Distinct 5-HT receptor subtypes regulate claustrum excitability by serotonin and the psychedelic, DOI

Abstract

Recent evidence indicates that neuronal activity within the claustrum (CLA) may be central to cellular and behavioral responses to psychedelic hallucinogens. The CLA prominently innervates many cortical targets and displays exceptionally high levels of serotonin (5-HT) binding. However, the influence of serotonin receptors, prime targets of psychedelic drug action, on CLA activity remains unexplored. We characterize the CLA expression of all known 5-HT subtypes and contrast the effects of 5-HT and the psychedelic hallucinogen, 2,5-dimethoxy-4-iodoamphetamine (DOI), on excitability of cortical-projecting CLA neurons. We find that the CLA is particularly enriched with 5-HT2C receptors, expressed predominantly on glutamatergic neurons. Electrophysiological recordings from CLA neurons that project to the anterior cingulate cortex (ACC) indicate that application of 5-HT inhibits glutamate receptor-mediated excitatory postsynaptic currents (EPSCs). In contrast, application of DOI stimulates EPSCs. We find that the opposite effects of 5-HT and DOI on synaptic signaling can both be reversed by inhibition of the 5-HT2C, but not 5-HT2A, receptors. We identify specific 5-HT receptor subtypes as serotonergic regulators of the CLA excitability and argue against the canonical role of 5-HT2A in glutamatergic synapse response to psychedelics within the CLA-ACC circuit.

1. Introduction

Psychedelic drugs have shown promising results in preclinical (Ly, Greb et al. 2018, Berquist and Fantegrossi 2021, Benvenuti, Colombo et al. 2023, Ferri, de Novais et al. 2023, Torrado Pacheco, Olson et al. 2023) and clinical studies (Johnson, Garcia-Romeu et al. 2014, 2015, Griffiths, Johnson et al. 2016, Johnson, Garcia-Romeu et al. 2017, Garcia-Romeu, Davis et al. 2019, Carhart-Harris, Giribaldi et al. 2021, Davis, Barrett et al. 2021, Gukasyan, Davis et al. 2022) to treat a range of psychiatric disorders, including treatment-resistant depression, post-traumatic stress and substance use disorders. The behavioral effects of psychedelic drugs have long been attributed to serotonin 2A receptors (5-HT2ARs) (Gonzalez-Maeso, Weisstaub et al. 2007, Gresch, Barrett et al. 2007, Nichols 2016, Ly, Greb et al. 2018, Vollenweider and Preller 2020, Shao, Liao et al. 2021, Cameron, Benetatos et al. 2023), particularly those in the prefrontal cortex, hippocampus, midbrain and thalamus (Lopez-Gimenez and Gonzalez-Maeso 2018). However, psychedelic drugs can be expected to activate 5-HT2ARs in other brain regions and have been clearly shown to bind non-5-HT2A receptors (Peroutka 1986, Pierce and Peroutka 1989, McKenna, Repke et al. 1990, Deliganis, Pierce et al. 1991, Halberstadt and Geyer 2011, Cameron, Benetatos et al. 2023) or act via signaling pathways entirely outside of the 5-HT receptor family (Moliner, Girych et al. 2023). The heterogeneity of 5-HT receptors, with 14 currently identified subtypes and multiple subtype splice variants, characterized by region-specific expression patterns (Acevedo-Triana, Leon et al. 2017) adds substantial complexity to theories of neuronal substrates underlying psychedelic effects and warrants detailed examination of serotonin and psychedelic signaling within discrete brain circuits.

Autoradiography studies have found that the highest density of binding of radiolabeled serotonin receptor ligands, including psychedelic hallucinogens (Pazos, Cortés et al. 1985, McKenna and Saavedra 1987, Nichols 2016) is observed in the claustrum (CLA), a subcortical brain region, wedged between the external capsule and the insula. Supporting a role for CLA as central to effect of psychedelics, recent fMRI data in humans show altered CLA activity and functional connectivity after systemic administration of psilocybin (Barrett, Krimmel et al. 2020). The CLA is densely innervated by serotonin fiber afferents from the raphe nuclei (Zingg, Dong et al. 2018, Narikiyo, Mizuguchi et al. 2020, Wong, Nair et al. 2021), and expresses subtypes for 5-HT1, 5-HT2, and 5-HT3 receptors (Mengod, Nguyen et al. 1990, Gehlert, Gackenheimer et al. 1991, Wright, Seroogy et al. 1995). However, expression of other 5-HT receptors, their relative abundance, cell-type localization, and 5-HT receptor impact on neuronal excitability within the CLA remain completely unknown.

Neuronal cell bodies within the CLA are thought to comprise predominantly glutamatergic projection cells (~80–90%) while parvalbumin, vasoactive intestinal polypeptide and somatostatin interneurons make up ~15% of the neuronal population with some sub-regional specificity (Braak and Braak 1982, Mathur 2014, Graf, Nair et al. 2020, Takahashi, Kobayashi et al. 2023). A prominent anatomical feature of the CLA projections neurons is their hyperconnectivity to a large number of cortical structures (Mathur 2014, Kim, Matney et al. 2016, Atlan, Terem et al. 2017, Wang, Ng et al. 2017, Qadir, Stewart et al. 2022, McBride, Gandhi et al. 2023, Wang, Wang et al. 2023). The most extensive CLA afferents target the prefrontal and cingulate cortices (Chia, Silberberg et al. 2017, Wang, Ng et al. 2017, White, Cody et al. 2017, Zingg, Dong et al. 2018, Chia, Augustine et al. 2020, Jackson, Smith et al. 2020), sparking an enthusiastic and on-going debate about the possible role of the CLA as a hub for regulation of higher order cognitive functions (Edelstein and Denaro 2004, Crick and Koch 2005, Mathur 2014, Liaw and Augustine 2023). Evidence suggests that the CLA may influence cortical activity via specific projections to cortical interneurons (Jackson, Karnani et al. 2018). Indeed, claustrocortical circuits have been implicated in reward processing (Terem, Gonzales et al. 2020), impulsivity (Liu, Wu et al. 2019), attention salience (Goll, Atlan et al. 2015, Smith, Watson et al. 2019), anxiety (Niu, Kasai et al. 2022), pain signaling (Xu, Zhang et al. 2022), sleep (Renouard, Billwiller et al. 2015, Narikiyo, Mizuguchi et al. 2020, Norimoto, Fenk et al. 2020), and conscious awareness (Crick and Koch 2005, Smythies, Edelstein et al. 2014).

In this study, we used whole-cell patch-clamp electrophysiology to compare the effects of 5-HT and the psychedelic hallucinogen, DOI, on glutamatergic transmission and excitability of CLA neurons that project to the ACC (CLA-ACC neurons). We explored contributions of distinct 5-HT receptors by profiling cell type-specific expression of mRNA for thirteen 5-HT receptor subtypes and by pharmacological manipulations of the most abundant 5-HT receptor subtypes. Our study provides a comprehensive characterization of CLA-ACC neuron regulation by 5-HT and highlights a prominent role for 5-HT2C signaling as an unexpected mediator of CLA-ACC excitability during psychedelic hallucinogen exposure.

2. Methods

2.1. ANIMALS

Male and female Sprague-Dawley rats (Rattus norvegicus), weighing 200–250 g, were obtained from Taconic Laboratories. Animals were individually housed, with food and water available ad libitum in the home cage. A 12 h light/dark cycle was used with the lights on at 7 A.M. All experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Kentucky.

