![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pharmacokinetics and Drug Metabolism, Amgen Inc., Thousand Oaks, California
(Received October 17, 2007; accepted November 12, 2007)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionConclusionReferences |
|---|
) to describe these systems. Since then, many examples have been published where milligram amounts of mammalian metabolites, sufficient for biological and other evaluations, have been produced by microbial fermentation methods [for recent examples, see Zmijewski et al. (2006) and Zhang et al. (2006)].
Microbial fermentation methods typically utilize a two-stage fermentation protocol (Goodhue CT, 1982). A common approach (Hilton MD, 1999) is to add substrate to a stage II culture in a sterile shake flask, incubate on an orbital shaker for a suitable period of time, quench, centrifuge or filter, and then extract supernatants to obtain the biotransformed products. This method has been successfully used to screen for microbes that make the metabolite of interest as well as to perform small-scale (<1 liter) biosynthesis with the identified microbe. One shortcoming of this configuration, however, is the destruction of the microbial biocatalyst when the reaction is quenched, which limits each preparation to a single reaction.
A hollow fiber cartridge (HFC) reactor (described in Materials and Methods) provides a configuration that isolates a relatively small volume of microbial culture from a larger reservoir of culture medium via a semipermeable membrane (Knazek et al., 1972). The membrane permits exchange of nutrients, substrate, and metabolites and provides two potential advantages of an HFC reactor over a shake-flask system. The excretory products that over time might otherwise inhibit the viability of the microorganism diffuse away from the culture and are diluted by the larger volume of medium in the reservoir. In addition, because the metabolites also diffuse through the capillaries into the circulating medium, they can be harvested without disturbing the biocatalyst. The biocatalyst may then be potentially reused many times over.
Diclofenac sodium is a nonsteroidal anti-inflammatory drug widely prescribed as an anti-inflammatory and antipyretic analgesic. Its principal phase-I metabolic products formed in vivo and in vitro in rats, dogs, baboons, and humans are 4'-hydroxy and 5-hydroxy metabolites (Stierlin et al., 1979; Leemann et al., 1993; Shen et al., 1999). Minor in vivo oxidative metabolites include 3'-OH and 4',5-dihydroxy diclofenac (Stierlin et al., 1979).
In this study, we report 1) the efficient turnover of diclofenac to its hydroxylated metabolites by the bacterium Actinoplanes sp. in a shake-flask incubation; 2) use of a shake-flask system to scale-up production of the 4'-hydroxy, 5-hydroxy, and 4',5-dihydroxy metabolites of diclofenac; 3) the efficient use of HFC reactors to produce turnover rates comparable to a shake-flask configuration; and 4) demonstration that an HFC reactor can be reused for many (n =>5 over 10 days) cycles.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
LC-MS/MS and NMR. The progress of biotransformations and purities of isolated metabolite were determined by LC-MS/MS analysis. Reverse-phase high performance liquid chromatography separations were carried out on an Agilent 1100 system (Agilent Technologies, Delaware, DE) consisting of a temperature-controlled autosampler, a binary pump, and a photodiode array detector in-line with a LCQ-DecaXPplus (Thermo Electron, Waltham, MA) ion-trap mass spectrometer. The mobile phase consisted of 10 mM ammonium formate in water at pH 5.0 (solvent A) and 10 mM ammonium formate in 95% MeOH/5% water (solvent B) with the following linear gradient system: flow rate of 0.3 ml/min; 0 to 2.5 min, 60% A; 2.5 to 9.5 min, 60 to 5% A; 9.5 to 11.5 min, 5% A; 11.5 to 12 min, 60% A. Separation was achieved on a C18 (symmetry, 2.1 x 100 mm, 5 µm; Waters, Milford, MA) column maintained at 40°C. An electrospray ionization source operated in negative ion mode was used to acquire LC-MS/MS data. Diclofenac turnover (measured by its depletion) and formation of its metabolites were approximated by measuring peak areas in extracted ion chromatograms for the pseudomolecular ions of each component (m/z 294: diclofenac; m/z 310: 4'-hydroxy/5-hydroxy; m/z 326: 4',5-dihydroxy metabolites).
