Dysfunction of the PI3K-Akt-GSK-3 pathway is a common feature in cell culture and in vivo models of prion disease

AIMS

Transmissible spongiform encephalopathies, also called prion diseases, are characterized by the cerebral accumulation of misfolded prion protein (PrP(SC) ) and subsequent neurodegeneration. However, despite considerable research effort, the molecular mechanisms underlying prion-induced neurodegeneration are poorly understood. Here, we explore the hypothesis that prions induce dysfunction of the PI3K/Akt/GSK-3 signalling pathway.

METHODS

We employed two parallel approaches. Using cell cultures derived from mouse primary neurones and from a human neuronal cell line, we identified common elements that were modified by the neurotoxic fragment of PrP(106-126) . These studies were then complemented by comparative analyses in a mouse model of prion infection.

RESULTS

The presence of a polymerized fragment of the prion protein (PrP(106-126) ) or of a prion strain altered PI3K-mediated signalling, as evidenced by Akt inhibition and GSK-3 activation. PI3K activation by the addition of insulin or the expression of a constitutively active Akt mutant restored normal levels of Akt and GSK-3 activity. These changes were correlated with a reduction in caspase activity and an increase in neuronal survival. Moreover, we found that activation of caspase 3, Erk and GSK-3 are common features of PrP(106-126) -mediated neurotoxicity in cellular systems and prion infection in the mouse cerebellum, while activation of caspase 12 and JNK was observed in cellular models.

CONCLUSIONS

Our findings in cell culture and in vivo models of prion disease demonstrate marked alterations to the PI3K/Akt/GSK-3 pathway and suggest that two additional pathways contribute to PrP-induced neurotoxicity as responsible of JNK and caspase 12 activation.

Role of PrP(C) Expression in Tau Protein Levels and Phosphorylation in Alzheimer's Disease Evolution

Alzheimer's disease (AD) is characterized by the presence of amyloid plaques mainly consisting of hydrophobic β-amyloid peptide (Aβ) aggregates and neurofibrillary tangles (NFTs) composed principally of hyperphosphorylated tau. Aβ oligomers have been described as the earliest effectors to negatively affect synaptic structure and plasticity in the affected brains, and cellular prion protein (PrP(C)) has been proposed as receptor for these oligomers. The most widely accepted theory holds that the toxic effects of Aβ are upstream of change in tau, a neuronal microtubule-associated protein that promotes the polymerization and stabilization of microtubules. However, tau is considered decisive for the progression of neurodegeneration, and, indeed, tau pathology correlates well with clinical symptoms such as dementia. Different pathways can lead to abnormal phosphorylation, and, as a consequence, tau aggregates into paired helical filaments (PHF) and later on into NFTs. Reported data suggest a regulatory tendency of PrP(C) expression in the development of AD, and a putative relationship between PrP(C) and tau processing is emerging. However, the role of tau/PrP(C) interaction in AD is poorly understood. In this study, we show increased susceptibility to Aβ-derived diffusible ligands (ADDLs) in neuronal primary cultures from PrP(C) knockout mice, compared to wild-type, which correlates with increased tau expression. Moreover, we found increased PrP(C) expression that paralleled with tau at early ages in an AD murine model and in early Braak stages of AD in affected individuals. Taken together, these results suggest a protective role for PrP(C) in AD by downregulating tau expression, and they point to this protein as being crucial in the molecular events that lead to neurodegeneration in AD.

Hsp90 activity is necessary to acquire a proper neuronal polarization

Chaperones are critical for the folding and regulation of a wide array of cellular proteins. Heat Shock Proteins (Hsps) are the most representative group of chaperones. Hsp90 represents up to 1-2% of soluble protein. Although the Hsp90 role is being studied in neurodegenerative diseases, its role in neuronal differentiation remains mostly unknown. Since neuronal polarity mechanisms depend on local stability and degradation, we asked whether Hsp90 could be a regulator of axonal polarity and growth. Thus, we studied the role of Hsp90 activity in a well established model of cultured hippocampal neurons using an Hsp90 specific inhibitor, 17-AAG. Our present data shows that Hsp90 inhibition at different developmental stages disturbs neuronal polarity formation or axonal elongation. Hsp90 inhibition during the first 3h in culture promotes multiple axon morphology, while this inhibition after 3h slows down axonal elongation. Hsp90 inhibition was accompanied by decreased Akt and GSK3 expression, as well as, a reduced Akt activity. In parallel, we detected an alteration of kinesin-1 subcellular distribution. Moreover, these effects were seconded by changes in Hsp70/Hsc70 subcellular localization that seem to compensate the lack of Hsp90 activity. In conclusion, our data strongly suggests that Hsp90 activity is necessary to control the expression, activity or location of specific kinases and motor proteins during the axon specification and axon elongation processes. Even more, our data demonstrate the existence of a "time-window" for axon specification in this model of cultured neurons after which the inhibition of Hsp90 only affects axonal elongation mechanisms.

Neurite outgrowth over resting and reactive astrocytes

A classic problem in CNS fiber regeneration is that the glial scar, generated after a lesion, is not crossed by regenerating axons. We know that reactive astrocytes are important in the formation of this barrier and that the barrier is not mechanical. However, its precise nature remains unclear. To study interactions of normal and reactive astrocytes with central neurites, we have attempted to create an in vitro model of the glial scar. We found the following: (1) Cultured astrocytes, independently of their lineage, morphology, immunological type and treatment with differentiating agents, induced profuse neurite outgrowth from various kinds of embryonic CNS neurons. The outgrowth was comparable to that elicited by laminin. (2) Membranes from isomorphic gliotic tissue (induced by deafferentation or excitotoxic injury and containing a large number of reactive astrocytes), inhibited central neurite outgrowth as powerfully as myelin. Reactive astrocyte membranes from areas of anisomorphic gliosis (following penetrating trauma) were permissive for neurite outgrowth, but growth was more limited than on cultured astrocyte membranes. (3) When given a choice, growing neurites actively avoided membranes from isomorphic gliosis (similar to myelin), while they seemed to follow anisomorphic membrane boundaries and crossed unhindered into membranes of cultured astrocytes. In conclusion, reactive glia seem to contain both inhibitory and neurite promoting molecules, the proportion of which depends on the way gliosis has been generated. For isomorphic reactive astrocytes the balance is inhibitory for central neurite outgrowth, while anisomorphic reactive astrocytes probably express inhibitory components at lower levels and the growth promoting factors predominate. Overall, our observations suggest that reactive astrocytes are still the major problem for axonal regeneration in the CNS.

