Function of COX-2 and prostaglandins in neurological disease.

Induction of COX-2 expression and enzymatic activity promotes neuronal injury in a number of models of neurological disease. Inhibition of COX-2 activity, either genetically or pharmacologically, has been shown to be neuroprotective in rodent models of stroke, Parkinson's disease, and amyotrophic lateral sclerosis. Inhibition of COX activity with nonsteroidal anti-inflammatory drugs (NSAIDs) reduces inflammation and amyloid accumulation in murine transgenic models of Familial Alzheimer's disease, and the use of NSAIDs decreases the risk of developing Alzheimer's disease in healthy aging populations. COX-mediated neuronal injury is presumed be due to downstream effects of one or more prostaglandin products including PGE2, PGD2, PGF2alpha, PGI2 (prostacylin) and TXA2 (thromboxane) that effect cellular changes through activation of specific prostaglandin receptor subtypes and second messenger systems. In this proceeding, we review recent data demonstrating effects of prostaglandin signaling on neuronal viability that are paradoxically protective, when taken in the context that COX-2 induces neuronal injury in the setting of excitotoxicity. Conversely, in the context of an inflammatory stimulus, the EP2 receptor enhances neuronal injury. These findings argue for an additional level of complexity in the prostaglandin response in neurological disease.

Additivity and potentiation of IGF-I and GDNF in the complete rescue of postnatal motor neurons.

BACKGROUND: Both growth and survival of motor neurons may depend on multiple neurotrophic factors. Individually, insulin-like growth factor I (IGF-I) and glial cell line-derived neurotrophic factor (GDNF) are potent neurotrophic/survival factors for postnatal motor neurons. METHODS: We used an organotypic spinal cord model of glutamatergic degeneration in ALS to investigate whether IGF-I and GDNF interact to enhance motor neuron survival, their trophic effect on choline acetyltransferase (ChAT) activity, and their effect on neurite outgrowth. RESULTS: We show that the combination of IGF-I and GDNF at active doses (1) is additively neuroprotective, (2) completely rescues rat motor neurons from chronic glutamate-mediated toxicity, and (3) additively upregulates motor neuron ChAT activity. Further, IGF-I, which by itself does not promote neurite outgrowth in this model, potentiates the neurite promoting action of GDNF. CONCLUSION: The results predict that IGF-I combined with GDNF may provide a better therapy for the treatment of motor neuron disorders such as amyotrophic lateral sclerosis (ALS) and spinal muscular atrophy.

Delayed application of IGF-I and GDNF can rescue already injured postnatal motor neurons.

IGF-I, GDNF, and other neurotrophic factors, when applied at the time of injury, can protect postnatal motor neurons from slow glutamate injury in organotypic spinal cord. However, in human spinal cord diseases, motor neuron injury is already established when treatment could begin. We tested whether neurotrophic factors can protect already-injured motor neurons, and whether combinations of factors can further lengthen the therapeutic time window. Our data show that during a 7--8 week process of slow neurodegeneration either IGF-I or GDNF treatment, though delayed up to 4 weeks, still allowed substantial rescue of already injured motor neurons. However, the combination of both factors additively provided better neuroprotection than either factor alone, even after a 4-week delay. This proof of principle is relevant to the potential of IGF-I and GDNF as therapy for acquired disorders affecting motor neurons.

Preclinical testing of neuroprotective neurotrophic factors in a model of chronic motor neuron degeneration.

