My Project

What I am studying

The hippocampus is a brain structure that is involved in memory formation, where inputs (contextual information) are processed here and converted to outputs (firing of pyramidal cells). The firing of pyramidal cells are regulated by more than 20 types of inhibitory interneurons, and we still don't know exactly how this works.

I am studying one type of interneuron, called the neurogliaform cells (NGFCs), to see how they function within the hippocampal circuit. NGFCs form synapses onto the distal tufts of pyramidal cell apical dendrites alongside excitatory inputs from the entorhinal cortex. NGFCs also express neuronal nitric oxide synthase (nNOS), are often synaptically coupled, and fire during theta oscillations in vivo.

Show figure of hippocampal interneuron subtypes

Peter Somogyi and Thomas Klausberger, The Journal of Physiology, 2005

What I have found, a summary

In this project, I describe a novel feature of synaptic communication between these interneuron types, hereafter referred to as the firing induced suppression of inhibition (FSI). Specifically, I found that when a theta-associated activity patterns were evoked in NGFCs from rodent hippocampal slices, the cells showed a transient reduction in unitary IPSP amplitude. My data suggest that FSI requires the backpropagation of action potentials, calcium influx through L-type calcium channels, nNOS activity within the dendrites of interneurons, and activation of NO-sensitive guanylyl cyclase (NOsGC) receptors, which are present on presynaptic terminals. My results also demonstrate the physiological impact of this phenomenon by showing that when FSI takes place the strength of incoming excitatory postsynaptic potentials onto NGFC is transiently sharpened. Specifically, FSI indirectly increased the amplitude of EPSPs. Thus FSI may enhance spatial and temporal summation of excitatory inputs to NGFCs, thus regulating their inhibition of pyramidal cells.

Here are the details

1. In vivo firing induces suppresion of inhibition (FSI)

I have recorded the unitary inhibitory postsynaptic potentials (uIPSPs) or currents (uIPSCs) from synaptically-coupled putative NGFCs or other interneurons with the soma in the SLM of the rat hippocampus in vitro. Neurons were recorded with a high chloride internal solution were used to increased the signal of these events and as a result these events were depolarising potentials or inward currents. A firing sequence (60 s duration, average frequency and CV values: 8.3 Hz, 0.75; stimulation protocol) was recorded in a rat NGFC in vivo during theta oscillations (Fuentealba et al., 2010) and which was injected in current clamp configuration in a postsynaptic interneuron of the SLM recorded in vitro. I have found that shortly after (125 or 250 ms) the firing of the postsynaptic cell from the in vivo sequence, there was a transient decrease in the amplitude of the uIPSPs recorded in current clamp or voltage clamp configuration. The kinetics of baseline and decreased uIPSCs evoked before and immediately after the stimulation protocol was similar, suggesting that presynaptic mechanisms were more likely than postsynaptic mechanisms to be responsible for the decrease of the synaptic responses.

In summary, I have found that in vivo NGFC firing during hippocampal theta rhythm induces a DSI-like event (FSI) recorded from pairs of interneurons with the soma in the SLM. The following experiments have been designed to study the cell types and mechanisms involved in this phenomenon.

NGFC in vivo firing pattern induces a transient suppression of synaptic inhibition
A, NGFC paired recording in vitro, current clamp mode, presynaptic action potentials (PRE, black traces) evoked depolarising uIPSPs in a postsynaptic cell recorded with an electrode filled with 84 mM Cl- solution (POST, left dark blue traces superimposed recorded 120 s or 60 s before the stimulation protocol). Injection of firing recorded in vivo from an NGFC for 60 s (stimulation protocol) in the postsynaptic NGFC recorded in vitro induced a transient decrease of the amplitude of the uIPSP (POST, middle trace), that returned to the baseline level 60 s and 120 s after the end of the stimulation protocol (right traces superimposed). B, NGFC paired recording, voltage clamp mode, presynaptic action currents (PRE, black traces) evoked inward uIPSCs in a postsynaptic cell recorded with an electrode filled with 84 mM Cl- solution (POST, left light blue traces superimposed recorded 120 s or 60 s before the stimulation protocol). Injection of in vivo NGFC firing for 60 s (stimulation protocol) in the postsynaptic NGFC induced a transient decrease of the amplitude of the uIPSC (middle trace), that returned to the baseline level 60 s and 120 s after the end of the stimulation protocol (right traces superimposed). Inset, scaled traces show no changes in the kinetics of the uIPSCs before and during FSI. C, mean uIPSP (CC, current clamp) or uIPSC (VC, voltage clamp) peak amplitude before and after the stimulation protocol (repeated sequentially three times) for the cell pairs shown in A and B; error bars are SEM. D, summary of normalised peak amplitudes of uIPSPs (CC, current clamp) or uIPSCs (VC, voltage clamp) before and after the stimulation protocol in all pairs showing FSI, error bars are SEM (p < 0.001, n = 23 for CC data, n = 25 for VC data). In the graphs of panels C and D of this and subsequent figures, 0 s indicates mean uIPSP or uIPSC amplitude detected 125 or 250 ms after the end of the stimulus protocol.
2. What happens during in vivo firing?