NMR spectra were acquired on a 600-MHz spectrometer equipped with a 5-mm cryoprobe (Bruker Instruments, Billerica, MA). Isolated metabolites were dissolved in 160 µl of methanol-d4 and transferred to 3-mm tubes. Complete structural assignment was carried out by analyzing the 1D 1H, 2D 1H/1H total correlation spectroscopy, 2D 1H/13C heteronuclear single quantum coherence, and heteronuclear multibond coherence data sets.
Microorganism Screening. Microbial cultures (shipped lyophilized or frozen) were reconstituted according to ATCC instructions and preserved in either nutrient or potato dextrose agar slants at 4 to 6°C. The following bacteria were screened for diclofenac (50 µM) turnover: Rhodococcus erythropolis (ATCC 4277), Streptomyces griseus (ATCC 13273), Nocardia corallina (ATCC 31338), Pseudonocardia autotrophica (ATCC 55293), Saccharopolyspora hirsuta (ATCC 20501), Pseudomonas sp. (ATCC 17483), and Actinoplanes sp. (ATCC 53771). For each bacteria, spores from slant/frozen stock were used to inoculate 25 ml of SOC medium in a 125-ml Erlenmeyer flask, and this stage I culture was incubated at 30°C for 3 days on an orbital shaker (250 rpm, MaxQ 4000; Barnstead, Dubuque, IA). A stage II culture was initiated by adding approximately 2 ml of stage I culture into 25 ml of fresh SOC medium in a 125-ml Erlenmeyer flask. The stage II culture was incubated at 30°C for 24 h, followed by addition of a stock solution of substrate. At selected time points, 1-ml aliquots were placed in polypropylene tubes with 0.5 ml of acetonitrile, vortex mixed, and centrifuged at 16,000g for 10 min, and supernatants were subjected to LC-MS/MS analysis.
The following fungi were also screened: Aspergillus ochraceus (ATCC 1008), Cunninghamella elegans (ATCC 9245), C. elegans PA-1 (ATCC 36112), Beauveria bassiana (ATCC 7159), Absidia blakesleeana (ATCC 10148b), Cunninghamella echinulata (ATCC 9244), C. echinulata (ATCC 11585b), and Rhizopus oryzae (ATCC 11145). The methods used for the bacterial incubations were also used for the fungal cultures except that all incubations were carried out in terrific broth complete medium at ambient temperature (
22°C).
Concentration Tolerance and Conversion Time Profiles. Experiments were performed to determine the tolerance of stage II cultures of Actinoplanes sp. to various concentrations of diclofenac and to determine the progression of metabolism. The culture methods were identical to those used for screening of bacterial metabolism. Aliquots of substrate stock solution were added to stage II cultures to achieve final concentrations of 50, 100, and 500 µM using a 50-mM stock and 1, 2, 5, and 10 mM using a 500-mM stock. One-milliliter aliquots were sampled at indicated time points and analyzed by LC-MS/MS as described previously.
The effect of using a diluted Actinoplanes sp. culture for the conversion of diclofenac was also studied at ambient (22°C) and incubation (30°C) temperatures. Fifteen milliliters of culture medium obtained 24 h after inoculation of the stage II culture was mixed with 135 ml of fresh SOC in 500-ml Erlenmeyer flasks. Appropriate amounts of substrate stock (400 mM) were added to achieve target final concentrations of 50 µM, 500 µM, 1 mM, and 2 mM. The time profile of the reaction was monitored via LC-MS/MS as described above.
Preparative-Scale Incubation and Chromatography. The scale-up procedure was similar to the microbial screening. Metabolites were produced in a 200-ml stage II Actinoplanes sp. culture (2 x 100 ml in 500-ml Erlenmeyer flasks, 200 rpm). Diclofenac sodium (1.25 ml of 400-mM stock added per flask; 5 mM final concentration) was added 24 h after stage II inoculation and incubated for 20 h. At the end of the incubation, the fermentation broth was adjusted to pH 5 with 3 N HCl and then extracted with ethyl acetate (8 x 50 ml). The pooled organic phases were dried over anhydrous MgSO4, and solvent was removed on a rotavap to yield a light-yellow oil. The crude oil was purified via flash column chromatography (CombiFlash system; Teledyne Isco, Lincoln, NE) on a 40 g RediSep silica cartridge with an isocratic mobile phase (hexanes/ethyl acetate, 6:4) to yield three metabolites. The isolated metabolites were characterized by LC-MS/MS and NMR.
