Skip to main content
NCBI home page
The Journal of Neuroscience logoLink to The Journal of Neuroscience
. 2021 Sep 29;41(39):8103–8110. doi: 10.1523/JNEUROSCI.0451-20.2021

Frequency-Dependent Synaptic Dynamics Differentially Tune CA1 and CA2 Pyramidal Neuron Responses to Cortical Input

Qian Sun 1,2,✉, Eric W Buss 1, Yu-Qiu Jiang 2, Bina Santoro 1, David H Brann 1, Daniel A Nicholson 3, Steven A Siegelbaum 1,✉
PMCID: PMC8482869  PMID: 34385360

Abstract

Entorhinal cortex neurons make monosynaptic connections onto distal apical dendrites of CA1 and CA2 pyramidal neurons through the perforant path (PP) projection. Previous studies show that differences in dendritic properties and synaptic input density enable the PP inputs to produce a much stronger excitation of CA2 compared with CA1 pyramidal neurons. Here, using mice of both sexes, we report that the difference in PP efficacy varies substantially as a function of presynaptic firing rate. Although a single PP stimulus evokes a 5- to 6-fold greater EPSP in CA2 compared with CA1, a brief high-frequency train of PP stimuli evokes a strongly facilitating postsynaptic response in CA1, with relatively little change in CA2. Furthermore, we demonstrate that blockade of NMDARs significantly reduces strong temporal summation in CA1 but has little impact on that in CA2. As a result of the differences in the frequency- and NMDAR-dependent temporal summation, naturalistic patterns of presynaptic activity evoke CA1 and CA2 responses with distinct dynamics, differentially tuning CA1 and CA2 responses to bursts of presynaptic firing versus single presynaptic spikes, respectively.

SIGNIFICANCE STATEMENT Recent studies have demonstrated that abundant entorhinal cortical innervation and efficient dendritic propagation enable hippocampal CA2 pyramidal neurons to produce robust excitation evoked by single cortical stimuli, compared with CA1. Here we uncovered, unexpectedly, that the difference in efficacy of cortical excitation varies substantially as a function of presynaptic firing rate. A burst of stimuli evokes a strongly facilitating response in CA1, but not in CA2. As a result, the postsynaptic response of CA1 and CA2 to presynaptic naturalistic firing displays contrasting temporal dynamics, which depends on the activation of NMDARs. Thus, whereas CA2 responds to single stimuli, CA1 is selectively recruited by bursts of cortical input.

Keywords: CA1, CA2, hippocampus, NMDAR, perforant path, synaptic

Introduction

Different types of pyramidal neurons (PNs) are often equipped with distinctive sets of synaptic and dendritic mechanisms to dynamically regulate information propagation (Spruston, 2008). In the hippocampus, we and others have previously demonstrated that CA1 and CA2 PNs use a distinct set of synaptic and dendritic mechanisms to integrate and amplify the information from entorhinal cortex (EC) (Chevaleyre and Siegelbaum, 2010; Sun et al., 2014; Dudek et al., 2016; Srinivas et al., 2017). For example, we found that the perforant path (PP) projection from EC layer II (LII) neurons onto CA2 PN distal apical dendrites in stratum lacunosum moleculare (SLM) is so powerful, despite its distal dendritic target, that a single PP stimulus can robustly excite CA2 PNs to reliably trigger full-blown action potentials (APs) (Sun et al., 2014). By contrast, the direct PP projection from EC layer III (LIII) PNs that targets the adjacent CA1 distal apical dendrites in SLM produces only a weak depolarization of the CA1 soma because of severe attenuation of the distal EPSP by the dendritic cable properties (Golding et al., 2005; Sun et al., 2014; Srinivas et al., 2017).

The higher efficacy of the distal inputs onto CA2 PNs compared with CA1 PNs results from a series of differences in synaptic and dendritic properties that facilitate and amplify the EC-to-CA2 information transfer. These include a denser cortical innervation of the distal apical dendrites, with nearly 3 times more glutamatergic synapses in SLM of CA2 than CA1, a more efficient dendritic propagation of EPSPs, and a unique dendritic geometry that enables more efficient propagation of dendritic Na+ spikes from distal dendrites to the CA2 soma (Sun et al., 2014; Srinivas et al., 2017).

Paradoxically, despite the severe attenuation of the distal CA1 EPSP, in vivo studies show that this pathway is crucial for spatial encoding (Brun et al., 2002, 2008; Nakashiba et al., 2008), place cell formation (Bittner et al., 2015), and performance of certain hippocampal-dependent memory tasks (Suh et al., 2011; Kitamura et al., 2014). Indeed, optogenetic inhibition of the EC-to-CA1 pathway in vivo significantly reduced firing rate in CA1 PNs, indicating that this pathway can robustly excite CA1 PNs in vivo (Kitamura et al., 2014).

Notably, previous studies that compare PP input with CA1 and CA2 have primarily used a single PP stimulus (Chevaleyre and Siegelbaum, 2010; Sun et al., 2014; Srinivas et al., 2017). However, the firing of EC LII and LIII neurons, which send PP projections to CA2 and CA1, respectively, is highly dynamic during behavior (Burgalossi et al., 2011; Tang et al., 2015). Indeed, the firing of short bursts of spikes (≥100 Hz) is a characteristic feature in freely moving animals performing behavioral tasks (Burgalossi et al., 2011; Tang et al., 2015).

