An immunophilin-binding assay for sirolimus

CLlNICALTHERAPEUTICS"/VOL.22,SUPPL.B,2000

An Immunophilin-Binding Diane L. Davis, MEd,l”

Assay for Sirolimus

and Steven J. Soldin, PhD’y2*4

‘Department qf Laboratory Medicine, Children k National Medical Center, ‘Biology Department, Catholic University of America, Washington, DC, 3Health Sciences Department, Salishury State University, Salishury, Maryland, and “Departmenh of Pediatrics and Pathology, George Washington Univers& School of Medicine, Washington, DC

ABSTRACT Objective: This review examinesthe performance of 4 assaysfor sirolimus in terms of their ability to meet 6 guidelines determined by a panel of experts. Background: Four methods have been described to date for the analysis of sirolimus concentrations in whole blood: high-performance liquid chromatography-mass spectrometry (HPLC-MS); microparticle enzyme immunoassay(MEIA); p70 S6 kinase inhibition; and an immunophilin-binding assay (IBA). Methods: A MEDLINE@ search of the literature was performed to identify relevant studies. Results: The HPLC methodssuffer from precision problemsbecauseof the substantial specimenpreparation required, and HPLC-MS methods are not practical for clinical use. Initial studiesof the MEIA have found overestimationof sirolimusconcentrationsthat may be causedby antibody cross-reactivity with sirolimus metabolites.Monitoring of sirolimus effects by p70 S6 kinase inhibition is as yet possibleonly theoretically, and the assayitself is not yet optimal. With the IBA, useof a T-cell protein that bindsto sirolimusand that may be the intracellular target of the drug as the assaybinding protein allows the assayto measure sirolimus selectively, even in the presenceof structurally similar metabolites. Conclusion: More than 200 clinical sampleshave beenanalyzed by the IBA, and correlation with HPLC values hasbeen good, with a regressionline slope near 1.0. In addition, the assayis easierto perform and more precisethan HPLC, and hasthe potential to be automated. Thus, the IBA appearsto have certain clear advantagesover the other assays. Key words: sirolimus, rapamycin, therapeutic drug monitoring, immunophilin-binding assay,radioreceptor assay,HPLC, MEIA. (Clin Ther. 2000;22[Suppl B]:B62-B70) Accepted for publication December Printed in the USA. Reproduction

B62

3, 1999. in whole or part

IS not permitted.

0149-291X/001$19.00

D.L. DAVIS AND S.J. SOLDIN

INTRODUCTION Sirolimus* is a macrolide compound isolated from Streptomyces hygroscopicus.’ The drug was initially studied for its antifungal properties, but substantialimmunosuppressiveeffects were noticed early in the trials.’ Although the drug had to be abandonedas an antifungal agent, it has since shown great promiseas an immunosuppressiveagent in transplant recipients. Indeed, clinical trials of the compound are currently under way in the European Union, and the drug hasreceived approval from the US Food and Drug Administration for use in renal transplantation. As with cyclosporine, sirolimus wholeblood trough concentrations should be maintained within a therapeutic window for efficacy.” Many immunosuppressive agentsare candidatesfor therapeutic drug monitoring, becausethere are seriousconsequencesof both undermedication(transplant rejection) and overmedication (infection and toxic side effects). Several studies have supported the assertionthat sirolimusbelongsin this category, because both rejection events and toxicity have been strongly associated with specific blood levels of sirolimus.ti It is therefore necessaryto establishdesirablelimits and methodsfor the therapeutic drug monitoring of sirolimus. In 1995, a panel of experts published consensusguidelines for the therapeutic monitoring of sirolimus.7It is useful to assesslaboratory methodsfor the analysisof sirolimusby the degreeto which they conform to these guidelines. Someof the important recommendationsof the guidelines are: (I) sirolimus metabolites have been “Trademark: Rapamm@ (Wyeth-Ayerst ries, Philadelphia. Pennsylvania).

