Faculty Bibliography

We report that neuronal overexpression of the endogenous inhibitor of calpains, calpastatin (CAST), in a mouse model of human Alzheimer's disease (AD) beta-amyloidosis, the APP23 mouse, reduces beta-amyloid (Abeta) pathology and Abeta levels when comparing aged, double transgenic (tg) APP23/CAST with APP23 mice. Concurrent with Abeta plaque deposition, aged APP23/CAST mice show a decrease in the steady-state brain levels of the amyloid precursor protein (APP) and APP C-terminal fragments (CTFs) when compared with APP23 mice. This CAST-dependent decrease in APP metabolite levels was not observed in single tg CAST mice expressing endogenous APP or in younger, Abeta plaque predepositing APP23/CAST mice. We also determined that the CAST-mediated inhibition of calpain activity in the brain is greater in the CAST mice with Abeta pathology than in non-APP tg mice, as demonstrated by a decrease in calpain-mediated cytoskeleton protein cleavage. Moreover, aged APP23/CAST mice have reduced extracellular signal-regulated kinase 1/2 (ERK1/2) activity and tau phosphorylation when compared with APP23 mice. In summary, in vivo calpain inhibition mediated by CAST transgene expression reduces Abeta pathology in APP23 mice, with our findings further suggesting that APP metabolism is modified by CAST overexpression as the mice develop Abeta pathology. Our results indicate that the calpain system in neurons is more responsive to CAST inhibition under conditions of Abeta pathology, suggesting that in the disease state neurons may be more sensitive to the therapeutic use of calpain inhibitors.

Objective: Determine the role of V2 antibodies (Abs) in protecting vaccinees from HIV infection after immunization with ALVAC/HIV and clade E and B recombinant gp120 (RV144 vaccine regimen). ELISA reactivity to a V1V2-gp70 fusion protein and a cyclic V2 peptide was assessed using coded plasma specimens from RV144 participants. Results showed that Ab responses were divided into tertiles (low, medium and high). Vaccinees with a high level of Abs to V1V2-gp70 had a lower rate of HIV infection than did those with a low level of Abs. Multivariate logistic regression analysis showed an estimated 71% reduction in the infecton rate for vaccinees with high vs. vaccinees with low Ab responses to V1V2-gp70 (p=0.02). Similarly, compared to the entire placebo group that was HIV- at 26 weeks, the estimated Vaccine Efficacy (VE) was 3% (p=0.91) for subjects with a low V1V2 Ab response, while the estimated VE was 58% (p=0.02) for vaccinees with high V1V2 Abs. The Abs measured by this assay appear to be specific for V2 because no V1 Abs were detected, and epitope mapping data suggest that the epitope recognized by vaccinees' Abs target the midloop region of V2. There was no statistically significant correlation between infection rate and Abs reactive with a cyclic V2 peptide containing the entire V2 loop of the clade E A244 strain. In the primary case control analysis the Ab response to V1V2-gp70 was statistically significantly inversely correlated with infection. While secondary analyses suggest that other immunologic responses (including Abs to other epitopes and cellular responses) may also play a role in protection, the data most strongly support the hypothesis that V2 Abs are important for protection

Newly synthesized neurofilaments or protofilaments are incorporated into a highly stable stationary cytoskeleton network as they are transported along axons. Although the heavily phosphorylated carboxyl-terminal tail domains of the heavy and medium neurofilament (NF) subunits have been proposed to contribute to this process and particularly to stability of this structure, their function is still obscure. Here we show in NF-H/M tail deletion [NF-(H/M)(tailDelta)] mice that the deletion of both of these domains selectively lowers NF levels 3-6 fold along optic axons without altering either rates of subunit synthesis or the rate of slow axonal transport of NF. Pulse labeling studies carried out over 90 days revealed a significantly faster rate of disappearance of NF from the stationary NF network of optic axons in NF-(H/M)(tailDelta) mice. Faster NF disappearance was accompanied by elevated levels of NF-L proteolytic fragments in NF-(H/M)(tailDelta) axons. We conclude that NF-H and NF-M C-terminal domains do not normally regulate NF transport rates as previously proposed, but instead increase the proteolytic resistance of NF, thereby stabilizing the stationary neurofilament cytoskeleton along axons.

The neurofilament light subunit (NF-L) binds to myosin Va (Myo Va) in neurons but the sites of interaction and functional significance are not clear. We show by deletion analysis that motor domain of Myo Va binds to the NF-L rod domain that forms the NF backbone. Loss of NF-L and Myo Va binding from axons significantly reduces the axonal content of ER, and redistributes ER to the periphery of axon. Our data are consistent with a novel function for NFs as a scaffold in axons for maintaining the content and proper distribution of vesicular organelles, mediated in part by Myo Va. Based on observations that the Myo Va motor domain binds to intermediate filament (IF) proteins of several classes, Myo Va interactions with IFs may serve similar roles in organizing organelle topography in different cell types

