KOR, D1/D2 Heteromers, and THC
This is another very long post, but this is my reasoning for why I feel like KOR antagonism is a great solution to helping THC users quit. This is a bit disorganized, but there is a TLDR at the end if you want to skip to that.
KOR on DA Terminals and D1–D2 Heteromers
https://pmc.ncbi.nlm.nih.gov/articles/PMC6596815/
https://pmc.ncbi.nlm.nih.gov/articles/PMC10871279/
https://pmc.ncbi.nlm.nih.gov/articles/PMC6061161/
So we know in the NAc and dorsal striatum, KORs are expressed on dopaminergic terminals supplying both D1 expressing and D2 expressing MSNs. And activation of KORs on these terminals inhibits DA release, reducing DA availability to stimulate D1R, D2R, and D1-D2 receptor heteromers in NAc. So dynorphins/KOR shift the local D1/D2 signaling balance to a hypodopaminergic/stress-vulnerable state. I also want to mention that KOR-mediated inhibition of DA is p38-independent.
https://pmc.ncbi.nlm.nih.gov/articles/PMC2978591/
So this paper shows in the NAc and broader striatum, a distinct population of MSNs co express D1Rs, D2Rs, dynorphin, and enkephalin. They also so that 100% of D1R/D2R co-expressing neurons also express DYN or ENK, so you have this unique D1-D2-DYN/ENK MSN sub pop. And within these neurons you have D1-D2 heteromers. In NAc, 91% of D1/D2 co expressing neurons show D1-D2 heteromer formation vs only 24% in the caudate putamen. So we have a region based enrichment of heteromers in NAc.
https://www.frontiersin.org/journals/neuroanatomy/articles/10.3389/fnana.2011.00031/full
We also know that D1-D2 heteromers co-localize with the presynaptic marker but not with the postsynaptic density marker. So this indicates a presynaptic heteromer pool, while D1/D2 homers can be found both pre and post synaptically. This presynaptic localization is observed on both local MSN collaterals and DYN/ENK projections into the GP and VP, so conceptually you have this presynaptic D1-D2 DYN/ENK network that can directly regulate transmitter release and synaptic integration across basal ganglia loops.
The D1–D2–DYN/ENK MSN Subpopulation and Heteromer Signaling
https://www.pnas.org/doi/10.1073/pnas.0604049104
We also know from this paper that D1-D2 heteromers are Gq/11 coupled. So the ligand they used in the paper showed a strong preference for activating D1-D2 heteromers over D1 or D2 homers, driving Gq/11 -> PLC -> IP3 -> Ca2+ signaling and CaMKIIa activation in the NAc. This is pretty clean because the Gq-linked CaMKIIa increase is not seen in D1-/- or D2-/- mice, and is locked by either a D1 or D2 antagonist, so you need simultaneous engagement of both receptors within the heteromer. We also know that repeated activation of D1-D2 heteromers has distinct plasticity like consequences in the NAc. So the pattern they showed with repeated activation was reduced total CaMKII expression in NAc and selective decrease of CaMKII-dependent GluR1-Ser831 phosphorylation, and not any noticeable effect on PKA-dependent GluR1-Ser831 phosphorylation. So this is pretty consistent with an inhibitory effect on CaMKII regulated AMPAR transmission enriched in the NAc D1-D2-DYN/ENK compartment, where D1-D2 heteromer density is the highest. This kinda supports the idea that repeated D1-D2 heteromer activation turns down AMPAR mediated excitatory drive onto these MSNs.The timescale here is pretty interesting too, acute heteromer activation upregulates CaMKIIa, and chronic activation triggers CaMKII downscaling.
The first paper also shows that the D1-D2-DYN/ENK system is highly plastic is hyperdopaminergic states. So repeated amphetamine exposure increases the proportion of D2 receptors in the high affinity D2high state in the striatum and specifically boosted the D2high fraction within the D1-D2 heteromer. It enhances the D1-D2 marker efficacy in the NAc but not the caudate putamen, so indicating increased D1-D2 heteromer density selectively in NAc cell bodies. In schizophrenic GPs, there was a 10x increase in affinity at D2R within the D1-D2 heteromer and a higher faction of heteromeric D2high. So you have this model where the heteromer is sensitized both in number and in signaling state by repeated DA elevation and in schizophrenia, and continued DA inputs create larger Gq/Ca2+ responses in D1-D2-DYN/ENK MSNs.
