Fenfluramine diminishes NMDA receptor-mediated seizures via its mixed activity at serotonin 5HT2A and type 1 sigma receptors

Abstract

Fenfluramine exhibits antiepileptic properties and thus diminishes epileptiform discharges in experimental animal models of Dravet syndrome. Fenfluramine is metabolized into norfenfluramine in vivo, which shows greater affinity and agonist activity at serotonin 5HT2 receptors (5HT2R) than fenfluramine. In this study, we found that fenfluramine and norfenfluramine disrupted the regulatory association of the sigma 1 receptor (σ1R) with NR1 subunits of glutamate N-methyl-D-aspartate receptors (NMDAR), an effect that was also produced by σ1R antagonists such as S1RA and prevented by σ1R agonists such as PPCC. The antagonists removed σ1R bound to NMDAR NR1 subunits enabling calcium-regulated calmodulin (CaM) to bind to those subunits. As a result, CaM may inhibit calcium permeation through NMDARs. The serotoninergic activity of fenfluramine at 5HT2AR, and likely also at 5HT2CR, collaborated with its activity at σ1Rs to prevent the convulsive syndrome promoted by NMDAR overactivation. Notably, fenfluramine enhanced the inhibitory coupling of G protein-coupled receptors such as 5HT1AR and cannabinoid type 1 receptor with NMDARs, thus allowing the more effective restrain of NMDAR activity. Thus, fenfluramine circumvents the negative side effects of direct NMDAR antagonists and may improve the quality of life of subjects affected by such proconvulsant dysfunctions.

Keywords: HINT1 protein; fenfluramine; glutamate N-methyl-D-aspartate receptor; seizures; type 1 sigma receptor.

Figure 1. Fenfluramine and norfenfluramine disrupt the association of σ1Rs with NR1 subunits of NMDA receptors

(A) In vitro assay to determine ligand activity at σ1Rs. NHS-activated Sepharose beads covalently coupled to a sequence of NR1 subunits containing the transmembrane region plus C0–C1–C2 cytosolic segments were incubated with excess σ1Rs (1:3). The unbound σ1Rs were washed out, and the NR1-coupled σ1Rs were exposed to serial concentrations of the ligands under study. The σ1Rs that remained attached to the NR1 subunits were then evaluated by SDS-PAGE and immunoblotting. (B) Diminishing effect of ligands on the association of σ1Rs with NR1 subunits. The assays were performed in the presence of 50 mM Tris–HCl (pH 7.5), 0.2% CHAPS and 2.5 mM calcium. Representative blots are shown. The ED₅₀ values were computed using the software SigmaPlot V.13. (C) The σ1R agonist PPCC diminishes the capacity of S1RA and fenfluramine to disrupt σ1R–NR1 association. (D) S1RA, fenfluramine and norfenfluramine, but not PPCC, decreased the association of σ1Rs with CaM binding site on NR1 subunits. Agarose-NR1 carrying the associated σ1Rs was incubated with 2.5 mM CaCl₂, in the absence or presence of 100 nM CaM. The assays were performed twice, and each point was duplicated. ^*Significant difference with respect to the control group, ϕ with respect to the group receiving only S1RA or fenfluramine; ANOVA, Dunnett multiple comparisons vs control group, p < 0.05.

Figure 2. Anticonvulsant effect of fenfluramine and S1RA in a mouse model of seizures induced by NMDAR over activation

(A) Behavioral alterations produced by icv administration of NMDA to mice pretreated (–24 h) with the opioid agonist morphine (10 nmol, icv). (B) Effects of S1RA, fenfluramine and norfenfluramine (3 nmol each, icv) on seizures induced by NMDA (300 pmol, icv) in morphine-primed mice. The drugs were injected icv 30 min before the administration of the NMDA receptor agonist. ^*Significant difference from the control group, which received morphine and 24 h later NMDA, but saline instead of treatment; ϕ indicates that 3 nmol icv PPCC or 3 nmol icv 4F 4PP prevented the diminishing effects of S1RA/fenfluramine/norfenfluramine on the corresponding NMDA-induced behavioral sign. (C) Effects of the treatments on the latency and duration of the seizure episodes. ^*Significant difference from the control group receiving saline. ANOVA, Dunnett multiple comparisons vs control group, p < 0.05. (A–C) Each bar indicates the percentage of mice showing the indicated sign and represents the mean ± SEM of 9 mice.

