A plant-derived cocaine hydrolase prevents cocaine overdose lethality and attenuates cocaine-induced drug seeking behavior

Abstract

Cocaine use disorders include short-term and acute pathologies (e.g. overdose) and long-term and chronic disorders (e.g. intractable addiction and post-abstinence relapse). There is currently no available treatment that can effectively reduce morbidity and mortality associated with cocaine overdose or that can effectively prevent relapse in recovering addicts. One recently developed approach to treat these problems is the use of enzymes that rapidly break down the active cocaine molecule into inactive metabolites. In particular, rational design and site-directed mutagenesis transformed human serum recombinant butyrylcholinesterase (BChE) into a highly efficient cocaine hydrolase with drastically improved catalytic efficiency toward (−)-cocaine. A current drawback preventing the clinical application of this promising enzyme-based therapy is the lack of a cost-effective production strategy that is also flexible enough to rapidly scale-up in response to continuous improvements in enzyme design. Plant-based expression systems provide a unique solution as this platform is designed for fast scalability, low cost and the advantage of performing eukaryotic protein modifications such as glycosylation. A Plant-derived form of the Cocaine Super Hydrolase (A199S/F227A/S287G/A328W/Y332G) we designate PCocSH protects mice from cocaine overdose, counters the lethal effects of acute cocaine overdose, and prevents reinstatement of extinguished drug-seeking behavior in mice that underwent place conditioning with cocaine. These results demonstrate that the novel PCocSH enzyme may well serve as an effective therapeutic for cocaine use disorders in a clinical setting.

1. Introduction

Cocaine abuse is a global problem with major medical and societal consequences. Despite this, patients suffering from acute toxicity (overdose) are only symptomatically treated and there is no FDA-approved treatment to decrease the likelihood of relapse in rehabilitated addicts. One therapeutic approach to counteract the toxic and addictive psychoactive effects of cocaine is to accelerate the drug’s metabolism using cocaine-hydrolyzing enzymes (Farrell et al., 2019; Zheng and Zhan, 2011). Butyrylcholinesterase (BChE, UniProtKB - P06276) is a native human serum enzyme that is responsible for cocaine inactivation in vivo, resulting in the transient nature of cocaine-induced euphoria (Inaba et al., 1978; Stewart et al., 1979; Stewart et al., 1977). However, the limited hydrolytic capacity of this promiscuous enzyme with (−)-cocaine, the psychoactive enantiomer of that drug, does not allow natural BChE to remove it rapidly enough from the human body in situations of overdose. But now, BChE variants have been rationally-designed to improve catalytic efficiency using molecular dynamics and modeling based on the atomically-resolved structure of the enzyme (PDB 1P0P). These efforts resulted in mutants with greatly improved catalytic efficiency against (−)-cocaine (Anker et al., 2012; Brimijoin et al., 2008; Carroll et al., 2011; Pan et al., 2005; Sun et al., 2002a; Sun et al., 2002b; Xue et al., 2013a; Xue et al., 2013b; Yang et al., 2009; Yang et al., 2010; Zheng et al., 2014). One such cocaine hydrolase, BChE variant (A199S/S287G/A328W/Y332G), derived from mammalian expression systems has been shown to accelerate cocaine metabolism in vivo and fully protect mice and rats from lethal doses of cocaine (Anker et al., 2012; Brimijoin et al., 2008; Carroll et al., 2011; Xue et al., 2013a; Xue et al., 2013b; Yang et al., 2010; Zheng et al., 2014; Zheng et al., 2008). This enzyme has also been shown safe and effective in humans to accelerate cocaine metabolism in Phase I and Phase II clinical trials (Cohen-Barak et al., 2015; Gilgun-Sherki et al., 2016; Shram et al., 2015; Zhang et al., 2017). In an effort to further improve the efficiency of cocaine hydrolysis, a pentavalent mutant (A199S/F227A/S287G/A328W/Y332G) was designed and was shown to be an even more effective cocaine hydrolase both in vitro and in vivo, and additional improved variants continue to be developed (Xue et al., 2013a; Xue et al., 2013b; Yang et al., 2009; Zheng et al., 2014).

An effective and practical enzyme-based therapy requires inexpensive production of the enzyme in large quantities that are easily amenable and scalable to change as new, more effective mutations are designed and discovered. A plant-based recombinant protein expression platform offers high levels of scalability, safety, and flexibility that are unachievable in mammalian cell-based systems. We have recently provided evidence that the full-length pentavalent BChE mutant can be expressed in plants at commercially viable levels and exhibits a > 2000 fold rise in catalytic efficiency against cocaine in vitro as compared to the WT enzyme (Larrimore et al., 2013; Larrimore et al., 2017). The plant-derived pentavalent mutant enzyme we rename PCocSH (for Plant-derived Cocaine Super-Hydrolase), proved to be enzymatically similar to the low-scale, mammalian-cell produced enzyme (Zheng et al., 2008; Zheng and Zhan, 2011).

