Design, Synthesis, and In Vivo Evaluation of C1-Linked 4,5-Epoxymorphinan Haptens for Heroin Vaccines

In our continuing effort to develop effective anti-heroin vaccines as potential medications for the treatment of opioid use disorder, herein we present the design and synthesis of the haptens: 1-AmidoMorHap (1), 1-AmidoMorHap epimer (2), 1 Amido-DihydroMorHap (3), and 1 Amido-DihydroMorHap epimer (4). This is the first report of hydrolytically stable haptenic surrogates of heroin with the attachment site at the C1 position in the 4,5-epoxymorophinan nucleus. We prepared respective tetanus toxoid (TT)–hapten conjugates as heroin vaccine immunogens and evaluated their efficacy in vivo. We showed that all TT–hapten conjugates induced high antibody endpoint titers against the targets but only haptens 2 and 3 can induce protective effects against heroin in vivo. The epimeric analogues of these haptens, 1 and 4, failed to protect mice from the effects of heroin. We also showed that the in vivo efficacy is consistent with the results of the in vitro drug sequestration assay. Attachment of the linker at the C1 position induced antibodies with weak binding to the target drugs. Only TT-2 and TT-3 yielded antibodies that bound heroin and 6-acetyl morphine. None of the TT–hapten conjugates induced antibodies that cross-reacted with morphine, methadone, naloxone, or naltrexone, and only TT-3 interacted weakly with buprenorphine, and that subtle structural difference, especially at the C6 position, can vastly alter the specificity of the induced antibodies. This study is an important contribution in the field of vaccine development against small-molecule targets, providing proof that the chirality at C6 in these epoxymorphinans is a vital key to their effectiveness.


Introduction
Approximately 130,000 Americans died from drug overdose due to heroin from 1999 to 2019, an average of more than four deaths for every 100,000 people [1]. The combined
With 6 in hand, we sought to introduce an amino group at the C1 position via nitration on an activated C1, followed by reduction of the nitro group. Many nitration conditions were attempted. Neither nitronium tetrafluoroborate in DMSO at room temperature nor sodium nitrate in acetic acid gave any nitration, while sodium nitrite in TFA decomposed the substrate 6. A classical HNO3/H2SO4 in nitromethane at −20 °C gave a mixture of the desired product 7, phenol 8, and over-nitrated product 9, likely the result of nitration of phenol 8 after loss of the hydrolytically unstable acetate (Scheme 2). Scheme 2. Over-nitration of intermediate 6.
In an effort to avoid over-nitration, we opted to alter our C3 protection strategy (Scheme 3). Reasoning that the C2 position could be rendered sterically inaccessible using a bulky C3 protecting group, acetate 6 was hydrolyzed to free phenol 10 through basic methanolysis and re-protected as a TBS ether (11). Nitration with bismuth nitrate gave the desired mono-nitrated compound 12 in moderate yield. After nitration, the TBS group was removed with TBAF to give phenol 8 at a yield of 92%, a 44% overall yield from phenol 10. With 6 in hand, we sought to introduce an amino group at the C1 position via nitration on an activated C1, followed by reduction of the nitro group. Many nitration conditions were attempted. Neither nitronium tetrafluoroborate in DMSO at room temperature nor sodium nitrate in acetic acid gave any nitration, while sodium nitrite in TFA decomposed the substrate 6. A classical HNO 3 /H 2 SO 4 in nitromethane at −20 • C gave a mixture of the desired product 7, phenol 8, and over-nitrated product 9, likely the result of nitration of phenol 8 after loss of the hydrolytically unstable acetate (Scheme 2).
With 6 in hand, we sought to introduce an amino group at the C1 position via nitration on an activated C1, followed by reduction of the nitro group. Many nitration conditions were attempted. Neither nitronium tetrafluoroborate in DMSO at room temperature nor sodium nitrate in acetic acid gave any nitration, while sodium nitrite in TFA decomposed the substrate 6. A classical HNO3/H2SO4 in nitromethane at −20 °C gave a mixture of the desired product 7, phenol 8, and over-nitrated product 9, likely the result of nitration of phenol 8 after loss of the hydrolytically unstable acetate (Scheme 2). Scheme 2. Over-nitration of intermediate 6.