Laborato-

shown to have no significant pharmacologic activity and thus should not be included in sirolimus measurements;(2) a same-day turnaround is optimal; (3) because-95% of sirolimus is sequesteredin red blood cells, whole blood collected in ethylenediamine-tetraaceticacid is the desired samplematrix; (4) between-day coefficients of variation (CV) should be 5 10%at concentrationsof 5 pg/L and ~5% at 40 pg/L; (5) the dynamic range of the assay should be between 1 and 75 kg/L; and (6) regressionstatistics generated by comparing methods should exhibit slopes between0.9 and 1.1, and an SE of the regression(SYiX)of <5 *g/L. It is also essentialto know the expected therapeutic range of a drug to determine the range of concentrationsthat will typically be measured.A definitive therapeutic range for sirolimus in humanshas not beenestablished;however, animal studies have shown an increasedincidence of rejection at concentrations ~10 kg/L and signs of toxicity at concentrations >60 kg/L.* A therapeutic rangeof 5 to 10 kg/L hasbeen suggestedfor steady-statetrough specimens.s However, recent evidence from 2 phase III trials of sirolimus suggeststhat the therapeutic window may be from 5 to 20 ng/mL (unpublished data). This review examinesthe performance of the assays used to determine levels of sirolimusin the lower concentrationrange. Relevant studieswere identified through a MEDLINE@ searchof the literature. To date, there are 4 published methods for sirolimus analysis: high-performance liquid chromatography-mass spectrometry (HPLC-MS), a microparticle enzyme immunoassay (MEIA), a pharmacodynamic assay for p70 S6 kinase activity, and an immunophilin-binding assay(IBA, also referred to as radioreceptor assay). B63

CLINICAL

Table I. Comparison of high-performance sirolimus concentrations.

Sample Size (mL)

Method Yatscoff

Napoli

et al’

and Kahan ‘”

Svensson

et al ’ ’

CV = coefficient

Throughput Estimates

Between-by cv (%)

methods for assaying

Sensitivity a.%/L) I .O (0.5 if 4.0.mL sample used)

Linearity or Highest Published Calibrator (kg/L) 250

2.0

230 min/injection

14.4 at IO (*g/L 9.8 at 50 kg/L

I.0

50 samples/24 h: 540 miniinjection

17at4&L 6 at 32 kg/L

I.0

50

I.0

IO miniinjection

14.2 at I kg/L 9.8 at 5 kg/L 5.6 at 40 )~g/lL

0.4

50

of variation.

The present review shows that, given the aforementioned criteria and the data published to date, the IBA appears to have clear advantages over the other methods. HIGH-PERFORMANCE CHROMATOGRAPHY

LIQUID

Data obtained with HPLC methods have been published by 3 research groups (Table I).“-” All HPLC methods have sensitivities well below the suggested therapeutic range using 2 I mL of whole blood, but all suffer from poor precision. Only the method described by Svensson et al” appears to be capable of meeting the goal of a CV of 510% at a concentration of 5 pg/L, with a CV of 9.8%; however, its CV of 5.6% at 40 kg/L does not meet the desired level of 55%. The precision of the other 2 methods also falls below the acceptable range. This problem is most likely the result of the extensive extraction/clean-up procedures necessary to prevent contaminants found in whole-blood B64

liquid chromatography

THERAPEUTICS’

specimens from interfering with the assay or ruining the analytical columns. Plasma cannot be substituted for whole blood in these methods to eliminate some of the variability of the extraction procedure, because only the more sensitive HPLC-MS methods have been shown to measure sirolimus levels in this fraction adequately. At this time, HPLC-MS methods are not deemed practical for clinical use.12 Other important limitations of the HPLC methods are the lengthy turnaround times and significant amount of hands-on time they require. All methods have extensive sample-preparation protocols, and the least amount of time required for each sample injection is 10 minutes.“-” The method reported by Napoli and Kahan”’ has injection times of 540 minutes, and the throughput is 50 samples per 24-hour period using an autosampler. Because any assay must include controls and calibrators, the number of patient samples processed using this technique is necessarily ~50. Same-day turnaround time is

