Abstract
Human butyrylcholinesterase (BChE) hydrolyzes cocaine to inactive metabolites. A mutant of human BChE, A328W, hydrolyzed cocaine 15-fold faster compared with wild-type BChE. Although the catalytic properties of human BChE secreted by Chinese hamster ovary (CHO) cells are identical to those of native BChE, a major difference became evident when the recombinant BChE was injected into rats and mice. Recombinant BChE disappeared from the circulation within minutes, whereas native BChE stayed in the blood for a week. Nondenaturing gel electrophoresis showed that the recombinant BChE consisted mainly of monomers and dimers. In contrast, native BChE is a tetramer. The problem of the short residence time was solved by finding a method to assemble the recombinant BChE into tetramers. Coexpression in CHO cells of BChE and 45 residues from the N terminus of the COLQ protein yielded 70% tetrameric BChE. The resulting purified recombinant BChE tetramers had a half-life of 16 h in the circulation of rats and mice. The 16-h half-life was achieved without modifying the carbohydrate content of recombinant BChE. The protective effect of recombinant wild-type and A328W mutant BChE against cocaine toxicity was tested by measuring locomotor activity in mice. Pretreatment with wild-type BChE or A328W tetramers at a dose of 2.8 units/g i.p. reduced cocaine-induced locomotor activity by 50 and 80%. These results indicate that recombinant human BChE could be useful for treating cocaine toxicity in humans.
Human BChE has a major role in detoxication of cocaine (Kalow and Grant, 2001). Treatment of rodents and cats with human or horse BChE protects from cocaine-induced hypertension, cardiac arrhythmia, hyperactivity, seizures, and lethality (Hoffman et al., 1996; Lynch et al., 1997;Mattes et al., 1997; Carmona et al., 1998). The amount of BChE present in human blood and tissues is insufficient to instantly detoxify inhaled or injected cocaine because BChE hydrolyzes the (−)-cocaine isomer slowly. By contrast, the (+)-cocaine isomer is hydrolyzed 2000- fold faster (Gatley, 1991; Xie et al., 1999) and has none of the physiological effects of (−)-cocaine. The lack of pharmacologic activity of (+)-cocaine is attributed to rapid hydrolysis by BChE (Gatley, 1991). A BChE with a faster rate of hydrolysis might render (−)-cocaine pharmacologically inactive.
One purpose of this work was to increase the catalytic efficiency of human BChE for hydrolysis of (−)-cocaine. A 4-fold increase in catalytic efficiency had been achieved with the A328Y mutant (Xie et al., 1999). In this report a further increase in catalytic efficiency is provided by the A328W mutant, which hydrolyzes (−)-cocaine 15-fold faster than does wild-type BChE. The laboratory of Stephen Brimijoin has engineered other cocaine-hydrolyzing mutants of BChE (Sun et al., 2001). A large aromatic residue at position 328 is expected to orient (−)-cocaine into position for attack by the active site serine (Xie et al., 1999).
A second goal was to produce the A328W mutant BChE in a form that would have a long residence time in the circulation of animals. Saxena et al. (1998) had shown that recombinant human BChE had a very short residence time in the circulation of mice, on the order of minutes, whereas native human BChE purified from plasma had a mean residence time of 46 h. A significant difference between recombinant and native BChE was that recombinant human BChE secreted by Chinese hamster ovary (CHO) cells consisted mainly of monomers and dimers (Blong et al., 1997;Saxena et al., 1998), whereas native BChE is a tetramer. The finding byBon et al. (1997) and Krejci et al. (1997) that a proline-rich peptide from the N terminus of the collagen-tail protein (COLQ gene) caused assembly of acetylcholinesterase into tetramers led us to the experiments that solved the problem. Coexpression of BChE with a 45-amino acid peptide encoded by the COLQ gene converted 70% of the recombinant BChE to tetramers (Altamirano and Lockridge, 1999). Tetrameric recombinant BChE was found to have a residence half-time of 16 h in the circulation of rats and mice. This long half-time was achieved without modifying the carbohydrate content.
