Endogenous 24S-hydroxycholesterol modulates NMDAR-mediated function in hippocampal slices

Abstract

N-methyl-D-aspartate receptors (NMDARs), a major subtype of glutamate receptors mediating excitatory transmission throughout the central nervous system (CNS), play critical roles in governing brain function and cognition. Because NMDAR dysfunction contributes to the etiology of neurological and psychiatric disorders including stroke and schizophrenia, NMDAR modulators are potential drug candidates. Our group recently demonstrated that the major brain cholesterol metabolite, 24S-hydroxycholesterol (24S-HC), positively modulates NMDARs when exogenously administered. Here, we studied whether endogenous 24S-HC regulates NMDAR activity in hippocampal slices. In CYP46A1(-/-) (knockout; KO) slices where endogenous 24S-HC is greatly reduced, NMDAR tone, measured as NMDAR-to-α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) excitatory postsynaptic current (EPSC) ratio, was reduced. This difference translated into more NMDAR-driven spiking in wild-type (WT) slices compared with KO slices. Application of SGE-301, a 24S-HC analog, had comparable potentiating effects on NMDAR EPSCs in both WT and KO slices, suggesting that endogenous 24S-HC does not saturate its NMDAR modulatory site in ex vivo slices. KO slices did not differ from WT slices in either spontaneous neurotransmission or in neuronal intrinsic excitability, and exhibited LTP indistinguishable from WT slices. However, KO slices exhibited higher resistance to persistent NMDAR-dependent depression of synaptic transmission induced by oxygen-glucose deprivation (OGD), an effect restored by SGE-301. Together, our results suggest that loss of positive NMDAR tone does not elicit compensatory changes in excitability or transmission, but it protects transmission against NMDAR-mediated dysfunction. We expect that manipulating this endogenous NMDAR modulator may offer new treatment strategies for neuropsychiatric dysfunction.

Keywords: 24S-hydroxycholesterol; CYP46A1 knockout mice; NMDAR; hippocampal slice.

Fig. 1.

Decreased N-methyl-d-aspartate receptor (NMDAR) positive tone in CYP46A1 knockout (KO) slices. A: endogenous 24S-hydroxycholesterol (24S-HC) was significantly reduced in KO tissue and slices [N = 3 wild-type (WT) and 3 KO animals; ***P < 0.001]. B: dual-component NMDAR and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) excitatory postsynaptic currents (EPSCs; red) were evoked in WT and KO slices at −30 mV. NMDAR EPSCs, measured 50 ms following peak EPSC, were completely blocked following 50 μM d-(−)-amino-5-phosphonopentanoic acid (d-APV) application (black). AMPAR EPSC peak amplitude was calculated after d-APV application. C: NMDAR/AMPAR ratio was compared between WT and KO slices (N = 15 cells from WT and 9 cells from KO slices; *P < 0.05).

Fig. 2.

Induced tonic NMDAR currents are smaller and represent decreased channel open probability in KO slices. A and B: WT and KO responses to dl-threo-β-Benzyloxyaspartic acid (TBOA; 50 μM) application in the presence of AMPAR and γ-aminobutyric acid receptor (GABAR) blockers {10 μM 2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide (NBQX) and 100 μM picrotoxin (PTX)}, followed by coapplication of 40 μM MK-801. The solid red line is a fit to a single-exponential decay. C: summary of TBOA induced current (N = 13 WT and 13 KO cells, *P < 0.05). D: time constant of current decay in the presence of MK-801 (n = 9 WT and 11 KO cells; **P < 0.01). The time constant was not obtainable in 4 WT and 2 KO cells due to a large, prolonged EPSC at the onset of MK-801 application or to seal loss in this subset of cells. Although the source of the contaminating current is not completely clear, it may represent network disinhibition upon initial MK-801 application.

Fig. 3.

NMDAR-induced spiking is reduced in CYP46A1 KO slices. A and B: excitatory postsynaptic potentials (EPSPs) and associated action potentials were evoked in WT (A) and KO (B) slices at −65 mV during a 10-pulse, 50-Hz presynaptic stimulus train. Neither NBQX nor d-APV was present at baseline. EPSPs were then reevoked with the same stimulus parameters after 50 μM d-APV wash-in. C: spike number at baseline in WT or KO neurons. D: input-output curve for spikes evoked by repetitive synaptic stimulation in WT and KO neurons at baseline. WT and KO neurons share similar maximum firing capacity at a saturated level of synaptic stimulation (N = 18 WT and 17 KO cells). E: normalized (after/before d-APV) spike number following d-APV wash-in showed a weaker d-APV effect on KO slices (N = 18 cells from WT and 17 cells from KO slices; *P < 0.05).