D.L. DAVIS AND S.J. SOLDIN

therefore possiblefor only a limited number of specimens. There is not enough information to assesswhether and to what degreesirolimus metabolites interfere in the HPLC assays. The method reported by Yatscoff et al” assessedthe performance of the method using only whole blood spiked with parent drug. The other 2 studies’0,1’assessed the method’s performance on clinical samples that could be expected to contain metabolites, and the authors did not indicate whether anything other than parent drug was quantified. However, definitive studies in which each HPLC method is usedto assaypurified individual metabolites have not yet been published. MICROPARTICLE IMMUNOASSAY

ENZYME

The recently describedMEIA method for sirolimusdeterminationisa semi-automated prototype.‘” Although turnaround time and samplethroughput estimateswere not reported, the procedure can be expected to have significant advantagesover HPLC in this regard. Other aspectsof the MEIA, however, could present significant problems. The authors usedcontrols with concentrations of 5, 1I, and 22 kg/L, and reported interday imprecision of ~12% using the prototype. This result suggests that precision is outside the range suggested by the consensusguidelines. The authors state that accuracy and precision were determined over a range of 5 to 22 pg/L, but linearity and sensitivity were not reported specifically. At this time, the primary concern is the interference of sirolimus metaboliteswith the assay.A total of 747 clinical samples were analyzed by HPLC-MS and the MEIA assay.The regressionequation gen-

erated by the comparisonof the 2 methods had a slopeof 1.41and an intercept of I .3 1. The reported slope is well beyond the 0.9 to I. I range recommendedby the consensus committee, indicating that sirolimus levels may be seriously overestimated by this method. The authors suggestthat the bias of the method may be due to crossreactivity of sirolimusmetaboliteswith the antibody. Given that in steadystate, trough samplesof sirolimus metabolites are believed to constitute 56% f 9% of all sirolimusspecies,sthis is a reasonablehypothesis.If it is proved that the bias is due to suchcross-reactivity, then this method is not a good candidate for routine clinical use, as inclusion of sirolimus metabolites in therapeutic drug monitoring is not recommendedat this time.7 INHIBITION ACTIVITY

OF P70 S6 KINASE

It hasbeen suggestedthat the most meaningful way to monitor attainment of adequate levels of sirolimus and other immunosuppressive drugs is to assessthe inhibition of their cellular targets in the lymphocyte. Becausestudies are still being conducted to ascertainthe clinical validity of monitoring inhibition of p70 S6 kinaseactivity rather than sirolimus’4 levels, this is not yet a legitimate alternative. In addition, the published method does not appear to offer any advantages in terms of turnaround time or easeof performance, and the authors themselves state that the robustnessof the assay as currently conducted is poor.l4 IMMUNOPHILIN-BINDING

ASSAY

The biologic target of sirolimus is assumedto be a protein molecule contained B65

CLINICAL

THERAPEUTICS’

Table II. The minor immunophilins. lmmunophilin 52 kDaZ2,j0 37 kDa2’ I4 kDa2” S-8 kDaF

Source

Cyclosporine (nmol)

Calf thymus Jurkat T cells Calf thymus Calf thymus; Jurkat T cells

Kd

64.0 NA NA NA

Tacrol imus Kd (nmol)

Sirolimus Kd (nmol)

2.0 4.5 I.8 0.8

I.1 0.8 40.0 0.08

Kd = dissociation constant: NA = not applicable. ‘Davis DL, Murthy JN, Soldin SJ. unpublished data