A third goal was to demonstrate the protective effect of recombinant human BChE against cocaine toxicity. Native wild-type BChE, purified from human or horse plasma, has previously been shown to protect mice, rats, and cats from cocaine toxicity (Hoffman et al., 1996; Lynch et al., 1997; Mattes et al., 1997; Carmona et al., 1998). However, recombinant human BChE has not previously been tested in animals. The protective effect of recombinant BChE tetramers was shown by measurement of locomotor activity in mice. Cocaine at a dose of 25 mg/kg i.p. induced high locomotor activity, but pretreatment with BChE reduced activity by 50 to 80%. The A328W mutant was more effective than wild-type BChE at reducing locomotor activity. These results indicate that recombinant wild-type and mutant BChE may be useful for treatment of cocaine toxicity in humans.
Materials and Methods
Mutagenesis.
The mutation A328W was introduced into human BChE by polymerase chain reaction as previously described (Xie et al., 1999). The cloned cDNA was completely sequenced to verify the presence of the A328W mutation and the absence of unwanted mutations.
Expression of Human BChE.
Human BChE was expressed in Chinese hamster ovary cells (CHO-K1; ATCC 61-CCL), stably transfected with plasmid pGS-BCHE wild-type or A328W as previously described (Xie et al., 1999). Selective pressure to retain the plasmid was provided by 50 μM methionine sulfoximine in the initial period and reduced to 25 μM for maintenance. Secreted BChE was collected into serum-free and glutamine-free culture medium, Ultraculture (BioWhittaker, Walkersville, MD, catalog number 12-725B), thus avoiding contamination by AChE present in fetal bovine serum. No antibiotics were added to the culture medium. The cells were grown in 1-liter roller bottles. Culture medium (150 ml per bottle) in the roller bottles was changed every 2 to 4 days. A roller bottle yielded BChE continuously for as long as 6 months. Each liter of culture medium contained 1 to 5 mg of BChE.
N Terminus of the Collagen-Tail.
The proline-rich attachment domain is a 17-residue peptide from the N terminus of the collagen-tail encoded by the COLQ gene (Bon et al., 1997; Krejci et al., 1997). It has two cysteines as well as five and three consecutive prolines in the sequence CCLLMPPPPPLFPPPFF. Bon et al. (1997) first reported that this peptide caused recombinant AChE monomers to assemble into tetramers. A clone encoding 117 amino acids of the N terminus of the rat collagen-tail was a gift of Dr. Eric Krejci. We modified this clone by PCR. The modified collagen-tail was called rQ45; it included 22 codons for the signal peptide, 45 for the N terminus of COLQ, and 8 for the FLAG epitope DYKDDDDK cloned into the mammalian expression plasmid, pRc/RSV (Invitrogen, Carlsbad, CA) (see Fig.1). The expression plasmid for rQ45 had a different promoter from the promoter for expression of BChE so that the two would not compete for transcription factors. The promoter for expression of rQ45 was Rous sarcoma virus long terminal repeat, whereas the promoter for expression of BChE was cytomegalovirus. The selectable marker in plasmid pRc/RSV-rQ45 is the NEO gene and in plasmid pGS is glutamine synthetase.
Coexpression of Human BChE and rQ45.
The tetramerization domain of human BChE is located at the C terminus where the 40 amino acids encoded by exon 4 are essential for formation of dimers and tetramers (Blong et al., 1997; Altamirano and Lockridge, 1999). Full-length BCHE cDNA encoding 574 amino acids of the mature protein was used for expression. Stable CHO cell lines expressing wild-type or A328W human BChE were transfected with pRc/RSV-rQ45. Clones expressing both BChE and rQ45 were selected in Ultraculture containing 25 μM methionine sulfoximine and 0.8 mg/ml G418 (geneticin). Cells were amplified in T150 flasks and finally in roller bottles for large-scale production of tetrameric BChE. After cells had coated the roller bottle, the G418 was no longer added to culture medium.
Purification of Recombinant BChE.
