Reducing neuronal inhibition restores locomotion in paralysed mice

Spinal-cord injury can render intact neuronal circuits functionally dormant. Targeted reduction of neuronal inhibition in the injured region has now enabled reactivation of these circuits in mice, restoring basic locomotion.
Grégoire Courtine is at the Center for Neuroprosthetics and the Brain Mind Institute, School of Life Sciences, Swiss Federal Institute of Technology (EPFL), 1015 Lausanne, Switzerland, and in the Department of Neurosurgery, Lausanne University Hospital.

Search for this author in:

When we decide to walk, the brain broadcasts commands through parallel neuronal pathways that cascade to executive centres in the lumbar region of the spinal cord1. A spinal-cord injury (SCI) scatters this exquisitely organized communication system, leading to severe locomotor deficits or paralysis2. Most SCIs spare islands of intact neural tissues below the injury, which contain nerve fibres that remain connected to executive centres. But for unclear reasons, these anatomically intact connections remain functionally dormant. Writing in Cell, Chen et al.3 demonstrate that reducing the excitability of inhibitory neurons within the injured region of the spinal cord enables these dormant connections to relay commands from the brain, and promotes partial recovery of locomotion in mice that have sustained an SCI that causes complete paralysis.

It was long assumed that restoring movement after an SCI would involve precisely reconstituting the circuit connectivity that was in place before the injury. However, evidence of spontaneous circuit reorganization after SCI has challenged this view4. For example, consider injuries in which the spine is severed on either side in two staggered places. Although descending pathways from the brain are all severed at either the first or second lesion, new contacts can form between the projections from neurons in the brain that reach the second cut and local neurons that lie between the cuts. These contacts establish ‘detour’ circuits that relay sufficient information to executive centres to restore basic locomotion5,6 (Fig. 1). But until now, experimental procedures that trigger the formation and activation of these detour relays either have not been clinically relevant5 or have acted only transiently6.

Figure 1 | Restoring balance in neuronal relay circuits. a, In a staggered spinal-cord injury (SCI), two partial cuts in the spine completely interrupt direct neuronal projections from the brain to executive centres that produce locomotion. Excitatory neurons (yellow) and inhibitory neurons (blue) located between the injuries can form relay circuits that, when active, pass brain-derived brain commands past the SCI. But Chen et al.3 show that, in mice with an SCI, excessive activity in inhibitory neurons increases inhibition of the circuits to a level that renders them dormant. Information is not passed to executive centres (indicated by dashed circle), and so the animals are paralysed. b, The authors find that treatment with a small molecule called CLP290 reduces the activity of the inhibitory neurons between the cuts. This restores the balance between excitation and inhibition, enabling relay circuits to pass information from the brain to executive centres and so leading to the recovery of basic locomotor movements such as walking.

Chen and colleagues set out to find a permanent way to render relay circuits functional following injury. They injected mice that had sustained a staggered SCI with a panel of compounds known to modulate neuronal activity. One of these compounds, a small molecule called CLP290, restored movement in the injured mice.

CLP290 activates a protein called KCC27, which is a transporter of potassium and chloride ions, and is responsible for maintaining a functional level of the latter in neurons. Neurotransmitter molecules such as GABA or glycine open chloride channels on the surface of target neurons, allowing chloride ions (Cl) to flow into the cells down a concentration gradient, resulting in neuronal inhibition. By pumping out Cl, KCC2 can control the concentration gradient and so limit how strongly target neurons are inhibited by these neurotransmitters.

SCIs lead to a decrease in the levels of KCC2 in neurons below the injury8. Chen et al. examined the effects of increasing KCC2 levels in their mice, either in the lumbar spinal cord below the SCIs or in the relay circuits between the staggered injuries. To do this, they used various genetically engineered mice, which enabled them to modulate KCC2 expression in the three main types of neuron in the spinal cord — inhibitory, excitatory and motor neurons. The authors found that increasing KCC2 expression in inhibitory neurons between the staggered lesions could reproduce the effects of CLP290 treatment, whereas no other manipulation could.

