Vol. 41, No. 1 123Biol. Pharm. Bull. 41, 123–131 (2018)

© 2018 The Pharmaceutical Society of Japan

Regular Article

Enantioselective Monoclonal Antibodies for Detecting Ketamine to Crack Down on Illicit UseIzumi Morita,a Hiroyuki Oyama,a Yui Kanda,a Mayumi Yasuo,a Aya Ito,a Masahiro Toyota,b Yoshinori Hayashi,c Takeshi Yokoyama,c and Norihiro Kobayashi*,a

a Kobe Pharmaceutical University; 4–19–1 Motoyama-Kitamachi, Higashinada-ku, Kobe 658–8558, Japan: b Department of Chemistry, Graduate School of Science, Osaka Prefecture University; Sakai, Osaka 599–8531, Japan: and c Department of Dental Anesthesiology, Faculty of Dental Science, Kyushu University; 3–1–1 Maidashi, Higashi-ku, Fukuoka 812–8582, Japan.Received September 20, 2017; accepted October 17, 2017

Ketamine (KT) is a chiral anesthetic agent, (R)- and (S)-enantiomers of which differ in their pharma-cological properties. KT has become one of the most commonly used illicit drugs in the world, thus, rapid and feasible on-site testing is required to crack down on the illicit use. Although immunochemical approach with specific antibodies is promising for this purpose, in practice anti-KT antibodies are difficult to obtain. We here disclose generation of monoclonal antibodies against KT. Mice were immunized with either (a) com-mercially-available or (b) in-house-prepared KT-albumin conjugates. Splenocytes from these mouse groups (a and b) were separately fused with P3/NS1/1-Ag4-1 myeloma cells. After standard screening and cloning, we established 5 hybridoma clones: 2 were derived from group-a mice [generating Ab-KT(a)#2 and #37] and 3 were from group-b mice [generating Ab-KT(b)#9, #13, and #45]. These antibodies exhibited practical per-formance in competitive enzyme-linked immunosorbent assay systems. When (±)-KT·hydrochloride (HCl) was used as the competitor, dose–response curves showed midpoint values of 30 and 70 ng/assay (a-series antibodies) and 2.0–3.0 ng/assay (b-series antibodies). Remarkably, the a-series antibodies were specific for (S)-KT·HCl, while the b-series antibodies were specific for (R)-KT·HCl. Ab-KT(a)#2 (Ka, 7.5×107 M−1) and Ab-KT(b)#45 (Ka, 7.7×108 M−1) exhibited the highest enantioselectivity for each group, and cross-reactivity with the (R)- and (S)-antipodes was 1.3 and 1.7%, respectively. The hybridomas established here are also valuable as a source of genetic information for the anti-KT antibodies, which is required for progressing to next-generation technologies using genetically engineered antibodies.

Key words ketamine; chiral drug; monoclonal antibody; enantioselectivity; enzyme-linked immunosorbent assay (ELISA); on-site testing

Ketamine (KT), a chiral phencyclidine derivative that func-tions as a noncompetitive N-methyl-D-aspartate receptor an-tagonist, has been used as an anesthetic agent in both humans and veterinary medicine.1) However, KT has become one of the most commonly used illicit drugs in the world, and is pop-ular particularly among young populations as a “club drug.”2,3) In Japan, KT is normally supplied as hydrochloride salt of a racemic mixture of (R)-KT and (S)-KT (Fig. 1A) for clinical use, both of which differ in their pharmacological proper-ties. The (S)-enantiomer shows more potent analgesic and anesthetic effects than the (R)-antipode. The (S)-enantiomer is responsible for the psychotomimetic effects whereas the (R)-antipode induces a state of relaxation.1,4,5) Therefore, the (S)-enantiomer is now commercially available in some countries, and its use has increased in recent years.1,5) After administra-tion in humans, KT is mostly biotransformed to norketamine (NKT), the N-demethylated active metabolite, which is then dehydrogenated to generate dehydronorketamine (DNKT) via hydroxylation of the cyclohexanone ring6,7) (Fig. 1B).

In Japan, KT has been controlled under the Narcotics and Psychotropics Control Law since 2007. Rapid, feasible, and reliable on-site testing is required to crack down on illicit KT use and distribution. Immunochemical approaches with specific antibodies are more suitable for this purpose than chromatographic methods because they require neither expen-sive and fixed instruments, nor time-consuming sample-pre-treatment steps. Indeed, several immunoassay kits for testing

KT are now available for purchase, using anti-KT antibodies, which would have been generated in these companies.8–11) Recently, advanced immunosensors for KT have also been reported. These studies involved the use of anti-KT antibodies for immobilization on electrodes12) or quartz crystal microbal-ance chips,13) which might have been donated or purchased from these companies.

However, it has been difficult in practice to obtain anti-KT antibodies when we plan for developing our own analytical methods. Although 1 patent report described the generation of polyclonal anti-KT antibodies against novel haptenic de-rivatives that the applicants synthesized,14) scientific reports describing the production of anti-KT antibodies are scarce. We therefore generated novel monoclonal antibodies against KT by immunizing mice with 2 kinds of haptenized immunogens, either a commercially available or an in-house-prepared KT-albumin conjugate: the use of immunogens of different types should enlarge the probability of obtaining useful antibodies. Interestingly, the resulting antibodies showed remarkable stereoselectivity for either the (R)- or (S)-KT enantiomer, and performed acceptably in enzyme-linked immunosorbent as-says (ELISAs) with practical sensitivity for determining circu-lating racemic KT contents.