Serum-free culture medium was collected from roller bottles over a period of months. The BChE-containing culture medium was stored at 4°C in sterile bottles. Ten to 20 liters of culture medium containing 50 to 100 mg of BChE were filtered through Whatman filter paper no. 1 (Whatman, Clifton, NJ) on a Buchner funnel, or through a coffee filter on a fritted glass funnel attached to a water aspirator, to remove cell debris. The filtered culture medium was loaded onto a 300- to 400-ml procainamide-Sepharose affinity column packed in a Pharmacia column XK50/30 (Pharmacia, Peapack, NJ). This column has a diameter of 5 cm, allowing a flow rate of 1 liter/h. All of the BChE activity was retained by the affinity gel. The column was washed with 20 mM potassium phosphate, 1 mM EDTA, pH 7, until the absorbance at 280 nm of the eluate was nearly zero. When the BChE on the column was wild-type BChE, the column was washed with 0.2 M NaCl in buffer to elute contaminating proteins. When the BChE was A328W, the column was washed with 0.6 M NaCl in buffer because A328W remained bound to the affinity gel at this salt concentration. The column was washed with buffer before eluting wild-type BChE with 1 M NaCl or A328W with 2 M NaCl containing 0.2 M choline chloride in 20 mM potassium phosphate, pH 7. The yield of BChE from this first step was 90 to 100%.
BChE can be eluted from the affinity column with inhibitors or poor substrates rather than NaCl. For example the following have been found to work: 0.2 M procainamide, 0.2 M procaine, 0.2 M decamethonium, 0.2 M acetyl-β-methylcholine, 0.2 M tetramethylammonium bromide, and 0.2 M succinyldicholine. A good substrate such as 0.2 M acetylcholine also elutes the enzyme, but the pH rapidly drops below 4 due to the release of acetic acid by hydrolysis of acetylcholine, and the low pH can inactivate the BChE if exposure is prolonged. We generally choose to elute with NaCl because inhibitors cannot be removed completely from BChE. This is a concern if the BChE is to be used for injection into humans.
The BChE was dialyzed against 20 mM Tris-Cl, pH 7.4, to reduce the salt concentration and then loaded onto 400 to 500 ml of DE52 (Whatman) ion exchanger packed in a Pharmacia XK50/30 column. The column was washed with 20 mM Tris-Cl, pH 7.4, until the absorbance of the eluate was nearly zero. BChE was eluted with 0.15 M NaCl in 20 mM Tris-Cl, pH 7.4. The BChE eluted as a shoulder ahead of a contaminating peak. The yield of BChE from this second chromatography step was about 70%. The cleanest fractions were 80 to 90% pure. Purity was estimated from specific activity and from gel electrophoresis. A specific activity of 720 units/mg was the standard for 100% pure wild-type BChE. Units of activity were measured with 1 mM butyrylthiocholine in 0.1 M potassium phosphate, pH 7.0, at 25°C. Protein concentration was measured by absorbance at 280 nm, where an absorbance of pure BChE, at 1 mg/ml, was 1.8.
Purified recombinant human BChE was dialyzed against phosphate-buffered saline and concentrated to 1 mg/ml in a Millipore Diaflo apparatus (Millipore Corp., Bedford, MA) fitted with a PM10 (particles < 10 μm in diameter) membrane. The dialyzed, concentrated BChE was filter-sterilized through a 0.2-μm filter and stored at 4°C. Although dilute BChE loses activity when it is frozen in the absence of a cryoprotectant such as glycerol, BChE concentrated to 1 mg/ml in phosphate-buffered saline can be frozen without loss of activity.
Purification of BChE from Human Plasma.
Native BChE was purified from human plasma by ion-exchange chromatography at pH 4.0, followed by affinity chromatography on procainamide Sepharose.
Affinity Gel.
Procainamide-Sepharose 4B affinity gel, with a 6 carbon spacer, was custom made by Yacov Ashani at the Israel Institute for Biological Research, Ness-Ziona, Israel. The concentration of bound procainamide was estimated to be 34 μmol/ml. Used affinity gel was recycled by washing on a fritted glass funnel with 0.5 M glacial acetic acid, followed by water. The washed gel was stored in the presence of 20% ethanol at 4°C. The gel has been reused repeatedly for several years, with no apparent loss in binding capacity.
kcat and Kmfor Cocaine, Butyrylthiocholine, and Benzoylcholine.