So you kinda have this picture of the D1-D2 heteromer as a Ca2+/CaMKII gate embedded specifically within DYN/ENK MSNs, and their gain and sensitivity is dynamically modulated by dopaminergic history and whatnot.
KOR–GRK3/β‑Arrestin–p38 and the Cdk5/Thr75 DARPP‑32 Brake
https://pubmed.ncbi.nlm.nih.gov/29354053/
We also have some data showing that activation of the D1–D2 heteromer increases pThr75-DARPP-32 and decreases cocaine-induced pThr34 in NAc D1-ENK neurons (heteromer-bearing), while sparing D1-only and ENK-only (D2-only) neurons, so shifting DARPP-32 into a PKA-inhibitory state that attenuates cocaine reward and pERK signaling. Functionally, that Thr75‑biased DARPP‑32 state in D1‑ENK heteromer‑bearing neurons is what mediates the heteromer’s ability to blunt cocaine reward and pERK signaling.
https://pmc.ncbi.nlm.nih.gov/articles/PMC2096730/
So we know KOR is expressed on striatal neurons and astrocytes and can recruit a B arrestin/GRK3-dependent p38 MAPK pathway. Phosphorylation of KOR Ser369 by GRK3 promotes B arrestin binding, receptor internalization, and assembly of the p38 MAPK signaling module. This paper essentially shows that dynorphin-KOR activation of p38 in NAc, cortex, and hippocampus is GRK3 dependent and required for stress induced immobility, so basically arguing that KOR-p38 coupling in the brain is explicitly a GRK3/B arrestin dependent stress-biased pathway
https://pmc.ncbi.nlm.nih.gov/articles/PMC2481272/
https://pmc.ncbi.nlm.nih.gov/articles/PMC4571610/
We also know that in VTA neurons, this pathway is directly linked to aversive behavior. KOR mediated inhibition of evoked DA release in NAc is preserved in p38 KO mice. This also shows that local norBNI in VTA and selective viral rescue rescue of WT KOR, but not a “p38-blind” KOR that still signals via Gi, in VTA DA neurons of KOR-null mice restores CPA. So I basically interpreted this as disscoiating presynaptic DA inhibition from aversive effects, and showing that KOR-Barrestin/GRK3-p38 signaling in DA somata, not just the GI coupled terminal inhibition, is required for the dysphoric component.
https://www.sciencedirect.com/science/article/pii/S0021925820379345
We know that KOR-p38 signaling in VTA DA neurons modifies GIRK channel kinetics and neuronal excitability. Repeated KOR activation increases tyrosine phosphorylation of Kir3.1 specifically in VTA DA neurons. And this adaptation is p38-dependent, so p38 KO mice failed to show the phosphorylation and attenuation of KOR responses with repeated exposure. So KOR-Barrestin-GRK3-p38 in DA neurons is a sort of anti-reward function that encodes aversive learning and stress dysphoria, and this is functionally distinct from p38-independent Gi inhibition of DA terminals.
https://pubmed.ncbi.nlm.nih.gov/10604473/
https://pmc.ncbi.nlm.nih.gov/articles/PMC547862/
https://pmc.ncbi.nlm.nih.gov/articles/PMC4409246/
https://pubmed.ncbi.nlm.nih.gov/17209049/
https://www.pnas.org/doi/10.1073/pnas.0308652100
https://journals.plos.org/ploscompbiol/article?id=10.1371%2Fjournal.pcbi.0020176
So we also know that p38 can act upstream of Cdk5 in neurons. In the striatum Cdk5 is a key modulator of DARPP-32, a central integrator of DA and glutamate signals. PKA phosphorylates DARPP-32 at Thr34, enabling DARPP-32 to inhibit PP1 and amplify D1R evoked signaling, and Cdk5 phosphorylates DARPP-32 at Thr75, which makes it a competitive inhibitor of PKA. Thr75-phosphorylated DARPP-32 suppresses D1R induced PKA activity, reduces phosphorylation of PKA substrates, and diminishes DA evoked ERK activation, so basically a PKA brake.