MATERIALS AND METHODS

KT and Its Derivatives (±)-KT·hydrochloride (HCl) was

* To whom correspondence should be addressed. e-mail: no-kobay@kobepharma-u.ac.jp

124 Vol. 41, No. 1 (2018)Biol. Pharm. Bull.

supplied by Central Customs Laboratory (Kashiwa, Japan). (R)-(−)-KT·HCl and (S)-(+)-KT·HCl (both with an enantio-meric excess of >99%) were supplied from our laboratory.15,16) Solutions of (±)-NKT·HCl and (±)-DNKT·HCl were obtained from Sigma-Aldrich (St. Louis, MO, U.S.A.). One immuno-genic conjugate with KT linked to bovine serum albumin (BSA), referred to hereafter as KT-BSA(a), was purchased from GenWay Biotech (San Diego, CA, U.S.A.).

Buffers The following buffers were used in this study: PB, 50 mM sodium phosphate buffer (pH 7.3); PBS, PB con-taining 9.0 g/L NaCl; G-PBS, PBS containing 1.0 g/L gelatin; T-PBS, PBS containing 0.050% (v/v) Tween 20; M-PBS, PBS containing 20 g/L skim milk.

In-House Preparation of the KT-BSA(b) Immunogenic Conjugate Another immunogenic conjugate, abbreviated as KT-BSA(b), was prepared as follows.

(a) Synthesis of Haptenic Derivative (±)-KT·HCl (80 mg) was dissolved in 5% Na2CO3 and extracted with CHCl3 to obtain the (±)-KT free base (Fig. 2A). Sodium hydride (55 mg) and 3-bromopropionic acid ethyl ester (150 μL) were added to the (±)-KT in N,N-dimethylformamide (1.5 mL), and the mix-ture was stirred at 60°C for 40 min. The mixture was diluted with water and then extracted with AcOEt. After removing the solvent, the residue was chromatographed on a silica gel (CHCl3/MeOH=15 : 1) to isolate ester (1) as a colorless oil (40.3 mg). [α]D −0.1 (c=0.40, MeOH). 1H-NMR (600 MHz, CD3OD) δ: 7.71 (1H, d, J=7.8 Hz, 3′-H), 7.46 (1H, t, J=7.8 Hz, 4′-H), 7.43 (1H, d, J=7.8 Hz, 6′-H), 7.36 (1H, t, J=7.8 Hz, 5′-H), 4.04 (2H, dd, J=14.4, 7.2 Hz, 10-H), 3.14 (1H, dd, J=14.4, 3.0 Hz, 3α-H), 2.44 (1H, ddt, J=12.8, 7.8, 5.3 Hz, 6-H), 2.26–2.39 (2H, m, 8-H), 2.07–2.11 (1H, m, 5-H), 1.96–2.03 (1H, m, 7-H), 1.97 (3H, s, 2-CH3), 1.75–1.78 (2H, m, 4α,β-H), 1.53–1.60 (2H, m, 3β-H, 7-H), 1.35–1.43 (1H, m, 5-H), 1.18 (3H, t, J=7.2 Hz, 11-H). 13C-NMR (151 MHz, CD3OD) δ: 213.6, 175.2, 136.8, 135.2, 132.5, 132.1, 130.8, 128.3, 72.7, 61.4, 48.4, 40.0, 38.7, 32.4, 28.8, 26.2, 23.0, 14.6. High resolution (HR)-MS (electrospray ionization (ESI)-Orbitrap) m/z: [M+H]+ Calcd for C18H25ClNO3 338.1517. Found 338.1514.

To a solution of ester 1 (40.3 mg) in MeOH (1.0 mL), 5% KOH (0.20 mL) was added and the mixture was stirred at 60°C for 30 min. The mixture was diluted with water and

extracted with CHCl3. The aqueous phase was acidified with HCl, and then run on a column packed with Supelpak-2. After washing with water, the adsorbed substances were eluted with MeOH. Removal of the solvent yielded the acid KT-COOH (2) as a colorless solid (14.3 mg). 1H-NMR (500 MHz, C5D5N) δ: 7.60 (1H, t, J=8.0 Hz, 6′-H), 7.42 (1H, t, J=8.0 Hz, 3′-H), 7.37 (1H, t, J=8.0 Hz, 5′-H), 7.26 (1H, t, J=8.0 Hz, 4′-H), 3.06 (1H, d, J=10.5 Hz, 3-H), 2.55–2.58 (1H, m, 6-H), 2.47–2.51 (1H, m, 7-H), 2.06 (3H, s, 2-CH3), 1.88–1.99 (2H, m, 5-H, 7-H), 1.52–1.63 (3H, m, 3-H, 4-H), 1.33–1.41 (1H, m, 5-H). 13C-NMR (125.5 MHz, C5D5N) δ: 212.5, 138.1, 136.2, 132.0, 131.1, 129.6, 127.5, 72.0, 48.2, 40.2, 37.9, 29.1, 26.7, 22.4. HR-MS (ESI-Orbitrap) m/z: [M+H]+ Calcd for C16H21ClNO3 310.1201. Found 310.1203.

(b) Conjugation of KT-COOH and Albumin KT-COOH was converted to its N-succinimidyl ester (3) according to a reported method.17) Crude ester 3 (6.6 mg, 16 µmol) was react-ed with BSA (Sigma-Aldrich) (15 mg, 0.23 µmol) in a mixture of 1,4-dioxane (1.5 mL), PB (1.5 mL), and pyridine (1.0 mL) with stirring at room temperature for 3 h, then overnight at 4°C. The reaction mixture was treated as described previ-ously (protein precipitation with acetone and subsequent di-alysis),18) and the desired conjugate KT–BSA(b) was obtained as a 1.0-mg/mL solution in saline. The KT:albumin-coupling molar ratio was determined to be 28, based on titration of the residual amino groups with sodium 2,4,6-trinitrobenzene sulfonate.19)

Preparation of Enzyme-Labeled KT β-Galactosidase (GAL)-labeled KT (KT-GAL) was prepared as described previously.20) Briefly, a solution of GAL (EC 3.2.1.23; Roche Diagnostics, Basel, Switzerland) (2.0 mg) in PB (200 µL) was added to the active ester 3 (46 µg) in 1,4-dioxane (200 µL), and the mixture was stirred at 4°C for 4 h. The solution was subjected to a PD-10 column (GE Healthcare Japan, Tokyo, Japan), which was equilibrated with PB–EtOH (4 : 1) and elut-ed with the same buffer. Fractions that showed GAL activity were collected and dialyzed for 2 d against cold PB. The re-sulting solution was adjusted to 45.7 µg/mL (in terms of GAL) in G-PBS and stored at 4°C until use.