(−)-Cocaine hydrochloride (purchased from Sigma-Aldrich, St. Louis, MO, after obtaining a controlled substance license from the U.S. Department of Justice) was dissolved in water to make a 0.1 M stock containing 34 mg/ml. Aliquots were frozen at −80°C, thawed once, and discarded. The rate of hydrolysis of (−)-cocaine was measured in the spectrophotometer at 240 nm, using an extinction coefficient of 6700 M−1 cm−1 for the difference in absorbance between cocaine and benzoic acid (Gatley, 1991). The temperature was 25°C, and the buffer was 0.1 M potassium phosphate, pH 7.0.
Butyrylthiocholine and benzoylcholinekcat andKm were measured in 0.1 M potassium phosphate buffer, pH 7.0, at 25°C as described (Xie et al., 1999). The Km was determined for benzoylcholine concentrations ranging from 12.5 to 60 μM.
Purified recombinant BChE was titrated with chlorpyrifos oxon (Chem Service Inc., West Chester, PA) for determination of active site concentration. Maximum rate of hydrolysis per active site, denoted askcat, was calculated by dividingVmax by the concentration of active sites.
BChE Activity Assay.
Serum samples were tested for BChE activity with 1 mM butyrylthiocholine and 0.5 mM 5,5′-dithio-bis-(2-nitrobenzoic acid), in 0.1 M potassium phosphate buffer, pH 7.0, at 25°C. A temperature-controlled Gilford spectrophotometer that interfaced via a MacLab data recorder (ADInstruments Pty Ltd., Castle Hill, Australia) to a Macintosh computer (Apple, Cupertino, CA) was used. Formation of the product was followed by the absorbance increase of 5-thio-2-nitrobenzoic acid at 412 nm, using a molar extinction coefficient of 13,600 M−1 cm−1 (Ellman et al., 1961). Activity is reported as units per milliliter, where 1 unit represents the hydrolysis of 1 μmol of butyrylthiocholine per min.
Nondenaturing Gel Electrophoresis
The relative amount of tetramers, dimers, and monomers was estimated on activity-stained nondenaturing polyacrylamide gels. Four to 30% polyacrylamide gradient gels were prepared in a Hoeffer SE600 gel apparatus (Hoeffer, San Francisco, CA; presently part of Pharmacia, Inc.). Electrophoresis was at 120 V constant voltage for 15 h at 4°C. It was important to keep the gels cold during electrophoresis to avoid degrading the heat-labile monomers and dimers (Blong et al., 1997). Gels were stained for BChE activity in the presence of 2 mM butyrylthiocholine iodide by the method of Karnovsky and Roots (1964). Band intensity was quantified with a BioImage 110S System (Millipore).
Elimination Time of Human BChE from Rats and Mice.
Animal studies were carried out in accordance with the Guide for the Care and Use of Laboratory Animals as adopted by the U.S. National Institutes of Health. Five male Sprague-Dawley rats weighing approximately 350 g were anesthetized by intraperitoneal administration of 60 mg/kg α-chloralose and 800 mg/kg urethane. A cannula made of Silastic tubing was implanted into the jugular vein. Blood for the zero time point was collected. Human BChE was delivered through the cannula to a dose of 1 mg/kg; that is, 0.35 ml of 1 mg/ml BChE (720 U/ml for wild-type BChE) was infused over a 45-s period. The cannula was flushed with 0.1 ml of saline, and a blood sample was collected 1 min after the flush with saline. Blood samples at 1, 3, 6, 12, 24, and 36 h were also collected through the cannula. The rat serum was tested for BChE activity.
Five male, strain 129Sv mice, weighing 23 to 26 g and 55 days of age, were injected intraperitoneally with recombinant human BChE. A maximum of 0.2 ml of BChE, containing 144 units of butyrylthiocholine hydrolyzing activity, was injected i.p. for a dose of 9 mg/kg. The hair was shaved off a hind leg with a scalpel, and the saphenous vein was punctured with a needle. About 10 to 50 μl of blood was collected into a capillary tube at 0, 10, 30, 60, and 90 min, and 3, 5, 7, 22, 30, 56, and 96 h, and transferred to a microcentrifuge tube. The serum was tested for BChE activity. Similarly, 0.2 ml of purified native human BChE containing 438 units was injected i.p. into five mice, at a dose of 27 mg/kg. Blood samples were collected at 0, 1, 2, 4, 6, 8, 10, 24, 34, 48, 72, 96, 168, and 216 h.
