Blocking interleukin-4 enhances efficacy of vaccines for treatment of opioid abuse and prevention of opioid overdose

Vaccines offer an option to treat heroin and prescription opioid abuse and prevent fatal overdoses. Opioid vaccines elicit antibodies that block opioid distribution to the brain and reduce opioid-induced behavioral effects and toxicity. The major limitation to the translation of addiction vaccines is that efficacy is observed only in subjects achieving optimal drug-specific serum antibody levels. This study tested whether efficacy of a vaccine against oxycodone is increased by immunomodulators targeting key cytokine signaling pathways involved in B and T cell lymphocyte activation. Blockage of IL-4 signaling increased vaccine efficacy in blocking oxycodone distribution to the brain and protection against opioid-induced behavior and toxicity in mice. This strategy generalized to a peptide-protein conjugate immunogen, and a tetanus-diphtheria-pertussis vaccine. These data demonstrate that cytokine-based immunomodulators increase efficacy of vaccines against small molecules, peptides and proteins, and identify IL-4 as a pharmacological target for improving efficacy of next-generation vaccines.

Vaccines are the single most effective intervention ever devised in medical practice 1 . Vaccines have the potential to protect against a variety of communicable and non-communicable diseases, including substance use disorders associated with abuse of stimulants and opioids. Worldwide, over 33 million people abuse drugs, whose health consequences include blood-borne pathogen transmission, cancer, and death 2 . In the US, 2.6 million people are dependent on heroin and prescription opioids including oxycodone and hydrocodone, costing more than $95 billion in crime, lost work productivity and health care annually 3 . Over 50,000 opioid-related fatal overdoses occur each year, making opioid abuse the leading cause of accidental death in the US, and prompting recent increases in stringency of prescription guidelines 3 . These data suggest that current interventions are insufficient to curb opioid abuse, and that more therapeutic options are needed 4 .
Vaccines and other biologics against substance use disorders have shown pre-clinical and clinical proof of concept, but no product has reached the market [5][6][7][8][9][10] . A major limitation to the translation of addiction vaccines is that clinical efficacy is observed only in the fraction of immunized subjects that display high levels of drug-specific antibodies 5,11 . This is a common shortcoming of vaccines consisting of purified protein subunits, or of proteins conjugated to peptides, carbohydrates, or synthetic small molecule haptens. For instance, the increased incidence in whooping cough is partially attributable to switching from the whole-cell pertussis vaccine to formulations containing the purified acellular pertussis subunit 12 . To increase efficacy, synthetic vaccines against substance use disorders are mixed with adjuvants or delivered via particle platforms 5 . However, a limited number of adjuvants or delivery systems are approved for human use because of potential side effects, and available adjuvants trigger innate immunity, but do not directly stimulate adaptive immunity 13,14 , suggesting that more effective immunization strategies are needed.
. Immunization success or failure hinges on the fine balance between these cytokines and other co-stimulatory factors, and their downstream effects on B and T cells.
This study tested interleukins as a pharmacological target to enhance vaccine efficacy against drugs of abuse and other targets. Immune checkpoint inhibitors, interleukins, and monoclonal antibodies (mAb) against interleukins or interleukin receptors have been shown to increase efficacy of CD8 + T cell-mediated immunotherapies 29,30 . Combining a HIV gag/pol vaccine with IL-4 receptor antagonism induced both HIV-specific CD8 + T cells, and gag-specific B cell immunity paired to enhancement of IgG 2a class switching in mice 31 . However, the effect of immunomodulators on subunit or conjugate vaccines that rely on CD4 + T cell-dependent B cell activation to generate antibodies remains largely unexplored. Since small molecules and biologic immunomodulators of IL-4, IL-2 and IL-7 are either approved or close to FDA approval for different indications, this study focused on testing the novel hypothesis that modulation of IL-4, IL-2 and IL-7 enhanced efficacy of the OXY-KLH vaccine against oxycodone. An αIL-4 mAb (Pascolizumab ® ) has been tested for allergies and pulmonary tuberculosis (NCT01638520), and other therapeutics targeting IL-4 (e.g., Altrakincept ® ), IL-4Rα (e.g., Dupilumab ® , AMG-317, and Pitrakinra ® ) are under development for treatment of atopic dermatitis, allergies, autoimmune disorders, or other indications 32 . An αCD25 mAb (anti-IL-2αR, Daclizumab ® ) is approved for prevention of kidney transplant rejection, treatment of melanoma and other cancers 33 . An αCD127 mAb (anti-IL-2αR) is being evaluated for treatment of autoimmune disorders (NCT02293161) and multiple sclerosis (NCT01808482). Repurposing of approved or commercially available immunomodulators can accelerate translation of this approach and development of clinically viable vaccine formulations.