Fig. 4.

Endogenous 24S-HC does not saturate oxysterol-mediated potentiation, and is selective for NMDAR EPSCs. A: NMDAR EPSCs were evoked at −30 mV in WT and KO slices in the presence of 100 μM PTX and 10 μM NBQX. After a 10-min stable baseline was established (black traces), SGE-301 was bath-applied for 15–30 min. After 6 min of SGE-301 perfusion, potentiation of EPSCs had stabilized and was averaged from traces in the next 10 min (red traces). B: summary of SGE-301 (10 μM) effect in WT and KO slices. Data from 10–15 min following drug onset were averaged for each cell and expressed relative to the average baseline responses (10 min) before drug delivery. A two-way ANOVA and Bonferroni post hoc analysis revealed a significant effect of SGE-301 (df = 1; F = 7.8; *P < 0.01) but no effect of genotype and no significant interaction between genotype and drug. (N = 10 cells from WT and 10 cells from KO in DMSO group; N = 13 cells from WT and 14 cells from KO slices in SGE-301 group; df = 1; F = 0.8; P > 0.2). C: AMPAR EPSCs were evoked at −70 mV in KO slices in the presence of 100 μM PTX and 50 μM d-APV. After a 10 min stable baseline was established (black trace), SGE-301 was bath-applied for 15–30 min. Potentiation was measured after 6 min of SGE-301 perfusion (red trace). D: GABAR IPSCs were evoked at 0 mV in KO slices in the presence of 10 μM NBQX and 50 μM d-APV. After a 10-min stable baseline was established (black trace), SGE-301 was bath-applied for 15–30 min. Potentiation was measured after 6 min of SGE-301 perfusion (red trace). E: normalized (after/before) peak amplitude was compared for NMDAR, AMPAR, and GABAR PSCs after 6 min SGE-301 application (N = 14 cells for NMDAR, 7 cells for AMPAR, and 11 cells for GABAR; Wilcoxon test ***P < 0.001).

Fig. 5.

Intrinsic excitability is unaffected by CYP46A1 deletion. A and B: a 500-ms current pulse, ranging from 0 to 800 pA to evoke trains of action potentials, was injected into WT or KO neurons held at −74 mV. C: the number of spikes evoked by injected current at certain intensity was compared between WT and KO neurons. D–F: parameters related to intrinsic excitability, including threshold of action potential firing, resting membrane potential, and membrane resistance, were compared between WT and KO neurons (N = 21 WT cells and 18 KO cells, P > 0.1).

Fig. 6.

Spontaneous transmission is not detectably altered in CYP46A1 KO slices. A and B: representative traces of sEPSCs and sIPSCs recorded from WT and KO neurons. C: frequency of sEPSC, sIPSC, and E/I ratio were compared between WT and KO neurons (N = 10 WT cells and 9 KO cells; P > 0.1).

Fig. 7.

NMDAR dependent LTP is unaltered, while synaptic depression induced by oxygen-glucose deprivation (OGD) is abrogated in CYP46A1 KO slices. A: field EPSPs (fEPSPs) were evoked at 50% maximal level. fEPSP slopes were compared before and after 100 Hz, 1 s tetanic stimulation in WT (open squares) and KO slices (open circles). At 60 min following tetanic stimulation, potentiation of fEPSPs was indistinguishable between WT and KO slices (N = 5 WT slices and 5 KO slices, P > 0.2). B: fEPSPs were evoked at 50% maximal level. fEPSP slopes were compared before/after 20 min OGD in WT (open squares) and KO slices (open circles), and KO slices treated with 1 μM SGE-301 before and during OGD for 40 min total (closed circles). OGD produced > 80% depression of fEPSPs (30′) from all the three groups. After 90 min OGD wash-out (120′), recovery of fEPSPs was significantly greater in KO slices, but not in WT slices or KO slices treated with 1 μM SGE-301 (N = 5 WT slices, 5 KO slices, and 5 KO slices treated with SGE-301. *P < 0.05; Kruskal-Wallis test and Dunn's post hoc analysis). Insets show representative traces from the various conditions at the indicated time points. Scale bar: 1 mV, 5 ms.