Autism-associated neuroligin-3 mutations commonly disrupt tonic endocannabinoid signaling

Abstract

Neuroligins are postsynaptic cell-adhesion molecules that interact with presynaptic neurexins. Rare mutations in neuroligins and neurexins predispose to autism, including a neuroligin-3 amino acid substitution (R451C) and a neuroligin-3 deletion. Previous analyses showed that neuroligin-3 R451C-knockin mice exhibit robust synaptic phenotypes but failed to uncover major changes in neuroligin-3 knockout mice, questioning the notion that a common synaptic mechanism mediates autism pathogenesis in patients with these mutations. Here, we used paired recordings in mice carrying these mutations to measure synaptic transmission at GABAergic synapses formed by hippocampal parvalbumin- and cholecystokinin-expressing basket cells onto pyramidal neurons. We demonstrate that in addition to unique gain-of-function effects produced by the neuroligin-3 R451C-knockin but not the neuroligin-3 knockout mutation, both mutations dramatically impaired tonic but not phasic endocannabinoid signaling. Our data thus suggest that neuroligin-3 is specifically required for tonic endocannabinoid signaling, raising the possibility that alterations in endocannabinoid signaling may contribute to autism pathophysiology.

Figure 1. Neuroligin-3 R451C substitution impairs GABAergic synaptic transmission in PV basket cell synapses

(A) Paired recordings of presynaptic APs in PV basket cells (upper traces) that produce unitary IPSCs in CA1 pyramidal cells (lower traces, V_holding = −70 mV). (B) Comparison of IPSC amplitudes (including failures) and of the percentages of successful transmissions induced by presynaptic APs applied at 1, 5, and 10 Hz in wild-type and R451C-mutant synapses. Open circles represent individual pairs (n_WT=14, n_R451C=27, Mann-Whitney RST, P<0.02 for all data sets). (C) The R451C KI mutation did not alter the half width of IPSCs (n_WT=14, n_R451C=23, Mann-Whitney RST, P=0.092). (D) Quantification of minimal IPSCs (amplitude of reliably occurring smallest IPSCs in each pair) suggest no change in quantal response in the R451C KI (n_WT=15, n_R451C=25, Mann-Whitney RST, P=0.235). (E) Additional paired-recordings show that IPSCs was independent of post-synaptic membrane voltage in R451C KI mice (n_WT=3, n_R451C=4). (F) The frequency of finding synaptically coupled pairs was not altered in R451C mice. (G) Neurolucida reconstructions of biocytin-filled basket cells show major reorganization in axonal and dendritic arbor of PV basket cells (str. = stratum, pyr. = pyramidale, rad. = radiatum). See also Figure S1.

Figure 2. Neuroligin-3 R451C substitution enhances GABAergic synaptic transmission in CCK basket cell synapses

(A) Paired recordings of presynaptic APs in CCK basket cells (upper traces) that produce unitary IPSCs in CA1 pyramidal cells (lower traces, V_holding = −70 mV). (B) Comparison of IPSC amplitudes (including failures) and of the percentages of successful transmissions induced by presynaptic APs applied at 1, 5, and 10 Hz in wild-type and R451C-mutant synapses. Open circles represent individual pairs (n_WT=8, n_R451C=17, Mann-Whitney RST, P=0.013 at 1 Hz IPSCs and P>0.08 in all other data sets). (C & D) No change in IPSC halfwidth (n_WT=8, n_R451C=15, Mann-Whitney RST, P=0.098) and no increase in the minimal IPSC amplitudes in R451C KI (n_WT=8, n_R451C=15, Mann-Whitney RST, P=0.5) suggest that the enhanced IPSC amplitudes in R451C KIs is not due to changes in quantal GABA receptor responses. (E) The frequency of finding synaptically coupled pairs was not altered in R451C mice. (F) Neurolucida reconstructions of biocytin-filled basket cells show no major reorganization in axonal and dendritic arbor of CCK basket cells.