Oestradiol regulates β-catenin-mediated transcription in neurones

Oestradiol acts in the brain by multiple mechanisms, including the regulation of transcriptional activity through classical oestrogen receptors, α and β, and by the activation of membrane/cytoplasm-initiated signalling cascades. In neuroblastoma cells, primary neurones in culture and in the brain in vivo, oestradiol activates the phosphoinositide 3-kinase/Akt/glycogen synthase kinase 3 signalling pathway by a mechanism involving oestrogen receptor α. Through this pathway, oestradiol regulates the stability of β-catenin, induces the translocation of β-catenin to the cell nucleus and regulates β-catenin-mediated transcription through the T cell factor/DNA complex. Genomic analyses in neuroblastoma cells have revealed that the set of genes regulated by oestradiol through β-catenin is not identical to that regulated by the Wnt signalling pathway, revealing a new mechanism for oestradiol signalling in neurones.

WIP is a negative regulator of neuronal maturation and synaptic activity

Wiskott-Aldrich syndrome protein (WASP) -interacting protein (WIP) is an actin-binding protein involved in the regulation of actin polymerization in cells, such as fibroblasts and lymphocytes. Despite its recognized function in non-neuronal cells, the role of WIP in the central nervous system has not been examined previously. We used WIP-deficient mice to examine WIP function both in vivo and in vitro. We report here that WIP(-)(/-) hippocampal neurons exhibit enlargement of somas as well as overgrowth of neuritic and dendritic branches that are more evident in early developmental stages. Dendritic arborization and synaptogenesis, which includes generation of postsynaptic dendritic spines, are actin-dependent processes that occur in parallel at later stages. WIP deficiency also increases the amplitude and frequency of miniature excitatory postsynaptic currents, suggesting that WIP(-)(/-) neurons have more mature synapses than wild-type neurons. These findings reveal WIP as a previously unreported regulator of neuronal maturation and synaptic activity.

Pharmacological inhibition of GSK-3 is not strictly correlated with a decrease in tyrosine phosphorylation of residues 216/279

Recent evidence suggests that intramolecular autophosphorylation is responsible for the tyrosine phosphorylation (pY) of residues 279 or 216 of glycogen synthase kinase-3 (GSK-3alpha or beta), an event that appears to play an important role in regulating this kinase. This provocative hypothesis was based on the capacity of certain nonselective GSK-3 inhibitors to alter both the activity of GSK-3 and its pY. Inhibitors of GSK-3 are not always capable of preventing this tyrosine phosphorylation, which may require an extended period of time. For example, although lithium chloride inhibits GSK-3 activity, this inhibition does not alter its pY content. Furthermore, even when GSK-3 activity is impaired, GSK-3 pY can still be modified by physiological or pharmacological agents. Taken together, these data indicate that GSK-3 kinase activity is not necessarily correlated with the extent of GSK-3 pY. We hypothesized that some as-yet-unidentified tyrosine kinases and phosphatases may also regulate this kinase.

The marine sphingolipid-derived compound ES 285 triggers an atypical cell death pathway

The isolation of new molecules from marine sources opens the door to their possible therapeutic use against tumors and other pathological conditions. Indeed, we recently defined the cytotoxicity of ES 285, obtained from the clam Mactromeris polynima, and its affects on the cells microfilament but not the microtubule network. Considering the analogy between ES 285 and sphingosine-related lipids, we wondered whether ES 285 might affect the activity of PKC at the intracellular level. While we anticipated that ES 285 might inhibit PKC, it turns out that in contrast it serves to activate PKC at the cellular level. Indeed, like other sphingosine-related lipids, ES 285 induces the phosphorylation of MARCKS. Additionally, we further examined the cytotoxicity of ES 285 to elucidate the molecular mechanisms through which this compound triggers apoptosis. When the influence of ES 285 on "cell death markers" was assessed, it became clear that ES285 activates caspase 3 and 12, and that it modified the phosphorylation of p53. In contrast, ES 285 does not affect other pathways widely implicated in regulating cell survival/apoptosis, such as JNK, Erks or Akt. Thus, these data suggest that ES 285-triggers an atypical cell death program when compared to other sphingosine-dependent apoptosis pathways.

Estradiol inhibits GSK3 and regulates interaction of estrogen receptors, GSK3, and beta-catenin in the hippocampus

Estrogens regulate a wide set of neuronal functions such as gene expression, survival and differentiation in a manner not very different from that exerted by neurotrophins or by growth factors. The best-studied hormonal action is the transcriptional activation mediated by estrogen receptors. However, the direct effects of estrogen on growth factor signaling have not been well clarified. The present data show that estradiol, in vivo, induces a transient activation of GSK3 in the adult female rat hippocampus, followed by a more sustained inhibition, as inferred from phosphorylation levels of Tau. Similar data was obtained from cultured hippocampal neurons when treated with the hormone. The transient activation was confirmed by direct measure of GSK3 kinase activity. In addition, our results show a novel complex of estrogen receptor alpha, GSK3, and beta-catenin. The presence of the hormone removes beta-catenin from this complex. There is a second complex, also affected by estradiol, in which Tau is associated with GSK3, beta-catenin, and elements of the PI3 kinase complex. Considering the role of GSK3 in neurodegeneration, our data suggest that part of the neuroprotective effects of estrogen may be due to the control of GSK3.