2.2. STEREOTAXIC INJECTIONS

Rats were anesthetized with isoflurane and CLA-ACC neurons were labeled by the retrograde AAV-hSyn-EGFP (Addgene #50465-AAVrg) injected bilaterally (2 μL/side) into the ACC at the following stereotaxic coordinates (in mm from Bregma): A/P: +0.3 M/L: ±0.9 D/V: −2.2) using a 2 μL Neuros syringe (Hamilton Company) at a rate of 0.2 μL/min. Animals recovered for 10–15 days before brains were extracted and sliced for electrophysiology experiments.

2.3. ELECTROPHYSIOLOGY

Brains were rapidly removed, and coronal slices (300 μm-thick) containing the CLA were cut using a vibratome (VT1200S; Leica Microsystems, Wetzlar, Germany) in an ice-cold aCSF cutting solution, containing the following (in mM): 93 NMDG, 2.5 KCl, 1.25 NaH₂PO₄, 30 NaHCO₃, 20 HEPES, 25 glucose, 5 Na-ascorbate, 2 thiourea, 3 Na-pyruvate, 10 MgSO₄, and 0.5 CaCl₂, 300–310 mOsm, pH 7.4 when continuously oxygenated with 95% O₂/5% CO₂. Slices were allowed to recover in the aCSF cutting solution at 34–36°C for 30 minutes, during which, increasing volumes of 2M NaCl (up to a total of 1 mL NaCl/37.5 mL aCSF) were added every 5 minutes as previously described (Ting, Lee et al. 2018). After recovery, the slices were transferred to a recording aCSF solution maintained at room temperature. Recording aCSF contained the following (in mM): 130 NaCl, 3 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, 10 glucose, 1 MgCl₂, and 2 CaCl₂, pH 7.2–7.4, when saturated with 95% O₂/5% CO₂. For electrophysiology recordings, recording pipettes were pulled from borosilicate glass capillaries (World Precision Instruments, Sarasota, FL) to a resistance of 4–7 MΩ when filled with the intracellular solution. The intracellular solution contained the following (in mM): 145 potassium gluconate, 2 MgCl₂, 2.5 KCl, 2.5 NaCl, 0.1 BAPTA, 10 HEPES, 2 Mg- ATP, and 0.5 GTP-Tris, pH 7.2–7.3 with KOH, osmolarity 280–290 mOsm. CLA-ACC neurons were viewed under an upright microscope (Olympus BX51WI) with infrared differential interference contrast optics and a 40x water-immersion objective. The recording chamber was continuously perfused (1–2 ml/min) with oxygenated recording aCSF warmed to 32 ± 1°C using an automatic temperature controller (Warner Instruments). CLA-ACC neurons were identified by eGFP fluorescence. To evaluate spontaneous excitatory postsynaptic currents (sEPSCs) CLA- ACC neurons were voltage-clamped at −70 mV. To evaluate RMP, rheobase, and action potential firing, the cells were current-clamped and 500 ms depolarizing current steps were applied every second in 10 pA increments. Bath-applied drugs were perfused into the recording chamber for 10 minutes before data collection. Some drugs were applied locally via the Y-tube perfusion system (Murase, Ryu et al. 1989) and the effects were measured within 3 minutes of drug application. All recordings were digitized at 20 kHz and lowpass-filtered at 2 kHz using a Digidata 1550B acquisition board (Molecular Devices, San Jose, CA) and pClamp11 software (Molecular Devices). Access resistance (10–30 MΩ) was monitored during recordings by injection of 10 mV hyperpolarizing pulses and data were discarded if access resistance changed >25% over the course of data collection.

2.4. QUANTITATIVE RT-PCR

Coronal brain slices were flash-frozen on dry ice and 2 mm punches were taken from slices containing the ACC and insula, while 1 mm punches were taken from the CLA (Disposable Biopsy Punch with Plunger, Integra Lifesciences, Princeton, NJ, USA). Punches from each area were pooled, so that one tissue sample contained punches from multiple slices from a single rat. Quantitative RT-PCR was performed as previously described (Cleck, Ecke et al. 2008). Briefly, RNA was isolated using the RNeasy Mini kit (Qiagen) and qPCR reactions were assembled using Thermo Scientific Maxima SYBR Green master mix along with 100nM primers (Integrated DNA Technologies, Coralville, IA). The mRNA levels were determined using the 2^−ΔΔCT method (Livak and Schmittgen 2001), and target genes were normalized to the housekeeping gene TATA-Binding Protein (TBP). A modified version of the 2^−ΔΔCT method was used to query RNA levels without creating a standard curve for each gene of interest. The equation below was used to estimate transcript levels of each gene of interest (GOI).

$$2\mathit{fold}\mathit{change}=2-\left({\mathrm{C}\mathrm{T}}_{\text{G}\mathrm{O}\mathrm{I}}-{\mathrm{C}\mathrm{T}}_{\mathrm{T}\mathrm{B}\mathrm{P}}\right)$$

The 2^−ΔΔCT method can exaggerate differences when normalizing to housekeeping genes with substantially different CT values. Therefore, in addition to TBP, Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was also included in each plate as another well-validated housekeeping gene. The average CT values of our GOIs were between 26.2 – 33.9 depending on GOI. TBP was chosen as a reference gene because it generated an average CT value of 27 across all plates, while the average CT value of GAPDH was 19.7. Forward and reverse primer sequences for each GOI are provided in the Supplemental Table 1.

2.5. RNASCOPE IN SITU HYBRIDIZATION

RNAscope experiments were conducted in mounted cryosections using the manufacturer’s detection kits, protocols, and previously published methods (Wang, Flanagan et al. 2012). Briefly, coronal sections were removed from the −80 °C cryostorage and immediately incubated in prechilled 4% PFA for 60 minutes at 4 °C. After incubation, the slides were rinsed twice in PBS and then dehydrated in increasing concentrations of ethanol (50%, 70%, and 100% ethanol for 5 min each) followed by an overnight incubation in 100% ethanol at −20 °C. The next day, RNAscope probes (ACD Bio, Newark, CA) were hybridized to the tissue slices in a hybridization oven for 2 hours, including both positive and negative controls for each individual probe. All probes were designed for rat as follows: Slc17a7 (solute-carrier family 17; 317001), Gad1-C2 (glutamate decarboxylase 1, 316401-C2), Htr1a-C3 (5-hydroxytryptamine receptor 1A, 404801-C3), Htr2a-C3 (5-hydroxytryptamine receptor 2A, 424551-C3), Htr2b-C3 (5-hydroxytryptamine receptor 2B, 426431-C3), Htr2c-C3 (5-hydroxytryptamine receptor 2C, 469321-C3). Following hybridization, the probes were amplified using the ACD Bio RNAscope kit and the fluorophores were introduced for transcript detection followed by DAPI staining. The slides were then coverslipped with ProLong^™ Gold Antifade mounting medium (Thermo Fisher) and allowed to dry overnight at room temperature in the dark. The next day, images were captured using 10X and 40X objectives on an Olympus FV3000 confocal microscope using the Fluoview software. All confocal images were acquired using identical imaging settings to quantify co-expression of 5-HT receptor mRNA with glutamatergic and GABAergic neuron markers and with the nuclear DAPI label. With each test slide, negative control slides were imaged at the same laser settings to confirm absence of non-specific signal in each laser channel. Images were acquired as ‘z-stacks’ and transformed into a 2-dimensional average-intensity projection in ImageJ for analysis.

