejection. However, its use is not devoid of side effects, making it important to maintain blood concentrations within therapeutic ranges. Several analytical methods are currently available for routine drug monitoring. However, these methods are based on use of the same monoclonal antibody, which also cross-reacts with some metabolites, resulting in overestimation of some blood concentrations. Even though this antibody appears appropriate
for therapeutic drug monitoring, no reference method measures only the parent drug, mainly because of the poor absorptivity of tacrolimus in ultraviolet light. We have developed a method displaying an increased specificity towards the unchanged drug, using conventional equipment available in most clinical laboratories. After chromatographic separation of the blood extract, the tacrolimus fraction is analyzed by an automated microenzyme immunoassay (MEIA) performed on the lMx analyzer (Abbott Labs.). This method is linear from 0 to 40 g/L, yields CVs from 8.5% to 18.2%, and has a detection limit of 5 g/L. Tacrolimus concentrations particle
obtained by HPLC-MEIA in hepatic and renal transplant patients are from 47.5% to 18.8% lower than those obtained by MEIA, according to liver function tests and metabolite accumulation, even though no significant differences were observed between the methods for drugfree blood samples supplemented with known amounts of tacrolimus. Indexing Terms: immunosuppressants/chromatography, liquid//rnmunoassay Tacrolimus (FK 506) is a macrolide characterized by similar but more potent immunosuppressive properties than cyclosporine (CsA), inhibiting cell-mediated and humoral immune responses (1). Currently under cliiiical investigation in organ transplantation, FK 506 has been shown to possess strong immunosuppressive ac-
Departments of ‘Clinical Chemistry, 2Hepatic Transplantation, and3 Kidney Transplantation, University Hospital St. Luc, University of Louvain, 10 Hippocrate Ave., B-1200 Brussels, Belgium. 4Address correspondence to this author at: Department of Clinical Chemistry, Laboratory of Therapeutic Drug Monitoring, University Hospital St. Luc, U.C.L. 10 Hippocrate Ave., B-1200 Brussels, Belgium. Fax hit + 3227643732. 5Nonstandard abbreviations: CsA, cyclosporin A (cyclosporime); MEIA, microparticle enzyme immunoassay; TDM, therapeutic drug monitoring; UV, ultraviolet; MLR, mixed lymphocyte reaction; and AST, aspartate aminotransferase. Received February 16, 1995; accepted June 12, 1995. 1292
CLINICAL CHEMISTRY, Vol. 41, No. 9, 1995
Reding,2
Jean
Paul Squlfflet,3
Tatiana
tivity in the mixed lymphocyte reaction (MLR), nearly 100 times more than CsA. The experience gained in the clinical use of FK 506 suggests quite a narrow therapeutic window (from 5 to 15 gi in whole blood) and reveals, mainly in the higher concentrations, some nephrotoxicity as one of its major side effects. These factors underscore the need for careful blood concentration monitoring in transplant patients to optimize FK 506 therapy. An ELISA has been developed by Fujisawa Pharmaceuticals (Osaka, Japan) (2), followed by a modified shortened version (3), and a kit produced by Incstar (Pro-TracTM FK 506 ELISA; Incstar, Stillwater, MN) to measure FK 506 in plasma or whole blood. A semiautomated immunoassay, namely, the microparticle enzyme immunoassay (MEIA; IMx#{174} Tacrolimus), has been developed by Abbott Diagnostics (Chicago, IL) (4) with the same monoclonal antibody, which allows the detection of FK 506 in whole blood for routine therapeutic drug monitoring (TDM). However, different studies (5-8) show that the monoclonal antibody used in both assays cross-reacts with some FK 506 metabolites, yielding a certain overestimation of the FK 506 concentrations. Some studies have been performed on plasma (5), on serum (6), and on microsomal preparations and purified metabolites (7). This overestimation is undesirable
for (e.g.)
pharmacokinetic
analyses,
even
though the assay may be useful for TDM purposes since some cross-reacting metabolites are active. Since the molecule does not have a satisfactory molar absorptivity at ultraviolet WV) wavelengths, it is not easily detected at therapeutic concentrations by HPLC methods involving UV detectors. To evaluate the concentrations of the unchanged drug, we developed an experimental method combining an HPLC procedure, to separate FK 506 from all other compounds, with a MEIA technique. The performance of the combined HPLC-MEIA procedure was determined with whole-blood samples supplemented with known amounts of FK 506, and results were compared with those by the original MEIA. In addition, 308 clinical blood samples drawn from renal or hepatic transplant patients were analyzed by both methods, allowing some comparison and correlation with the clinical
Materials
status.
