Direct evidence for decreased presynaptic inhibition evoked by PBSt group I muscle afferents after chronic SCI and recovery with step-training in rats

Abstract

Key points: Presynaptic inhibition is modulated by supraspinal centres and primary afferents in order to filter sensory information, adjust spinal reflex excitability, and ensure smooth movement. After spinal cord injury (SCI), the supraspinal control of primary afferent depolarization (PAD) interneurons is disengaged, suggesting an increased role for sensory afferents. While increased H-reflex excitability in spastic individuals indicates a possible decrease in presynaptic inhibition, it remains unclear whether a decrease in sensory-evoked PAD contributes to this effect. We investigated whether the PAD evoked by hindlimb afferents contributes to the change in presynaptic inhibition of the H-reflex in a decerebrated rat preparation. We found that chronic SCI decreases presynaptic inhibition of the plantar H-reflex through a reduction in PAD evoked by posterior biceps-semitendinosus (PBSt) muscle group I afferents. We further found that step-training restored presynaptic inhibition of the plantar H-reflex evoked by PBSt, suggesting the presence of activity-dependent plasticity of PAD pathways activated by flexor muscle group I afferents.

Abstract: Spinal cord injury (SCI) results in the disruption of supraspinal control of spinal networks and an increase in the relative influence of afferent feedback to sublesional neural networks, both of which contribute to enhancing spinal reflex excitability. Hyperreflexia occurs in ∼75% of individuals with a chronic SCI and critically hinders functional recovery and quality of life. It is suggested that it results from an increase in motoneuronal excitability and a decrease in presynaptic and postsynaptic inhibitory mechanisms. In contrast, locomotor training decreases hyperreflexia by restoring presynaptic inhibition. Primary afferent depolarization (PAD) is a powerful presynaptic inhibitory mechanism that selectively gates primary afferent transmission to spinal neurons to adjust reflex excitability and ensure smooth movement. However, the effect of chronic SCI and step-training on the reorganization of presynaptic inhibition evoked by hindlimb afferents, and the contribution of PAD has never been demonstrated. The objective of this study is to directly measure changes in presynaptic inhibition through dorsal root potentials (DRPs) and its association with plantar H-reflex inhibition. We provide direct evidence that H-reflex hyperexcitability is associated with a decrease in transmission of PAD pathways activated by posterior biceps-semitendinosus (PBSt) afferents after chronic SCI. More precisely, we illustrate that the pattern of inhibition evoked by PBSt group I muscle afferents onto both L4-DRPs and plantar H-reflexes evoked by the distal tibial nerve is impaired after chronic SCI. These changes are not observed in step-trained animals, suggesting a role for activity-dependent plasticity to regulate PAD pathways activated by flexor muscle group I afferents.

Keywords: H-reflex; dorsal root potential; presynaptic inhibition; spinal cord injury; step-training.

Figure 1.. Experimental set-up.

A, The PBSt nerve, tibial nerve and cut dorsal rootlet were mounted on bipolar hook electrodes. The cord dorsum potential (CDP) was recorded using a monopolar silver ball electrode placed at the L4-L5 dorsal root entry to determine the activation threshold for the most excitable afferent fiber. Dorsal root potentials (DRPs) were recorded in response to a stimulation to the tibial or PBSt nerves. H-reflexes were evoked by a stimulation to the tibial nerve and obtained using bipolar wire electrodes inserted under the plantar surface of the foot. In the conditioning protocol, PBSt stimulation was used as the conditioning stimulus and preceded the tibial stimulation. B, Typical DRP recording displaying a long-lasting upward deflection. By convention, DRP negativity is represented upwards and CDP negativity downwards. The latency-to-peak was determined as the time between the onset of the afferent volley (CDP) and the maximal DRP amplitude from baseline to peak.

Figure 2.. Chronic SCI and step-training specifically modulates DRPs evoked by PBSt group I muscle afferents.

