Motor networks below the site of spinal cord injury (SCI) and their reconfiguration after loss of central inputs are poorly understood but remain of great interest in SCI research. Harley et al. (J Neurophysiol 113: 3610–3622, 2015) report a striking locomotor recovery paradigm in the leech Hirudo verbana with features that are functionally analogous to SCI. They propose that this well-established neurophysiological system could potentially be repurposed to provide a complementary model to investigate basic principles of homeostatic compensation relevant to SCI research.
- homeostatic plasticity
- motor network
- spinal cord injury
following spinal cord injury (SCI), it is clear that locomotor networks below the site of injury change as a result of the loss of descending input, but many of the cellular and network-level mechanisms of this compensation remain poorly understood. This injury-induced plasticity when left undirected is not always beneficial (Beauparlant et al. 2013). Consequently, a variety of approaches are being explored that assist with functional recovery during a critical period for compensation after SCI. Therapeutic interventions promoting favorable outcomes in spinal locomotor CPGs after injury include electrical, mechanical, and pharmacological stimulation (Ghosh and Pearse 2015; Rossignol et al. 2007; Tillakaratne et al. 2010). One promising approach involves the use of step training to practice cycles of ambulatory or locomotor-like activity through the use of treadmills or other assistance. Proprioceptive or other sensory feedback may play a role in stabilizing functional circuitry (Edgerton et al. 2008), but resolving the network mechanisms remains problematic. One fairly well-characterized injury-induced change following SCI involves network responses to the loss of descending serotonergic inputs, from the brain. SCI removes this modulation, and 5-HT transporters and receptors below the site of injury become dysregulated. Various avenues are being actively explored to influence these processes, from pharmacologically supplying receptor agonists or 5-HT precursors to grafting fetal serotonergic neurons to the spinal cord below the injury (Ghosh and Pearse 2015).
Despite these advances and interest in promoting positive outcomes after injury, many mechanistic questions remain unresolved and are difficult to study directly. Full descriptions of network architecture, changes to synaptic connections, single-cell physiology, and gene expression are lacking. One of the limitations in work exploring compensatory changes and their impacts following SCI is the lack of a good model system for basic research in recovery of decentralized locomotor networks. Although mammalian models are the most obvious in terms of their translational potential, the complexity of the mammalian spinal locomotor system and its poorly understood circuitry makes experimentation difficult (Guertin 2014).
Until now, the crustacean stomatogastric ganglion (STG) has been used as an effective preparation to study the effects of decentralization and the mechanisms of subsequent motor network compensation (e.g., Hamood et al. 2015; Zhang and Golowasch 2011). This CPG network comprises unambiguously identifiable cell types with known connectivity and depends on modulatory inputs from higher order centers. The STG has successfully been employed to study cellular and network level compensatory changes with the use of molecular and electrophysiological methods. Studies have examined ionic conductances, network output, neuromodulator sensitivity, and single-cell quantification of mRNA. Although this system has its advantages, it also has limitations in its applicability to recover locomotor behavior. Compensation observed in this network occurs spontaneously without external feedback, and the rhythmic motor patterns it generates do not control locomotion in the animal.
Recent work in the leech system by Harley et al. (2015) describes a novel invertebrate experimental model for understanding recovery of locomotor function following decentralization and could potentially serve as a model to investigate basic principles of homeostatic plasticity in this context. Its long history as a neurophysiological preparation, combined with a recently published transcriptome analysis (Hibsh et al. 2015), leave this model system well-positioned to rapidly gain traction to experimentally pose detailed questions about the molecular and electrophysiological mechanisms involved in this robust recovery. Although the leech system contains a larger number of neurons than the STG, these networks are still far more tractable than the vertebrate spinal cord. Additionally, using an intact and otherwise healthy animal could allow direct connections to be made between mechanisms of neural plasticity and functional behavioral recovery.
Descending fibers from the brain of leeches normally provide input to initiate, modulate, and coordinate locomotion such as crawling or swimming, and specific roles for identifiable cells have been characterized (Kristan et al. 2005; Puhl et al. 2012). Below the brain, the nervous system of Hirudo verbana is structured around a chain of 21 segmental ganglia and a compound terminal ganglion. Oscillatory networks are present in each segmental ganglion, and each innervates the musculature to mediate motor behaviors in that segment (Puhl and Mesce 2010). In their study, Harley et al. (2015) surgically severed connectives between the brain and segmental ganglia, confirming previous findings that the acute loss of these inputs abolishes crawling behavior and promotes swimming, which is normally inhibited by the brain. Spontaneous crawling was not observed 1 day posttransection, and attempts to evoke crawling were ineffective. Bouts of swimming are observed in uninjured leeches, but the duration of these bouts was observed to increase ∼10-fold after surgery.
