Hallucinogens recruit specific cortical 5-HT(2A) receptor-mediated signaling pathways to affect behavior

Hallucinogenic compounds such as psilocybin, mescaline and LSD profoundly alter perception, emotion and cognition and share high affinity for the serotonin 5-HT2A receptor (2AR). Genetic or pharmacological blockade of 2AR prevents hallucinogen effects in multiple species, yet a key paradox remains: some compounds that are 2AR agonists, for example lisuride and ergotamine, lack hallucinogenic activity in humans. The neuronal substrates and signalling mechanisms that explain why only a subset of 2AR agonists produce hallucinogenic effects have therefore remained unclear. González-Maeso and colleagues set out to identify the cellular and molecular basis for this distinction. They compared hallucinogenic (HC) and nonhallucinogenic (NHC) 2AR agonists across behavioural, transcriptional, electrophysiological and pharmacological assays in mice, and used a genetic rescue strategy to restore 2AR expression selectively to cortical neurons of 2AR-deficient (htr2A -/-) mice. The study aims to determine whether distinct 2AR-mediated signalling pathways in cortical neurons account for hallucinogen-specific effects and whether cortical 2AR expression is sufficient to restore those effects.

This experimental study used adult male 129S6/SvEv mice, including wild-type (htr2A +/+), heterozygous (htr2A +/−) and knockout (htr2A −/−) lines. To test the sufficiency of cortical 2AR, the investigators restored receptor expression selectively to forebrain glutamatergic neurons by crossing htr2A −/− mice carrying a floxed transcriptional stop cassette with Emx1-Cre mice, yielding htr2A −/−:Emx1-Cre +/− animals with cortical 2AR expression. Autoradiography and radioligand binding ([125I]DOI, [3H]ketanserin and [125I]LSD assays) were used to confirm regional 2AR expression. A panel of hallucinogenic and nonhallucinogenic 2AR agonists was administered intraperitoneally at reported doses (examples: LSD 0.24 mg/kg, R-lisuride 0.4 mg/kg, DOI 2 mg/kg, mescaline 20 mg/kg). Behavioural assays included the head twitch response (HTR), ear scratch response (ESR), locomotion, grooming, rearing and body temperature; HTR and ESR were evaluated 15 min after injection for 20 min. For transcriptome profiling, cortex was collected 60 min after injection for quantitative real-time PCR (qRT-PCR) measuring 19 transcripts previously identified as 2AR-responsive; fold-changes were log2-transformed and analysed by principal components analysis (PCA). Primary cortical cultures (E15–17) provided an in vitro system comprising 90%–95% neurons; cultures were used for fluorescent in situ hybridisation (FISH) and qRT-PCR. Pharmacological inhibitors included the PLCβ inhibitor U73122, pertussis toxin (PTX) to inactivate Gi/o proteins, the Src-family kinase inhibitor PP2 and the PI3-kinase inhibitor LY294002. Electrophysiology employed whole-cell patch-clamp recordings from layer V pyramidal neurons in somatosensory cortex slices (P10–20 mice); voltage ramps (−110 to −30 mV) were used to examine drug-induced changes in voltage-gated currents in the presence of TTX. FISH localised induction of transcripts to neurons expressing 2AR mRNA. Statistical analyses included two-factor ANOVA with Bonferroni post hoc tests where reported; some group sizes are noted in the text and figures (e.g. n = 4–8 per qRT-PCR group, n = 4–6 per behavioural group, n = 16 cells for some electrophysiology experiments).