in the cytosol of T cells. The class of cytosolic proteins that binds sirolimus and related immunosuppressivedrugs is called immunophilins,ts and the isolation of the precise immunophilin that mediates the biologic effects of sirohmushas beenvigorously pursued to help elucidate the exact inhibition pathway. The binding properties of this immunophilin can also be exploited in clinical assaysfor sirolimus.‘s An important theoretical advantage of using the correct immunophilin rather than an antibody is that the protein shouldbind both parent drug and its metabolitesin direct proportion to their biologic activity. Thus, interference from inactive metabolites could be eliminated.‘” Two major immunophilins have been isolated that bind cyclosporine (cyclophilin)th and tacrolimusisirolimus (FKbinding protein). “.” Both are present in micromolar concentrations in the cytosol and are unlikely biologic targets for drugs that are effective in nanomolarconcentrations.‘” Our group has therefore pursued the identification of low-abundance immunophilins. We have successfully isolated 4 minor proteins capable of binding sirolimus; a summary of their binding characteristicsis contained in Table II.20-22 B66

Using the 52-kDa minor immunophilin, our group was able to develop an IBA for sirolimus,2” and to date we have used the technique on 221 patient samples previously assayed by HPLC.2’m25The assay works well with 200- and 400-pL sample sizes, requiring less volume than is used for HPLC assays, and allows duplicate or triplicate runs. The precision of the assayalso surpassesthat of the HPLC and MEIA assays, and comes close to the consensus recommendations with CVs of 12.9%, 9.2%, 8.5%, and 5.9% at concentrations of 2.5, 7.5, 12.5, and 20 p.glL, respectively.2’ Also, SY,uis ~5, as suggestedby the consensusguidelines.2’ These favorable precision data are probably the result of a short extraction procedure involving a single-step direct extraction with absolute methanol. Because sample preparation is simpler and becausesamples incubate together for 30 minutes (as opposedto being individually injected in an HPLC), we estimatethat I person can analyze

30 to 50 samples

during

an S-

hour period, a substantialadvantage over HPLC.‘j This also makes the same-day turnaround time recommended by the committee a reasonableexpectation.

D.L.

DAVIS

AND

S.J. SOLDIN

0 0

10

20

30

40

50

HPLC (kg/L)

Figure. Comparisonof high-performanceliquid chromatography(HPLC) andan immunophilinbinding assay(IBA) for sirolimus.2”The regressionequation is y = 1.09x - 0.522, and the correlation coefficient is 0.977. The sensitivity of the assayis 1.O kg/L, well within the tolerance for clinical use.23 Linearity was excellent using standards ranging from 2.5 to 40 p,g/L, and no interference was detected from drugs commonly coadministeredwith sirolimus (cyclosporine, dexamethasone, prednisone, methotrexate).*” Perhapsthe most notable advantage of this IBA over the publishedMEIA method is the lack of significant interference from sirolimus metabolitesand the good correlation of the method with HPLC. To date,

all sirolimus metabolites tested have ~10% of the biologic activity of the parent molecule,26-28 so significant inclusion of metabolite concentrations in any measurement is undesirable. In 2 separate studies,23.2sour group has compared HPLC sirolimus levels with those obtained with IBA and found regressionline slopeswithin the consensusrange of 0.9 to 1.1 ( 1.09 in I996*” and 0.92 in 19992s), as shown in the figure. In addition, as shown in Table III, we have usedthe IBA to test individual purified sirolimus

Table III. Cross-reactivity of major sirolimusmetaboliteswith the 52-kDa immunophilin.29

Sirolimus

Metabolite,

25 kg/L*

7-0-demethyl sirolimus 4 I Odemethyl sirolimus 32,41-O-demethyl sirolimus (C9-C23)-OH sirolimus (C I -C8 or C32-C36)-OH sirolimus

Cross-Reactivity of IBA with lmmunophilin (‘%)

52.kDa

20.0 21.2 13.6
IBA = immunophilin-binding assay. *Numbering system begins at triene

