Experimental Procedures

Materials.

The budesonide epimers, 22R and 22S, were provided by AstraZeneca Pharmaceuticals LP (Södertälje, Sweden) and were 97.5 and 96.5% pure, respectively. [³⁵S]3′-phosphoadenosine 5′-phosphosulfate (PAPS) (2 Ci/mmol) was purchased from PerkinElmer Life Sciences (Boston, MA). LK6DF silica gel 60A thin-layer chromatography (TLC) plates were purchased from Whatman Inc. (Clifton, NJ).

Methods.

Sulfotransferase assays

Budesonide sulfation activity was assayed using both nonradioactively labeled 22R and 22S budesonide. [³⁵S]PAPS was included in the reactions, and [³⁵S]budesonide sulfate was resolved by TLC as described previously (Falany et al., 1994). Initial reactions involved screening 22R and 22S for activity with each SULT at 25 μM and 150 μM budesonide. Positive controls using known substrates at optimal concentrations for each SULT isoform analyzed were included in the reactions. Typical reactions contained 1 to 20 μM budesonide in 50 mM Tris-HCl, pH 7.4, and 20 μM [³⁵S]PAPS in a final volume of 62.5 μl and were linear with respect to time and protein. Control reactions were identical except that budesonide was not added. The substrates were resolved using an 85:15:5 methylene chloride, methanol, ammonium hydroxide solvent system. After TLC, autoradiography was used to localize sulfated products. The conversion to radioactive product was determined by scintillation counting after scraping the specific radioactive budesonide product bands from the TLC plate. Control reactions contained all products of the reaction mixture except for substrate to detect background sulfation. The bacterial expression, purification, and isolation of the expressed human SULTs, monoamine-sulfating form of phenol SULT (MPST, SULT1A3, accession number P50224), phenol-sulfating form of phenol SULT (PST-1, SULT1A1,AAA99892), dehydroepiandrosterone SULT (DHEA-ST, SULT2A1, Q06520), SULT1B2 (BAA24547), estrogen SULT (EST, SULT1E1, AAB34601), SULT2B1a (AAC78553), and SULT2B1b (AAC78554) has been reported previously (Falany et al., 1989; Wilborn et al., 1993; Falany et al., 1995;Ganguly et al., 1995; Her et al., 1998; Wang et al., 1998).

Product analysis by mass spectrometry.

Sulfated budesonide products were identified by negative ion electrospray mass spectrometry on a PE-Sciex API III triple quadrupole mass spectrometer (PerkinElmerSciex, Boston, MA). The sulfated budesonide bands were scraped from the TLC plate, extracted with deionized water, lyophilized, dissolved in methanol, and then injected into a 25 μl/min flow of 50:50 acetonitrile/water containing 0.1% formic acid. A Harvard Apparatus model 22-syringe pump (Harvard Apparatus Inc., Holliston, MA) was used to deliver the flow to the electrospray interface. Tandem or MS-MS spectra were obtained by selecting the parent ions with the first quadrupole. The selected parent ions enter the second quadrupole where, upon collision with argon gas, fragment ions are produced. The fragment ions are scanned by the third quadrupole to give the MS-MS spectra.

Molecular Modeling.

Construction of the sulfate products of budesonide

SYBYL version 6.4 molecular-modeling package (Tripos Inc., St. Louis, MO) was used for all the modeling procedures. For the calculation of the minimum energies of the two epimers and four singly sulfated products of budesonide, each molecule was assigned Gasteiger-Hückel charges and minimized by the Powel algorithm using the Tripos force field with a gradient termination of 0.01 kcal/mol as the convergence criteria.

Modification of DHEA-ST structure.

The structure of DHEA-ST, [protein database bank (PDB) file 1EFH] was minimized after including PAPS, DHEA, and hydrogen atoms. For the addition of PAPS, the structure of a proposed mimic of the transition state intermediate of the SULT reaction, PAP-vanadate, was extracted from (PDB file 1BO6) (Kakuta et al., 1998) that contained the coordinates for the estrogen sulfotransferase-PAP-vanadate complex. Replacing the vanadate with a sulfuryl group modified the PAP-vanadate moiety to result in a PAPS molecule. The PAPS molecule was then fit into the position occupied by PAP in the active site of the DHEA-ST model using the Multifit option in SYBYL 6.4. For the addition of DHEA, the estradiol molecule was extracted from the (PDB file 1AQU) (Kakuta et al., 1997) that contained the coordinates for the estrogen sulfotransferase-PAP-estradiol complex. The structure of estradiol was modified to result in a DHEA molecule. The DHEA molecule was fit into the substrate-binding pocket of the DHEA-ST structure based on structural analogy with estradiol and estrogen sulfotransferase. The estrogen sulfotransferase-PAP-estradiol complex was used because it contains a substrate that was cocrystallized with the enzyme, unlike the crystal structure for DHEA-ST (Kakuta et al., 1998; Pedersen et al., 2000). Hydrogen atoms and Kollman-all charges were added to the protein whereas the atoms of DHEA and PAPS were assigned Gasteiger-Hückel charges. DHEA and PAPS were held as fixed aggregates during the protein minimization by the Powel algorithm using the Tripos force field with a gradient termination of 500 iterations as the convergence criteria.