Relapse into cocaine use can occur even after prolonged abstinence especially if a subject is exposed to cocaine-associated environmental cues or samples the drug again (Gawin and Kleber, 1984; O’Brien et al., 1998). Laboratory studies both in humans and animals find that a single priming injection of cocaine increases motivation for more cocaine (de Wit and Stewart, 1981; Jaffe et al., 1989; Preston et al., 1993). An animal model used to demonstrate this effect is the reinstatement of extinguished cocaine conditioned place preference (CPP) (Itzhak and Martin, 2002; Mueller and Stewart, 2000). Cocaine CPP measures motivation to seek cocaine as an animal’s preference for spending time in an environment where it previously experienced the psychoactive effects of cocaine over a distinctly different environment previously experienced in a non-drug state. During the conditioning phase of this model, animals form an association between the effects of cocaine and the environment, which subsequently motivates them to spend more time in the cocaine-paired compartment when they are given free-choice access to both compartments during a preference test. To use this model to measure the motivational effects of a cocaine priming injection, the cocaine CPP is first extinguished by returning the animal to the cocaine-associated environment without inducing a cocaine state. During this extinction procedure, the animal learns that the environment is no longer predictive of cocaine psychoactive effects, and the expression of CPP therefore decreases. Subsequently, animals are given a priming injection of cocaine prior to a preference test, which produces motivational effects that reinstate cocaine CPP (Mueller and Stewart, 2000).

Here, we demonstrate for the first time, the efficacy of this plant-derived PCocSH enzyme in vivo for its ability to protect mice from the effects of lethal cocaine overdose, reverse the effects of acute cocaine toxicity, and prevent the motivational effects of a cocaine priming injection.

2. Materials and methods

2.1. Protein expression and purification

Construction of the plant-expression optimized pentavalent mutant of full-length BChE (PCocSH) was previously described (Larrimore et al., 2013; Larrimore et al., 2017). The PCocSH enzyme was transiently expressed in WT Nicotiana benthamiana plants using the magnICON vector system based on deconstructed tobacco mosaic virus. All PCocSH prepared for animal studies was derived from plasmid pTM783 and has a C-terminal 6× histidine tag (Hx6). The recombinant PCocSH was extracted as described previously (Larrimore et al., 2013; Larrimore et al., 2017). The preparation was subject to Concanavalin A (ConA) purification and was eluted with five stepwise increasing concentrations of methyl-α-D-mannopyranoside (0.05 M, 0.1 M, 0.2 M, 0.5 M, and 1 M). Eluates showing similar degrees of purity were pooled, concentrated, and dialyzed against 20 mM sodium phosphate buffer, pH 7.5. The partially-purified PCocSH was then subject to final polishing using batch procainamide (Sigma, catalogue no. P2240) affinity chromatography to remove any remaining contaminants. Proteins were released by stepwise elution as follows: 0.05 M NaCl, 0.5 M NaCl, 1 M NaCl, 1 M NaCl and 0.2 M Procainamide HCl, 1 M NaCl and 0.3 M Procainamide HCl. Eluates were dialyzed against PBS, pH 7.4 and eluates of similar purity were pooled and concentrated. After concentration, 0.02% sodium azide (NaN₃) was added to prevent unwanted organismal growth during storage, which was subsequently dialyzed out against PBS, pH 7.4 prior to use for animal experiments. Although the 6×H tag in the construct was intended to facilitate purification in case the mutant BChE would be impaired in its procainamide binding, it turned out to be unnecessary (data not shown).

2.2. Enzyme biochemical characterization

The amount and concentration of PCocSH in final purified product was determined both based on enzymatic activity and absorbance of purified material at 280 nm using a NanoDrop 2000 (Thermo Scientific) or DU 640 Spectrophotometer (Beckman Coulter). The molecular weight (65,957 Da) and extinction coefficient (ε = 133,310 M⁻¹ cm⁻¹) were determined based on the amino acid sequence of PCocSH (assuming fully cleaved ER-directing signal peptide) using the ProtParam tool from the Expert Protein Analysis System (ExPASy) Bioinformatics Resource Portal (Wilkins et al., 1999). Purified preparations of PCocSH were fractionated using size exclusion Alliance high-pressure liquid chromatography (HPLC, Waters) with a Shodex KW-803 column (8 × 300 mm, Kawasaki) as recently described (Larrimore et al., 2017). Enzymatic activity of the HPLC fractions was assayed by modified Ellman assay as described previously (Larrimore et al., 2013; Larrimore et al., 2017). Purified PCocSH and WT N. benthamiana proteins were resolved by SDS-PAGE on 8% polyacrylamide gels followed by silver staining or chemiluminescence western blotting using rabbit polyclonal anti-human BChE antibodies (a kind gift from Dr. Oksana Lockridge) and horse radish peroxidase secondary antibodies (Santa Cruz Biotechnology).

2.3. Animals

Black C57BL/6 male mice weighing 25–30 g at the start of the experiment were housed 5 mice per cage with free access to food and water in a temperature-controlled colony room with a reversed 12-h light:dark cycle and allowed at least one week to acclimate upon arrival to the housing facility.

All animal experiments were performed in accordance to the NIH guide for the care and use of laboratory animals (8th edition) and protocols approved by the Institutional Care and Use Committee of Arizona State University. Moribund animals, and all animals by the end of the observation period were euthanized by CO₂ asphyxiation followed by a secondary method of euthanasia (cervical dislocation). Moribundity was defined as prolonged seizures (≥120 s), or more than two shorter seizures within 2 min.

