Inhibition of sulfamethoxazole hydroxylamine formation by fluconazole in human liver microsomes and healthy volunteers*

Inhibition of sulfamethoxazole hydroxylamine formation by fluconazole in human liver microsomcs and healthy volunteers

Sulfamethoxazole toxicity is putatively initiated by the formation of a hydroxylamine metabolite by cytochromes P450. I f this reaction could be inhibited, t o , c i ty may decrease. We have studied--in vitro and in vivo--fluconazole, ketoconazole, and cimetidine as potentially suitable clinical inhibitors of sulfamethoxazole hydroxylamine formation. Both fluconazole and ketoconazole in human liver microsomal incubations competitively inhibited sulfamethoxazole N-hydroxylation, with the inhibitory constant (K 0 values of 3.5 and 6 ~mol /L , respectively. Cimetidine exhibited a mixed type of inhibition of sulfamethoxazole hydroxylamine formation in human liver microsomes, with ICs0 values (the concentration required to decrease hydroxylamine formation by 50%) of 80 and 800 ~mol /L , the lower value being observed when cimetidine was preincubated with microsomes and reduced nicotinamide adenine dinucleotide phosphate. In an in vivo study in six healthy volunteers the inlfibition of the cytochrome P450- mediated generation of the toxic metabolite in the presence of fluconazole was shown by a 94% decrease in the area under the plasma concentration-time curve of sulfamethoxazole hydroxylamine. In contrast, the recovery of hydroxylamine in urine decreased by only 60%. Total clearance of sulfamethoxazole was decreased 26% by fluconazole, most likely because of the inhibition of unidentified P450 elimination pathways. There was close agreement between the predicted (87%) and observed inhibition (94%) of sulfamethoxazole hydroxylamine formation in vivo. Similarly, there was close agreement between in vivo and in vitro I~ values--l .6 and 3.5 Ixmol/L, respectively. (CHN PmutstaCOL THER 1996;59:332-40.)

A s h o k e K . M i t r a , P h D , a K e n n e t h E . T h u m m e l , P h D , T h o m a s F. K a l h o r n , BS ,

E v a n D . K h a r a s c h , M D , P h D , J a s h v a n t D . U n a d k a t , P h D , a n d J o h n T . S l a t t e r y , P h D Seattle, Wash.

The arylamine antibiotic sulfamethoxazole is fre- quently used in combination with tr imethoprim for the t reatment and prophylaxis of Pneumoo, stis carinii pneumonia a common opportunistic infection in human acquired immunodeficiency syndrome

From the Department of Pharmaceutics, the Department of Anesthesiology, and the Department of Medicinal Chemistry, University of Washington.

Supported by grants AI 27664 and GM 32165 from the National Institutes of Health (Bethesda, Md.).

Received for publication March 30, 1995; accepted Sept. 15, 1995.

Reprint requests: John T. Slattery, PhD, Department of Pharma- ceutics, Box 357610, University of Washington, Seattle, WA 98195-7610.

~'Present address: Procter & Gamble Company, Miami Valley Laboratories, Cincinnati, OH 45239.

Copyright © 1996 by Mosby-Year Book, Inc. 0009-9236/96/$5.00 + 0 13/1/69387

3 3 2

(AIDS). Although the drug is very effective for this indication, its use is associated with adverse reactions sufficiently severe to require discontinuation of therapy for active infection in 57% of patients with AIDS who have Pnemnocystis carinii pneumonia. "- The most common side effects in this patient popu- lation are leukopenia and rash; less common are fever (which often accompanies rash), thrombocyto- penia, azotemia, and elevations of alanine transam- inase or alkaline phosphatase. 3-5 These adverse effects are observed even when sulfamethoxazole is used alone. 6"7

It has been suggested that the pathogenesis of these side effects depends on the formation of reac- tive arylating metabolites of sulfamethoxazole, 8 and the hydroxylamine metabolite of sulfamethoxazole has been implicated in a variety of the drug's adverse reactions. 9"13 Sulfamethoxazole hydroxyl-

CLINICAL PHARMA(;()LOGY & THERAI'EUTiCS VOLUMI" 59. NUMBER 3 M i t r a et al. 333

amine is excreted in urine after administration of the drug to humans and is formed in human liver microsomes in a cytochrome P450-catalyzed reaction. H Recently, we reported the involvement of CYP3A4 and CYP2C isoforms in the N-hydroxylation of sulfamethoxazole in human liver microsomes.~5 However, no information on the clinically applicable inhibitors of the formation of sup famethoxazole hydroxylamine has been reported.

The purpose of this investigation was to identify a clinically applicable inhibitor of the N-hydroxylation of sulfamethoxazole in vitro and to evaluate its effect in healthy individuals.

