The purpose of this study was to determine whether the catabolism of brevican is involved in mechanisms of neural plasticity in the hippocampus, and to accomplish this, synaptic input to the OML of the dentate gyrus was denervated by excitotoxic lesion in the lateral EC. Two days after lesion, synaptic input into the OML was significantly reduced and this was accompanied by an increase in the production of full length, intact brevican. At seven days, while brevican levels returned to baseline, a significant increase in the ADAMTS-derived, C-terminally truncated, brevican fragment was observed during this initial, sprouting and reinnervation period. This implies that there was an increase in ADAMTS activity in the OML during the highly plastic, regenerative phase. However at thirty days post-lesion, there was complete reinnervation of the OML on the ipsilateral side, as synaptic density, brevican and ADAMTS activity were not different from the contralateral side at this time point. These results indicate that the ADAMTSs and their substrate, brevican, that is abundant in the CNS, have a regulatory function in neural plasticity and support earlier data that had demonstrated important actions for the ADAMTSs in plasticity after seizure-induced hippocampal lesion .
Previous studies have examined the role of matrix-altering proteases in synaptic plasticity after CNS lesion. The expression of the matrix metalloproteinases (MMPs), MMP-9 and MMP-2, have been shown to be increased in various regions of the hippocampus after seizure-induced lesion [21, 23, 27, 28]. MMP-3 concentrations were elevated in the molecular layer of the dentate after traumatic brain injury . More specifically after ECL, administration of a non-selective MMP inhibitor was able to diminish sprouting and synaptogenesis in the dentate OML , suggesting a direct proteolytic role for the MMPs in this process. In adults, most MMP expression and activity is low and maintained throughout adulthood. After injury and during the recovery and regenerative phase, however, there is increased activity of MMPs derived from glia and neurons that is thought to facilitate axonal reinnervation, sprouting and/or synaptogenesis. Nonetheless, the mechanism(s) of action of the MMPs and the potential substrates on which they act to promote neural plasticity have yet to be determined in these models. More recently, the activity and expression of the PG-degrading, ADAMTSs have been shown to be elevated in the OML after kainite-induced lesion. In contrast to the absence of a defined substrate for the MMPs, a selective ADAMTS-derived, brevican fragment was localized to the OML after seizure-induced lesion in the rat . In the present study, a similar ADAMTS-derived brevican fragment was localized to the OML of the mouse after discrete denervation of the perforant path, suggesting a critical role in neural plasticity for the proteolytic turnover of brevican. Thus, the ability to localize and quantitate the ADAMTS specific, proteolytic product of brevican provides a means to indirectly estimate ADAMTS activity during times of neural plasticity and synaptogenesis.
The expression of brevican was shown previously to be up-regulated in the OML, the area of denervation after ECL in the rat , however, in contrast to the transient production observed here in the mouse, expression of immunoreactive brevican remained elevated compared to the non-lesioned side for almost 6 months after injury. Neurocan is a lectican that is expressed at high levels during early development but it was found to be up-regulated and synthesized by astrocytes in the OML after ECL. It was suggested that neurocan and possibly brevican may act to maintain the boundary of the denervated dentate after ECL , yet these complex molecules may be multifunctional during periods of neural plasticity. Each of the lecticans, exhibit a characteristic pattern of expression during development, with neurocan and versican V1 highly expressed in the brain of the fetus and neonate, whereas aggrecan, versican V2 and brevican increase expression during the period of synaptic stabilization in the adult and expression remains high throughout adulthood [13, 31]. Each of the lecticans is thought to bind to tenascin R and hyaluronic acid (with varying affinities) forming a multi-molecular lattice of ECM . It may be that proteolytic cleavage of the lectican loosens the lattice to promote neurite growth and synaptogenesis. Classically, the highly negatively charged CS chains on the lecticans inhibits neurite outgrowth, but proteolytic cleavage of the core protein may allow more movement of these chains and actually promote plasticity of neurons. This is a testable notion and preliminary data indicates that the ADAMTSs promote neurite outgrowth and other measures of neural plasticity in vitro (our unpublished observations).
