Cocaine chemogenetics blunts drug-seeking by synthetic physiology

Abstract

Chemical feedback is ubiquitous in physiology but is challenging to study without perturbing basal functions. One example is addictive drugs, which elicit a positive-feedback cycle of drug-seeking and ingestion by acting on the brain to increase dopamine signalling^1-3. However, interfering with this process by altering basal dopamine also adversely affects learning, movement, attention and wakefulness⁴. Here, inspired by physiological control systems, we developed a highly selective synthetic physiology approach to interfere with the positive-feedback cycle of addiction by installing a cocaine-dependent opposing signalling process into this body-brain signalling loop. We used protein engineering to create cocaine-gated ion channels that are selective for cocaine over other drugs and endogenous molecules. Expression of an excitatory cocaine-gated channel in the rat lateral habenula, a brain region that is normally inhibited by cocaine, suppressed cocaine self-administration without affecting food motivation. This artificial cocaine-activated chemogenetic process reduced the cocaine-induced extracellular dopamine rise in the nucleus accumbens. Our results show that cocaine chemogenetics is a selective approach for countering drug reinforcement by clamping dopamine release in the presence of cocaine. In the future, chemogenetic receptors could be developed for additional addictive drugs or hormones and metabolites, which would facilitate efforts to probe their neural circuit mechanisms using a synthetic physiology approach. As these chemogenetic ion channels are specific for cocaine over natural rewards, they may also offer a route towards gene therapies for cocaine addiction.

Fig. 1. Cocaine-gated ion channels.

a, Synthetic physiology schematic for using cocaine chemogenetics to reverse reinforcement after ingestion of an addictive drug that binds to a chemogenetic receptor targeted to a neural circuit. The diagram was adapted with permission from ref. , AAAS. b, The chemical structures of cocaine and ACh. c, Cocaine (yellow) binds to AChBP at the interface of two protomer subunits (cyan and pink). Homologous α7 nAChR pre-protein residues and numbering are shown (modified from Protein Data Bank (PDB): 2pgz). d, The potency of cocaine and ACh agonism of chimeric channels comprising single-mutation α7 nAChR LBDs and IPDs from either 5HT3 or GlyR. The red asterisks highlight chimeric channels with cocaine agonism and reduced ACh potency. e, The relationship between ACh potency and cocaine potency at α7–5HT3 mutant chimeric channels: L141G (1), L141G G175K (2), L141G G175K Y217F (3), L141G G175K Y115F (4), L141G G175K Y210F (5), L141G G175K Y210F Y217F (6). Unmod., unmodified α7–5HT3. f, Coca-5HT3 potency for cocaine (left, n = 8 separate dose–responses) and ACh (right, n = 8 separate dose–responses) compared with unmodified α7–5HT3 (n = 3 separate dose–responses each). g, Displacement of [³H]ASEM at coca-5HT3 by cocaine or ASEM. n = 3 replicate curves per drug. Data are mean ± s.e.m. Source Data

Fig. 2. Functional properties of excitatory and inhibitory cocaine-gated ion channels.

a,b, Cocaine-activated (a) and ACh-activated (b) currents from coca-5HT3. c, Depolarization in hippocampal neurons expressing coca-5HT3 (n = 11 cells) or GFP (n = 9 cells). d, Action-potential firing in a hippocampal neuron expressing coca-5HT3 in response to cocaine. Downward deflections indicate current injections to monitor membrane properties. e, The fold change in current necessary to elicit an action potential (rheobase) in neurons expressing coca-5HT3 (n = 9 cells) or GFP (n = 4 cells) in response to 3 µM cocaine. Statistical analysis was performed using a two-sided t-test; P = 0.023. f,g, Cocaine-activated (f) and ACh-activated (g) currents from coca-GlyR in HEK293 cells. h, Membrane potential assay of coca-GlyR potency for cocaine (left; n = 7 separate dose–responses) and ACh (right; n = 8 separate dose–responses) compared with unmodified α7–GlyR (n = 3 separate dose–responses each). i, Cocaine inhibits excitability to current injection of a hippocampal neuron expressing coca-GlyR. j, The input resistance of hippocampal neurons expressing coca-GlyR (n = 8 cells) or GFP (n = 13 cells) in the presence of cocaine. Recovery after cocaine removal (WASH). k, The fold change of current that elicits an action potential (rheobase) in neurons expressing coca-GlyR (n = 10 (3 µM) and n = 6 (10 µM) cells) or GFP (n = 4 (3 µM) and n = 7 (10 µM) cells) in response to cocaine (two-way ANOVA, transgene, F_1,23 = 6.14, P = 0.02; dose, F_1,23 = 0.69; P = 0.41; transgene × dose, F_1,23 = 0.01, P = 0.93). l, The membrane properties of neurons expressing coca-5HT3 (n = 11 cells), coca-GlyR (n = 13 cells) or GFP (n = 14 cells). ANOVA for capacitance, F_2,35 = 0.040, P = 0.96; Kruskal–Wallis ANOVA by ranks for resistance: H₂ = 1.23, P = 0.54; Kruskal–Wallis ANOVA by ranks for resting potential: H₂ = 3.58, P = 0.17. Data are mean ± s.e.m. Not significant (NS), P > 0.05. Source Data