A cAMP-activated pathway, including PKA and PI3K, regulates neuronal differentiation

Neuronal differentiation is a complex process in which many different signalling pathways may be involved. An increase in the intracellular levels of cyclic AMP (cAMP) has been shown to induce neuronal differentiation and also to cooperate with NGF to induce PC12 neurite outgrowth in a Ras-dependent manner. However, the neuritogenic activities associated with cAMP are still not well understood. The purpose of this study was to investigate the potential neuritogenic activities mediated by cAMP. For this purpose, we used the human neuroblastoma cell line SH-SY5Y. These neuroblastoma cells respond to cAMP by forming neurite-like extensions. We tried to identify some essential pathways involved in the cAMP-induced neurite elongation of these cells. Our results indicated that PKA is transiently activated in this elongation model. When we blocked PKA activity, elongation did not take place. Similarly, PI3K also plays an essential role because when we blocked this kinase activity, there was no neurite elongation. Indeed, over-expression of the p110-catalytic subunit or an activating form of the p85-regulatory subunit (p65) is able to induce some degree of neurite extension. Moreover, our results showed that when elongation is initiated, PI3K is still essential for maintenance of the neuronal morphology, whereas PKA or MAPK (ERKs or p38) activation does not appear to be necessary during this process.

Semaphorin 3C preserves survival and induces neuritogenesis of cerebellar granule neurons in culture

Semaphorins (sema) constitute a family of molecules sharing a common extracellular domain (semaphorin domain). This family includes several types of secreted and membrane-associated molecules that are grouped into eight subclasses (subclasses 1-7 and viral semaphorins). Subclass 3 semaphorins are secreted molecules involved in axonal guidance, mainly through repulsive gradients and induction of growth cone collapse. More recently sema 3 molecules have been identified as positive factors in dependence of the type of neurons. Besides their axonal guidance function, some semaphorins have been implicated in apoptosis and survival. We investigated the effect of sema3C on survival and neurite outgrowth of rat cerebellar granule neurons (CGNs) in culture. 3T3 cells were stably transfected with sema3C. Several clonal lines were established and tested for their neuritogenic activity and one, S3C-8, was selected for the bulk of experiments. S3C-8 was co-cultured with CGNs. Sema3C enhanced CGN viability as assessed in co-cultures of CGNs with monolayers of S3C-8 in comparison with co-cultures of CGNs with control mock-transfected 3T3 cells. Moreover sema3C induced neuritogenesis of cultured CGNs, which express neuropilin-1 and -2. S3C-8 cells, overexpressing sema3C, were significantly more neuritogenic for CGN than poly l-lysine (PLL), a positive substrate for CGNs, as assessed by the measurement of the length of neurites and confirmed by Tau expression along the time of culture. CGNs co-cultured with S3C-8, showed up-regulation of the expression of axonal microtubule-associated proteins (MAPs) such as Tau, phosphorylated MAP2C and mode I-phosphorylated MAP1B compared with neurons cultured on control 3T3 cells. We also found increased expression of a specific marker of neuronal cell bodies and dendrites, high molecular weight MAP2 (HMW-MAP2). Interestingly, there was no accompanying up-regulation of a marker enriched within the neuronal somatodendritic domain, mode II-phosphorylated MAP1B. These data support the idea that secreted sema3C favors survival and neuritogenesis of cultured CGNs.

Altered mitogen-activated protein kinase signaling, tau hyperphosphorylation and mild spatial learning dysfunction in transgenic rats expressing the β-amyloid peptide intracellularly in hippocampal and cortical neurons

The pathological significance of intracellular Abeta accumulation in vivo is not yet fully understood. To address this, we have studied transgenic rats expressing Alzheimer's-related transgenes that accumulate Abeta intraneuronally in the cerebral and hippocampal cortices but do not develop extracellular amyloid plaques. In these rats, the presence of intraneuronal Abeta is sufficient to provoke up-regulation of the phosphorylated form of extracellular-regulated kinase (ERK) 2 and its enzymatic activity in the hippocampus while no changes were observed in the activity or phosphorylation status of other putative tau kinases such as p38, glycogen synthase kinase 3, and cycline-dependent kinase 5. The increase in active phospho-ERK2 was accompanied by increased levels of tau phosphorylation at S396 and S404 ERK2 sites and a decrease in the phosphorylation of the CREB kinase p90RSK. In a water maze paradigm, male transgenic rats displayed a mild spatial learning deficit relative to control littermates. Our results suggest that in the absence of plaques, intraneuronal accumulation of Abeta peptide correlates with the initial steps in the tau-phosphorylation cascade, alterations in ERK2 signaling and impairment of higher CNS functions in male rats.

Regulation of neuronal cytoskeleton by lysophosphatidic acid: role of GSK-3

Neurite retraction is a crucial process during nervous system development and neurodegeneration. This process implies reorganization of the neuronal cytoskeleton. Some bioactive lipids such as lysophosphatidic acid (LPA) induce neurite retraction. The reorganization of the actin cytoskeleton during neurite retraction is one of the best-characterized effects of LPA. However, less information is available regarding the reorganization of the microtubule (MT) network in response to LPA in neuronal cells. Here, we first give an overview of the roles of cytoskeleton during neurite outgrowth, and subsequently, we review some of the data from different laboratories concerning LPA-induced cytoskeletal rearrangement in neuronal cells. We also summarize our own recent results about modifications of MTs during LPA-induced neurite retraction. We have shown that LPA induces changes in tubulin pools and increases in the phosphorylation levels of microtubule-associated proteins (MAPs), such as Tau. Tau hyperphosphorylation in response to LPA is mediated by the activation of glycogen synthase kinase-3 (GSK-3). The upregulation of GSK-3 activity by LPA seems to be a general process as it occurs in diverse neuronal cells of different species in correlation with the neurite retraction process.