2.4. Intravenous enzyme and Intraperitoneal drug administration

For intravenous (i.v.) injections of highly purified PCocSH or vehicle control (PBS, pH 7.4), mice were placed in a commercial mouse restrainer. The exposed tail was cleaned with an alcohol wipe and a U-100 insulin syringe (28G1/2) needle (Becton Dickinson) was used for injection into one of the tail veins. The i.v. injection volume did not exceed 5 μL/g body weight. The (−)-cocaine hydrochloride (RTI International Triangle Park, NC, USA) was dissolved in sterile 0.9% saline and filtered through a 0.2 μm membrane. Mice were released from restraint as soon as the injection was complete and sterile gauze and pressure were applied to the injection site to stop any bleeding. For intraperitoneal (i.p.) injections of cocaine or vehicle (0.9% saline), the same kind of syringe as described above was used. Injection volume never exceeded 10 μL/g per mouse.

2.5. Protection experiment

Vehicle control (1× PBS, pH 7.4) or purified PCocSH (3 or 10 mg/kg) was administered i.v. (tail) 5 min before i.p. administration of varying doses of (−)-cocaine. The presence or absence of convulsions, lethality, and general observations were recorded for 60 min after cocaine administration or, if applicable, until euthanasia criteria were met. First, the dose response to cocaine in absence of any enzyme protection was tested using the following cocaine doses: 0, 30, 55, 75 and 100 mg/kg. Higher varying doses of cocaine (100, 180, 325, 585, and 1055 mg/kg) were needed to observe toxicity when PCocSH (3 or 10 mg/kg) was administered 5 min prior. Group size for each experimental treatment was n = 6.

The median lethal dose (LD50) for cocaine was calculated for each group by least-squares regression using the four-parameter dose-response model (GraphPad Prism) with the span fixed at 0–100%. The R² values were 0.979, 0.998 and 0.999 for, respectively, control mice and mice pre-treated with lower and higher doses of PCocSH. Statistical differences between the values were tested by the Extra Sum-of-squares F Test (GraphPad Prism).

2.6. Rescue from acute cocaine toxicity

For the rescue study, cocaine (100 mg/kg) was given i.p. Immediately upon the onset of cocaine-induced convulsions, mice (n = 6 per experimental group at the beginning of the experiment) were i.v. injected with PCocSH (3 or 10 mg/kg mouse) or PBS control.

2.7. Reinstatement of extinguished conditioned place preference

CPP experiments were conducted using Plexiglas two-compartment chambers as described previously (Der-Ghazarian et al., 2017). Each chamber contained a removable partition dividing the chamber into two equal-sized compartments, each measuring 35 × 24 × 31 cm high. One compartment had a wire 1 × 1 cm grid floor and alternating black and white vertical stripes (2 cm wide) on the walls. The other had a parallel bar floor (5 mm diameter) and alternating black and white horizontal stripes (2 cm wide) on the walls. Both compartments had corncob bedding beneath the wire floors. The chambers were located in a testing room, dimly lit with two overhead lamps, each with a 25 W bulb. A camera (Panasonic WV-CP284, colour CCTV, Suzhou, China) was mounted 101 cm above the center of the apparatus to record the sessions. A WinTV 350 personal video recorder (Hauppage, NJ, USA) captured live video and encoded it to MPEG streams. A modified version of TopScan Software (Clever Sys., Inc. Reston, VA, USA) tracked the animal’s body parts (e.g., nose, head, center of body, forepaws, base of tail, etc.) to estimate the position of the whole body when some parts were not in view in order to yield measures of time spent in each compartment.

The CPP procedure is described in detail elsewhere (Fig. 1) (Der-Ghazarian et al., 2017). Briefly, it involved the following phases (Fig. 1A): 1) habituation, during which mice were simply exposed to the chamber to habituate them to novelty stress, 2) initial preference test, during which the mice had free-access to both sides of the chamber (Fig. 1B, left), 3) conditioning, during which mice were confined to their initially least preferred compartment after receiving an injection of cocaine and during alternating sessions were confined to their initially preferred compartment after receiving saline (i.e., placebo) (Fig. 1B, right), 4) CPP test during which mice were again given free-access to both compartments; CPP was operationally defined as an increase in time spent in the drug-paired compartment after conditioning relative to before conditioning, 5) extinction, during which mice were confined to the drug-paired compartment without receiving cocaine while also being confined to the placebo-paired compartment during alternating sessions, 6) extinction test, during which mice were again given free-access to both compartments and extinction was verified as a decrease in time spent in the drug-paired compartment relative to the CPP test, and finally, 7) reinstatement test for which mice were given a priming injection of cocaine prior to testing for reinstatement of CPP, defined as an increase in time spent in the drug-paired side relative to the extinction test. These experimental phases are detailed as follows (Fig. 1C):

Fig. 1.

