Pannexin 1 regulates bidirectional hippocampal synaptic plasticity in adult mice

Abstract

The threshold for bidirectional modification of synaptic plasticity is known to be controlled by several factors, including the balance between protein phosphorylation and dephosphorylation, postsynaptic free Ca(2+) concentration and NMDA receptor (NMDAR) composition of GluN2 subunits. Pannexin 1 (Panx1), a member of the integral membrane protein family, has been shown to form non-selective channels and to regulate the induction of synaptic plasticity as well as hippocampal-dependent learning. Although Panx1 channels have been suggested to play a role in excitatory long-term potentiation (LTP), it remains unknown whether these channels also modulate long-term depression (LTD) or the balance between both types of synaptic plasticity. To study how Panx1 contributes to excitatory synaptic efficacy, we examined the age-dependent effects of eliminating or blocking Panx1 channels on excitatory synaptic plasticity within the CA1 region of the mouse hippocampus. By using different protocols to induce bidirectional synaptic plasticity, Panx1 channel blockade or lack of Panx1 were found to enhance LTP, whereas both conditions precluded the induction of LTD in adults, but not in young animals. These findings suggest that Panx1 channels restrain the sliding threshold for the induction of synaptic plasticity and underlying brain mechanisms of learning and memory.

Keywords: LTD; LTP; NMDA receptors; hippocampus; mice; pannexin 1; synaptic plasticity.

Figure 1

Decreased expression of Panx1 transcript and protein in the adult mice brain. (A) Quantitative RT-PCR analysis of relative abundance of Panx1 mRNA in the cerebral cortex (Cx), hippocampus (Hip), and cerebellum (Cer) from young wild type (y-WT, gray), young Panx1 knockout (y-KO, blue), adult wild type (a-WT, black) and adult Panx1 knockout mice (a-KO, green). The transcript values were normalized to the levels of cyclophilin (Cyp1) (N = 4). (B) Representative Western blot showing the expression levels of Panx1 protein in homogenates of cerebral cortex (Cx), hippocampus (Hip) and cerebellum (Cer) of young or adult animals. β-actin expression in samples was used for loading control. (C) Quantification of Panx1 protein expression by densitometry analysis of bands from four independent Western blots (N = 4), including the one shown in (B). Values were normalized to β-actin loading control. (D) Western blots of plasma membrane biotinylated proteins of hippocampal slices probed with anti-Panx1 and anti-Cx43 antibodies. (E) Quantification of protein expression by densitometry analysis of Panx1 bands from three independent Western blots (N = 3) like the one shown in (D). All data are plotted as mean ± SEM related to results of y-WT animals. Statistical differences were calculated using 2 way-ANOVA, followed by post hoc Bonferroni’s test. * p < 0.05 vs. y-WT Cx; § p < 0.05 vs. y-WT Hip; € p < 0.05 vs. y-WT Cer.

Figure 2

Absence and blockade of Panx1 channels alter hippocampal synaptic transmission in adult, but not in young mice. (A1) Cartoon depicting stimulus-record electrode configuration to record synaptic activity in Sc-CA1 synapse in hippocampal slices. (A2) representative FP at different stimulus intensities for young wild type (y-WT, gray line), young Panx1 knockout (y-KO, blue line), adult wild type (a-WT, black line), adult wild type plus probenecid 100 µM (a-WT+Pbncd, dotted line) and adult Panx1 knockout mice (a-KO, green line). (A3–A5) Input-output curves showing the relationship between FP slope (A3), fiber volley amplitude (A4) and stimulus intensity; and fiber volley amplitude and FP slope (A5). An increased FP slope was observed in a-WT+Pbncd and a-KO mice compared to either y-WT or a-WT mice. (B1) Representative FP traces at interstimulus intervals of 100 ms. (B2) Paired-pulse facilitation (PPF) of the FP at various interstimulus intervals. No significant differences were observed between WT and KO mice. (C1) Absence and blockade of Panx1 channels significantly increased glutamate release and spillover in adult animals. DL-TBOA (TBOA 10, µM) was perfused after 20 min of basal transmission. (C2) Averaged increments in basal synaptic transmission induced by TBOA. The values in parentheses indicate the number of hippocampal slices (left) and the number of animals (right) used. All data are plotted as mean ± SEM. Statistical differences were calculated using ANOVA, followed by post hoc Bonferroni’s test. Asterisks indicate statistical significance of the observed differences. *p < 0.05 vs. y-WT; § p < 0.05 vs. a-WT.

Figure 3

Increased long term potentiation (LTP) and absent long term depression (LTD) in the Schaffer collateral–CA1 pathway from adult Panx1 knock-out and WT mice treated with Pbncd. (A1) Representative traces of FPs recorded 1 min before (a) and 60 min after (b) TBS. (A2) LTP obtained in slices from young wild type (y-WT, gray line), young Panx1 knockout (y-KO, blue line), adult wild type (a-WT, black line), adult wild type plus probenecid 100 µM (a-WT+Pbncd, red line) and adult Panx1 knockout mice (a-KO, green). Long-term potentiation was induced by the delivery of TBS at the time indicated by the arrow. (A3) Magnitude average of LTP determined as responses between 50 and 60 min after TBS. Long-term potentiation was significantly different for a-KO mice compared to y-WT, y-KO and a-WT mice. (B1) Representative traces of FPs recorded 1 min before (a) and 60 min after (b) PP-LFS. (B2) Long-term depression obtained in slices from y-WT (gray line), y-KO (blue line), a-WT (black line), a-WT plus probenecid 100 µM (a-WT+Pbncd, dotted line) and a-KO hippocampal slices (green line). Long-term depression was induced by the delivery of PP-LFS at the time indicated by the line. (B3) Magnitude average of LTD determined as responses between 50 and 60 min after PP-LFS. Long-term depression was absent in a-WT+Pbncd and in a-KO mice. The values in parentheses indicate the number of hippocampal slices (left) and the number of animals (right) used. All data are plotted as mean ± SEM. Statistical differences were calculated using ANOVA, followed by post hoc Tukey test. Asterisks indicate statistical significance of the observed differences (*p < 0.05).