MATERIAL AND METHODS

Chemicals. Sulfamethoxazole hydroxylamine was provided by Dr. Polly Sager of the National Institute of Allergy and Infectious Diseases (Bethesda, Md.). Ketoconazole was purchased from United States Pharmacopeial Convention Inc. (Rockville, Md.). Fluconazole (Diflucan) was obtained from Pfizer Inc. (New York, N.Y.). Cimetidine, sulfamethoxazole, reduced nicotinamide adenine dinucleotide phosphate (NADPH), and reduced glutathione were purchased from Sigma Chemical Co. (St. Louis, Mo.). All other chemicals were of analytic grade and were obtained commercially.

Microsomes and incubations. Procurement, prepa- ration, and storage of human livers from organ do- nors has been described previously, l~' Human liver microsomes were prepared from four livers by dif- ferential centrifugation. ~7 Microsomal protein was determined by the method of Lowry et al. I~ with use of bovine serum albumin as the standard.

Sulfamethoxazole N-hydroxylation was evaluated in a shaking water bath at 37 ° C in a reaction mixture (0.5 ml) that contained 0.5 to 1.0 mg microso- real protein, 15 mmol/L Hepes buffer (pH 7.4), 1.0 mmol/L NADPH, 1 mmol/L glutathione, and 0.5 mmol/L sulfamethoxazole dissolved in dimethyl sulfoxide (final dimethyl sulfoxide concentrations in reaction mixtures were <-0.8%). Glutathione was added to stabilize the sulfamethoxazole hydroxylamine and to prevent its autooxidation to nitroso- sulfamethoxazole. ~9 Fluconazole was added directly to the microsomai suspension, substrate was added 3 minutes later, and cofactor was added after an additional 2 minutes of preincubation. Ketoconazole and cimetidine were initially dissolved in acetone, appropriate amounts were transferred to the incu- bation vials, and the acetone was evaporated under a stream of nitrogen. For ketoconazole, microsomes

were added to the residue of ketoconazole and preincubated at 37 ° C for 3 minutes before the addition of substrate. Cofactor was added 2 minutes after the addition of substrate. Cimetidine was preincubated with microsomes two ways: (1) with 0.5 mmol/L NADPH for 15 minutes, with the reaction initiated by the addition of substrate and an additional 0.5 mmol/L NADPH, or (2) in a separate experiment, with microsomes but without NADPH for 15 minutes, the reaction was initiated by the addition of substrate and cofactor. Reactions were terminated after 10 minutes by the addition of 300 ill ice-cold acetonitrile. The terminated reaction mixture was maintained on ice for 5 minutes to precipitate protein and then centrifuged at 3100g for 20 minutes at 4 ° C. Aliquots (10 ~zi) of the supernatant were analyzed for sulfamethoxazole hydroxylamine.

Sulfamethoxazole hydroxylamine was assayed by an HPLC method with electrochemical detection. A reversed-phase Cis column was used (Econosphere 5 Ixm, 4.6 x 250 ram, Alltech Associates, Deerfield, Ill.) coupled to a Hewlett-Packard 1050 series chro- matograph (Hewlett-Packard Company, Palo Alto, Calif.). The mobile phase was water/acetonitrile/ acetic acid/triethylamine (75:25:1.0:0.05, vol/vol) at a flow rate of !.0 ml/min. Retention time of sulfamethoxazole hydroxylamine was 5.5 minutes as monitored by a single-cell electrochemical detector (Hewlett-Packard mode[ 1049A) at a potential of +0.3 volts. The slopes of sulfamethoxazole hydroxylamine for standard curves run daily were linear over the range of 5 to 25 pmol, with within- and between-day coefficients of variation of 4.3% and 7%, respectively.

In vivo studies. Six nonsmoking, healthy male volunteers (age range, 26 to 38 years; health assessed by interview) who gave written consent for the study were included. The investigational protocol was ap- proved by the University of Washington Human Subjects Committee. None of the subjects received any medication before or during the course of the study.

Because the half-life (tl/2) of fluconazole is approximately 40 hours, all subjects first received a single oral dose of a combination product that contained 800 mg sulfamethoxazole and 160 mg trimethoprim, and the pharmacokinetics of sulfamethoxazole, acetylsulfamethoxazole, and sulfamethoxazole hydroxylamine were assessed through serial blood samples and urine collection as described below. Fif- teen days later, each received 400 mg fluconazole (2 x 200 mg Difluean tablets) 24 hours before again receiv-

334 M i t r a et al. CLINICAl . P H A R M A C O L O G Y & "I+HERAPEUTI(~ ~,

MARCH 1996

-1o 2"0 t 1.6

~, 1.2 ere

"d

0o8"

0.4"

0.0 - 1 0 0

m 1.0 m b l SIVlX * 0.6 m M SMX • 0A mbl SMX

1 0 2 0 3 0 40

KgrOCONAZOLE (~'B

e,e

011

lit , lo+++ • I 0.4 mM SMX I

4

2

0 - 1 0 0 10 2 0 3 0 4 0

A

F L U C O N A Z O L E (pM)

B

Fig. 1. Inhibition of sulfamethoxazole N-hydroxylation by the antifungal drugs ketoconazole and fluconazole. Ap- parent Ki values from the Dixon plots were 6 ~mol/L for ketoconazole (A) and 3.5 i~mol/L for fluconazole (B); in both cases, inhibition was competitive. Values indicate means of duplicate experiments. SMX, Sulfamethoxazole, SMA-HA, sulfamethoxazole hydroxylamine.

ing sulfamethoxazole-trimethoprim. Fluconazole (200 mg) was ingested every 24 hours after the first dose for the duration of the study. No food or beverage was allowed f~om 2 hours before to 2 hours after ingestion of sulfamethoxazole-trimethoprim.