and Methods
Materials
Tacrolimus (as vials of 100 g/L in methanol) provided by Fujisawa Pharmaceuticals.
kindly
was The
following solvents were used in the extraction procedure or in the HPLC method: acetonitrile, methanol, hexane, propan-2-ol, all HPLC grade, from Lab-Scan (Dublin, Ireland). We used as chromatographic column a Spherisorb CN (3 m X 15 cm) from Alltech (Deerfield, IL), and 1.5-mL polypropylene Eppendorf#{174} (Hamburg, Germany) microtubes. The chromatographic equipment consisted of a ternary pump (Model 9010; Varian, Walnut Creek, CA), a diode-array detector (Model 1040 M; Hewlett-Packard, Palo Alto, CA), and a data processor computer (Hewlett-Packard, Model 9153C and HP 9000 series 300 computer). We used the programmable column heater from the Marathon (Spark, The Netherlands) Autolnjector. A column/valve mounting module was fastened directly to the side of the pump, with a Rheodyne valve (Model 7125) and a 50-FL injection loop. FK 506 concentrations were obtained by MEIA on the IMx analyzer according to the manufacturer’s instructions (Abbott), including daily quality controls. Blood samples (-1.5 mL), drawn into EDTA tubes, were obtained from 20 liver transplant recipients (12 pediatric/8 adult) and 12 adult kidney transplant recipients at different times. The pediatric hepatic transplant population included seven girls and five boys, median age 2 years (0.7-13 years), who underwent primary transplantation between December 1992 and November 1993 for biliary atresia, Byler disease, and some metabolic disease. The adult liver transplant population involved compassionate treatments, and the kidney transplant population included patients of age 18 years or more with endstage kidney disease. The blood samples were stored at 4 #{176}C immediately after collection and analyzed within 48 h. Whole blood (drug free), obtained from healthy volunteers at the local blood bank, was used either to prepare supplemented blood calibrators/controls, specimens to evaluate the recovery yield, or, in the last step of the combined procedure, to reconstitute FK 506 blood samples before MEIA analysis. This study was approved by the local ethical committee.
Methods
HPLC-MEIA FK 506 concentrations in whole blood were deterby a combined HPLC-MEIA of four major steps: 1) Pretreatment of the whole-blood sample. 2) Isocratic normal-phase HPLC separation of the extract. 3) Collection of the peak corresponding to FK 506. 4) Quantification by MEIA. Acetonitrile (400 L) is added to 200 L of whole blood in Eppendorfmicrotubes. The mixture is agitated 15 mm, then centrifuged 5 mm at 9500g. The clear supernate is transferred into a second microtube for evaporation at 40 #{176}C. The dry residue is reconstituted with 70 L of propan-2-ol, agitated 30 s, and injected (50-p.L injection loop) into the CN column heated at 40 #{176}C. The mobile phase, a mixture of hexane/propan2-ol (72/28 by vol), is pumped through the column at a flow rate of 1.8 mlimin. The retention time of FK 506 is -4.5 mm, which is controlled regularly throughout the study by injecting a large amount of FK 506 detectable by UV light. The fraction corresponding to FK 506 is collected between 4.1 and 4.9 mm, evaporated, and reconstituted with 100 L of drug-free whole blood. Such a sample is treated with 200 L of the tacrolimus whole-blood precipitation reagent for MEIA analysis according to the IMx procedure. The calibration curve was determined by supplementing drug-free blood with FK 506 to obtain 0, 6.25, 10, 12.5, 20, 25, 30, and 40 pg/L, following a procedure similar to that for the patient’s blood sample. mined
Assay Validation Inter- and intraassay
CVs were obtained by assaying human blood samples supplemented with 15 and 25 gfL FK 506. In addition, the reproducibility of the method was assessed with pooled blood, collected from different patients taking FK 506, analyzed daily as frozen aliquots. The sensitivity of the combined method is defined as the lowest measurable concentration that can be distinguished from zero with 95% confidence. The recovery yield was determined (n 10) for the whole procedure, at 15 and 251.gfL, as the ratio drug-free
=
MEIA. This immunoassay (4) involves the anti-FK 506 monoclonal antibody provided by Fujisawa Pharmaceuticals in the ELISA. This antibody presents good correlation between MLR immunosuppressive activity and cross-reactivity (8). This assay requires 100 L of whole blood, and 24 samples can be completed in -40 min after a rapid organic extraction by adding 200 L of tacrolimus whole-blood precipitation reagent. The capture reagent consists of latex microparticles to which FK 506 antibodies are attached. The enzyme is the FK 506 alkaline phosphatase conjugate reagent, and the substrate is 4-methylumbelliferyl phosphate. The ratio at which methylumbelliferone is generated on the matrix is inversely proportional to the concentration of FK 506 in the sample. The detection limit of this assay is 5 pg/L (4). A 10% CV for the quality controls was considered acceptable to validate the data.