Figure 3. Neuroligin-3 KO does not alter GABAergic transmission in PV basket cell synapses

(A & B) Paired-recording data show that IPSC amplitudes and transmission rates were unaltered in NL3 KO mice compared to WT littermates (n_WT=12, n_KO=8, Mann-Whitney RST, P>0.32 in all data set). (C & D) Quantification of IPSC halfwidth and minimal IPSC amplitudes suggest no changes in postsynaptic GABA receptor subunit composition. (E) The frequency of finding connected pairs was similar in NL3 WT and KO mice.

Figure 4. Neuroligin-3 KO enhances GABAergic synaptic transmission in CCK basket cell synapses similar to the R451C KI

(A & B) Paired recording data show that IPSC amplitudes and transmission rates were enhanced in CCK basket cell to CA1 pyramidal neuron synapses at multiple AP firing frequencies (n_WT=28, n_KO=36, Mann-Whitney RST, P=0.12 at 10 Hz IPSCs, and P<0.03 for all other data set). (C) Increase in IPSC halfwidth in KO suggest possible subunit reorganization of GABA receptor subunits in NL3 KO (n_WT=28, n_KO=35, Mann-Whitney RST, P=0.021). (D & E) No change in minimal IPSC amplitudes (n_WT=28, n_KO=35, Mann-Whitney RST, P=0.885), and in the frequency of finding connected pairs between CCK basket cells and pyramidal cells.

Figure 5. The NL3 R451C KI mutation lowers the probability of GABA release from PV basket cell synapses

(A) Averaged PV basket cell IPSCs (same data as in Fig. 1) are plotted against their corresponding averaged success rates (WT data were pooled from wild-type littermates of R451C KI and NL3 KO mice). Data were fitted to the equation $[\text{PSC}=\mathrm{Q}\cdot \mathrm{N}\cdot [1-\sqrt[\mathrm{N}]{1-\text{Successes}}]$ to estimate the mean quantal size (Q) and number of release sites (N) for each synapse population. Solid lines indicate best fit (black: WT, blue: R451C KI). Inset shows the distribution of individual data points. (B) Computer simulations of PV basket cell IPSCs. Simulation results for WT (open black circles) and R451C KI (open blue circles) were not significantly different (in mean IPSCs and successes) from their corresponding experimental IPSCs datasets when P_R was set to 0.23 and 0.11, respectively, in the model (see main text for further parameters). (C) Light microscopy analysis of the bouton density of PV basket cell axons. Left: example of axonal segments for axons in WT and R451C KI mice. Right: summary data from WT (n=7) and R451C KI (n=8) mice. P=0.152, Mann-Whitney RST. (D) Bath application of μ-opioid receptor antagonist CTAP (500 nM) in paired recording experiments between PV basket and pyramidal cells in R451C KI mice (n=4 pairs). Averaged time course (left) and time averaged means (right) of the 4 recordings did not show statistically significant effect of μ-opioid receptor antagonist on IPSCs. (E) Bath application of M2 muscarinic-receptor antagonist AF-DX 116 (10 μM) in paired recording experiments between PV basket and pyramidal cells in R451C KI mice (n=4 pairs). Averaged time course (left) and time averaged means (right) of the 4 recordings did not show statistically significant effect of μ-opioid receptor antagonist on IPSCs. Averaged data presented as mean ± s.e.m. See also Figure S2 and S3.

Figure 6. Neuroligin-3 KO and R451C KI mutations impair tonic endocannabinoid signaling

