Exercise-induced Changes In Synaptic Plasticity In The Spinal Cord Following Peripheral Nerve Transection

[Background]Peripheral nerve injury produces profound changes in central nervous system (CNS) circuitry. More than 100,000 new traumatic peripheral nerve injuries occur in the US each year. Only about 10%of these patients ever recover full function (Frostick et al.,1998; Scholz et al.,2009). Following transection of a peripheral nerve, a withdrawal of nearly half of the synaptic inputs from the somata and proximal dendrites of the axotomized motoneurons is found (Alvarez et al.,2010; Blinzinger and Kreutzberg,1968). This synaptic stripping is not reversed, even if peripheral axon regeneration is successful and muscle fiber reinnervation occurs. The permanent withdrawal of the terminals of muscle spindle primary afferent neurons onto motoneurons is thought to play an important role in the functional loss of the stretch reflex (Alvarez et al.,2010) or the attenuation of its electrical analog, the H reflex (English et al.,2007), in reinnervated muscles. Synaptic stripping could be an important factor contributing to poor functional recovery found clinically. Treatments that influence synaptic stripping might contribute to improved functional outcomes following injuries to peripheral nerves. The contribution of the research in this proposal will be to investigate the mechanism of exercise as one such treatment.Exercise has a marked effect on synaptic stripping. If mouse sciatic nerves are cut and the animals are exposed to two weeks of moderate daily treadmill exercise, the anticipated reduction in synaptic inputs to the axotomized motoneurons is not noted (English et al.,2011). Based on this seminal finding, treatments such as exercise might be used to influence axotomy-induced synaptic plasticity in the spinal cord. However, the nature of the effect of exercise and the cellular basis for its effect on CNS plasticity, which will influence the variables that must be considered in designing an effective exercise protocol to influence synaptic stripping, are not yet clear. Our contribution here will be to begin to fill these gaps in our knowledge and by doing so increase the translational potential of exercise as a treatment strategy.Synaptic stripping is a hallmark of the early stages of several neurodegenerative diseases. Withdrawal of synaptic inputs from CNS neurons also is found at the early stages of several neurodegenerative diseases, including Alzheimer’s disease (Small,2008), Parkinson’s disease (Day et al., 2006), and Huntington’s disease (Moreno-Lopez et al.,2011) and from motoneurons in amyotrophic lateral sclerosis (ALS) (Zang et al.,2005), spinal muscular atrophy (Mentis et al.,2011), and other less common motoneuron diseases (Moreno-Lopez et al.,2011). The elimination of synapses in each of these disorders also utilizes cellular mechanisms that are similar to those found after peripheral nerve transection (Moreno-Lopez et al.,2011). Thus, treatments, such as exercise, which affect synaptic stripping following peripheral nerve injury, might be valuable in addressing similar effects related to a number of chronic degenerative diseases.In this project we will study four inter-related and fundamental aspects of the effect of exercise on synaptic stripping using peripheral nerve transection as a model system. Our expected findings will be basic, but we also anticipate that our results will be applicable to public health issues associated with peripheral nerve injuries and for the development of potential treatments for neurodegenerative diseases.[Methods]The Model System-The model system that we will use is transection injury to the mouse sciatic nerve. Within the first week following such an injury, nearly half of the synaptic inputs on the somata of axotomized motoneurons are withdrawn (Blinzinger and Kreutzberg,1968; Titmus and Faber,1990). If left untreated, this withdrawal will be permanent (Alvarez et al.,2010). This observation has been made for different nerves in several different species (Moreno-Lopez et al.,2011). We have shown that if mice are treated with moderate ex-ercise, in the form of daily treadmill training begun two days after peripheral nerve transection, the anticipated loss of synaptic inputs onto the axotomized motoneurons is not observed (English et al.,2011). We will use this model and the power of mouse genetics to dissect the mechanism underlying this effect of exercise.In Aim 1, we will determine whether this effect of exercise is to restore stripped synaptic inputs, rather than block the initial