GPR56/ADGRG1 is associated with response to antidepressant treatment

Abstract

It remains unclear why many patients with depression do not respond to antidepressant treatment. In three cohorts of individuals with depression and treated with serotonin-norepinephrine reuptake inhibitor (N = 424) we show that responders, but not non-responders, display an increase of GPR56 mRNA in the blood. In a small group of subjects we also show that GPR56 is downregulated in the PFC of individuals with depression that died by suicide. In mice, we show that chronic stress-induced Gpr56 downregulation in the blood and prefrontal cortex (PFC), which is accompanied by depression-like behavior, and can be reversed by antidepressant treatment. Gpr56 knockdown in mouse PFC is associated with depressive-like behaviors, executive dysfunction and poor response to antidepressant treatment. GPR56 peptide agonists have antidepressant-like effects and upregulated AKT/GSK3/EIF4 pathways. Our findings uncover a potential role of GPR56 in antidepressant response.

Trial registration: ClinicalTrials.gov NCT00635219 NCT00599911 NCT01140906 NCT02209142.

Fig. 1. GPR56 mRNA is related to antidepressant response.

a In the discovery cohort, 237 patients in a major depressive episode were randomized to double-blind treatment with either duloxetine (n = 112) or placebo (n = 125), for up to 8 weeks. Using two-class paired significant analysis of microarray (SAM) with correction for multiple testing (FDR < 1%) in patients who responded to duloxetine, GPR56 mRNA is the most significantly upregulated mRNA in whole blood after duloxetine treatment, based on fold change and q-value (FC = 1.19, q-value < 0.01). General linear model (GLM) demonstrated a time x treatment interaction, F(1,199)=8.468, p = 0.004, that confirms the specificity of GPR56 mRNA increase in responders to duloxetine. b In the first replication cohort (Montréal), patients were treated with citalopram in an open-label trial. GPR56 mRNA in whole blood demonstrated an increase only in responders. GLM demonstrated a time x group interaction, F(1,61)=4.27, p = 0.043 (not adjusted for multiple testing). c In the second replication cohort (Marseille), psychiatrically healthy subjects and patients with depression were included in a naturalistic design. In patients who responded and achieved remission after 30 weeks of treatment (responders-remitters, n = 20), GPR56 mRNA is not different at inclusion and 2 weeks, however was then overexpressed at 8 weeks in comparison to others (n = 44) (two-sided t-test t = 2.085, p = 0.049) and remained stably overexpressed over a 30-week follow-up (n = 18 responders-remitters in comparison to others n = 31) two-sided t-test t = 3.076, p = 0.005); Linear Mixed model (F(1,230.199) = 14.79, p = 0.0001). Bars represent mean. Error bars represent standard error of the mean. **p < 0.01, *p < 0.05. Source data are provided as a Source Data file.

Fig. 2. Unpredictable chronic mild stress (UCMS) and antidepressant response dysregulate Gpr56 mRNA in blood and CNS in mice.

a Gpr56 expression was analysed by qRT-PCR in blood and several brain regions, including the dorsal and ventral hippocampal areas (HD and HV, respectively), the prefrontal cortex (PFC) and the Nucleus Accumbens (NAcc) in non-stressed or stressed mice (exposed to UCMS) and receiving vehicle or fluoxetine. Mice treated by fluoxetine were classified as “responders” or “non-responders” according to behavioral tests. b In whole blood, a one-way ANOVA showed between group differences for Gpr56 expression (F = 6,150, p = 0.001). Blood Gpr56 mRNA expression was decreased in mice subjected to UCMS, while reversal of depressive-like behaviors with fluoxetine was paralleled by normalization of blood Gpr56 mRNA expression in responder mice (post hoc analysis p < 0.01). c In brain, a two-way ANOVA between group and brain regions showed a significant interaction between brain region and mice group (F(9,151) = 3.112; p = 0.0018). Post hoc analysis demonstrated a specific PFC effect, a decrease of Gpr56 in PFC between stressed and non-stressed mice, with a reversal effect of antidepressant only in responder mice. Sample numbers vary between tissues due to removal of poor quality RNA samples from the analyses. Bars represent mean. Error bars represent standard error of the mean. **p < 0.01. Graph represents Box and Whiskers Min to Max. Source data are provided as a Source Data file.

Fig. 3. Gpr56 regulates depressive-like behaviors.

a Mice were injected bilaterally in the pre-frontal cortex with control (CTL) or lentivirus-Gpr56 constructs, as well as with CTL or lentivirus-sh-Gpr56 (inhibitor) constructs. b Overexpression of Gpr56 (n = 15) was associated with lower immobility time in the TST in comparison to control animals (n = 12), two-sided t = 3.07, p = 0.005. c Downregulation of Gpr56 (n = 8) produced increased immobility time, i.e. depressive-like behavior in the tail suspension test (TST) in comparison to control animals (n = 11), two-sided t = 2.203, p = 0.048. d Fluoxetine decreased immobility in control animals (n = 5), while this effect was strongly attenuated in sh-Gpr56-virus infused animals (n = 6); two-way ANOVA for repeated measures; treatment × group interaction F(1,9) = 6.80, p = 0.028; main effects for group F (1,9) = 25.4, p = 0.001 (as compared to control) and for treatment F(1,9) = 6.80 p = 0.001 (as compared to vehicle). e GPR56 agonist P7 has antidepressant-like effects. GPR56 agonist P7 peptide infused bilaterally with escalating doses in PFC decreases immobility time and demonstrates antidepressant-like effects, ANOVA F(1,4) = 4.88 p = 0.008. In comparison to vehicle (n = 13), both 1 mM (n = 7) and 2 mM (n = 4) doses demonstrate a significant decrease of immobility time (p < 0.01 and p = 0.01 respectively, two-sided post hoc test). Bars represent mean. Error bars represent standard error of the mean. **p < 0.01, *p < 0.05. Source data are provided as a Source Data file.

Fig. 4. GPR56 expression is altered in the prefrontal cortex from post-mortem brain tissue of individuals with depression and is related to pre-frontal cortex functioning in patients.

a GPR56 expression was measured by qRT-PCR in post-mortem brain tissue (BA44). Expression was lower in individuals with depression (MDD, n = 49) in comparison to psychiatrically healthy controls (n = 26). FC = 0.56, two-sided U = 385, Z = −2.81, p = 0.005. Graph represents Box and Whiskers Min to Max. b Changes in Stroop interference score, a neuropsychological test that involves pre-frontal cortex function, were correlated with changes in GPR56 expression in whole blood following antidepressant treatment (n = 12 from discovery cohort, Pearson coefficient of correlation = −0.71, two-sided p = 0.009). Reduction of interference score was associated with an improvement of Stroop interference test, i.e., an improvement in pre-frontal cortex functioning. Source data are provided as a Source Data file.