We provided recent experimental evidence that coincident unitary events sum slightly sublinearly when targeting closely located postsynaptic sites. Simultaneous activation of many co-aligned inputs might lead to more significant nonlinear interactions especially in compartments of relatively small diameter. The axon initial segment of pyramidal cells has a limited volume and it receives inputs only from a moderate number of axo-axonic interneurons. We recorded the interaction of unitary axo-axonic inputs targeting a layer 4 pyramidal cell and determined the exact number and position of synapses mediating the effects. Both axo-axonic cells established three synaptic release sites on the axon initial segment of the postsynaptic cell which received a total of 19 synapses. The summation of identified inhibitory postsynaptic potentials (IPSPs) was slightly sublinear (9.4%) and the time course of sublinearity was slower than that of the IPSPs. Repeating the experiment while holding the postsynaptic cell in voltage clamp mode showed linear summation of inhibitory postsynaptic currents (IPSCs), suggesting that a local decrease in driving force could contribute to the sublinear summation measured in voltage recordings. The results indicate that moderate sublinearity during the interaction of neighboring inputs might be preserved in cellular compartments of relatively small volume, even if a considerable portion of all afferents converging to the same domain is simultaneously active.
Electron microscopic determination of the sites of synaptic junctions provided by the functionally tested converging afferents revealed that the distance between distinct simultaneously active inputs influences the degree of linearity of summation (Tamás et al., 2002). Supporting predictions made by cable theory (Jack et al., 1975; Segev et al., 1995; Koch, 1999), nonlinear interactions were recorded between inputs targeting closely located postsynaptic sites. Compartment specific interactions between afferents targeting closely situated membrane domains have already been detected between two inputs, although sublinearity was moderate. Simultaneous activation of many co-aligned inputs might lead to more significant nonlinear interactions, especially in compartments of relatively small diameter, but simultaneous recording of a significant fraction of all inputs targeting the same domain is difficult. The axon initial segment (AIS) of pyramidal cells appears to be the postsynaptic compartment of choice for these experiments since it receives inputs exclusively from axo-axonic interneurons and it was estimated that only three to six presynaptic axo-axonic cells (AACs) converge onto the same pyramidal cell (Somogyi, 1989).
AACs are local circuit neurons, which are unique elements of the cortical microcircuit (Somogyi et al., 1998). Their widespread occurrence in a variety of cortical regions suggests that AACs contribute to fundamental operations in cortical processing. Several electrophysiological and anatomical characteristics of AACs are similar to other GABAergic interneurons (Soriano and Frotscher, 1989; Gulyas et al., 1993; Buhl et al., 1994; Kawaguchi, 1995; Pawelzik et al., 1999; Maccaferri et al., 2000; Krimer and Goldman-Rakic, 2001), but AACs are distinct, providing uniform innervation of principal neurons exclusively targeting their axon initial segments (Somogyi, 1977). Axo-axonic interneurons elicit fast inhibitory postsynaptic potentials (IPSPs) showing paired pulse depression in their target cells (Buhl et al., 1994; Pawelzik et al., 1999; Maccaferri et al., 2000), but how axo-axonic inputs interact is not known. Here we present a case study of two axo-axonic inputs converging onto a pyramidal cell; this unique example resulted from >1700 triple and/or quadruple recordings with 188 convergent inputs targeting the same postsynaptic cell.