The data for elimination of BChE from the circulation were fitted to a double-exponential equation in Sigma Plot (Jandel Scientific, Chicago, IL). The equation is described by Kronman et al. (2000) (see legend to Table 3) as well as by Saxena et al. (1998).
Locomotor Activity.
A motion detector was made by the Instrument Shop at the University of Nebraska Medical Center. It consisted of a red-light-emitting diode that sent a beam of light through the cage wall, a photodiode detector on the opposite side of the cage, and a microprocessor. When the mouse passed through the beam, an event was recorded by the microprocessor. After a defined time, the total number of events was printed. The beam was set up in a dimly lit box containing the home cage in a sound-proof room. The cage contained no bedding because kicked-up bedding could trigger the beam counter. The mouse was acclimated for 1 h before beam breaks were counted.
Adult, male, strain 129Sv mice, 52 to 94 days of age, weighing 24 to 35 g and averaging 28.2 ± 3.0 g, were from a colony at the University of Nebraska Medical Center. Mice received either cocaine alone, saline alone, or BChE followed 1 h later by cocaine. Each mouse received cocaine only once. The dose of cocaine was 25 mg/kg i.p. The dose of BChE was 2.8 units/g i.p., where units were defined as micromoles of butyrylthiocholine hydrolyzed per minute. Animals were not returned to their home cage until at least 16 h after cocaine dosing because cocaine-treated animals displayed aggressive behavior for several hours after treatment.
Results
Catalytic Activity of A328W.
The A328W mutant of BChE was tested for activity with (−)-cocaine, butyrylthiocholine, and benzoylcholine. Table 1 shows that A328W hydrolyzed (−)-cocaine 15-fold faster than did wild-type BChE as measured by kcat. The binding affinity was nearly the same for both enzymes, theKm value being 10 μM for A328W and 7 μM for wild-type BChE. Both enzymes hydrolyzed butyrylthiocholine and benzoylcholine much more rapidly than cocaine. However, thekcat value for A328W was lower for these substrates compared with thekcat for wild-type BChE.
Tetramers of Recombinant BChE.
When BChE was expressed in stable CHO cell lines in the absence of the peptide from COLQ, the BChE consisted predominantly of dimers and monomers, as demonstrated on nondenaturing gel stained for BChE activity. Only 10% of the BChE activity was a tetramer (Blong et al., 1997; Altamirano and Lockridge, 1999). However, coexpression of 45 residues of the N terminus of the rat collagen-tail, called rQ45, resulted in 70% BChE tetramers (Fig.2).
Elimination Half-Life in Rats.
When a purified preparation of recombinant human BChE monomers and dimers was injected into the jugular vein of rats, the half-life was 2 min (Fig.3), a result in agreement with that ofSaxena et al. (1998). No detectable human BChE was present in rat serum after 1 h. In contrast, when a preparation containing 70% tetramers (wild-type BChE rQ45 or A328W rQ45) was injected into rats, the elimination was biphasic, with 81% disappearing with a half-life of 26 ± 3 min and 19% disappearing with a half-life of about 16 ± 6 h (Fig. 3). The mean residence time was 1245 min. Significant amounts of human BChE activity were still present in the rat 24 h after i.v. injection of tetrameric BChE. The 1 to 2.8 U/ml of BChE activity found after 24 h is a 40- to 50-fold increase above the endogenous BChE activity of 0.02 to 0.07 U/ml in rat serum.
The nondenaturing gel in Fig. 4 shows that tetramers of BChE remained in the circulation longer than monomers and dimers. Thus, the human BChE remaining after 24 h was essentially all tetrameric. This result is consistent with the results in Fig. 3, which showed rapid clearance of monomers and dimers of BChE. The lane labeled “0 min” contains 2 μl of rat serum taken before the rat was injected with human BChE. No bands showed up in this lane, supporting the finding that rat serum contains very little endogenous BChE activity.