Participation of structural microtubule-associated proteins (MAPs) in the development of neuronal polarity

Several lines of evidence have indicated that changes in the structure of neuronal cytoskeleton provide the support for the dramatic morphological changes that occur during neuronal differentiation. It has been proposed that microtubule-associated proteins can contribute to the development of this phenomenon by controlling the dynamic properties of microtubules. In this report we have characterized the effect of the combined suppression of MAP1B and tau, and MAP1B and MAP2 on neuronal polarization in cultured hippocampal cells grown on a laminin-containing substrate. We have taken advantage of the use of a mouse line deficient in MAP1B expression obtained by the gene trapping approach. In addition to this engineered mice line we used the antisense oligonucleotide approach to induce the suppression of tau or MAP2, in wild type and MAP1B-deficient neurons. Together these results show a synergistic role for MAP1B/MAP2 and MAP1B/TAU.

Fibronectin modulation by A beta amyloid peptide (25-35) in cultured astrocytes of newborn rat cortex

Fibronectin appears to be present in Senile Plaques of Alzheimer's disease brains. These senile or neuritic plaques are surrounded by dystrophic neurites, activated microglia and reactive astrocytes. The purpose of this work was to establish if a direct correlation exists between the production of Fibronectin (FN) by astrocytes and the presence of amyloid, analysing the modification of this protein produced after the treatment of cultured astrocytes with amyloid peptide (25-35). Our data showed that the addition of previously polymerised A beta-peptide to cultured astrocytes induced a marked increase in FN immunoreactivity that is in part dependent on phosphatases 2A or phosphatase 1, since was partially inhibited by okadaic acid. The increased amount of FN did not appear to be associated to any specific single isoform of which are mainly present in the rat brain. Our data suggest that in vivo FN accumulated in senile plaques may be the result, at least in part, of the response of reactive astrocyte to the presence of amyloid peptide. The importance of FN up-regulation in vivo, as part of a 'positive' response of the astrocytes to produce molecules that favours neurite outgrowth, is discussed.

Integrating retroviral cassette extends gene delivery of HSV-1 expression vectors to dividing cells

Retroviral vectors have long been used in a wide variety of gene transfer applications but have certain drawbacks, such as small cargo size, limited tropism, and low titers. HSV expression vectors overcome these disadvantages, but, because they persist in target cells as nonreplicative episomes, they are not retained in all the progeny of dividing cells. Chimeric HSV/AAV products that can mediate transgene integration in human mitotic cells have been constructed, but, to date, genetic modification of dividing cells in animal models using HSV products has not been possible. Here, we report the construction of hybrid HSV/retroviral vectors that exhibit up to 50-fold higher transgene integration efficiency compared to vectors containing only HSV-1 components. Efficient integration of a retroviral transgene cassette encoding pac in human cells required expression of the Moloney murine leukemia virus gag-pol genes, but in murine cells, could also be mediated by endogenous activities, albeit at a lower level. Gene delivery was equally efficient in BHK21, a cell line resistant to retroviral infection, and transgene retention and expression were observed to be stable for least one month in Hs683 human glioma cells. These vectors have wide applications for the genetic modification of many cell types.

The inhibition of phosphatidylinositol-3-kinase induces neurite retraction and activates GSK3

It has been extensively described that neuronal differentiation involves the signalling through neurotrophin receptors to a Ras-dependent mitogen-activated protein kinase (MAPK) cascade. However, signalling pathways from other neuritogenic factors have not been well established. It has been reported that cAMP may activate protein kinase (PKA), and it has been shown that PKA-mediated stimulation of MAPK pathway regulates not only neuritogenesis but also survival. However, extracellular regulated kinases (ERKs) mediated pathways are not sufficient to explain all the processes which occur in neuronal differentiation. Our present data show that: in cAMP-mediated neuritogenesis, using the SH-SY5Y human neuroblastoma cell line, there exists a link between the activation of PKA and stimulation of phosphatidylinositol 3-kinase (PI3K). Both kinase activities are essential to the initial elongation steps. Surprisingly, this neuritogenic process appears to be independent of ERKs. While the activity of PI3K is essential for elongation and maintenance of neurites, its inhibition causes retraction. In this neurite retraction process, GSK3 is activated. Using both a pharmacological approach and gene transfer of a dominant negative form of GSK3, we conclude that this induced retraction is a GSK3-dependent process which in turn appears to be a common target for transduction pathways involved in lysophosphatidic acid-mediated and PI3K-mediated neurite retraction.

Perinatal lethality of microtubule-associated protein 1B-deficient mice expressing alternative isoforms of the protein at low levels

Microtubule-associated protein 1B (MAP1B) has been implicated in axogenesis in cultured cells. To gain insight into the functions that MAP1B plays in vivo, we analyzed a strain of Map1B mutant mice generated by a gene trapping approach. Homozygous mice die on the first day after birth, probably due to a severe abnormal development of the nervous system. They present alterations in the structure of several brain regions. The normal Map1B gene yields different protein isoforms from alternatively spliced transcripts. The smaller isoforms were present in wild type, hetero-, and homozygous mice, but their expression was higher in the mutants than in the wild-type. Moreover, trace amounts of MAP1B protein were also observed in Map1B homozygous mutants, indicating an alternative splicing around the gene trap insertion. Thus, the Map1B gene trapped mutation reported in this work did not generated a null mutant, but a mouse with a drastic deficiency in MAP1B expression. Analyses of these mice indicate the presence of several neural defects and suggest the participation of MAP1B in neuronal migration.