Schematic representation of behavioral studies testing prevention of reinstatement of cocaine-conditioned place preference (CPP) in mice treated with PCocSH. (A) Timeline of CPP experiment (B) Diagram of chambers used in CPP studies. The chambers have two compartments separated by a removable partition. (C) Overview of treatments given during the CPP study. During the habituation and the initial preference test (Days 1–2), and subsequent CPP, post-extinction, and reinstatement tests (Day 13, 28 and 29), mice had free access to both compartments of the chamber by removing the middle partition. During the 8-day long conditioning phase, cocaine-conditioned mice received alternating sessions (1 session/day) in their initially least preferred compartment (i.e., drug-paired side) immediately after a cocaine injection (10 mg/kg, i.p.) and in their initially preferred compartment immediately after a saline injection. Control mice received only saline injections immediately prior to all sessions, but otherwise were treated the same as the cocaine-conditioned mice. Following two days of rest (Days 11–12, mice kept in their vivarium cages) and the CPP test (Day 13), a two week-long extinction phase commenced (Days 14–27). During this phase, all mice were injected daily with saline prior to their confinement to either compartment on alternating days. The post-extinction test was conducted on Day 28 and was followed by the reinstatement test (Day 29) in which mice received a priming injection of cocaine or saline after receiving an i.v. injection of PCocSH or saline.

Habituation day (Day 1). Mice were transported to the conditioning room and allowed free excess to both sides of the apparatus for 10 min to habituate them to the novel environment.

Pre-conditioning/baseline test (Day 2). Initial baseline side preference was assessed by allowing mice to freely explore both sides of the chamber without a middle partition for 15 min with time spent in each compartment being recorded. Time spent on each side of the chamber was used to determine side preference for each mouse. All subsequent tests were identically conducted, and, except where noted mice received no further treatment on test days.

Conditioning phase (Days 3–10). Based on the baseline preference data, each animal’s least preferred side was designated as the drug-paired compartment and the preferred side was designated as the placebo-paired compartment. Mice were divided into groups that either received saline or cocaine (10 mg/kg, i.p.) immediately before they were confined to the drug-paired side for 30 min. For alternating sessions, all mice received saline immediately before being confined to the placebo-paired side for 30 min. Note that the initially least preferred compartment was designated as the “drug-paired” compartment a priori for saline controls even though this group received only saline. Mice underwent eight daily sessions, each held at the same time of day across all eight consecutive days. The order of drug and placebo sessions was drug first then placebo for half of the mice and placebo then drug for the other half. To equate the cocaine conditioned and saline control groups, the difference between time spent in the preferred minus nonpreferred compartment during the baseline preference test and the specific side preferred were equally distributed among the groups.

CPP test (Day 13). Three days after the last conditioning session, mice were tested for CPP by conducting a preference test as previously done for the pre-conditioning test. The 72-h interval between the last conditioning day and the test eliminated residual cocaine and recovery influences (Benuck et al., 1987) and allowed the consolidation of long-term memory of drug reward (Kuo et al., 2007). Test was conducted identically to the previous pre-conditioning test.

Extinction (Days 14–27). In order to extinguish cocaine-CPP, mice received repeated exposures to the drug-paired compartment without cocaine while also receiving exposure to the placebo-paired compartment on alternate days to avoid relative novelty on the subsequent test day. Extinction sessions occurred at the same time of day across four-teen consecutive days. Mice received a saline injection prior to each session and were subsequently placed into the appropriate compartment for 20 min.

Post-extinction test: (Day 28). To determine extinction of cocaine-CPP, mice were tested on the day following the two-week long extinction period and without further treatment. Test was conducted identically to the previous pre-conditioning and CPP tests.

Cocaine-induced reinstatement test: (Day 29). On the day after the post-extinction test, mice were assessed for cocaine-induced reinstatement of CPP with and without treatment with PCocSH. To this end, mice were placed into a restrainer and injected i.v. through the tail vein with saline or enzyme at a 3 mg/kg dose at a volume of 5 μL/g body weight. Five minutes later, a priming injection of cocaine (10 mg/kg) was administered (i.p.) at a volume of 10 μL/g body weight. The number of mice per group were as follows: vehicle control (no PCocSH) with vehicle control (no cocaine) (n = 9), vehicle control (no PCocSH) with cocaine (n = 10), PCocSH with vehicle control (no cocaine) (n = 12), and PCocSH with cocaine (n = 12). After the priming injection, the animals were immediately placed into the apparatus for 15 min with unrestricted access to both compartments. Locomotor activity and time spent in the cocaine- and saline-paired compartments were recorded.

2.8. Statistical analysis

For rescue experiments, statistical analyses were carried out using the GraphPad Prism software. Log-rank (Mantel-Cox) test was used to determine significance of the difference between treatment and control group survival and recovery curves (Fig. 5C and 4D). Animal subjects still alive by the experiment endpoint time (Fig. 5C, 60 min) are censored.

Fig. 5.

Rescue from cocaine-induced toxicity. (A) Timeline of rescue experiment. Mice were treated first with cocaine, monitored for appearance of cocaine toxicity symptoms (occurring within the first 5 min of the challenge for all subjects) and then treated with vehicle or with one of two doses PCocSH. Each mouse was challenged only once and was euthanized 60 min after cocaine administration or upon moribundity. (B) Symptom severity displayed by each mouse at the end of the 60-min experiment is shown. Symptom score was 0, asymptomatic; 1, decreased motor activity; 2, tremors or fasciculation; 3, convulsions; 4, death. Individual mice (represented by symbols) were i.p. injected with cocaine in saline (100 mg∕kg). Within approximately 1 min of onset of cocaine-induced seizures, PCocSH at 3 mg/kg (blue) or 10 mg/kg (red) or vehicle control (black) was delivered i.v., and animals were observed for 60 min. (C) Survival curves of mice treated with vehicle control (n = 5), PCocSH 3 mg/kg (n = 6) or 10 mg/kg (n = 5). Comparison by the Log-rank (Mantel-Cox) test between low dose and control (p = .0016) and high dose and control (p = .0027) were significant. (D) Time to recovery from cocaine-induced seizures following PCocSH treatment. Time (in minutes) from the onset of cocaine-induced convulsions until upright posture and normal motor activity resumed is shown. Comparison by the Log-rank (Mantel-Cox) test between low dose and control (p = .001) and high dose and control (p = .0016) were significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4.