In this study, we used protocols using high-frequency bursts or naturalistic EC firing patterns to examine further how behaviorally relevant patterns of PP input may differentially regulate temporal dynamics of cortico-hippocampal information processing in CA1 and CA2 PNs. Surprisingly, we found that the enhancement in depolarization evoked by a PP stimulus in CA2 relative to CA1 PNs largely disappears in response to burst firing of presynaptic neurons. This can, in part, be explained by a substantially more robust frequency- and NMDAR-dependent temporal summation in CA1, compared with CA2. These different synaptic dynamics produce different temporal patterns of postsynaptic spike firing, with a CA2 PN firing early during a burst of presynaptic inputs versus a CA1 PN firing later during the burst. Together with our previous studies (Sun et al., 2014; Srinivas et al., 2017), these findings suggest that CA1 and CA2 PNs use distinct sets of synaptic and dendritic mechanisms to dynamically regulate cortico-hippocampal information flow.

Materials and Methods

Mice

Wild-type (WT) C57BL6J mice were obtained from The Jackson Laboratory. Both male and female WT mice were used in the experiments. The procedures described were conducted in accordance with National Institutes of Health regulations and approved by the Institutional Animal Care and Use Committees of Columbia University, the New York State Psychiatric Institute, and Case Western Reserve University.

Hippocampal slice preparation

Transverse hippocampal slices were prepared from 5- to 8-week-old C57BL6J mice as described previously (Sun et al., 2014). In brief, animals were anesthetized and killed by decapitation in accordance with institutional regulations. Hippocampi were dissected out, and transverse slices from the dorsal hippocampus (400 µm thickness) were cut on a vibratome (Leica Microsystems, VT1200S) in ice-cold dissection solution containing the following (in mm): 10 NaCl, 195 sucrose, 2.5 KCl, 10 glucose, 25 NaHCO3, 1.25 NaH2PO4, 2 Na-pyruvate, 0.5 CaCl2, and 7 MgCl2. The slices were then incubated at 33°C in ACSF (in mm: 125 NaCl, 2.5 KCl, 20 glucose, 25 NaHCO3, 1.25 NaH2PO4, 2 Na-pyruvate, 2 CaCl2, and 1 MgCl2) for 20-30 min and then kept at room temperature for at least 1.5 h before transfer to the recording chamber. Cutting and recording solutions were both saturated with 95% O2 and 5% CO2, pH 7.4. All electrophysiological recordings were performed at 31°C-32°C.

Electrophysiological recordings

Whole-cell recordings were obtained from PNs using the “blind” patch-clamp technique. Synaptic potentials were recorded in current-clamp mode with the membrane held at ∼−70 mV with a patch pipette containing the following (in mm): 135 K-gluconate, 5 KCl, 0.1 EGTA-Na, 10 HEPES, 2 NaCl, 5 MgATP, 0.4 Na2GTP, and 10 Na2-phosphocreatine, pH 7.3 (280-290 mOsm); 0.2% biocytin was routinely included in the intracellular solution. Series resistance was monitored throughout each experiment; neurons with a series resistance >25 mΩ were excluded from analysis. Extracellular field potentials were recorded with a patch pipette containing 1 m NaCl. Synaptic potentials were evoked by monopolar stimulation with a patch pipette filled with 1 m NaCl and located in SLM of the CA1 region. Neurons were held at ∼–70 mV for current-clamp recordings. Resting membrane potential was measured immediately on break-in. EPSPs were recorded in the presence of GABAA and GABAB receptor antagonists (1 μm SR 95531 and 2 μm CGP 55845, Tocris Bioscience). Extracellular field recordings of population spikes were recorded in stratum pyramidale in mid-CA1 (CA1b) and CA2.

Biocytin staining

The streptavidin staining was performed as described previously (Sun et al., 2017). Briefly, after each recording, slices were fixed at 4°C for at least 24 h in 4% PFA in PBS, pH 7.3. After fixation, slices were treated with PBS containing normal goat serum (5%) and 0.5% Triton X-100 for 2 h at room temperature. Slices were subsequently incubated in PBS containing streptavidin, AlexaFluor-594 conjugate (1:500), and 0.1% Triton X-100 for 2 d at 4°C. Subsequently, slices were rinsed in PBS several times and processed through increasing concentrations of glycerol, and then embedded in mounting media (Fluorogel, Electron Microscopy Sciences).

Statistics

Statistical comparisons were performed using two-tailed, unpaired Student's t test, one-way ANOVA, or repeated-measures ANOVA followed by Bonferroni's multiple comparisons test in GraphPad Prism 9, unless otherwise noted in the figure legends. Results are expressed as mean ± SEM. p < 0.05 is considered statistically significant.

Results

Distinct frequency-dependent temporal dynamics in CA1 versus CA2 PNs in response to PP stimulation

We recorded from CA1 and CA2 PNs under current clamp in response to delivery of one or more stimuli to the PP inputs using a stimulating electrode placed in SLM (Fig. 1A–C). We first confirmed that a single PP stimulus caused only a small depolarization of the CA1 soma, with a maximal value of 2.65 ± 0.33 mV (mean ± SEM, n = 9 cells; 40 V stimulus), far below the threshold required for firing an AP (Fig. 1C). By contrast, the same protocol produced a 5- to 6-fold greater depolarization in CA2 PNs (Fig. 1C), confirming our previous reports (Chevaleyre and Siegelbaum, 2010; Sun et al., 2014; Srinivas et al., 2017).

Figure 1.

Figure 1.

