Transient hypofunction of NMDARs during early postnatal development impairs the frequency-dependent synaptic filtering in the LPP – DG synapse.
A specialized feature of the LPP – DG synapse is that it operates as a low-pass filter (Quintanilla et al., 2022). That is, while low frequencies, such as those in the theta range, are efficiently transferred to the DG, the high frequency patterns, i.e., the gamma range, are greatly attenuated upon arrival to the DG. Therefore, in the next experiment, we examined the frequency-dependent filtering property of the LPP – DG synapse by delivering stimulation frequencies relevant to hippocampal cognitive processing, such as the theta (5 Hz), beta (20 Hz), and low gamma (50 Hz) stimulation ranges (Engel and Fries, 2010; Colgin, 2016).
Consistent with Quintanilla et al. (2022), in control slices, a theta train (10 pulses at 5 Hz) induced sustained facilitation all over the evoked synaptic responses (n = 9 slices / 6 animals; traces and black symbols in Figures 8a1 and 8b1). The traces in Figure 8c1 also depict the facilitated response, which contrasts the amplitude of the first response vs. the last response (S1/S10). In contrast, the MK-801-treated slices exhibited greater facilitation in response to 5 Hz stimulation (n = 7 slices / 7 animals). The green symbols of the scatter plot in Figure 8a1 show the increased facilitation of the LPP fEPSPs in MK-801-treated slices compared to control slices (RM two-way ANOVA, treatment effect: F(1, 14) = 7.553, P< 0.05; green traces in Figure 8b1), and the increased facilitation in the amplitude of last response vs. first response is depicted in the green traces in Figure 8c1.
On the other hand, in response to 20 Hz stimulation, control slices exhibited increased facilitation on the first part of the train (second to fifth response), which reverted in the last five responses, generating a Gaussian-type facilitation distribution (n = 8 slices / 6 animals; symbols and black traces in Figure 8a2-c2). Contrary to the ability to attenuate the synaptic responses, the MK-801-treated slices exhibited greater facilitation. i.e., decreased filtering ability. However, this phenomenon, illustrated in Figure 8a2, did not show significant differences compared to control slices (RM two-way ANOVA, treatment effect: F(1, 15) = 3.386, P > 0.05; n = 8 slices / 7 animals for MK-801). The representative traces are depicted in Figures 8b2 and 8c2.
The diminished filtering ability of the LPP – DG synapse in MK-801-treated slices was further revealed by stimulation in the gamma frequency range (50 Hz). Under this experimental paradigm, the control slices exhibited enhanced facilitation in the initial portion of the stimulation train (second to fourth response). The facilitation reverted in the middle section and was strongly attenuated in the final section of the train (n = 8 slices / 6 animals; symbols and black traces in Figure 8a3-c3). Although filtering ability was present in the MK-801-treated slices, there was still an easily identifiable decrease under this condition compared to control slices (RM two-way ANOVA, treatment effect: F(1, 14) = 5.285, P < 0.05, n = 8 slices / 7 animals for MK-801; green symbols in Figure 8a3 and green traces in Figures 8b3 and 8c3). The dysregulation in frequency-dependent filtering of LPP in MK-801-treated slices can also be observed in the scatter plot graphs, which show the change in amplitude between the first response and tenth response for each slice at 5, 20, and 50 Hz train (Figure 8d-f), demonstrating reduced synaptic filtering in MK-801-treated slices. These results show that the synaptic property of low-pass filtering of the LPP – DG synapse is dramatically altered in response to the transient hypofunction of NMDARs during early postnatal development.