Protection from cocaine-induced toxicity. (A) Timeline of protection experiment. PCocSH (3 or 10 mg per kg mouse body weight) or vehicle control was delivered (i.v.) 5 min prior to the administration of varying doses of cocaine (i.p). Acute symptoms in control animals arose within the first 5 min following the cocaine challenge. Experiment was concluded 60 min post cocaine administration or upon moribundity (see Materials and Methods). (B) Dose response curves of cocaine-induced convulsions and tremors and (C) lethality. Each data point represents the % of mice (n = 6), unless otherwise noted in main text) for each dosing condition. Cocaine doses are expressed in mg cocaine per kg mouse body weight and plotted on a logarithmic scale.

For CPP experiments, we first examined whether there were any differences in the amount of time spent in the initially least preferred compartment among the saline-conditioned groups that were or were not treated with enzyme. As expected, there were no differences among these groups, and therefore to simplify further analyses, these groups were combined to comprise a saline control group. Time spent in the initially least preferred compartment was then analyzed using a mixed factor ANOVA with group as a between-subjects factor (Saline Controls, No PCocSH Cocaine-conditioned, and PCocSH Cocaine-conditioned) and test day as a repeated measure (baseline, CPP, extinction, reinstatement). Interactions were further analyzed using smaller ANOVAs and t-tests with Bonferroni correction for multiple comparisons. For all ANOVAs, Mauchly’s test of sphericity was conducted and when this test indicated heterogeneity of variance, the degrees of freedom were adjusted using the Greenhouse-Geiser correction.

3. Results

3.1. Pretreatment with PCocSH protects against cocaine-induced toxicity

The aim of this study was to develop a plant-based cocaine-hydrolyzing enzyme toward the generation of an anti-cocaine therapeutic (Fig. 2).

Fig. 2.

Depiction of the platform for the production of clinically relevant quantities of plant-derived BChE variants. (1) The gene of interest (in this study PCocSH) is codon-optimized for plant expression and relevant site-directed mutations are made. The gene is then cloned into a transient plant expression vector system such as magnICON (T-DNA construct is drawn to scale; length bar corresponds to 1000 bp). (2) This system is transformed into Agrobacteria tumefaciens which readily infects plants. (3) To infiltrate the bacteria into the intercellular spaces of the leaf tissue, entire plants are submerged in bacteria solution and a vacuum is applied. (4) Following infiltration, plants are returned to the growth chamber or greenhouse until the day post inoculation resulting in peak recombinant protein expression (established during small-scale time-course experiments). (5) On the peak expression day leaf tissue is homogenized, and total soluble proteins are filtered and clarified. (6) PCocSH can be purified by sequential Concanavalin A and procainamide purification steps. (7) Following dialysis into a suitable buffer, the purified PCocSH product can be stored for several years at 4 °C and used for testing.

In order to test our hypothesis that PCocSH could counter cocaine-induced acute toxicity, we first sought to determine the timeline of the intoxication and the cocaine dose-response in the presence or absence of the plant-derived enzyme. First, highly purified PCocSH was prepared from plant material. Purified preparations of PCocSH were fractionated using size exclusion high-pressure liquid chromatography (HPLC) (Fig. 3A) and enzymatic activity of the HPLC fractions was assayed by modified Ellman assay (Fig. 3B). Purified PCocSH or crude WT N. benthamiana proteins were resolved by SDS-PAGE followed by silver staining (Fig. 3C) or western blotting using anti-human BChE antibodies (Fig. 3D).

Fig. 3.

Purification of PCocSH for animal studies. (A) Size exclusion-HPLC analysis and (B) enzymatic activity of fractions collected during SE-HPLC analysis. Monomer, dimer, and tetramer fractions are indicated with 1°, 2°, and 4° respectively. Blue dextran (2000 kDa), beta amylase (200 kDa), albumin (66 kDa) and carbonic anhydrase (29 kDa) were included as size references and respective peak elution times are indicated by arrows. Samples for enzymatic activity were collected from the HPLC (0.5 mL) every 1 min. (C) Silver stain and (D) western blot of pooled, purified preparations of PCocSH used in animal studies alongside uninfiltrated wildtype (WT) Nicotiana benthamiana control plant extract. Please note that monomeric plant-derived BChE is glycosylated with high mannose glycans thereby raising its apparent molecular mass.

This highly purified PCocSH (3 or 10 mg/kg) or vehicle control (PBS, pH 7.4) were i.v.-administered to two groups of mice (n = 6) that were challenged five minutes later with varying concentrations of cocaine (delivered i.p.).