Fig. 3. Cocaine chemogenetics in LHb neurons suppresses cocaine-seeking.

a, Schematic of brain reinforcement circuits engineered to produce artificial feedback through cocaine-gated excitatory channels that oppose the natural suppression of LHb activity resulting from elevated dopamine (DA). Blue, dopamine; red, glutamate; green, GABA. DSt, dorsal striatum; NAc, nucleus accumbens. The diagram was created in BioRender. Sternson, S. (2025) https://BioRender.com/wrre5mw. b, Schematic of dose–response curves for drug self-administration and the potential effects of interfering with reinforcement mechanisms. c, Organization of the cocaine self-administration experiment. inf., infusion. d, Left, image of a coronal rat brain section expressing coca-5HT3-IRES-mCherry in the LHb (scale bar, 1 mm). Right, magnified image of the region indicated by the dashed box (scale bar, 200 µm). e,f, Acquisition of food (e; number of pellets per 3 h) and cocaine infusions per 3 h (f; 0.5 mg per kg per infusion) in rats with coca-5HT3 in the LHb (n = 11 rats) and controls (n = 15 rats). g,h, Rats with coca-5HT3 in the LHb (n = 11 rats) showed significantly lower cocaine infusions per 3 h at 0.125 mg per kg and 0.25 mg per kg (g; P = 0.01 and P = 0.04, respectively, Holm–Šidák post hoc comparison) and intake per 3 h at doses that maintain high-effort self-administration compared with the controls (h; n = 15 rats; mixed-effects ANOVA, dose × transgene interaction, P = 0.02). *P < 0.05. Data are mean ± s.e.m. Source Data

Fig. 4. Cocaine chemogenetics in LHb neurons suppresses cocaine-induced increases in extracellular dopamine in the NAc.

a, Schematic showing the stereotaxic delivery of AAV5-Syn::coca-5HT3-IRES-mCherry into the LHb of rats and [¹⁸F]fallypride PET procedures. The diagram was created in BioRender. Michaelides, M. (2025) https://BioRender.com/yvwycvv. b, Schematic showing the displacement of [¹⁸F]fallypride by cocaine detected using PET. The diagram was created in BioRender. Sternson, S. (2025) https://BioRender.com/9fog7n9. c, Coronal PET brain images co-registered to a magnetic resonance imaging template showing significant cocaine-induced voxel-wise decreases in [¹⁸F]fallypride binding in the ventral striatum of control rats (n = 9 rats) and in rats with coca-5HT3 in the LHb (n = 5 rats). d, Time–activity curves of [¹⁸F]fallypride binding (standardized uptake value (SUV)) in the ventral striatum (squares) and cerebellum (triangles) calculated as the percentage of the baseline (60–90 min) for control (n = 9 rats, black) and coca-5HT3 (n = 5 rats, red) rats, showing that cocaine selectively displaces [¹⁸F]fallypride binding (that is, increases dopamine) in the ventral striatum of control rats (three-way RM-ANOVA, time × transgene × brain region interaction, F_25,300 = 3.83, P < 0.001). e, Schematic of the stereotaxic delivery of AAV5-Syn::coca-5HT3-IRES-mCherry and AAV9-Syn::GRAB_DA and the experimental timeline of fibre photometry procedures. The diagram was created in BioRender. Michaelides, M. (2025) https://BioRender.com/hmnai5g. f, Left, image of a coronal rat brain section showing the optic fibre tract and GRAB_DA expression (scale bar, 1 mm). Right, magnification of the area indicated on the left (scale bar, 400 µm). g,h, GRAB_DA response time course with saline and cocaine (5 mg per kg) injection in control rats (g) (n = 5 rats; two-way RM-ANOVA, treatment main effect, F_1,5 = 47.77, P = 0.001; time main effect, F_2.2,11.1 = 1.84, P = 0.2; treatment × time interaction, F_2.1,10.4 = 10.86, P = 0.002) and rats with coca-5HT3 in the LHb (h) (n = 6 rats; two-way RM-ANOVA, treatment main effect, F_1,4 = 3.58, P = 0.13; time main effect, F_1.5,6.1 = 0.73, P = 0.48; treatment × time interaction, F_2.5,10.2 = 2.03, P = 0.17). i, The AUC of time-course measurements with saline- and cocaine-induced (5, 10 and 20 mg per kg) GRAB_DA responses. Data are mean ± s.e.m. Source Data