On each occasion, sulfamethoxazole was ingested, venous blood (5 ml) was collected over sodium eth- ylenediaminetetraacetic acid, and 1 mmol/L glutathione at 1, 2, 3, 4, 5, 6, 8, 10, 12, 24, and 31 hours after ingestion of sulfamethoxazole-trimethoprim through a heparin lock (the first 3 ml were discard- ed). All blood samples to be measured were centri-

fuged, and the plasma was frozen at -70°C until analyzed. Each subject emptied his bladder before ingestion of sulfamethoxazole, and urine was collected from 0 to 72 hours over 2 gm ascorbic acid to stabilize the hydroxylamine. All collection bottles were maintained at 4 ° C throughout the period of collection. Urine volumes were measured for each collection interval, and aliquots were frozen at -80 ° C until analyzed.

Plasma and urine samples were analyzed for sulfamethoxazole, N-acetylsulfamethoxazole, and sui- famethoxazole hydroxylamine by a reversed-phase HPLC method as described above, with minor modifications, including the insertion of an ultraviolet detector (254 nm; Hewlett-Packard series 1050) in series with the electrochemical detector. Plasma (0.5 ml) was combined with 500 i~1 ice-cold acetonitrile, vortexed briefly, and allowed to stand over ice for 5 minutes to precipitate protein before centrifugation at 3300g for 25 minutes at 4 ° C. Twenty p+l of the supernatant was injected by a Hewlett-Packard 1050 autosampler (Hewlett-Packard Company) and analyzed as described. Sulfamethoxazole and metabolites from urine were assayed by dilution of 500 t~1 of urine with 500 !~1 of ice cold acetonitrile, centrifugation at 3300g for 20 minutes at 4 ° C, and injection of 5 to 15 ~1 of the supernatant. Calibration curves for calculating the concentrations of experimental data points were drawn with spiked plasma that contained sulfamethoxazole (1 to 50 I~g/mi), N-acetylsulfamethoxazole (1 to 15 Ixg/mi), and sulfamethoxazole hydroxylamine (25 to 150 ng/ml). The coefficient of variation ranged between 5.1% and 8% for sulfamethoxazole and its metabolites. Plasma concentrations of fluconazole were measured in samples from four of the six volunteers by use of an established HPLC method with minor modifications, z° The coefficient of variation for the fluconazole assay at the 1 txg/ml lower limit of detection was 8.9%. At higher concentrations (2 to 50 l~g/ml), it ranged from 3.2% to 7.1%. Plasma volume from the two other volunteers was not sufficient for analysis of fluconazole concentration.

Sulfamethoxazole total plasma clearance (CL) was calculated as the ratio of dose to total area under the plasma concentration-time curve (AUC), with extrapolation based on apparent t~/2. Sulfa- methoxazole tl/2 was determined from the 8- to 31-hour time points. Apparent formation clearance (CLf) values of the N-acetylated and N-hydroxylated metabolites were calculated as the product of the fraction of the dose of the respective metabolite

M i t r a et al. 335

o~

found in the urine and total plasma clearance of the drug. This calculation assumes that all the metabolite formed is excreted in urine. The in vivo inhibitory constant (Ki) for fluconazole and fractional inhibition of hydroxylamine formation were calculated by two methods. The first is based on the apparent formation clearance of the hydroxylamine, assuming competitive inhibition (consistent with the fit to the in vitro data) from the relationship21:

CLdCL~.I = 1 + [ I ~ i

in which CL r and CLr. ~ are the formation clearances of the N-hydroxytamine metabolite in the absence and presence of the inhibitor, respectively, and [I] is the time-averaged [AUC(0-24)/24 hours] plasma concentration of fluconazole. Fractional inhibition was defined as follows:

(CLf - C~.,)/CI~

The second method was based on the AUC of the hydroxylamine (AUCsMx-HA) and the mass balance statement:

(AUCsMx.HA)(CLsMx-HA) = fmsMX-HA(DsMx)

in which SMX-HA is sulfamethoxazole hydroxylamine, CL is clearance, fm is fraction metabo- lized, D is dose, and SMX is sulfamethoxazoIe. Based on both this statement of mass balance and the assumption that the elimination clearance of the hydroxylamine is not altered by fluconazole administration:

Fractionalinhibition = [ ( A U C s M x . H A ) c ( C L s M x ) C - -

(AUCsM×-I-IA)Ilu(CLsMx)nu]/(A UCsMx.HA)c(CLsMx)c

in which the subscripts C and flu refer to the control (no fluconazole) and fluconazole cases, respectively. In vivo K i is calculated by substitu- tion of the ratio of (AUCsMx_HA)c(CLsMx)c/ (AUCsMx_nA)fl~(CLsMx)nu for CLf/CLrl as described for the first method. Renal clearance was determined as the ratio of the amount recovered in the urine over the 72 hours of urine collection to the plasma AUC.