Open in a new tab

PP-CA2 synapses display less frequency-dependent summation, compared with PP-CA1 synapses. A, Left, A sample biocytin-filled CA1 PN and the placement of stimulating electrode in SLM. Right, Membrane voltage changes in response to indicated constant depolarizing or hyperpolarizing current injections. S, Stimulating electrode; R, recording electrode; SC, Schaffer collaterals. B, Left, A sample biocytin-filled CA2 PN and the placement of stimulating electrode in SLM. Right, Membrane voltage changes in response to indicated constant depolarizing or hyperpolarizing current injections. S, Stimulating electrode; R, recording electrode; SC, Schaffer collaterals. C, Sample traces (top) and summary data (bottom) of PP EPSP input-output relation recorded in CA1 (n = 9 cells) or CA2 (n = 5-8 cells) PNs in response to single stimuli of indicated intensities. Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,29) = 269.3). ***p < 0.001. Inhibition was blocked with 1 μm gabazine (SR-95531) and 2 μm CGP 55845A, unless otherwise noted. D, Sample traces (left) and summary data (right) of EPSPs in CA1 (n = 6 or 7 cells) and CA2 (n = 4-7 cells) PNs in response to a train of 5 PP stimuli at 50 Hz. Error bars indicate SE. Significance of differences in CA1 and CA2 EPSPs: repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,51) = 43.47). ***p < 0.001. E, The ratios of CA2 to CA1 peak EPSP amplitude plotted against stimulating intensity in response to a single stimulus or 50 Hz train of 5 stimuli. Error bars indicate SE.

Next, we examined the synaptic input-output relationship in response to a high-frequency burst of stimuli (5 pulses at 50 Hz) (Fig. 1D,E). This produced a marked temporal summation of the depolarization in CA1 PNs, as previously reported (Jarsky et al., 2005). Surprisingly, the burst produced much less temporal summation in CA2. As a result, despite the initially much greater amplitude of the PP EPSP in CA2 compared with CA1 at the beginning of the burst, by the end of the train of 40 V stimulation, the net depolarization in CA1 (17.3 ± 1.8 mV, n = 6 cells) was similar to that in CA2 (20.9 ± 2.6 mV, n = 4 cells; p = 0.26 for CA1 compared with CA2; Fig. 1D,E).

These results, combined with our previous findings of powerful PP-driven dendritic Na+ spikes in CA2 (Sun et al., 2014), led us to predict that high-frequency firing in EC will give rise to contrasting dynamics of AP output in CA1 and CA2. To test this, we delivered a train of high-frequency stimuli with the intensity of 40-48 V (5 pulses at 50 Hz) to the PP to elicit AP output in CA1 or CA2 (Fig. 2A,B). We found that CA1 PNs required repetitive high-frequency stimulation to trigger spikes, whereas CA2 PNs fired in response to the first stimulus, leading to a markedly greater spike latency in CA1 than that in CA2 (Fig. 2A–C). Extracellular recordings from CA1 and CA2 pyramidal cell layers further confirmed the contrasting temporal dynamics of AP output, with an increasing population spike amplitude in CA1 versus a decreasing population spike in CA2 during a train of 5 stimuli (Fig. 2D–F).

Figure 2.

Figure 2.

Open in a new tab

Distinct temporal dynamics of AP output in CA1 versus CA2 PNs in response to bursts of strong PP stimulation. A, Overlay of sample voltage traces from CA1 and CA2 PNs in response to a train of PP stimuli (5 pulses at 50 Hz). Note the contrasting spike latency and temporal summation between CA1 and CA2 PNs. Arrows indicate the electrical stimuli. B, Sample traces of APs in response to a train of 5 PP stimuli at 50 Hz in three additional trials obtained from the same cells as shown in A. C, Plots of mean (bars) and individual (circles) AP latency in CA1 (n = 7 cells) and CA2 (n = 7 cells) PNs in response to the train of 50 Hz PP stimuli. Error bars indicate SE. ***p < 0.001 (unpaired t test). D, Schematic diagram of a transverse hippocampal slice and the placement of stimulating and recording electrodes for extracellular field recording experiments shown in E and F. E, Sample traces of population spikes recorded in the CA1 and CA2 pyramidal cell layer in response to a train of PP stimuli (5 pulses at 50 Hz). Arrowheads indicate population spikes. F, Normalized population spike amplitude as a function of PP stimulus number during the train for CA1 and CA2. Population spike amplitudes from a given experiment were normalized to the largest population spike amplitude during the 5 stimuli, determined separately for CA1 (n = 6 slices) and CA2 (n = 6 slices). Error bars indicate SE.

To more closely mimic in vivo conditions, we elicited AP output using a naturalistic stimulating sequence derived from in vivo EC neuron firing patterns in awake behaving animals, based on a previous report (Giocomo et al., 2011). This naturalistic stimulating sequence consists of 18 stimuli with a mean frequency of 10 Hz and the instantaneous frequencies ranging from 2 to 89 Hz. This protocol revealed distinct dynamics of AP output, in which CA2 fired with a short latency after EC stimulation while CA1 fired only following bursts of input activity (Fig. 3A,B). Moreover, naturalistic stimulating sequences evoked a higher spike frequency in CA2 (9.9 ± 2.2 Hz, n = 8 cells) than in CA1 (4.7 ± 0.7 Hz, n = 7 cells, p = 0.049) (Fig. 3C). Interestingly, however, a significantly larger fraction of spike events in CA1 occurred during burst firing (defined as interspike intervals < 20 ms) than in CA2 (47.5% in CA1 vs 25% in CA2; Fig. 3D,E). Together, PP naturalistic stimulation leads to distinct dynamics of AP output between CA1 and CA2, with CA2 firing more single spikes immediately after EC input but with a smaller fraction of spikes occurring in bursts compared with CA1.

Figure 3.

Figure 3.