2.6. DATA ANALYSIS AND STATISTICS

Cells from 3–5 animals were analyzed for electrophysiology experiments in each experimental condition. Amplitude and decay time of sEPSCs were computed from an average of 50 – 100 events in each cell. sEPSC decay time was expressed as a weighted double exponential fit to the decay phase of an average sEPSC as previously described (Ortinski, Lu et al. 2004). sEPSC charge transfer was calculated as a product of sEPSC amplitude and sEPSC decay time. sEPSC frequencies were analyzed from 20 s long trace segments in each cell. Rheobase was recorded as the magnitude of the smallest current in a series of 10 pA depolarizing current steps that induced an action potential spike. The firing rate was measured as the frequency of action potentials during each depolarizing current step. The RMP was calculated with no injected current in current-clamp configuration as the mean potential in a 1–5 second window. Membrane resistance and capacitance were determined from currents elicited by brief hyperpolarizing voltage pulses (−10 mV). RNAscope confocal images were analyzed by measuring the integrated density of mRNA puncta of interest as a fraction of integrated density of puncta contained within a mask of Slc17a7-, GAD1-, or DAPI-positive label. All analyses were completed using Clampfit 11.1 (Molecular Devices), ImageJ, and Microsoft Excel. Statistical comparisons were performed in Microsoft Excel or GraphPad Prism 8–10, using two-tailed paired or unpaired Student’s t-tests, 1-way ANOVA, or 2-way ANOVAs as indicated. All data were expressed as mean ± SEM.

3. Results

3.1. 5-HT decreases synaptic and membrane excitability of CLA-ACC neurons

To determine the effects of 5-HT on CLA-ACC neuron physiology, retrogradely labeled, eGFP-positive, CLA-ACC neurons (Figure 1A) were targeted for recordings of spontaneous excitatory post-synaptic currents (sEPSCs) before and during bath-application of 5-HT (30 μM). We found that application of 5-HT decreased sEPSC amplitude (t₁₁ = 6.401, p <0.0001, paired Student’s t-test) (Figure 1B–D), but did not change sEPSC decay time (t₁₁ = 0.090_, p = 0.9303, paired Student’s t-test) (Figure 1E). The resulting sEPSC charge transfer was significantly decreased in the presence of 5-HT (t₁₁ = 4.237, p = 0.0014, paired Student’s t-test) (Figure 1F). Additionally, application of 5-HT significantly decreased the frequency of sEPSCs (t₁₁ = 3.373, p = 0.0053, paired Student’s t-test) (Figure 1G). In the same CLA-ACC neurons, we also examined whether 5-HT impacted intrinsic membrane excitability and action potential generation. We found that 5-HT caused hyperpolarization of the resting membrane potential (RMP) (t₁₁ = 3.032, p = 0.0114, paired Student’s t-test) (Figure 2A), but observed no differences in membrane resistance (t₁₁ = 1.582, p = 0.1419, paired Student’s t-test) (Figure 2B) or membrane capacitance (t₁₁ = 0.1760, p = 0.8635, paired Student’s t-test) (Figure 2C). Bath application of 5-HT also increased the rheobase current (t₁₁ = 5.266, p = 0.0003, paired Student’s t-test) (Figure 2D) and decreased the action potential firing rate in response to depolarizing current steps (Figure 2E–F) (F_1,12 = 28,08, p = 0.0002, 2-way RM ANOVA). These results indicate that 5-HT broadly reduces excitability of glutamatergic synapses onto the CLA-ACC neurons, hyperpolarizes CLA-ACC neuron membranes, and suppresses action potential generation.

Figure 1: 5-HT decreases synaptic glutamate transmission on CLA-ACC neurons.

(A) Representative 2X (i-ii), 10X (iii), and 40X (iv) confocal images of the retrograde AAV-eGFP expression. Injection of the virus into the ACC highlights a large population of the CLA-ACC neurons. A_iii is an expansion of the dotted square in A_ii. A_iv is an expansion of the dotted square in A_iii. Scale bars are 500 μm in A_i-ii, 100 μm in A_iii, and 20 μm in A_iv. INS, insula; EC, external capsule. (B) Representative traces of sEPSCs recorded from a CLA-ACC neuron before (blue) and during (red) bath perfusion of 5-HT (30 μM). (C) Example sEPSC averages before and during 5-HT application in a CLA-ACC neuron (left) are amplitude-normalized (right) to compare sEPSC duration. (D-G) Amplitude, duration (tau), charge transfer, and frequency of sEPSCs before and during 5-HT application in all recorded CLA-ACC neurons. **p < 0.01; ****p < 0.0001, paired Student’s t-tests. n = 12 neurons from 4 rats.

Figure 2: 5-HT inhibits CLA-ACC membrane excitability.

(A-D) RMP, membrane resistance, membrane capacitance, and rheobase of CLA-ACC neurons before and during 5-HT application. (E) Representative traces of action potentials at rheobase (top) and at maximal (300 pA) injected current before (blue) and during (red) 5-HT perfusion. (F) Inhibitory effect of 5-HT exposure on average action potential firing rates in CLA-ACC neurons across the range of depolarizing current injections. *p < 0.05; ***p < 0.001, paired Student’s t-tests (A-D) 2-way RM ANOVA (F). n = 12 neurons from 4 rats.

3.2. Elevated levels of 5-HT2C mRNA in the CLA

We used qPCR analysis to gain insight into the specific receptor subtypes that may be responsible for the 5-HT effects seen in our electrophysiology data. Expression levels of 13 different 5-HT receptor subtypes in the CLA were compared to the nearby insula and the ACC, as a way to gauge CLA-specific enrichment of 5-HT receptors subtypes. Highly expressed subtypes were defined as those with average arbitrary fluorescence (afu) ≥ 1.0, moderately expressed – as those with afu ≥ 0.5 and < 1.0, and those with afu < 0.5 were defined as having low regional expression levels. Within the CLA, we found high levels of 5-HT2A and 5-HT2C expression, moderate levels of expression of 5-HT1A, 5-HT1B, and 5-HT5A RNA, and low levels of expression of other 5-HT receptor subtypes (Figure 3A–B, Suppl. Table 2). Expression of 5-HT3B mRNA was below our level of detection in all regions (data not shown). Among the highly expressed subtypes, we found that 5-HT2C expression in the CLA was ~5-fold larger than in either the ACC or the insula, whereas 5-HT2A levels in the CLA were ~40% larger than in the ACC, but comparable to expression in the insula (Suppl. Table 2). Among the moderately expressed genes, region-specific differences were observed for 5-HT1A and 5-HT1B, but not for 5-HT5A (Suppl. Table 2). Among the genes with low expression levels, unique CLA expression patterns were observed for 5-HT1D, 5-HT6, and 5-HT7, relative to expression levels of these genes in the ACC and the insula (Suppl. Table 2).

Figure 3: qPCR results of 5-HT subtype expression.

(A) Bar graphs showing mRNA expression levels of the low to moderately expressed 5-HT receptor subtypes in the CLA and insula (INS), normalized to Tbp expression. (B) Bar histogram of the highly expressed 5-HT2A and 5-HT2C receptor mRNA in CLA and INS. Small black squares and triangles indicate mRNA levels within individual animals. Tissue punches from 4–8 brain slices were collected from each region and pooled for each individual animal. N = 16 rats (8 male & 8 female).

Since our samples included tissue from both male and female rats, we evaluated potential sex differences in the expression of the 5-HT receptor mRNA. Significant main effects of sex were found for 5-HT1A, 5-HT2A, 5-HT5A, and 5-HT7 across the three regions of interest. However, pairwise comparisons revealed that 5-HT1A was the only gene differentially expressed between males and female at the regional level, with greater 5-HT1A expression in female than in male ACC (Suppl. Figure 1). Within the CLA, expression levels of all 5-HT subtypes were comparable between males and females.