A, Example of DRPs evoked by the tibial (top) or PBSt (bottom) nerves at different stimulation intensities (1.2, 1.5, 2, or 5T) in an Acute SCI, Chronic SCI and SCI + Step-training animal. B, Example of the individual sigmoid function that was fitted to input-output relationships for each animal. Overall, all animals displayed curves that tightly fitted a sigmoid function (P < 0.001) with R² ranging from 0.93-0.98 after acute SCI, 0.94-0.99 after chronic SCI and 0.90-0.99 after SCI + Step-training. C-E, Chronic SCI did not significantly affect the amplitude of Tib-DRP_max (P = 0.538) and PBst-DRP_max (P = 0.371) whether the animals were step-trained or not (C). The stimulation intensity to reach 50% of DRP_max (s50) when evoked by the tibial (P = 0.004) or PBSt nerve (P < 0.001) was different across groups (D). Chronic SCI significantly increased s50 as compared to acute SCI (tibial, P = 0.020; PBSt, P < 0.001). Step-trained animals had similar s50 than Chronic SCIs when DRPs were evoked by the tibial nerve (P = 0.225 vs. Chronic SCI) but lower s50 when evoked by PBSt (p<0.001) with values similar to acute SCI (P = 0.182). The slope of PBST-DRPs input-output relationship was different across groups only when evoked by PBSt (P < 0.001), not tibial nerve (P = 0.067) (E). Chronic SCI decreased the slope (P < 0.001), while step-training had no further effect (P = 0.100 vs. Chronic SCI). Results are expressed as mean ± SD, one-way ANOVA followed by Holm-Sidak post hoc test. *P < 0.05; ** P < 0.01; ***P < 0.001. Acute SCI, n = 8; Chronic SCI, n = 7; SCI + Step-training, n = 6. CDP, cord dorsum potential; F-G, Sigmoid function resulting from group averages is illustrated with 95% confidence interval for Tib- and PBSt-DRPs.

Figure 3.. Disruption in the transmission of group I PBSt afferents in PAD pathways after chronic SCI is improved by step-training.

A, Representative traces of DRPs evoked by a stimulation to PBSt nerve at group I afferent strength (4p, 1.5T, 250Hz) in an Acute SCI, a Chronic SCI and a SCI + Step-training animal. For reference purpose, the dotted line illustrates DRP_max in the same animal. B, The amplitude of DRPs evoked by PBSt group I afferents was different across groups (P < 0.001). Chronic SCI decreased DRP amplitude (P = 0.002, vs. Acute SCI). This decrease was not observed after step-training with values similar to acute SCIs (P = 0.058) but significantly larger than chronic SCI animals (P < 0.001). Results are expressed as mean ± SD, one-way ANOVA followed by Holm-Sidak post hoc test ** P < 0.01; *** P < 0.001. Acute SCI, n = 8; Chronic SCI, n = 7; SCI + Step-training, n = 6.

Figure 4.. Step-training improves the inhibition induced by PBSt group I afferents on L4-DRPs evoked by the tibial nerve after chronic SCI.

A, Conditioning-test protocol used to estimate the level of inhibition evoked by a conditioning stimulation to PBSt group I afferents (4p, 250Hz, 1.5T) on Tib-DRPs (1p, 0.2 Hz, 1.2-1.8T). B, DRPs evoked by the tibial nerve when preceded (black) or not (gray) by a conditioning stimulation to PBSt (C-T: 50 ms) in an Acute SCI, a Chronic SCI and a SCI + Step-training animal. The inhibitory effect of the conditioning stimulation is not present after chronic SCI. C, Overall, chronic SCI significantly decreases the strength of inhibition at C-T intervals ranging from 10 to 60ms as compared to acute SCI (P < 0.05, gray stars). This decrease was not observed in step-trained animals with values similar to acute SCIs and inhibition significantly larger than chronic SCI animals at intervals ranging from 10 to 50ms (P < 0.05, green stars). Two-way RM ANOVA followed by Holm-Sidak post hoc test, *p<0.05; **p<0.01; ***p<0.001. Acute SCI, n = 8; Chronic SCI, n = 7; SCI + Step-training, n = 6. C-T: Conditioning-Test.