In an effort to avoid over-nitration, we opted to alter our C3 protection strategy (Scheme 3). Reasoning that the C2 position could be rendered sterically inaccessible using a bulky C3 protecting group, acetate 6 was hydrolyzed to free phenol 10 through basic methanolysis and re-protected as a TBS ether (11). Nitration with bismuth nitrate gave the desired mono-nitrated compound 12 in moderate yield. After nitration, the TBS group was removed with TBAF to give phenol 8 at a yield of 92%, a 44% overall yield from phenol 10. In an effort to avoid over-nitration, we opted to alter our C3 protection strategy (Scheme 3). Reasoning that the C2 position could be rendered sterically inaccessible using a bulky C3 protecting group, acetate 6 was hydrolyzed to free phenol 10 through basic methanolysis and re-protected as a TBS ether (11). Nitration with bismuth nitrate gave the desired mono-nitrated compound 12 in moderate yield. After nitration, the TBS group was removed with TBAF to give phenol 8 at a yield of 92%, a 44% overall yield from phenol 10.
While effective, we thought that a more stable protective group at the C3 position would aid the nitration step, and that codeine, rather than morphine, would be a better starting material. We found that nitration with methyl ether gave a much better yield than that of acetate and TBS ether-protected substrate. Compound 15 was synthesized at a yield of 74% over 3 steps following the procedure used to prepare 6 (Scheme 4) [23].
With compound 15 as the substrate, the nitration using the nitric acid and sulfuric acid mixture in nitromethane at −20 to −10 • C still gave an over-nitrated product, while nitration with bismuth nitrate yielded the desired nitro compound 16 at a yield of 88% without over-nitration. The absolute configuration of 16 was confirmed by X-ray crystallographic analysis ( Figure 2). The 1-nitro compound 16 was reduced with formamidinesulfinic acid (FSA) in basic solution to afford aniline 17. The methyl ether was cleaved with BBr 3 and the resulting hydrobromide salt was used directly without further purification due to its instability in solution. The resulting aminophenol was coupled with 3-(tritylthio)propanoic acid using O-(benzotriazol-1-yl)-N,N,N ,N -tetramethyluronium tetrafluoroborate (TBTU) as the coupling reagent. Finally, the ester resulting from coupling  Figure 2. X-ray crystallographic structure determination of the absolute configuration of 16. Thermal ellipsoids in the plot are at the 50% probability level.
With compound 15 as the substrate, the nitration using the nitric acid and sulfuric acid mixture in nitromethane at −20 to −10 °C still gave an over-nitrated product, while nitration with bismuth nitrate yielded the desired nitro compound 16 at a yield of 88% without over-nitration. The absolute configuration of 16 was confirmed by X-ray crystallographic analysis ( Figure 2). The 1-nitro compound 16 was reduced with formamidinesulfinic acid (FSA) in basic solution to afford aniline 17. The methyl ether was cleaved with BBr3 and the resulting hydrobromide salt was used directly without further purification due to its instability in solution. The resulting aminophenol was coupled with 3-(tritylthio)propanoic acid using O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (TBTU) as the coupling reagent. Finally, the ester resulting from coupling with excess 3-(tritylthio)propanoic acid was hydrolyzed with NH4OH to give the desired hapten 1 at a yield of 31% (Scheme 4).