FSI is temporally-correlated with the postsynaptic firing, but is postsynaptic cell firing required for FSI? I have studied this question by using a subthreshold protocol and found that a subthreshold protocol was unable to induce FSI in all cell pairs tested. This result suggests that FSI was initiated by the postsynaptic neuron firing, indicating that active processes were likely to be involved. According to this result, Action potential backpropagation (bAP) could be one of the active processes that occur, namely bAP travels along the postsynaptic dendrites in order to trigger FSI. To test this hypothesis, I have visualised bAPs in interneurons of the SLM by using single cell voltage imaging and they show robust action potential backpropagation

Backpropagating action potential along the dendrites of NGFCs or other interneurons of the SLM
A, Right image, NGFC (filled with biocytin, visualised after biocytin-streptoavidin-Cy3 reaction) showing characteristic stellate dendrites (objective x20). Left image, the same neuron in A, which was loaded with JPW-1114 and voltage imaging was performed from dendritic sites included in the red box; in particular, ΔF/F signals at the sites indicated (1-4, 4x4 pixels) were analysed. B, Left, A depolarising current pulse (800 pA, 3 ms) induced a somatic action potential (AP) recorded in current clamp (green trace) and the corresponding ΔF/F signals from the site 1-4 in A are illustrated. The red vertical bar denotes the AP peak recorded from the soma. B, Middle traces, fractional changes of voltage fluorescence (ΔF/F) in response to a ~ 10 mV membrane potential hyperpolarisation lasting 350 ms (average of 9 traces) injected into the soma and recorded from sites 1-4 of the dendrites; somatic current clamp recording is illustrated on the top. Right traces, ΔF/F of the bAPs peak amplitude calibrated with the corresponding ΔF/F induced by the hyperpolarising current pulse recorded from regions 1-4 of the dendrites. C, summary graph showing bAP peak amplitude recorded at different dendritic sites normalised to the bAP amplitude recorded at 20-30 μm (shortest distance) from the soma (light blue, calibrated; dark blue, uncalibrated; n = 24). D, group data of calibrated bAP peak amplitude shown in C. Note that bAPs occurs in NGFCs and in other interneurons of the SLM (non-NGFC) without any significant decrement (bAP peak amplitude detected at dendrites >70 μm or ~20 μm away from the soma was not significantly different, p > 0.5, n = 24). E, Top trace, effect of 1 μM TTX on an AP recorded in current clamp evoked by a depolarising current pulse (800 pA x 3 ms, top traces). Bottom trace, effect of 1 μM TTX on ΔF/F bAP recorded from a dendrite of an NGFC. Blue traces, control; Red traces, TTX. Note that TTX blocks the ΔF/F of the bAP, whereas it spares ΔF/F of the subthreshold depolarisation evoked by the somatic subthreshold depolarising pulse.

A possible function for bAP is to activate downstream enzymes that are responsible for the synthesis of retrograde signalling molecules. According to this idea, I have tested the hypothesis that the postsynaptic bAPs may elicit an increase in the dendritic calcium concentration. Using calcium imaging I have found that cell pairs showing FSI have higher dendritic calcium level.

Sustained elevation of Ca2+ in the dendrites of postsynaptic interneurons triggers FSI
A, ΔF/F Ca2+ signal from an NGFC with FSI (A1) or without FSI (A2) filled with OG5N (blue trace) associated with an AP (green trace, single trial). B, Same as A (in the same neurons) but for a train of ten APs at 100 Hz (single trial); note the increased decay time compared to the calcium signal associated with single AP; the red trace is the summation of 10 template ΔF/F Ca2+ signals associated with the single AP (note the faster decay time). C, ΔF/F Ca2+ signals (blue traces) associated with the stimulation protocol (green traces, single trial) from the dendrites of two NGFC cells, one exhibiting FSI (C1) and the other not showing FSI (C2); the red traces are again the linear summation of single AP template ΔF/F Ca2+ signals; note the larger supralinear summation of ΔF/F Ca2+ signals in the cell exhibiting FSI. D, Columns and error bars are the mean and the SEM of the difference of ΔF/F dendritic Ca2+ signal occurring 60 s and 10 s after the onset of the stimulation protocols in postsynaptic neurons with FSI (blue column) and in cells without FSI (red column); this value in the two populations of neurons was significantly different (n = 6, p < 0.05). E, calibration of OG5N ΔF/F Ca2+ signal in terms of MF2 ΔF/F Ca2+ signal: plot of OG5N ΔF/F peak vs. MF2 ΔF/F peak associated with different protocols (spade, 5 APs at 100 Hz; square, 10 APs at 100 Hz; triangle, 20 APs at 100 Hz, n = 6) in the same cells. The linear fit gives a conversion factor of 2.5; given that 1% for MF2 corresponds to ~250 nM, 1% for OG5N corresponds to ~100 nM.