Docking Procedure.

The automated docking package FlexiDock, a module of SYBYL 6.4, was used in the docking studies. Individual budesonide epimers were constructed, given Gasteiger-Hückel charges, and optimized by energy minimization using the Powel algorithm. These structures were fit into the position occupied by DHEA in the DHEA-ST-PAPS complex using the Multifit option in SYBYL 6.4. The resultant structure was used as the initial conformation for the docking procedure. This methodology provided a set of solutions that represented the optimum ligand geometry in the active site of the DHEA-ST model.

Results

Sulfation Activity.

Budesonide sulfation was initially analyzed using expressed isoforms of the major members of both the SULT1 and SULT2 families. The SULT1 family was represented by PPST-1, MPST, EST, and SULT1B2. The SULT2 family was represented by DHEA-ST, SULT2B1a, and SULT2B1b. Budesonide sulfation activity was assayed under varying conditions, and budesonide sulfation was only observed with DHEA-ST. This was the only SULT that demonstrated budesonide activity that was greater than background. Figure 2 is a representative figure of sulfation activity shown by only DHEA-ST. In preliminary studies, the epimeric forms of budesonide displayed differential activity, which is shown in Table 1 by the rate of budesonide sulfation activity of the 22R epimer being approximately 3.5-fold greater than that with the 22Sepimer. Using expressed purified DHEA-ST and 20 μM PAPS, the apparentK_m values for the 22R and 22S epimers were 20.9 ± 3.9 μM and 11.1 ± 3.0 μM, respectively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Representative TLC autoradiograph showing sulfation of 25 μM 22R-budesonide with seven human SULTs.

Lane 1, 22R-budesonide with DHEA-ST; lane 2, no substrate with DHEA-ST, lane 3, 22R-budesonide with SULT2B1a; lane 4, 22R-budesonide with SULT2B1b; lane 5, 22R-budesonide with PPST-1; lane 6, 22R-budesonide with MPST; lane 7, 22R-budesonide with EST; lane 8, 22R-budesonide with SULT1B2.

Mass Spectrometry.

MS-MS analysis of the products of the budesonide sulfation reaction detected only the formation of monosulfated budesonide, which indicates that both hydroxyl groups on budesonide were not sulfated at the same time. The scan showed peaks corresponding to the nonsulfated parent compound and the negative ion peak for budesonide monosulfate. Since budesonide possesses hydroxyl groups at the 11 and 21 positions, the site of sulfation was investigated. Further analysis of budesonide sulfate by MS-MS was attempted; however, fragmentation of the steroid nucleus was not sufficient for identification of specific sulfated fragments.

Modeling.

Since the site of budesonide sulfation could not be determined by MS, a modeling approach was undertaken to define which hydroxyl group on budesonide is closest to the active site when the epimers bind to the SULT enzyme. The minimum energies of the four possible singly sulfated products of the two epimers of budesonide are shown in Table2. For both the 22S and 22R epimers, sulfate conjugation at position 21 generates more stable sulfated products than conjugation at position 11.

The availability of the coordinates for the structure of DHEA-ST (Pedersen et al., 2000) allows for the analysis of the probable interaction of the budesonide epimers with DHEA-ST in the presence of PAPS. Using FlexiDock methodology, only position 21 of budesonide showed binding in a suitable geometry for sulfate transfer. In none of the binding models did position 11 associate closely with the active site histidine (Pedersen et al., 2000). Comparison of the binding of the budesonide epimers indicated that the 22S epimer (Fig.3A) could be docked in the active site with better geometry than the 22R epimer (Fig. 3B). The results are consistent with the lowerK_m for the 22S isomer, which might indicate higher affinity for the 22S isomer (Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

22(S) (Fig. 3A) and 22(R) (Fig. 3B) budesonide docked into the coordinates of the active site of human DHEA-ST.

This enzyme structure was modeled with analogy to the coordinates for the mouse estrogen sulfotransferase (PDB code 1AQU). The labeled areas of the figure include the sulfate group of PAPS, the 11-hydroxyl, and the 21-hydroxyl groups of the budesonide epimers.

