Corticospinal axons make direct synaptic connections with spinal motoneurons innervating forearm muscles early during postnatal development in the rat

Abstract

Direct connections between corticospinal (CS) axons and motoneurons (MNs) appear to be present only in higher primates, where they are essential for discrete movement of the digits. Their presence in adult rodents was once claimed but is now questioned. We report that MNs innervating forearm muscles in infant rats receive monosynaptic input from CS axons, but MNs innervating proximal muscles do not, which is a pattern similar to that in primates. Our experiments were carefully designed to show monosynaptic connections. This entailed selective electrical and optogenetic stimulation of CS axons and recording from MNs identified by retrograde labelling from innervated muscles. Morphological evidence was also obtained for rigorous identification of CS axons and MNs. These connections would be transient and would regress later during development. These results shed light on the development and evolution of direct CS-MN connections, which serve as the basis for dexterity in humans. Recent evidence suggests there is no direct connection between corticospinal (CS) axons and spinal motoneurons (MNs) in adult rodents. We previously showed that CS synapses are present throughout the spinal cord for a time, but are eliminated from the ventral horn during development in rodents. This raises the possibility that CS axons transiently make direct connections with MNs located in the ventral horn of the spinal cord. This was tested in the present study. Using cervical cord slices prepared from rats on postnatal days (P) 7-9, CS axons were stimulated and whole cell recordings were made from MNs retrogradely labelled with fluorescent cholera toxin B subunit (CTB) injected into selected groups of muscles. To selectively activate CS axons, electrical stimulation was carefully limited to the CS tract. In addition we employed optogenetic stimulation after injecting an adeno-associated virus vector encoding channelrhodopsin-2 (ChR2) into the sensorimotor cortex on P0. We were then able to record monosynaptic excitatory postsynaptic currents from MNs innervating forearm muscles, but not from those innervating proximal muscles. We also showed close contacts between CTB-labelled MNs and CS axons labelled through introduction of fluorescent protein-conjugated synaptophysin or the ChR2 expression system. We confirmed that some of these contacts colocalized with postsynaptic density protein 95 in their partner dendrites. It is intriguing from both phylogenetic and ontogenetic viewpoints that direct and putatively transient CS-MN connections were found only on MNs innervating the forearm muscles in infant rats, as this is analogous to the connection pattern seen in adult primates.

Figure 1. Monosynaptic EPSCs elicited in cervical MNs by electrical stimulation of the CST

A, schematic drawing of the experimental arrangement. Transverse cervical spinal cord slices of (C7–C8) were prepared from P7–P9 animals. The CST fibres were electrically stimulated, and whole‐cell recordings were made from MNs retrogradely labelled with Alexa 488‐CTB injected into distal or proximal target muscles. Inset: photomicrograph of fluorescently labelled MNs. B, optical imaging of CST‐evoked postsynaptic potentials recorded in the hemilateral spinal cord. Responses occurred 13.2 ms after the stimulation (left) and disappeared after electrolytic lesion (right). The black oblique bar on the spinal cord surface is a string holding the slice. C, neurofilament‐immunostaining of the slice in B, where the hemilateral part is shown (the yellow dotted rectangle). % ΔF/F, fluorescence change by stimulation / baseline fluorescence (%). The electrolytic lesion is restricted to within the dorsal column. The area enclosed by the white rectangle, which includes the dorsal column, is magnified in the right panel. D, representative traces of CS‐EPSCs recorded from MNs labelled with CTB‐Alexa 488 injected into forearm muscles voltage clamped at –90 mV. Each black trace is the average of 10 traces (original traces are shown in grey). The upper traces were recorded in normal ACSF, the lower ones in ACSF containing high concentrations of divalent cations (HDC; 7 mm Ca²⁺, 3 mm Mg²⁺). E, membrane currents recorded at –90 mV from a proximal muscle MN where monosynaptic CS‐EPSCs were not recorded. The upper 5 traces were recorded without stimulation. The middle 5 traces were recorded before, during and after 10 repetitive stimuli (arrows, interval 10 ms) over a span of 170 ms. The lower 5 traces were recorded after the perfusate was changed from normal ACSF to HDC. F, schematic drawing of the location of the stimulating electrodes. The distance of the dorsal edge of the CST from the tip of the stimulating electrode for CST (2) was nearly 175 μm. The location of Ia stimulation in the middle part of the dorsal column was 175 μm (1) from the dorsal edge of the CST. The location of the electrode shift over the midline to the contralateral CST is also shown (see text). G, responses to stimulation of the middle part (1) (ascending fibres, including Ia fibres) and the ventralmost part (the CST) of the dorsal column. Both types of stimulation evoked monosynaptic EPSCs, but with different latencies, in forearm MNs under HDC‐ACSF. The latency of responses to CST stimulation was significantly shorter than that to Ia stimulation (insets).