Following an analogous procedure to the synthesis of hapten 1, nitration of 20 with bismuth nitrate hydrate gave 21 in modest yield (Scheme 5). Utilizing a similar reduction/deprotection sequence that was used in the synthesis of 1 afforded a highly unstable amino alcohol that degraded upon standing and during purification. Attempts to take the crude alcohol through the acylation/hydrolysis sequence afforded the desired hapten 2, but the sequence was inconsistent, and the final hapten was extensively contaminated. Hypothesizing that the issue was oxidative instability and difficulty purifying the highly polar amino phenol, we sought to avoid having a free amino phenol present during the course of the synthesis, which necessitated replacing the methyl ether with a protecting group more amenable to late-stage removal. Cleavage of the C3-methyl (BBr 3 , −78-25 • C) followed by protection of the phenol as the tert-butyldiphenyl silyl ether gave 23 in good yield. Unfortunately, the silyl ether was labile under the strongly basic conditions of the FSA reduction. Switching to a tin chloride-mediated sodium borohydride reduction gave 24 followed by a TBTU coupling of the crude amine with the propionic acid linker affording amide 25 in moderate yield over two steps. Removal of the silyl group with tetrabutylammonium fluoride gave hapten 2 at a yield of 56%, a 6% overall yield from 18. Following an analogous procedure to the synthesis of hapten 1, nitration of 20 with bismuth nitrate hydrate gave 21 in modest yield (Scheme 5). Utilizing a similar reduction/deprotection sequence that was used in the synthesis of 1 afforded a highly unstable amino alcohol that degraded upon standing and during purification. Attempts to take the crude alcohol through the acylation/hydrolysis sequence afforded the desired hapten 2, but the sequence was inconsistent, and the final hapten was extensively contaminated. Hypothesizing that the issue was oxidative instability and difficulty purifying the highly polar amino phenol, we sought to avoid having a free amino phenol present during the course of the synthesis, which necessitated replacing the methyl ether with a protecting group more amenable to late-stage removal. Cleavage of the C3-methyl (BBr3, −78-25 °C) followed by protection of the phenol as the tert-butyldiphenyl silyl ether gave 23 in good yield. Unfortunately, the silyl ether was labile under the strongly basic conditions of the FSA reduction. Switching to a tin chloride-mediated sodium borohydride reduction gave 24 followed by a TBTU coupling of the crude amine with the propionic acid linker affording amide 25 in moderate yield over two steps. Removal of the silyl group with tetrabutylammonium fluoride gave hapten 2 at a yield of 56%, a 6% overall yield from 18.
(tritylthio)propanamide, 1-AmidoDihydroMorHap, Figure 1) was synthesized from hydromorphone (Scheme 6). Reductive amination of hydromorphone with benzylamine gave secondary amine 26 in good yield. Hydrogenolysis of the benzyl group gave an intermediate (27) which underwent acylation to intermediate 28. Basic hydrolysis of the aryl acetate 28 gave phenol 29 in 70% over three steps. The phenol was protected as the TBS ether (30) and nitrated under acidic conditions (NaNO2/TFA) to give the C1-nitro compound 31 (Scheme 6). Removal of the phenolic protecting group gave 32, and the nitro moiety was reduced to the unstable amino compound 33 which was carried through a coupling/hydrolysis protocol to give hapten 3 in approximately 30% overall yield from hydromorphone. Removal of the phenolic protecting group gave 32, and the nitro moiety was reduced to the unstable amino compound 33 which was carried through a coupling/hydrolysis protocol to give hapten 3 in approximately 30% overall yield from hydromorphone.

Scheme 7. Synthesis of hapten 4 (1-AmidoDihydroMorHap epimer).
Compound 16 was subjected to BBr3 demethylation of the C3-methoxy moiety to give 8, followed by formation of 34 by reduction of both the C-ring olefin and the C1-nitro to an amine using Pd/C. Coupling of the linker 3-(tritythio)propanoic acid and TBTU to the amine gave the desired compound hapten 4 at a yield of 36% over 3 steps from 16.