B67

CLINICAL

metabolites at concentrations far above those that would be expected in steadystate trough clinical samples (25 p.g/L).2” It has been determined that the major metabolite is 4 1-O-demethyl sirolimus,’ and its level of interference was only 2 1.2%. Thus, it is reasonable to conclude that unless a patient has a significantly altered metabolite pattern, interference at therapeutic concentrations is low with this method. CONCLUSIONS We believe that, based on the available data, the method that comes closest to meeting the consensus guidelines is the IBA. The MEIA appears to overestimate sirolimus concentrations unacceptably, and the p70 S6 kinase inhibition assay has not yet shown good correlation with clinical events. The HPLC methods have precision problems. In addition, the IBA appears to have advantages over HPLC in terms of ease of performance, turnaround time, and sample size. Another important aspect of the IBA is that it can be automated and potentially improved beyond the statistics stated herein. Immunophilins theoretically can be cloned, and binding proteins for the assay could be mass-produced and used at a higher purity than was employed in these studies. We performed all pipetting, separation-column preparation, and elutions manually, and it is conceivable that automation could improve precision with more consistent pipetting and better reagent preparation and handling. Given that HPLC methods are probably already at maximal precision and assay efficiency, it seems realistic to expect that the IBA may be the best assay for blood sirolimus concentrations.“” B68

THERAPEUTICS

ACKNOWLEDGMENTS This work hasbeen partially funded over the past 12 years by Abbott Laboratories (Abbott Park, Illinois); Roche Laboratories (Nutley, New Jersey); and Bayer Corporation (Tarrytown, New York). Address correspondence to: Steven J. Soldin, PhD, Department of Laboratory Medicine, Children’s National Medical Center, 1I 1 Michigan Avenue NW, Washington, DC 20010. REFERENCES Sehgal SN, Baker H, VCzina C. Rapamycin (AY-22,989). a new antifungal antibiotic: II. Fermentation, isolation and characterization. J Amtihiot. 1975:28:727-732. Morris R. Rapamycins: Antifungal, antitumour, antiproliferative and immunosuppressive macrolides. Trmnspht Rei: 1992; 6:39. Abstract. MacDonald A, Scarola J, Burke JT, Zimmerman JJ. Clinical pharmacokinetics and therapeutic drug monitoring of sirolimus. C/in Tkr. 2000;22:(Suppl B):B 101-B 121. Fryer J, Yatscoff RW. Pascoe EA. Thliveris J. The relationship of blood concentrations of rapamycin and cyclosporine to suppression of allograft rejection in a rabbit heterotopic heart transplant model. 7ico7.spl~ititcrtior7. 1993;55:340-345. Yakimets WJ, Lakey JR, Yatscoff RW, et al. Prolongation of canine pancreatic islet a-

lograft survival with combinedrapamycin and cyclosporine therapy at low doses. Rapamycin efficacy is blood level related. Trcmsplcrtltation. 199.136: 1293-l 298.

G-angerDK, CromwellJW, ChenSC, et al. Prolongation

of renal allograft survival

D.L. DAVIS AND S.J. SOLDIN

in a large animal model by oral rapamycin monotherapy. 7’ransplanration. 1995;59: 183-l 86.

15. Soldin SJ. Immunophilins: Their properties and their potential role in therapeutic drug monitoring. Ther Drug Monir. 2000; 22:44-46.

7. Yatscoff RW, Boeckx R, Holt DW, et al. Consensus guidelines for therapeutic drug monitoring of rapamycin: Report of the consensus panel. Ther Drug Monit. 1995; I7:676-680.

16. Takahashi N, Hayano T, Suzuki M. Peptidyl-prolyl cis-trans isomerase is the cyclosporin A-binding protein cyclophilin. Nature. 1989;337:473475.

8. Trepanier DJ, Gallant H, Legatt DF, Yatscoff RW. Rapamycin: Distribution, pharmacokinetics and therapeutic range investigations: An update. Clin Biochetn. 1Y98;3 I :34s-35 1.

17. Harding MW, Galat A, Uehling DE, Schreiber SL. A receptor for the immunosuppressant FKS06 is a cis-trans peptidylprolyl isomerase. Nature. lY8Y;341:758760.

P. Y Yatscoff RW, Faraci C, Bolingbroke Measurement of rapamycin in whole blood using reverse-phase high-performance liquid chromatography. Ther Drug Mkt. 3992;14:138-141.

