Profiling the Impact of mGlu7/Elfn1 Protein Interactions on the Pharmacology of mGlu7 Allosteric Modulators

Abstract

The group III metabotropic glutamate receptors (mGlu receptors) are predominantly expressed presynaptically throughout the central nervous system (CNS) where they regulate the release of glutamate and GABA. These receptors have recently been shown to be anchored by transsynaptic expression of the laminin proteins ELFN1 and ELFN2. In particular, the mGlu₇ receptor is localized at presynaptic active zones from pyramidal cells to somatostatin-containing interneurons with postsynaptic ELFN1, and this interaction drives the rapidly facilitating nature of these synapses in the hippocampus and cortex. Interestingly, individuals with mutations in ELFN1 or GRM7 genes present with attention-deficit hyperactivity disorder and epilepsy, and knockout mice of each of these proteins develop seizures with very similar time courses. In the current manuscript, we explore the hypothesis that the pharmacology of positive and negative allosteric modulators (PAMs and NAMs) of mGlu₇ might be changed in the presence of ELFN1. These results showed that, across a range of NAMs, we observed similar efficacy in the presence of ELFN1. For PAMs, we observed decreased maximal potentiation when ELFN1 was present, but all examined compounds were still able to potentiate receptor signaling regardless of ELFN1 expression. Finally, we confirm that a tool PAM with mGlu₇ activity is able to potentiate responses at pyramidal cell-somatostatin interneuron synapses where ELFN1 is expressed. These results suggest that, for the modulators shown here, native tissue activity should be retained in the presence of ELFN1 expression.

Keywords: ELFN1; ELFN2; allosteric modulator; interneuron; mGlu7; mGlu8; metabotropic glutamate; somatostatin.

1

(A) Schematic of interactions between mGlu₇ (blue) and Elfn1 (peach/green). Glutamate or the surrogate agonist L-AP4 is represented by red circles, Venus flytrap domain (VFD) allosteric modulator by pink triangles, and 7-transmembrane domain (7TMD) allosteric modulator by purple stars. (B) Mapping of relevant binding sites and glycosylation sites on mGlu₇ (blue). Glycosylation/Elfn1 sites (orange), XAP044 binding sites (light pink), XAP044/LSP4-2022 binding site overlap (dark pink), and l-AP4/glutamate binding sites (red) are differentiated by color. Prepared using PyMOL Molecular Graphics System, Version 3.1 Schrödinger, LLC. Binding sites based on refs − .

2

Assay design and workflow. (Left panel) Cells expressing rat mGlu₇/GIRK (pink) were cocultured with pcDNA3 (blue) or Elfn1 (purple). (Middle panel) Activity was assessed using a thallium flux assay where the compound or vehicle was added at t = 1 s followed by the agonist at t = 141 s. The slope value for each kinetic trace was calculated for the time window of 145–155 s. (Right panel) The effect on the concentration–response curve in the presence (purple) or absence (blue) of Elfn1 can be seen by a change in efficacy, potency, and/or basal response. Note that while the decrease in basal response was reported for other mGlu receptors (Dunn et al., 2018), we did not observe this effect in the assay performed here.

3

Coculture of mGlu₇/GIRK-expressing HEK293 cells with HEK293 cells expressing Elfn1 results in a decreased maximal agonist response and a rightward shift in agonist potency. (A) Increasing concentrations of the mGlu₇ agonist l-AP4 were applied to mGlu₇/GIRK cells cocultured with cells expressing a pcDNA3 vector control (black symbols) or an Elfn1/pcDNA3 construct (white symbols) and thallium flux through GIRK channels was measured. (B) Maximal responses to l-AP4 were significantly reduced in the presence of Elfn1 (black versus white bars, ***p < 0.001, Students t test). (C) pEC₅₀ of l-AP4 was significantly reduced in the presence of Elfn1 (**p < 0.01, Students t test). Data represent eight individual experiments performed in duplicate or triplicate and are shown as mean ± standard error of the mean (SEM).

4

Compound structures.