Hollow Fiber Cartridge Reactor Procedures. The composition of an HFC reactor and its operation in a closed-cycle loop is depicted in Fig. 1. An HFC consists of a bundle of capillaries composed of a semipermeable membrane (molecular cutoff varies with the membrane material) and an extracapillary space (ECS) all enclosed in a polycarbonate cylinder. The semipermeable membrane permits exchange of O2, nutrients, substrate, metabolites, and cellular excretory products (with molecular weights below the cutoff) while retaining the microorganism in the ECS. The capillary bundle is potted on both ends to polyurethane adapters resulting in a common flow path when connected to an external reservoir. The ECS, between the capillary membranes and inner walls of the cartridge, is loaded with a microbe culture. A medium (containing O2, nutrients, and the substrate) is then continuously circulated at a constant flow rate using a peristaltic pump.
|
The ECS in each HFC was then loaded with the microbial culture, and time profiles for diclofenac turnover were determined. Two sterile 50-ml syringes, one empty and another loaded with 40 ml of stage II Actinoplanes sp. culture, were attached to the two ECS ports. The microbial suspension was flushed slowly through the ECS space to the other syringe, back and forth, for 2 min. The cartridge ports were wiped with 70% ethanol and closed. The external reservoir volumes remained the same as described above. The system was equilibrated by starting the flow of medium at 10 ml/min for 5 min, and then diclofenac stock solution (50-µM final concentration in the cellulose and Teflon HFC reactors; 2 mM in the polysulfone HFC reactor) was added to the external reservoir. Aliquots (1 ml) were sampled at different times and analyzed by LC-MS/MS. The same experiment was also conducted with cellulose membrane HFC reactor at 2 mM diclofenac concentration.
Experiments to evaluate the reusability of the HFC reactors were performed at two substrate concentrations (50 and 500 µM) in cellulose membrane cartridges. The microbial suspensions were loaded and the reactors were equilibrated as described above. Nineteen microliters of 400 mM diclofenac stock (50 µM final concentration) was then spiked into the reservoir. The reactions proceeded for 15 h at ambient temperature (22°C), after which the contents of the external reservoir were removed and analyzed by LC-MS/MS. The reactors were washed by circulating 100 ml of fresh SOC broth for 1 h, and then 135 ml of fresh SOC broth was placed in each reservoir. Medium was recirculated for 5 min, and then aliquots were collected and analyzed to ensure minimal carryover from the preceding reactions. The medium was then amended with diclofenac stock, and formation of the reaction products was monitored. This cycle was repeated a total of 7 times. The experimental procedure to evaluate the reusability at 500-µM concentrations of diclofenac was slightly modified. The external reservoir containing SOC medium was placed on a hot-plate stirrer with its temperature set to 34°C so that the temperature of the medium (measured at the reservoir with the pump set at a flow rate of 10 ml/min) was 30°C; 188 µl of 400 mM diclofenac stock was added and the reaction proceeded for 36 h. The reaction was repeated four more times for a total of five cycles with two wash cycles between each reaction.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
|
Shake-Flask Concentration Tolerance and Time Profiles. Time profiles of diclofenac turnover when incubated at various concentrations (50 µM to 5 mM) with Actinoplanes sp. at 30°C in a shake-flask are shown in Fig. 2. At concentrations of up to 1 mM, substrate turnover was rapid and complete within 5 h. The turnover at 2 and 5 mM was complete in 10 and 24 h, respectively. LC-MS/MS analysis indicated the reaction products predominantly consisted of 4'-OH (Rt 10.6 min) and 5-OH (Rt 10.8 min) diclofenac in ratios ranging from 8:2 to 7:3 for all of the tested concentrations. At later time points, 4',5-dihydroxy diclofenac (Rt 8.2 min) was also formed. Turnover at a substrate concentration of 10 mM was dramatically diminished (0.5% over 48 h, data not shown), and substantial cell death was observed.
|
1:0.65:0.1. Collision-induced fragmentation of the m/z 310 ion resulted in a major ion at m/z 266 (–44, loss of CO2) and minor ions at m/z 230 (loss of CO2 and HCl) and 194 (loss of CO2 and 2HCl). MS3 analysis of the m/z 266 ion resulted in secondary ions at m/z 230 and 194. The proton NMR spectra of the 4'-OH and 5-OH diclofenac metabolites matched well with previously published data (Stierlin et al., 1979; Shen et al., 1999; Tang et al., 1999). A complete list of identifiable proton and carbon chemical shifts is available in Table 2.