Extended Data Fig. 1. Chimeric ion channels that are activated by cocaine.

a, Chimeric ion channels were created from the ligand binding domain (LBD) of the α7 nAChR that was spliced to either the IPD of 5HT3 or GlyR and produced channels for neuron activation or inhibition, respectively. Mutated chimeric channel subunits homopentamerize to form LGICs. Mutations in the LBD conferred novel cocaine agonism and reduced ACh sensitivity. Created in BioRender. Sternson, S. (2025) https://BioRender.com/2uq2ikb. b, Structural elements that correspond to the nicotine pharmacophore are highlighted in red. Acetoxymethyl ester portion of cocaine that is proposed to adversely affect cocaine agonism of the unmodified α7 nAChR is highlighted in blue. Structural relationship to tropisetron, an α7 nAChR agonist, is shown on the right. c, Cocaine binding to AChBP (modified from PDB: 2pgz). Two views of cocaine (yellow) bound to AChBP at the interface of two protomer subunits (cyan & pink) showing 4 amino acid side chains in grey that are from homologous α7 nAChR human sequence and are mutated in coca-5HT3. Homologous α7 nAChR pre-protein numbering is shown.

Extended Data Fig. 2. Activity of choline, cocaine, metabolites, and other drugs on coca-5HT3.

a, Dose response of choline on coca-5HT3 chimeric ion channel (n = 4 separate dose responses). b, Dose response for cocaine and two cocaine metabolites on coca-5HT3 (n = 1 dose response each). c, Dose response for cocaine and other amine-containing addictive drugs on coca-5HT3 (n = 1 dose response each). d, Comparison of dose responses for nicotine on unmodified α7-5HT3 (n = 6 separate dose responses) and coca-5HT3 (n = 3 separate dose responses). Data is shown as mean ± sem. Source Data

Extended Data Fig. 3. Activity of cocaine and physiologically relevant amines on coca-5HT3.

a, Endogenous amines (n = 1 dose response per amine). b, Clinically used drugs with structural similarity to cocaine and two common anaesthetics (cocaine, n = 5 separate dose responses; benztropine, n = 3 separate dose responses; methylatropine, n = 3 separate dose responses; all other drugs, n = 1 dose response). Data is shown as mean ± sem. Source Data

Extended Data Fig. 4. Competitive binding of cocaine and ASEM against [³H]ASEM binding in membranes from cells expressing coca-GlyR.

Dose-response curves were fitted using a two-site binding model. Data are shown as mean (n = 3 replicate curves for ASEM and n = 2 replicate curves for cocaine). Source Data

Extended Data Fig. 5. Membrane properties of LHb neurons expressing coca-5HT3.

LHb neurons from mice expressing coca-5HT3-IRES-mCherry (n = 16 cells) or negative control mCherry (n = 25 cells) showing a, Capacitance (Shapiro Wilk normality test P < 0.05, two-sided Mann-Whitney Rank Sum Test), b, Membrane Resistance (Shapiro Wilk normality test P < 0.05, two-sided Mann-Whitney Rank Sum Test), c, and Resting Membrane Potential (Shapiro Wilk normality test P = 0.97, two-sided t-test). Data is shown as mean ± sem, statistical tests are two-sided. Source Data

Extended Data Fig. 6. Control experiments to validate lack of motivation effects of coca-5HT3 LHb expression in the absence of cocaine.