In this study, we found that IL-4 depletion enhanced OXY-KLH's efficacy against oxycodone and increased the subset of vaccine responders. The increased efficacy correlated with higher IgG 2a titers and a mixed Th 1 /Th 2 response. This strategy generalized to a peptide-protein conjugate immunogen and a vaccine for infectious diseases. Our data demonstrate that targeting of IL-4 is a safe and effective strategy to improve efficacy of vaccines against substance use disorders and other unmet medical needs.

Results
Vaccine efficacy is enhanced by modulators of IL-4 signaling. First, this study tested whether modulation of either Th 2 signaling through IL-4 or Th 1 signaling through IL-2 improved efficacy of the OXY-KLH vaccine. In sequential rounds of screening, independent mouse cohorts were immunized with lead vaccine formulations and challenged with increasing doses of oxycodone (2.25-10 mg/kg). These drug challenges model clinically relevant oxycodone doses and resulting plasma concentrations found in recreational opioid users and individuals in pain management regimens [34][35][36][37][38][39][40] . As predetermined screening criteria, new leads had to be more effective than OXY-KLH in aluminum adjuvant to be tested in subsequent rounds. We predicted that blocking signaling through IL-2 or IL-7, or promoting IL-4 signaling, would increase efficacy through increased GC formation. In the first round, OXY-KLH was administered in combination with immunomodulators including recombinant IL-4 (rIL-4), rIL-4+ recombinant IL-21 (rIL-21), or with antibodies against the IL-2 receptor (αCD25). Immunocomplexes of interleukins and anti-interleukin mAb (IL:αIL) are potentially more effective than interleukins alone 41 and have longer half-life 42,43 . For example, IL-2:αIL-2 and IL-7:αIL-7 potentiate cellular immunity against cancer by enhancing activation of the IL-2 or IL-7 receptor α chain (IL-2αR or IL-7αR) 44 , whereas immunocomplexes consisting of IL-4 and a neutralizing mAb against IL-4 (αIL-4, clone 11B11) showed increased activity at the IL-4α receptor and protection against parasites 45 . Thus, we also tested whether αIL-4 or the IL-4:αIL-4 complex enhanced vaccine efficacy.
Surprisingly, co-administration of OXY-KLH and rIL-4 did not increase Ab titers or vaccine efficacy ( Fig. 1A-C), and co-administration of OXY-KLH with both IL-4 and IL-21 did not increase vaccine efficacy. Contrary to our expectations, blocking IL-2 signaling with the αCD25 mAb or co-administration of αCD25 with IL-4 did not enhance OXY-KLH efficacy. Instead, both the αIL-4 mAb and the rIL-4:αIL-4 immunocomplex showed significantly increased efficacy compared to OXY-KLH alone in blocking a dose of 2.25 mg/kg oxycodone (Fig. 1B,C). Finally, co-administration of OXY-KLH with the αCD127 mAb targeting the IL-7 receptor (Suppl. Fig. S1), or with the co-stimulatory molecules programmable death ligand 1 or ICOS ligand (Suppl. Fig. S2) did not increase vaccine efficacy.

Combining OXY-KLH and αIL-4 mAb blunted oxycodone-induced behavioral and toxic effects.