All values are expressed as mean values +__ SD. The data were analyzed by the paired Student t test for differences between treatments before and after fluconazole, with a = 0.05.

RESULTS

In vitro studies. Ketoconazole and fluconazole each competitively inhibited sulfamethoxazole hydroxylamine formation, with K i values of 6 and 3.5

0.2-

O

E ¢- 0.125.

0.15 -

0.1

0.05 . . . . ! . . . . ! . . . . !

0 500 1000 1500

0.15 -

O. I "

0.075 -

0.05 . . . . I • ' ' • I . . . . I . . . . I

0 25 50 75 100

ClMETIDINE (gM)

¢-

C

C L I N I C A L P H A I G M A C O L O G Y & T H E R A P E U T I C . ' ;

V O L U M E 5 9 , N U M B E R 3

Fig. 2. Cimetidine inhibition of sulfamethoxazole hydroxylamine (SMX-HA) formation in microsomes from HL- 123. Microsomes were preincubated with cimetidine without reduced nicotinamide adenine dinucleotide phosphate (NADPH) (top panel) or with NADPH (bottom panel), as described in the Material and Methods section.

ixmol/L, respectively (Fig. 1). Inhibition of hydroxylamine formation by cimetidine was mixed under both conditions of preincubation (cimetidine plus microsomes with and without cofactor). Inhibition by cimetidine was therefore quantitated as the concentration required to decrease hydroxylamine formation by 50% (IC5o). I5o was tenfold less when cimetidine was preincubated with microsomes in the presence of cofactor (I5o = 80 ~mol/L) than when cofactor was omitted from the preincubation (15o = 800 ixmol/L; Fig. 2), consistent with previous observations. 22-25

The ratio of cimetidine, ketoconazole, and fluconazole peak concentrations anticipated in plasma

336 M i t r a e t al. CLINICAL PHAI~IACOLOGY & THEI~'~,I'Et.rI'ICS

MARCI-I 1996

Table I. Effect of potential clinical inhibitors on sulfamethoxazole hydroxylamine formation in human liver microsomes

Ki C~,. r h*hibito," (wnol/L) Ki/K,, , (txrnol/L)* Ct,.l/Ki

Cimetidine 80"t 0.20 7.0 0.09 Ketoconazole 6 0.015 0.40 0.07 Fluconazole 3.5 0.009 23.4 6.7

K i values were determined in liver 123. K i, Inhibitory constant; K,,,. Michaelis-Menten constant; ICso. the con-

centration required to decrease hydroxylamine formntion by 50%: NADPH, reduced nicotinamide adenine dinucleotide phosphate.

*Co. f . Free peak plasma concentration estimated from dose and volume of distribution for cimetidine and ketoconazole and from concentrations observed in this study for fluconazole. 33 A peak concentration of 30 ~mol/L for 400 mg fluconazole has been reported, similar to the value we observed. 34

t ls . , Cimetidine preineubated with microsomcs and NADPH.

during use of conventional doses to the Michaelis- Menten constant (Kin) of sulfamethoxazole (0.4 mmol/L), determined previously ~5) is shown in Table I. The results indicate that, on the basis of the ratio of estimated unbound peak plasma concentration at a standard therapeutic dose to Ki, fluconazole appears likely to yield the greatest inhibition in vivo, followed by ketoconazole. Cal- culation of Ki/K m ratios from values obtained from in vitro microsomal incubations of HL-123, a liver that ranks 91% in CYP3A4 and 78% in CYP2C9 ~ among the 29 livers in the liver bank (calculated as a percentage of the liver with the highest content of that form), also indicates the effectiveness of fluconazole (0.009) over ketoconazole (0.015) and cimetidine (0.20).