The C2 band in Figs. 2 and 4 is a dimer formed by covalent bonding of one subunit of BChE and one of albumin (Masson, 1989). Purified BChE does not contain C2. Human serum contains C2. The C2 band in rat serum appeared only after the purified human BChE had been injected into the rat.
Peak Activity in Blood.
Five mice received intraperitoneal injections of the same purified BChE preparation (recombinant wild-type human BChE rQ45 containing 70% tetramers) that had been given to rats. Figure 5 shows that the human BChE rapidly entered mouse blood, where peak activity was achieved about 1 h after injection into the peritoneal cavity. Peak activity in mouse blood was 45-fold above the endogenous activity of 1 to 1.4 U/ml. The activity remained high for hours. After 7 h, it was still 20-fold over endogenous activity. Hoffman et al. (1996) had previously found that native human BChE injected i.p. reached peak activity in mouse blood 1 h after injection. Based on these results, we decided to allow 1 h to elapse between i.p. injection of BChE and injection of cocaine in locomotor activity assays.
Elimination Half-Life in Mice for Recombinant BChE.
Figure6 shows that recombinant BChE disappeared from mouse blood in a biphasic manner, with 57% disappearing with a half-life of 48 ± 20 min and 43% with a half-life of 15.6 ± 2 h. The mean residence time was 1269 min. The pharmacokinetics of the purified human BChE rQ45 were similar in rats and mice.
Blood samples, taken from the mouse various times after i.p. injection of human BChE rQ45, were subjected to gel electrophoresis on a nondenaturing gradient gel. Each lane in Fig.7 received 2 μl of mouse serum. The gel shows that all sizes of human BChE entered mouse blood from the peritoneum. Comparison of lanes 0 and 10 min shows that the human BChE bands migrated more slowly on the gel than the corresponding mouse bands, so that a doublet of tetramer bands is visible. Mouse BChE has 574 amino acids and 7 glycan chains, whereas human BChE has 574 amino acids and 9 glycan chains. The faster migration of mouse BChE is explained by its lower molecular weight. Tetramers of human BChE remained in the mouse blood longer than dimers and monomers, a result similar to that observed in rat blood. The last lane of this gel shows the pattern of bands in the recombinant human BChE rQ45 before it was injected into the mouse. The monomer band is broader, suggesting that some of the diffuse forms did not enter the mouse blood. Another point to notice is that the purified human BChE rQ45 has no C2 band. The C2 band forms in the mouse probably by combining mouse albumin with one subunit of human BChE.
Elimination Half-Life in Mice for Native Human BChE.
Figure8 shows the BChE activity in mouse blood after an i.p. injection of 438 units of native human BChE tetramers (0.6 mg of BChE per mouse). High amounts of human BChE were present in mouse blood 1 to 10 h after injection. The mean residence time was 56.6 h. These results show that the residence time of native human BChE in the mouse circulation is 2.7-fold longer than the residence time of recombinant human BChE.
Cocaine Toxicity Measured as Locomotor Activity.
A dose of 25 mg/kg cocaine i.p. caused increased locomotor activity in mice. Their behavior was markedly different from that of control mice (n = 32) injected with saline or BChE alone, who ambled back and forth for a few minutes, sniffed, defecated, urinated, then settled down and went to sleep (Fig. 9). Locomotor hyperactivity was therefore taken to be a good indicator of cocaine toxicity.
The protective effect of human BChE against cocaine toxicity was tested by pretreating mice with either recombinant wild-type BChE rQ45 (n = 10) or A328W rQ45 (n = 10) or native wild-type BChE (n = 15) at a dose of 2.8 units/g (units measured with 1 mM butyrylthiocholine). One hour after i.p. injection of BChE, the mice were injected i.p. with cocaine at a dose of 25 mg/kg.
Figure 9 shows the number of beam breaks per 10 min as a function of time after injection of cocaine. Mice that received cocaine alone (no BChE) were hyperactive immediately after injection of cocaine; their activity increased during the first 10 min and remained high for the next 60 min. Mice pretreated with recombinant wild-type BChE, or with native wild-type BChE, were 50% less active, and mice pretreated with recombinant A328W were 80% less active (Table2). These results indicate that both wild-type BChE and A328W have a protective effect against cocaine.