Many neurotrophic factors have been shown to enhance survival of embryonic motor neurons or affect their response to injury. Few studies have investigated the potential effects of neurotrophic factors on more mature motor neurons that might be relevant for neurodegenerative diseases. Using organotypic spinal cord cultures from postnatal rats, we have demonstrated that insulin-like growth factor-I (IGF-I) and glial-derived neurotrophic factor (GDNF) significantly increase choline acetyltransferase (ChAT) activity, but brain-derived neurotrophic factor (BDNF), neurotrophin-4 (NT-4/5), and neurotrophin-3 (NT-3) do not. Surprisingly, ciliary neurotrophic factor (CNTF) actually reduces ChAT activity compared to age-matched control cultures. Neurotrophic factors have also been shown to alter the sensitivity of some neurons to glutamate neurotoxicity, a postulated mechanism of injury in the neurodegenerative disease, amyotrophic lateral sclerosis (ALS). Incubation of organotypic spinal cord cultures in the presence of the glutamate transport inhibitor threo-hydroxyaspartate (THA) reproducibly causes death of motor neurons which is glutamate-mediated. In this model of motor neuron degeneration, IGF-I, GDNF, and NT-4/5 are potently neuroprotective, but BDNF, CNTF, and NT-3 are not. The organotypic glutamate toxicity model appears to be the best preclinical predictor to date of success in human clinical trials in ALS.

Pigment epithelium-derived factor (PEDF) protects motor neurons from chronic glutamate-mediated neurodegeneration.

Although pigment epithelium-derived factor (PEDF) is a neurotrophic factor that may aid the development, differentiation, and survival of adjacent neural retinae, the wider distribution of PEDF mRNA in the central nervous system suggested to us that this factor could have pleiotropic neurotrophic and neuroprotective effects on nonretinal neurons. We examined the distribution of PEDF mRNA and its transcript in the spinal cord. By immunohistochemistry and western blot analysis using an antihuman PEDF antiserum of known specificity, we found that PEDF protein is present in spinal cord, cerebrospinal fluid, and skeletal muscle and that its mRNA appears concentrated in motor neurons of the human spinal cord. These observations indicate that PEDF could have potential autocrine and paracrine effects on motor neurons, as well as being target-derived. We analyzed the pharmacologic utility of PEDF in a postnatal organotypic culture model of motor neuron degeneration and proved it is highly neuroprotective. The effect was biologically important, significantly sparing the spinal cord's gross organotypic morphological appearance and preserving motor neuron choline acetyltransferase (ChAT). PEDF alone did not increase ChAT, indicating that the observed effect is neuroprotective, not merely an upregulation of motor neuron ChAT. Further, PEDF preserved motor neuron number, proving a survival effect. We hypothesize that PEDF may play important roles in the survival and maintenance of spinal motor neurons in their neuroprotection against acquired insults in postnatal life. It should be developed further as a therapeutic strategy for motor neuron diseases such as amyotrophic lateral sclerosis (ALS).

Neuroprotective utility and neurotrophic action of neurturin in postnatal motor neurons: comparison with GDNF and persephin.

Neurturin and persephin are recently discovered homologs of glial cell line-derived neurotrophic factor (GDNF). Here, we report that neurturin, like GDNF, increases the choline acetyltransferase activity of normal postnatal motor neurons, induces neurite outgrowth in spinal cord, and potently protects motor neurons from chronic glutamate-mediated degeneration. Persephin, in contrast, does not appear to have neurotrophic or neurite-promoting effects on mature motor neurons and may instead worsen the glutamate injury of motor neurons. This pattern in the TGF-beta family suggests certain receptor specificities, requiring at least the Ret/GFRalpha-1 receptor complex. The results predict potential benefit of neurturin, but not persephin, in the treatment of motor neuron disorders and spinal cord diseases.

Multiple roles of neural cell adhesion molecule, neural cell adhesion molecule-polysialic acid, and L1 adhesion molecules during sensory innervation of the otic epithelium in vitro.

To explore the role of cell adhesion molecules in the innervation of the inner ear, antibody perturbation was used on histotypic co-cultures of the ganglionic and epithelial anlagen derived from the otocyst. When unperturbed, these tissues survived and differentiated in this culture system with outgrowth of fasciculated neuronal fibers which expressed neural cell adhesion molecule and L1. The fibers exhibited target choice and penetration, then branching and spreading within the otic epithelium as individual axons. Treatment of the co-cultures, or of the ganglionic anlagen alone, with anti-neural cell adhesion molecule or anti-L1 Fab fragments produced a defasciculation of fibers but did not affect neurite outgrowth. In the co-cultures this defasciculation was accompanied by a small increase in the number of fibers found in inappropriate tissues. However, the antibodies did not prevent fiber entry to the otic epithelium. In contrast, removal of polysialic acid from neural cell adhesion molecule with endoneuraminadase-N, while producing a similar fiber defasciculation, also increased the incidence of fibers entering the epithelium. Nevertheless, once within the target tissue, the individual fibers responded to either Fab or to desialylation by spreading out more rapidly, branching, and growing farther into the epithelium. The findings suggest that fasciculation is not essential for specific sensory fibers to seek out and penetrate the appropriate target, although it may improve their tracking efficiency. Polysialic acid on neural cell adhesion molecule appears to limit initial penetration of the target epithelium. Polysialic acid as well as neural cell adhesion molecule and L1 function are involved in fiber-target interactions that influence the arborization of sensory axons within the otic epithelium.