Ketoconazole and fluconazole were further assessed as inhibitors of sulfamethoxazole hydroxylamine formation in microsomes from three additional livers: HL-119, HL-122, and HL-123. These livers ranked 76%, 100%, and 95%, respectively, for CYP3A4 content among the 29 livers in the liver bank, and 24%, 95%, and 86%, respectively, for CYP2C9 content.* Total concentrations of ketoconazole (15 ~mol/L) and fluconazole (20 l~mol/L) ex- pected to be obtained by usual therapeutic doses also inhibited sulfamethoxazole hydroxylamine formation. In incubations without inhibitor, the velocity of sulfamethoxazole hydroxylamine formation (nanomoles of product/milligram of protein/minute; mean ___ SD) was 0.090 ___ 0.02. In the presence of

*These isozymes have been shown previously to account for the oxida t ion of su l f ame thoxazo le to the hydroxylamine in h u m a n

l iver m i c r o s o m e s ? 5

fluconazole, the velocity was diminished to 0.036 _+ 0.08; in the presence of ketoconazole, the velocity was to 0.04 _ 0.08. On the basis of these observations, fluconazole was chosen for evaluation in subjects because of its negligible plasma protein binding and long tl/2 (free ketoconazole concentration is l % of total).

In vivo studies. The plasma concentration-time profiles of sulfamethoxazole and its acetyl and hydroxylamine metabolites in subject 1 are provided in Fig. 3. Differences in the time course of sulfamethoxazole and its acetyl conjugate concentrations in plasma between treatments were relatively small in this subject. However, there was a 79% decrease in AUC of sulfamethoxazole hydroxylamine in the presence of fluconazole (3.6 versus 0.10 l~g • hr/ml) in this subject. The effect of fluconazole on the pharmacokinetics of a single dose of sulfamethoxazole in all subjects is shown in Table II. The mean AUC of sulfamethoxazole was increased by 25%, and SMX clearance was decreased by 26% (p < 0.05). Sulfamethoxazole tl/2 was lengthened from 9.3 _ 3.3 to 12.9 +__ 2.3 hours (p < 0.005) in the presence of fluconazole. The mean AUC for the sulfamethoxazole hydroxylamine was decreased by 93%. Although the AUC of the acetyl conjugate was unchanged, its tt/2 was more than doubled. Most unexpected was the apparent 6.6-fold increase in renal clearance of the hydroxylamine, whereas the renal clearances of the acetyl conjugate and sulfamethoxazole were unchanged.

Table II also shows that 54% of a dose of sulfamethoxazole was recovered as parent drug, the acetyl conjugate, and the hydroxylamine in the absence of fluconazole and that recovery of these compounds increased to 83% in the presence of fluconazole. Also, the recovery of the hydroxylamine in urine decreased by approximately 60%, whereas that of the acetylconjugate increased by 70%. Table III shows that ftuconazole inhibited the apparent (discussed below) formation clearance of sulfamethoxazole hydroxylamine by 69%, whereas that of acetylsulfamethoxazole was increased by 32%.

Fractional inhibition of hydroxylamine formation and in vivo K i calculated from the apparent formation clearance and hydroxylamine AUC are shown in Table IV for the four subjects in whom fluconazole concentration was measured. The mean predicted fractional inhibition of hydroxylamine for- mationIassuming competitive inhibition, observed time-averaged fluconazole concentrations, and the in vitro Ki--was within 8% of the value observed

CI.IN1CAL f'HA1U','IACOI.OGY & THE1L'\PEUT1CS \;t)LUME 59, NL MBl.:.l't. 3 M i t r a et al. 337

I00

:~ 10

[-, z

z 0.1 O

.<

0.01, .<

0.001

" - . . . . .

,,.~'~o, ° Q"# ""0,.

• " " ! " " " I " ' " I " ' " I " " " I " ' " I " " " I " " "

4 8 12 16 20 24 28 32

~ * - - S M X

........ A ......... S M X - A c

. . . . * . . . . S M X - H A

. . . . O - - - S M X f

. . . . ~ . . . . S M X - A c f

. . . . O . . . . S M X - H A f

T 1 M E , hours

Fig. 3. Effect of fluconazole (f) on the plasma concentration-time profiles of sulfamethoxazole (SMX), N-acetylsulfamethoxazole (SMX-Ac), and sulfamethoxazole hydroxylamine (SMX- HA) in subject 1.

Table II. Effect of f luconazole on exposure to sulfamethoxazole, acetylsulfamethoxazole, and sulfamethoxazole hydroxylamine

A UC CL/F CL,~ Urine recovery Species (lag • hr/ml ) t i/2 ( hr ) (mlhnin ) * ( ml/min ) (% dose)

Without fluconazole SMX 459 +- 185 9.3 -+ 3.3 32.9 -+ 12 3.60 - 2.33 10.1 -+ 4.5 SMX-Ac 201 -+ 96 8.7 _+ 1.3 - - 36.3 - 2.33 42.9 --- 5.1 SMX-HA 4.71 +_ 2.7 10.1 --- 3.8 - - 38.2 --- 20.0 1.26 -+ 0.5

With fluconazole SMX 574 _ 135t 12.9 --+ 2.3~: 24.2 -+ 5.191" 2.38 +-- 1.13 9.6 --- 3.5 SMX-Ac 279 +_ 148 22.6 _-. 3.5§ - - 47.2 -+ 35.0 73.0 +__ 11.3§ SMX-HA 0.35 - 0.3 9.8 --- 5.9 - - 290 -+ 200 0.5 --- 0.25

AUC, Area under the plasma concentration-time curve; tl~, half-life; CI..JF, apparent oral clearance; CL,, remd clearance; SMX, sulfamethoxazole; SMX-Ac, acetylsulfamethoxazole: SMX-HA, sulfamethoxazole hydroxylamine.