The mammalian-derived enzyme was shown to be protective against a lethal 180 mg/kg dose of cocaine at a low enzyme dose of approximately 1 mg/kg in mice when delivered 1 min before cocaine administration (Xue et al., 2011; Zheng et al., 2008). In a separate study, a dose-response curve for seizures was generated using a higher enzyme dose of 10 mg/kg in mice when given 10 min prior to cocaine administration (Brimijoin et al., 2008). We sought to administer the enzyme at 5 min prior to cocaine administration and determine if there was a dose-dependent protection afforded by the enzyme; therefore, we chose to test both a low (3 mg/kg) and high (10 mg/kg) enzyme dosing regimen for this study (Fig. 4A).

Mice that were pre-treated with the low dose of PCocSH were almost fully protected from a cocaine dose (100 mg/kg) that killed 100% of mice in the vehicle control group (Fig. 4B). In fact, only one mouse of the 6 challenged with this cocaine dose showed signs of acute toxicity (Fig. 4B). This mouse seized for > 60 s and was subsequently euthanized. None of the other mice (5/6) had any seizures, tremors, or other signs of acute toxicity. At 585 mg/kg cocaine, all mice seized for > 60 s and were euthanized.

Mice were then treated with a higher enzyme dose of 10 mg/kg. Mice given this treatment did not exhibit any signs of cocaine-induced seizures or tremors at 100 or 180 mg/kg cocaine, which was 100% lethal in unprotected mice (Fig. 4B). At 180 mg/kg the mice initially had slightly decreased motor activity, which subsided within a few minutes. At 353 mg/kg most mice began to show neurologic symptoms (dragging feet, dystaxia, knuckle walking) but only 1 mouse exhibited seizures. At an even higher dose of 585 mg/kg, all mice (100%) exhibited neurologic symptoms described above as well as labored breathing and decreased locomotion. These mice had decreased responsiveness to stimuli, dyspnea, and decreased locomotion indicative of phase II acute cocaine toxicity rather than the premorbid phase III state. At this dose, 50% of the mice (3/6) recovered during the observation time and only three became moribund and had to be euthanized.

Mice that were pre-treated with the low dose of PCocSH were almost fully protected from a cocaine dose (100 mg/kg) that killed 100% of mice in the vehicle control group (Fig. 4C). With a protective PCocSH dose of 10 mg/kg, LD100 was reached only at 1055 mg/kg of cocaine, a > 10-fold higher dose compared to unprotected control mice. The calculated LD50 for the lower and higher PCocSH doses were, respectively, 179 mg/kg and 585 mg/kg as compared to 68 mg/kg in the unprotected controls (95% confidence intervals were, respectively, 166–193, 584–586 and 65–71 mg/kg, P < .0001). These results are similar to the ability of a mammalian (CHO cell) derived albumin-fused BChE variant (A199S/S287G/A328W/Y332G) to protect rats (Brimijoin et al., 2008) and to the protection afforded by the bacterially-derived cocaine esterase (CocE) enzyme to protect mice (Ko et al., 2007) and rats (Cooper et al., 2006).

3.2. PCocSH rescues mice from cocaine overdose

The protection experiment (Section 3.1) allowed us to determine the dose that produced acute-toxicity responses to cocaine in protected and unprotected mice. To simulate a more clinically relevant scenario, we then determined if PCocSH provided protection when delivered after cocaine-challenged subjects become symptomatic. To evaluate the potential for PCocSH to rescue mice from cocaine overdose, mice received the LD100 dose of cocaine as determined above (100 mg/kg, i.p.). Immediately upon onset of convulsions, mice were treated by i.v. injection through the tail vein with either low dose (3 mg/kg), high dose (10 mg/kg) of PCocSH or vehicle control (Fig. 5A).

As expected, all of the mice challenged by cocaine experienced severe effects of cocaine toxicity within 5 min of the challenge, except for one control mouse that displayed no severe symptoms and therefore was excluded from the study. The interval between cocaine administration and onset of convulsions was 4.8 ± 0.5 min (n = 17, mean ± SEM, standard error of the mean). One animal from the PCocSH (10 mg/kg) group died later while being restrained prior to receiving further treatment and was subsequently removed from the study.

The 100 mg/kg dose of cocaine was lethal to all 5 control (vehicle-treated) mice that exhibited cocaine-induced convulsions and were euthanized upon moribundity. Strikingly, all of the mice (n = 6) that were treated with 3 mg/kg PCocSH soon after the onset of seizures survived (Fig. 5B). In all PCocSH-treated subjects, convulsions ceased within 1 min of treatment. Thus, PCocSH can reverse the acute effects of cocaine-induced toxicity in mice when given soon after the onset of convulsions. Moreover, four out of six mice treated with the lower dose of PCocSH fully recovered by the end of the observation period with only two of the mice (2/6) continuing to experience some mild symptoms (Fig. 5B). Delivery of the enzyme for these two mice took nearly 3 times longer than the other 4 mice due to the technically challenging nature of conducting an i.v. injection on a seizing animal. This prolonged time of seizing before receiving the full dose of the enzyme likely led to the lethargy observed in these mice.

Similar results of rescue were observed for mice (n = 5) treated with the high dose of PCocSH (10 mg/kg) (Fig. 5C). All of these mice ceased convulsing in < 1 min and all mice resumed upright posture in < 10 min (Fig. 5D). Mice that received the full 10 mg/kg dose of PCocSH after onset of convulsions induced by 100 mg/kg cocaine exhibited an extremely rapid return to a normal state. In the time it took to replace the animal in its cage (< 30 s), the mouse began to walk normally and resumed normal behavior.