It would be intuitive to expect that downregulation of KCC2 following injury would decrease the activity of neurons, by increasing the GABA- or glycine-mediated Cl influx. However, KCC2 downregulation actually leads to neuronal excitation if the lack of this transporter increases the concentration of Cl in the cells to such a level that GABA- or glycine-mediated opening of Cl channels causes an efflux, rather than an influx, of Cl8. This mechanism enables GABA and glycine to excite neurons during development. After birth, KCC2 upregulation in neurons reduces intracellular Cl concentrations, transforming excitation into inhibition.

Similarly, the authors found that injury-mediated downregulation of KCC2 in inhibitory neurons located between the staggered lesions increased the cells’ activity, and thus their ability to inhibit local relay circuits, rendering these circuits dormant. CLP290 restored the functionality of relay circuits by preventing downregulation of KCC2 and thus maintaining the balance between inhibition and excitation (Fig. 1). Moreover, the researchers showed that CLP290 did not affect the growth of new neuronal projections — recovery was triggered merely by restoring this balance in relay circuits.

Finally, Chen et al. demonstrated through two experiments that brain commands were being transmitted through the reactivated relays in CLP290-treated mice. First, electrically induced signals in the brain’s cortex were relayed to the motor neurons below the injury, resulting in the activation of hindlimb muscles. Second, neurons between the staggered lesions were more active in response to locomotor activity in treated than in non-treated mice.

The authors’ results are important. Together, they show that excessive neuron-mediated inhibition of relay circuits in the injured spinal cord is a key mechanism in preventing anatomically intact but functionally dormant neuronal pathways from contributing to movement after an SCI.

What are the clinical implications of the findings? Chen et al. used a small molecule that is well tolerated in mice, even at a high concentration6. However, the relevance of this treatment for humans with severe SCI remains unclear. The authors’ experimental model poorly mimics severe spinal-cord contusions commonly found in humans, in which nearly all the connections from putative relay neurons above the injured site are interrupted9. It therefore remains unclear whether Chen and colleagues’ results could be reproduced after a clinically relevant SCI. Indeed, because CLP290 does not promote the growth of new neural connections, this treatment would be expected to be effective only after an SCI that spares a substantial proportion of nerve fibres. In addition, SCI causes a cascade of detrimental changes, so effective treatments must target multiple facets of spinal-cord repair and recovery4, but CLP290 targets a single mechanism.

However, this type of orally administered pharmacological treatment is particularly attractive in combination with complementary strategies — notably, with interventions that promote the formation of relays in the spinal cord. For example, Chen and colleagues predict that CLP290 treatment will act synergistically with rehabilitative training, especially electrical spinal-cord stimulation, which promotes relay formation6. Alternatively, neural stem cells grafted into the injured spinal cord can enable reconstitution of relays across an SCI in monkeys10. Reducing neuron-mediated inhibition in the vicinity of the grafted relays could aid the functional integration of the relays into the host’s neuronal networks.

We are reaching an exciting time in SCI medicine, when multiple interventions that have strong synergistic potential are approaching clinical applications. There are now realistic opportunities to develop treatments that improve recovery after SCI in humans.

Nature 561, 317-318 (2018)

doi: 10.1038/d41586-018-06651-3


  1. 1.

    Arber, S. & Costa, R. M. Science 360, 1403–1404 (2018).

  2. 2.

    Fawcett, J. W. et al. Spinal Cord 45, 190–205 (2007).

  3. 3.

    Chen, B. et al. Cell 174, 521–535 (2018).

  4. 4.

    Sofroniew, M. V. Nature 557, 343–350 (2018).

  5. 5.

    Courtine, G. et al. Nature Med. 14, 69–74 (2008).

  6. 6.

    van den Brand, R. et al. Science 336, 1182–1185 (2012).

  7. 7.

    Gagnon, M. et al. Nature Med. 19, 1524–1528 (2013).

  8. 8.

    Boulenguez, P. et al. Nature Med. 16, 302–307 (2010).

  9. 9.

    Asboth, L. et al. Nature Neurosci. 21, 576–588 (2018).

  10. 10.

    Rosenzweig, E.S. et al. Nature Med. 24, 484–490 (2018).

Download references

Competing Financial Interests

G.C. holds various patents in relation to the present work. He is a founder and shareholder of GTX Medical, a company developing a therapeutic intervention for clinical applications directly in relation to the reported results.

Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.