The marine compound spisulosine, an inhibitor of cell proliferation, promotes the disassembly of actin stress fibers

Spisulosine is a novel antiproliferative (antitumoral) compound of marine origin. In this work the molecular target for this toxic agent has been analyzed. In the presence of spisulosine, cultured cells change their morphology, first acquiring a fusiform morphology, and later becoming rounded without focal adhesions. Analysis of the cytoskeleton of treated cells indicate the absence of actin stress fibers.

Increasing neurite outgrowth capacity of beta-amyloid precursor protein proteoglycan in Alzheimer's disease

Progressive cerebral deposition of beta-amyloid peptide either in blood vessels or around neurites is one of the most important features of Alzheimer's disease (AD). The beta-peptide, known as Abeta or A4, is produced by proteolytic cleavage of the amyloid precursor protein (APP). Two APP processing pathways have been proposed as physiological alternatives; only one of which leads to the production of Abeta or amyloidogenic peptides. However, we have little information regarding these processing pathways in the brain, or on whether posttranslational modifications such as glycosylation affect APP processing in vivo. Furthermore, the physiological function(s) of this protein in nervous tissue remains unclear, although modulatory roles in cell adhesion and neuritic extension have been suggested. It has been reported that APP may be glycosylated as a proteoglycan. We purified this APP population from human brain, and our data indicate that PG-APP supports neurite extension of hippocampal neurons. Neurons grown on this substratum showed an increased capacity to elongate neurites and increased neuritic "branching" compared to culture on laminin. These effects were enhanced with PG-APP samples obtained from AD brains. Our results suggest that this APP population may act as a neurite outgrowth and branching promoter and may thus play a role in some pathological conditions. These findings may have significant implications in understanding normal brain development and pathological situations (such as AD).

The neurite retraction induced by lysophosphatidic acid increases Alzheimer's disease-like Tau phosphorylation

The bioactive phospholipid lysophosphatidic acid (LPA) causes growth cone collapse and neurite retraction in neuronal cells. These changes are brought about by the action of a cell surface receptor coupled to specific G proteins that control morphology and motility through the action of a group of small GTPases, the Rho family of proteins. Many studies have focused on actin reorganization modulated by Rho-GTPases, but almost no information has been obtained concerning microtubular network reorganization after LPA-induced neurite retraction. In the present study, we demonstrate an increase in site-specific Alzheimer's disease-like Tau phosphorylation during LPA-induced neurite retraction in differentiated SY-SH5Y human neuroblastoma cells. The phosphorylation state of Tau was inferred from its immunoreactivity with antibodies that recognize phosphorylation-sensitive epitopes. The effects of specific kinase inhibitors indicate that this phosphorylation is mediated by glycogen synthase kinase-3 (GSK-3). In support of this idea, we observed an increase of GSK-3 activity upon growth cone collapse. Our results are consistent with the hypothesis that activation of GSK-3 occurs in the Rho pathway and may represent an important link between microtubules and microfilaments dynamics during neuritogenesis and in pathological situations such as Alzheimer's disease.

Expression of presenilin 1 in nervous system during rat development

Alzheimer's disease (AD) is a polygenic disorder involving at least four different genes. Among them, missense mutations in the presenilins segregate with the vast majority of early onset cases of familial AD. To elucidate possible function(s) of presenilin 1 (PS1), we have studied its expression during the development of the rat nervous system. Analysis by in situ hybridization showed expression of PS1 in a variety of cell types and tissues during development, with prominent expression in the nervous system. During late embryogenesis, the ventricular zone presented the highest levels of expression, paralleling the pattern previously reported for Notch. Later, during postnatal development, we observed a peak of PS1 expression at postnatal day 10, particularly in the cerebellum and hippocampus, a time when proliferation and migration are still ongoing and synapse formation is being completed. We propose that presenilins participate in at least two different developmental processes: (1) one involved in neurogenesis and skeleton formation during embryonic development, probably involving coordinate expression with Notch, and (2) a second one in the postnatal central nervous system, perhaps involved in neuritogenic and/or synaptogenic stages, most likely playing a role in amyloid precursor protein processing and amyloid beta production.

Up-regulation of Eph tyrosine kinase receptors after excitotoxic injury in adult hippocampus

The molecular mechanisms underlying the response to injury in the central nervous system are incompletely understood. Many cell activation systems may be involved. Tyrosine kinase receptors and their ligands play key roles in cell activation throughout life. The Eph family of tyrosine kinase receptors/ ligands are developmentally regulated and have been implicated in neural pathfinding. However, nothing is known about their role in the adult brain. We have used a model of central nervous system lesion in the rat, in which intraventricular injection of kainate was performed. This produced neuronal death in the CA3-CA4 fields and glial activation in the hippocampus. Highly degenerate primers, corresponding to the catalytic domain of the tyrosine kinase family, were used for reverse transcription-polymerase chain reaction of pooled RNA extracted from injured hippocampi. The amplified products were cloned and 100 clones (arbitrarily named TK1-TK100) were examined and inserts sequenced. We obtained four clones containing inserts which belong to the Eph receptor family. Two of these inserts (TK17 and TK63) were EphA4 and the other were EphB2 (TK25) and EphA5 (TK23). We performed in situ hybridization, and we found our clones to be present in all fields of the hippocampus, their expression being mainly neuronal. Three days after lesion, prominent expression appeared in CA1 as compared to the same field in the non-treated contralateral hippocampus. We performed northern blot analysis for quantification, and found that, three days after injury, the values decreased to 33 +/- 4%, 33 +/- 1% and 46 +/- 1% of control values for TK63 (EphA4), TK25 (EphB2) and TK23 (EphA5), respectively. Neuronal death in CA3-CA4 might account for this fact. Later, five days post-injury, the expression increased to 63 +/- 3%, 71 +/- 1% and 111 +/- 5% of control values, respectively. This increase was due to an up-regulation of these genes in the hippocampal neurons that survive after the injury, as indicated by in situ hybridization.