To further explore the clinical relevance of this approach, we tested whether immunization blocks oxycodone-induced respiratory depression and bradycardia, which are common side effects contributing to overdose-related death and complications 46 . Immunized mice were also tested for oxycodone analgesia on a hotplate as a measure of vaccine efficacy in blocking opioid central effects 15,16 . OXY-KLH was effective in preventing oxycodone-induced depression of heart (Fig. 2D) and respiratory (Fig. 2E) rates as well as blocking oxycodone-induced analgesia over increasing doses (Fig. 2F). Co-administration of OXY-KLH and αIL-4 mAb potentiated the effect of the vaccine on respiratory depression evoked by a cumulative dose of 4.5 mg/kg oxycodone. Additionally, the efficacy of naloxone in reversing opioid-induced analgesia, bradycardia, and respiratory depression was preserved in both OXY-KLH and OXY-KLH plus αIL-4 mAb groups compare to KLH control. These data support that opioid vaccines can be combined with naloxone to rescue from fatal overdose.  The lead formulation of OXY-KLH plus αIL-4 mAb is effective across clinically-relevant immunization regimens and attenuated oxycodone-induced physiological and behavioral effects. BALB/c mice were immunized with OXY-KLH or KLH on days 0, 14 and 28, and tested for vaccine efficacy a week after the third immunization. Mice received KLH or OXY-KLH s.c. or i.m., whereas αIL-4 mAb was administered i.p., i.v., or i.m. In a first cohort, mice were challenged with 5 mg/kg oxycodone: (A) oxycodone-specific IgG titers, (B) oxycodone brain concentrations, and (C) subjects that showed more than 50% reduction in oxycodone distribution to the brain compared to the KLH group. Data are from two independent experiments and expressed as mean ± SEM. Percentages (%) above bars indicate reduction compared to KLH. One-way ANOVA paired with Tukey's multiple comparisons test or unpairedtwo-tailed t-test. Chi-square test (χ 2 = 62.66, df = 6). *p < 0.05, **p < 0.001, ****p < 0.0001 compared to KLH, and brackets to indicate significance between groups. In a second cohort, immunized mice were tested for opioid-induced respiratory depression, heart rate and antinociception. Oxycodone was administered s.c. every 15 min at increasing doses and the doses listed are the cumulative dose received. Naloxone 0.1 mg/kg s.c. was administered 15 min after the final oxycodone dose to reverse opioid effects. (D) heart rate, (E) breath rate, and (F) antinociception on a hot plate. Data are mean ± SEM. Data shown are from one experiment (n = 8). Two-way ANOVA paired with Tukey's multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001 compared to KLH. # p < 0.05, ## p < 0.01 for OXY-KLH compared to OXY-KLH plus anti-IL-4 mAb.
To gain insight in the possible mechanism underlying the effect of the αIL-4 mAb on vaccine efficacy, we next immunized IL-4-deficient (IL-4 −/− ) mice with OXY-KLH. IL-4 −/− mice showed increased OXY-KLH efficacy, which was associated with increased production of IgG 2a and IgG 3 (Fig. 3F-H), in agreement with previous reports that indicate that immunization or infection of IL-4 KO mice elicit increased IgG 2a antibodies 49 . Similarly, when OXY-KLH was administered in combination with an αIL-4 mAb that is non-neutralizing 42 , vaccine efficacy was not affected (Suppl. Fig. S6). Together, these results show that the effect of the neutralizing mAb, clone 11B11, is due to a reduction in IL-4 signaling.