18. Bierer BE, Mattila PS, Standaert RF, et al. Two distinct signal transmission pathways in T lymphocytes are inhibited by complexes formed between an immunophilin and either FK506 or rapamycin. Proc Nut1 Aud SC; U S A. 1990;87:9231-9235.

10. Napoli KL. Kahan BD. Routine clinical monitoring of sirolimus (rapamycin) whole-blood concentrations by HPLC with ultraviolet detection. Clitz Chem. 1996; 42: 1943-1948. II. Svensson J-O, Brattstriim C, SIwe J. Determination of rapamycin in whole blood by HPLC. Ther Drug Mkt. lYY7;19: 112-l 16. 12 Kahan BD, Murgia MG, Slaton J, Napoli K. Potential applications of therapeutic drug monitoring of sirolimus immunosuppression in clinical renal transplantation. Ther Drug Mot&. 1995;17:672-675. 13 Salm P, Taylor P, Pillans P. Evaluation of a microparticle enzyme immunoassay (MEIA) for whole blood sirolimus determination in renal transplant recipients. Ther Drug Monir. 1999;2 1:442. Abstract. 14. Yatscoff RW, Aspeslet LJ, Gallant HL. Pharmacodynamic monitoring of immunosuppressive drugs. C/in Chem. 1998; 44:428a32.

19. Schreiber SL. Chemistry and biology of the immunophilins and their immunosuppressive ligands. Science. 1991;25 I :283-287. 20. Murthy JN, Chen Y, Soldin SJ. Identification of a 14 kDa FK-506irapamycin binding immunophilin from calf thymus. C/in Biochrm. lYY4;27:357-365. 21. Mm-thy JN, Goodyear N, Soldin SJ. Identification of a 37 kDa tacrolimus, sirolimus and cyclosporine binding immunophilin possessing glyceraldehyde 3-phosphate dehydrogenase activity isolated from the Jurkat T cell line. Clin Biochem. 1997; 30:129-133. 22. Palaszynski EW, Donnelly JG, Soldin SJ. Purification and characterization of cyclosporine and FK-506 binding proteins from a human T-helper cell line. Clin Biochem. 199 I ;24:63-70. 23. Goodyear N, Napoli KL, Murthy JN, et al. Radioreceptor assay for sirolimus. C/in Biochetn. 1996;29:457~!60. B69

CLINICAL

24. Goodyear N, Napoli KL, Murthy JN, et al. Radioreceptor assay for sirolimus in patients with decreased platelet counts. Clirz Biochem. I997;30:539-543. 25. Davis DL, Murthy JN, Soldin SJ. Comparison of steady-state trough sirolimus samples by HPLC and a radioreceptor assay. Clin Biochern. 2000;33:3 l-36. 26. Christians U, Sattler M, Schiebel HM, et al. Isolation of two immunosuppressive metabolites after in vitro metabolism of rapamycin. Drug Met& Dispos. 1992~20: 186-191. 27. Wang C, Lim H, Chan K, et al. High performance liquid chromatographic isolation, spectroscopic characterization and immunosuppressive activities of three major metabolites from the plasma of rats re-

B70

THERAPEUTICS”

ceiving rapamycin (sirolimus) orally. J Liq Chrotnntogr Relat Technol. 1995; 18: 2559-2568. 28. Wang C, Chan K, Schiksinis R, et al. High performance liquid chromatographic isolation, spectroscopic characterization and immunosuppressive activities of two rapamycin degradation products. J Liq ChrotnntoXr Relat Tedno/. 1994;17:3383-3392. 29. Davis DL, Murthy JN, Soldin SJ. Minor immunophilin binding of tacrolimus and sirolimus metabolites. Clirl Biochem. 2000;33: l-6. 30. Soldin SJ. Receptor (immunophilin-binding) assay for immunosuppressive drugs. Tiler Drrtg Morzit. 1995: 17:574-576.