a, Sucrose (1%) drinking preference in rats with coca-5HT3 expression in LHb (n = 6 rats) and in controls (n = 6 rats). b, Total sucrose intake (g/kg) (two-way RM ANOVA; transgene main effect (F(1, 11) = 0.18, p = 0.68), session main effect (F(5, 55) = 13.41, p < 0.001), transgene × session interaction (F(5, 55) = 0.61, p = 0.69)), (c) active lever presses (two-way RM ANOVA; transgene main effect (F(1, 11) = 0.24, p = 0.63), session main effect (F(5, 55) = 10.37, p < 0.001), transgene × session interaction (F(5, 55) = 0.72, p = 0.61)), and (d) number of sucrose pellets delivered during operant experiments in rats with coca-5HT3 expression in LHb (n = 5 rats) and in controls (n = 8 rats). All experiments are in the absence of cocaine. e,f, Open-field locomotor activity in rats with coca-5HT3 expression in LHb (n = 6 rats) and in controls (n = 6 rats) showing (e) total distance travelled over the 1 h session (two-sided unpaired t-test, t = 0.21; p = 0.84) and (f) within session time course of distance travelled. Experiment is in the absence of cocaine. Data is shown as mean ± sem. Source Data

Extended Data Fig. 7. Dopamine levels for cocaine treatment and at baseline.

a, Time-activity curves of [¹⁸F]fallypride binding (SUV; standardized uptake value) in the ventral striatum and cerebellum for control (n = 9 rats) and coca-5HT3 expressing rats (n = 9 rats). b, Baseline striatal dopamine concentrations in rats with coca-5HT3 expression in LHb (n = 6 rats) and in controls (n = 6 rats) in the absence of cocaine (unpaired t-test, P = 0.31). Data is shown as mean ± sem. Source Data

Extended Data Fig. 8. Inhibitory cocaine chemogenetics in VTA dopamine neurons suppresses cocaine-induced increases in extracellular dopamine in the NAc.

a, Schematic showing the stereotaxic delivery of AAV1-Syn::FLEX-Coca-GlyR-IRES-mCherry into the VTA of TH-cre rats and [¹⁸F]fallypride PET procedures. Created in BioRender. Michaelides, M. (2025) https://BioRender.com/v8hpd3n. b, Coronal PET brain images co-registered to MRI showing significant cocaine-induced voxel-wise decreases in [¹⁸F]fallypride binding in the ventral striatum of control rats (n = 6 rats) and in rats with coca-GlyR in VTA dopamine neurons (n = 9 rats). c, Time-activity curves of [¹⁸F]fallypride binding (SUV; standardized uptake value) in the ventral striatum (squares) and cerebellum (triangles) calculated as the percent of baseline (60–90 min) for control (n = 6 rats, black) and coca-GlyR rats (n = 9 rats, orange) showing that cocaine selectively displaces [¹⁸F]fallypride binding (i.e., increases dopamine) in the ventral striatum of control rats. Data is shown as mean ± sem. Source Data

Extended Data Fig. 9. Influence of LHb coca-5HT3 expression on cocaine-induced dopamine levels in NAc.

a-d, Time course (z-score) of GRAB_DA response in control rats treated with saline or cocaine (a, c) (10 mg/kg: two-way RM ANOVA; treatment main effect (F(1, 5) = 32.2, p = 0.002), time main effect (F(2.6, 13.1) = 4.56, p = 0.2), treatment × time interaction (F(1.7, 8.8) = 21.45, p = 0.005, 20 mg/kg: two-way RM ANOVA; treatment main effect (F(1, 5) = 13.61, p = 0.014), time main effect (F(1.4, 7.1) = 2.3, p = 0.17), treatment × time interaction (F(1.4, 7.2) = 5.09, p = 0.04) and coca-5HT3 rats treated with saline or cocaine (b, d) (10 mg/kg: two-way RM ANOVA; treatment main effect (F(1, 4) = 5.05, p = 0.08), time main effect (F(1.4, 5.6) = 1.12, p = 0.36), treatment × time interaction (F(1.1, 8.8) = 1.24, p = 0.33, 20 mg/kg: two-way RM ANOVA; treatment main effect (F(1, 4) = 0.41, p = 0.55), time main effect (F(2.4, 9.7) = 0.52, p = 0.63), treatment × time interaction (F(1.3, 5.4) = 0.37, p = 0.63). GRAB_DA response in the medial nucleus accumbens shell measured using fibre photometry in rats with coca-5HT3 in the LHb (n = 5 rats) and in controls (n = 6 rats). Data is shown as mean ± sem. Source Data

Extended Data Fig. 10

a,b, Raw photometry data from a representative control (a) and LHb-coca-5HT3 rat (b) in response to saline and 5 mg/kg cocaine injections. The blue trace shows the 465 nm signal (dopamine), and the red trace shows the 405 nm isosbestic channel.