3.3. Cell-type specificity of 5-HT1 and 5-HT2 mRNA levels.

Our qPCR results suggested an outsized role for 5-HT2C-mediated signaling in the CLA, whereas the majority of published studies indicate that serotonergic regulation of neuronal excitability relies on 5-HT1A (Araneda and Andrade 1991, Andrade 2011, Halberstadt and Geyer 2011, Meunier, Amar et al. 2013, Rojas and Fiedler 2016, Carhart-Harris and Nutt 2017, Cameron, Benetatos et al. 2023) and 5-HT2A (Araneda and Andrade 1991, Aghajanian and Marek 1997, Aghajanian and Marek 1999, Zhou and Hablitz 1999, Andrade 2011, Crunelli and Di Giovanni 2015, Carhart-Harris and Nutt 2017, Berthoux, Barre et al. 2019) signaling in other brain regions. We used RNAscope to evaluate whether 5-HT1A and 5HT2A-C were differentially expressed in glutamatergic and GABAergic cells of the CLA which may contribute to the ability of 5-HT to control synaptic excitability and action potential output of CLA-ACC neurons. We found that 30.05 ± 0.6345 % of DAPI-marked nuclei in the CLA colocalized with the glutamatergic neuron marker, Slc17a7, while only 7.193% ± 0.4357 of DAPI label colocalized with the GABA-neuron marker, Gad1 (Figure 4A–C). Assuming that glutamate and GABA markers comprise the majority of neurons in the CLA and do not substantially co-localize with each other, these fractions translate into 81% and 19% of glutamate and GABA neurons, respectively, as contributing to the overall CLA neuron population. Within the overall DAPI-positive cell population (both neurons and non-neurons), we calculated similar levels of 5-HT1A (17.86 ± 0.8591%), 5-HT2A (24.23 ± 1.008%), and 5-HT2C (22.26 ± 0.6263%) puncta (Figure 4D). Within the population of glutamatergic, Slc17a6-positive, neurons we found significantly more 5-HT2C (47.92 ± 0.8661%) than 5-HT1A (42.77 ± 0.9632%) puncta (t₃₁ = 3.983, p = 0.0004, unpaired Student’s t-test) with intermediate levels of 5-HT2A (42.77 ± 0.9632%) puncta that were not significantly different from the other two subtypes (Figure 4E). There was less overall colocalization of 5-HT subtypes with the GABAergic neurons. The 5-HT1AR in particular, was only detected in 5.645% ± 0.5450 of GAD1-positive cells, which was significantly less than both the 5-HT2A (14.08% ± 1.890, t₃₄ = 4.288, p=0.0001, unpaired Student’s t-test) and the 5-HT2C (20.37 ± 2.024; t₃₄ = 7.025, p <0.0001, unpaired Student’s t-test) puncta (Figure 4F). There were also significantly more 5-HT2C, than 5-HT2A puncta (t₃₄ = 2.270, p = 0.0297, unpaired Student’s t-test) localized to the GABAergic neurons of the CLA. No 5-HT2B fluorescence was detected in RNAscope experiments (data not shown), consistent with the qPCR findings. Overall, these results show that the majority of 5-HT receptor expression in the CLA is on glutamatergic neurons with 5-HT2C expressed at particularly elevated levels. Additionally, we show that the GABAergic neurons of the CLA express mostly the 5-HT2 subtypes, again with higher levels of 5-HT2C than 5-HT1A or 5-HT2A subtypes.

Figure 4: Cell-type specificity of 5-HT1A and 5-HT2A,C mRNA levels.

(A) Representative 10X (top) and 40X (bottom) confocal images of overlaid glutamatergic (Slc17a7, green), 5-HT receptor (pink), and DAPI (cyan) mRNA puncta. Select areas that co- express all three labels are indicated by arrowheads. Scale bars are 200 μm and 50 μm for 10X and 40X images, respectively. INS, insula; EC, external capsule. (B) Same as (A), except that GABAergic neuron probe (GAD1) is used in the green channel. (C) Bar graphs summarizing integrated density of glutamatergic and GABAergic puncta as a fraction of DAPI-labeled area. (D) Integrated density of 5-HT receptor puncta as a fraction of DAPI-labeled area. (E) Integrated density of 5-HT receptor puncta as a fraction of Slc17a7 (Glu⁺)-labeled area. (F) Integrated density of 5-HT receptor puncta as a fraction of GAD1 (GABA⁺)-labeled area. Small grey dots in C-F represent individual 40X field of view values. *p < 0.05; **p < 0.01; ****p < 0.0001, unpaired Student’s t-tests. n = 15–18 fields of view from 5–6 brain slices from 4 rats (2 males & 2 females).

3.4. 5-HT1A and 2C underlie the 5-HT effects on CLA-ACC neuron excitability

To interrogate 5-HT receptor subtype contributions to the synaptic effects of 5-HT at CLA-ACC neurons, we performed slice recording experiments with local application of subtype-selective 5-HT receptor antagonists. We chose three antagonists: WAY 100,635 (5-HT1A antagonist, 10 μM), MDL 100907 (5-HT2A antagonist, 300 nM), and SB 242,084 (5-HT2C antagonist, 10 μM), based on our mRNA data and documented role of 5-HT1A as the principal mediator of inhibitory effects of 5-HT on membrane excitability outside the CLA (Lanfumey and Hamon 2000, Carhart-Harris and Nutt 2017). Our results showed that inhibition of 5-HT1A signaling reversed the effects of 5-HT on sEPSC amplitude (t₁₃ = 2.590, p = 0.0224, paired Student’s t-test) and decreased sEPSC duration (t₁₃ = 2.333, p = 0.0364, paired Student’s t-test) (Figure 5A,Bi–Di), but had no effect on sEPSC frequency (t₁₃ = 1.931, p = 0.0756, paired Student’s t-test) (Figure 5A, Ei). In contrast, inhibition of 5-HT2C signaling reversed the effects of 5-HT on sEPSC frequency (t₁₂ = 3.005, p = 0.011, paired Student’s t-test), but had no effect on sEPSC amplitude, duration, or charge transfer (Figure 5A,Biii–Eiii). Surprisingly, inhibition of 5-HT2A was without effect on any of the measured sEPSC parameters (Figure 5A,Bii,Cii,Dii). Inhibition of GABA_A receptors with picrotoxin (PTX) (100 μM) did not eliminate the effects of 5-HT (30 μM) and SB 242084 (10 μM) on sEPSC amplitude (F_{2, 20} = 3.637, p = 0.045, 1-way RM ANOVA), frequency (F_{2, 20} = 6.320, p = 0.0075, 1-way RM ANOVA), or RMP (F_{2, 15} = 9.493, p = 0.0022, 1-way RM ANOVA) (Supplemental Figure 2). In particular, SB 242,084 continued to reverse the effects of 5-HT on sEPSC frequency as observed in the absence of picrotoxin, suggesting that these synaptic effects occur independently of local GABAergic inhibition. Altogether, our results suggest that 5-HT2C receptor contribution to 5-HT regulation of glutamatergic synapses onto the CLA-ACC neurons involves pre-synaptic mechanisms that may impact probability of glutamate release, whereas 5-HT1A receptors contribute to post-synaptic effects of 5-HT that alter the amplitude and duration of synaptic response to glutamate.

Figure 5: 5-HT1A and 5HT-2C receptors mediate different components of the 5-HT synaptic effects.