Discussion

In this report, DHEA-ST has been shown to be the SULT isoform primarily responsible for the sulfation of budesonide in human tissues. DHEA-ST is capable of conjugating both the 3α- and 3β-hydroxyls of hydroxysteroids, the 3-phenolic hydroxyl of estrogens, as well as the 17β-hydroxyl of testosterone and β-estradiol (Falany et al., 1989). This observation is consistent with the report of Pacifici et al. (1994) that human liver budesonide sulfation was inhibited by testosterone and correlated with testosterone sulfotransferase activity. To date, DHEA-ST is the only human SULT reported to sulfate testosterone.

Budesonide is administered by inhalation for the treatment of asthma and is also used in nasal preparations for the treatment of rhinitis and in oral preparations for the treatment of inflammatory bowel disease (e.g., Crohn's disease; Spencer and McTavish, 1995). DHEA-ST is highly expressed in both the human adrenal cortex and liver and is present in the gastrointestinal tract indicating a probable role in the first-pass and systemic metabolism of budesonide (Her et al., 1996;Tashiro et al., 2000). DHEA-ST expression in human lung is less well characterized. Luu-The et al. (1995) did not detect DHEA-ST message in adult lung RNA, whereas Hume et al. (1996) report the immunolocalization of DHEA-ST in lung tissue of young children. Therefore, the role of DHEA-ST in the metabolism of budesonide in different tissues still needs to be investigated.

Budesonide is administered as an equal mixture of two epimers, 22R and 22S (Ryrfeldt et al., 1984). The 22R epimer has a 2- to 3-fold more potent anti-inflammatory response and has a different pharmacokinetic profile (Brattsand et al., 1982). The greater potency of the 22R epimer is possibly due to a greater affinity for the glucocorticoid receptor (Derendorf et al., 1998). As shown in Table 1, DHEA-ST displays a 2-fold lower apparent K_m for the 22Sepimer as compared with the 22R; however, the sulfation activity of the 22R epimer is 3.5-fold greater than the 22S epimer.

Attempts to identify the site of sulfate conjugation using MS-MS were unsuccessful. Sufficient fragmentation of the steroid nucleus was not obtained to allow for definitive identification of daughter fragments. The availability of the structural coordinates for EST and DHEA-ST (Kakuta et al., 1998; Pedersen et al., 2000) allowed for a molecular-modeling approach to predict the geometry of budesonide binding in the active site. Consistent with the activity data, the 11-hydroxyl position did not bind in the active site close to the active site histidine (Pedersen et al., 2000). Also, the presence of the two angular methyl groups on the steroid backbone should provide considerable steric hindrance to sulfation of the 11-hydroxyl group. In accordance with these observations, it has been shown that budesonide and hydrocortisone can be selectively acetylated chemically at the 21-hydroxyl group without affecting the more sterically hindered 11β-hydroxyl group (Paulson and Lindberg, 1991; Lindberg et al., 1992). Moreover, budesonide is known to be conjugated with fatty acids in lung and liver apparently at the 21-hydroxyl group (Tunek et al., 1997).

Studies of the metabolism of budesonide have not reported the formation of large amounts of sulfate conjugate; however, phase II metabolism has not been the focus of these reports (Andersson et al., 1982; Edsbacker et al., 1983, 1987; Jonsson et al., 1995). Previous reports of the in vitro metabolism of epimeric budesonide by human liver microsomes have identified phase I metabolites resulting from the action of the oxidative cytochrome P450 enzyme family and principally, CYP3A (Jonsson et al., 1995). The two major metabolites of epimeric budesonide from in vitro studies using human liver homogenates are 6β-hydroxybudesonide and 16α-hydroxyprednisolone, although 16α-hydroxyprednisolone was not a detectable metabolite of the 22S epimer (Jonsson et al., 1995). Although sulfation might not be a major pathway in budesonide metabolism and inactivation following inhalation, it may have a significant role in budesonide metabolism following oral administration, especially if time-release formulations are used to increase delivery to the colon (Spencer and McTavish, 1995). Absorption from the gastrointestinal tract can also occur from the fraction of the given dose that is deposited in the mouth and swallowed after inhalation (Newman et al., 1981; Davies, 1982).

In conclusion, this study has shown the ability of DHEA-ST, a major human hepatic hydroxysteroid SULT, to catalyze the sulfation of both epimers of budesonide. Sulfation apparently occurs only at the 21-hydroxyl moiety and shows differences in rate and affinity for the two epimers. Although, the CYP3A enzymes have the most important role in the budesonide deactivation, the significance of sulfation in the deactivation of budesonide may become more important in situations in which the CYP3A isoforms are inhibited by concomitant medication or by dietary factors (Ameer and Weintraub, 1997).