Open in a new tab

Distinct temporal dynamics of AP output in CA1 versus CA2 PNs in response to PP stimulation using a naturalistic stimulating pattern. A, B, Sample traces (A) and summary data (B) showing AP output from CA1 (n = 7 cells) and CA2 (n = 8 cells) PNs in response to PP stimulation using a naturalistic stimulating pattern. A, Orange and green arrows indicate the timing of the first spike during the first burst firing for CA1 and CA2 PNs, respectively. The spike of CA2 PNs precedes that of CA1 PNs. Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,234) = 47.14). *p < 0.05. **p < 0.01. ***p < 0.001. C, Mean firing frequency for CA1 and CA2 PNs in response to a naturalistic stimulating pattern shown in A and B. Error bars indicate SE. *p < 0.05 (unpaired t test). D, E, Histogram (D) and cumulative probability (E) of the percentage of events of interspike intervals for CA1 (n = 59 events) and CA2 (n = 144 events) in response to a naturalistic stimulating pattern. Events with interspike interval < 20 ms are defined as bursting firing. ***p < 0.001 (Kolmogorov–Smirnov test).

More robust frequency-dependent temporal summation at PP-CA1 synapses than PP-CA2 synapses

To further characterize the dynamics of temporal summation of PP inputs to CA1 versus CA2 neurons, we varied the stimulus frequency from 2 to 50 Hz (Fig. 4). Because PP-evoked EPSPs in CA2 were 5- to 6-fold greater than CA1 (Fig. 1C), we matched initial baseline EPSPs between CA1 (1.45 ± 0.23 mV) and CA2 (1.76 ± 0.32 mV, p = 0.43) by adjusting PP stimulation intensity in CA1 (16-24 V) and CA2 (4-8 V). We observed a steep, sigmoidal relationship between temporal summation (measured as the EPSP5/EPSP1 ratio) and stimulating frequency, with a sharp rise in CA1 EPSP amplitude at stimulating frequencies of 20-30 Hz (Fig. 4A–C). Remarkably, the EPSP5/EPSP1 ratio was as large as 7 in response to stimulating frequencies of 30-50 Hz at PP-CA1 synapses (Fig. 4B,C). In sharp contrast, temporal summation at PP-CA2 synapses was less than that at CA1 synapses at all frequencies (Fig. 4D,E), and less than half that seen in CA1 with a 50 Hz train (CA1: EPSP5/EPSP1 = 7.11 ± 0.95, n = 10 cells; CA2: EPSP5/EPSP1 = 3.18 ± 0.29, n = 8 cells; p < 0.001; Fig. 4E).

Figure 4.

Figure 4.

Open in a new tab

Contrasting frequency-dependent summation at PP-CA1 versus PP-CA2 synapses. A, Sample traces of CA1 PN EPSPs in response to trains of 10 pulses with increasing stimulation frequency as indicated. B, Mean EPSP amplitude during trains shown in A. Peak amplitude after each stimulus is normalized to the peak of the first EPSP (n = 8-14 cells). Error bars indicate SE. C, Summary data for CA1 PNs showing the ratio of the fifth EPSP to the first EPSP plotted against stimulation frequency (n = 10-12 cells). Error bars indicate SE. D, Sample traces of EPSPs in CA1 versus CA2 PNs in response to trains of 5 PP stimuli at 5 or 30 Hz. E, Summary data of the ratio of the fifth EPSP to the first EPSP plotted against the stimulating frequency in CA1 (n = 10-12 cells) and CA2 (n = 8 cells). Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,40) = 55.90). ***p < 0.001. F, Sample raw (top) and scaled (bottom) traces of EPSPs in CA2 in response to trains of 5 weak versus strong PP stimuli at 50 Hz. G, Summary data of the ratio of the fifth EPSP to the first EPSP plotted against the stimulating frequency in response to weak (n = 8 cells) or strong stimulation (n = 4 cells) in CA2. Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,18) = 30.52). **p < 0.01.

Might the weaker stimulation intensity in CA2, relative to CA1, affect the level of temporal summation? We addressed this by applying a stimulation intensity to CA2 (16-24 V) that matched the stimulation intensity used in CA1 (Fig. 4A–E). As expected, strong PP stimulation produced a large initial baseline EPSP in CA2 (8.90 ± 0.86 mV, n = 4 cells). Interestingly, however, strong PP stimulation led to significantly reduced temporal summation compared with weaker stimulation in CA2 with a 50 Hz train (weak stimulation: EPSP5/EPSP1 = 3.18 ± 0.29, n = 8 cells; strong stimulation: EPSP5/EPSP1 = 2.02 ± 0.26, n = 4 cells; p < 0.01; Fig. 4F,G). Together, we conclude that temporal summation in CA2 was significantly weaker than that in CA1, regardless of the strength of stimulation.

NMDAR activation plays a key role in the robust frequency-dependent temporal summation at PP-CA1, but not PP-CA2, synapses

What are the mechanisms underlying the stark contrasting temporal summation between PP-CA1 and PP-CA2 synapses? As activation of NMDARs is known to be sensitive to repetitive stimulation (Herrero et al., 2000; Augustinaite and Heggelund, 2007; Hunt and Castillo, 2012), we first examined the effect of bath application of D-APV (50 μm), an antagonist of NMDARs, on both PP-CA1 and PP-CA2 synapses. Bath application of APV produced a small, but not significant, reduction of the initial PP EPSPs in CA1 (APV: 77.5 ± 8.5% of control, p = 0.088, n = 6 cells), and reduced PP EPSPs in CA2 (APV: 74.6 ± 5.3% of control, p = 0.002, n = 6 cells) (Sun et al., 2014). Thus, we adjusted the stimulation intensity to match the initial EPSPs between control and APV conditions. The antagonist significantly reduced EPSP summation during high-frequency stimulation (20-50 Hz), transforming the dependence of EPSP5/EPSP1 ratio from a sigmoidal to linear relationship in CA1 PNs (Fig. 5A,B), with no effect on the paired-pulse ratio (control: 2.00 ± 0.07, n = 6 cells; APV: 2.05 ± 0.08, n = 9 cells; p = 0.7, unpaired Student's t test). In sharp contrast, bath application of APV had little effect on temporal summation in CA2 PNs in response to high-frequency stimuli at 30 or 50 Hz (Fig. 5G,H).