In both IL-4 depleted mice and IL-4 −/− mice, greater vaccine efficacy against oxycodone was associated with increased IgG 2a and IgG 3 antibody titers. This finding is consistent with other reports that active immunization against drugs of abuse benefits from a mixed Th 1 /Th 2 response 50-54 , but its biological relevance remains unknown. We hypothesized that IgG-dependent effector functions may be involved at least in part in vaccine efficacy against drugs of abuse. IgG 2a and IgG 3 activate various effector functions, including phagocytosis through the Fcγ receptors I-IV 55 . To test the hypothesis that antibody-mediated phagocytosis is required for vaccine efficacy against drugs of abuse, mice immunized with OXY-KLH were administered liposome-embedded clodronate, an in vivo macrophage poison 56 , and then challenged with 5 mg/kg oxycodone. Treatment with clodronate reduced the efficacy of OXY-KLH in preventing oxycodone distribution to the brain (Fig. 4I). This is the first report of IgG-dependent phagocytosis as a potential mechanism underlying efficacy of addiction vaccines. Future studies will further explore the role of IgG subclasses in mediating the effect of polyclonal and monoclonal antibodies against opioids and other drugs of abuse.
Co-administration of OXY-KLH and αIL-4 mAb is potentially safe. All mice that received the combination of OXY-KLH s.c. and αIL-4 i.p. were evaluated for potential adverse effects. Mice receiving this formulation exhibited no significant difference in baseline nociception or body weight as compared to mice that received KLH control (Suppl . Table S1). An independent cohort of BALB/c mice immunized with KLH, OXY-KLH, αIL-4, or OXY-KLH plus αIL-4 underwent a clinical toxicology (Suppl . Table S2) and histopathology assessment ( Fig. 4 and Suppl. Fig. S7) to evaluate the safety profile of this approach. Across groups, serum protein, bilirubin, urea nitrogen, and cholesterol concentrations as well as alanine aminotransferase (ALT) activity were within the reference interval. In all groups, elevations in aspartate aminotransferase (AST) activity compared to reference values were noted and, in light of the unremarkable ALT activity and the liver histopathology, is not likely the result of mAb treatment. The histopathology of treated animals revealed mild to occasionally moderate vacuolar change within the centrolobular hepatocytes that was similar in severity across all treatment groups with no significant pathology in the spleen. However, lungs from all groups displayed mild to occasionally moderately increased numbers of interstitial inflammatory cells, principally lymphocytes with fewer neutrophils with concurrent congestion. Occasionally, these regions of inflammation were centered on lymphocyte-rich lymphatic vessels. The severity of the lung pathology was similar across the four treatment groups. No significant pathology was identified in other organs (Suppl. Fig. S7).

IL-4 modulation does not interfere with post-immunization hapten-specific B cell early expansion, but increases peptide-specific T cell differentiation. Since co-administration of OXY-KLH and αIL-4
increased vaccine efficacy against oxycodone, we sought to gain a better understanding of the effects of IL-4 depletion on GC in vivo in our model. In mice immunized with OXY-KLH, depletion of IL-4 did not affect early expansion of OXY-specific antibody-secreting and GC B cell subsets (Fig. 5A,B, and Suppl. Fig. S8 for analysis detail). These data suggest that hapten-specific B cells responses were at least in part IL-4-independent at this early stage.
The formation and persistence of GC requires functional and anatomical synergy between GC B cells and GC-Tfh 57 . Here we tested the effect of IL-4 depletion on T cell responses to immunization with a model immunogen, 2W peptide conjugated to streptavidin-phycoerythrin (SA-PE). The 2W SA-PE immunogen allows tracking of 2W-specific T cells by 2Wp:MHCII tetramer technology paired to flow cytometry to assess T cell differentiation, and models peptide-protein vaccines. Co-administration of 2W SA-PE plus αIL-4 increased Tfh differentiation toward GC-Tfh, the critical T cell helper subset devoted to sustain cognate GC B cell expansion (Fig. 5C-F, and Suppl. Fig. S9-10 for analysis detail). At 7 days after immunization, αIL-4 increased antigen-specific non-Tfh expansion (Fig. 5C), which may explain IgG subclass shift toward IgG 2a and IgG 3 observed in mice immunized with OXY-KLH plus αIL-4 (Fig. 3). Most importantly, αIL-4 shifted T cell differentiation toward Tfh and GC-Tfh (Fig. 5F). These data support the use of mAb against IL-4 to increase efficacy of vaccines consisting of carrier proteins conjugated to peptide antigens, such as cancer-derived or Alzheimer's β amyloid-derived peptides.