(A) Representative traces of sEPSCs during application of 5-HT (30 μM) and 5-HT with receptor antagonists for 5-HT1 (WAY 100,635, 10 μM (+WAY)), 5-HT2A (MDL 100,907, 300 nM (+MDL)), and 5-HT2C (SB 242,084, 10 μM (+SB)) receptors. (B-E) Average sEPSC amplitude, duration (tau), charge transfer, and frequency during application of 5-HT and 5-HT with receptor antagonists in individual CLA-ACC neurons. *p < 0.05, paired Student’s t-tests. n = 12–15 cells from 4–5 rats for each 5-HT receptor antagonist.

We then examined 5-HT receptor subtype contribution to passive and active membrane properties. Interestingly, inhibition of either the 5-HT1A or the 5-HT2C receptors attenuated 5-HT induced hyperpolarization of the RMP (5HT1A: t₁₁ = 3.360, p = 0.0064; 5HT2C: t₅ = 4.070, p = 0.0096, paired Student’s t-tests), while inhibition of 5HT2A was without effect on RMP in these cells (Figure 6A–C). None of the subtype-selective antagonists reversed 5-HT effects on rheobase (Figure 6D–F) and none of the antagonists attenuated inhibitory effects of 5-HT on action potential firing (Figure 6G). In fact, both 5-HT2A and 5-HT2C antagonists further reduced 5-HT-mediated inhibition of action potential firing rates across a range of depolarizing potentials (Figure 6H–J). These data indicate that in addition to the pre-synaptic effects of 5-HT2C on glutamatergic synapses, 5-HT2C receptors may contribute to post-synaptic regulation of membrane excitability by 5-HT. Finally, 5-HT suppression of action potential output of CLA-ACC neurons appears to rely on a more complex regulatory scheme that is preserved in the presence of subtype-specific antagonists.

Figure 6: 5-HT1A and 5-HT2C receptors both contribute to 5-HT hyperpolarization of the membrane potential.

(A-F) RMP and rheobase values during application of 5-HT (30 μM) and 5-HT with antagonists for the 5-HT1A (WAY 100,635, 10 μM), 2A (MDL 100,907, 300nM), or 2C (SB 242,084, 10 μM) receptors in individual CLA-ACC neurons. (G) Representative traces of 5-HT antagonist effects on action potential output in response to rheobase current (left) and maximal current (300 pA, right) injections. Red traces are during 5-HT application. Colored traces are during 5-HT+antagonist application. (H-J) Receptor antagonist effects on action potential output across the range of depolarizing current injections. Significant main effects of drug treatments from 2-way RM ANOVAs are as follows: I) F_1,6=20.98, p=0.0038; J) F_1,5=12.14, p=0.0176. *p < 0.05; **p < 0.01, paired Student’s t-tests (A-F), 2-way RM ANOVA (H-J). n = 6–10 neurons from 3–5 rats for each receptor antagonist.

3.5. The psychedelic 5-HT2R agonist, DOI, excites CLA-ACC neurons via 5-HT2C receptors

We were surprised by lack of effects of 5-HT2A inhibition on CLA-ACC neuron excitability, especially given our mRNA data indicating robust 5-HT2A expression in the CLA and published evidence of CLA involvement in cellular and behavioral responses to serotonergic psychedelics, that are thought to rely heavily on 5-HT2A signaling. To examine the relative contribution of 5-HT2A and 5-HT2C receptors to the effects of a psychedelic drug exposure, we performed slice recording experiments before and after bath application of DOI (10 μM). DOI is a psychedelic partial agonist of 5-HT2A and 5-HT2C receptors with about 3.5-fold higher affinity for the 5-HT2AR (racemic K_i for 5-HT2A = 0.68 nM, 5-HT2C = 2.38 nM) (Boess and Martin 1994). During DOI application, CLA-ACC neurons showed significant increases in sEPSC amplitude (t₈ = 3.027, p = 0.0164, paired Student’s t-test), frequency (t₈ = 4.689, p = 0.0016, paired Student’s t-test), and charge transfer (t₈ = 2.664, p = 0.0286, paired Student’s t-test), while the sEPSC decay time remained unchanged (Figure 7A–E). These synaptic effects were not accompanied by significant changes in RMP or rheobase at the 10 μM concentration (Figure 7F,G), but action potential firing rate of CLA-ACC neurons was significantly suppressed during DOI exposure (F_1,5 = 72.57, p = 0.0004, 2-way RM ANOVA) (Figure 7H–I).

Figure 7: The psychedelic 5-HT2 receptor agonist, DOI, increases synaptic transmission, but suppresses action potential output in CLA-ACC neurons.

(A) Example traces of sEPSCs before and during DOI application. (B-G) DOI effects on sEPSC amplitude, duration (tau), charge transfer, and frequency as well as on RMP and rheobase current in individual CLA-ACC neurons. (H) Representative traces of action potentials at rheobase current (left) and at maximal injected current (right) in CLA-ACC neurons before and during DOI application. (I) DOI effect on action potential output across the range of depolarizing currents. *p< 0.05; **p < 0.01; ***p < 0.001, paired Student’s t-tests (B-G), 2-way RM ANOVA (I). n = 7–9 neurons from 2–3 rats.

We next evaluated whether regulation of glutamatergic synapse excitability would be observed at lower concentrations of DOI. Application of 300 nM DOI increased sEPSC amplitude, charge transfer, and frequency (amplitude: t₁₀ = 3.690, p = 0.0042; charge: t₁₀ = 2.282, p = 0.0456; frequency: t₁₀ = 2.669, p = 0.0235, paired Student’s t-tests). Qualitatively similar potentiation of sEPSCs was observed with 1 μM DOI (amplitude: t₉ = 3.407, p = 0.0078; charge: t₉ = 3.940, p = 0.0034; frequency: t₉ = 2.548, p = 0.0313, paired Student’s t-tests) (Supplemental Figure 3A–D). Unlike the 10 μM dose, however, both 300 nM and 1 μM DOI concentrations caused a hyperpolarization of the RMP (300 nM: t₇ = 4.387, p = 0.0032; 1 μM: t₇ = 10.15, p < 0.0001, paired Student’s t-tests) (Supplemental Figure 3E,J). Altogether, these findings show that selective activation of 5-HT2 signaling in the CLA by DOI facilitates glutamatergic transmission at CLA-ACC neurons, in stark contrast to the attenuating synaptic effects of non-selective 5-HT receptor activation by 5-HT. However, action potential output is suppressed by DOI and 5-HT alike.

To probe whether the effects of DOI relied on 5HT2A or 5HT2C receptor activation, CLA-ACC neurons were pre-incubated with the bath-applied antagonist of one of either receptor subtype prior to application of DOI. During inhibition of 5-HT2A by MDL 100907 (300 nM), application of DOI increased sEPSC amplitude (t₆ = 2.464, p = 0.0488, paired Student’s t-test) and frequency (t₆ = 3.084, p = 0.0216, paired two-tailed t test, n = 7), but had no effect on decay time or charge transfer (Figure 8A–D, M). The RMP was hyperpolarized (t₈ = 3.429, p = 0.0090, paired Student’s t-test), but there was no significant difference in rheobase current (Figure 8E,F). The action potential firing rate was suppressed (F_{1, 7} = 6.62, p = 0.0369, 2-way RM ANOVA) similar to the effects of DOI in the absence of 5-HT2A inhibition (Figure 8N,O). A different pattern of results was observed in the presence of the 5-HT2C antagonist, SB 242,084 (10 μM). In this case, DOI application did not change the sEPSC amplitude, duration, charge transfer, or frequency (Figure 8G–J, P), indicating that the 5-HT2CR mediates the synaptic effects of DOI. However, RMP was hyperpolarized (t₁₀ = 5.466, p = 0.0003, paired Student’s t-test, Figure 8K) and rheobase current was significantly increased (t₅ = 3.091, p = 0.0271, paired Student’s t-test, Figure 8L). These changes were accompanied by a significant attenuation of the action potential firing rate (F_{1, 5} = 36.28, p < 0.01, 2-way RM ANOVA, Figure 8Q, R).