Immunization, Cell Fusion, and Monoclonal Anti-body Production All experiments on animals were carried

Fig. 1. Structures of KT Enantiomers (A) and the Major Metabolic Pathway of KT in Humans (B)

Vol. 41, No. 1 (2018) 125Biol. Pharm. Bull.

out in Kobe Pharmaceutical University in accordance with guidelines and regulations established in the university, and all experimental protocols were approved by the Institutional Animal Care and Use Committee.

Female BALB/c and A/J mice (8 weeks of age) were immu-nized with KT-BSA(a) or KT-BSA(b) (25 µg/mouse) biweekly for a total of 4 or 5 times, as summarized in Fig. 3. After screening of the titer for anti-KT antibodies in serum, 1 or 2 mice that showed stronger immune responses were adminis-tered intraperitoneal and intrasplenic injections21) of the con-jugate (25 µg total). After 3 d, splenocytes were collected from the mice and cell fusion22) with P3/NS1/1-Ag4-1 (NS1) my-eloma cells23) was performed according to our standard pro-cedure.18,24,25) Briefly, splenocytes (1–3×108 cells) were fused with 1/5 number of NS1 cells using a 40% polyethylene glycol 4000 solution containing dimethyl sulfoxide and poly-L-argi-nine. Fused cells were cultured for approximately 2 weeks at 37°C in HAT (hypoxanthine-aminopterin-thymidine) medium supplemented with 10% briclone (DS Pharma Biomedical, Osaka, Japan) under 5% CO2/95% air. Culture supernatants were then screened by ELISA, as described below. Hybrid-omas producing anti-KT antibodies were expanded in HT (hypoxanthine-thymidine) medium and then cloned by limit-ing dilution. Finally, the cloned hybridomas were cultured for 7–10 d, and the monoclonal antibodies in the supernatant were characterized. The heavy and light chain isotypes were deter-mined using the ImmunoPure Monoclonal Antibody Isotyping Kit II (Thermo Fisher Scientific, Waltham, MA, U.S.A.).

ELISAs The ELISA procedures used in this study are illustrated in Fig. 4. In both assays described below, the con-centrations of antibodies were adjusted to give bound enzyme activities at B0 (the reactions without standard KT·HCl or its analogs) of approximately 1.0–1.5 absorbance units after a 30-min enzyme reaction.

(a) Characterization of Anti-KT “a-Series” Antibodies De-

rived from Mice Immunized with KT-BSA(a) Costar 96-well microplates (#3590; Corning, Corning, NY, U.S.A.) were coated overnight at room temperature with 1.0 µg/mL KT-BSA(a) in 0.10 M carbonate buffer (pH 8.6) (100 µL/well) and blocked with M-PBS at 37°C for 60 min. After washing the wells 3 times with T-PBS, various concentrations of KT·HCl (or analogs) dissolved in G-PBS (50.0 µL/well) and a monoclo-nal anti-KT antibody (the hybridoma supernatants were used) diluted with G-PBS or an antiserum diluted with PBS con-taining 5.0% BSA (100 µL/well) were added, mixed, and in-cubated at 37°C for 60 min. Subsequently, wells were washed and probed with peroxidase (POD)-labeled goat anti-mouse immunoglobulin G (IgG) (Fc-specific) antibody (160 ng/mL; Jackson ImmunoResearch, West Grove, PA, U.S.A.) diluted in G-PBS (100 µL/well). After incubation at 37°C for 30 min, the wells were washed, and the captured POD activity was deter-mined colorimetrically at 490 nm using o-phenylenediamine as a hydrogen donor.26)

(b) Characterization of Anti-KT “b-Series” Antibodies De-rived from Mice Immunized with KT-BSA(b) Costar micro-plates (#3590) were coated overnight at 4°C with a 5.0-µg/mL solution of affinity-purified goat anti-mouse IgG antibody (Jackson ImmunoResearch) in PBS (100 µL/well). After wash-ing 3 times with PBS, the wells were blocked with M-PBS at 37°C for 60 min. The wells were washed 3 times with T-PBS, and then a monoclonal antibody (the hybridoma supernatants were used) or antiserum against KT diluted with G-PBS (100 µL/well) was added to the wells. After incubation at 37°C for 60 min, the solutions were removed and the wells were washed similarly. Then, 0.50 µg/mL KT-GAL (100 µL/well) and various concentrations of KT·HCl (or analogs) (50.0 µL/well), both diluted in G-PBS, were added, mixed, and incu-bated at 37°C for 60 min. The wells were washed, and the cap-tured GAL activity was determined colorimetrically at 405 nm using o-nitrophenyl β-D-galactopyranoside as the substrate.20)

Fig. 2. Synthesis of the Haptenic Derivative Used to Prepare the KT-BSA (b) ConjugateSynthesis pathway (A) and the HMBC and NOESY correlations for the compound 1 (B). The most stable conformation of compound 1 calculated, viewed from 2 dif-

ferent angles, are also shown (C). A restricted Hartree–Fock SCF calculation was performed at the 3–21G* level using the Spartan program (Spartan’14 for Windows; Wavefunction, Irvine, CA, U.S.A.).