Properties of the novel intermediate filament protein synemin and its identification in mammalian muscle.

We examined specific properties of highly purified synemin (230 kDa), recently identified as a novel intermediate filament (IF) protein, from avian smooth muscle. Soluble synemin in 10 mM Tris-HCl, pH 8.5, appears as approximately 11-nm-diameter globular structures by negative-stain and low-angle shadow electron microscopy. Chemical crosslinking and SDS-PAGE analysis indicate that soluble synemin molecules contain two 230-kDa subunits. The pH- and ionic strength-dependent solubility properties of synemin are similar to those of the type III IF protein desmin, but under physiological-like conditions in which desmin self-assembles into long approximately 10-nm-diameter IFs, synemin self-associates into complex, approx 15- to 25-nm-diameter globular structures. Calpain digestion demonstrated that synemin is extremely proteolytically labile. Western blot analysis, with monospecific polyclonal antibodies against avian synemin, shows the presence of the reactive 230-kDa synemin band in samples of adult avian skeletal, cardiac, and smooth muscle and of two reactive bands at approximately 225 kDa (major) and approximately 195 kDa in adult porcine skeletal, cardiac, and smooth muscle. Partial purification of synemin from porcine smooth muscle also resulted in fractions highly enriched in the approximately 225- and approximately 195-kDa polypeptides. Conventional immunofluorescence and immunoconfocal microscopy of isolated myofibrils and of frozen sections also demonstrated, for the first time, that synemin is present in all three adult porcine muscle cell types and is colocalized with desmin in skeletal and cardiac muscle cells at the myofibrillar Z-lines.

New growth of axons in the cochlear nucleus of adult chinchillas after acoustic trauma.

This study determined the effect of acoustic overstimulation of the adult cochlea on axons in the cochlear nucleus. Chinchillas were exposed to an octave-band noise centered at 4 kHz at 108 dB sound pressure level for 1.75 h. One chinchilla was never exposed to the noise, and several others had one ear protected by an ear plug or prior removal of the malleus and incus. Exposure of unprotected ears caused loss of inner and outer hair cells and myelinated nerve fibers, mostly in the basal half of the cochlea. Cochlear nerve fiber degeneration, ipsilateral to the exposed ears, was traced to regions of the cochlear nucleus representing the damaged parts of the cochlea. In silver impregnations of a deafferented zone in the posteroventral cochlear nucleus, the concentration of axons decreased by 43% after 1 month and by 54% after 2 months. However, by 8 months, the concentration of thinner axons, with diameters of less than 0.46 microm, increased by 46-90% over that at 2 months. The concentration of axons with larger diameters did not change. Between 2 and 8 months small axonal endings appeared next to neuronal cell bodies. This later increase of thinner axons and endings is consistent with a reactive growth of new axons of relatively small diameter. The emergence of small perisomatic boutons suggests that the new axons formed synaptic endings, which might contribute to an abnormal reorganization of the central auditory system and to the pathological changes that accompany acoustic overstimulation.

Differential expression of N-methyl-D-aspartate receptor in the cochlear nucleus of the mouse.