stripping, by delaying the start of exercise until synapses have already been stripped. Our use of the term "restore" has two implications and both will be investigated. We will evaluate whether exercise can restore any withdrawn synaptic contacts onto motoneurons using immunoreactivity to the generalized synaptic vesicle protein, SV2. However, restoration of synaptic inputs also implies that the relative composition of synaptic inputs from different sources will be the same as found in intact mice. Whether exercise affects synaptic inputs from varied sources equally is not known. We will identify the relative amounts of these synapses from different sources, exploiting the knowledge that synaptic terminals onto motoneurons from different sources contain different vesicular transporter proteins (Oliveira et al.,2003; Todd et al.,2003). Because clarifying this issue is so fundamental to understanding any cellular mechanism for the effect of exercise, we have chosen to study it first.In each of the next three Aims, we will evaluate how exercise impacts a comprehensive model of the cellular basis for axotomy-induced synaptic stripping developed by (Moreno-Lopez et al.,2011). Peripheral axotomy results in the production of nitric oxide (NO) in motoneurons, and associated reactive astrocytes, cells that do not normally produce it. This NO has two actions. Diffused NO acts directly on presynaptic terminals, activating a signaling pathway eventually leading to microtubule disruption and withdrawal. This NO also acts indirectly on the synaptic inputs. Within the axotomized motoneurons it blocks the secretion of BDNF at synaptic sites, which eventually leads to effects on cytoskeletal actin in the afferent neural terminals and mechanical destabilization of synaptic inputs. Several aspects of the direct action of NO are supported by experimental evidence. Blocking either NO production with neuronal nitric oxide synthase (nNOS) inhibitors (Sunico et al.,2005) or the effects of direct downstream signaling components (Sunico et al.,2010) blocks nerve crush-induced synaptic stripping. Inducing nNOS activity in motoneurons using viral constructs induces synaptic stripping in intact animals (Montero et al.,2010). Similar changes in nNOS expression have been observed in different mouse models of neurodegenerative diseases (Sasaki et al.,2001). The indirect, BDNF-dependent part of this model, although well grounded at a molecular level (Jovanovic et al.,1996), and based on hypotheses in the literature, that retrograde signals from motoneurons maintain synaptic inputs onto them (Ottem et al.,2010; Titmus and Faber,1990), has been subjected to less rigorous testing. The effects of exercise on different aspects of this part of the model will be evaluated in each of the last three Specific Aims of this project.Expression of BDNF is known to be elevated in the spinal cord with exercise (Gomez-Pinilla et al.,2002; Wood et al.,2011). This increased BDNF expression in post-synaptic motoneurons with exercise could overcome the proposed NO-induced block of its secretion and stabilize synapses. In Aim 2, we will use conditional BDNF knockout mice to test this prediction. One might predict that exercise could influence motoneuron secretion of BDNF by blocking the axotomy-induced increase in nNOS expression. We will evaluate nNOS expression in exercised mice in Aim 2. Any effect of exercise on stripped synaptic inputs onto motoneurons may involve the synthesis of cytoskeletal, growth promoting, and presynaptic proteins destined for the axon terminals in presynaptic afferent neurons. The effects of exercise are found rapidly so that we postulate that exercise stimulates the synthesis of these proteins locally, in the afferent axons that form synapses onto motoneurons. In Aim 3, we will investigate this postulate by examining the activation of translation-associated proteins by exercise and by evaluating the effects of exercise when axonal protein synthesis is compromised. Finally, in Aim 4, we will evaluate the role of motoneuron activity as the essential feature of these pre-and post-synaptic effects of exercise on synaptic stripping using optogenetic techniques. Synthesizing the results of the first three Aims, we will assay the extent and composition of synaptic inputs to axotomized motoneurons and markers of pre-and post-synaptic cellular aspects of the model when motoneuron activity is stimulated or blocked.Each of the Aims will use a similar assay to measure synaptic coverage on axotomized motoneurons. Motoneurons will be marked bilaterally by injecting a retrograde