https://www.sciencedirect.com/science/article/pii/S0006322319301519
We also know that psychostimulants and stuff that enhances Cdk5 activity elevates pThr75-DARPP-32 in the NAc and dorsal striatum, with a decrease in Thr34 and attenuated D1R/PKA-dependent responses. Increased Cdk5/p35 activity after chronic cocaine or p35 overexpression produces higher pThr75, lower pThr34, lower ERK signaling, and reduced behavioral responsiveness to DA agonists, whereas Cdk5 inhibition or Thr75Ala mutation augments D1R signaling and shows antidepressant‑like effects. Activation of the D1–D2 heteromer in NAc has itself been shown to increase pThr75‑DARPP‑32 and reduce pThr34, directly biasing DARPP‑32 toward a D1‑antagonistic configuration
So my theory is that sustained KOR-p38 activation in striatal MSNs enhances Cdk5/p35 activity through known p38->p35 interactions, and promotes DARPP-32 Thr75 phosphorylation, further suppressing D1R/PKA gain on top of the D1-D2 heteromer’s own tendency to increase Thr75. So under this concept KOR-p38 signaling would be another intrinsic PKA brake that biases plasticity away from D1 dependent signaling and towards an anti-reward state. So we have a concept like: KOR activation -> p38 -> p35/Cdk5 -> increased DARPP-32 Thr75, which hasn’t been directly demonstrated in striatal MSNs but the individual links are established, so just my idea.
https://pubmed.ncbi.nlm.nih.gov/39190549/
https://www.nature.com/articles/1300604
Aside from that, this study is pretty interesting. So we know activation of the D1-D2 heteromer in NAc engages a Ca2+ CaMKIIa cascade that drives BDNF mRNA and protein up regulation, and this heteromer induced BDNF increase contributes to structural and synaptic remodeling in NAc. We also have some literature showing Cdk5 and DARPP-32 states shape BDNF driven adaptations, with altered Cdk5/Thr75 signaling modifying how DA inputs translate into BDNF dependent plasticity and mood related behavior.
But this study shows that sustained KOR activation epigenetically suppresses BDNF expression. Chronic KOR agonist treatment reduces total BDNF mRNA in cortical and limbic regions by decreasing permissive histone acetylation and increasing repressive marks at BDNF promoters, effects which are prevented by KOR antagonists.
So my thinking is a dynorphin high/constant KOR activation state, D1-D2 heteromer linked CaMKIIa signaling in D1-D2-DYN/ENK MSNs may drive BDNF transcription, while KOR-p38 clamps BDNF output, so selectively blocking the heteromer’s plasticity arm even as its Cdk5/Thr75-DARPP-32 anti reward arm remains active.
So putting all this together, what are KOR antagonists expected to do to the D1-D2 heteromer?
We have the simple answer of preventing inhibition of DA release in NAc and dorsal striatum, increasing phasic and tonic DA available to stimulate D1R, D2R, and D1-D2 heteromers. So your DA bursts can recruit both canonical D1R-Gs/AC/PKA signaling and heteromer-Gq/Ca2+ signaling in a more dynamic task dependent manner.
And by preventing KOR-GRK3/Barrestin activation, KOR antagonists block KOR driven p38 activation in DA neurons and striatal cells and would be expected to reduce any p38-dependent drive onto Cdk5/p35 and DARPP-32 Thr75. So chronic KOR blockade should shift the DARPP‑32 phosphorylation balance away from Thr75 and back toward the D1‑dependent Thr34 state, removing PKA inhibition and restoring the D1R–cAMP/PKA/Thr34‑DARPP‑32/ERK reward cascade. So in D1-D2 heteromer bearing MSNs, this would move the system from a globally PKA suppressed state(from heteromer driven Thr75 and KOR-p38 driven Thr75) to one where PKA and CaMKII signals can coexist, so the heteromer’s Gq output can function as an actual integrator of DA tone rather than purely anti-reward. We have some data showing chronic KOR antagonism in NAc not only improves depressive‑like behaviors but also normalizes BDNF, pERK, and synaptic markers (PSD95, synaptophysin) in withdrawal models, which fits with the idea that it lifts both the Cdk5/Thr75 brake and the BDNF clamp
https://pmc.ncbi.nlm.nih.gov/articles/PMC10448166/
Over longer timelines, chronic KOR antagonism would prevent recurrent KOR-p38 stress signaling. So if KOR tone normally constrains the expression of D1-D2 heteromer linked CaMKII BDNF programs, sustained KOR blockade would be expected to allow for more adaptive heteromer driven plasticity, where CaMKII and BDNF support controlled remodeling rather than the anhedonic circuit. So within this idea, the amount of behavioral improvements with KOR antagonists should scale with the pre-existing “anti-reward” features: so a combination of high KOR tone, elevated pThr75-DARPP-32, high D1D2 heteromer expression, stress induced p38 activation, etc.