Figure 5.

Figure 5.

Open in a new tab

Differential involvement of NMDAR activation in frequency-dependent temporal summation at PP synapses at CA1 and CA2 PNs. A, Sample traces of CA1 PN EPSPs in the absence or presence of the NMDAR antagonist D-APV (50 μm) in response to 5 pulses of 50 Hz PP stimulation. B, Summary data showing the CA1 PN EPSP5/EPS1 ratio plotted against stimulation frequency in the absence or presence of APV (control: n = 10-12 cells; APV: n = 7 cells). Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,34) = 35.66). ***p < 0.001. C, Sample subthreshold CA1 PN EPSPs driven by a naturalistic train of PP stimuli in the absence or presence of APV. D, Mean CA1 PN EPSPs during naturalistic stimulation train normalized by response to first stimulus. Responses averaged from 6 and 9 cells in the absence or presence of APV, respectively, using the same temporal stimulus pattern. Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,108) = 100.1). **p < 0.01. ***p < 0.001. E, Sample traces showing suprathreshold EPSP-driven spikes using the same temporal pattern of naturalistic stimulation as in C and D with a higher stimulus intensity. Traces shown in the absence or presence of APV. F, Quantification of firing probability using conditions shown in E (control: n = 6 cells; APV: n = 9 cells). Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,108) = 100.6). Error bars indicate SE. ***p < 0.001. G, Sample traces of CA2 PN EPSPs in the absence or presence of D-APV (50 μm) in response to 5 pulses of 50 Hz PP stimulation. H, Summary data showing the CA2 PN EPSP5/EPSP1 ratio plotted against stimulation frequency in the absence or presence of APV (control: n = 8 cells; APV: n = 5 cells). Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,24) = 9.597). *p < 0.05. I, Sample subthreshold CA2 PN EPSPs driven by a train of PP stimuli using a naturalistic EC firing pattern in the absence or presence of APV. J, Mean CA2 PN EPSPs during naturalistic stimulation train normalized by response to first stimulus. Responses averaged from 7 and 6 cells for control and APV, respectively, using the same temporal stimulus pattern. Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,197) = 18.47). *p < 0.05. K, Sample traces showing suprathreshold EPSP-driven spikes using the same temporal pattern of naturalistic stimulation as in I and J with a higher stimulus intensity in CA2 PNs. Traces shown in the absence or presence of APV. L, Quantification of firing probability using conditions shown in K (control: n = 8 cells; APV: n = 6 cells). Error bars indicate SE. Repeated-measures ANOVA followed by Bonferroni's multiple comparisons test (F(1,90) = 15.44).

To assess the relevance of NMDAR-dependent EPSP summation to in vivo patterns of synaptic input, we stimulated the PP in CA1 using a naturalistic set of subthreshold stimuli of 16-24 V intensity, as described above. We found this naturalistic stimulating pattern also caused pronounced summation, with a maximum ratio of the last to first EPSP of a burst equal to 7.0 ± 1.4 in CA1 (n = 7 cells; Fig. 5C,D), that was capable of evoking short bursts of postsynaptic spikes with strong stimulation intensity of 48 V (Fig. 5E,F). Bath application of APV markedly reduced the summation to natural spike sequences (the maximum ratio of EPSPn/EPSP1: control = 7.0 ± 1.4, n = 7 cells; APV = 3.1 ± 0.3, n = 8 cells; p = 0.01), and completely blocked the ability of CA1 PNs to fire APs in response to PP stimulation up to 56 V (Fig. 5E,F).

By contrast to results in CA1, the same naturalistic pattern of PP stimulation of 16-24 V intensity produced much smaller temporal summation in CA2 (the maximum ratio of EPSPn/EPSP1 = 2.0 ± 0.1, n = 7 cells; p = 0.01, relative to CA1). Moreover, bath application of APV had very little effect on CA2 temporal summation (in APV, the maximum ratio of EPSPn/EPSP1 = 1.8 ± 0.1, n = 6 cells; Fig. 5I,J). In addition, unlike CA1, APV failed to block the ability of CA2 PNs to fire spikes in response to 48 V PP stimulation (Fig. 5K,L). Together, these results suggest that activation of NMDARs plays a key role in robust temporal summation at PP-CA1 synapses, but makes a minimal contribution at PP-CA2 synapses.