Sulphated glycosaminoglycans prevent the neurotoxicity of a human prion protein fragment

Although a number of features distinguish the disease isoform of the prion protein (PrPSc) from its normal cellular counterpart (PrPC) in the transmissible spongiform encephalopathies (TSEs), the neuropathogenesis of these diseases remains an enigma. The amyloid fibrils formed by fragments of human PrP have, however, been shown to be directly neurotoxic in vitro. We show here that sulphated polysaccharides (heparin, keratan and chondroitin) inhibit the neurotoxicity of these amyloid fibrils and this appears to be mediated via inhibition of the polymerization of the PrP peptide into fibrils. This provides a rationale for the therapeutic effects of sulphated polysaccharides and suggests a rapid in vitro functional screen for TSE therapeutics.

BetaA amyloid peptide (25-35) induced APP expression in cultured astrocytes

Alzheimer's disease is characterized by an accumulation of senile or neuritic plaques surrounded by activated microglia and reactive astrocytes, the cell processes of which are frequently in contact with the amyloid core. The major component of this amyloid deposit is the amyloid peptide (betaA or betaA4). These reactive glia are characterized by their hypertrophic phenotype and by the overexpression of some molecules such as glial fibrillary acidic protein and the amyloid precursor protein (APP). The purpose of this work was to analyze whether APP expression was modified in astrocytes by the presence of betaA peptide. To study this, the effects of beta-Amyloid (25-35) on cultured astrocytes were analyzed and compared with those of a scrambled peptide. Our data indicated that the addition of previously polymerized betaA peptide induced a marked morphological change from a flat, polygonal shape to a stellated, process-bearing morphology. This change occurred with an increase in APP immunoreactivity that is dependent of phosphatases PP2A or PP1, since it was inhibited by okadaic acid. Upregulation of APP protein expression appears to be mainly nontranscriptional, because the increase of APP protein precedes the increase of mRNA expression. The analysis of several APP isoforms indicated that this increment is not due to changes of a single isoform. Our data may correlate with some in vivo reports of astrocytic APP induction after brain insult, suggesting an important role for betaA peptide in the initial process and/or maintenance of the reactive phenotype in vivo.

Amyloid precursor protein proteoglycan is increased after brain damage

The beta-amyloid peptide (Abeta or A4) is produced by proteolytic cleavage from amyloid precursor protein (APP). The progressive cerebral deposition of this peptide is one of the most important features of Alzheimer's disease. From the study of normal and transfected cells, two APP processing pathways have been proposed as physiological alternatives. One of these can produce Abeta or amyloidogenic peptides, whereas the second does not. However, it is not completely clear how APPs are post-translationally modified, proteolytically processed and metabolized in the brain. We report here that APPs also exist as proteoglycan, chondroitin-sulfate (ChS). We have identified in normal rat brain a complex pool of 8 to 130 kDa ChS-core proteins. The main portion of these proteoglycan (PGs) APPs contains complete amyloidogenic sequence, suggesting a novel proteolytic processing of APP from the amino-terminal to the transmembrane region. This population appears augmented after brain damage. These findings may have significant implications in understanding the initial deposition and kinetics of amyloid aggregation in a pathological situation like Alzheimer's disease.

Okadaic acid modulates the cytoskeleton changes induced by amyloid peptide (25-35) in cultured astrocytes

Amyloid beta-protein (25-35) (betaA) induced a marked morphological change in astrocytes, changing their flat polygonal shape into a stellate process-bearing morphology. The changes induced by betaA were concentration and time-dependent, whereas the addition of a scrambled peptide did not alter astrocyte morphology. We discard the possibility of betaA-astrocytes being type II-like astrocytes. We also analysed the influence of the presence of kinase and phosphate inhibitors on this morphological change. Our data indicate that the betaA-induced phenotype was not affected by the inhibition of protein tyrosine kinase or tyrosine phosphatases. Only the addition of okadaic acid to astrocytes prevented the morphological transformation from flat to stellate shape, induced by betaA (25-35). Inhibition of the stellate phenotype by okadaic acid was initiated at a concentration of 10 nM which suggested that either phosphatase 2A or 1 plays an important role in the betaA astrocytic transformation.

beta-Amyloid peptide induced cytoskeletal reorganization in cultured astrocytes

The effects of beta-amyloid (25-35) (betaA) on cultured astrocytes from rat cortex were studied and compared with those of a scrambled peptide and with untreated cultures. Single addition (from 5 to 200 microg/ml) of betaA peptide induced a marked morphological change in astrocytes, changing their flat polygonal shape into stellate process-bearing morphology. The changes induced by betaA were concentration and time-dependent. The addition of the scrambled peptide did not alter cell viability in comparison with untreated astrocyte cultures. However, cell viability was dose-dependently decreased by betaA. A subpopulation of betaA-treated astrocytes showed an increase in glial fibrillary acidic protein (GFAP) and Vimentin (Vim) immunostaining while other reactive astrocyte markers such as S100beta, MAP2, and ApoE remained unaltered or undetectable. The morphological changes in betaA-treated astrocytes appeared to be mainly due to a cytoskeletal reorganization, since the total amounts of GFAP and Vim proteins were not essentially modified. These results strongly suggest that astrocytes are another cellular target of the effects of betaA and this may be relevant to understanding the neuropathology of Alzheimer's disease.