tracer, cholera toxin B (CTB), into the right and left gastrocnemius and tibialis anterior muscles (1μl each). Muscle injections, rather than nerve soaks or nerve injections will be used to avoid synaptic stripping that might be initiated by damage to peripheral axons associated with these methods. Three days later the right sciatic nerve will be transected in the mid-thigh and the cut segments of nerve will not be repaired to discourage reinnervation. The left sciatic nerve will remain intact. Training or motoneuron activation will follow. Serial transverse cryostat sections through the lumbar spinal cord of euthanized mice will be reacted with different antibodies to synaptic vesicle proteins to study synaptic inputs to retrogradely labeled axotomized motoneurons from different sources. Labeling of motoneurons with CTB will be amplified using an antibody to CTB.Thin (0.6μm) optical sections will be obtained from antibody-stained histological sections containing CTB labeled motoneurons using a laser scanning confocal microscope. Images will be selected for study if the labeled motoneurons contain a distinct nuclear shadow and are filled with tracer sufficiently to identify their cell boundaries. Synaptic terminals contacting the motoneuron somata and proximal dendrites containing the different marker proteins will be identified and the extent of their coverage on the motoneuron will be measured. For all treatment groups, the person performing the analyses of synaptic coverage always will be blinded to the experimental conditions associated with the neurons being studied while performing the measurements-i.e. the microscope slides will be coded.Twenty motoneurons on each side of each spinal cord will be studied. Six mice will be studied in each treatment group. Based on a power sample size estimate, a=0.05, power=0.8, and N=6, a real significant difference in synaptic coverage of as little as 4.6%can be detected. In preliminary experiments absolute differences in synaptic coverage with exercise ranged from 6%(for VGLUT1) to 29%(for SV2), themselves a reduction of more than half the coverage by the same type of terminal found in controls. Average synaptic coverage of motoneurons by terminals of different origins will be compared between different groups of mice using a one-way ANOVA. Post-hoc paired testing will be performed if the omnibus ANOVA is significant.[Results] (1) Determine whether exercise results in a restoration of synaptic inputs in the spinal cord following peripheral nerve transection. We have found exercise could restore synaptic inputs into areas where synapses had been stripped. By initiating exercise at different times after nerve transection, we also will begin to establish a therapeutic window for the effectiveness of exercise as a treatment.(2) Investigate the role of brain derived neurotrophic factor (BDNF), in exercise-induced changes in synaptic stripping following peripheral nerve transection. Axotomy-induced reduction in BDNF signaling is an important part of the proposed cellular mechanism of synaptic stripping. Exercise is known to increase the expression of BDNF in spinal motoneurons. We have found motoneuron BDNF is required for the effectiveness of exercise on axotomy-induced synaptic stripping.(3) Examine the requirement for axonal protein synthesis in exercise-induced changes in synaptic plasticity in the spinal cord following peripheral nerve transection. The effect of exercise on synaptic stripping may involve the synthesis of cytoskeletal, growth promoting, and synaptic proteins by afferent neurons connecting to motoneurons. We have found the effects of exercise on axotomy-induced synaptic stripping are dependent upon local protein synthesis in afferent axons. Local protein synthesis in axons requires mRNA transport by zip-code binding proteins. In adult mice heterozygous null for zip-code binding protein 1 (ZBP1 het), axonal protein synthesis is compromised and axon regeneration in injured peripheral nerves is impaired. Some important moleculars, downstream regulators of axonal protein synthesis in synaptic terminals surrounding axotomized motoneurons in exercised and untrained mice.[Conclusions]At the end of this project, we have established pre- and post-synaptic aspects of the mechanism by which exercise exerts its effect on an important form of synaptic plasticity in the spinal cord. Understanding more about that mechanism will result in an advancement of the translational potential of exercise as a treatment for peripheral nerve injuries that might also be applied to treatments of neurodegenerative diseases.