Blockade of NMDARs has no effect on burst-induced facilitation of extracellular field EPSP (fEPSP) measured in SLM of CA1

The effect of bath application of APV can be attributed to the blockade of postsynaptic NMDARs (Golding et al., 2002; Takahashi and Magee, 2009) or presynaptic NMDARs, as shown in the SC terminals in CA1 (McGuinness et al., 2010; Zhang et al., 2016) and mossy fiber terminals in CA3 (Lituma et al., 2021). One possible mechanism underlying strong temporal summation in CA1 is through presynaptic facilitation of glutamate release mediated by the activation of presynaptic NMDARs. To address this possibility, we measured the effect of burst stimulation on the extracellular fEPSP in SLM of CA1 (Fig. 6A), which provides a measure of the EPSC. Although burst stimulation produced a strongly facilitation fEPSP, bath application of APV had little impact on this facilitation (Fig. 6B,C). This is in contrast to the strong suppression by NMDAR blockade of the temporal summation we typically observed in the intracellularly recorded EPSP (e.g., Fig. 4A–C). Thus, temporal summation during burst stimulation in CA1 is unlikely to result from enhanced glutamate release, but rather is likely to result from enhanced dendritic integration of the excitatory responses that requires NMDAR activation.

Figure 6.

Figure 6.

Open in a new tab

No effect of bath application of APV on burst-induced facilitation of fEPSP measured in SLM of CA1. A, Experimental setup. Both stimulating and recording electrodes were placed in SLM of CA1b with ∼200 µm apart. B, Sample extracellular field recording traces in response to a train of PP stimuli (30 Hz). Dashed lines indicate the level of the first fEPSP. C, Summary data showing the CA1 fEPSP5/fEPSP1 ratio plotted against stimulation frequency in the absence or presence of APV (control: n = 6 slices; APV: n = 4 slices). Error bars indicate SE. Two-way ANOVA followed by Bonferroni's multiple comparisons test (F(1,18) = 7.095).

Discussion

In this study, we found that high-frequency bursts or naturalistic firing sequences of presynaptic PP input lead to differential temporal summation at CA1 and CA2 postsynaptic targets. In CA1, robust temporal summation enables a brief burst of PP stimuli to elicit a large postsynaptic depolarization sufficient to evoke AP output, whereas a single PP stimulus produces only a weak subthreshold response. In contrast, a single PP stimulus to CA2 can elicit a large postsynaptic depolarization sufficient to elicit spike output, and this response is only modestly enhanced by weak temporal summation in response to a high-frequency burst of PP stimuli. As a result of these different dynamics, CA2 PNs responded rapidly and precisely to PP input, whereas CA1 PNs required high-frequency bursts to elicit APs that occurred probabilistically late in a train. We further showed that the differential activation of NMDARs at synapses formed between PP inputs and their CA1 and CA2 PN targets contributes to the contrasting temporal dynamics.

Distinct temporal dynamics of cortico-hippocampal information transfer in CA1 versus CA2 PNs

Our study provides further insight into the differential transfer of information at the PP to CA1 compared with CA2 PNs. Our previous work focused on a set of dendritic and synaptic mechanisms by which a single PP stimulus elicits a 5- to 6-fold larger depolarization at the soma of CA2 compared with CA1 PNs. This includes the finding that there are 3 times as many dendritic spines in CA2 distal dendrites compared with CA1 (Srinivas et al., 2017). In addition, CA1 and CA2 dendrites differ in their passive and active integrative properties, enhancing the somatic response to dendritic Na+ spikes in CA2 compared with CA1 (Sun et al., 2014). As a result, PP stimuli are able to evoke large suprathreshold EPSPs that trigger dendritic Na+ spikes that propagate to CA2 soma to trigger APs (Sun et al., 2014; Srinivas et al., 2017). Our present results extend these findings by showing that a greater temporal summation at PP-CA1 synapses normalizes some of the differences with CA2 in terms of overall excitatory drive. However, the distinct dendritic and synaptic mechanisms give rise to strikingly contrasting dynamics of information transfer between PP inputs and their CA1 and CA2 targets. Whereas CA2 PNs fire precisely and reliably in response to EC input, CA1 PNs require a burst of EC activity to trigger spike output. As CA2 PNs are capable of firing APs immediately following the firing of presynaptic EC neurons, this mechanism should in principle enable CA1 PNs to more effectively integrate CA2-to-CA1 input with EC-to-CA1 input. Moreover, the higher firing rate in CA2 in response to naturalistic stimulating sequence is consistent with in vivo recordings, which also reported a higher mean firing rate in CA2 than CA1 (Mankin et al., 2015; Donegan et al., 2020).

The finding that PP inputs are strongly attenuated by the CA1 dendritic cable properties (Golding et al., 2005) has long been difficult to reconcile with more recent evidence that these inputs have a powerful influence on behaviors and CA1 firing in vivo. Lesion and genetic manipulations indicate that the PP-CA1 inputs contribute to spatial working memory, trace fear conditioning (Suh et al., 2011; Kitamura et al., 2014), contextual learning (Basu et al., 2016), and de novo place cell formation (Bittner et al., 2015, 2017). Indirect evidence that PP inputs can drive CA1 firing in vivo comes from findings that CA1 place field firing is largely maintained following surgical or chemical lesions of the CA3-CA1 input (Brun et al., 2002, 2008) or genetic silencing of CA3 (Nakashiba et al., 2008). Our finding of powerful temporal summation of PP inputs to CA1, together with previous findings that PP inputs can trigger CA1 dendritic spikes and dendritic plateau potentials (Golding et al., 2002; Jarsky et al., 2005; Takahashi and Magee, 2009; Bittner et al., 2015; Kim et al., 2015), suggests that the PP inputs can be effectively tuned to drive CA1 output and influence behavior depending on input dynamics.