IL-4 modulation increases efficacy of tetanus-diphtheria-pertussis vaccine.
Current acellular pertussis vaccines are safer but not more effective than whole-cell pertussis vaccines. Here we found that co-administration of the αIL-4 mAb with TDaP (GSK©) improved TT-specific serum IgG titers over time (Fig. 5G), and increased TT-specific IgG 2a titers (Fig. 5H). In addition, an in vitro assay was used to evaluate Ab-mediated protection of Vero cells from diphtheria toxin. Co-administration of αIL-4 with TDaP significantly increased the subset of immunized subjects that produced antibodies that protected cells from the toxin (Fig. 5I  and Fig. S10). These data suggest that depletion of IL-4 is a strategy that can be used to enhance efficacy of vaccines for multiple applications.

Discussion
This study tested the novel hypothesis that the efficacy of vaccines for substance use disorders is enhanced by immunomodulators that target key interleukin signaling pathways involved in critical B and T cell functions underlying antibody production. Here, we found that combining a candidate vaccine against the prescription opioid oxycodone with an IL-4 neutralizing mAb enhanced vaccine efficacy in blocking oxycodone distribution to the brain, and increased the subset of immunized subjects that displayed vaccine efficacy. Both the OXY-KLH vaccine and the combination of OXY-KLH with αIL-4 mAb were effective in blocking oxycodone-induced behavioral effects and preventing respiratory depression and bradycardia in mice challenged with increasing doses of oxycodone. The effect of OXY-KLH on preventing oxycodone-induced adverse effects that contribute to overdose death supported findings from another report that oxycodone vaccines prevented oxycodone overdose mortality 58 . In this study, both the OXY-KLH vaccine and the combination of OXY-KLH with αIL-4 mAb did not interfere with reversal by naloxone of respiratory depression and bradycardia. These data replicated and extended evidence that OXY-KLH did not interfere with naloxone and extended-release naltrexone in mice and rats 20 , supporting the notion that opioid vaccines can be safely combined with current pharmacotherapy against opioid abuse and overdose. The combination of OXY-KLH with αIL-4 mAb was effective across multiple routes of administration, and no significant adverse effects were observed, indicating that modulation of IL-4 is a potentially safe method that can be used to enhance vaccine efficacy against a variety of targets or in at-risk populations.
Co-administration of the OXY-KLH vaccine plus αIL-4 mAb or immunization of IL-4 −/− mice yielded a more balanced Th 1 /Th 2 response as demonstrated by the presence of IgG 1 as well as IgG 2a and IgG 3 in contrast to a primarily Th 2 (i.e., IgG 1 only) response in mice immunized with OXY-KLH in alum. While other groups have shown that more effective formulations of addiction vaccines elicit a more balanced Th 1 /Th 2 response 50-53 , the biological relevance of IgG 2a and IgG 3 in the context of addiction vaccines is still unknown. Our data support the hypothesis that blockage of IL-4 induced a shift toward IgG 2a and IgG 3 , and that presence of IgG 2a increased vaccine efficacy against oxycodone. For instance, IgG 2a is a key mediator for efficacy of influenza vaccines, and it is possible that IgG 2a also elicits antibody-dependent effector functions that increase removal or inactivation of drugs of abuse. In this study, ablation of macrophages using clodronate in vivo reduced OXY-KLH vaccine efficacy against oxycodone, indicating a possible role for antibody-mediated phagocytosis of opioids. Future studies will focus on addressing the contribution of antibody-dependent effector functions to efficacy of vaccines and mAb against opioid abuse and overdose.
This study suggested that IL-4 signaling and specific IgG subclasses mediate efficacy of addiction vaccines in mice. We previously reported that the frequency of naïve and early antigen-specific B and T cell subsets correlated with individual vaccine efficacy against nicotine or opioids in mice [23][24][25] . Greater GC formation was also associated with increased vaccine efficacy against oxycodone in mice 23 . In addition to offering insight into the immunobiology of addiction vaccines, these data provide benchmarks for pre-clinical and clinical development of addiction vaccines. The frequency of antigen-specific B and T cells in human blood paired with analysis of IL-4 genetic polymorphisms and drug-specific IgG subclasses may also provide biomarkers predictive of vaccine efficacy in human subjects.
