Monosynaptic convergence of C- and Adelta-afferent fibres from different segmental dorsal roots on to single substantia gelatinosa neurones in the rat spinal cord

Abstract

Although it is known that each spinal cord segment receives thin-fibre inputs from several segmental dorsal roots, it remains unclear how these inputs converge at the cellular level. To study whether C- and Adelta-afferents from different roots can converge monosynaptically on to a single substantia gelatinosa (SG) neurone, we performed tight-seal recordings from SG neurones in the entire lumbar enlargement of the rat spinal cord with all six segmental (L1-L6) dorsal roots attached. The neurones in the spinal cord were visualized using our recently developed oblique LED illumination technique. Individual SG neurones from the spinal segment L4 or L3 were voltage clamped to record the monosynaptic EPSCs evoked by stimulating ipsilateral L1-L6 dorsal roots. We found that one-third of the SG neurones receive simultaneous monosynaptic inputs from two to four different segmental dorsal roots. For the SG neurones from segment L4, the major monosynaptic input was from the L4-L6 roots, whereas for those located in segment L3 the input pattern was shifted to the L2-L5 roots. Based on these data, we propose a new model of primary afferent organization where several C- or Adelta-fibres innervating one cutaneous region (peripheral convergence) and ascending together in a common peripheral nerve may first diverge at the level of spinal nerves and enter the spinal cord through different segmental dorsal roots, but finally re-converge monosynaptically on to a single SG neurone. This organization would allow formation of precise and robust neural maps of the body surface at the spinal cord level.

Figure 1. Tight-seal recording from SG neurones in the entire spinal lumbar enlargement

A, preparation on the entire lumbar enlargement of the spinal cord with preserved unilateral six dorsal roots, L1–L6. The roots were stimulated through suction electrodes. The neurones in the SG were viewed using oblique illumination by infrared LED. B, an image of an SG neurone within the spinal cord. Depth, 43 μm. C, cross sections cut from the fixed spinal cord. The sections of segments L3, L4 and L5 are shown from the preparation used for recordings from the L4 SG neurones. White lines show the border of the grey matter. The border of the grey matter in the cut part of the L4 section is shown as a symmetrical projection of the intact contralateral half. The region of the SG exposed for the patch-clamp recording is indicated by a white arrowhead in segment L4. The images were taken using room illumination (condenser illumination and LED were off) and digitally inverted. D, distribution of ascending and descending primary afferents in the dorsal white matter of a 27-day-old rat. Parasagittal sections from medial (left) and lateral (right) dorsal white matter were from regions indicated on the cross section of the dorsal horn (middle). The cut dorsal root is indicated by an asterisk. In the medial dorsal white matter, several biocytin-labelled thin afferents are indicated by filled arrowheads. In the narrow lateral white matter (right) no stained afferents were detected. Dotted lines show the border between the white and grey matter (in D and E). E, parasagittal section of medial dorsal white matter showing a thin primary afferent entering the grey matter. Parts of the image were taken in different focal planes. F, a 100-μm-thick parasagittal section of the spinal cord with two biocytin-labelled SG neurones. Two dotted lines show the dorsal and ventral borders of the white matter in the focal plane (top of the section). Continuous line shows the dorsal border of the white matter from the bottom of the section.

Figure 2. Testing the stimulation frequency in isolated dorsal roots

A, recordings of C-fibre CAPCs in isolated dorsal roots (L4, conduction distance, 4.5 mm). Each family of traces has 15 consecutive (individual) recordings of CAPCs activated at 1 Hz and 0.1 Hz. Fast A-components are truncated. B, Aδ-fibre CAPCs in isolated dorsal roots (L5, conduction distance, 10 mm). Each of 15 traces shown at 10 Hz and 1 Hz stimulation is an average of 20 individual traces. In experiments with Aδ-fibre CAPCs, a protocol with 15 stimulations at a given frequency was repeated 20 times (with 4 s intervals between protocols) and the corresponding episodes were averaged. C, examples of mono- and polysynaptic EPSCs recorded in an SG neurone. Each group is a superposition of 10 consecutive recordings. Holding potential was −70 mV. D, examples of recordings from an SG neurone held at −70 and −80 mV. Both EPSCs and IPSCs were seen at −70 mV, whereas only EPSCs were seen at −80 mV.

Figure 3. Recording from the L4 SG neurone with dominating C-input

Left, recordings of EPSCs elicited in one L4 SG neurone by stimulation of L1–L6 segmental dorsal roots using 1 ms pulses (100 μA). Holding potential was −70 mV. Monosynaptic C- and Aδ-fibre-mediated EPSCs are indicated by filled and open triangles, respectively. The triangles also show the time moment for which the latency analysis was done. Middle, roots with monosynaptic connections (L4, L5 and L6) were also stimulated with 50 μs pulses (100 μA). Holding potential was −70 mV. Here and in the following figures, recordings are shown as a superposition of 10 consecutive traces for roots with monosynaptic inputs (indicated by triangles) and 5 consecutive traces for roots without monosynaptic input. Right, schematic drawing of monosynaptic innervation of this SG neurone by C- and Aδ-afferents originating from the L4, L5 and L6 segmental dorsal roots.

Figure 4. Recording from the L4 SG neurone with dominating Aδ-input

Left, recordings of EPSCs elicited in an L4 SG neurone by the L1–L6 root stimulation with 1 ms pulses. Holding potential was −70 mV. Middle, roots with monosynaptic connections (L3, L4, L5 and L6) were also stimulated with 50 μs pulses. Holding potential was −80 mV. Here and in the following figures, the voltage-gated Na⁺ currents activated by EPSCs were truncated. Right, suggested organization of monosynaptic innervation of this SG neurone by C- and Aδ-afferents from the L3, L4, L5 and L6 segmental dorsal roots.

Figure 5. An SG neurone from the segment L3 with only C-fibre input

Left, recordings of EPSCs evoked in one L3 SG neurone by the L1–L6 root stimulations with 1 ms pulses. Holding potential was −70 mV. Middle, roots giving monosynaptic inputs (L2 and L3) were also tested with 50 μs pulses. Holding potential was −70 mV. Right, suggested organization of monosynaptic C-fibre inputs to this SG neurone from the L2 and L3 segmental dorsal roots.

Figure 6. An SG neurone from the segment L3 with dominating Aδ-fibre input

Left, EPSCs recorded in one L3 SG neurone after the stimulations of L1–L6 dorsal roots with 1 ms pulses. Holding potential was −70 mV. Middle, roots giving monosynaptic inputs (L3, L4 and L5) were also tested with 50 μs pulses. Holding potential was −70 mV. Right, suggested organization of monosynaptic C- and Aδ-fibre inputs to this SG neurone from the L3, L4 and L5 segmental dorsal roots.

Figure 7. Total inputs to the L4 and L3 SG neurones from the L1–L6 roots

Histograms showing the numbers of total inputs observed in the SG neurones located in spinal segments L4 and L3. These histograms include all cells with a monosynaptic input from at least one dorsal root. Neurones were included only if the preparation had all 6 roots undamaged.

Figure 8. Proposed model for the thin afferent convergence on to an SG neurone

Re-convergence model, three converging fibres originate from one cutaneous reference point, ascend in a common peripheral nerve and diverge at the level of segmental spinal nerves, in order to re-converge again on to an SG neurone. The Re-convergence model was developed to explain formation of precise and robust somatotopic maps and to comply with observations of Takahashi et al. (2003). Convergence model, in this alternative model three centrally converging afferents originate from different terminal branches of a common peripheral nerve. In the Convergence model, the peripheral divergence strongly increases the cutaneous area projecting to a single SG receptive field.