Differential cellular response after glutamate analog hippocampal damage

The tissue response after brain damage implicates the cellular "activation" of astrocytes and microglia. This glial response is referred as reactive gliosis. Using immunohistochemical markers, we have analyzed the neuronal and glial response to some neurotoxic-induced lesions. We have compared the effects of two glutamate analogs, AMPA and kainic acid, with those of traumatic injury. Our data showed that the time-course of appearance, the relative contribution of and the behavior of reactive astrocytes and microglial cells were clearly different after AMPA or kainic acid administration. The immunoreactivity associated with microglia response, with respect to the immunoreactivity associated with reactive astrocytes, was higher after AMPA damage than after kainic acid treatment. In both cases, however, glial cells were more abundant than after traumatic lesions. Interestingly, the CA1 pyramidal neurons affected by AMPA and some cortical neurons affected by traumatic injury responded with an overexpression of amyloid precursor protein, whereas no neuronal response was detected after the kainic acid treatment. Our data suggest that the gliotic response is highly specific to the type of insult and heterogeneous depending on the brain area affected.

Neurite outgrowth inhibitors associated with glial cells and glial cell lines

'Reactive' astrocytes and 'activated' microglial cells are the major cellular components of gliotic tissue, one of the most serious obstacles to axonal regeneration in mammalian central nervous system grey and white matter. The appearance of reactive glial cells after a lesion in the CNS correlates with the expression of molecules, like proteoglycans, capable of preventing neurite outgrowth. Co-cultures of embryonic neurones with glial cells and glial cell lines, that might share characteristics with reactive astrocytes and microglial cells, show that while cultured astrocytes promote neurite outgrowth, plasma membranes of C6 glioma and microglial cells express neurite inhibitory activities with proteoglycan-like characteristics, similar to those expressed by the gliotic tissue associated inhibitors. These results suggest that in vivo microglial cells might be at least one of the sources of proteoglycans with neurite outgrowth inhibitory properties.

Differences between reactive astrocytes and cultured astrocytes treated with di-butyryl-cyclic AMP

The long-standing idea that astrogliosis acts as a barrier for regenerating axons could be tested if an in vitro model of reactive astrocytes were available. The morphology and intermediate filament content of cultured perinatal astrocytes treated with di-butyryl-cyclic-AMP are reminiscent of reactive astrocytes evoked by injury. Thus, they have been proposed as a reactive astrocyte in vitro model. However, we show here that di-butyryl-cyclic-AMP-treated astrocytes are much closer to untreated neonatal cells than to reactive astrocytes in vivo, when using other immunohistochemical markers of living reactive glia (i.e. EGF receptor or laminin). Furthermore, living di-butyryl-cyclic-AMP-treated astrocytes and untreated, flat, epithelioid cells, as well as their purified plasma membranes, had similar neurite outgrowth promoting properties, whereas membranes from gliotic tissue enriched in reactive astrocytes inhibited neurite outgrowth. Our observations indicate that di-butyryl-cyclic-AMP treatment leads, at best, to a morphological model of reactive cells that does not share many properties of reactive astrocytes in vivo.

Characterization of a neurite outgrowth inhibitor expressed after CNS injury

Reactive gliosis, a general response to injury in the central system grey and white matter, represents a serious obstacle to axonal regeneration in mammals. In culture, myelin-free plasma membranes from normal rat brain tissue promoted neurite outgrowth, whereas myelin-free membranes purified from injured tissue were inhibitory. The inhibitory activity could be solubilized by detergent, was sensible to glycosaminoglycan lyase digestion and eluted with an apparent molecular weight of 160-220 kDa in gel filtration chromatography. When presented as a surface-bound molecule, the inhibitor prevented neurite initiation; when added in a soluble form to growing neurites, it induced their retraction. These results provide cellular and molecular evidence supporting the classical view that, in the mammalian central nervous system, damage-evoked gliosis correlates with the expression of molecules capable of preventing neurite outgrowth.

Abnormal meiotic spindles cause a cascade of defects during spermatogenesis in asp males of Drosophila

Since spermatogenesis in Drosophila is a series of interconnected and interdependent steps and most of the spermatogenic events take place in the absence of transcription, failures in a given stage can give rise to a cascade of defects later on. The asp locus of Drosophila melanogaster codes for a non-tubulin component implicated in proper spindle structure and/or function (Ripoll et al. 1985). Homozygous asp males exhibit abnormal meiotic spindles giving rise to altered segregation of chromosomes and mitochondria and failures in cytokinesis. Postmeiotic spermatogenic stages of asp males show a series of alterations that we interpret as due to the previously occurring defective meiosis because meiotic spindles are the only microtubular structure altered in mutant testes. The most conspicuous alterations are: (i) variable size of nuclei and nebenkerns of early spermatids, which are also multinucleate instead of having single and uniformly sized nuclei; (ii) elongating spermatids in which abnormal-sized mitochondrial derivatives elongate alongside more than one axoneme; (iii) failures in the individualization process, where abnormal spermatids remain syncytial, and seem to be eliminated during the coiling stage.

Effect of specific proteolytic cleavages on tubulin polymer formation

The capacity for self-polymerization and shape of the tubulin polymers assembled after digestion with trypsin, Pronase, chymotrypsin, subtilisin, Staphylococcus aureus proteinase V8 and proteinase K were investigated. Digestion with trypsin, Pronase or chymotrypsin resulted in a decrease in the ability of tubulin for self-assembly, whereas limited proteolysis with subtilisin, S. aureus proteinase V8 or proteinase K resulted in an increase in such ability. The shape of the assembled polymers varied from typical microtubules (after the treatment with trypsin or Pronase) to sheets (after the treatment with chymotrypsin) and from hooked microtubules with a constant polarity (after the treatment with subtilisin) to the disappearance of a defined polarity of such polymers (after the treatment with S. aureus V8 proteinase or proteinase K). These results indicate that the tubulin C-terminal regions are involved in the regulation of microtubule polymerization, shape, directional growth and lateral interactions between tubulin protofilaments.