Evaluation of TT-Hapten Conjugates In Vivo
The haptens were separately conjugated to tetanus toxoid (TT) carrier protein using the previously optimized method [18,23,24,27]. Vaccines were adsorbed to aluminum hydroxide and mixed with Army Liposome Formulation (ALF43) adsorbed to aluminum hydroxide as an adjuvant. Immunogenicity was assessed by immunizing mice and collecting sera at weeks 0, 3, and 6. Antibodies to the immunizing haptens were measured using ELISA that used BSA-hapten conjugates as coating agents. Results showed that all of the haptens induced high antibody endpoint titers (>10 5 after the second vaccine dose) against their respective antigens ( Figure 3b). This suggests that all of the vaccine candidates tested are immunogenic.
Multiple reports have suggested that antibody endpoint titers may not predict efficacy of a vaccine against small molecules such as drugs of abuse [5,23]. To this end, we assessed the efficacy of each vaccine candidate by in vivo hot plate assay. This assay measures the time it takes for the mouse to respond to a pain stimulus (heat). This has been used in the past to evaluate vaccines to drugs of abuse [27,28]. Our results showed that among the conjugates tested, only TT-2 and TT-3 gave low % MPE values, suggesting that mice were protected from the antinociceptive effects of heroin ( Figure 3c). The other Scheme 7. Synthesis of hapten 4 (1-AmidoDihydroMorHap epimer).
Compound 16 was subjected to BBr 3 demethylation of the C3-methoxy moiety to give 8, followed by formation of 34 by reduction of both the C-ring olefin and the C1-nitro to an amine using Pd/C. Coupling of the linker 3-(tritythio)propanoic acid and TBTU to the amine gave the desired compound hapten 4 at a yield of 36% over 3 steps from 16.

Evaluation of TT-Hapten Conjugates In Vivo
The haptens were separately conjugated to tetanus toxoid (TT) carrier protein using the previously optimized method [18,23,24,27]. Vaccines were adsorbed to aluminum hydroxide and mixed with Army Liposome Formulation (ALF43) adsorbed to aluminum hydroxide as an adjuvant. Immunogenicity was assessed by immunizing mice and collecting sera at weeks 0, 3, and 6. Antibodies to the immunizing haptens were measured using ELISA that used BSA-hapten conjugates as coating agents. Results showed that all of the haptens induced high antibody endpoint titers (>10 5 after the second vaccine dose) against their respective antigens ( Figure 3b). This suggests that all of the vaccine candidates tested are immunogenic.
Multiple reports have suggested that antibody endpoint titers may not predict efficacy of a vaccine against small molecules such as drugs of abuse [5,23]. To this end, we assessed the efficacy of each vaccine candidate by in vivo hot plate assay. This assay measures the time it takes for the mouse to respond to a pain stimulus (heat). This has been used in the past to evaluate vaccines to drugs of abuse [27,28]. Our results showed that among the conjugates tested, only TT-2 and TT-3 gave low % MPE values, suggesting that mice were protected from the antinociceptive effects of heroin ( Figure 3c). The other conjugates, TT-1 and TT-4, carrying epimeric haptens showed no significant difference with the unvaccinated controls ( Figure 3c). While all TT-hapten conjugates induced high antibody endpoint titers to the targets, only TT-2 and TT-3 showed protection against heroin, suggesting that antibody titers may not predict the efficacy of heroin vaccines in vivo. This is consistent with previous reports [23,29].
conjugates, TT-1 and TT-4, carrying epimeric haptens showed no significant difference with the unvaccinated controls ( Figure 3c). While all TT-hapten conjugates induced high antibody endpoint titers to the targets, only TT-2 and TT-3 showed protection against heroin, suggesting that antibody titers may not predict the efficacy of heroin vaccines in vivo. This is consistent with previous reports [23,29]. On week 10, mice received 1.0 mg/kg heroin (s.c.) and nociception was measured using the hot plate assay set at 54 °C. Response was reported in terms of %MPE. Results shown are the mean ± sem. One-way analysis of variance (ANOVA) with Dunnett's correction for multiple comparisons was used to determine statistical significance: ****, p < 0.0001, ***, p = 0.0001, vs. control group.

