Inhibitory coupling between inhibitory interneurons in the spinal cord dorsal horn

Abstract

Local inhibitory interneurons in the dorsal horn play an important role in the control of excitability at the segmental level and thus determine how nociceptive information is relayed to higher structures. Regulation of inhibitory interneuron activity may therefore have critical consequences on pain perception. Indeed, disinhibition of dorsal horn neuronal networks disrupts the balance between excitation and inhibition and is believed to be a key mechanism underlying different forms of pain hypersensitivity and chronic pain states. In this context, studying the source and the synaptic properties of the inhibitory inputs that the inhibitory interneurons receive is important in order to predict the impact of drug action at the network level. To address this, we studied inhibitory synaptic transmission in lamina II inhibitory interneurons identified under visual guidance in spinal slices taken from transgenic mice expressing enhanced green fluorescent protein (EGFP) under the control of the GAD promoter. The majority of these cells fired tonically to a long depolarizing current pulse. Monosynaptically evoked inhibitory postsynaptic currents (eIPSCs) in these cells were mediated by both GABAA and glycine receptors. Consistent with this, both GABAA and glycine receptor-mediated miniature IPSCs were recorded in all of the cells. These inhibitory inputs originated at least in part from local lamina II interneurons as verified by simultaneous recordings from pairs of EGFP+ cells. These synapses appeared to have low release probability and displayed potentiation and asynchronous release upon repeated activation. In summary, we report on a previously unexamined component of the dorsal horn circuitry that likely constitutes an essential element of the fine tuning of nociception.

Figure 1

Characterization of the GAD65-EGFP+ neurons. A. Photomicrograph of a lumbar transverse section of the spinal cord showing the distribution of EGFP expression. Scale bar = 200 μm. B. immunostaining for parvalbumin (red) and GFP (green) shows some co-localization in lamina II (yellow). Scale bar = 50 μm. C. Examples of EGFP+ neurons, reconstructed from neurobiotin/Lucifer Yellow filled cells. a-b: islet cells, c: vertical cell, d-e: unclassified cells. All cells are shown in the parasagittal plane. The majority of the dye filled cells were islet cells (9/15), one cell was vertical and the rest (5/15) could not be classified. D. Firing patterns in response to 500 ms-long depolarizing and hyperpolarizing current steps from a holding potential of -65 mV. Left: The response of a tonic firing neuron to current steps of -30 pA and +40 pA. Middle: A neuron that fired one or few spikes at the onset of the depolarization. Responses to current steps of -30 pA and +40 pA. The Inset shows the average instantaneous firing frequency (f) plotted against stimulus intensity (I). The f-I curve for this neuron being linear, it was classified as tonic firing according to previously established criteria [20]. Right: Response of a single spiking neuron to -30 pA and +90 pA current steps.

Figure 2

Lamina II inhibitory neurons receive both GABA_AR- and glycine receptor-mediated inhibitory inputs. IPSCs evoked by focal electrical stimulation while recording at a holding potential (HP) of 0 mV using CsMeSO₃-filled micropipettes (low [Cl^-]_i). The traces represent averages of 10 consecutive responses. These IPSCs displayed complex kinetics. Administration of strychnine (0.5 μM; Left) blocked the fast decaying component, indicating that it was mediated by activation of glycine receptors, while administration of SR95531 (gabazine, 10 μM; Right) blocked the slow decaying component, indicating that is was mediated by activation of GABA_A receptors. The relative contribution of the two components varied from cell to cell: the fast component comprised 78% (range 54%–94%) of the peak, while the slow component 22% (range 6%–46%) Inset: IPSCs evoked by focal stimulation while recording at a holding potential of -70 mV using CsCl-filled micropipettes (high [Cl^-]_i) to record inward currents (Scale bars = 25 pA, 50 ms). Recordings were performed in the presence of strychnine and addition of SR95531 abolished all remaining inward currents indicating that the evoked responses were entirely mediated by GABA_A and glycine receptors.

