147 Synaptic loss in the MS spinal cord Figure 2. Immunostaining for synaptic proteins, and synaptic bouton area quantification. (A–C) Myelin basic protein (MBP) immunostained sections from control (A) and multiple sclerosis (MS) cases (B, C) were reviewed. MS anterior horns were classified as nonlesional (B) or lesional (C) based on the presence of myelin on MBP sections. (D–I) Synaptophysin intensity was quantified in control (D, G), nonlesional (E, H), and lesional gray matter sections (F, I). (J, K, M–R) The area of synaptic boutons (white arrowheads in J, arrows in M and N) was assessed on highmagnification sections using (1) synapsin-1 in control (J) and MS sections (K), and (2) synaptophysin in control (M, O, Q) and MS sections (N, P, R) . (M–R) The area of synaptic boutons around an average of 5 neurons was measured by outlining their boundaries, as shown in red. (L) Negative controls were run during synapsin-1 immunostaining. (S–X) Negative (S, U, W) and positive (T, V, X) controls were run for synaptophysin immunostaining. Scale bars = 1mm in A–F, 50μm in G–L, 10μm in J–R, 500μm in S and T, 100μm in U and V, and 50μm in W and X. Black rectangles in S and T are magnified in U and V; green rectangles in S and T are magnified in W and X. relationship between synaptic bouton area and the number of neurons was significant (p <0.001) in controls only. The relationship between synaptic bouton area and (1) GM area (p <0.001) and (2) CSA (p =0.005) was only significant in controls. Discussion Axonal loss has been considered the main substrate of chronic disability in pwMS8, with approximately 60% of long spinal cord axons being lost approximately 30years after diagnosis10. Although the figure of 60% indicates extensive axonal loss, comparison with other spinal cord pathologies raises important questions about additional factors involved in the progressive disability accrual of pwMS. Animal models of traumatic spinal cord injury suggest preservation of as little as 8% of axons may enable useful limb movements13, and the case of a 38-year-old man who regained lower limb control sufficient to walk (with assistance) after a near-complete traumatic spinal cord transection14 indicates that a much smaller number of preserved long tract axons than the 40% detected in pwMS may suffice for meaningful leg function. The pwMS from which our samples were obtained, however, were invariably wheelchair bound due to loss of lower limb function. Thus, loss of long motor tracts is unlikely to fully explain the degree of motor dysfunction, particularly of the lower limbs, in pwMS. Given that synaptic loss was detectable throughout the cord, including the cervical level, which is associated with upper limb control, it is unlikely that this loss is solely due to a lack of sensory input and/or motor output secondary to immobility. Synaptic loss in MS has previously been described in the cerebral GM, including the hippocampus, and the cerebellar dentate nucleus11,12,15. Although our results are not inconsistent with a possible effect of demyelination on synaptic survival, the key finding is of substantial synaptic loss throughout the spinal cord, corroborating recent work describing extensive synaptic spine density reduction in both myelinated and demyelinated neocortex16. We recently confirmed that CSA is not associated with axonal density10. Against this backdrop, the correlation between several measures of spinal cord area and synaptic bouton areas in control sections highlights the nature of the spinal cord as a self-regulatory network17, rather than merely a conduit for long axonal tract fibers. Neuronal damage and loss in pwMS can be extensive11. In a recent stereological study, we reported that after nearly 30years disease duration, pwMS will have lost almost 40% of their neocortical neurons irrespective of the presence of demyelination18. The degree of neuronal loss in the current study (47%) is higher than previously reported19, perhaps reflecting
RkJQdWJsaXNoZXIy MTk4NDMw