Determination of Drug Sequestration In Vitro
Anti-heroin vaccines are thought to act by inducing antibodies that sequester the drugs in the periphery [4][5][6]. The sequestered drug is prevented from crossing the bloodbrain barrier due to increase in apparent size thus effectively blocking the drug's physiological effects. In vivo, heroin quickly hydrolyzes to 6-acetyl morphine (6-AM) and morphine. Studies suggested that the physiological effects of heroin are mainly due to heroin Balb/c mice (n = 10 per treatment group) were immunized with 50 µL of the TT-hapten formulation (10 µg TT-hapten containing ALF 50 mM MLV-PHAD and 30 µg Alhydrogel) at weeks 0, 3, and 6 and bled at weeks 0, 3, 6, and 8. (b) Anti-hapten ELISA of sera using BSA-hapten conjugates as coating agents. (c) Hot plate nociception assay. The color legends in (c) are the same as in (b). On week 10, mice received 1.0 mg/kg heroin (s.c.) and nociception was measured using the hot plate assay set at 54 • C. Response was reported in terms of %MPE. Results shown are the mean ± sem. One-way analysis of variance (ANOVA) with Dunnett's correction for multiple comparisons was used to determine statistical significance: ****, p < 0.0001, ***, p = 0.0001, vs. control group.

Determination of Drug Sequestration In Vitro
Anti-heroin vaccines are thought to act by inducing antibodies that sequester the drugs in the periphery [4][5][6]. The sequestered drug is prevented from crossing the blood-brain barrier due to increase in apparent size thus effectively blocking the drug's physiological effects. In vivo, heroin quickly hydrolyzes to 6-acetyl morphine (6-AM) and morphine. Studies suggested that the physiological effects of heroin are mainly due to heroin and 6-AM with little contribution from other downstream metabolites such as morphine and the glycosylated metabolites [5,30].
In vitro drug binding experiments using equilibrium dialysis and liquid chromatography tandem mass spectrometry (ED-LC-MS/MS) were performed to assess the drug sequestering potential of mice sera [22]. First, heroin and its two bioactive metabolites, 6-AM, and morphine were tested. Results showed that mice sera at week 8 after the first dose only TT-2 provided the highest binding to heroin (Figure 4a,b). TT-1 and TT-3 showed slightly (fraction bound~0.4) but significantly higher than before the first dose (week 0). We arbitrarily defined a fraction bound value of <0.5 at low serum dilutions (1:200 or less) as weak binding. A similar trend was observed for the metabolite 6-AM where TT-2 and TT-3 showed the highest binding (fraction bound >0.4) (Figure 4c,d). All conjugates tested did not bind morphine (Figure 4e,f). These results are consistent with the in vivo efficacy data (Figure 3c), where only TT-2 and TT-3 induced protective effects against heroin.
In vitro drug binding experiments using equilibrium dialysis and liquid chromatography tandem mass spectrometry (ED-LC-MS/MS) were performed to assess the drug sequestering potential of mice sera [22]. First, heroin and its two bioactive metabolites, 6-AM, and morphine were tested. Results showed that mice sera at week 8 after the first dose only TT-2 provided the highest binding to heroin (Figure 4a,b). TT-1 and TT-3 showed slightly (fraction bound ~0.4) but significantly higher than before the first dose (week 0). We arbitrarily defined a fraction bound value of <0.5 at low serum dilutions (1:200 or less) as weak binding. A similar trend was observed for the metabolite 6-AM where TT-2 and TT-3 showed the highest binding (fraction bound >0.4) (Figure 4c,d). All conjugates tested did not bind morphine (Figure 4e   We also tested the in vitro sequestration of drugs used to treat OUD, namely, buprenorphine, methadone, naloxone, and naltrexone. Results indicated that TT-1, TT-2, and TT-4 did not induce antibodies that can bind any of these drugs ( Figure 5). Only TT-3 induced antibodies that weakly bound buprenorphine (Figure 5b).
