The origins of spinal cord injury can be traced back more than 4500 years, to the oldest known trauma text: the ancient Egyptian Edwin Smith Papyrus. It was here that spinal cord injury was first described by clinicians, who, even then, described it as an injury ‘not to be treated’ - an attitude that was to last for millennia. Only after breakthroughs in imaging, medicine, and rehabilitation in the 20th century did we begin to develop a greater understanding of the mechanisms behind the debilitating, often paralysing injury, which affects between 250,000 to 500,000 people worldwide each year. We now understand that although the peripheral nervous system (PNS) can regenerate fully after damage, the central nervous system (CNS) cannot, and this is why CNS trauma carries such a poor prognosis. It is only by picking apart the reasons for this dichotomy that researchers are now able to develop therapies incorporating the rapidly expanding fields of gene therapy and stem cell research to stimulate regeneration, finally offering hope of treatment to an injury historically thought to be untreatable.
In order to better understand the inability of the CNS to heal, it is important to first consider the mechanisms behind successful regeneration in the PNS, as illustrated in Fig. 1. If a PNS neuron’s axon is damaged, it will regrow at a rate of around 1 mm a day in small neurons and 5 mm a day in larger ones  (note that if the cell body is damaged, regeneration is impossible). Immediately after axotomy (the severing of the axon), supporting glial cells such as Schwann cells recruit macrophages by releasing cytokines, and accompany them to the site of the injury to clear away debris - for example, by the phagocytosis of myelin. The distal stump (the end of the neuron not attached to the cell body) then undergoes Wallerian degeneration, a process that takes around 24 hours and results in the complete fragmentation of the axon. The endoplasmic reticulum degrades, mitochondria disintegrate and microtubules are depolymerised. The endoneurium (layer of tissue around the myelin sheath) remains intact, however, to provide a conduit to guide growing axons in a later stage of regeneration. This involves the proximal end (the end of the neuron attached to the cell body) sprouting axons with growth cones on their ends, which produce a protease that further digests debris on its journey to reinnervation through the endoneurial, or basal laminar tube, along which Schwann cells assemble in ordered longitudinal columns called ‘bands of Büngner’ to preserve the channel and help guide the axon to its target. Schwann cells and macrophages also upregulate neurotrophic factors such as nerve growth factor (NGF), while the PNS neuron itself upregulates regeneration-associated genes (RAGs). This intracellular encouragement, structural guidance, and expression of growth factors all create a favorable environment for regeneration and eventual reinnervation [2,3].
However, the general mechanism detailed above does not occur for a damaged CNS neuron. This is not entirely due to the CNS neurons’ inherent inability to regenerate – on the contrary, when placed into a permissive environment of a peripheral nerve graft, they are able to grow long distances , thus indicating that the explanation lies in the CNS environment. Indeed, research has shown that the difference between PNS-CNS regeneration is likely to be due to their different glial cell populations and their reactions to injury. The CNS’s glial cells are not Schwann cells, but oligodendrocytes and astrocytes.
Oligodendrocytes require axon signals to survive , and therefore undergo apoptosis (cell death) and fail to recruit macrophages to clear debris after injury, as Schwann cells do in the PNS. Instead, the CNS must rely on the action of microglia (the CNS analogue of macrophages in the peripheral immune system), which are slower than macrophages and may also fail to function. This causes the distal stump to degenerate slower than in a PNS neuron, resulting in the accumulation of inhibitory myelin debris. This is exacerbated by the formation of a glial scar (see Fig. 2), which axons cannot cross. It consists mainly of reactive astrocytes, which undergo heavy proliferation after CNS injury and form a dense network of gap junctions which act as a physical barrier to regrowth. Compelling proof of this inhibitory environment hypothesis comes from studies in which axotomy of the dorsal root ganglion neurons is followed by regeneration within its peripheral environment, but further growth is arrested at the PNS-CNS interface, also known as the dorsal root entry zone (DREZ). This is a phenomenon which has been known since the early twentieth century, and for which ultrastructural analysis has shown astrocytes in the glial scar to be responsible .
Although this barrier serves the beneficial purpose of preventing cytokine release from the injury site, causing further damage to surrounding tissue, it is the main physical inhibitor of CNS regeneration. Furthermore, the CNS glia do not produce neurotrophic factors, and instead produce factors that inhibit remyelination. Oligodendrocytes express myelin-associated inhibitors (MAI) such as Nogo-A, ephrin-B3 and Semaphorin-4D, and the astrocytes in the astroglial scar produce chondroitin sulfate proteoglycans (CSPGs), such as neurocan  (as well as providing a physical barrier in the form of the glial scar) (Fig. 3). CNS neurons themselves are also at fault, as it has been found that they upregulate RAGs less than do PNS neurons. Therefore, a mixture of a poor intracellular regenerative response and a hostile extracellular environment serve to inhibit axonal regrowth and restoration of function in the CNS.