The bis-mono-oxygenated metabolite 4',5-dihydroxy diclofenac (25.3 mg; 7.7% yield) was isolated as a beige solid, and it formed [M-H]– ions at m/z 326, 328, and 330. MS/MS analysis of the m/z 326 ion resulted in a major fragment at m/z 282 (–44, loss of CO2) and a minor fragment at m/z 246 (loss of CO2 and HCl). MS3 analysis of the m/z 282 ion resulted in peaks at m/z 246 and 210 (successive loss of two HCl). The NMR data, listed in Table 2, confirmed that bishydroxylation had occurred at the 4' and 5 positions.
Hollow Fiber Cartridge Reactions. The nonspecific binding of diclofenac to HFC reactors with three different semipermeable membrane materials was evaluated in the absence of microorganisms in ECS. Nonspecific binding to a polysulfone membrane HFC reactor was rapid and high with only approximately 18% of nominal concentrations of diclofenac unabsorbed at 1 h and available for biotransformation. The nonspecific binding to Teflon and cellulose membrane HFC reactors was substantially lower: 98.5% and >85% diclofenac remaining in both at 1 and 5 h, respectively.
Diclofenac was turned over in all three HFC reactors at ambient temperature to its 4'-OH and 5-OH metabolites, and the results are summarized in Table 3. The time for complete substrate depletion, however, differed widely. Actinoplanes sp. in a cellulose membrane HFC turned over 50 µM diclofenac rapidly: 95% conversion in 15 h and 100% in 24 h. Diclofenac was metabolized at the slowest rate in the Teflon HFC reactor, probably because its ECS can hold the smallest volume (2 ml) of microorganism culture. In the polysulfone HFC, 95% of substrate (at 2 mM nominal concentration) had disappeared in 96 h. However, a proportionally lower amount of metabolites formed. This is a result of the combination of irreversible binding to the polysulfone membrane and metabolic consumption. Diclofenac at a 2-mM substrate concentration in a cellulose HFC reactor was not turned over probably because of lethality to cells as observed in the shake-flask experiment.
|
Incubations were conducted in shake flasks using the amounts of stage II culture that were used in the cellulose-membrane HFC reactors. The results are presented in Fig. 3A. The amounts of stage II culture were matched in a way to provide a more direct comparison of the two techniques. The cellulose HFC reactor holds 15 ml of stage II bacterial culture in a total reaction volume of 150 ml, which is an approximately 10-fold lower microbial density than was used in the shake-flask cultures used to determine the time course of turnover shown in Fig. 2. At ambient temperature, 50 µM diclofenac was completely consumed in 24 h and 500 µM diclofenac required 72 h. Figure 3B shows results from the same incubations conducted under identical conditions, except temperature was 30°C. The reaction rate was approximately 3- to 4-fold higher at higher temperature. At 1-mM or higher concentrations, the substrate turnover was dramatically lower, with <1% turnover observed at 72 h. The 4'-OH and 5-OH metabolites were observed approximately in the same ratio as in earlier incubations.
|
Time profiles of 50 µM diclofenac conversion in a shake-flask and a cellulose-membrane HFC reactor under identical conditions (same amount of stage II culture, 22°C) are shown in Fig. 4. The progression of turnover seems to be very similar for both techniques.
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
) or alkyl carbon (Chen et al., 1993; So et al., 1995; Haag et al., 1998; Hsu et al., 2002); no aromatic hydroxylations have been previously reported.