[FK R(%)
506]PL
r,n1 X ir Zi
OJUJsupp1emented
x100
where [FK 506],, represents the FK 506 concentration obtained by MEIA after HPLC separation, and [FK 506lsupplemented represents the known FK 506 concentration in the whole blood before extraction. The factor 2 in the denominator arises from the fact that the initial 200 pL of the blood specimen is finally reconstituted in 100 gAL used in the MEIA procedure. The linearity of the assay was assessed by linear regression analysis on mean stratified data from supplemented blood samples (n 51) in the range 0-40 /Lg/L. Analytical recovery was assessed with blood samples =
CLINICAL CHEMISTRY, Vol. 41, No. 9, 1995
1293
from eight different hepatic transplant patients taking FK 506, each blood sample being divided into four 200ILL aliquots. The FK 506 concentration (x) was measured in the first aliquot by HPLC-MEIA, and the three others were supplemented with different known amounts of FK 506 to obtain final theoretical blood concentrations ofc1 x + 6.25 g.LgIL,c2 x + 12.5 g1.gfL, and c3 x + 25 gLgfL subsequently analyzed again by HPLC-MEIA. The extrapolated concentration (c,,) should be virtually the same as the first measured concentration (x). The specificity of the method was first evaluated with 26 drugs and 9 endogenous compounds tested for potential interference with the assay (3). Moreover, supplemented blood samples and blood samples from transplant patients on FK 506 therapy were analyzed by both HPLC-MEIA and MEIA (Fig. 1). A discrepancy between concentrations obtained from both methods for patients’ samples would suggest either the presence of an endogenous component in patients that might have interfered with the direct MEIA but been removed by HPLC, or, more likely, the presence of metabolites detected by MEIA. To better determine the role of the liver function in the accumulation process of FK 506 metabolites, we categorized blood samples in four groups according to hepatic tests such as aspartate aminotransferase (AST) and conjugated serum bilirubin concentrations. Group I consisted of liver transplant patients with a severe hepatic dysfunction corresponding to day 1 posttransplantation. Groups II and IH consisted of the same patients, respectively 3 days and -4 weeks (before discharge from hospital) posttransplantation. Group IV represented kidney transplant patients selected for their normal liver function (normal ranges: AST 5-25 U/L; conjugated bilirubin 0-3 mg/L) at steady state a few days after the onset of FK 506 therapy. =
=
=
45 40
35
y
0.9742x R’
30
=
+
0.0522
0.9987
25
20
Statistical analysis was performed by using the paired Student’s t-test to compare FK 506 concentrations measured by HPLC-MEIA and MEIA in hepatic and renal transplant patients. Results The absolute recovery yield of the combined procedure (HPLC-MEIA) was independent of tacrolimus concentration and has been evaluated as 51% ± 6% (mean ± SD). Linear regression analyses performed between observed HPLC-MEIA values (y) and direct MEIA concentrations (x) with either supplemented blood samples ranging from 0 to 40 p.gfL (n 51; 7 stratified data) or blood samples from kidney-transplanted patients (n 54; 7 stratified data) and livertransplanted patients (n 254; 14 stratified data) are ifiustrated in Fig. 1. The detection limit of the MEIA beings p.gfL, we could evaluate no concentration below this point. The results obtained from the recovery test are particularly satisfactory for concentrations >10 g/L: The following concentrations (x) 3.9, 6.5, 8.7, 11.4, 11.4, 15.3, 16.8, and 21.5 corresponded to the extrapolated concentrations (c0) 4.6, 7.6, 6.0, 10.5, 11.2, 13.9, 16.5, and 21.0. Inter- and intraassay performances of the HPLC-MEIA assay are illustrated in Table 1, which shows CVs of 8.5% to 18.2% and 9.8% to 10.1%, respectively. These values are obviously higher than by MEIA alone. The reproducibility test performed with the pooled blood showed a mean value of 7.9 ± 1.4 tgfL and an interassay CV of 17.1%. None of the drugs and clinical chemistry analytes tested (Table 2) displayed significant interference in the assay at their respective ranges of concentration. As shown in Table 3, analyses of drug-free blood samples supplemented with 15 or 25 pgfL FK 506 yielded similar values independent of the method (HPLC-MEIA or MEIA). However, whole-blood samples drawn from hepatic transplant patients on FK 506 therapy showed significantly lower values by HPLC-MEIA (Table 3 and Fig. 1). Mean FK 506 concentrations obtained by this specific assay in group I (day 1 posttransplantation), group II (day 3), and group III were 47.5%, 31.4%, and 29.9% lower than those obtained by MEIA, respectively. This strongly suggests a participation of some cross-reacting FK 506 metabolites in the MEIA analy=