Microtubule-associated proteins present in different developmental stages of Drosophila melanogaster

Microtubule-associated proteins (MAPs) have been isolated from different development stages of Drosophila melanogaster and characterized by their association to tubulin, but not to tubulin lacking its 4-kD carboxy terminal region (S-tubulin), and by their ability to promote tubulin polymerization. Following these criteria some peptides of Mr 255, 205, and 180 kD were identified as MAPs. By means of immunological analogy we have identified a peptide related to mammalian brain MAP known as tau factor.

Tubulin phosphorylation by casein kinase II is similar to that found in vivo

Purified brain tubulin subjected to an exhaustive phosphatase treatment can be rephosphorylated by casein kinase II. This phosphorylation takes place mainly on a serine residue, which has been located at the carboxy-terminal domain of the beta-subunit. Interestingly, tubulin phosphorylated by casein kinase II retains its ability to polymerize in accordance with descriptions by other authors of in vivo phosphorylated tubulin. Moreover, the V8 phosphopeptide patterns of both tubulin phosphorylated in vitro by casein kinase II and tubulin phosphorylated in vivo in N2A cells are quite similar, and different from that of tubulin phosphorylated in vitro by Ca/calmodulin-dependent kinase II. On the other hand, we have found an endogenous casein kinase II-like activity in purified brain microtubule protein that uses GTP and ATP as phosphate donors, is inhibited by heparin, and phosphorylates phosphatase-treated tubulin. Thus it appears that a casein kinase II-like activity should be considered a candidate for the observed phosphorylation of beta-tubulin in vivo in brain or neuroblastoma cells.

Phosphorylation of alpha-tubulin carboxyl-terminal tyrosine prevents its incorporation into microtubules

Insulin receptor kinase phosphorylated tubulin in an insulin-dependent fashion. Two different populations of phosphotubulin were found. In tubulin dimers containing tyrosine at the carboxyl-terminal of their alpha subunit, phosphate was incorporated in that residue, and the phosphorylated protein did not assemble into polymers. In tubulin dimers lacking this tyrosine residue, phosphate was incorporated into different tyrosine residues located in other parts of the molecule, and the phosphoprotein retained its capacity to polymerize.

Localization of the phosphorylation sites for different kinases in the microtubule-associated protein MAP2

The phosphorylation of microtubule-associated protein 2 (MAP2) by four different kinases was studied in vitro to determine whether MAP2 is phosphorylated in its tubulin binding region or in the microtubule projection portion. Fragments corresponding to both regions of MAP2 were produced not only by chymotrypsin or trypsin digestion, but also using pepsin, a broad chain-specificity protease, a result supporting previous notions of the two-domain structure of MAP2. The position of these two functional domains was determined with respect to the carboxy terminal of the molecule, by labeling MAP2 exclusively at the carboxy terminal and subjecting it to pepsin digestion. The results suggested that the projection region of MAP2 contained the carboxy terminal of the protein. A phosphorylation map was constructed by subjecting phosphorylated MAP2 to enzymatic digestion using Staphylococcus aureus V8 protease or to chemical cleavage using N-chlorosuccinimide. The results indicated that all four kinases phosphorylated MAP2 in a 42-kilodalton peptide that contained the tubulin binding region but differed in the level and localization of the sites at which they phosphorylated the projection of MAP2.

Location of the regions recognized by five commercial antibodies on the tubulin molecule

Through limited proteolysis of the tubulin molecule with trypsin, chymotrypsin, subtilisin, or pronase we have mapped the regions recognized by five commercial antibodies. Two of them recognized a sequence between amino acids 340 and 400 near the C terminal of the alpha or beta subunits, one recognized a sequence between amino acids 120 and 150 present in both subunits, and another one probably recognized a conformational epitope. Simultaneously we have confirmed the results obtained for other antibodies which recognized a region previously mapped on the tubulin molecule.

Phosphorylation of tubulin enhances its interaction with membranes

Tubulin, the main component of intracellular microtubules, is also a major protein in subcellular membrane preparations and can interact with biological and artificial membranes in vitro. Of particular interest is the association of tubulin with postsynaptic junctional lattices enriched in a polypeptide of relative molecular mass (Mr) 50,000 (50K), recently identified as the major subunit of the calmodulin-dependent protein kinase. Phosphorylation of tubulin with a calmodulin-dependent protein kinase similar to that found in postsynaptic densities inhibits its ability to self-assemble into microtubules in a reversible fashion. This involves the phosphorylation of residues in its 4K carboxy-terminal region, a domain that seems to regulate its self-assembly. The results presented here suggest that the phosphorylation of tubulin with this kinase enhances its ability to interact with membranes. This effect is reversible.

Phosphorylation of tubulin by a calmodulin-dependent protein kinase

Calmodulin-dependent protein kinase was purified from porcine brain cytosol through sequential steps involving acid precipitation, DEAE-chromatography, and calmodulin-Sepharose chromatography. The purified enzyme contained a major Mr 50,000 and a minor Mr 60,000 peptide. Porcine brain tubulin was a major substrate for this kinase. Under optimal conditions 2.6 mol of phosphate were incorporated per mol of tubulin. The kinase phosphorylated both tubulin subunits at their carboxyl-terminal region. Limited proteolysis, using trypsin and chymotrypsin, of phosphorylated and unphosphorylated tubulins resulted in different cleavage patterns as determined by peptide mapping. Phosphorylated tubulin was unable to bind to microtubule-associated protein or to polymerize, but regained its assembly capacity after phosphatase treatment.

The combined use of limited proteolysis and differential peptide staining for protein characterization

A method to characterize a protein by peptide mapping is described. It involves a combination of limited proteolysis and differential staining with the dye 'stains all' which allows the identification of the proteolytic fragments by their size and colour. When this procedure was used for tubulin, specific acidic fragments were identified and localized in the molecule.