It is by targeting these factors that combine to prevent regeneration that researchers have begun to make headway in developing treatments to promote it. For example, neurotrophin (NT) treatment showed some success in a study in which rats were injured on dorsal roots, and osmotic mini-pumps were immediately implanted to infuse neurotrophic factors such as NGF, GDNF and NT3 over the next week . This allowed the dorsal roots to overcome the transitional zone between the PNS and CNS that had previously been shown to be insurmountable, restoring some function. Yet delivery by a mini-pump is inconvenient for long-term treatment and does not stimulate the production of NTs in the spinal cord itself and can cause more parenchymal damage than other treatments such as viral-vector based gene therapy .
This type of therapy involves injecting gene-edited viruses which express NTs along the DREZ, leading to near-normal recovery of function . However, this also causes the sprouting of non-injured neurons, which, although it can enhance recovery, may cause chronic pain . Nonetheless, research has demonstrated strategically using a gradient of both neurotrophic factors and growth inhibitory factors can prevent hyperinnervation. Alongside using NGF to encourage growth across the DREZ, the use of increasing concentrations of the inhibitory Semaphorin-3A can prevent the growth of the neuron into unwanted areas, allowing for directed growth. NT treatment, either through protein delivery or gene therapy, has therefore been shown to have potential therapeutic applications.
Another option is to, instead of increasing growth factors, decrease factors inhibitory to growth. The genetic deletion of the aforementioned MAI protein Nogo-A promotes growth and enhances recovery after spinal cord injury  (although phenotypic expression varies among individuals) by targeting anti-Nogo-A monoclonal antibodies .
Besides manipulating the extracellular environment, the intracellular hindrances to regeneration can also be manipulated. For example, induced RAG over-expression in CNS neurons has been shown to promote sensory axon regeneration . Inhibitory intracellular signaling pathways can also be manipulated to promote regeneration. A key example is the RhoA/Rho-kinase pathway. When activated, the protein RhoA activates the protein kinase 2, which, in turn, regulates the dynamics of the cytoskeleton and results in the cessation of neurite growth. C3 transferase, an enzyme that deactivates RhoA, has been shown to promote axonal sprouting and motor function in mouse models , and kinase 2 knockout mice (mice in which the gene to produce kinase 2 is deleted) also showed functional recovery after spinal cord injury .
Perhaps the answer may not lie in changing the extrinsic or intrinsic, but rather replacing the damaged neuron altogether. Stem cell transplants have received much attention from the media and general population, and they do show some promise for morphologically replacing neurons or glial cells in the damaged CNS, sometimes even altered with therapeutic genes (for example, to over-produce neurotrophins ). This is due to their multipotency, meaning they can differentiate into the appropriate neuronal or glial subpopulation. When oligodendrocytes were replaced by embryonic stem cells in a rat with spinal cord injury, the rat’s locomotion improved . Yet this study is one of very few studies that have shown a functional use for stem cells in spinal cord injury, and it is unclear whether the stem cells actually differentiated into functional cells that contributed to structural reorganisation, or whether they simply secreted factors that aided the pre-existing cells to recover. In any case, despite the need for more rigorous studies as this new field develops, its preliminary success does show that future therapies may not need to be derived from the nervous system itself.
Thanks to our newly found understanding of multifactorial causes that facilitate PNS regeneration and hinder CNS regeneration, the latter, once an unachievable goal, is now a viable possibility. Targeting the glia that create physical barricades and inhibitory cues, as well as boosting the intrinsic regenerative capacity of the neuron itself, and even perhaps replacing them altogether with stem cells are all therapies that have shown promise in in vivo experimentation. Yet extrapolating these results to the inherently complex adult human CNS would be naive, and for each successful experiment, one must question its implications carefully. What role does neural plasticity – the compensation of undamaged systems for the function of damaged systems – play? To what extent does recovery of function correlate with anatomical and structural recovery? Perhaps most importantly – how applicable is the therapy for clinical use? The answers to these questions will only be found with a more rigorous evaluation of all these different therapies, facilitated by ongoing technological advances. In the future, we may finally be able to bring new life not only to the damaged spinal cord, but also to the lives that have been irrevocably impaired by it.
This author has not yet uploaded a bio.