Can oral midazolam predict oral cyclosporine disposition?

Abstract

Effective cyclosporine therapy is confounded by large interindividual differences in oral bioavailability and a narrow therapeutic window. Because cytochrome P450 (CYP) 3A-mediated first-pass metabolism contributes to this unpredictable bioavailability, an in vivo oral CYP3A phenotyping probe could be a valuable tool in optimizing cyclosporine therapy. Based on similarities in the metabolic kinetics of cyclosporine and midazolam by the liver and intestinal mucosa, we evaluated whether midazolam oral clearance would predict cyclosporine oral clearance when the two drugs were administered to 20 medically stable kidney transplant recipients. Despite earlier findings in liver transplant recipients who displayed a strong correlation between the systemic clearances of midazolam and cyclosporine, there was a weak correlation between their oral clearances in the current group of subjects (r_s=0.50, P=0.03). Differing extents of intestinal first-pass metabolic extraction between the two drugs, inhibition of midazolam metabolism by cyclosporine at the level of the intestine, and/or P-glycoprotein-mediated intestinal efflux of cyclosporine (but not midazolam) may account for this poor correlation. We conclude that although oral midazolam is unlikely to be clinically useful as a probe for cyclosporine disposition, its utility in the prediction of other orally administered CYP3A substrates cannot be out ruled.

Introduction

Cyclosporine A (cyclosporine) is a cyclic undecapeptide that has revolutionized the area of immunosuppressant therapy for solid organ transplantation since its introduction in the early 1980s (Beveridge and Calne, 1995; Ponticelli et al., 1996). Although we have gained considerable clinical experience in the use of cyclosporine over the years, management of therapy is still not without its difficulties. Not only does the drug have a narrow therapeutic window, it is notorious for exhibiting large interindividual variation in its oral bioavailability. That for the conventional formulation, Sandimmune^®, ranges from less than 5 to 89% (Ptachcinski et al., 1986). Consequently, suboptimal dosing is common and can lead to either graft rejection or nephrotoxicity. Factors that can contribute to cyclosporine’s unpredictable oral bioavailability include low aqueous solubility and poor intestinal permeability, variable hepatic and intestinal first-pass metabolism, and intestinal P-glycoprotein-mediated active efflux (Benet et al., 1996). Although the microemulsion formulation, Neoral^®, has improved absorption properties over Sandimmune^® (Noble and Markham, 1995), it still appears to suffer from significant and variable first-pass metabolism and active intestinal efflux, as evidenced by an average oral bioavailability that ranges from 21 to 73% (Wallemacq et al., 1997; Chueh and Kahan, 1998; Ku et al., 1998). Since cyclosporine is eliminated almost entirely by metabolism — less than 1% of an administered dose appears in the bile and urine as unchanged drug (Ptachcinski et al., 1986) — many investigators have focused on this contributing factor to interindividual variability in cyclosporine oral bioavailability.

Cyclosporine is biotransformed to more than 20 metabolites, all of which retain the cyclic structure (Christians and Sewing, 1993). Kronbach et al. (1988), using human liver microsomes, identified cytochrome P450 3A (CYP3A)¹ as the principal enzyme subfamily catalyzing the formation of the three primary metabolites observed in vivo (AM1, AM9, and AM4N). In vivo and in vitro investigations later revealed that CYP3A in the small intestine, in addition to the liver, participates in the first-pass metabolism of cyclosporine (Kolars et al., 1991; Webber et al., 1992). Moreover, CYP3A in the small intestine may be more susceptible to the effects of inhibitors (Watkins, 1992; Gomez et al., 1995; Lown et al., 1997; Gorski et al., 1998) and inducers (Hebert et al., 1992; Wu et al., 1995; Fromm et al., 1996; Holtbecker et al., 1996) compared to CYP3A in the liver. Therefore, a priori assessment of combined hepatic and intestinal CYP3A activity could aid in the optimization of cyclosporine dosage regimens as well as in the identification of individuals most at risk to interactions involving cyclosporine and known (e.g., ketoconazole, itraconazole, diltiazem, erythromycin, grapefruit juice, rifampin, phenytoin) (Watkins, 1992; Compana et al., 1996; Ku et al., 1998; Foradori et al., 1998) or as yet unknown CYP3A activity modulators.

A clinically applicable method to assess an individual’s ability to metabolize CYP3A substrates requires an enzyme-selective probe. Several are described in the literature and include the intravenous erythromycin breath test (Watkins et al., 1989); the oral dapsone recovery ratio (May et al., 1992, May et al., 1994); the oral dextromethorphan/3-methoxymorphinan urinary ratio (Jones et al., 1996); and the urinary ratio of endogenous 6β-hydroxycortisol to free cortisol (Bienvenu et al., 1991; Horsmans et al., 1992). Each has applications and recognized limitations (Watkins, 1994; Streetman et al., 2000). For example, although the erythromycin breath test has been used successfully to assess hepatic CYP3A levels (Cheng et al., 1997; Lown et al., 1997; Hirth et al., 2000), it cannot be used to assess intestinal levels as intravenous administration avoids significant metabolism by the gut. Dapsone is administered orally, thus formation of its hydroxylamine metabolite is likely a reflection of both intestinal and hepatic CYP3A activities. However, another CYP isoform, CYP2E1, was found to account for the majority of dapsone hydroxylamine formation at concentrations observed in vivo, raising questions of CYP3A selectivity (Mitra et al., 1995). Dextromethorphan is also administered orally, but in addition to the polymorphically expressed CYP2D6 being the major enzyme responsible for the overall metabolism of this drug, at least a 72-h urine collection is required (11-day if the subject is a CYP2D6 poor metabolizer) (Jones et al., 1996), rendering this test impractical (Kashuba et al., 1999). Finally, for reasons not well understood, the urinary ratio of 6β-hydroxycortisol to free cortisol appears to only be a sensitive marker for CYP3A induction and is a poor measure of basal CYP3A activity (Ged et al., 1989). Collectively, these findings indicate the need for a versatile, yet specific, CYP3A probe.

Our laboratory and others have investigated the utility of midazolam as an in vitro and in vivo probe for both intestinal and hepatic CYP3A metabolic activity (Gorski et al., 1994; Lown et al., 1994; Thummel et al., 1994, Thummel et al., 1996; Paine et al., 1996, Paine et al., 1997). Like cyclosporine, midazolam is eliminated almost entirely by metabolism (Dundee et al., 1984) and principally by CYP3A (Fabre et al., 1988; Kronbach et al., 1989). In addition, orally administered midazolam undergoes significant intestinal and hepatic first-pass metabolism (Thummel et al., 1996; Paine et al., 1996). Hence, midazolam has the desirable characteristics of a CYP3A phenotyping probe that should predict intravenous and oral cyclosporine disposition. Indeed, midazolam clearance was found to correlate strongly with cyclosporine clearance after the two CYP3A substrates were administered intravenously to the same group of liver transplant recipients (r=0.81, P<0.001) (Thummel et al., 1994).

Based on these earlier findings, the presystemic and systemic clearance characteristics of orally administered midazolam could parallel those of orally administered cyclosporine. Accordingly, we conducted a study in kidney transplant recipients to determine whether the oral clearance of midazolam would predict the oral clearance of cyclosporine.