Corticosterone stimulates synthesis of 2-arachidonoylglycerol via putative membrane-bound glucocorticoid receptors and inhibits GABA release via CB1 cannabinoid receptors in the ventrolateral periaqueductal gray

Abstract

The ventrolateral periaqueductal gray (vlPAG) plays a critical role in pain modulation. GABAergic neurotransmission within the vlPAG regulates the descending pain pathway. This study investigates the mechanisms through which corticosterone (CORT) modulates GABA release in the vlPAG via putative membrane-associated glucocorticoid receptors (mbGRs). Superfusion of CORT decreases evoked inhibitory postsynaptic currents in a mbGR- and CB1 cannabinoid receptor (CB1R)-dependent manner. Using a depolarization-induced suppression of inhibition protocol to test the effects of CORT on the endocannabinoid system, we find that CORT-mediated signaling enhances 2-arachidonoylglycerol synthesis that is inhibited by the diacylglycerol lipase inhibitor, DO34. CORT prolongs CB1R activation through a Gα_s and protein kinase A-dependent pathway, whereas early depolarization-induced suppression of inhibition-initiated endocannabinoid activation of CB1Rs is independent of protein kinase A. These results highlight the critical role of CORT in the vlPAG in engaging endocannabinoid pathways to inhibit GABA release. The results indicate that CORT activation of putative mbGRs promote activation of the descending pain modulatory pathway through CB1R-mediated inhibition of GABA release in the vlPAG. SIGNIFICANCE STATEMENT: This study provides evidence that corticosterone activates putative membrane glucocorticoid receptors to increase levels of 2-arachidonoylglycerol to activate presynaptic CB1 cannabinoid receptors. These findings reveal mechanisms by which stress modulates the ventrolateral periaqueductal gray and the descending pain circuit.

Keywords: Corticosterone; Diacylglycerol lipase; Endocannabinoid; Periaqueductal gray; Protein kinase A.

Figure 1.. CORT-activated mbGRs indirectly inhibit GABA release via CB1R activation in vlPAG.

(A) Representative eIPSC traces at baseline: NBQX, 5 µM (black); CORT, 1 µM (blue), and RU486, 3 µM (red). (B) Inhibition of eIPSCs by CORT (1 µM) is blocked by pre-incubation of aCSF solution containing RU486 (3 µM) and rimonabant (RIM; 3 µM). The CB1R agnist WIN 55212–2 inhibts eIPSCs in the presence of RU486. One-way ANOVA F_(3,33)= 17.44, P < 0.0001. Dunnett’s posthoc test (p values on graph). (C) Percentage inhibition of eIPSCs by DEX:BSA (1 µM) compared to BSA (1 µM) alone, and DEX-BSA in the presence of RU486 (3 µM) and rimonabant (RIM; 3 µM). One-way ANOVA F_(3,25)= 15.48, P < 0.0001. Dunnett’s posthoc test (p values on graph). Error bars represent the SEM, dots indicate individual recordings, and numbers represent the number of rats per experiment.

Figure 2.. CORT prolongs CB1R activation by stimulating 2-AG synthesis.

(A) Summary of DSI (+20 mV for 5 s) in vlPAG slices from naive rats without drug exposure (black line; n= 20 recordings from 10 rats) and after CORT 1 µM (blue line; n= 18 recordings from 9 rats). (B) Quantification of inhibition of eIPSC amplitudes during early and late time windows (respectively after 5–25 s and 30–50 s after the depolarization step). Two-way repeated-measures ANOVA; main effect of drug: F_(1,36) = 5.94, p = 0.02; main effect of time: F_(1,36) = 13.82, P = 0.0007; interaction: F_(1,36) = 7.77, p = 0.0084; Sidak’s multiple comparisons, p values on graph. (C) Quantification of inhibition of eIPSC amplitudes during early and late time windows on slices exposed to CORT 1 µM with or without co-exposure of DO34 1 µM (n= 10 recordings from 5 rats). Two-way repeated-measures ANOVA; main effect of drug: F_(1,18) = 50.71, p < 0.0001; main effect of time: F_(1,18) = 1.81, p = 0.20; interaction F_(1,18) = 5.25, p = 0.034; Sidak’s multiple comparisons, p values on graph. Error bars represent SEM, dots indicate individual recordings, and numbers represent the number of rats represented per bar.

Figure 3.. CORT stimulates DAGL activity via a PKA-dependent mechanism.

(A) Summary of DSI (+20 mV for 5 s) from vlPAG recordings in the absence (white circles; n = 15 recordings from 5 rats) and presence of PKI 300 nM in the recording electrode (white squares; n = 13 recordings from 4 rats). (B) Quantification of inhibition of eIPSC amplitudes during early and late time windows. (C-D) Summary of DSI after 5–10 min bath application of CORT 1 µM (circles; n = 17 recordings from 6 rats), with PKI 300 nM in the recording electrode (squares; n = 12 recordings from 4 rats), and with PKI 1 µM added in perfusion solution with CORT (triangle; n= 11 recordings from 3 rats). Panels A-D were analyzed with Two-way repeated-measures ANOVA; main effect of treatment: F_(4,63) = 3.14, p = 0.02; main effect of time: F_(1,63) = 53.24, P < 0.0001; interaction: F_(4,63) = 1.69, P = 0.16; Tukey post hoc test, p values on graphs. (E). Representative eIPSCs showing blockade of CORT-mediated inhibition after incubation of slice in PKI (1 µM). (F) Compiled percent inhibition of eIPSCs by CORT in the absence and presence of PKI and a Gs peptide inhibitor (One-way ANOVA, F_(2,15) = 26.64, P < 0.0001; Dunnett’s post-hoc test, p values on graph. (G) sIPSC frequencies are not different at baseline and after bath application of PKI 1 µM from slices of naive rats (Paired t-test, t₆ = 1.09, p = 0.32). (H) Representative traces of sIPSCs in absence and presence of PKI. Error bars represent the SEM, dots indicate individual recordings, and numbers represent the number of rats represented per bar.