The projection from the EC to the hippocampus is called the perforant path , and is thought to be involved in long-term potentiation and learning and memory [33–35]. The ECL model to study neural and synaptic plasticity denervates up to 80% of the input to the outer two-thirds of the molecular layer of the dentate gyrus , and due to sprouting of surviving fiber systems will reinnervate nearly fully. This model was developed more than thirty years ago in the rat , yet it has been only relatively recently that the technique was employed in mice to take advantage of transgenic models [38, 39]. Surprisingly, there are some differences in the projections from the EC to the hippocampal formation between rats and mice . For example, input to the dentate molecular layer from the contralateral EC is absent in the mouse, yet these contralateral fibers are responsible for much of the sprouting after ECL in the rat. In addition, the width of the inner molecular layer, that contains associational-commissural fibers, is thinner in the mouse than in the rat, causing an increase in the relative width occupied by the middle and OMLs, layers innervated mainly by EC fibers. In the mouse, the middle and OML occupy closer to four-fifths of the total, rather than two-thirds as seen in the rat. Moreover, three layers can be clearly differentiated in an untreated mouse, but not in a rat, using synaptophysin immunohistochemistry , and following ECL this laminar feature is lost (Fig. 5B). Synaptophysin immunochemistry has been one of the more common techniques, among many, to quantitate the loss and reinnervation of input into the ipsilateral molecular layer of the dentate gyrus after unilateral ECL . Optical density of the synaptophysin signal in the contralateral OML is measured using the sum (or average) of the gray levels of the pixels in this region, and this value is used as a "normal" value to the ipsilateral side. However, with this technique, the ipsilateral dentate usually shows only a 10–30% reduction in signal compared to the contralateral side at seven days after lesion, a time when sprouting has begun . Clearly this absolute value does not reflect the extent to which fibers are actually lost in the OML after ECL. Thus, we decided to develop a fresh tissue, needle punch dissection technique that could be limited to the dentate gyrus in the mouse. This way, biochemical assays could be conducted on the tissue to measure overall synaptophysin immunoreactivity by ELISA. Using this method, at two and seven days after lesion, there was greater than a 40% reduction in synaptophysin levels in the OML, a value which at least may closer reflect the absolute loss of fibers after ECL. The major disadvantage of this method is that the dissected tissue also includes the granule layer and the hilus of the dentate, regions where input is not lost after ECL. At the same time using this technique, tissue containing the lesion itself may be collected and assayed biochemically or processed for histochemistry to monitor the extent of the lesion in the EC.
The present results suggest that the lecticans and the proteases that cleave the lecticans play a regulatory role in neural plasticity after ECL. There are several potential mechanisms by which this substrate – protease pair may modulate neural plasticity, one of which was described above. Significant changes were observed in the abundance of the different isoforms of brevican including the expression of the C-terminally-truncated ADAMTS-derived brevican fragment during plastic events in the hippocampus. A genetic approach to study the individual lecticans and ADAMTSs could reveal the individual contributions for each of the molecules involved in neural plasticity after ECL, however, there is considerable redundancy among these molecules. For example, there appears to be a compensatory increase in the expression of neurocan in the brain of the brevican null mouse.  In addition, several of the ADAMTSs exhibit proteoglycanase activity, and of these, at least ADAMTS1, 4 and 9 appear to be expressed in the nervous system (unpublished observations). Whether there are compensatory changes in the expression of any of these molecules in the brain in ADAMTS null mice remains to be determined. Nonetheless, significant protective effects toward arthritic changes were demonstrated just recently in a single mutant, the ADAMTS5 null mouse . Should it turn out that these proteases play a significant role in plasticity related mechanisms in the nervous system, it will be interesting to examine how removing this regulatory action will impact development, sprouting after lesion, learning and memory and other plasticity related mechanisms in the adult.