"Ratio of clearance to fraction of dose available to the site of sampling. t"p < 0.05, paired t test. ~p < 0.005. §p < 0,001.

when calculated from hydroxylamine A U C but was 30% greater than the value calculated f rom apparent format ion clearance, Similarly, in vivo K~ estimated f rom hydroxylamine A U C was comparab le to that observed in vitro (mean, 1.6 ixmol/L; range, 0.5 to 3.6 txmol/L in vivo versus 3.5 ixmol/L in vitro), whereas that es t imated f rom the apparent format ion clearance had a mean four times the in vitro value, with a range two to eight times the in vitro value.

D I S C U S S I O N We evaluated three potential inhibitors o f cyto-

chrome P450-ca ta lyzed sulfamethoxazole hydroxylamine format ion- - f luconazole , ketoconazole , and c imet id ine- - in vitro in human liver microsomes. All three compounds inhibited microsomal oxidation of sulfamethoxazole, with f luconazole the most potent . Al though ketoconazole is also an effective inhibitor o f sulfamethoxazole hydroxylamine format ion in

3 3 8 M i t r a e t aL CLINICAL PHARMACOLOGY & THERAPEUTICS

MARCH 1996

Table lIl. Effect of fluconazole treatment on apparent sulfamethoxazole metabolite formation clearances

Apparent formation clearance (ml/min)

Acetyl sulfamethoxazole Sulfamethoxazole hydroxylamine

Without With Without With Subject No. fluconazole fluconazole fluconazole fluconazole

1 8.19 9.66 0.308 0.154 2 17.0 23.4 0.357 0.098 3 12.2 16.1 0.245 0.077 5 11.3 14.2 0.483 0.098 6 13.7 22.6 0.413 0.105 7 18.4 20.1 0.413 0.154

Mean -+ SD 13.5 _+ 3.8 17.8 _ 5.32* 0.369 +_ 0.085 0.114 --- 0.032t

*p = 0.017. tp < 0.001.

Table IV. Comparison of predicted (from in vitro) and observed in vivo inhibition of sulfamethoxazole hydroxylamine formation by fluconazole

Fluconazole % Inhibition time-averaged in vivo* In vivo K i (ta~nol/L ) concentration Predicted

Subject No. ( l~,nol/L ) A UC CL I % inhibition A UC CLf

1 29.0 89 50 89 3.57 29.0 2 21.2 98 72 86 0.545 8.02 3 23.5 93 69 87 1.67 10.8 5 19.9 97 75 85 0.614 6.78

Mean --- SD 23.4 __. 4.2 94 +-- 4.1 66.5 - 11 87 -+ 1.7 1.60 -+ 1.4 13.7 --- 10 95% CI - - - - - - 84-90 - - - -

CLf, Formation clearance; CI, confidence interval. *Data in the AUC and CLr columns were calculated from the AUC and CLr of the hydroxylamine, respectively.

vitro (Fig. 3), fluconazole inhibits both CYP3A42~ and CYP2C927 in vivo, isoforms that produce sulfamethoxazole hydroxylamine. 15 Results from the same study also suggested that high plasma protein binding of ketoconazole and low binding of fluconazole were important determinants of the differen- tial inhibition of hepatic drug metabolism of these two drugs in vivo. Cimetidine was a much more potent inhibitor of sulfamethoxazole hydroxylamine formation in vitro when preincubated with NADPH. However, even when preincubated with NADPH and microsomes, cimetidine appeared to be less potent an inhibitor of sulfamethoxazole hydroxylamine formation in vitro than either fluconazole or ketoconazole.

Fluconazole administration to healthy volunteers decreased the AUC of sutfamethoxazole hydroxylamine by 94%. However, the decrease in urinary

recovery of the hydroxylamine was only 60%. These changes were accompanied by a decrease in the clearance of sulfamethoxazole, an increased formation clearance of acetylsulfamethoxazole, an en- hanced recovery of parent and total metabolites in urine, an increased recovery of acetylsulfamethoxazole, and a 6.6-fold increase in the apparent renal clearance of sulfamethoxazole hydroxylamine. The increase in urinary recovery of acetylsulfamethoxazole is consistent with the increase in the formation clearance of acetylsulfamethoxazole and would contribute to the increased recovery of dose as sulfamethoxazole and assayed metabolites in urine. The decrease in sulfamethoxazole clearance is apparently the result of inhibition of oxidation of sulfamethoxazole to a metabolite other than those assayed. The recovery we obtained in the control arm of the study was similar to that reported by others

CLINICAL PHARMACOLOGY & THERAPEUTICS VOLUME 59, NUMBER 3 M i t r a et al. 339

who assayed the same compounds, 28 and no other metabolite has been reported that accounts for a significant fraction of dose. However, if the apparent clearance of sulfamethoxazole by pathways other than those identified in Table II is calculated, it is found to be diminished by approximately 73% in the presence of fluconazole.