We also tested the in vitro sequestration of drugs used to treat OUD, namely, buprenorphine, methadone, naloxone, and naltrexone. Results indicated that TT-1, TT-2, and TT-4 did not induce antibodies that can bind any of these drugs ( Figure 5). Only TT-3 induced antibodies that weakly bound buprenorphine (Figure 5b).  The facial recognition hypothesis [16,17] predicts that TT-2 and TT-3 would induce protection in vivo because of their resemblance to the target dug, heroin. Indeed, we observed that mice sera induced by these conjugates bound heroin and 6-AM (Figure 4a-d).
Interestingly, the epimers of haptens 1 and 4 failed to show binding to heroin or 6-AM.
Heroin and 6-AM have an absolute configuration opposite that of the epimeric haptens 1 and 4 (Figure 4a-d). This suggested that the configuration of the acetamide in the hapten at the C6 position is essential to induce antibodies that would bind heroin and 6-AM. This observation reinforced the facial recognition hypothesis that even subtle difference in the hapten structure, in this case, chirality at the C6 position, can vastly alter the specificity of the induced antibodies. The chirality of haptens and its effect on efficacy have been previously demonstrated for methamphetamine vaccines [5]. This was further visualized by constructing the Molecular Mechanics 2 (MM2)-optimized 3D models of the haptens and the target drugs ( Figure 6). It was observed that heroin and 6-AM resembled more closely the structures of haptens 2 and 3, than haptens 1 and 4, with emphasis on the substituent at the C6 position. The carbon-carbon double bonds in the morphinan ring did not significantly alter the 3D orientation of the haptens. The antibodies induced by TT-2 and TT-3 were notably weak (Figure 4). This may be because neither of these haptens exactly mimicked the three-dimensional orientation of the target drugs. Importantly, none of the conjugates induced antibodies that can bind morphine, methadone, naloxone, or naltrexone. This observation was expected because the structure of the haptens is distinct from the structure of these drugs [17]. Interestingly, we observed that TT-3 induced antibodies that weakly recognized buprenorphine (fraction bound~0.4 at 1:100 serum dilution). When modeled, the structure of buprenorphine has some resemblance to hapten 3 (Figure 1), where the methoxy group in the C6 position of buprenorphine is projected away from the linker attachment site, and thus a part of the "face" recognized by the immune system. This structural feature is unique to buprenorphine among the therapeutics tested (buprenorphine, naloxone, naltrexone, and methadone). This difference may be attributed to the absence of the hydroxyl group at the C9 position that is present in naloxone and naltrexone. This hydroxyl group induced geometric change in those two molecules that cannot occur in buprenorphine, leaving the latter with a unique "face" that slightly mimicked that of hapten 3. Naloxone and naltrexone share the same epoxymorphinan ring structure as the haptens and buprenorphine, but their structures in three-dimensional space suggest no significant similarity to the "face" of the haptens exposed to the immune system for recognition. Collectively, these results suggested that linker attachment at the C1 position is feasible without altering the cross-reactivity of the antibodies against OUD therapeutics. It is important that anti-opioid vaccines induce antibodies that will not cross-react with therapeutic drugs since the ultimate goal of a vaccine is to complement existing treatments [4][5][6].
The present study has limitations. First, this study only evaluated a vaccine based on the core morphinan structure with the linker attachment site at the C1 position. The effect of chirality at C6 in the epoxymorphinan ring in the hapten on the quality of the vaccineinduced antibodies should also be investigated where the linker attachment is on another site in the epoxymorphinan nucleus that projects a distinctly different face of the hapten. Second, this study did not attempt to evaluate the functional changes in the vaccineinduced antibodies. Recently, we have demonstrated that affinity maturation takes place. Conducting this same study for the TT-hapten conjugates described here in a future study might provide valuable insights as to how hapten-specific antibody affinity changes with time. Third, since only an in-bred mouse strain was used in this study, it would be interesting to see how these vaccines behave in other animal models, such as rats, rhesus macaques, and rabbits.