Using pharmcology I have found this calcium influx is indeed through calcium channels (L-type) and not from stores.

The L-type Ca2+ blocker nimodipine inhibits FSI
A, Presynaptic action currents (PRE, black traces) evoked inward uIPSCs in a postsynaptic NGFC recorded with an electrode filled with 84 mM Cl- solution (POST, blue traces). Injection of the stimulation protocol in the postsynaptic cell induced a transient depression of the uIPSC amplitude (middle blue trace), that returned to control level 60 s after the end of the stimulation protocol (right blue trace). Application of the L-type Ca2+ channel blocker nimodipine (10 μM) inhibited FSI (red traces). B, quantification of the uIPSC peak amplitude that occurred before or during the bath application of nimodipine for the data shown in A (dark symbols denote values immediately after stimulation protocol). C, normalised data (mean and SEM) for all uIPSCs studied with this protocol (FSI control was 70.9 ± 6.0% and 96.7 ± 3.5% with nimodipine, p < 0.01, n = 4).
3. Which messenger(s) is responsible for FSI?

Using pharmacology I have found that nitric oxide is responsible for FSI. FSI is blocked by the application of the nNOs inhibitor L-NAME, NOs-GC antogonist ODQ, and potentiated by the nitric oxide precursor L-arginine.

The nNOS inhibitor L-NAME blocks FSI
A, Presynaptic APs (PRE, black traces) elicited depolarising uIPSPs (control, left traces superimposed) recorded with an electrode filled with 84 mM Cl- solution showing FSI immediately after stimulation protocol applied to a postsynaptic NGFC (control, blue middle trace). The uIPSP recovered 60 s after the stimulation protocol (blue trace, right). Application of the nNOS inhibitor L-NAME (200 μM) blocked FSI (red traces). B, quantification of the uIPSP peak amplitude that occurred before or after the stimulation protocol (dark symbols) in control and in the presence of L-NAME for the data shown in A. C, normalised data (mean and SEM, shaded areas) for all uIPSPs or uIPSCs studied with this protocol (FSI control was 76.6 ± 2.5% and 95.1 ± 2.4% with L-NAME, p < 0.001, n = 10).
The NO receptor antagonist ODQ blocks FSI
A, Presynaptic action currents (PRE, black traces) evoked inward uIPSCs in a postsynaptic NGFC recorded with an electrode filled with 84 mM Cl- solution (POST, blue traces). Injection of the stimulation protocol in the postsynaptic cell induced a transient depression of the uIPSC amplitude (middle blue trace), that returned to control level 60 s after the end of the stimulation protocol (right blue trace). Application of the NO receptor antagonist ODQ (10 μM) blocked FSI (red traces). B, quantification of the uIPSC peak amplitude that occurred before or in the presence of ODQ for the data shown in A (dark symbols denote values immediately after stimulation protocol). C, normalised data (mean and SEM) for all uIPSPs or uIPSCs studied with this protocol (FSI control was 71.1 ± 3.7% and 96.8 ± 2.3% with ODQ, p < 0.01, n = 8).
4. What are the physiological implications of FSI?

To test the physiological role of FSI on single cell integration, I have injected a train of EPSPs as synaptic conductance by using dynamic clamp (dEPSPs) in postsynaptic NGFCs. Since NGFCs have short dendrites and are biophysically compact, the somatic injection of unitary dEPSPs should represent a realistic representation of EPSPs evoked by stimulation of one or a few fibres present in the SLM, such as the perforant path from the entorhinal cortex. I observed that when FSI occurred, the amplitude of the dEPSPs was significantly increased compared to dEPSPs elicited in the presence of control synaptic inhibition before or after FSI (note, normal chloride internal solution was used). Thus, FSI transiently increased the size of the dEPSPs thereby promoting the temporal/spatial integration of the uEPSPs elicited by stimulation of perforant path/other excitatory fibres present in the SLM.

FSI transiently modulates EPSPs impinging onto NGFCs
A, An EPSP train was injected as synaptic conductance using dynamic clamp (POST, dEPSPs, blue trace, - 60 s) in a postsynaptic NGFC while an action potential in a presynaptic NGFC (PRE, black trace) elicited a coincident uIPSP. Application of the stimulation protocol in the postsynaptic cell evoked a transient, smaller uIPSP (FSI). As a result, a significant enhanced depolarisation level was reached by several dEPSPs of the train (red trace), the depolarisation level of the EPSPs returned to the baseline 60 s after FSI (green trace, 60 s). B, C, normalised values of the dEPSPs amplitude during FSI (red bars) or 60 s after FSI (recovery, green bars) versus before FSI (blue histograms), for the events shown above (B) or for all recorded cell pairs (C). In the presence of FSI, the first three dEPSPs had significantly larger amplitude than before FSI (p < 0.05, n = 8).