Figure 8: 5-HT2C receptors mediate DOI effects on glutamatergic synapses.

(A-F) DOI (10 μM) effects on synaptic and membrane excitability measures after pre-incubation with the 5-HT2A antagonist, MDL 100,907 (MDL, 300nM) in individual CLA-ACC neurons. Note that potentiation of sEPSC amplitude and frequency is maintained by DOI. (G-L) Same as (A-F) except after preincubation with the 5-HT2C antagonist, SB 242,084 (SB, 10 μM). Note the absence of DOI effects on sEPSC amplitude and frequency. (M) Representative sEPSC traces in the presence of MDL100,907 or MDL 100,907 + DOI. (N) Representative action potentials in response to maximal injected current in the presence of MDL100,907 or MDL 100,907 + DOI. Red traces are during MDL 100,907 + DOI application. Blue traces are during application of MDL 100,907 alone. (O) Inhibition of 5-HT2A signaling does not eliminate inhibitory effect of DOI on action potential output. (P) Representative sEPSC traces in the presence of SB 242,084 or SB 242,084 + DOI. (Q) Representative action potentials in response to maximal injected current in the presence of SB 242,084 or SB 242,084 + DOI. Red traces are during SB 242,084 + DOI application. Blue traces are during application of SB 242,084 alone. (R) Inhibition of 5-HT2C signaling does not eliminate inhibitory effect of DOI on action potential output. Scale bars are 30 pA by 5 s in (panels M and P) and 40 mV by 200 ms (panels N and O). *p < 0.05; **p < 0.01; ***p < 0.001, paired Student’s t-tests (A-L), 2-way RM ANOVA (O,R). n = 6–11 neurons from 6 animals.

In a separate cohort of animals, CLA-ACC neurons were patched in regular aCSF before application of a lower concentration of SB 242,084 (1 μM) followed by DOI (1 μM) to evaluate the possibility of constitutive 5-HT2C activity influencing DOI effects. We found that application of SB 242,084 by itself had no effect on sEPSC parameters or the RMP (Supplementary Figure 4). However, the lower concentration of SB 242,084 was still sufficient to prevent the DOI-mediated increase in sEPSC amplitude and frequency (Supplementary Figure 4).

Overall, these results support the conclusion that the synaptic effects of the psychedelic drug, DOI, in CLA-ACC neurons are mediated by 5-HT2C, rather than 5-HT2A, receptor activation. As with 5-HT experiments, however, DOI effects on passive membrane properties and action potential output of CLA-ACC neurons cannot be attributed to activation of a single 5-HT subtype.

4. Discussion

Despite the promising clinical (2015, Griffiths, Johnson et al. 2016, Johnson, Garcia-Romeu et al. 2017, Garcia-Romeu, Davis et al. 2019, Vollenweider and Preller 2020, Carhart-Harris, Giribaldi et al. 2021, Davis, Barrett et al. 2021, Gukasyan, Davis et al. 2022) and preclinical (Cameron, Benson et al. 2018, Ly, Greb et al. 2018, Berquist and Fantegrossi 2021, Benvenuti, Colombo et al. 2023, Ferri, de Novais et al. 2023, Torrado Pacheco, Olson et al. 2023) results, the mechanisms by which psychedelics achieve their therapeutic effects remain a mystery. Much of the pre-clinical psychedelic research has focused on 5-TH2A, and, to a lesser extent, 5-HT1 receptors in pyramidal neurons of the prefrontal cortex (Gonzalez-Maeso, Weisstaub et al. 2007, Gresch, Barrett et al. 2007, Halberstadt and Geyer 2011, Nichols 2016, Ly, Greb et al. 2018, Riga, Llado-Pelfort et al. 2018, Shao, Liao et al. 2021, Erkizia-Santamaria, Alles-Pascual et al. 2022, Cameron, Benetatos et al. 2023). At the same time, recent functional connectivity analyses support a central role for the CLA as mediator of network response to psychedelics (Barrett, Krimmel et al. 2020, Doss, Madden et al. 2022), bolstering early arguments of substantial CLA influence based on intensity of 5-HT receptor binding across the brain (Pazos, Cortés et al. 1985, McKenna and Saavedra 1987, Nichols 2016). Our results further extend these findings by confirming dense 5-HT receptor expression in the CLA and by evaluating CLA levels of mRNA for thirteen 5-HT subtypes. Among these subtypes, we find that CLA has markedly increased expression of 5-HT2C receptors relative to both the neighboring insula and to cortical (ACC) tissue. Consistent with this, our functional characterization with slice electrophysiology indicates a prominent contribution of 5-HT2C, but not 5-HT2A, receptors in regulation of glutamatergic synapse excitability of CLA-ACC neurons. The 5-HT2C regulation of CLA synaptic excitability is intriguingly bi-directional, contributing to both 5-HT-mediated suppression of sEPSCs and to sEPSC potentiation by the serotonergic psychedelic, DOI.

4.1. Cellular distribution of 5-HT receptor subtypes in the CLA

Our qPCR data show that 5-HT2A and 5-HT2C receptors are the most abundantly expressed subtypes in the CLA, whereas 5-HT1A, 1B, 5A and 6 are expressed at moderate levels. However, it must be noted that our mRNA data may not reflect levels of receptor protein that could be expressed in different proportions. The mRNA for 5-HT1A, 2A, and 2C receptors was found to co-localize with 40–50% of the Slc17a7 mRNA label, suggesting that a substantial fraction of CLA glutamatergic neurons express other 5-HT receptor subtypes or do not express 5-HT receptors at all. The extent to which 5-HT subtypes are desegregated between individual CLA neurons remains unclear as we have not attempted to examine co-localization of mRNA for distinct 5-HT receptor subtypes within individual cells. Similarly, our mRNA data does not allow us to determine whether 5-HT receptors subtypes are uniformly distributed across all glutamatergic projection neurons or concentrated at projection neurons onto specific targets. However, our functional data with both 5-HT and with DOI applications support the interpretation that multiple 5-HT receptor subtypes exert a combined influence on CLA excitability at least within the population of glutamatergic CLA-ACC neurons. This observation aligns with findings of co-localized expression of 5-HT1 and 5-HT2 receptor subtypes in the rat cortex (Wedzony, Chocyk et al. 2008, Nocjar, Alex et al. 2015).

We found that GABAergic cells of the CLA express markedly lower levels of 5-HT1A, 2A, and 2C mRNA. 5-HT2A and 2C were detected at ~15–20% of GAD1-positive neurons while 5-HT1A expression was minimal and co-expressed with ~5% of GAD1 label. Qualitatively similar results have been reported in the primate and human cortex. The mRNA for 5-HT1A was identified in ~13–30% of GAD-positive neurons in the dorsolateral prefrontal and orbitofrontal cortices (de Almeida and Mengod 2008), a substantially lower fraction than the ~45–87% of GABAergic cells of the prefrontal cortex that express mRNA for the 5-HT2A receptor subtype (de Almeida and Mengod 2007). These studies have noted unequal distribution of 5-HT receptor subtypes among parvalbumin- and calbindin- positive interneurons of the cortex (de Almeida and Mengod 2007, de Almeida and Mengod 2008), while CLA has been shown to express parvalbumin, somatostatin, and other types of interneurons (Braak and Braak 1982, Takahashi, Kobayashi et al. 2023). It will be informative to examine whether 5-HT receptor subtypes in the CLA are also preferentially targeted to distinct interneuron populations.