The discrepancy between the effect of fluconazole on hydroxylamine AUC and urinary recovery and the very large increase in apparent renal clearance of the hydroxylamine is consistent with the formation of the metabolite in the kidney with elimination directly into the urine. Because this means of formation of hydroxylamine is apparently not inhibited by fluconazole, it is likely to be catalyzed by an enzyme other than cytochrome P450. Sulfa- methoxazole hydroxylamine has previously been shown to be formed by several peroxidases, including prostaglandin synthetase, 29 an enzyme present in the kidney. 3°32 Although this process apparently contributes significantly to the recovery of sulfamethoxazole hydroxylamine in urine (particularly when cytochrome P450 is inhibited), it apparently contributes little to the metabolite in plasma, given the 93% decrease in hydroxylamine AUC.

Because the estimate of hydroxylamine formation clearance includes the contribution of hydroxylamine formation in the kidney, which does not appear to add to the AUC of the metabolite, the comparison between in vivo and in vitro K~ is more appropriately made on AUC than on formation clearance. This comparison assumes that there is no change in renal clearance of the hydroxylamine brought to the kidney in the blood. On the basis of this comparison, the agreement between the extent of inhibition predicted from in vitro studies and in vivo AUC is very good.

In summary, this investigation provides evidence through in vitro and in vivo experiments that fluconazole is an effective clinical inhibitor of sulfamethoxazole N-hydroxylation. Inhibition of N-hydroxylation of sulfamethoxazole by this tria- zole antimycotic may decrease the susceptibility of patients who are positive for the human immunodeficiency virus to the treatment-limiting side effects of sulfamethoxazole. We also provide in- direct evidence that sulfamethoxazole hydroxylamine may be formed in the kidney, presumably by prostaglandin synthetase, and that this site of formation does not contribute significantly to exposure through the plasma.

References

1. Lee BL, Medina I, Benowitz NL, Jacob P, Wofsy CB, Mills J. Dapsone, trimethoprim and sulfamethoxazole plasma levels during treatment of Pneumocystis pneumonia in patients with acquired immunodeficiency syndrome: evidence of drug interactions. Ann Intern Med 1989;110:606-11.

2. Medina I, Mills J, Leoung G, et aI. Oral therapy for pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome: a controlled trial of trimethoprim-sulfamethoxazole versus dapsone- trimethoprim. N Engl J Med 1990;323:776-82.

3. Kovacs JA, Hiementz JW, Macher AM, et al. Pneu- mocystis carinff pneumonia: a comparison between patients with the acquired immunodeficiency syndrome and patients with other immunodeficiencies. Ann Intern Med 1984;100:66-71.

4. Gordon FM, Simon GL, Wofsy CB, Mills J. Adverse reactions to trimethoprim-sulfamethoxazole in patients with the acquired immunodeficiency syndrome. Ann Intern Med 1984;100:495-9.

5. Mandell GL, Sande MA. Antimicrobial agents (cont): sulfonamides, trimethoprim-sulfamethoxazole, quino- Iones, and agents for urinary tract infections. In: Gil- man AG, Rail TW, Nies AS, Taylor P, eds. The pharmacological basis of therapeutics. 8th ed. New York: Pergamon Press, 1990:1047-64.

6. Mandell GL, Sande MA. Antimicrobial agents (cont): drugs used in the chemotherapy of tuberculosis and leprosy. In: Gilman AG, Rail TW, Nies AS, Taylor P, eds. The pharmacological basis of therapeutics. 8th ed. New York: Pergamon Press, 1990:1146-64.

7. Hughes WT, Kennedy W, Dugdale M, et al. Preven- tion of pneumocystis carinii pneumonitis in AIDS patients with weekly dapsone [Letter]. Lancet 1990;336: 1066.

8. Park BK, Coleman JW, Kitteringham NR. Drug disposition and drug hypersensitivity. Biochem Pharma- col 1987;36:581-90.

9. Shear NH, Spielberg SP, Grant DM, Tang BK, Kalow W. Differences in metabolism of sulfonamides predis- posing to idiosyncratic toxicity. Ann Intern Med 1986; 105:179-84.

10. Rieder M J, Uetrecht J, Cannon M, Miller M, Spiel- berg SP. Diagnosis of sulfonamide hypersensitivity reactions by in vitro "rechallenge" with hydroxylamine metabolites. Ann Intern Med 1989;110:286-9.

11. Kiese M. The biochemical production of ferrihemoglobin-forming derivatives from aromatic amines and mechanisms of ferrihemoglobin formation. Pharma- col Rev 1966;18:1091-161.