In both the low and high enzyme groups, at least 67% of the mice regained upright posture in 1 min or less following enzyme treatment, compared to 0% of the mice in the control group (Fig. 5D). All (100%) of the mice treated with complete doses of PCocSH (entire 3 or 10 mg/kg) which had exhibited cocaine-induced seizures were saved from subsequent death compared to 0% of mice which seized and received vehicle control.

3.3. PCocSH prevents reinstatement of cocaine-conditioned place preference

In addition to acute cocaine overdose, another highly problematic issue of cocaine use is relapse after a period of abstinence, which is often initiated by an increase in motivation for cocaine that occurs after consuming a small amount of cocaine (de Wit and Stewart, 1981; Jaffe et al., 1989; Preston et al., 1993). As our plant-derived enzyme was efficient enough to protect and rescue mice from lethal doses of cocaine, we next chose to investigate its potential therapeutic use in preventing cocaine-seeking behavior induced by a cocaine priming injection using the mouse model of reinstatement of extinguished cocaine-conditioned place preference (CPP) (Fig. 1).

To analyze the CPP data, we first determined whether the two groups of saline control mice (animals that were mock-conditioned with saline rather than with cocaine and either received the enzyme or vehicle prior to their post-extinction priming with cocaine, named here PCocSH and No PCocSH) could be combined based on the predicted lack of enzyme effect in these groups. A mixed factor ANOVA (four test days by and enzyme condition) of time spent in the drug-paired compartment verified that there was no effect of enzyme nor interaction with enzyme. Therefore, all saline CPP mice formed a single control group, which increased power in the control condition and simplified the experimental design to avoid the potential problem of spurious higher order interactions (Fig. 6).

Fig. 6.

Time spent in the initially least preferred side over the course of the CPP study. Dashed line at 450 s represents 50% of the total test time. *Represents a difference from respective baseline and from Saline Controls. #Represents a difference from respective extinction test. +Represents a difference from No PCocSH, Cocaine-conditioned group. Bars represent means ± SEM, standard error of the mean. The number of mice per group were as follows: vehicle control (no PCocSH) with vehicle control (no cocaine) (n = 9), vehicle control (no PCocSH) with cocaine (n = 10), PCocSH with vehicle control (no cocaine) (n = 12), and PCocSH with cocaine (n = 12).

Subsequently, a mixed factor repeated measures ANOVA [four test days (baseline, CPP, post-extinction and reinstatement tests) by group (controls, PCocSH-Cocaine, and No PCocSH-Cocaine)] of time spent in the drug-paired compartment was performed. The analysis indicated significant main effects of test day [F(2.28, 86.72) = 21.5, p < .001] and group [F(2,38) = 8.35, p < .001], and a test day by group interaction [F(4.56, 86.72) = 4.62, p < .001]. To analyze the interaction, separate repeated measures ANOVAs of time spent in the drug-paired side for each group across test days were conducted. There was no significant effect of test day for the control group [F(3, 57) = 2.727, p = .052], nor did this group switch preference to the initially nonpreferred side as they still spent less than 50% of the total test time on that side across all tests. The No PCocSH-Cocaine group showed changes in the time spent in the drug-paired side across tests [F(3, 27) = 21.95, p = .001]. CPP was evident as an increase in time spent in the drug-paired compartment relative to baseline [t(9) = −9.75, p < .001] with values over 50% of the total test time, indicating a switch in preference to the drug-paired side. This cocaine CPP was significantly attenuated by extinction training as expected [CPP test compared to extinction, t(9) = 5.35, p < .001] and the cocaine priming injection reinstated CPP [extinction compared to reinstatement, t(9) = −2.68, p < .0125]. Similarly, the PCocSH-Cocaine group showed changes in the time spent in the drug-paired compartment across tests [F(1.49,14.92) = 12.06, p = .005], including a significant switch in preference to the drug-paired compartment [CPP test compared to baseline, t(10) = −10.51, p < .001] that diminished as a result of extinction training [CPP test compared to extinction, t(10) = 5.34,p < .001]; however, the cocaine priming injection failed to reinstate the CPP in this group. We also compared groups to each other at each test day. The only test indicating significant differences among the groups was the CPP test [F(2,38) = 15.69, p < .001], where both cocaine-conditioned groups spent more time in the drug-paired compartment compared to saline controls, and the magnitude of CPP was greater in the PCocSH-Cocaine group compared to the No PCocSH-Cocaine group [t(19) = 2.46, p < .025] (Fig. 6). The difference in magnitude of initial CPP between the latter two groups was unexpected because they had undergone identical procedures at this point in the experiment. Because these groups displayed a different magnitude of initial CPP, it is important to rely on within group comparisons of test results when interpreting the results, where each subject serves as its own control. Accordingly, we conclude that the cocaine prime reinstated CPP in the No PCocSH-Cocaine group because they showed an increase in time spent on the drug-paired side on the reinstatement test compared to the extinction test, but had no effect in the PCocSH-Cocaine group because they showed no change across these two tests.