Figure 3

GABA_AR- and glycine receptor-mediated mIPSCs in Lamina II inhibitory neurons. A. Recordings of mIPSCs in presence of 1 μM TTX at a holding potential of 0 mV using CsMeSO₃-filled micropipettes. mIPSC frequency was 0.38 ± 0.06 Hz (n = 9). The traces on the left show mIPSCs in control conditions (top and middle), displaying different decay kinetics, and in the presence of strychnine (bottom) displaying only slow decay kinetics. The traces on the right are averages of 100 consecutive mIPSCs in control (top) and in the presence of strychnine (bottom). Administration of strychnine abolished the fastest component of the complex decay kinetics of the average mIPSCs confirming that this component was mediated by glycine receptors as previously established [26,27]. Inset: Histogram showing the distribution of GABA_AR-mediated mIPSC decay τ in the presence of strychnine. B. Three individual mIPSCs showing characteristic different decay kinetics: fast decay (left; absent in strychnine thus presumably glycinergic); slow decay (middle; GABA_A-mediated); mixed fast and slow (both glycine and GABA_A-mediated). The proportion of fast, slow or mixed mIPSCs varied across cells. Fast decaying mIPSC proportions ranged from 17% to 93%, slow decaying mIPSCs from 5% to 78% and mixed mIPSCs from 1% to 10%. Five out of nine neurons showed predominately (> 70%) fast mIPSCs, two out of nine were having predominately (> 70%) slow mIPSCs. C. Cumulative probability plot of the slow mIPSC decay τ (black) and the decay τ of the slow component from mixed mIPSCs (red), no significant difference between the two populations was found (p > 0.05; Kolmogorov-Smirnov test). All recordings we performed in the presence of 10 μM CNQX and 40 μM APV.

Figure 4

Different forms of plasticity at inhibitory synapse onto lamina II inhibitory neurons. Recordings were made in the presence of APV, CNQX and strychnine. A. Summation of evoked IPSCs (eIPSCs) caused by a train of 12 focal stimuli (dots at bottom) at 20 Hz. Stimulus artefacts are partially blanked. Note the occurrence of numerous sIPSCs at the end of the train indicative of asynchronous release. B. The normalized amplitude of the first eIPSC evoked by each train is plotted for the initial five consecutive repetitions of the trains (at 0.05 Hz; black). The amplitude of the first eIPSC in each train was normalized to the average of five IPSCs evoked by single stimuli (at 0.1 Hz) that preceded the train stimulation protocol. The values plotted are the means ± SEM from 20 neurons. The graph shows a significant (p < 0.05, Friedman test) increase in the eIPSC amplitude after the initial train (train number 1). Asterisks denote significant differences in subsequent values from the initial train (train number 1; Student-Newman-Keuls posthoc test; p < 0.05). For comparison, the normalized amplitudes of five consecutive IPSCs and the 13^th IPSC evoked by single stimuli are plotted (red). This indicates that the potentiation of the eIPSCs after the first train (of 12 stimuli) does not occur with accumulation of the same number of stimuli at low frequency. C. Left: Normalized amplitudes of the initial four eIPSCs within the first (black) and fifth (red) train. The values plotted are the means ± SEM from 20 neurons. Right: Ratios between the second eIPSC and the first eIPSC (eIPSC₂/eIPSC₁) from individual trains are plotted against the normalized amplitude of the first eIPSC. The data shows negative correlation between these values (linear fit: r = -0.53, p < 0.05), indicating depression of the second eIPSC when the first was potentiated and vice versa. D. Peristimulus time histogram showing the change in sIPSC frequency before and after each train (arrow; n = 15 trains from one cell).

Figure 5

Inhibitory coupling between pairs of lamina II inhibitory interneurons. A. Example recordings from a pair of synaptically connected neurons. Top traces are from the presynaptic neuron showing a train of action potentials evoked by a 100 ms depolarizing current step. Bottom traces show simultaneous recordings of IPSCs from the postsynaptic cell. Note on the right, the release failure in response to the first action potential. Most IPSCs in this pair displayed fast decay kinetics with a slow component also visible in some events. B. Graph showing the probability p of release for each of the first five presynaptic action potentials in the train calculated from the pair of neurons shown in A (p defined as number of IPSCs per number of corresponding action potential in 50 consecutive trains delivered at 0.1 Hz). Note the dramatic increase in release probability in response to the 2nd action potential in the train. C. Probability of release for each of the first 4 action potentials in the train averaged for all 3 synaptically connected pairs studied. Values are means ± SEM. D. The probability p of release per train calculated for each train from the pair of neurons shown in A. The plot was smoothed with a running average of four values. Note the transient increase in probability after the initial trains.