Behavioural screening identified the head twitch response (HTR) and ear scratch response (ESR) as measures that distinguished hallucinogenic from nonhallucinogenic 2AR agonists in mice. HTR was reliably and robustly elicited by all hallucinogens tested (DOI, DOM, DOB, psilocin, mescaline, LSD) in htr2A +/+ mice and was absent in htr2A −/− mice. Nonhallucinogenic 2AR agonists (ergotamine, R- and S-lisuride) failed to produce HTR in either genotype. ESR was less consistent and not normally distributed across animals. Transcriptional profiling of 19 2AR-regulated transcripts in somatosensory cortex produced a reproducible ‘‘HC signature’’ that separated hallucinogens from nonhallucinogens by PCA. Induction of c-fos correlated with 2AR agonist activity generally, whereas egr-1 and egr-2 induction most robustly predicted hallucinogenic behavioural activity. The HC transcriptional signature was absent in htr2A −/− mice and present in wild-type animals and in primary cortical neurons. FISH indicated that both LSD and R-lisuride induced c-fos in 2AR-expressing cortical neurons (layers II/III and V), but only LSD induced egr-2 in layer V 2AR-positive neurons; R-lisuride did not induce egr-2 in vivo or in cultured 2AR-expressing neurons. Tetrodotoxin (TTX) blockade of action potentials in primary culture did not prevent induction of egr-1 and egr-2, indicating that these transcriptional responses are intrinsic to 2AR-expressing neurons rather than secondary to circuit activity. Dose–response experiments showed that R-lisuride induced c-fos over a range of doses both in vitro (1–10 μM) and in vivo (0.4–0.8 mg/kg) but failed to induce egr-1 or egr-2 even at doses up to 10-fold higher than those that elicited c-fos. In contrast, LSD induced c-fos, egr-1 and egr-2 across comparable in vitro and in vivo dose ranges; transcript induction was absent in htr2A −/− samples. Electrophysiological recordings from layer V pyramidal neurons revealed heterogeneous drug effects on voltage-gated currents. Small and medium current changes occurred in neurons from both genotypes, indicating 2AR-independent effects. A distinctive large current response (average charge ~334 ± 97 pC, reversal potential about −67.5 ± 2.4 mV) was seen in ~25% of neurons from htr2A +/+ mice treated with LSD but was never observed with R-lisuride or in htr2A −/− tissue. Binding studies indicated that R-lisuride competes with [125I]LSD for the same 2AR binding site. Functionally, pretreatment with R-lisuride blocked LSD-induced large currents, prevented LSD-mediated induction of egr-1 and egr-2 while leaving c-fos induction intact, and abolished LSD-induced HTR. Co-administration experiments argued against dopamine receptor involvement: D1-like (SKF82958) and D2-like (quinpirole) agonists did not induce egr-1/egr-2 or affect LSD-induced HTR, and coadministration with LSD did not reproduce the inhibitory interaction seen with R-lisuride. Knockout of 5-HT1A receptors also did not enhance hallucinogen responses, arguing that 5-HT1A coactivation cannot explain lisuride’s nonhallucinogenic profile. Pharmacological dissection of signalling pathways showed that PLCβ inhibition (U73122) abolished gene responses to both LSD and R-lisuride, indicating reliance on classical Gq/11–PLCβ signalling. However, PTX (Gi/o blockade) greatly attenuated LSD-elicited transcript responses but not those induced by R-lisuride, implicating Gi/o proteins selectively in the HC response. Src-family kinase inhibition with PP2 converted the LSD-induced transcriptome to resemble that elicited by R-lisuride; PI3-kinase inhibition (LY294002) had no effect. Regional analyses found LSD induction of egr-1 and egr-2 in cortical areas and olfactory bulb but not in thalamus or cerebellum, whereas c-fos was activated more broadly. Together these data indicate that hallucinogen-specific signalling requires coactivation of Gi/o and Src in addition to Gq/11–PLCβ, and that cortical pyramidal neurons are the pertinent cellular substrate. Finally, genetic restoration of 2AR expression to cortical glutamatergic neurons in htr2A −/−:Emx1-Cre +/− mice rescued the LSD-induced HTR, LSD-dependent large electrophysiological currents and HC-specific induction of egr-1 and egr-2 to levels comparable to controls. R-lisuride remained inactive in the rescue animals, indicating that cortical 2AR expression is sufficient for hallucinogen-specific signalling and behaviour.

González-Maeso and colleagues interpret these findings to mean that hallucinogenic and nonhallucinogenic 2AR agonists act on the same population of cortical pyramidal neurons but stabilise distinct receptor conformations that engage different intracellular signalling cascades. Both ligand classes engage the canonical Gq/11–PLCβ pathway, as shown by blockade with U73122, yet hallucinogens uniquely recruit pertussis toxin-sensitive Gi/o proteins and Src-family kinases; Src inhibition converted the hallucinogen transcriptome to a nonhallucinogen-like pattern and PTX selectively attenuated hallucinogen-driven gene induction. The authors therefore propose a biased-agonism (‘‘agonist trafficking’’) model of 2AR function in which hallucinogens preferentially stabilise receptor states that couple to Gi/o and Src in addition to Gq/11, producing the characteristic transcriptional, physiological and behavioural fingerprint of hallucinogens. The investigators argue against several alternative explanations. Pharmacological and genetic experiments did not support a dominant role for dopamine receptors or 5-HT1A coactivation in explaining lisuride’s lack of hallucinogenic activity. Rescue of 2AR expression only in cortical glutamatergic neurons of htr2A −/− mice was sufficient to restore hallucinogen-specific signalling and HTR, whereas restoring 2AR selectively to thalamus did not, which argues that cortical (not thalamocortical terminal) 2AR populations mediate behavioural effects. The intrinsic nature of the cortical response was reinforced by primary-culture experiments showing HC-specific gene induction in the absence of action potentials or thalamic input. The authors acknowledge unresolved issues and limitations. Src inhibition produced a more complete suppression of the hallucinogen transcriptome than PTX, suggesting additional, PTX-insensitive mechanisms linking 2AR to Src may exist and remain to be identified. They also note prior discrepancies in cellular localisation studies—previous reports suggested c-fos induction in non-2AR-expressing cells, whereas their FISH-based approach found induction predominantly in 2AR-expressing neurons; methodological differences may account for this divergence. Finally, while cortical layer V pyramidal neurons are highlighted as plausible mediators of hallucinogen action because of their output/gating role, the precise downstream circuit-level consequences remain to be detailed. In sum, the authors conclude that the unique psychoactive profile of classical hallucinogens can be explained by biased 2AR signalling in cortical pyramidal neurons, a mechanism that may have broader relevance for understanding receptor-specific drug actions and for dissecting pharmacological mechanisms underlying neuropsychiatric disorders.