=
=
a
1. ComparatIve
Table
15 y
10
=
0.4625x
+
3,41 55
methods
R’=0.9085
Mean
performances of analytical used In this study. CV, %
a SD, pglL
5 Method
0 0
10
20
30
40
50
Intraassay
1294
CLINICAL CHEMISTRY, Vol. 41, No. 9, 1995
At 25 pglL
At 15 gsg/L At 25 pg/L
1.7 23.5 ± 2.3 (10)
10.1
9.8
14.7 ± 2.7 24.7 ± 2.1 (18) (16)
18.2
8.5
16.0 ± 1.4 23.9 ± 1.4 15.3 ± 1.5 24.1 ± 1.4
8.8
5.9 5.8
17.2 ±
(10
Mean MEIA conc. pgIL Fig. 1. Linear regression analyses of HPLC-MEIA data and direct MEIA data obtained from supplemented FK 506 samples (#{149}; n = 51 stratified in 7 mean concentrations), kidney transplant samples (A; n = 54 stratified in 7 mean concentrations), and liver transplant samples (0; n = 254 stratified in 14 mean concentrations). Stratification was done to maintain each stratified subgroup with a homogeneous number of data.
At 15 zg/L
HPLC-MEIA
Interassay MEIA (n
=
20 each)
Intraassay Interassay
No. of assays is listed In parentheses.
9.7
sis, decreasing when liver function improves. Wholeblood samples drawn from steady-state renal transplant patients selected for their normal liver function had mean FK 506 concentrations 18.8% lower by __________________________________________________ Table 2. Drugs and biological analytes tested to assess interference In the HPLC-MEIA. Blood conc, mg/L
Drug name Erythromycin
2.5 and 5 30
Amikacin Vancomycin
50
Gentamicin Tobramycin Netilmicin Theophylhne Phenobarbital Valproic acid Carbamazepine
10 10 10
20 40 75 12 25
Phen “oin
16 100
Salicylate bigitoxin
500
Digoxin Disopyramide Procainamide and NAPA Quinidine
0.003 4 10 and 15 4 1 5 5 0.75 and 1.50 0.1 and 1 0.1 and 1
o.oz
Lidocaine Methylprednisolone Cyclosponne Azathioprine
Diltiazem
Plasma conc, g/l.
Biological analytes Bilirubin Protein (from proteinemia) p-Lipoprotein 1
0.300 88
2.55
-Acid glycoprotein
1.91
Triglycerides
4.220 and 8.640
Rheumatoid factor
1/512 (Waaler Rose)
Gammopathies IgA
2.3
1gM lgG
0.2 2.6
Table 3. Mean difference between known amount Mean AST, U/L Supplemented Group I Group II Group III Group IV
describmg, besides the FK 506 peak, a cross-reacting and active metabolite present in some patients (particularly those with severe hepatic impairement). The quantitative HPLC-MEIA method used in this study shows an increased specificity towards the parent drug as compared with the direct MEIA. However, although most metabolites should elute later because of their increased hydrophilic property, and should therefore be separated from the tacrolimus fraction collected at -4.5 mm, a partial and most likely minor contamination of this tacrolimus fraction by some metabolites cannot be excluded. Further studies are required to assure and confirm a total specificity. The difference between both assays appears consistent with another recent report (13), which compared FK 506 plasma concentrations obtained either by ELISA or by HPLCELISA. This group observed similar or even higher discrepancies in the case of liver dysfunction, which is not surprising, since both immunoassays involve the same monoclonal antibody. Our data show clearly the role played by hepatic function and particularly by the elimination function of the liver in the accumulation process of metabolites, which could reflect either a global increase of the whole pattern of metabolites or a specific increase of some cross-reacting metabolites. This aspect should be kept in mind when interpreting a FK 506 blood concentration in transplanted patients with different clinical status. Fig. 1 illustrates clearly
HPLC-MEIA and MEIA results in drug-free blood samples supplemented of FK 506 (15 or 25 g/L), and according to liver function.
with
% dIfference between HPLC-MEIA and MEIA
Mean conjugated blllrubln, mgIL
n
Mean
SD
Range
P
70.5 62 5 1
54 58 97 99 54
1.2 47.5 31.3 29.9 18.8
14.5 26 20.3 16.5 18
0-21 0-83 0-82 0-69 0-50
NS