With regard to sub-regional specificity, the CLA can be provisionally divided into shell and core regions based on cell type distribution, circuit mapping, and electrophysiological characteristics (Mathur 2014, Chia, Augustine et al. 2020, Erwin, Bristow et al. 2021, Marriott, Do et al. 2021). Previous studies indicate that CLA core is enriched with the parvalbumin-positive cell bodies (Mathur 2014, Chia, Augustine et al. 2020, Marriott, Do et al. 2021), which may influence the impact of 5-HT receptors on CLA synaptic excitability, assuming that dense expression of cell bodies translates into greater local synapse density. In our data, 5-HT1A receptor mRNA appeared equally distributed throughout the CLA, whereas 5-HT2A and 5-HT2C expression was localized more toward the CLA center. However, we have not performed quantitative comparisons of 5-HT expression between CLA core and shell since the rules for delineation of the two regions are somewhat ambiguous. Finally, we would like to emphasize that despite evidence of sex differences in behavioral sensitivity to 5-HT receptor agonists (Tyls, Palenicek et al. 2016, Jaster, Younkin et al. 2022, Effinger, Quadir et al. 2023), our data indicate similar CLA levels of mRNA for all the examined 5-HT receptors between male and female rats.

4.2. Synaptic effects of 5-HT and DOI

While significant innervation of the CLA by various neuromodulators has been noted (Zingg, Dong et al. 2018, Wong, Nair et al. 2021), the impact of 5-HT on CLA neuron excitability has not been previously investigated. Based on our findings of substantial 5-HT receptor co-expression with glutamate neuron markers, we evaluated 5-HT receptor role in regulation of glutamatergic synaptic transmission. Previous studies in pyramidal neurons of the prefrontal cortex have shown that 5-HT (Aghajanian and Marek 1997) and DOI (Aghajanian and Marek 1999, Ekins, Brooks et al. 2023) increase sEPSC amplitude and frequency, however, 5-HT has been reported to decrease EPSCs in the cingulate cortex (Tanaka and North 1993, Tian, Yamanaka et al. 2017) and DOI has been shown to inhibit firing of neurons in the locus coeruleus (Chiang and Aston-Jones 1993) and dorsal raphe nucleus (Boothman, Allers et al. 2003). We found robust reduction of sEPSC amplitude and frequency by 5-HT in CLA-ACC neurons, suggesting that 5-HT receptors regulate glutamatergic synapses in CLA via both post- and pre-synaptic mechanisms. Suppression of sEPSC amplitude by 5-HT could be reversed by the 5-HT1A antagonist, WAY 100,635 and suppression of sEPSC frequency was sensitive to the 5-HT2C antagonist, SB 242,084. Unexpectedly, application of the serotonergic psychedelic, DOI, increased sEPSC amplitude and frequency, and both of these effects were eliminated in the presence of the 5-HT2C antagonist. These findings indicate that the presumably pre-synaptic mechanisms underlying 5-HT regulation of sEPSC frequency uniquely involve 5-HT2C receptor activation, whereas the presumably post-synaptic pathways underlying changes in sEPSC amplitude may involve cooperative activation of 5-HT1A and 5-HT2C receptors. Related to this, heterodimers between diverse subtypes of serotonin receptors, including those coupled to distinct intracellular signaling cascades (e.g. 5-HT1+5-HT2) have been demonstrated in various models (Maroteaux, Bechade et al. 2019). It remains to be determined whether such heterodimers exist in CLA and what effect they might have on excitatory synaptic transmission.

Our picrotoxin experiments (Supplemental Figure 2) suggest that GABAergic interneurons in the CLA did not impact our sEPSC results. GABAergic regulation of sEPSCs would require either innervation of glutamatergic terminals within CLA by local interneurons or extensive connectivity between the CLA-ACC projection neurons that are influenced by local GABA signals. While CLA interneurons have been shown to synapse extensively onto the CLA projection cells, synaptic connections between CLA projection neurons themselves are very sparse (Kim, Matney et al. 2016). The axo-axonal synapses between interneurons and glutamatergic terminals have been demonstrated in some brain areas (Cover and Mathur 2021), however, axonal arbors of CLA interneurons appear to largely target other interneurons or the CLA projection cells (Kim, Matney et al. 2016).

With respect to the origin of glutamatergic terminals that may underlie the 5-HT and DOI effects on sEPSC frequency, recent circuit mapping data show that there are several brain regions that innervate CLA-ACC neurons, most prominently the frontal cortices (Qadir, Stewart et al. 2022). Roughly half of CLA-ACC neurons respond to stimulation of the ACC, prelimbic cortex, and orbitofrontal cortex, while over 30% respond to insula stimulation, and less than 5% respond to inputs from motor and visual cortices, indicating weak, but existing synaptic connections from sensorimotor regions (Chia, Augustine et al. 2020).

The most difficult question is how 5-HT2C receptors can mediate both the suppression of sEPSC frequency by 5-HT and the potentiation of sEPSC frequency by DOI that we observe. Generally speaking, the concept of functional selectivity at G-protein coupled receptors where different ligands at the same receptor activate different signaling cascades is relatively well-established (Zhou and Bohn 2014, Nichols 2016, Kwan, Olson et al. 2022). Although empirical evidence is limited in specific regard to 5-HT2C regulation of neurotransmitter release that could modulate sEPSC frequency, 5-HT2C signaling can be expected to increase intracellular Ca²⁺ levels through canonical G_q-coupled intracellular pathways that would promote Ca²⁺ dependent release of glutamate. The rise in intracellular Ca²⁺ may stimulate pre-synaptic Ca²⁺-activated potassium (BK) channels that would oppose Ca²⁺-mediated signaling by hyperpolarizing neuronal membrane. Indeed, both 5-HT2C coupling to BK channels (Wang, Wang et al. 2020) and BK-mediated suppression of neurotransmitter release (Griguoli, Sgritta et al. 2016) have been previously described. BK channel gating is strongly sensitive to Ca²⁺ concentrations within a narrow 0.1–100 μM range (Lorenzo-Ceballos, Carrasquel-Ursulaez et al. 2019). Given this evidence, our data support the prediction that activation of 5-HT2C receptors results in differential recruitment of presynaptic BK signaling, sufficient to overcome G_q-stimulated release of glutamate in the case of 5-HT, but not in the case of DOI exposure.