Figure 5.. H-reflex excitability is impaired following chronic SCI and normalized by step-training.

A, Example of M-wave and H-reflex recruitment curves in an Acute SCI, a Chronic SCI and a SCI + Step-training animal. Sigmoid functions were individually fitted to the recruitment curves. Overall, all animals displayed curves that tightly fitted a sigmoid function (P < 0.001) with R² ranging from 0.93-0.99 in acute SCIs, 0.90-0.99 in chronic SCIs and 0.91-0.99 in SCI + Step-training animals. B-C, The sigmoid function resulting from group averages is illustrated with 95% confidence interval for both the M-wave (B) and the H-reflex (C). D-F, Chronic SCI did not alter the H_max and M_max amplitude (D) (P = 0.454 and P = 0.595 respectively) and the stimulation intensity to reach 50% of the maximal amplitude (E) (P = 0.609 and P = 0.405 respectively) whether the animals were step-trained or not. However, the slope of the H-reflex sigmoid function was significantly different across groups (F) (P = 0.004). Chronic SCI increased the steepness of the slope of the recruitment curve for the H-reflex as compared to acute SCI (P = 0.002) and step-trained animals (P = 0.015), while step-training shifted the slope toward values similar to acute SCIs (P = 0.465). The slope of the M-wave recruitment curve was similar across groups (P = 0.113). One-way ANOVA followed by Holm-Sidak post hoc test. * P < 0.05; ** P < 0.01. Acute SCI, n = 8; Chronic SCI, n = 7; SCI + Step-training, n = 6. MT, motor threshold.

Figure 6.. Step-training improves H-reflex inhibition after chronic SCI.

A, Representative recordings of H-reflexes evoked by a stimulation to the tibial nerve with or without a conditioning stimulation to PBSt in an Acute SCI, a Chronic SCI and SCI + Step-training animal at C-T intervals of 10ms and 50ms. B, There was a significant difference in the strength of inhibition of the H-reflex between groups (F_2,171=32.056; p<0.001), C-T interval (F_9,171=29.865; P < 0.001), and an interaction between groups and C-T intervals (F_18,171=2.506; P = 0.001). For clarity purpose, only significance between groups, but not intragroup is illustrated. Chronic SCI significantly decreased the depression of the H-reflex as compared to acute SCI at intervals ranging from 10-150ms (P < 0.01). In addition, the strength of inhibition in step-trained animals was larger than in non-trained animals after chronic SCI (P < 0.05, green stars) and similar to acute SCIs at similar C-T intervals. Two-way RM ANOVA followed by Holm-Sidak post hoc test. C-T, Conditioning-Test. C, The amplitude of the unconditioned H-reflex (% of M_ma) was similar between experimental groups (P = 0.552). One-way ANOVA. D, There was a significant decrease in postsynaptic (C-T intervals 10-20ms; F_2,18 = 15.567, P < 0.001) and presynaptic inhibition (C-T intervals 50-150ms; F_2,18 = 16.597, P < 0.001). Chronic SCI decreased the strength of postsynaptic (P < 0.001, gray stars) and presynaptic inhibition (P < 0.001, gray stars), unless the animals were step-trained (respectively P = 0.002 and P < 0.001, green stars). One-way ANOVA followed by Holm-Sidak post hoc test. *P < 0.05; ** P < 0.01; *** P < 0.001. Acute SCI, n = 8; Chronic SCI, n = 7; SCI + Step-training, n = 6.

Figure 7.. Chronic SCI disrupts the relationship between DRP amplitude and H-reflex inhibition unless the animal is step-trained.

The linear relationship between the amplitude of the H-reflex and DRP conditioned by a stimulation to PBSt after acute SCI (R²: 0.244, P < 0.001) is not present after chronic SCI (R²: 0.0150, P = 0.301), unless the animals were step-trained (R²: 0.264, P < 0.001). Regression Analysis.