Glial fibrillary acidic protein, GFAP, is a major intermediate filament protein of glial cells and major cytoskeletal structure in astrocytes. The entorhinal cortex has a key role in memory function and is one of the first brain areas to reveal hallmark structures of Alzheimer's disease and therefore provides an ideal tissue to investigate incipient neurodegenerative changes. Here we have analyzed age- and disease-related occurrence and composition of GFAP in the human entorhinal cortex by using one- and two-dimensional electrophoresis, Western blots and immunocytochemistry combined with confocal microscopy. A novel monoclonal antibody, GF-02, was characterized that mainly reacted with intact GFAP molecules and indicated that more acidic and soluble GFAP forms were also more susceptible to degradation. GFAP and vimentin increased with aging and in Alzheimer's disease (AD). Two-dimensional electrophoresis and Western blots revealed a complex GFAP pattern, both in aging and AD with different modification and degradation forms. Immunohistochemistry indicated that reactive astrocytes mainly accumulated in relation to neurofibrillary tangles and senile plaques in deeper entorhinal cortex layers. GFAP may be used as an additional but not exclusive diagnostic tool in the evaluation of neurodegenerative diseases because its levels change with age and respond to senile plaque and tangle formation.

Neurofilament (NF) proteins consist of three subunits of different molecular weights defined as NF-H, NF-M, and NF-L. They are typical structures of the neuronal cytoskeleton. Their immunocytochemical distribution during postnatal development of cat cerebellum was studied with several monoclonal and polyclonal antibodies against phosphorylated or unmodified sites. Expression and distribution of the triplet neurofilament proteins changed with maturation. Afferent mossy and climbing fibers in the medullary layer contained NF-M and NF-L already at birth, whereas NF-H appeared later. Within the first three postnatal weeks, all three subunits appeared in mossy and climbing fibers in the internal granular and molecular layers and in the axons of Purkinje cells. Axons of local circuit neurons such as basket cells expressed these proteins at the end of the first month, whereas parallel fibers expressed them last, at the beginning of the third postnatal month. Differential localization was especially observed for NF-H. Depending on phosphorylation, NF-H proteins were found in different axon types in climbing, mossy, and basket fibers or additionally in parallel fibers. A nonphosphorylated NF-H subunit was exclusively located in some Purkinje cells at early developmental stages and in some smaller interneurons later. A novel finding is the presence of a phosphorylation site in the NF-H subunit that is localized in dendrites of Purkinje cells but not in axons. Expression and phosphorylation of the NF-H subunit, especially, is cell-type specific and possibly involved in the adult-type stabilization of the axonal and dendritic cytoskeleton.

For accurate and quantitative immunohistochemical localization of antigens it is crucial to know the solubility of tissue proteins and their degree of loss during processing. In this study we focused on the solubility of several cytoskeletal proteins in cat brain tissue at various ages and their loss during immunohistochemical procedures. We further examined whether fixation affected either solubility or immunocytochemical detectability of several cytoskeletal proteins. An assay was designed to measure the solubility of cytoskeletal proteins in cryostat sections. Quantity and quality of proteins lost or remaining in tissue were measured and analyzed by electrophoresis and immunoblots. Most microtubule proteins were found to be soluble in unfixed and alcohol fixed tissues. Furthermore, the microtubule proteins remaining in the tissue had a changed cellular distribution. In contrast, brain spectrin and all three neurofilament subunits were insoluble and remained in the tissue, allowing their immunocytochemical localization in alcohol-fixed tissue. Synapsin I, a protein associated with the spectrin cytoskeleton, was soluble, and aldehyde fixation is advised for its immunohistochemical localization. With aldehyde fixation, the immunoreactivity of some antibodies against neurofilament proteins was reduced in axons unveiling novel immunogenic sites in nuclei that may represent artifacts of fixation. In conclusion, protein solubility and the effects of fixation are influential factors in cytoskeletal immunohistochemistry, and should be considered before assessments for a quantitative distribution are made.

This study focused mainly on changes in the microtubule cytoskeleton in a transgenic mouse where beta-galactosidase fused to a truncated neurofilament subunit led to a decrease in neurofilament triplet protein expression and a loss in neurofilament assembly and abolished transport into neuronal processes in spinal cord and brain. Although all neurofilament subunits accumulated in neuronal cell bodies, our data suggest an increased solubility of all three subunits, rather than increased precipitation, and point to a perturbed filament assembly. In addition, reduced neurofilament phosphorylation may favor an increased filament degradation. The function of microtubules seemed largely unaffected, in that tubulin and microtubule-associated proteins (MAP) expression and their distribution were largely unchanged in transgenic animals. MAP1A was the only MAP with a reduced signal in spinal cord tissue, and differences in immunostaining in various brain regions corroborate a relationship between MAP1A and neurofilaments.

A panel of novel monoclonal antibodies was tested on the human entorhinal cortex for the recognition of age- and disease-related changes of neurofilament proteins (NF). Several antibodies identified phosphorylated NF-H subunit, which occurred preferentially in those aged between 60 and 80 years and were localized in degenerating neurons. Such neurons also contained neurofibrillary tangles, but neurofilament aggregates did not co-localize with tangles, nor did the quantity nor the number of NF-positive neurons correlate with the severity of Alzheimer's disease. This points to a susceptibility of NF in a subset of neurons for phosphorylation- and metabolically related morphological changes during neurodegeneration.