Figure 3. Effects of fenfluramine and S1RA on phosphorylation associated with the NMDAR/CaMKII pathway

The mice received icv fenfluramine or S1RA at the doses indicated, and after 1 h, the animals were euthanized and their cerebral cortices obtained for ex vivo examination. An activator of NMDARs was administered 1 h before euthanasia, 15 min before S1RA or fenfluramine. The effects of these treatments on NMDAR-related molecular changes were determined by western blot analysis of cortical synaptosomal membranes, (A–C) Regulatory phosphorylations on NMDAR NR1 subunits (S: serine), and NR2B subunits (Y: tyrosine). (D) Phosphorylation of threonine (T) 286 CaMKII. Immunosignals (average optical density of the pixels within the object area/mm²) are expressed as the change relative to the control group (C), which is assigned an arbitrary value of 1. Actin was used as a loading control, and the immunosignals did not differ by more than 10% (Supplementary Figure 2). Each bar represents the mean ± SEM of 6 mice. ^*Significant difference with respect to the control group (C), which received only saline; ϕ significant effect of S1RA or fenfluramine on the activity of NMDA; ANOVA, Dunnett multiple comparisons vs control group, p < 0.05.

Figure 4. Effect of fenfluramine and S1RA on the overactivity of NMDARs promoted by NMDA in morphine-primed mice

(A) Morphine-primed mice exhibited a high level of CaMKII auto phosphorylation in response to NMDA. This effect was not evoked by fenfluramine or S1RA, which instead diminished P-T286 CaMKII levels in those morphine and NMDA-treated mice. The assay was repeated at least twice. The immunosignals are expressed relative to the control (C, value of 1), which received saline instead of the treatments indicated. (B) Effect of the different treatments on NMDAR phosphorylation and on P-T286 CaMKII. All substances were given icv, and the mice were killed at the post-drug intervals indicated, e.g., 30 min after NMDA and 24 h after morphine. The combined treatments started with morphine, and 24 h later, fenfluramine or S1RA was administered 45 m before NMDA. ^* Significant difference with respect to the group that received only morphine; ϕ significant difference respect to the morphine-primed group that received only NMDA 24 h later; ANOVA, Dunnett multiple comparisons vs control group, p < 0.05. (C) Effect of the σ1R agonist PPCC on the diminished effect of fenfluramine and S1RA on CaMKII hyper phosphorylation evoked by NMDA. The mice were primed with morphine, and 24 h later, the effects of the different drugs on P-T286 CaMKII were evaluated. ^*Significant difference with respect to the morphine-primed group that received saline instead of the drug and NMDA 24 h later; ϕ significant difference with respect to the group that received fenfluramine or S1RA but not PPCC; ANOVA, Dunnett multiple comparisons vs control group, p < 0.05. (A–C) Each bar represents the mean ± SEM of 6 mice. Details as in Figure 3 and Supplementary Figure 3.

Figure 5. Fenfluramine transfers control of NMDARs and HINT1 proteins to other GPCRs such as 5HT1ARs, CB1Rs or MORs. Comparative study with morphine and S1RA

Fenfluramine (B, E, F, G) and morphine (A, C, D) were both icv injected into the mice at 10 nmol; S1RA (H) was icv injected at 3 nmol. At the intervals indicated, the animals were euthanized to obtain the synaptosomal cortical fraction. The GPCRs were immunoprecipitated (IP), and the associated NR1 subunits and HINT1 protein were assessed by western blot analysis. Each bar represents the mean ± SEM of 8 mice. ^*Significant difference with respect to the control group, which received saline instead of the drug; ANOVA, Dunnett multiple comparisons vs control group, p < 0.05. Details are in the Methods section.

Figure 6. Modeling fenfluramine activity at 5HT2A/C and σ1 receptors. Influence on negative control of NMDAR function by the 5HT1A receptor

(A) A certain pool of 5HT2A/C receptors is complexed with NMDARs via HINT1 proteins and σ1Rs. Fenfluramine binds and activates 5HT2A/C receptors, promoting PKC/Src priming of the coupled NMDARs. The binding of antagonists such as fenfluramine to σ1Rs prevents σ1Rs from binding to NMDARs [37], enabling HINT1 transfer from the GPCR to NMDAR NR1 subunits. The phosphorylation of 5HT2A/CR cytosolic residues inhibits its re-association with HINT1 to recruit further NMDAR activity. Because HINT1 allows CaM to access NR1 subunits [36], the inactive HINT1-NMDARs can now couple with other GPCRS such as CB1Rs and 5HT1ARs. (B) While 5HT1ARs inhibit NMDAR activity, 5HT2A/CRs promote NMDAR function directly and furthermore, promote it indirectly through the negative control of 5HT1AR signaling. Exaggerated activation of NMDARs may promote seizures. Antagonists of σ1Rs (At) remove σ1Rs from NR1 subunits and enable negative control by CaM; the incidence of convulsive episodes diminishes. The combined activity of fenfluramine as agonist at 5HT2R and antagonist at σ1R uncouples 5HT2Rs from positive control of NMDAR function and increases the inhibitory coupling of GPCRs, such as the 5HT1AR, to NMDAR function. In addition, fenfluramine, as a σ1R antagonist, facilitates CaM inhibition of already primed and GPCR-freed NMDARs. Fenfluramine, by triggering these mechanisms, efficaciously reduces the incidence of epileptogenic episodes. A and B, ^*indicates activation; B, ^*indicates higher activity than^*.