Glutamate is used in the cochlear nucleus as a neurotransmitter by cochlear nerve synapses and by local circuits of granule cell axons. In the present study, immunocytochemistry and in situ hybridization were used to identify different types of neurons expressing N-methyl-D-aspartate receptor subunit I (NMDAR1) in the mouse cochlear nucleus. N-Methyl-D-aspartate receptor subunit 1 was expressed in most neuronal types, but granule cells in the dorsal cochlear nucleus had little, if any, expression, unlike their heavily labeled counterparts in the small cell shell and cerebellum. The findings do not support an analogy between the dorsal cochlear nucleus and the cerebellar cortex. In the cochlear nucleus the most heavily labeled structures were dendrites in the small cell shell and superficial dorsal cochlear nucleus, including the fusiform cell apical dendrites, which are targets of granule cell axons. However, fusiform cell basal dendrites, which are the synaptic sites of cochlear nerve fibers, did not express N-methyl-D-aspartate receptor subunit 1. Thus different parts of the fusiform cells can have different subunits in their glutamate receptors. Also branches of the same cochlear nerve axons projecting to the octopus, stellate, and bushy cells of the ventral cochlear nucleus can use N-methyl-D-aspartate receptor, while their branches to fusiform cells cannot. Each cochlear nucleus neuron type has a characteristic level of N-methyl-D-aspartate receptor subunit 1 expression. Each type differs in its auditory response properties, which may depend on synaptic activities requiring different glutamate subunit patterns.

Difference in expression of phosphorylated tau epitopes between sporadic inclusion-body myositis and hereditary inclusion-body myopathies.

Sporadic inclusion-body myositis (s-IBM) and the hereditary inclusion-body myopathies (h-IBMs) are severe and progressive muscle diseases, characterized pathologically by vacuolated muscle fibers containing paired-helical filaments (PHFs). An interesting feature of the s- and h-IBM muscle phenotype is its striking similarity to Alzheimer-disease (AD) brain. We immunostained muscle biopsies of 9 s-IBM patients, 9 autosomal-recessive h-IBM patients, 1 autosomal-dominant h-IBM patients, and 18 normal and disease-controls with several antibodies known to react with the hyperphosphorylated tau of AD-PHFs. Those included SMI-31, SMI-310, PHF-1, and AT8. In both s- and h-IBM, virtually all vacuolated muscle fibers had strongly immunoreactive inclusions with SMI-31, and by immuno-electronmicroscopy SMI-31 was exclusively localized to PHFs. Approximately 40 to 50% of both s- and h-IBM vacuolated muscle fibers were also immunoreactive with AT8 antibody. To the contrary, in h-IBM, there was no immunoreactivity with SMI-310 and PHF-1 antibodies, whereas in s-IBM the vacuolated muscle fibers had strong immunoreactivity with those two antibodies. By immunoelectronmicorscopy, SMI-310 and PHF-1 also were localized to PHFs. Within s-IBM muscle fibers, the structures immunoreactive with SMI-310 were congophilic, whereas h-IBM muscle fibers did not have congophilia. Our studies: (a) demonstrate a distinct difference between s-IBM and the h-IBMs in regard to expression of immunoreactive phosphorylated tau and congophilia; (b) demonstrate a new "diagnostic duo" combination of SMI-31 and SMI-310 antibodies for identifying and distinguishing s-IBM and the h-IBMs; and (c) provide another close similarity of pathologic phenotypes between s-IBM muscle and AD brain, suggesting that similar cellular pathogenic mechanisms may be active in both diseases.

NMDA receptor expression in the mouse cerebellar cortex.

A detailed, light microscopic study on the distribution of the N-methyl- D-aspartate receptor subunit 1 (NMDAR1) was carried out with immunohistochemistry and in situ hybridization on the cerebellar cortex of the mouse. With a monoclonal antibody, labeling of Purkinje cell bodies varied from intense to negative, while heavy dendritic staining was limited to the proximal dendrites (unlike the rat, which also had heavily stained distal dendrites). In the granular layer, the cell bodies and and the dendritic shafts of Golgi II cells were only moderately stained, but very intense labeling was associated with granule cell bodies, and with their dendrites and dendritic endings in the glomeruli. The mossy and climbing fibers were negative. In situ hybridization with a cRNA probe showed levels and spatial distributions of NMDAR1 mRNA consistent with the immunolabeling pattern, in that signals were strongest in the granular and Purkinje cell layers and relatively low or absent in the molecular layer and white matter. The findings are consistent with the hypothesis that NMDAR1 may be especially well concentrated at the synaptic target sites of the mossy and climbing fibers. In the mouse, NMDAR1 at the parallel fiber sites associated with Purkinje cell spiny branchlets may differ from the rat in its level of expression or in its molecular configuration.