126 Vol. 41, No. 1 (2018)Biol. Pharm. Bull.

Preparation of Fab Fragments and Determining Their Antigen-Binding Parameters Fab fragments of Ab-KT(a)#2 and Ab-KT(b)#45 were prepared with the Pierce Fab Prepara-tion Kit (Thermo Fisher Scientific). Affinity and dissociation rate constants (ka and kd) and equilibrium affinity and dis-sociation constants (Ka and Kd) against KT residues in the “homologous KT-BSA” (the conjugate from which the anti-body was derived) were determined for these Fab fragments at

25°C using BLItz (ForteBio, Fremont, CA, U.S.A.), a bio-layer interferometry (BLI) sensor. Streptavidin-coated biosensor tips, which were saturated with biotin-labeled KT-BSA, pre-pared by reaction with EZ-Link NHS-LC-Biotin (Thermo Fisher Scientific), were dipped into 4.0-µL antibody solutions in G-PBS [10, 20, 50, or 100 nmol/L for Ab-KT(a)#2; 50, 100, 200, or 500 nmol/L for Ab-KT(b)#45]. Association of the anti-bodies with the KT residues was monitored for 300 s, and then

Fig. 4. Schematic Illustrations of the ELISA Systems Using the a-Series (A) and b-Series (B) Monoclonal Antibodies and Antisera

Fig. 3. Schematic Illustrations of the Immunization of Animals, Cell Fusion, and Production of Monoclonal Antibodiesa) Each mouse was immunized with subcutaneous injection of 25 µg KT-BSA at multiple sites on the back with an emulsion of Freund’s complete adjuvant (FCA) or

incomplete adjuvant (FIA) and saline (1 : 1, 0.2 mL). The selected mice were received the final immunization with 25 µg KT-BSA dissolved in saline (see text). b) The #5 mouse that showed the strongest humoral response was dead, thus the splenocytes were collected only from the #1 mouse that showed the second strongest response.

Vol. 41, No. 1 (2018) 127Biol. Pharm. Bull.

dissociation was measured for 300 s in G-PBS.

RESULTS

Production of KT-BSA Immunogenic Conjugates Be-cause of a lack of immunogenicity, KT molecules had to be linked with suitable carrier macromolecules to generate anti-KT antibodies. A previous patent report described the syn-thesis of 4 haptenic derivatives each having a different linker structure for coupling with the carrier.14) Three of them (type A; Fig. 5) had the linker on the nitrogen, and another (type B; Fig. 5) had the linker on the carbon at the position 6 (C6: the α-carbon of the cyclohexanone ring). One of the type-A com-pounds was coupled to BSA and used to immunize sheep. The resulting antisera reacted almost equally to KT and NKT (Fig. 1B): this binding property is reasonably explained based on the nitrogen-masked hapten structure of the conjugate.

For our hapten synthesis, (±)-KT was alkylated with a β-bromoester to introduce a linker structure (Fig. 2A). Protons and carbons of the compound 1 were assigned by 1H–1H cor-relation spectroscopy (COSY) and 1H–13C heteronuclear single quantum coherence (HSQC) and multiple bond coherence (HMBC) (Fig. 2B) experiments. The relative configuration of compound 1 was established as depicted in Fig. 2A on the basis of nuclear Overhauser effect correlated spectroscopy (NOESY) experiments (Fig. 2B). For a better understanding

of the conformation of 1, ab initio (HF/3–21G*) calculations were performed. The most stable conformation of 1 (Fig. 2C) was in good agreement with its spectral data. Interestingly, the ester substituent on C6, which was deduced to a trans configuration with respect to the o-chlorophenyl group, is bending downwards to face the o-chlorophenyl group. Sa-ponification of the ethyl ester 1 provided KT-COOH 2, which was then converted to the N-succinimidyl ester 3: this was obtained as a solid and could be stored stably for years at 4°C. The ester 3 was reacted with BSA and GAL separately, to prepare immunogen [KT-BSA(b)] and enzyme-labeled KT (KT-GAL), respectively.

Generation of Monoclonal Anti-KT Antibodies Outline of the monoclonal antibody production is illustrated in Fig. 3. BALB/c mice, the first strain chosen as the spleen donor during the hybridoma preparation, were immunized with 2 immunogens: the commercially available hapten-carrier con-jugate [KT-BSA(a)] and the in-house-prepared immunogen [KT-BSA(b)] described above. A/J mice were also selected for immunization with the latter immunogen, based on our previ-ous success in preparing hybridomas secreting practical anti-bodies against several small molecules.20,27) The immunization schedule for this study is summarized in Fig. 3. Seven days after the fourth or fifth immunization, blood was collected individually from each mouse, and the titers of the anti-KT antibodies in the sera were examined by the ELISA system, as illustrated in Fig. 4 (see in Materials and Methods as well).

For sera from the mice immunized with KT-BSA(a) (a-series sera), it was necessary to use the same conjugate, i.e., KT-BSA(a), to capture the target antibodies because no in-formation for the chemical structure of haptenic derivative was available. The antigen–antibody binding reaction was performed in assay buffer containing 5.0% BSA to reduce hapten-independent binding due to accompanying anti-BSA antibodies in the sera. The antibody dilution curves are shown in Fig. 6A. Unusually high binding was observed, most of which might have been due to the anti-BSA antibodies that were not completely prevented. However, subsequent inhibi-tion testing by addition of the (±)-KT·HCl standard indicated

Fig. 5. Two Different Types of Haptenic Derivatives Described in a Previous Patent Report14)

“X” refers to functional groups used for conjugation with carrier proteins, and “R” refers to spacers that linked the KT moiety and functional group X.

Fig. 6. Antibody Dilution Curves for the a-Series (A) and b-Series (B) AntiseraThe a- and b-series antisera were assessed by the ELISA (A) and (B), which are shown in Figs. 4A and B, respectively.