4. Discussion

Cocaine asserts its initial neuromodulatory effects primarily through inhibition of monoamine (dopamine, serotonin and norepinephrine) reuptake in the brain, leading to the persistence of these neurotransmitters in their respective synapses (Connors and Hoffman, 2013). Overexcitation of these brain circuitries rapidly result in complex cascades of effects within the central nervous system and beyond (e.g. the cardio-vascular system) (Connors and Hoffman, 2013; Kim and Park, 2019). Because of the pleiotropic nature of the response to cocaine, treating acute cocaine intoxication by directly muting effects of the excess neurotransmitters (i.e. a pharmacodynamic approach) is problematic (Connors and Hoffman, 2013; Skolnick et al., 2015). Instead, tackling the drug directly by neutralizing or metabolizing it (i.e. a pharmacokinetic approach) has shown more promise (Connors and Hoffman, 2013; Zhang et al., 2017; Zheng and Zhan, 2012). It is possible to sequester cocaine away from the brain by systemic administration of cocaine-binding proteins in the periphery. Indeed, the idea has been explored to utilize antibodies with high affinity to cocaine, either passively by infusion of purified monoclonal antibodies (e.g. Wetzel et al., 2016) or by eliciting the in-vivo production of endogenous antibodies through active vaccination (Fox et al., 1996; Martell et al., 2009). That approach suffers from two main drawbacks. First, the affinities to cocaine of the antibodies tested are generally not high enough to compete with the natural targets. Second, the stoichiometric nature of the interactions between antibodies and their cognate determinants necessitate administration of high levels of exogenous protein or elicitation of unrealistically high-titer antibody responses to effectively treat cocaine overdose. Lower cocaine blood levels are associated with low dose recreational use of the drug and antibody therapy may find usefulness under such circumstances (Martell et al., 2009). However, the elicitation of strong humoral responses against cocaine in rodent models involved administration of the antigen at a dose 200-fold higher even than the large dose chosen for the experiments discussed in this article (Fox et al., 1996). Moreover, the necessity of using a very strong adjuvant like Freund’s adjuvant precludes the feasibility of this approach in humans (Fox et al., 1996). Indeed, a recent clinical trial in which volunteers were immunized five times with cholera toxin B subunit chemically attached to the succinylnorcocaine hapten with alum as an adjuvant, demonstrated elicitation of anti-cocaine antibodies at sufficiently high levels in only a third of the subjects and even these humoral responses decayed quickly resulting in minimal effect on cocaine use by the subjects (Martell et al., 2009).

The use of highly efficient catalytic scavengers would solve much of the problems associated with the stoichiometric scavengers described above. While catalytic antibodies were identified, their catalytic efficiencies were vastly inferior to those of natural or engineered hydrolytic enzymes (Wenthur et al., 2017; Zhang et al., 2017). Recent clinical trials have reported the effectiveness of cocaine-hydrolase enzyme therapy to reduce or prevent cocaine overdose (Gorelick, 2012; Shemesh-Darvish et al., 2018; Shram et al., 2015). The A199S/S287G/A328W/Y332G variant, also known as TV-1380, was recently reported to protect monkeys from the cardiac effects of cocaine, and prevented the formation of toxic byproducts that can form when cocaine is taken along with alcohol (Shemesh-Darvish et al., 2018). Despite these advances, translational progress of this promising therapeutic has been limited by the current low-yield, high-cost production system and the inability to quickly scale-up production when new, more efficient enzymes are designed (Zheng et al., 2014; Zheng et al., 2008). The current recombinant biologic expression strategies involve labor-intensive lentiviral-based mammalian cell-based production platforms which are not amenable to rapid generation upon identification of new mutants and require rigorous biosafety measures to be followed, sterile tissue culture conditions and costly media and serum. Here, we report the production of a plant-produced cocaine-hydrolase variant of BChE able to protect and rescue mice from the toxic effects of cocaine, and prevent reinstatement of extinguished drug-seeking behavior in mice that underwent place conditioning with cocaine.

The work presented here relied on our use of a cost-effective, sustainable and safe source of recombinant BChE (Fig. 2). The system utilizes disarmed plant viral vectors to express and accumulate to relatively high levels (hundreds of milligrams to more than a gram of recombinant protein per kg of biomass) of PCocSH and similar variants of BChE (Geyer et al., 2008; Larrimore et al., 2013; Larrimore et al., 2017). Among the advantages of such a transient expression system are reduced production costs, similar or cheaper drug-processing costs, as well as flexibility of production scale to suit production needs (Chen et al., 2015; Mor, 2015; Topp et al., 2016). Moreover, plants constitute a malleable production system that can be engineered to have the recombinant protein product presenting fully sialylated glycans that can support the necessary pharmacokinetic profile of the therapeutic (Schneider et al., 2014a; Schneider et al., 2014b).

Our results provide proof-of-principle that a plant-derived enzyme, PCocSH, can prevent cocaine toxicity when provided prior to cocaine exposure, and also rescue animals from cocaine overdose when given soon after the onset of cocaine-induced convulsions. Animals treated with PCocSH prior to receiving cocaine at doses nearly 20-fold higher than the typical rewarding cocaine dose of 30 mg/kg are protected from advanced stages of acute cocaine toxicity. When given after lethal doses of cocaine, 100% of seizing mice receiving complete doses of PCocSH were saved from subsequent death, compared to 0% of mice which seized and received vehicle control. Further, this plant-derived enzyme can prevent reinstatement of drug seeking behavior in mice. Future studies will aim to establish pharmacokinetic parameters to determine in vivo stability for long term therapeutic applications and examine the effects of PCocSH on cocaine self-administration.