12. Cribb AlE, Spielberg SP. Hepatic microsomal metabolism of sulfamethoxazole to the hydroxylamine. Drug Metab Dispos 1990;18:784-7.

13. Cribb AE, Spielberg SP. An in vitro investigation of

340 M i t r a et al. CI.INICAI~ PHARMACOI.OGY & TH EI~'~d~EUTICS

MARCH 1996

predisposition to sulfonamide idiosyncratic reactions in dogs. Vet Res Commun 1990;14:241-52.

14. Cribb AE, Spielberg SP. Sulfamethoxazole is metab- olized to the hydroxylamine in humans. CuN PHARMA- COL THER 1992;51:552-6.

15. Mitra AK, Kalhorn TF, Thummel KE, Unadkat JD, Slattery JT. Metabolism of arylamines by human liver microsomal cytochrome P-450(s) [Abstract]. XSSX 1993;4:148.

16. Thummel KE, Kharasch ED, Podoll T, Kunze K. Human liver microsomal enflurane deftuorination catalyzed by cytochrome P-450 2El. Drug Metab Dis- pos 1993;21:350-7.

17. Thummel KE, Lee CA, Kunze KL, Nelson SD, Slat- tery JT, Oxidation of acetaminophen to N-acetyl-p- benzoquinone imine by human CYP3A4. Biochem Pharmacol 1993;45:1563-9.

t8. Lowry OH, Rosenbrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem 1951;193:265-75.

19. Cribb AE, Miller M, Leeder JS, Hill J, Spielberg SP. Reactions of the nitroso and hydroxylamine metabolites of sulfamethoxazole with reduced glutathione. Implications for idiosyncratic toxicity. Drug Metab Dispos 1991;19:900-6.

20. Hosotsubo KK, Hosotsubo H, Nishijima MK, Okada T, Taenaka N, Yoshiya I. Rapid determination of serum levels of a new antifungal agent, fluconazole, by high performance liquid chromatography. J Chro- matogr 1990;529:223-8.

21. Shaw PN, Houston JB. Kinetics of drug metabolism inhibition: Use of metabolic concentration-time profiles. J Pharmacokinet Biopharm 1987;15:497-510.

22. Shear NH, Spielberg SP, Grant DM, Tang BK, Kalow W. Differences in metabolism of sulfonamides predis- posing to idiosyncratic toxicity. Ann Intern Med 1986; 105:179-84.

23. Somogyi A, Gugler R. Drug interactions with cimetidine. Clin Pharmacokinet 1982;7:23-41.

24. Jensen JC, Gugler R. Cimetidine interaction with

liver microsomes in vitro and in vivo: involvement of an activated complex with cytochrome P-450. Bio- chem Pharmacol 1985;34:2141-6.

25. Tingle MD, Coleman MD, Park BK. The effcct of preincubation with cimetidine on the N-hydroxylation of dapsone by human liver microsomes. Br J Clin Pharmacol 1991;32:120-3.

26. Morita K, Konishi H, Shimakawa H. Fluconazole: a potent inhibitor of cytochrome P-450 dependent drug metabolism in mice and humans in vivo. Comparative study with ketoconazole. Chem Pharm Bull (Tokyo) 1992;40:1247-51.

27. Blum RA, Wilton JH, Hilligoss DM, et al. Effect of fluconazole on the disposition of phenytoin. CuN PHARMACOL THER 1991;49:420-5.

28. Vree TB, Hekster YA, Baars AM, Damsma JE, van der Kleijn E. Pharmacokinetics of sulfamethoxazole in man. Clin Pharmacokinet 1978;3:319-29.

29. Cribb AE, Miller M, Tesoro A, Spielberg SP. Peroxidase-dependent oxidation of sulfonamides by monocytes and neutrophils from humans and dogs. Mol Pharmacol 1990;38:744-51.

30. Larsson R, Ross D, Berlin T, Olsson LI, Molddus P. Prostaglandin synthase catalyzed metabolic activation of p-phenetidine and acetaminophen by microsomes isolated from rabbit and human kidney. J Pharmacol Exp Ther 1985;235:475-80.

31. Tai HH, Tai CL, Hollander CS. Biosynthesis of pros- taglandins in rabbit kidney medulla. Biochem J 1976; 154:257-64.

32. Cavallo T. Fine structural localization of endogenous peroxidase activity in inner medullary interstitial cells of the rat kidney. Lab Invest 1974;31:458-64.

33. Benet LZ, Williams RL. Design and optimization of dosage regimens: pharmacokinetic data. In: Gilman AG, Rail TW, Nies AS, Taylor P, eds. The pharmacological basis of therapeutics. 8th ed. New York: Pergamon Press, 1990:1650-1735.

34. Debruyne D, Ryckelynck JP. Clinical pharmacokinetics of fluconazole. Clin Pharmacokinet 1993;24:10-27.

Documents

Inhibition of sulfamethoxazole hydroxylamine formation by fluconazole in human liver microsomes and healthy volunteers*