Human muscle macrophages express beta-amyloid precursor and prion proteins and their mRNAs.

Well-characterized antibodies against beta-amyloid precursor protein (beta APP) and prion protein (PrP), and specific cRNA probes, were used to localize beta APP and PrP and their mRNAs in human muscle macrophages. Macrophages present in muscle biopsies of 51 patients with various neuromuscular disorders showed accumulation of beta APP and PrP, and strongly expressed beta APP and PrP mRNAs. These were present in all muscle macrophages unrelated to their localization within the muscle tissue or diagnosis. Our study provides the first demonstration that human muscle resident macrophages synthesize and accumulate beta APP and PrP. We suggest that those proteins play a role in biology of muscle macrophages, including their participation in inflammatory and immune responses.

Alpha 1-antichymotrypsin is strongly immunolocalized at normal human and rat neuromuscular junctions.

alpha 1-Antichymotrypsin (alpha 1-ACT) is an early-stage acute-phase plasma protein and a serpin that preferentially inactivates chymotrypsin, cathepsin G, and chymase. Using immunofluorescence with four rabbit polyclonal and two monoclonal specific antibodies against human alpha 1-ACT, we have localized alpha 1-ACT at human and rat neuromuscular junctions (NMJs). Strong alpha 1-ACT immunoreactivity (IR) was present at all NMJs identified by bound alpha-bungarotoxin (alpha-BT). alpha 1-ACT immunoreactivity typically extended slightly deeper into the muscle fiber than alpha-BT, and it closely co-localized with immunoreactivities of post-synaptic desmin, beta-amyloid precursor protein, and dystrophin at the same double- or triple-labeled NMJs. Topography of alpha 1-ACT-IR was the same at human and rat NMJs. The muscle non-junctional sarcolemma was either not immunoreactive or was only very slightly so. When the primary antibody was omitted, absorbed, or replaced by a non-immune serum, there was no immunostaining. Thus, alpha 1-ACT is a novel component of the NMJ. Although its role in the postsynaptic domain of the NMJ is unknown, it might be involved in the interaction between the presynaptic and postsynaptic components and/or inhibit excessive or unwanted serine proteases that may exist in the region of the NMJ.

Twisted tubulofilaments of inclusion body myositis muscle resemble paired helical filaments of Alzheimer brain and contain hyperphosphorylated tau.

We immunostained muscle biopsies of 8 patients with sporadic inclusion body myositis (S-IBM), 7 patients with autosomal recessive hereditary inclusion body myopathy (H-IBM) (both diseases being characterized by similar muscle fiber vacuoles containing inclusions), and 11 normal and disease controls. We used the following well-characterized antibodies against tau protein: Tau-1, Alz-50, and anti-paired helical filament (PHF) antiserum. By light microscopy, in all S-IBM muscle biopsies virtually all vacuoles immunoreactive for ubiquitin and beta-amyloid protein also contained inclusions immunoreactive with Alz-50 and anti-PHF antiserum. With tau-1 antibody, strong immunoreactivity in the vacuoles was obtained only after dephosphorylation of muscle sections. By electronmicroscopy, all three antibodies immunodecorated exclusively cytoplasmic twisted tubulofilaments (TTFs). In H-IBM, virtually all ubiquitin and beta-amyloid-positive muscle fiber vacuoles contained inclusions immunoreactive with anti-PHF antiserum, but in only 40% of those fibers were the inclusions immunoreactive with Alz-50. In six H-IBM patients there were no tau-1 immunoreactive inclusions in any of their vacuolated muscle fibers; in one patient, 24% of the vacuolated fibers had tau-1 immunoreactivity. By demonstrating that hyperphosphorylated tau, which is characteristic of Alzheimer brain PHFs, is a component of S-IBM-muscle TTFs (which are also ultrastructurally similar to PHFs), our study: 1) provides the first demonstration of abnormally accumulated tau in nonneural tissue and 2) suggests that the cytopathogenesis in Alzheimer brain and S-IBM muscle may share some similar mechanisms. Whether the difference in tau immunoreactivity between S-IBM and most of the H-IBM patients reflects a difference in genetically determined transcriptional or posttranslational modifications of tau protein or other factors remains to be determined.