As nearly all functionally identified EC principal cells, including grid cells and head direction cells, produce burst firing when animals perform behaviorally relevant tasks (Burgalossi et al., 2011; Tang et al., 2015), this mechanism should be particularly effective in acting as a high-pass filter to enhance information transfer through the EC-to-CA1 pathway during such behaviors. By contrast, the suprathreshold response of CA2 PNs to PP input is less sensitive to the firing rate of presynaptic EC neurons. Rather, the weaker frequency-dependent amplification in CA2 PNs, in combination with previously identified synaptic and dendritic integrative properties (Sun et al., 2014; Srinivas et al., 2017), enables CA2 PNs to respond to EC input more quickly and faithfully. How these contrasting temporal dynamics process cortical information through the hippocampal circuitry in vivo remains to be explored in future studies.

Mechanisms underlying contrasting frequency-dependent temporal summation of PP-CA1 versus PP-CA2 synapses

Our data suggest that NMDAR activation is essential for regulating the distinctive dynamics of CA1 versus CA2 PNs. However, the mechanism by which NMDARs enhance temporal summation at PP-CA1, relative to PP-CA2, synapses remains unclear. Activation of both presynaptic and postsynaptic NMDARs can influence postsynaptic responses and plasticity (McGuinness et al., 2010; Hunt and Castillo, 2012; Dore et al., 2017; Lituma et al., 2021). We found that bath application of APV reduced the size of the EPSP by 20%-25% in response to a single PP stimulation in both CA1 and CA2 PNs, indicating comparable expression levels of postsynaptic NMDARs between CA1 and CA2. Previous studies suggest that activation of presynaptic NMDARs at certain synapses, including SC-CA1 and mossy fiber synapses in the hippocampus (McGuinness et al., 2010; Lituma et al., 2021), facilitates glutamate release from presynaptic terminals. However, our finding that the PP fEPSP in CA1 is not enhanced during burst stimulation suggests that the large temporal summation seen in CA1 is not because of enhanced glutamate release from PP terminals. Thus, we propose that distal apical dendrites in CA1, relative to CA2, are equipped with a distinct unknown mechanism that requires NMDAR activation to enhance postsynaptic dendritic integration. Whether the relevant NMDARs are located in the postsynaptic CA1 neuron or presynaptic PP terminals remains to be determined.

Footnotes

This work was supported by National Institutes of Health Grants R01MH104602 and R01MH106629 to S.A.S. and Grant K01MH117444 to Q.S. We thank S.A. Hussaini for providing the raw data of the in vivo EC firing sequences; and S. Hassan for helpful discussions and comments on the first version of the manuscript.

The authors declare no competing financial interests.