4.3. Membrane excitability during 5-HT and DOI

The coupling between 5-HT receptors and various K⁺ channels has been previously raised as a mechanism for 5-HT receptor regulation of membrane excitability. It has been proposed, for example, that hyperpolarizing and depolarizing effects of 5-HT on cingulate cortex neurons could be attributed to, respectively, 5-HT1-mediated increase and 5-HT2-mediated decrease of an unidentified potassium conductance (Tanaka and North 1993). Consistent with a potassium channel activation theory, we routinely observed an outward shift of holding current (5–230 pA) in the CLA-ACC neurons during 5-HT and DOI exposures (data not shown). In our data, 5-HT-induced hyperpolarization of the RMP could be attenuated by inhibition of the 5-HT1AR and 5-HT2CR, indicative of cooperative regulation of RMP by these receptor subtypes. Further supporting receptor cooperativity are our observations that DOI (at 10 μM) was without effect on RMP when applied alone but hyperpolarized the RMP when applied in the presence of either the 5-HT2A or the 5-HT2C antagonist. Which potassium channel may account for these effects? Signaling at 5-HT receptors has been linked to inhibition of G-protein-coupled inward rectifiers (Yamamoto, Hatano et al. 2014, Montalbano, Corradetti et al. 2015), activation of BK channels (Wang, Wang et al. 2020) and, most recently, activation of M-type potassium channels (Ekins, Brooks et al. 2023). Both depolarizing and hyperpolarizing effects on neuronal membranes can be expected from such a mix of potassium channel coupling. However, regulation of CLA-ACC neuron membrane excitability by 5-HT receptors may not be limited to potassium channels. For example, DOI exposure has been found to result in phosphorylation of the voltage-gated sodium channels, inhibiting dendritic, but not somatic, action potentials in the prefrontal cortex (Carr, Cooper et al. 2002). We observed that both 5-HT and DOI robustly inhibited somatic action potentials in CLA-ACC neurons. None of the 5-HT subtype-selective antagonists that we tested were able to alleviate this suppression. In fact, all antagonists, with the exception of 5-HT1, further suppressed action potential output. We do not have a clear explanation for the latter finding. Distinct time-courses of 5-HT receptor activation and desensitization may play a role as suggested by previous studies (Aghajanian and Marek 1997, Zhou and Hablitz 1999, Backstrom, Price et al. 2000).

4.4. Absence of 5-HT2A effects

One striking outcome of our experiments is that 5-HT2A receptors did not contribute to synaptic or intrinsic membrane excitability in CLA-ACC neurons, despite the prominent expression of 5-HT2A mRNA in the CLA. The CLA projection neurons innervate multiple brain targets (Kim, Matney et al. 2016, Graf, Nair et al. 2020, Erwin, Bristow et al. 2021, Qadir, Stewart et al. 2022), but the relative expression of 5-HT2A in these circuits is currently not known. Nevertheless, it is unlikely that 5-HT2A receptors are specifically excluded from the CLA-ACC neurons. Our electrophysiological recordings did not distinguish between CLA-ACC neurons along the rostrocaudal CLA axis, and were, if anything, biased toward CLA core where 5-HT2A expression appeared denser than in CLA periphery. More likely is the possibility that the 5-HT2A receptors in CLA-ACC cells provoke cellular responses that were not examined in the current study. While synaptic and membrane excitability effects of 5-HT activation have been reported in various cell types, several non-conventional cellular mediators of psychedelic drug action have been suggested in recent years. For example, psychedelic-induced remodeling of dendrites may differ depending on recruitment of plasma membrane and intracellular 5-HT2A receptors (Vargas, Dunlap et al. 2023) and may specifically require BDNF interaction with trkB receptors (Moliner, Girych et al. 2023). Sequestration of 5-HT2A to plasma membrane or intracellular receptors pools or recruitment of BDNF signaling pathways by serotonergic psychedelics in the CLA remains to be explored.

5. Conclusion

To summarize, our findings emphasize that the effects of the serotonergic psychedelic, DOI, on CLA-ACC neurons are distinct from the effects of the broader receptor pool activation by 5-HT and highlight an essential role for 5-HT2C signaling in regulation of CLA excitability. Progress in understanding psychoactive and therapeutic effects of psychedelics will benefit from further characterization of receptors, signaling cascades, and cellular interactions in this underexplored brain area.

Supplementary Material

1

Supplemental Figure 1. Sex differences in 5-HT receptor gene expression across CLA, ACC, and insula. Blue columns – ACC, red columns – CLA, green columns – INS. Two-way ANOVAs were conducted to evaluate difference in individual 5-HT receptor mRNA levels with sex and region as factors. Only significant main effects of sex are indicated on the graphs. In cases of significant main effects of sex, the only significant post-hoc comparison (Sidak’s) was for 5-HT1A levels in the ACC. Small colored circles, squares, and triangles are receptor expression levels in individual animals.

Supplemental Figure 2. Interneurons do not contribute to the glutamatergic synaptic effects of 5-HT or SB 242,084. (A-E) sEPSC amplitude, tau, charge, frequency, and RMP effects of 5-HT (30 μM) and SB 242,084 (10 μM) on CLA-ACC neurons recorded in the presence of bath picrotoxin (PTX) (100 μM). (F) Representative amplitude traces of CLA-ACC neurons recorded in regular aCSF (blue), during 5-HT (red), and during 5-HT with SB (green). (G) Representative frequency traces of CLA-ACC neurons recorded in regular aCSF (blue), after 5-HT (red), and after 5-HT with SB (green). *p < 0.05, **, p<0.01, Tukey’s post-hoc after 1-way RM ANOVA. n = 11 cells from 4 animals.

Supplemental Figure 3. Lower concentrations of DOI increase sEPSC transmission. (A-E) Effects of 300 nM DOI on sEPSC amplitude, tau, charge, frequency, and RMP. (F-J) Effects of 1 μM DOI on sEPSC amplitude, tau, charge, frequency, and cell RMP. (K-L) Representative amplitude traces recorded in regular aCSF (blue) and after DOI (red). (M-N) Representative frequency traces recorded in regular aCSF (blue) and after DOI (red). (O-S). A summary of sEPSC and RMP regulation by different DOI concentrations. The effects at each concentration are expressed as percent change from baseline values prior to DOI application in each cell. *p < 0.05; **p < 0.005; ***p < 0.0005; ****p < 0.0005, paired Student’s t-tests. n = 11 cells from 3 animals in the 300 nM group and n = 10 cells from 3 animals in the 1 μM group. 10 μM data is from Figure 7.

Supplemental Figure 4. Low dose of SB 242–084 does not affect CLA-ACC excitability on its own, but continues to block DOI synaptic effects. (A-E) sEPSC amplitude, tau, charge, frequency, RMP effects of SB 242,084 (1 μM) and DOI (1 μM) on CLA-ACC neurons. (F) Representative amplitude traces of CLA-ACC neurons recorded in regular aCSF (blue), after SB (red), and after SB with DOI (green). (G) Representative frequency traces of CLA-ACC neurons recorded in regular aCSF (blue), after SB (red), and after SB with DOI (green). 1-way RM ANOVA indicated no significant effects between groups on any of the measured excitability parameters. n = 11 cells from 4 animals.

Supplemental Table 1. Forward and reverse primer sequences for qPCR experiments. 5-htr1–7, 5-HT receptors; Tbp, TATA-binding protein.

Supplemental Table 2. mRNA levels of individual 5-HT receptor genes in the CLA, ACC, and insula (INS). Transcript expression for genes of interest (GOI) is quantified as arbitrary fluorescence units (afu). N is number of animals. CLA vs ACC: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001, unpaired Student’s t-tests. CLA vs INS: ^#p < 0.05; ^##p < 0.01; ^###p < 0.001; ^####p < 0.0001, unpaired Student’s t-tests.

Highlights.

• The claustrum expresses large amounts of 5-HT2A and 5-HT2C receptors.

• Nearly half of claustrum glutamatergic neurons express 5-HT2 receptors, while only ~15–20% of GABAergic interneurons express 5-HT2 receptors.

• Serotonin inhibits glutamatergic synaptic events and intrinsic excitability of CLA-ACC neurons through the 5-HT1A and 5-HT2C receptors.

• The psychedelic drug, DOI excites CLA-ACC glutamatergic post-synaptic events via the 5-HT2C and not the 5-HT2A receptors.

Data Availability

Data will be made available on request.