128 Vol. 41, No. 1 (2018)Biol. Pharm. Bull.

the presence of anti-KT antibodies. The antigen–antibody binding ability of sera from mice immunized with KT-BSA(b) (b-series sera) was monitored with the homologous haptenic derivative (KT-COOH in Fig. 2A) labeled with β-galactosidase (KT-GAL): in this case, anti-BSA antibodies, if any, should not cause positive signals. The dilution curves showed that all mice generated anti-KT antibodies (Fig. 6B), and inhibi-tion test results suggested that the antibodies might provide practical dose-dependent responses against (±)-KT·HCl (data not shown).

Based on these data, mice were selected as spleen donors (Fig. 3) and their splenocytes were fused with NS1 myeloma cells.22,23) Three fusion experiments (Fig. 3) afforded 2 anti-body-secreting hybridoma clones (#2 and #37) derived from mice immunized with KT-BSA(a) and 3 clones (#9, #13, and #45) derived from mice immunized with KT-BSA(b). The “a-series antibodies” from the former 2 hybridoma clones, designated Ab-KT(a)#2 and Ab-KT(a)#37, were composed of γ1 heavy chains and κ light chains. The “b-series anti-bodies” from the latter 3 clones, designated Ab-KT(b)#9, Ab-KT(b)#13, and Ab-KT(b)#45, were composed of γ1 heavy chains and λ light chains. We note that these 5 antibodies had different amino acid sequences both for the heavy and light chain variable domains (data not shown).

Characterization of the Monoclonal Anti-KT Antibodies by ELISA

(a) Sensitivity Although the (S)-enantiomer is available in some countries,1) racemic mixtures of KT·HCl are supposed to be still circulating for illicit use. Therefore, ELISAs were first performed using (±)-KT·HCl as the analyte with our monoclonal antibodies. The assay systems using the a-series or b-series antibodies are shown in Fig. 4. Dose–response curves are shown in Fig. 7. The midpoint, which is the dose of (±)-KT·HCl that inhibits antibody binding to the immobilized- or enzyme-labeled KT by 50% (the same meaning as IC50: a useful index of assay sensitivity), was 30 and 70 ng/assay for the a-series antibodies [Ab-KT(a)#2 and Ab-KT(a)#37, respec-tively] (Fig. 7A). The b-series antibodies [Ab-KT(b)#9, Ab-

KT(b)#13, and Ab-KT(b)#45] generated more sensitive dose–response curves with much lower midpoint values (3.0, 2.0, and 2.1 ng/assay, respectively) (Fig. 7B). The limit of detection (LOD) was determined for the most effective antibodies in each series [i.e., Ab-KT(a)#2 and Ab-KT(b)#45], selected based on the sensitivity and enantioselectivity (see below). The LODs for Ab-KT(a)#2 and Ab-KT(b)#45 were 1.5 and 0.21 ng/assay, based on the amount of (±)-KT·HCl required to give a bound absorbance 2 standard deviations below the aver-age (n=10) of the B0 absorbance. The unit “X g/assay” used in this study means that a total of X g (mass) of analyte or cross-reactive analogs was added to the assay chamber (microwells) for the competitive antigen–antibody reactions.

(b) Specificity Enantioselectivity of these antibodies was then examined using (R)-KT·HCl and (S)-KT·HCl standards, and evaluated as the cross-reactivity in ELISAs, where the reactivity of (±)-KT·HCl was taken as 100% (Fig. 8). The a-series antibodies bound preferentially to the (S)-enantiomer, as shown by cross-reactivity that exceeded 100%. Remarkably, Ab-KT(a)#2 showed ca. 200% cross-reactivity: when reactivi-ty to the (S)-enantiomer was set at 100%, cross-reactivity with the (R)-antipode was 1.3%. In contrast, the b-series antibodies were specific for the (R)-enantiomer as shown by >160% re-activity. The most (R)-specific antibody, Ab-KT(b)#45, showed 1.7% cross-reactivity with the (S)-antipode when the reactivity of the (R)-enantiomer was normalized to 100%.

The a-series antibodies significantly cross-reacted with NKT·HCl (ca. 30%), but showed almost negligible values for DNKT·HCl (<4%). The b-series antibodies, Ab-KT(b)#9, #13, and #45, showed rather fluctuating but significant cross-reac-tivity with NKT·HCl (35, 25, and 7.0%, respectively), while <4% cross-reactivity was observed with DNKT·HCl. Cross-reactivity with some compounds that are structurally unre-lated to KT was examined for Ab-KT(a)#2 and Ab-KT(b)#45. Both antibodies showed negligible cross-reactivity towards p-acetamidophenol, acetylsalicylic acid, creatine, creatinine, caffeine, and uric acid (<1 and <0.1%, respectively).

(c) Binding Parameters The association and dissocia-

Fig. 7. Dose–Response Curves of ELISAsThe a-series monoclonal antibodies were used for measuring (±)-KT·HCl and (S)-KT·HCl (A), and the b-series monoclonal antibodies were used for measuring

(±)-KT·HCl and (R)-KT·HCl (B). The vertical bars indicate the standard deviation (n=4).

Vol. 41, No. 1 (2018) 129Biol. Pharm. Bull.

tion rate constants (ka and kd, respectively) and the equilib-rium affinity and dissociation constants (Ka and Kd, re-spectively) of Ab-KT(a)#2 and Ab-KT(b)#45 (examined as the Fab fragment) against the KT-residues were as follows: Ab-KT(a)#2, ka=3.0×105 M−1 s−1, kd=4.0×10−3 s−1, Ka=7.5× 107 M−1, Kd=1.3×10−8 M; Ab-KT(b)#45, ka=1.1×106 M−1 s−1, kd=1.5×10−3 s−1, Ka=7.7×108 M−1, Kd=1.3×10−9 M. Both anti-bodies showed the practical nanomolar-order affinity. Ab-KT(b)#45 exhibited a 10-fold greater Ka value than Ab-KT(a)#2, consistent with the fact that this antibody showed more sensitive dose–response curves in the ELISA (Fig. 7). We observed extremely high ka values (>105 M−1 s−1), which are within the maximum range of ka values deduced for an-tigen–antibody reactions based on the diffusion coefficients of the reactant molecules.28) This property should allow quick binding against immobilized antigens, which is a great ad-vantage for on-site analytical systems involving immunochro-matographies or immunosensors.