In the locomotor experiments by Mattes et al. (1997), the same rats were retested repeatedly with cocaine. This strategy did not work for us. A single injection of cocaine made mice supersensitive to cocaine for as long as 2 weeks. This phenomenon of sensitization is known in the literature (Carey and Gui, 1998). To avoid the complications of sensitization to cocaine, we used a naive mouse for each experiment.
Discussion
Protection against Cocaine Toxicity.
Both wild-type recombinant human BChE and the A328W mutant BChE protected mice from cocaine-induced hyperactivity. The A328W mutant gave more protection, consistent with its higher catalytic activity toward cocaine. Native human wild-type BChE, purified from human plasma, gave a level of protection similar to that of recombinant wild-type human BChE.
Our results with recombinant wild-type BChE agree with the results of others who had used native BChE. Mattes et al. (1997), using native human BChE, and Carmona et al. (1998), using horse BChE, found that pretreatment with BChE attenuated cocaine-induced locomotor activity in rats.
A 22-g mouse contains a total of 17 units of BChE activity (0.3 nmol) in its body, 60% in the intestine and liver, and 20% in plasma (Duysen et al., 2001). The dose of human BChE in the protection experiments was 2.8 units/g which for a 22-g mouse is 3-fold over endogenous BChE content. If the same dose of BChE is required to protect a human, about 270 mg of purified BChE will be needed to protect a 70-kg person.
Recombinant BChE with a 16-h Half-Life in the Circulation.
This is the first report in which recombinant human BChE has been used in protection experiments. Previous work has used native BChE purified from plasma for protection against cocaine toxicity (Hoffman et al., 1996; Lynch et al., 1997; Mattes et al., 1997; Carmona et al., 1998) and for protection against organophosphorus toxicants (Broomfield et al., 1991; Doctor et al., 1991; Brandeis et al., 1993; Raveh et al., 1997).
The problem of short residence time in the circulation of animals had to be solved before recombinant BChE could be studied for its protective properties. Monomers and dimers of recombinant human BChE disappeared within minutes from the circulation of mice (Saxena et al., 1998). On the other hand, native human serum BChE is a tetramer, and its half-life in the circulation is 7 to 12 days in humans, or 2 days in the circulation of mice. We therefore aimed to make BChE tetramers. The breakthrough that made it possible to produce recombinant BChE tetramers was the finding by the laboratory of Jean Massoulié in Paris that a 17-residue proline-rich peptide was required for assembly into tetramers. Bon et al. (1997) reported that assembly of rat AChE into tetramers required the N-terminal peptide from the collagen-tail. Since AChE and BChE are homologous in many features, it was reasonable to expect that this same collagen-tail peptide would facilitate assembly of BChE as well. When the collagen-tail peptide was coexpressed with BChE in CHO cells, the resulting BChE was a tetramer (Altamirano and Lockridge, 1999). Without the collagen-tail peptide, the recombinant BChE consisted mainly of monomers and dimers.
The half-life of recombinant human BChE tetramers was about 16 h in rats and mice. This is an improvement of 480-fold over the half-life of BChE monomers and dimers, which were cleared from the circulation within 2 min. Native human BChE had a 2.7-fold longer residence time in the circulation of mice than recombinant BChE.
Carbohydrate Content of Recombinant BChE.
Saxena et al. (1998)found that recombinant human BChE monomers and dimers secreted by CHO cells are underglycosylated. Recombinant BChE has only fiveN-glycans, whereas native human BChE has nine. The recombinant BChE had a ratio of sialic acid to galactose of about 1, suggesting that nearly all galactose residues were capped with sialic acid. Saxena et al. (1998) concluded that the capping of galactose with sialic acid by itself is not sufficient to confer circulatory stability and that high molecular weight is also important. Our results support the conclusion that the rapid clearance of BChE monomers and dimers is not due to incomplete sialylation but to their small size of 85 and 170 kDa. It was not necessary to modify the carbohydrate content of the recombinant BChE to attain a half-life of 16 h. Assembly into tetramers achieved the desired goal. This contrasts with recombinant AChE expressed in human embryonic kidney 293 cells (Kronman et al., 2000), where a half-life of 15 h was obtained only after modification of the carbohydrate content and after assembly into tetramers. The importance of large molecular size was confirmed byCohen et al. (2001), who attached polyethylene glycol to AChE monomers and found that high molecular weight monomers had a half-life of 26 h in the circulation of mice, even though only 60% of theN-glycans were sialylated.