Experimental procedures applied in this study were published earlier (Tamás et al., 2002) and were carried out in accordance with the guidelines of NIH and University of Szeged. Briefly, a young (P24) Wistar rat was anaesthetized by the i.p. injection of ketamine (30 mg/kg) and xylazine (10 mg/kg) and, following decapitation, coronal slices (350 µm thick) were prepared from the somatosensory cortex. Slices were incubated at room temperature for 1 h in a solution composed of (in mM): 130 NaCl; 3.5 KCl; 1 NaH2PO4; 24 NaHCO3; 1 CaCl; 3 MgSO4; and 10 d(+)-glucose; saturated with 95% O2 and 5% CO2. The solution used during recordings differed only in that it contained 3 mM CaCl2 and 1.5 mM MgSO4. Micropipettes (5–7 MΩ) were filled with (in mM): 126 K-gluconate; 4 KCl; 4 ATP-Mg; 0.3 GTP-NA2; 10 HEPES; 10 phosphocreatine;and 8 biocytin (pH 7.25, 300 mOsm). Somatic whole-cell recordings were obtained at ∼36°C from three concomitantly recorded cells visualized in layer 4 by infrared differential interference contrast videomicroscopy (Zeiss Axioskop microscope Hamamatsu CCD camera, Luigs & Neumann Infrapatch set-up and two HEKA EPC 9/double patch-clamp amplifiers). Signals were filtered at 8 kHz, digitized at 16 kHz and analyzed with PULSE software (HEKA). Presynaptic cells were stimulated with brief (2 ms) suprathreshold pulses delivered at 7 s intervals, to minimize intertrial variability. Presynaptic cells were stimulated in cycles containing single presynaptic cell activations and synchronous and asynchronous dual presynaptic activation. For synchronous presynaptic activation, action potentials were timed to synchronize maximal unitary postsynaptic current amplitudes measured prior to the experiments testing convergence. Membrane potentials were corrected for junction potentials. Voltage clamp recordings were terminated and excluded from analysis when series resistance was >20 MΩ. Traces were excluded from the analysis if spontaneous PSPs occurred 20 ms before or 100 ms after the activation of identified responses; this process resulted in the elimination of <10% of events in a particular paradigm. All traces were offset to align their baselines for the period from –20 to 0 ms prior to the onset of current injections into the presynaptic neuron. The amplitude of postsynaptic event was defined as the difference between the peak amplitude and the baseline value measured 0–20 ms prior to presynaptic activation. Data for analysis of summation were used only from epochs in which the postsynaptic response remained stationary, i.e. the mean amplitude of 10 consecutive events remained within ±10% of the mean of the first 10 events of the epoch. The difference between the algebraic sums of single input responses and the recorded summed response was calculated during postsynaptic responses and expressed as a percentage of the maximal amplitude of the calculated response at the given time point. The resulting waveform is used as a measure of the degree of linearity over time.
The overall rate of connectivity in our slices between neighbouring neurons depends on the type of connection examined. In connections from interneurons to pyramidal cells the rate of connectivity is ∼0.23, meaning that one out of four or five potential postsynaptic cells receive IPSPs. The major problem in the case of AACs is that , to our knowledge, there is no electrophysiological fingerprint specific to AACs. The perisomatic morphology, fast spiking firing pattern and membrane parameters such as time constant and input resistance of AACs are similar to those of basket cells and some interneurons targeting the dendritic region of postsynaptic cells. Therefore, identification of AACs during recording is difficult and the experimenter has to rely on post hoc identification methods. The extact ratio of AACs among interneurons or among fast spiking cells is not known, but we encounter 1 AAC in ∼20 fast spiking cells in our specimens.
Visualization of biocytin and correlated light- and electron microscopy was performed as described (Tamás et al., 1997). Three-dimensional light microscopic reconstructions were carried out using Neurolucida (MicroBrightfield) with a 100× objective; synaptic distance and compartmental volume measurements were aided by Neuroexplorer (MicroBrightfield) software. Three-dimensional electron microscopic reconstructions were performed from 184 serial sections with Convert, Align, Trace and MergeWRL software tools from Synapse Web (http://synapses.mcg.edu/tools/index.stm).
We recorded the interaction of unitary axo-axonic inputs targeting a layer 4 pyramidal cell and determined the exact number and position of synapses mediating the effects. The two presynaptic AACs had a fast spiking firing pattern and the postsynaptic pyramidal cell showed a regular spiking behavior (Fig. 1A–C). When activating either AACs separately, the postsynaptic pyramidal cell responded with short-latency inhibitory postsynaptic currents (IPSCs) or IPSPs; both presynaptic cells evoked current and voltage responses of similar amplitude (50.1 pA, 58.8 pA and 0.85 mV, 0.90 mV, respectively; Fig. 1D–F). Application of a paired pulse protocol (interval, 60 ms) produced depression of the second responses (paired pulse ratios, 79 and 65%). Simultaneous activation of the two AACs produced a linear summation of IPSCs (101.4%) when comparing the peak amplitude of experimentally recorded compound responses (110.5 pA) with the calculated sum of individual inputs (109.0 pA) and the linearity of IPSC summation was maintained during the entire time course of responses as evidenced also by the almost perfectly overlapping recorded and calculated traces (Fig. 1D,E). Repeating the experiment in current clamp mode, however, indicated moderately sublinear summation (9.4%) as the experimentally recorded maximal amplitude of compound IPSP (1.58 mV) was smaller than those of corresponding algebraic sums of individual IPSPs (1.75 mV; Fig. 1F,G). The time course of the degree of linearity was distinct from the kinetics of IPSPs. Both individual and compound IPSPs showed exponential decays (τ = 19.2 ± 2.1 ms), but the kinetics of linearity could not be adequately fitted with conventional functions as the degree of linearity platoed around 92% for ∼30 ms before decaying back to baseline (half-width, 62.3 ms).