DISCUSSION

Production of practical hybridoma-based monoclonal antibodies targeting small molecules (i.e., haptens) is not straightforward and requires considerable expertise. To gen-erate specific antibodies, suitable haptenized immunogens should be used with characteristic functional groups on the target hapten externally exposed, being adequately distant from the carrier macromolecules.20) In addition, using haptens with smaller molecular mass usually increases the difficulty in generating high-affinity antibodies,29,30) although some exceptional instances have been reported: e.g., monoclonal antibodies against 3-methylindole (Mr 131.2) with Ka values of >108 M−1 were produced previously.31) This difficulty mainly arises from the fact that such small compounds (Mr<300, as

a rough indication), linked with carrier macromolecules, often generate antibodies that recognize epitopes composed of both hapten residues and linker structures bridging the hapten and carrier (and sometimes, even part of the carrier structure). Consequently, such antibodies cannot tightly capture the un-modified hapten molecule alone. It is therefore not surprising that monoclonal antibodies against the small KT molecule (Mr 237.7) are not readily available.

Initially, when exploring methods to synthesize KT deriva-tives for linking with carriers, we attempted the reaction of (±)-KT with O-(carboxymethyl) hydroxylamine under standard conditions, but failed to obtain the desired oxime. Then, we performed a reaction with ethyl 3-bromopropionate, which mainly substituted on the α-carbon of the carbonyl group (compound 1; Fig. 2A), not on the nitrogen in the methylamino group. Thus, the haptenic derivative KT-COOH 2, which was obtained after subsequent saponification, was the one catego-rized as a type-B hapten (Fig. 5). Subsequent saponification of compound 1 provided the haptenic derivative 2 (KT-COOH).

Mice were immunized with 2 immunogens: the commer-cially available hapten-carrier conjugate [KT-BSA(a)], and the in-house-prepared conjugate linking KT-COOH and BSA [KT-BSA(b)]. Both conjugates successfully induced desirable humoral immune responses with a standard immunization schedule using Freund’s adjuvant, yielding hybridoma clones secreting anti-KT antibodies. Three functional groups contain-ing heteroatoms (carbonyl, methylamino, and o-chlorophenyl groups) and 2 rigid ring structures (phenyl and cyclohexanone rings) are present in KT. Despite its rather small molecular size, these features might have allowed for stable complex for-mation with the surface receptors of relevant B-cells.

The a-series antibodies, Ab-KT(a)#2 and Ab-KT(a)#37 derived from KT-BSA(a), were specific to the (S)-enantiomer of KT·HCl, while the b-series antibodies, Ab-KT(b)#9, Ab-KT(b)#13, and Ab-KT(b)#45 derived from KT-BSA(b), pref-erably bound to the (R)-enantiomer. As is well recognized, specificity of anti-hapten antibodies is closely related with the structure of the hapten-carrier conjugate used for immuniza-tion. Unfortunately, the chemical structure of the commer-cially-available KT-BSA(a) is not publicly available. However, such distinct differences in specificity of the resulting anti-bodies suggested that the KT derivative used to prepare KT-BSA(a) was structurally rather different from KT-COOH used to prepare KT-BSA(b). Because KT-COOH is racemic, both the (R)- and (S)-KT residues should have been linked to BSA. In the initial event of the antibody production in vivo, indi-vidual B-cells should recognize a single hapten residue, i.e., either the (R)- or (S)-KT residue on the carrier molecule. For KT-BSA(b), the (R)-KT redidue might have exhibited stronger immunogenicity, resulting in the preferential proliferation of plasma cells secreting (R)-KT-specific antibodies.

The a-series and b-series antibodies discriminated NKT·HCl somewhat comparably (ca. 30% for the a-series and 7.0–35% for the b-series, setting values obtained with (±)-KT·HCl at 100%). These findings indicated that the im-munogen KT-BSA(a) employed a haptenic derivative that was different from the type A compounds (i.e., N-substituted derivatives; Fig. 5), considering the specificity of antibodies previously generated from a type A haptenic derivative (men-tioned above).14) The KT molecule has a limited number of positions where substituents can be feasibly introduced. The

Fig. 8. Cross-Reactivity of the Monoclonal Antibodies with KT-Related Compounds

Cross-reactivity was determined by the 50% displacement method40) according to the following equation: Cross reactivity (%)=(X/Y)×100; where X is the mid-point (ng/assay) of (±)-KT·HCl, and Y is the midpoint (ng/assay) of a KT analog, cross-reactivity of which is to be determined.

130 Vol. 41, No. 1 (2018)Biol. Pharm. Bull.

haptenic derivative linked in the KT-BSA(a) conjugate, which induced (S)-KT-specific antibodies, might have been a deriva-tive of (S)-KT that have a linker structure on α-carbon of the cyclohexanone like our KT-COOH. If this is true, then, the re-sults for the a-series antibodies shown above indicate that the (S)-KT residue is not suitable for inducing high-affinity anti-bodies: this possibility is compatible with the model proposed above to explain why KT-BSA(b) induced (R)-KT-specific antibodies. Such distinct differences in immunogenic proper-ties between a pair of enantiomers have been recognized for several chiral drugs as reviewed previously.32)