CHO Cells.
CHO cells are the preferred cell type for expression of human glycoproteins of potential therapeutic value such as erythropoietin, human growth factor, and tissue plasminogen activator (Jenkins and Curling, 1994). Our decision to produce recombinant human BChE in CHO cells was made after we compared growth conditions and yield of BChE in CHO cells and in 293 human embryonic kidney cells. The 293 cells are used to produce human AChE (Kronman et al., 2000). We found that the yield of human BChE was similar in both cell lines. Both cell lines secreted human BChE to a maximum of 5 mg/l. Both cell lines produced monomers and dimers and very few tetramers.Kronman et al. (2000) had to modify the 293 cells to produce α2,6-sialyltransferase, an enzyme deficient in 293 cells, to fully sialylate the AChE. Although CHO cells are also deficient in α2,6-sialyltransferase (Jenkins and Curling, 1994), the recombinant BChE produced by CHO cells was fully sialylated (Saxena et al., 1998), and the tetrameric BChE produced by CHO cells had a half-life of 16 h without introducing sialyltransferase. An additional advantage of CHO cells is their ability to grow in the absence of fetal bovine serum.
Purification of Recombinant BChE.
Recombinant BChE secreted by CHO cells was collected over a period of months. The culture medium was stored at 4°C until 100 mg of BChE had accumulated. No loss of BChE activity resulted during the storage period. A two-step purification procedure consisting of affinity chromatography on procainamide-Sepharose followed by ion-exchange chromatography on DE52 resulted in highly purified BChE tetramers. The simplicity of the purification protocol makes it possible to produce gram quantities of highly purified recombinant BChE.
Potential for Use of Human BChE in People.
Purified BChE has a history of use in the clinic. The literature reports 134 patients who received partially purified human BChE for treatment of prolonged neuromuscular block by the muscle relaxants succinylcholine and mivacurium or for treatment of organophosphorus pesticide poisoning (Table 3). The BChE injected into these patients was a 5% pure preparation from human plasma sold by Behringwerke (Marburg, Germany). It is a dry concentrate of “cholase”. The patients recovered and had no side-effects from the BChE treatment. All of the reports in Table 3 are from Europe or Saudi Arabia, as human BChE is not yet approved for human use in the United States.
In conclusion, human BChE can now be produced in gram quantities by expression in CHO cells. The recombinant BChE consists predominantly of tetramers, and these have an elimination half-life of 16 h in rodents. Both wild-type and A328W mutant human BChE have the potential to be clinically useful for treating cocaine toxicity in humans.
Acknowledgments
We thank Anu Singh from the laboratory of Kenneth Dretchen at Georgetown University, Washington, D.C., for collecting blood from rats she had injected with BChE. We also thank Dr. Eric Krejci at Centre National de la Recherche Scientifique, Ecole Normale Supérieure, Paris, France, for the gift of a clone encoding 117 amino acids of the N terminus of the rat collagen-tail.
Footnotes
-
This work was supported by U.S. Army Medical Research and Materiel Command Grant DAMD17-01-2-0036 and National Institutes of Health/National Institute on Drug Abuse Grant R01 DA011707 (to O.L.), and by a Center Grant to University of Nebraska Medical Center from the National Cancer Institute, Grant CA36727. The opinions or assertions contained herein belong to the authors and should not be construed as the official views of the U.S. Army or the Department of Defense.
-
DOI: 10.1124/jpet.102.033746
- Abbreviations:
- BChE
- butyrylcholinesterase
- CHO
- Chinese hamster ovary
- AChE
- acetylcholinesterase
- RSV
- Rous sarcoma virus
- G418
- geneticin
- 293 cell
- human embryonic kidney 293cell
- Received January 29, 2002.
- Accepted April 15, 2002.
- The American Society for Pharmacology and Experimental Therapeutics