(A) Representative paired recordings (upper traces) and normalized time-courses (lower left panel) demonstrate that bath application of 10 μM AM251 enhances IPSCs in WT, but not in NL3 KO mice. Lower right panel: IPSC changes (failures included) in each paired-recording experiment (`control': average data for minutes 1–5, `AM251': for minutes 6–10; n_WT=9, P=0.004; n_NL3KO=11, P=0.268, paired T-test). (B) Left panel: time-courses of AM251 wash-in suggest that the lack of effect of AM251 on IPSCs was due to the failure of AM251 in increasing the number of successful transmissions. Right panel: AM251 reliably increased the number of successes in WT, but not in NL3 KO mice (n_WT=9, P<0.001; n_NL3KO=11, P=0.79, paired T-test). (C) Averaged CCK basket cell IPSCs (same data as in Fig. 4) are plotted against their corresponding averaged success rates (WT data were pooled from wild-type littermates of R451C KI and NL3 KO mice). Data were fitted to the equation $[\text{PSC}=\mathrm{Q}\cdot \mathrm{N}\cdot [1-\sqrt[\mathrm{N}]{1-\text{Successes}}]$ to estimate the mean quantal size (Q) and number of release sites (N) for each synapse population. Solid lines indicate best fit (black: WT, red: NL3 KO). Inset shows the distribution of individual data points. (D) Computer simulations of CCK basket cell IPSCs. Simulation results for WT (open black circles) and NL3 KO (open red circles) were not significantly different (in mean IPSCs and successes) from their corresponding experimental IPSCs datasets when P_R was set to 0.26 and 0.12, respectively, in the model (see main text for further parameters). (E) Light microscopy analysis of the bouton density of CCK basket cell axons. Left: example of axonal segments for axons in WT and NL3 KO mice. Right: summary data from WT (n=6) and NL3 KO (n=7) mice. P=0.779, t-test. (F) Time-course of the effect of the AM251 wash-in on extracellulary evoked IPSCs (eIPSC; left panel), and averaged data in each experiment (right panel) show increase in eIPSC amplitude in WT, but not in NL3 KO mice (V_pyramidal= −70 mV, 1 Hz stimulation, in the presence of 5 μM NBQX and 10 μM D-AP5; n_WT=6, P=0.008; n_NL3KO=10, P=0.63, paired t-test). (G) Time-course of the effect of the CP945598 wash-in on extracellulary evoked IPSCs (eIPSC; left panel), and averaged data in each experiment (right panel) show increase in eIPSC amplitude in WT, but not in NL3 KO mice (V_pyramidal= −70 mV, 1 Hz stimulation, in the presence of 5 μM NBQX and 10 μM D-AP5; n_WT=15, P=0.0005; n_NL3KO=18, P=0.41, paired t-test). (H & I) Paired recordings of IPSC amplitudes and success rates in response to 10 μM AM251 in R451C KI mice. Left panels: time-course of the experiments. Right panels: absolute changes in each pair (n_WT=6, P=0.07; n_R451C=10, P=0.072, paired T-test). (J) Time-course of the effect of 10 μM AM251 wash-in on extracellularly evoked EPSCs (eEPSC; left panel), and averaged data in each experiment (right panel) in WT and NL3 KO mice (V_pyramidal= −70 mV, 1 Hz stimulation, in the presence of 50 μM picrotoxin; n_WT=11, P>0.05; n_NL3KO=8, P>0.05, paired t-test). See also Figure S2, S3 and S4.

Figure 7. Neuroligin-3 is not required for phasic short-term endocannabinoid signaling (DSI) or long-term endocannabinoid-dependent synaptic plasticity (i-LTD)

(A) Paired recordings show that DSI induced by phasic endocannabinoid signaling was unaffected in NL3 KO (left panel: example of DSI, note the transient suppression of IPSCs after brief depolarization in the pyramidal cell; right panel: averaged time-course of DSI in WT and NL3 KO). (B) Deletion of NL3 does not affect the magnitude or time-course of the endocannabinoid-dependent I-LTD (V_pyramidal = +10 mV, inter-stimulus interval 20 s, [Cl_pipette] = 4 mM, in presence of 5 μM NBQX and 10 μM D-AP5).

Figure 8. Schematic summary diagram of the effects of the NL3 KO and R451C substitution on three different synapses on pyramidal neurons in the CA1 region of the hippocampus

The diagram depicts a pyramidal neuron (green) receiving inputs from Schaffer collateral fibers and two different types of basket cell neurons (PV, parvalbumin; CCK, cholecystokinin). The changes observed in NL3 R451C knockin and KO mice are described on the left.