Figure 4

Blockade or absence of Panx1 channels facilitates the induction of LTP. (A) Long-term potentiation induced by one TBS (TBS₁). TBS1 induced a stable LTP in adult Panx1 knockout (a-KO, green square), but transient LTP in adult wild type (a-WT, black circle), which returned to baseline after 30 min. A second TBS (TBS2) applied to the same synapses elicited significantly more potentiation in a-KO but not in a-WT mice. The blockade of Panx1 with Pbncd added 20 min before the application of the TBS2, induced a significant enhancement of potentiation in a-WT (a-WT+Pbncd, red circle). (B) Magnitude averages of LTP were determined as responses between 30 and 35 min after the first TBS (open bar) and 50 and 60 min after the second TBS (filled bar). (C) Long-term potentiation induced by TBS was completely blocked by the incubation of APV 50 µM. (D) Magnitude average of LTP determined as responses between 50 and 60 min after TBS. The values in parentheses indicate the number of hippocampal slices (left) and the number of animals (right) used. *p < 0.05 for TBS1 vs. TBS2; § p < 0.05 between TBS1; € p < 0.05 between TBS2. All data are plotted as mean ± SEM.

Figure 5

Absence or blockade of Panx1 channels modifies the threshold for inducing excitatory hippocampal synaptic plasticity. (A) One tetanus (100 Hz, 1 s) induced significantly more potentiation in adult wild type plus Pbncd (a-WT+Pbncd, white circle) and adult Panx1 knockout (a-KO, green circle) compared to adult wild type (a-WT, black) mice. (B) 10 Hz stimulation for 1.5 min evoked significant potentiation in a-WT+Pbncd and a-KO mice compared to a-WT mice. (C) 5 Hz stimulation for 3 min produced potentiation in a-WT+Pbncd and a-KO mice, whereas it produced a slight depression in a-WT. (D) 1 Hz stimulation for 15 min induced a reliable LTD in WT mice, whereas it elicited a mild potentiation in a-WT+Pbncd and a-KO mice. (E) 30 µM of NMDA for 5 min induced LTD in all groups. (F) Summary of data for synaptic plasticity at different frequencies of stimulation. The values in parentheses indicate the number of hippocampal slices (left) and the number of animals (right) used. All data are plotted as mean ± SEM.

Figure 6

Hypothesis for the role of Panx1 channels in the regulation of excitatory synaptic plasticity. (1) Calcium Hypothesis. Upon neuronal activity, NMDAR activation trigger calcium influx into dendritic spine. Depending on the kinetics and magnitude of the calcium concentration increments, kinases or phosphatases are activated promoting the insertion or remotion of AMPARs mediating LTP and LTD respectively (Lüscher and Malenka, 2012). NMDAR activation, probably through depolarization of post-synaptic membrane, activates Panx1 channels (Thompson et al., 2008) producing more calcium influx and ATP release by Panx1 channels. ATP released by Panx1 channels may activate ionotropic (P2X) and metabotropic (P2Y) purinergic receptors. P2X depolarizes the membrane and allows calcium entry, whereas P2Y controls calcium release from intracellular stores such as endoplasmic reticulum (ER) through the activation of G protein (Gα), therefore both receptors may contribute to the increase in the cytosolic calcium concentration (Collo et al., ; Yamazaki et al., ; Wang et al., ; Abbracchio et al., 2009). In addition, Panx1 channels could interact with specific NMDAR subunits exerting a modulatory effect over these receptors. We also speculate that Panx1 channels could facilitate the localization or function of specific NMDAR subunit in the post-synaptic membrane, therefore in the absence of Panx1, a greater activation of kynases instead phosphatases could be due to a change in the composition of GluN2 subunits, leading a change in the kinetics and calcium influx through NMDARs. (2) Activation of pre-synaptic Adenosine receptors Hypothesis. In the synapses, ATP release through Panx1 channels could be converted into adenosine (Ado), which in turn can activate P1 purinergic receptors (adenosine receptors). The activation of A1 adenosine receptor (A1), located in the presynaptic terminals reduces the release of glutamate. Therefore, in the absence of Panx1 the depletion of extracellular ATP and adenosine promotes an increase in the neurotransmitter release (Prochnow et al., 2012). (3) Regulation of synaptic cleft pH Hypothesis. Extracellular ATP hydrolysis also generates protons and phosphate that produce a decrease in the extracellular pH (Vroman et al., 2014). This acidification could inhibit both, voltage gated calcium channels (VGCC) present in the presynaptic terminal, and NMDAR on postsynaptic membrane, reducing glutamate release and NMDA receptor activation. The absence of Panx1 channels or their inhibition prevents the release of ATP, stops the production of phosphate buffer and produce the alkalization of the synaptic cleft, which increase the release probability and the activation of NMDARs.