Spinal cord injury: plasticity, regeneration and the challenge of translational drug development Armin Blesch 1 and Mark H. Tuszynski 1, 2 1 Department of Neurosciences-0626, University of California, San Diego, La Jolla, CA 92093, USA 2 Veterans Administration Medical Center, San Diego, CA 92165, USA Over the past three decades, multiple mechanisms limiting central nervous system regeneration have been identified. Here, we address plasticity arising from spared systems as a particularly important and often unrecognized mechanism that potentially contributes to functional recovery in studies of ‘regeneration’ after spinal cord injury. We then discuss complexities involved in translating findings from animal models to human clinical trials in spinal cord injury; current strategies might be too limited in scope to yield detect- able benefits in the complex and variable arena of human injury. Our animal models are imperfect, and the very variability that we attempt to control in the course of conducting rigorous research might, ironi- cally, limit our ability to identify the most promising therapies in the human arena. Therapeutic candidates are most likely to have a detectable effect in human trials if they elicit benefits in severe contusion and larger animal models and pass the test of independent replica- tion. Introduction Efforts to enhance central nervous system (CNS) regen- eration have been underway for 100 years [1,2]. The past 30 years, in particular, have contributed enormous pro- gress in identifying mechanisms that limit the plasticity and regeneration of the injured adult CNS; these have been the subject of extensive previous reviews [3,4]. To summarize, mechanisms underlying the limited capacity of the adult CNS to regenerate are multiple and are both intrinsic and extrinsic to the neuron. At sites of spinal cord injury, axons are confronted with an array of inhibitory molecules present in both the extracellular matrix (inhibi- tory chondroitin sulfate proteoglycans or CSPGs) [5,6] and in adult myelin (including nogo, myelin-associated glyco- protein, oilgodendrocyte myelin glycoprotein, netrin, semaphorin and ephrin) [7,8]. Reactive astrocytosis stabil- izes the injured zone and is essential in limiting the spread of injury [9], yet reactive astrocytes also form a ‘glial scar’ beyond which axons might have difficulty extending [5]. Finally, within the injured neuron, a set of injury-associ- ated signals, growth-associated proteins (for example GAP-43, CAP-23, b-tubulin and potentially thousands of genes [10,11]) and post-translational events are likely to be required to re-enter an active growth state to ‘sense’ extracellular matrix cues and cell adhesion molecules and to regenerate toward a target [12]. And, in contrast to events supporting successful regeneration after periph- eral nerve injury, neither growth factors nor permissive extracellular matrices are produced at sites of CNS injury to support and stimulate regeneration. Collectively, these mechanisms limit the regenerative potential of the injured adult CNS. These mechanisms are multiple and complex. A priori, it seems unlikely that manipulating a single mechanism would be capable of eliciting growth of sufficient magnitude and specificity to reinnervate an appropriate set of target neurons beyond a lesion site to generate functional recov- ery. Yet, several reports in the past 15 years put forth evidence of positive anatomical and functional outcomes after rodent spinal cord injury. Assuming that these reports are factually and interpretatively accurate, how can targeting of a single factor in such a complex process generate functional recovery? It is tempting to speculate that additional mechanisms might influence the outcome of these therapeutic studies. Recent work on spontaneous adaptive and maladaptive plasticity in the injured brain and spinal cord offer insight into the potential nature of these mechanisms. Spontaneous plasticity in the CNS After selective denervation of various brain regions, spontaneous sprouting of spared inputs has been identified in sensory, cerebellar, hippocampal and motor cortical regions [13–17]. Indeed, compensatory collateral sprouting of spared systems was identified two decades ago in the hippocampi of humans suffering from Alzheimer’s disease [15]. In the spinal cord, spontaneous sprouting of spared sensory inputs after dorsal root or peripheral nerve lesions was also identified several years ago [14,18] and has been directly associated in some models with the development of chronic pain, an indication of maladaptive plasticity [19]. More recent reports describe spontaneous and apparently adaptive plasticity in the rodent spinal cord after injury. The crossed phrenic nerve phenomenon has been described [20], in which spinal cord hemisection above the phrenic nucleus in mid-to-upper cervical spinal cord segments activates a latent, crossed bulbospinal pathway that can generate respiration when the contralateral phrenic nerve Review Corresponding author: Tuszynski, M.H. (mtuszynski@ucsd.edu). 0166-2236/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tins.2008.09.008 Available online 30 October 2008 41