Tropomodulin is highly concentrated at the postsynaptic domain of human and rat neuromuscular junctions.

Tropomodulin is expressed in skeletal muscle and recent studies suggest that tropomodulin is associated with synaptic membranes. Therefore, we have examined neuromuscular junctions by immunofluorescence analysis for tropomodulin localization. Anti-tropomodulin antibodies generated against either the whole protein or a 15-amino acid peptide fragment label human neuromuscular junctions as demonstrated by colocalization with alpha-bungarotoxin label. Tropomodulin labeling also colocalizes with desmin and beta-amyloid proteins which are concentrated in the postsynaptic region. Comparable results are found with immunostaining of neuromuscular junctions in rat diaphragm muscle. These immunofluorescence results suggest that tropomodulin is an integral component of the cytoskeletal lattice associated with the postsynaptic region of neuromuscular junctions.

Prion protein is abnormally accumulated in inclusion-body myositis.

In muscle biopsies of 8 sporadic inclusion-body myositis (S-IBM) and 4 hereditary inclusion-body myopathy (H-IBM) patients, vacuolated muscle fibers contained within their vacuoles strongly immunoreactive inclusions with 2 polyclonal and 1 monoclonal antibodies against prion protein (PrP). By light-microscopy, PrP deposits co-localized with beta-amyloid protein (A beta) and ubiquitin (Ub). By immuno-electronmicroscopy, both PrP and A beta were present on amorphous material and on 6-10 nm amyloid-like fibrils; and PrP and Ub co-localized on cytoplasmic twisted tubulofilaments (TTFs) and on amorphous material. Our study provides the first demonstration of abnormally accumulated PrP in pathological tissue other than brain, and it suggests that PrP may play a role in the pathogenesis of IBM.

Prion protein is strongly immunolocalized at the postsynaptic domain of human normal neuromuscular junctions.

Using three well-characterized polyclonal and monoclonal antibodies against prion protein (PrP), we demonstrated a strong concentration of PrP at human neuromuscular junctions (NMJs). Applying double and triple fluorescence-labeling, we found that PrP immunoreactivity exactly co-localized with alpha-bungarotoxin (alpha-BT) identified acetylcholine receptors, as well as with the high junctional concentrations of beta-amyloid precursor protein, beta-amyloid protein, desmin, ubiquitin and dystrophin. Therefore, PrP was considered to be located on the postsynaptic muscle membrane. At all NMJs identified by bound alpha-BT, strong PrP immunoreactivity was obtained with all PrP antibodies. This appears to be the first demonstration of PrP concentrated at human NMJs.

Strong immunoreactivity of alpha 1-antichymotrypsin co-localizes with beta-amyloid protein and ubiquitin in vacuolated muscle fibers of inclusion-body myositis.

In 10 of 10 inclusion-body myositis (IBM) patients, including 1 hereditary case, vacuolated muscle fibers contained large or small cytoplasmic inclusions immunoreactive for alpha 1-antichymotrypsin (alpha 1-ACT). All IBM muscle biopsies had characteristic cytoplasmic tubulo-filaments by electron microscopy. None of 17 control muscle biopsies contained the alpha 1-ACT immunoreactive inclusions characteristic of IBM. In vacuolated muscle fibers, alpha 1-ACT immunoreactive inclusions colocalized with beta-amyloid protein and ubiquitin immunoreactivities. Our study provides the first demonstration of alpha 1-ACT accumulations in abnormal human muscle, and it suggest that, as in Alzheimer's disease and Down's syndrome, alpha 1-ACT may be involved in the pathogenesis of IBM.