In the competitive ELISA, the best antibody in the a-series, Ab-KT(a)#2, facilitated antigen measurements at ca. 10–1000 ng/assay for (±)-KT·HCl and ca. 5.0–500 ng/assay for (S)-KT·HCl (Fig. 7A). The best antibody in the b-series, Ab-KT(b)#45, enabled more sensitive measurements with the range of ca. 0.20–50 ng/assay for (±)-KT·HCl, and ca. 0.010–50 ng/assay for (R)-KT·HCl (although different ELISA protocols were used for the a-series and b-series antibodies) (Fig. 7B). Ab-KT(b)#45 also exhibited dose–response curves for (S)-KT·HCl, due to cross-reaction, covering the range of ca. 5.0–1000 ng/assay (not shown in the Figure), which was similar to that obtained with Ab-KT(a)#2. Presently, KT for illegal drug use is supposed to be distributed as (±)-KT·HCl or (S)-KT·HCl. Considering the measurable ranges shown above, Ab-KT(b)#45 is the most useful antibody among the 5 antibodies developed here for on-site analysis aimed cracking down on illicit KT use. The combined use of Ab-KT(a)#2 and Ab-KT(b)#45 might enable separate monitoring of (R)- and (S)-KT disposition after administration of the racemic KT mixture in clinical and preclinical studies. Very recently, (R)-KT has been shown to exert longer-lasting antidepressant effects than (S)-KT in animal models of depression.33) The (R)-enantiomer might be promising as a new antidepressant drug, and the (R)-specific Ab-KT(b)#45 might help the devel-opment researches.

CONCLUSION

Here, we provide the first description of producing mono-clonal anti-KT antibodies with practical binding properties. The present antibodies will be useful for constructing various immunochemical systems suitable for on-site analyses in-volving immunochromatographies and immunosensors. The hybridoma clones that we established are also valuable as a source of genetic information for these antibodies, which is necessary for extending next-generation technologies using genetically engineered antibodies.34–36) The genes encoding the variable domains of the antibodies referred above have already been cloned, and relevant single-chain Fv fragments (scFvs) that with KT-binding ability have been prepared in other laboratories (unpublished data). In vitro evolution on these scFvs might create improved antibodies with higher af-finity enabling more sensitive assays.37–40)

Acknowledgments We thank Drs. Atsuko Takeuchi and Chisato Tode (Kobe Pharmaceutical University) for determin-ing MS and NMR spectra. (±)-KT·HCl was supplied from Central Customs Laboratory (Kashiwa, Japan). This work was supported in part by Ushio Inc. (Chiyoda-ku, Tokyo, Japan).

Conflict of Interest The authors declare no conflict of interest.

REFERENCES

1) Mion G, Villevieille T. Ketamine pharmacology: an update (phar-macodynamics and molecular aspects, recent findings). CNS Neuro-sci. Ther., 19, 370–380 (2013).

2) Liu Y, Lin D, Wu B, Zhou W. Ketamine abuse potential and use disorder. Brain Res. Bull., 126, 68–73 (2016).

3) Han E, Kwon NJ, Feng LY, Li JH, Chung H. Illegal use patterns, side effects, and analytical methods of ketamine. Forensic. Sci. Int., 268, 25–34 (2016).

4) White PF, Ham J, Way WL, Trevor AJ. Pharmacology of ketamine isomers in surgical patients. Anesthesiology, 52, 231–239 (1980).

5) WHO. Critical review of Ketamine. 34th Expert Comm. Drug De-pend., 1–30 (2006).

6) Desta Z, Moaddel R, Ogburn ET, Xu C, Ramamoorthy A, Venkata SLV, Sanghvi M, Goldberg ME, Torjman MC, Wainer IW. Stere-oselective and regiospecific hydroxylation of ketamine and norket-amine. Xenobiotica, 42, 1076–1087 (2012).

7) Dinis-Oliveira RJ. Metabolism and metabolomics of ketamine: a toxicological approach. Forensic Sci. Res. J., 2, 2–10 (2017).

8) Huang MH, Wu MY, Wu CH, Tsai JL, Lee HH, Liu RH. Perfor-mance characteristics of ELISAs for monitoring ketamine exposure. Clin. Chim. Acta, 379, 59–65 (2007).

9) “Teste Ketamine | Toxicologia | Neogen.”: ‹http://toxicology.neogen.com/pt/ketamine-forensic›

10) “Ketamine Drug Test Kit, Test for Ketamine-RapidTest.”: ‹http://www.rapidtest.com/index.php?i=Drug-Tests&id=643&cat=24›

11) “ELISA for Forensic Matrices | Immunalysis.”: ‹https://immunalysis.com/products/elisa/elisa-for-forensic/›

12) Li Q, Tang W, Wang Y, Di J, Yang J, Wu Y. Electrochemilumi-nescence immunosensor for ketamine detection based on poly-amidoamine-coated carbon dot film. J. Solid State Electrochem., 19, 2973–2980 (2015).

13) Yang Y, Tu Y, Wang X, Pan J, Ding Y. A label-free immunosensor for ultrasensitive detection of ketamine based on quartz crystal mi-crobalance. Sensors, 15, 8540–8549 (2015).

14) McConnell RI, Benchikh E, Fitzgerald SP, Lamont JV. Haptens, immunogens, antibodies and conjugates to ketamine and its metabo-lites. U.S. Patent 20030224447 (2003).

15) Yokoyama R, Matsumoto S, Nomura S, Higaki T, Yokoyama T, Ki-yooka S. Enantioselective construction of nitrogen-substituted qua-ternary carbon centers adjacent to the carbonyl group in the cyclo-hexane ring: first asymmetric synthesis of anesthetic (S)-ketamine with high selectivity. Tetrahedron, 65, 5181–5191 (2009).

16) Yokoyama T, Yokoyama R, Nomura S, Matsumoto S, Fujiyama R, Kiyooka S. Synthesis of (S)-ketamine via [1,3]-chirality transfer of a stereocenter created by enantioselective aldol reaction. Bull. Chem. Soc. Jpn., 82, 1528–1532 (2009).

17) Hosoda H, Sakai Y, Yoshida H, Miyairi S, Ishii K, Nambara T. The preparation of steroid N-hydroxysuccinimide esters and their reactivities with bovine serum albumin. Chem. Pharm. Bull., 27, 742–746 (1979).