Materials and Methods
All reagents were obtained from commercial sources and used without further purification. Melting points were determined on a Mettler Toledo MP70 Melting Point System (Mettler Toledo, Columbus, OH, USA) and are uncorrected. Proton nuclear magnetic resonance ( 1 H NMR, 400 MHz) and carbon nuclear magnetic resonance ( 13 C NMR, 100 MHz) spectra were recorded on a Varian 400 wide-bore spectrometer (Varian, Palo Alto, CA, USA) in CDCl3 (unless otherwise noted) with the values given in ppm and

Immunizations and Animal Challenge
All animal studies were conducted under an approved animal use protocol in an Association for Assessment and Accreditation of Laboratory Animal Care International (AAALACI)-accredited facility in compliance with the Animal Welfare Act and other federal statutes and regulations relating to animals. Experiments involving animals adhered to the principles stated in the Guide for the Care and Use of Laboratory Animals, 8th edition [34]. Female BALB/c mice, 6-7 weeks of age (n = 10 per group, for a total of 50 mice) were immunized intramuscularly (i.m.) on weeks 0, 3, and 6 using 50 µL of the vaccine formulation. This dose contained 10 µg of TT−hapten (based on the protein content of the protein−hapten conjugate), 20 µg of PHAD ® in ALF43 [18,23], and 30 µg of aluminum in aluminum hydroxide (Alhydrogel ® ) in DPBS pH 7.4. Mice were bled for ELISA assay at weeks 0, 3, 6, and 8. Challenge studies were performed at week 10.
Mice received 1.0 mg/kg of heroin (s.c.), and this route has been used previously to evaluate anti-opioid vaccines [27,28]. Antinociceptive effects were assessed 15 min after each heroin injection using the hot plate assay [35]. This assay involved placing the mouse on a hot plate set at 54 • C. The latency to a nociceptive response was recorded, defined as the time elapsed until the mouse licked or flicked its hind paw. The latency times were measured with a cutoff time of 30 sec to prevent injury. Antinociception, measured as %MPE, was calculated using Equation (1)

Drug Binding Analysis
Drug binding was measured using ED-LC-MS/MS as described previously [22,27,28]. Briefly, mice sera were diluted with 0.05% BSA in DPBS, pH 7.4 (ED buffer) containing 3-5 mg/mL NaF and 5 nM of a drug. The following drugs were tested: heroin, 6-A.M., morphine, methadone, naloxone, buprenorphine, and methadone. An aliquot (100 µL) was seeded into sample chambers of the ED plate and the buffer chamber was filled with 300 µL of ED buffer. The plate was incubated at 4 • C and 300 rpm for 24 h in a thermomixer. Aliquots (90 µL) from sample and buffer chambers were pipetted out and analyzed by LC-MS/MS.

Data Analysis
Graphical and statistical analyses were performed using Prism 9 (GraphPad Software, San Diego, CA, USA). All results were reported as the mean ± standard error of the mean (sem). The chemical structures and 3D optimization were made using ChemDraw 18.1. The built-in MM2 method was used for geometry optimization and energy minimization.

Conclusions
In this study, we described the synthesis of haptens 1-4, (Figure 1), produced the individual TT-hapten conjugates and evaluated their efficacy both in vivo and in vitro. We showed that only haptens 2 and 3 can induce protection against the effects of heroin in vivo. The epimeric congeners of these haptens, haptens 1 and 4, failed to protect mice from the effects of heroin. We also showed that the in vivo efficacy is consistent with the results of the in vitro drug sequestration assay. Only TT-2 and TT-3 yielded antibodies that bound heroin and 6-AM. None of the TT-hapten conjugates induced antibodies that cross-reacted with morphine, methadone, naloxone, or naltrexone, and only TT-3 interacted weakly with buprenorphine. These results highlight the importance of judicious hapten design toward effective heroin vaccines.