White matter disorders can involve injury to myelin or axons but the respective contribution of each to clinical course is difficult to evaluate non-invasively. Here, to develop a paradigm for further investigations of axonal pathology by MRI, we compared two genetic mouse models exhibiting relatively selective axonal or myelin deficits using quantitative MRI relaxography of the transverse relaxation times (T2) in vivo and ultrastructural morphometry. In HM-DKO mice, which lack genes encoding the heavy (NF-H) and medium (NF-M) subunits of neurofilaments, neurofilament content of large myelinated axons of the central nervous system (CNS) is markedly reduced in the absence of changes in myelin thickness and volume. In shiverer mutant mice, which lack functional myelin basic protein, CNS myelin sheath formation is markedly reduced but neurofilament content is normal. We observed increases in T2 in nearly all white matter in shiverer mice compared to their wild type, while more subtle increases in T2 were observed in HM-DKO in the corpus callosum. White matter T2 was generally greater in shiverer mice than HM-DKO mice. Ultrastructural morphometry of the corpus callosum, which exhibited the greatest T2 differences, confirmed that total cross-sectional area occupied by axons was similar in the two mouse models and that the major ultrastructural differences, determined by morphometry, were an absence of myelin and larger unmyelinated axons in shiverer mice and absence of neurofilaments in HM-DKO mice. Our findings indicate that T2 is strongly influenced by myelination state and axonal volume, while neurofilament structure within the intra-axonal compartment has a lesser effect upon single compartment T2 estimates

The ultrastructural view of the axonal cytoskeleton as an extensively cross-linked network of neurofilaments (NFs) and other cytoskeletal polymers contrasts with the dynamic view suggested by axonal transport studies on cytoskeletal elements. Here we reconcile these perspectives by showing that neurons form a large NF network along axons which is unequivocally stationary, metabolically stable, and maintained by NFs and nonfilamentous subunit assemblies undergoing slow transport by intermittent rapid movements and pauses. In mouse primary cortical neurons transfected with EGFP-NFL, formation of this stationary NF network requires a critical level of NFs, which explains its absence in NF-poor developing neurons studied previously. Most NFs at proximal axon regions were in a stationary structure coexisting with a smaller pool of moving EGFP-NFL assemblies that were mainly nonfilamentous. Distally along the same axon, EGFP-labeled NFL was much less abundant, and we detected only short filaments moving bidirectionally by slow transport (rapid movements and pauses) as previously described. In living mice, >25% of radiolabeled newly synthesized NFs remained in optic axons after slowly transported NFs had exited. Retained NF remained fixed over several months in a nonuniform distribution and exhibited exceptionally slow turnover (t(1/2) >2.5 months), implying that, at steady state, >90% of NFs in mature optic axons comprise the stationary cytoskeleton and <10% are undergoing slow transport. These findings reconcile in vitro and in vivo axonal transport observations, showing that slowly transported NFs or subunit oligomers are precursors to a highly stable stationary cytoskeletal network that supports mature axons

The prospect of using cell-based interventions (CBIs) to treat neurological conditions raises several important ethical and policy questions. In this target article, we focus on issues related to the unique constellation of traits that characterize CBIs targeted at the central nervous system. In particular, there is at least a theoretical prospect that these cells will alter the recipients' cognition, mood, and behavior-brain functions that are central to our concept of the self. The potential for such changes, although perhaps remote, is cause for concern and careful ethical analysis. Both to enable better informed consent in the future and as an end in itself, we argue that early human trials of CBIs for neurological conditions must monitor subjects for changes in cognition, mood, and behavior; further, we recommend concrete steps for that monitoring. Such steps will help better characterize the potential risks and benefits of CBIs as they are tested and potentially used for treatment.

Neurofilament medium (NF-M) is essential for the acquisition of normal axonal caliber in response to a myelin-dependent 'outside-in' trigger for radial axonal growth. Removal of the tail domain and lysine-serine-proline (KSP) repeats of NF-M, but not neurofilament heavy, produced axons with impaired radial growth and reduced conduction velocities. These earlier findings supported myelin-dependent phosphorylation of NF-M KSP repeats as an essential component of axonal growth. As a direct test of whether phosphorylation of NF-M KSP repeats is the target for the myelin-derived signal, gene replacement has now been used to produce mice in which all serines of NF-M's KSP repeats have been replaced with phosphorylation-incompetent alanines. This substitution did not alter accumulation of the neurofilaments or their subunits. Axonal caliber and motor neuron conduction velocity of mice expressing KSP phospho-incompetent NF-M were also indistinguishable from wild-type mice. Thus, phosphorylation of NF-M KSP repeats is not an essential component for the acquisition of normal axonal caliber mediated by myelin-dependent outside-in signaling

Increased activity of calpains is implicated in synaptic dysfunction and neurodegeneration in Alzheimer's disease (AD). The molecular mechanisms responsible for increased calpain activity in AD are not known. Here, we demonstrate that disease progression is propelled by a marked depletion of the endogenous calpain inhibitor, calpastatin (CAST), from AD neurons, which is mediated by caspase-1, caspase-3, and calpains. Initial CAST depletion focally along dendrites coincides topographically with calpain II and ERK 1/2 activation, tau cleavage by caspase-3, and tau and neurofilament hyperphosphorylation. These same changes, together with cytoskeletal proteolysis and neuronal cell death, accompany CAST depletion after intrahippocampal kainic acid administration to mice, and are substantially reduced in mice overexpressing human CAST. Moreover, CAST reduction by shRNA in neuronal cells causes calpain-mediated death at levels of calcium-induced injury that are sublethal to cells normally expressing CAST. Our results strongly support a novel hypothesis that CAST depletion by multiple abnormally activated proteases accelerates calpain dysregulation in AD leading to cytoskeleton disruption and neurodegeneration. CAST mimetics may, therefore, be neuroprotective in AD