The novelty of the recent study comes in the longitudinal approach to tracking the behavior of leeches as they recovered from this perturbation. Interestingly, although the severed connections do not regrow, both crawling and swimming behaviors recover over a period of days to weeks. To ensure that regrowth of neural projections that normally coordinate crawling was not responsible, the entire region containing the somata and integrative regions of projection neurons were removed. Crawling consists of cycles of elongation and contraction, normally coordinated in an anterior-to-posterior “wave” of activity, which the investigators measured by video analysis. Although crawling behavior reemerged, it was not immediately apparent which anatomical region was responsible for the initiation and intersegmental coordination of this movement. Some animals that had recovered crawling following the first surgery were used to test this further. By performing a second transection at the midpoint along the ventral nerve cord in the same animal, the study found that the anterior section retained the ability to crawl, whereas it was lost in the posterior half. Once again, a similar process of dysregulation followed by recovery occurred in the posterior half of the animal. Essentially, these behaviors reemerge in two separate sections of the same animal. The authors propose what they call the “lead ganglion” hypothesis, interpreting that the most anterior ganglion behind the lesion appears to be reconfigured to take command of these behaviors. This would implicate networks with only a few hundred paired neurons in the homeostatic plasticity underlying this compensation. This is illustrated by the summary schematic in Fig. 1.
With two separate control centers for body movement, mechanistic differences between swimming and crawling networks become important. Much of the coordination of swimming and crawling can be explained by direct communication between oscillatory elements in each segmental ganglion. Coordination during swimming, but not crawling, is known to include an indirect pathway for coordination between ganglia via proprioceptive feedback from the periphery. Consistent with this knowledge, swimming animals seem to have some degree of coordination between the two halves of their body governed by different systems. However, when these same animals are crawling, the two halves function largely independently from one another, implying that if sensory feedback does play a role it is not sufficient to coordinate movement across the mid-body transection. This does not entirely rule out a role for sensory feedback, which may help to explain the faster recovery observed in the posterior half. After the second surgery, the anterior half of the animal continues to engage in crawling or crawl-like activity, which would be expected to induce more sensory stimulation in the posterior half. This explanation remains to be tested, but the clear contrast between two distinct behaviors in the same animal may provide a useful platform to explore the role of sensory feedback in the reconfiguration of injured motor networks.
Finally, the study notes substantial variability across animals in recovery rate. This variability may prove informative, particularly if rates or degrees of recovery can be linked to other physiological factors. Understanding this recovery process based on clearly elucidated mechanisms and directing the recovery process with specific interventions remain as future challenges.
Whereas the leech preparation obviously has differences from the vertebrate spinal locomotor networks, there are functional similarities as well. Descending projections from higher order centers initiate and coordinate movement in both systems. In the leech, this involves neurons such as the R3b-1 command neuron, which activates crawling, analogous to the role of reticulospinal neurons in vertebrate locomotion (Puhl et al. 2012). Neuromodulation with biogenic amines influences and shapes the motor output in both the segmental ganglia of the leech and the vertebrate spinal cord (Puhl and Mesce 2010). Harley et al. (2015) argue that these functionally analogous features and fundamental principles underlying compensation could allow this experimentally tractable system to be repurposed to provide a complementary approach to understanding general principles of homeostatic plasticity and the mechanisms underlying compensation in decentralized motor networks.
B. J. Lane is supported by National Institute of General Medical Sciences Molecular Biology Training Grant T32 GM008396.
No conflicts of interest, financial or otherwise, are declared by the author.
B.J.L. prepared figure; B.J.L. drafted manuscript; B.J.L. edited and revised manuscript; B.J.L. approved final version of manuscript.
I thank my advisor, David J. Schulz, and the Schulz laboratory. I also thank Emily C. Dabe for comments on an earlier version of the manuscript.
- Copyright © 2016 the American Physiological Society