Combining immunomodulators with active immunization was also effective at increasing GC-Tfh differentiation in response to a model peptide-protein conjugate vaccine, as well as increasing the efficacy of the TDaP vaccine. This suggests the potential for broader application of immunomodulators to enhance efficacy of subunit or conjugate vaccines that may benefit from a mixed Th 1 /Th 2 -type response or from a specific IgG subclass distribution. This strategy may generalize to sensitive populations including pregnant women, elderly, or immunocompromised individuals known to be poor responders to vaccines. This study provided additional pre-clinical proof of efficacy that opioid vaccines offer a viable approach to the treatment of opioid abuse and prevention of opioid fatal overdoses.

Ethics statement. Studies were approved by the Minneapolis Medical Research Foundation and the
University of Minnesota Institutional Animal Care and Use Committees. All procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals (8th Edition, National Academies Press).
Vaccines and immunogens. The oxycodone-based hapten was synthesized from oxycodone HCl, and included a tetraglycine linker 16  Recombinant interleukin 4 (30,000 IU and 60,000 IU) and rIL-4 plus rIL-21 (60,000 IU total) were administered s.c. at days 0, 14 and 28 in combination with each immunization. The rIL-4:αIL-4 mAb immunocomplex was mixed prior to injection (30,000 IU and 0.5 mg of mAb) and administered once s.c. on day 0. Monoclonal antibodies were administered (0.5 mg/mouse/injection) two days prior and one day after the first immunization. The αIL-4 mAb was administered either i.p., i.v., or i.m. as noted in each figure. Analysis of the OXY-specific B cell repertoire was performed a week after the first immunization, whereas Ab characterization and oxycodone challenges were initiated a week after the third immunization. An independent cohort of BALB/c mice immunized with KLH or OXY-KLH was first analyzed for anti-oxycodone Ab titers, given liposome-embedded clodronate (Liposoma, Amsterdam, The Netherlands) to poison macrophages in vivo 56 , and then challenged with oxycodone. In some experiments, C57Bl/6 mice were immunized s.c. with 25 µg of the immunogen PE-SA 2W adsorbed on alum adjuvant, and analysis of the 2W-specific T cell repertoire was performed a week after the first immunization. In an independent study, male BALB/c mice were immunized s.c. on days 0, 14 and 28 with 25 µl of TDaP diluted in 0.01 M phosphate-buffered saline to a final volume of 100 µl, and given either saline or αIL-4 mAb i.p. two days prior and one day after the first immunization. Antibody analysis. Standard 96-well ELISA plates were coated with unconjugated chicken ovalbumin (OVA) or OXY hapten conjugated to OVA, at 5 ng/well in carbonate buffer at pH 9.6 and blocked with 1% gelatin 23 . Similarly, plates were coated with 50 ng/well of TT or OVA. Primary antibodies were incubated with goat-anti-mouse IgG, IgG 1 , IgG 2a , or IgG 3 conjugated to horseradish peroxidase to measure titers of oxycodoneor TT-specific IgG and individual IgG subclasses (Alpha Diagnostic International, Inc., San Antonio, TX).

Analysis of antigen-specific B and T cells.
Lymph nodes and/or spleen were collected at seven days after the first immunization. Using a two-step procedure, OXY-specific B cells were first isolated by magnetic enrichment using a Quadromacs Separator (Miltenyi Biotec, Auburn, CA) and biotinylated oxycodone-based haptens conjugated to PE-SA or a decoy reagent, anti-PE mAbs conjugated to magnetic particles (Miltenyi Biotec) and then analyzed by multi-parameter flow cytometry 23,25 (and Suppl. Fig. S8). Similarly, antigen-specific T cells were first isolated by magnetic enrichment using 2W conjugated to soluble MHCII receptors (2Wp:MHCII) tetramerized on SA-PE and then analyzed by flow cytometry using standard surface and intracellular markers 23,59 . 2Wp:MHCII-restricted Tfh and GC-Tfh were further characterized as described 60 (and Suppl. Fig. S9). Analysis was conducted on a LSRFortessa (BD Biosciences, San Jose, CA), and analyzed using FlowJo (FlowJo LLC, Ashland, OR).