5

Three distinct antagonists that interact with different domains on the mGlu₇ receptor show similar profiles in the absence and presence of Elfn1. Top row = orthosteric antagonist, LY341495; middle row = Venus flytrap domain-interacting antagonist, XAP044; bottom row = transmembrane domain negative allosteric modulator, ADX71743. Red = pcDNA + dimethyl sulfoxide (DMSO), black = Elfn1 + DMSO, blue = pcDNA + antagonist, green = Elfn1 + antagonist. DMSO vehicle or 10 μM antagonist was added to mGlu₇/GIRK cells cocultured with cells expressing a pcDNA3 vector control or an Elfn1/pcDNA3 construct. Increasing concentrations of l-AP4 were added and thallium flux through GIRK channels was measured. (A, E, I) Full concentration–response curves. (B, F, J) Maximal responses were quantified by averaging the responses of the top two concentrations of l-AP4. Paired t test between vehicle and antagonist conditions, *p < 0.05; **p < 0.01; ****p < 0.0001; ns = not significant. (C, G, K) Percent inhibition between the pcDNA (blue) and Elfn1 (green) conditions were measured. Paired t test between conditions, *p < 0.05, ns = not significant. (D, H, L) Full concentration–response curves of antagonists were performed in the presence of 600 μM l-AP4, and responses were normalized to the response of 600 μM l-AP4 in the presence of pcDNA3 alone. Data represent three experiments performed in duplicate or triplicate and are shown as mean ± SEM.

6

Structure–activity relationships of a range of mGlu₇ NAMs are maintained in the presence of Elfn1. Red = pcDNA + DMSO, black = pcDNA + antagonist, blue = Elfn1 + DMSO, green = Elfn1 + antagonist. DMSO vehicle or 10 μM antagonist was added to mGlu₇/GIRK cells cocultured with cells expressing a pcDNA3 vector control or an Elfn1/pcDNA3 construct. Increasing concentrations of l-AP4 were added, and thallium flux through GIRK channels was measured for a range of NAMs. Data represent three experiments performed in duplicate or triplicate and are shown as mean ± SEM.

7

Across all compounds, there is a strong correlation in blocking mGlu₇-mediated response in the absence or presence of Elfn1. The maximal responses of all NAMs examined in Figures and were plotted for activity in the presence of pcDNA3 (x-axis) or Elfn1 (y-axis) and a linear regression was performed. r ² = 0.88, ****p < 0.0001.

8

All examined PAMs potentiate mGlu₇ responses in the presence or absence of Elfn1. Red = pcDNA3 + DMSO, blue = pcDNA3 + PAM, black = Elfn1 + DMSO, green = Elfn1 + PAM. DMSO vehicle or 10 μM PAM was added to mGlu₇/GIRK cells cocultured with cells expressing a pcDNA3 vector control or an Elfn1/pcDNA3 construct. Increasing concentrations of l-AP4 were added and thallium flux through GIRK channels was measured. Data represent three experiments performed in duplicate or triplicate and are shown as mean ± SEM.

9

All tested PAMs potentiate EC₂₀ l-AP4 responses in the presence of Elfn1 and VU0422288 is active with glutamate. Increasing concentrations of (A) VU6005649, (B) VU0155094, (C) VU0422288, and (D) VU6027459 were added to mGlu₇/GIRK cell cocultures with either pcDNA3 (black) or Elfn1 (white) and thallium flux was measured. (E) Fold potentiation of an EC₂₀ concentration of l-AP4 was calculated, showing that all PAMs exhibited lower maximal levels of potentiation when Elfn1 is present. Unpaired t tests between vehicle and PAM conditions, *p < 0.05, **p < 0.01 between the pcDNA (black) and Elfn1 (white) conditions. (F) VU0422288 potentiates glutamate responses in the presence of Elfn1. **p < 0.01, ****p < 0.0001. Data represent three experiments performed in duplicate or triplicate and are shown as mean ± SEM.

10

The PAM VU0422288 is active in potentiating responses at SST-INs. (A) Left, Example of averaged traces of 5-pulse-evoked EPSCs recorded from SST-INs in the PFC during baseline (BL, black) and application of 30 μM l-AP4 (light red); and right, during baseline (black), 10 μM VU0422288 (VU288, green), and a combination of VU0422288 and 30 μM l-AP4 (dark red). Scale bars: 50 pA/20 ms. (B) Bar graphs summarizing the normalized amplitude of EPSC₁ in baseline and l-AP4 conditions (left, ****p < 0.0001, paired t test, n = 8), and in baseline, VU288, and a combination of VU0422288 and l-AP4 conditions (right). Repeated measures (RM) one-way analysis of variance (ANOVA), F(1.572, 6.288) = 73.09, ****p < 0.0001, with post hoc Dunnett’s test; middle, ***p < 0.0005, n = 5. (C) Summary of percentage change of EPSC₁ amplitude during application of 30 μM l-AP4 in the absence and presence of VU0422288 (*p < 0.05, unpaired t test; n = 8 and 5, respectively).