Chronic THC, D1–D2 Heteromers, Dynorphin/KOR, and Why KOR Antagonists Help
https://pmc.ncbi.nlm.nih.gov/articles/PMC6971351/
https://pmc.ncbi.nlm.nih.gov/articles/PMC6763924/
https://pubmed.ncbi.nlm.nih.gov/37519471/
So one of the biggest reasons I started thinking about KOR antagonism in the context of D1-D2 heteromers is because of literature showing that chronic THC use greatly increases the number of D1-D2 heteromer expressing neurons in the NAc by 50%, and increased heteromer density per neuron 3x, with parallel increases seen in the caudate. CBD co administration blocked many of these effects. The same THC treated animals showed increased pCaMKIIa in NAc, increased pThr75-DARPP-32 in NAc, and decreased pThr34-DARPP-32, so basically all the heteromer anti-reward signals. Also worth noting is that pTrkB increased without a commensurate increase in BDNF protein, so BDNF receptor signaling is engaged but the downstream output is being constrained somewhere. The THC treated MSNs also showed a reprogrammed coexpression of D1/D2/SP/A2AR, which looks really similar to the D1-D2-DYN/ENK sub pop I talked about earlier.
Following this, they directly tested the consequences of daily THC plus spontaneous withdrawal. They saw chronic THC induced anhedonia and anxiogenic like behaviors that persisted through withdrawal. In the NAc, THC increased D1-D2 heteromer expression and BDNF/TrkB activity, and these changes were accompanied by increased dynorphin expression and KOR. Also, disruption of the D1-D2 heteromer reversed both the behavioral changes and the dynorphin/KOR neuroadaptations. So you have this really direct loop: THC -> increased D1-D2 heteromer -> increased dynorphin -> increased KOR activation -> anhedonia and anxiety.
So integrating all of this with the model I’ve been discussing, you get a compounding set of anti-reward adaptations from chronic THC use in NAc D1-D2-DYN/ENK MSNs that map really well onto the KOR-p38-Cdk5/Thr75-BDNF framework. So during THC withdrawal you have a circuit locked into high dynorphin/KOR tone, elevated Thr75-DARPP-32, suppressed D1R/PKA/ERK signaling, and constrained BDNF output, all reinforcing each other to maintain anhedonia and anxiety. KOR antagonism seems extremely well suited for this because it targets the overlap between these adaptations. You would simultaneously:
- Relieve dynorphin mediated DA inhibition at terminals
- Prevent continued KOR-p38 driven Cdk5/Thr75 loading in the MSNs
- Lift the BDNF restriction
And allow the heteromers Ca2+/CaMKII/BDNF arm to emerge as adaptive plasticity. And this benefit would scale with heteromer density, the depth of the Thr75 change, and the duration of dynorphin/KOR engagement. So it seems like heavy long term THC users would stand to benefit the most from antagonism, especially when trying to quit.
TLDR: Chronic THC use and withdrawal drive a D1–D2 heteromer–dynorphin/KOR–Cdk5/Thr75‑DARPP‑32 anti‑reward state in NAc that causes anhedonia and anxiety, especially in heavy users. KOR antagonism is well suited here because it can restore DA release, relieve the Cdk5/Thr75 PKA brake, and lift BDNF repression, allowing the D1–D2 heteromer’s Ca²⁺/CaMKII/BDNF arm to support adaptive plasticity instead of persistent dysphoria.