A 200-ml scale-up of a shake-flask incubation and isolation of the reaction products yielded three metabolites: 4'-hydroxy, 5-hydroxy, and 4',5-dihydroxy diclofenac. One hundred sixty-nine milligrams of 4'-hydroxy diclofenac was obtained, and a 54% yield was significantly higher than previously reported bioreactor scale-ups. Previously obtained yields were 15% from Epicoccum nigrum (Webster et al., 1998), 26% from human P450 2C9 expressed in Sf21 insect cells (Rushmore et al., 2000), and 35% from human P450 2C9 expressed in an Escherichia coli system (Vail et al., 2005). 5-Hydroxy diclofenac has been previously reported to be formed in fungal incubations (Webster et al., 1998) but was never isolated. It has been isolated from human urine (Stierlin et al., 1979) as well as synthesized (Bort et al., 1996; Tang et al., 1999; Kenney et al., 2004). 4',5-Dihydroxy diclofenac has also been isolated from human urine (Stierlin et al., 1979); however, a microbial or synthetic route has not been previously reported.
The tolerance of Actinoplanes sp. to high concentrations of diclofenac (up to 5 mM; Fig. 2) in a shake flask nominally seemed to be higher than in the HFC reactor (only up to 500 µM; Fig. 3). The microbial content was, however, approximately 10 times higher in the shake flask than in the HFC reactor. When the lower microbial content in the HFC reactor was compensated, the rate of 50 µM diclofenac conversion was observed to be similar in both the shake flask and cellulose HFC reactor (Fig. 4).
Hollow fiber cartridge reactors were demonstrated to turn over substrate and produce metabolites with similar efficiency to shake-flask reactors. They also offer several potential advantages compared with shake-flask configurations. As was demonstrated, the biocatalyst in an HFC reactor can be reused many times over. In contrast, the biocatalyst in a shake flask can be used only once because the quenching process kills the microbes. The purification of metabolites biosynthesized in an HFC reactor is also theoretically easier because the products can be harvested from the external medium, which is devoid of the cellular matrix.
An HFC reactor potentially offers a readily available system to generate metabolites for multiple compounds over time because there is no theoretical reason to believe that the substrate would have to be the same in each reaction cycle. The recirculation of medium (containing the substrate) in a closed-loop system allows for multiple passes for the substrate to be turned over. The cartridges employed in this report had an ECS volume between 2 and 15 ml and were sufficient to achieve a 100% turnover of diclofenac (up to 500 µM) in a closed-loop system. Larger volume cartridges are also commercially available (cellulose HFC with a 70 ml ECS; FiberCell Systems) to facilitate scaled-up reactions.
The one drawback observed with an HFC reactor is the potential for nonspecific binding of substrate. Nonspecific binding could occur in any of the components in the reactor. The three cartridges used in this study all had the same length of silicone tubing and the same cartridge shell. The only differences were in the material that comprised the semipermeable membrane. Differences in the membrane material were probably responsible for the differences observed in nonspecific binding. Binding to Teflon and cellulose membranes was substantially less than to the polysulfone membrane. The extent to which other substrates (e.g., hydrophobic or basic compounds) might nonspecifically bind to these materials was not determined.
To our knowledge, this work represents the first example of an HFC reactor used to produce mammalian P450 metabolites using microbial cultures. Previously published examples employing HFC reactors for biochemical reactions include conversions of glucose to ethanol using entrapped Saccharomyces cerevisiae (Inloes et al., 1983), L-histidine to urocanic acid by Pseudomonas fluorescens (Kan and Shuler, 1978), and 2-methyl-1,3-propane diol to (R)-β-hydroxyisobutyric acid by Acetobacter sp. (Leon et al., 2001).
|
Conclusion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
ABBREVIATIONS: HFC, hollow fiber cartridge; SOC, superoptimal medium C; DIC, diclofenac; 4'-OH-DIC, 4'-hydroxy diclofenac; 5-OH-DIC, 5-hydroxy diclofenac; 4',5-di-OH-DIC, 4',5-dihydroxy diclofenac; ECS, extracapillary space; LC-MS/MS, liquid chromatography tandem mass spectrometry; NMR, nuclear magnetic resonance spectroscopy; P450, cytochrome P450.
1 Current affiliation: Department of Medicinal and Natural Products Chemistry, University of Iowa, Iowa City, Iowa. 
2 Current affiliation: Drug Metabolism and Pharmacokinetics, Celgene Corporation, Summit, New Jersey.
Address correspondence to: Dr. Raju Subramanian, One Amgen Center Dr. M/S 30E-2-B, Amgen Inc., Thousand Oaks, CA 91320. E-mail: rajus{at}amgen.com
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