18) Kobayashi N, Kato Y, Oyama H, Taga S, Niwa T, Sun P, Ohtoyo M, Goto J. Anti-estradiol-17β single-chain Fv fragments: generation, characterization, gene randomization, and optimized phage display. Steroids, 73, 1485–1499 (2008).

19) Habeeb AF. Determination of free amino groups in proteins by trinitrobenzenesulfonic acid. Anal. Biochem., 14, 328–336 (1966).

20) Kobayashi N, Imazu T, Ebisawa A, Shimada K. Production and characterization of monoclonal antibodies against a novel 1α,25-dihydroxyvitamin D3-bovine serum albumin conjugate linked through the 11α-position. J. Steroid Biochem. Mol. Biol., 63, 127–137 (1997).

Vol. 41, No. 1 (2018) 131Biol. Pharm. Bull.

21) Kussie PH, Albright G, Linthicum DS. Production and characteriza-tion of monoclonal idiotypes and anti-idiotypes for small ligands. Methods Enzymol., 178, 49–63 (1989).

22) Köhler G, Milstein C. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature, 256, 495–497 (1975).

23) Köhler G, Howe SC, Milstein C. Fusion between immunoglobulin-secreting and nonsecreting myeloma cell lines. Eur. J. Immunol., 6, 292–295 (1976).

24) Kobayashi N, Banzono E, Shimoda Y, Oyama H, Kunihiro T, Morita I, Ohta M. A monoclonal antibody-based enzyme-linked im-munosorbent assay for human urinary cotinine to monitor tobacco smoke exposure. Anal. Methods, 3, 1995–2002 (2011).

25) Oyama H, Morita I, Kiguchi Y, Miyake S, Moriuchi A, Akisada T, Niwa T, Kobayashi N. Gaussia luciferase as a genetic fusion partner with antibody fragments for sensitive immunoassay monitoring of clinical biomarkers. Anal. Chem., 87, 12387–12395 (2015).

26) Kobayashi N, Katsumata H, Katayama H, Oiwa H, Goto J, Take-uchi Y. A monoclonal antibody-based enzyme-linked immunosor-bent assay of ursodeoxycholic acid 3-sulfates in human urine. J. Steroid Biochem. Mol. Biol., 72, 265–272 (2000).

27) Kobayashi N, Oiwa H, Goto J. Production and characterization of group-specific monoclonal antibodies recognizing nonamidated, glycine- and taurine-amidated ursodeoxycholic acid 7-N-acetylglu-cosaminides. J. Steroid Biochem. Mol. Biol., 64, 171–177 (1998).

28) Foote J, Eisen HN. Kinetic and affinity limits on antibodies pro-duced during immune responses. Proc. Natl. Acad. Sci. U.S.A., 92, 1254–1256 (1995).

29) Chappey ON, Sandouk P, Scherrmann JM. Monoclonal antibodies in hapten immunoassays. Pharm. Res., 9, 1375–1379 (1992).

30) Chappey O, Debray M, Niel E, Scherrmann JM. Association con-stants of monoclonal antibodies for hapten: heterogeneity of fre-quency distribution and possible relationship with hapten molecular weight. J. Immunol. Methods, 172, 219–225 (1994).

31) Tuomola M, Harpio R, Mikola H, Knuuttila P, Lindström M, Mukkala VM, Matikainen MT, Lövgren T. Production and char-

acterisation of monoclonal antibodies against a very small hapten, 3-methylindole. J. Immunol. Methods, 240, 111–124 (2000).

32) Got PA, Scherrmann JM. Stereoselectivity of antibodies for the bio-analysis of chiral drugs. Pharm. Res., 14, 1516–1523 (1997).

33) Fukumoto K, Toki H, Iijima M, Hashihayata T, Yamaguchi J, Hashimoto K, Chaki S. Antidepressant potential of (R)-ketamine in rodent models: comparison with (S)-ketamine. J. Pharmacol. Exp. Ther., 361, 9–16 (2017).

34) Handbook of therapeutic antibodies. (Dübel S ed.) Wiley-Black-well, Hoboken, New Jersey (2010).

35) Kobayashi N, Oyama H. Antibody engineering toward high-sensi-tivity high-throughput immunosensing of small molecules. Analyst, 136, 642–651 (2011).

36) Abe R, Ohashi H, Iijima I, Ihara M, Takagi H, Hohsaka T, Ueda H. “Quenchbodies”: Quench-based antibody probes that show antigen-dependent fluorescence. J. Am. Chem. Soc., 133, 17386–17394 (2011).

37) Kobayashi N, Oyama H, Kato Y, Goto J, Söderlind E, Borrebaeck CAK. Two-step in vitro antibody affinity maturation enables estradiol-17β assays with more than 10-fold higher sensitivity. Anal. Chem., 82, 1027–1038 (2010).

38) Oyama H, Yamaguchi S, Nakata S, Niwa T, Kobayashi N. “Breed-ing” diagnostic antibodies for higher assay performance: a 250-fold affinity-matured antibody mutant targeting a small biomarker. Anal. Chem., 85, 4930–4937 (2013).

39) Oyama H, Morita I, Kiguchi Y, Banzono E, Ishii K, Kubo S, Watanabe Y, Hirai A, Kaede C, Ohta M, Kobayashi N. One-shot in vitro evolution generated an antibody fragment for testing urinary cotinine with more than 40-fold enhanced affinity. Anal. Chem., 89, 988–995 (2017).

40) Morita I, Oyama H, Yasuo M, Matsuda K, Katagi K, Ito A, Tatsuda H, Tanaka H, Morimoto S, Kobayashi N. Antibody fragments for on-site testing of cannabinoids generated via in vitro affinity matu-ration. Biol. Pharm. Bull., 40, 174–181 (2017).