The somata of all three cells were located in layer 4 of the cortex (Fig. 2A,B). The dendrites of pyramidal cell and AAC1 were fully recovered, the distal dendrites of AAC2 could not be traced further than ∼100 µm from the soma. Most dendrites of the AACs originated from the upper or lower pole of the cell body and formed radially elongated fields. The axons of each cell were recovered and could be reconstructed in detail. Similar to previously established characteristics (Somogyi, 1977), both AACs gave rise to clusters of radially oriented small axonal branches. Electron microscopic sampling of postsynaptic targets taken from non-overlapping parts of the axonal arborizations identified both presynaptic cells as AACs since both cells established synaptic junctions exclusively on AISs (n = 13 and 14 for AAC1 and AAC2, respectively). High order branches of both presynaptic axons fromed close appositions with the AIS of the postsynaptic pyramidal cell and both AAC1 and AAC2 established three synaptic junctions in neighboring clusters at distances of 19 ± 3 and 40 ± 13 µm from the soma (Fig. 2C–G). Synapses formed by the two presynaptic cells were, on average, 20 ± 12 µm from each other and the distance between the nearest synapses of distinct sources was 6 µm. We determined the total number synaptic junctions (n = 19) on the axon initial segment of the postsynaptic cell between the axon hillock and the first branch point in serial electron microscopic sections. The complete three-dimensional reconstruction of the postsynaptic AIS and all presynaptic terminals showed that 11 unidentified presynaptic boutons formed 13 synapses in addition to the five identified boutons and six identified synaptic junctions.
Our results provide a direct experimental example that the summation of inputs targeting the AIS follows slightly sublinear summation in cortical neurons in vitro. The results indicate that the degree of sublinearity during the interaction of neighboring inputs might be similar in cellular compartments of different volume. In the postsynaptic pyramidal cell, the volume of the AIS measured from the hillock to the origin of the first axonal collateral was ∼47 µm3, which corresponded to ∼1.3% of the somatic volume (∼3500 µm3). It should be noted, however, that the soma may provided a sink for chloride ions entering the AIS during the experiments maintaining the gradient and therefore promoting linearity of summation. This could be of significant importance since we have showed that the degree of sublinearity between IPSPs targeting the soma and/or proximal dendrites was highly correlated with the amplitude of inputs suggesting the contribution of a drop in local driving force to sublinearity (Tamás et al., 2002). Ultimately, experiments addressing the summation of convergent inputs targeting the same dendritic spine could test interactions minimizing source/sink factors influencing summation properties (Shepherd and Brayton, 1987; Qian and Sejnowski, 1990; Koch, 1999). Furthermore, given that the synapses arriving from the two presynaptic cells formed separate proximal and distal clusters on the AIS, our recordings suggest that on-the-path shunting mechanisms (Koch, 1999) were suppressed in the case studied here. This could be due to the relatively high density of voltage gated ion channels on the AIS decreasing the significance of the synaptic conductance changes. The involvement of voltage activated currents, such as de-inactivation of sodium conductances by hyperpolarization, might be responsible for the difference in kinetics between the time course of linearity and the postsynaptic potentials.
One of the key factors shaping summation properties of any neuron is the background activity of synaptic conductances. The number of simultaneously active synapses on a particular cellular compartment might change dramatically depending on the behavioral state of the animal. However, the AIS receives inputs solely from AACs and the six synapses identified in this study represent 32% of all synapses which targeted the recorded pyramidal cell and ∼20–40% of all synaptic junctions targeting this compartment on average (DeFelipe et al., 1985; Somogyi et al., 1985; Somogyi, 1989; Farinas and DeFelipe, 1991). This suggests that, even if a significant portion of all afferents converging to the same domain is simultaneously active, linear or moderately sublinear summation might be conserved in the integration of inputs as detected at the soma.
We thank Professor P. Somogyi for comments on the manuscript and É. Tóth for technical help. This work was supported by the James S. McDonnell Foundation (EESI grant No 97-39), the Wellcome Trust, the Hungarian Scientific Research Fund (D32815) and the Hungarian Ministry of Education (FKFP 0106/2001). G.T. held a János Bolyai scholarship and J.Sz. was a Boehringer Ingelheim PhD scholar during part of this project.