Vaccine efficacy against oxycodone distribution to the brain, oxycodone-induced behavioral and toxic effects. Independent cohorts of mice immunized with either KLH, OXY-KLH, or OXY-KLH in combination with immunomodulators were challenged with increasing doses of oxycodone (2.25-10 mg/kg, s.c.). At 30 min after the oxycodone challenge, mice were euthanized and oxycodone serum and brain concentrations measured by gas chromatography as described 23 . In a separate study, immunized mice were tested for vaccine efficacy in blocking opioid-induced respiratory depression, heart rate, and anti-nociception. Oxycodone-induced analgesia was first assessed on a hotplate as described 23 , and then arterial oxygen saturation (SaO 2 ), breath rate, and heart rate were measured with a MouseOx oxymeter (STARR Life Sciences Corp., Oakmont, PA). Oxycodone was administered s.c.
Clinical pathology. Mice were immunized s.c. with either KLH or OXY-KLH and given either saline or αIL-4 mAb i.p. as described above. A week after the third immunization, BALB/c mice were euthanized with CO 2 to perform histopathology and clinical pathology. Serum was collected by centrifugation of clotted whole blood at 7500 rpm for three minutes at 4 °C and collected in lithium heparin (VACUTAINER, Becton, Dickinson and Co., Franklin Lakes, NJ). Frozen samples were sent to Marshfield Labs (Marshfield, WI) for analysis of protein, bilirubin, alanine aminotransferase (ALT), urea nitrogen, total cholesterol, and aspartate aminotransferase (AST). Organs were harvested, fixed in 10% neutral buffered formalin (methanol free, ultrapure EM grade, Polysciences, Inc., Warminster, PA), and embedded in paraffin using standard methods. Thereafter, four-micron tissue sections were cut and stained with hematoxylin and eosin (H&E). Histopathologic evaluation was performed by a board certified veterinary pathologist (DMS) who was blinded to their treatment status.
In vitro diphtheria toxin protection assay. Mouse serum samples were first diluted 1:10 in RPMI containing 10% FBS and then serially diluted in 384-well plates using Bio-Tek Precision 2000 liquid-handlers (Bio-Tek, Winooski, VT). Forty µL of 100 ng/mL Diphtheria toxin (List Biological Laboratories, Campbell, CA) was added to each sample well and incubated for 1 h at 37 °C. Twenty µL of Vero cells (in Phenol Red-free DMEM containing 10% FBS [fDMEM], at a density 2 × 10 5 cells/mL) were then added to each well and incubated 24 h at 37 °C. Following incubation, cells were washed using a plate washer (Bio-Tek, Winooski, VT), and 40 µL of fDMEM containing 0.15 mM resazurin was added to each well and incubated 16 hrs. Plates were read on a Spectramax M2 (Molecular Devices, Sunnyvale, CA) using ratiometric reads at 530/620 nm to evaluate cell viability. The IC 50 for Vero cell survival was calculated, and the subset of immunized subjects that generated protective titers was defined as the proportion of samples in each group with IC 50 >10. Data Analysis. Mean titers, number of B and T cells, oxycodone concentrations, breath rate, hearth rate, and delay time to nociception across groups were analyzed by one-or two-way ANOVA followed by Tukey's multiple comparisons test. Unpaired two-tailed t-tests were used to analyze data between two groups. Chi-square or Fisher's test were used to analyze the subset (%) of immunized mice that showed vaccine efficacy against oxycodone or diphtheria-protecting titers.
Data availability. Supplementary information is available in the online version of the paper. Source data used to generate all figures or any other information are available from the corresponding author upon request.