Figure 2. Distribution of monosynaptic CS‐EPSC‐positive and ‐negative motoneurons

A, summary graphs showing the numbers of CS monosynaptic EPSC‐positive and ‐negative neurons innervating distal (forearm) and proximal (pectoralis major and serratus anterior) muscles. B, locations of the monosynaptic CS‐EPSC‐positive (red) and ‐negative (blue) neurons examined using electrical stimulation. Red and blue dots indicate the centres of the somata of those neurons. Stars indicate the neurons where monosynaptic CS‐EPSCs were evoked by optogenetic stimulation. The fine dotted lines indicate the borders of the forearm, pectoralis major and serratus anterior MN groups, and were drawn based on the findings obtained after injection of CTB‐Alexa 488 into each muscle group, with reference to the atlas of MN pools (Watson et al. 2012).

Figure 3. Optogenetic stimulation of CS axons

A, schematic drawing similar to Fig. 1 A . AAV vector (pAAV‐CaMKIIα‐hChR2(H134R)‐EYFP) was injected into the unilateral cortex on P0. CTB‐Alexa 488 was injected into the muscle groups on P4–P6. The slice was illuminated with blue LED light through an optical fibre. Inset: EYFP‐ChR2‐containing axons at the C8 level in a transverse cord section stained with anti‐GFP antibody. The dotted line indicates the contour of the spinal cord. B, the illumination durations and the evoked EPSCs in a MN in normal ACSF. The duration of the light pulse (blue bars) was progressively shortened to the indicated times. Open circles indicate definite monosynaptic responses, filled circles those that may be polysynaptic (see text). Illumination of 1 ms evoked a no response (*). C, monosynaptic EPSCs in HDC‐ACSF obtained from another MN. Latencies generally increased in HDC solution, where membrane excitability was lower (Frankenhaeuser & Hodgkin, 1977). EPSC amplitude also depended on the level of ChR2 expression.

Figure 4. Close apposition of synaptophysin‐ECFP‐containing CS terminals to a MN

A, CS axons containing ECFP were stained with anti‐GFP antibody and detected with a 2nd antibody conjugated with Texas Red. This image was made from 26 Z‐stack images. MNs were retrogradely labelled with CTB‐Alexa 488 (green) in the C7 spinal cord on P7. The white line indicates the contour of the spinal cord. B, CS axons labelled with Texas Red. To eliminate non‐specific staining near the surface, Z projections of 5 serial images were made from the same Z‐stack images used in A. The dotted line indicates the border between the white and grey matter. C, MN pools located in the dorsolateral ventral horn. Shown is an expanded image of the area within the white box in A. D, close contact between a proximal dendrite of a MN and a bouton‐like putative presynaptic terminal of a CS axon (arrow). Shown is an expanded image of the area within the white box in C.

Figure 5. Close apposition between CS axons and MN dendrites where monosynaptic CS‐EPSCs were recorded

A, bright field micrograph of a transverse slice of hemilateral spinal cord where monosynaptic EPSC responses were recorded from a cell stained red by Neurobiotin‐NeutrAvidin‐Texas Red injected during the recording. B, magnification of the area within the white rectangular box in A. Shown is immunostaining of ChR2‐EYFP‐containing CS axons (green) and the traced image of the dendritic arbor of the MN (red). Ca–c, images taken from the white box C in B. Ca, fluorescence image of NeutrAvidin‐Texas Red labelling of Neurobiotin injected during recording. Cb, the soma was also retrogradely labelled with Alexa 488‐CTB injected into a forearm muscle. Cc, merged image of Ca and b. D and E, close apposition of the distal (D) and proximal (E) dendrites of a recorded cell to ChR2‐EYFP‐containing axons: Da and Ea, images of Neurobiotin‐NeutrAvidin‐Texas Red fluorescence; Db and Eb, immunostaining for ChR2‐EYFP; Dc and Ec, the merged images. These are magnified images of the small white boxes labelled D and E in B.

Figure 6. Close apposition of CS axon and PSD‐95 puncta on a MN dendrite where monosynaptic CS‐EPSCs were recorded

A, immunostaining of a ChR2‐EYFP‐containing CS axon. B, immunostaining of PSD‐95. C, merged image of A and B. D, fluorescence image of streptavidin‐Alexa 546 labelling of Neurobiotin. E, merged image (A, B and D). Each image is a single optical scan. The three spots are on a MN making contact with ChR2‐expressing axons expressing PSD‐95 (arrows).