References

  1. Augustinaite S, Heggelund P (2007) Changes in firing pattern of lateral geniculate neurons caused by membrane potential dependent modulation of retinal input through NMDA receptors. J Physiol 582:297–315. 10.1113/jphysiol.2007.131540 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Basu J, Zaremba JD, Cheung SK, Hitti FL, Zemelman BV, Losonczy A, Siegelbaum SA (2016) Gating of hippocampal activity, plasticity, and memory by entorhinal cortex long-range inhibition. Science 351:aaa5694. 10.1126/science.aaa5694 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bittner KC, Grienberger C, Vaidya SP, Milstein AD, Macklin JJ, Suh J, Tonegawa S, Magee JC (2015) Conjunctive input processing drives feature selectivity in hippocampal CA1 neurons. Nat Neurosci 18:1133–1142. 10.1038/nn.4062 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bittner KC, Milstein AD, Grienberger C, Romani S, Magee JC (2017) Behavioral time scale synaptic plasticity underlies CA1 place fields. Science 357:1033–1036. 10.1126/science.aan3846 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brun VH, Otnass MK, Molden S, Steffenach HA, Witter MP, Moser MB, Moser EI (2002) Place cells and place recognition maintained by direct entorhinal-hippocampal circuitry. Science 296:2243–2246. 10.1126/science.1071089 [DOI] [PubMed] [Google Scholar]
  6. Brun VH, Leutgeb S, Wu HQ, Schwarcz R, Witter MP, Moser EI, Moser MB (2008) Impaired spatial representation in CA1 after lesion of direct input from entorhinal cortex. Neuron 57:290–302. 10.1016/j.neuron.2007.11.034 [DOI] [PubMed] [Google Scholar]
  7. Burgalossi A, Herfst L, von Heimendahl M, Forste H, Haskic K, Schmidt M, Brecht M (2011) Microcircuits of functionally identified neurons in the rat medial entorhinal cortex. Neuron 70:773–786. 10.1016/j.neuron.2011.04.003 [DOI] [PubMed] [Google Scholar]
  8. Chevaleyre V, Siegelbaum SA (2010) Strong CA2 pyramidal neuron synapses define a powerful disynaptic cortico-hippocampal loop. Neuron 66:560–572. 10.1016/j.neuron.2010.04.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Donegan ML, Stefanini F, Meira T, Gordon JA, Fusi S, Siegelbaum SA (2020) Coding of social novelty in the hippocampal CA2 region and its disruption and rescue in a 22q11.2 microdeletion mouse model. Nat Neurosci 23:1365–1375. 10.1038/s41593-020-00720-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dore K, Stein IS, Brock JA, Castillo PE, Zito K, Sjostrom PJ (2017) Unconventional NMDA receptor signaling. J Neurosci 37:10800–10807. 10.1523/JNEUROSCI.1825-17.2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dudek SM, Alexander GM, Farris S (2016) Rediscovering area CA2: unique properties and functions. Nat Rev Neurosci 17:89–102. 10.1038/nrn.2015.22 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Giocomo LM, Hussaini SA, Zheng F, Kandel ER, Moser MB, Moser EI (2011) Grid cells use HCN1 channels for spatial scaling. Cell 147:1159–1170. 10.1016/j.cell.2011.08.051 [DOI] [PubMed] [Google Scholar]
  13. Golding NL, Staff NP, Spruston N (2002) Dendritic spikes as a mechanism for cooperative long-term potentiation. Nature 418:326–331. 10.1038/nature00854 [DOI] [PubMed] [Google Scholar]
  14. Golding NL, Mickus TJ, Katz Y, Kath WL, Spruston N (2005) Factors mediating powerful voltage attenuation along CA1 pyramidal neuron dendrites. J Physiol 568:69–82. 10.1113/jphysiol.2005.086793 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Herrero JF, Laird JM, López-García JA (2000) Wind-up of spinal cord neurones and pain sensation: much ado about something? Prog Neurobiol 61:169–203. 10.1016/s0301-0082(99)00051-9 [DOI] [PubMed] [Google Scholar]
  16. Hunt DL, Castillo PE (2012) Synaptic plasticity of NMDA receptors: mechanisms and functional implications. Curr Opin Neurobiol 22:496–508. 10.1016/j.conb.2012.01.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jarsky T, Roxin A, Kath WL, Spruston N (2005) Conditional dendritic spike propagation following distal synaptic activation of hippocampal CA1 pyramidal neurons. Nat Neurosci 8:1667–1676. 10.1038/nn1599 [DOI] [PubMed] [Google Scholar]
  18. Kim Y, Hsu CL, Cembrowski MS, Mensh BD, Spruston N (2015) Dendritic sodium spikes are required for long-term potentiation at distal synapses on hippocampal pyramidal neurons. Elife 4:e06414. 10.7554/eLife.06414 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kitamura T, Pignatelli M, Suh J, Kohara K, Yoshiki A, Abe K, Tonegawa S (2014) Island cells control temporal association memory. Science 343:896–901. 10.1126/science.1244634 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lituma PJ, Kwon HB, Alvina K, Lujan R, Castillo PE (2021) Presynaptic NMDA receptors facilitate short-term plasticity and BDNF release at hippocampal mossy fiber synapses. Elife 10:e66612. 10.7554/eLife.66612 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Mankin EA, Diehl GW, Sparks FT, Leutgeb S, Leutgeb JK (2015) Hippocampal CA2 activity patterns change over time to a larger extent than between spatial contexts. Neuron 85:190–201. 10.1016/j.neuron.2014.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. McGuinness L, Taylor C, Taylor RD, Yau C, Langenhan T, Hart ML, Christian H, Tynan PW, Donnelly P, Emptage NJ (2010) Presynaptic NMDARs in the hippocampus facilitate transmitter release at theta frequency. Neuron 68:1109–1127. 10.1016/j.neuron.2010.11.023 [DOI] [PubMed] [Google Scholar]
  23. Nakashiba T, Young JZ, McHugh TJ, Buhl DL, Tonegawa S (2008) Transgenic inhibition of synaptic transmission reveals role of CA3 output in hippocampal learning. Science 319:1260–1264. 10.1126/science.1151120 [DOI] [PubMed] [Google Scholar]
  24. Spruston N (2008) Pyramidal neurons: dendritic structure and synaptic integration. Nat Rev Neurosci 9:206–221. 10.1038/nrn2286 [DOI] [PubMed] [Google Scholar]
  25. Srinivas KV, Buss EW, Sun Q, Santoro B, Takahashi H, Nicholson DA, Siegelbaum SA (2017) The dendrites of CA2 and CA1 pyramidal neurons differentially regulate information flow in the cortico-hippocampal circuit. J Neurosci 37:3276–3293. 10.1523/JNEUROSCI.2219-16.2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Suh J, Rivest AJ, Nakashiba T, Tominaga T, Tonegawa S (2011) Entorhinal cortex layer III input to the hippocampus is crucial for temporal association memory. Science 334:1415–1420. 10.1126/science.1210125 [DOI] [PubMed] [Google Scholar]
  27. Sun Q, Srinivas KV, Sotayo A, Siegelbaum SA (2014) Dendritic Na+ spikes enable cortical input to drive action potential output from hippocampal CA2 pyramidal neurons. Elife 3:e04551. 10.7554/eLife.04551 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Sun Q, Sotayo A, Cazzulino AS, Snyder AM, Denny CA, Siegelbaum SA (2017) Proximodistal heterogeneity of hippocampal CA3 pyramidal neuron intrinsic properties, connectivity, and reactivation during memory recall. Neuron 95:656–672. 10.1016/j.neuron.2017.07.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Takahashi H, Magee JC (2009) Pathway interactions and synaptic plasticity in the dendritic tuft regions of CA1 pyramidal neurons. Neuron 62:102–111. 10.1016/j.neuron.2009.03.007 [DOI] [PubMed] [Google Scholar]
  30. Tang Q, Ebbesen CL, Sanguinetti-Scheck JI, Preston-Ferrer P, Gundlfinger A, Winterer J, Beed P, Ray S, Naumann R, Schmitz D, Brecht M, Burgalossi A (2015) Anatomical organization and spatiotemporal firing patterns of layer 3 neurons in the rat medial entorhinal cortex. J Neurosci 35:12346–12354. 10.1523/JNEUROSCI.0696-15.2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Zhang K, Xu T, Yuan ZM, Wei ZS, Yamaki VN, Huang MF, Huganir RL, Cai X (2016) Essential roles of AMPA receptor GluA1 phosphorylation and presynaptic HCN channels in fast-acting antidepressant responses of ketamine. Sci Signal 9:ra123. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from The Journal of Neuroscience are provided here courtesy of Society for Neuroscience

RESOURCES