Characterization of the neurotrophic interaction between nerve growth factor and secreted alpha-amyloid precursor protein.

The expression and secretion of amyloid precursor protein (beta APP) is increased in rat cerebral cortices that have been denervated by subcortical lesions of the nucleus basalis of Meynert. The physiological role of the secreted beta APP in response to this injury has not been established. We have previously shown that secreted beta APP produced by alpha-secretase activity (sAPP(alpha)) potentiates the neuritogenic activity of nerve growth factor (NGF) in vitro on naive PC12 cells. In this investigation, we have further characterized the neurotrophic interaction of NGF and sAPP(alpha) using differentiated PC12 cells and rat primary cortical neurons. NGF required the expression of beta APP to maintain a neuronal phenotype. Reduction of endogenous beta APP expression by introduction of antisense oligonucleotides in the presence of NGF resulted in loss of neurites from differentiated PC12 cells but no apparent cell death. Addition of exogenous sAPP(alpha) (60--200 pM) potentiated the protective activity of NGF in serum-deprived differentiated PC12 cells as determined by retention of neurites and cell viability. In addition, exogenous sAPP(alpha) increased neuron viability in both short-term (3 days) cortical neuron cultures grown in the absence of serum and in long-term (9 days) cultures grown with serum. Disruption of the insulin signaling pathway by reduction of IRS-1 expression inhibited the ability of sAPP(alpha) to potentiate neurotrophic activity. These observations suggest that sAPP(alpha) acts as an injury-induced neurotrophic factor that interacts with NGF to enhance neuronal viability using the insulin signaling pathway.

Differential effects of transforming growth factor-beta(s) and glial cell line-derived neurotrophic factor on gene expression of presenilin-1 in human post-mitotic neurons and astrocytes.

Mutations in the presenilin-1 gene are linked to the majority of early-onset familial Alzheimer's disease cases. We have previously shown that the expression of transforming growth factor-beta is altered in Alzheimer's patients, compared to controls. Here we examine presenilin- expression in human post-mitotic neurons (hNT cells), normal human astrocytes, and human brain tumor cell lines following treatment with three isoforms of transforming growth factor-beta, or glial cell line-derived neurotrophic factor, a member of the transforming growth factor-beta superfamily. As the NT2/D1 teratocarcinoma cell line is treated with retinoic acid to induce differentiation to hNT cells, presenilin-1 messenger RNA expression is dramatically increased. Furthermore, there is a 2-3-fold increase in presenilin-1 messenger RNA expression following treatment of hNT cells with growth factors and similar results are found by Western blotting and with immunohistochemical staining for presenilin-1 protein. However, treatment of normal human astrocytes with cytokines results in minimal changes in presenilin-1 messenger RNA and protein. Interestingly, the expression of presenilin-1 in human U87 MG astrocytoma and human SK-N-SH neuroblastoma cells is only increased when cells are treated with glial cell line-derived neurotrophic factor or transforming growth factor-beta3. These findings suggest that endogenous presenilin-1 gene expression in human neurons can be induced by growth factors present in normal and diseased brain tissue. Cytokines may play a major role in regulating expression of presenilin-1 which may affect its biological actions in physiological and pathological conditions.

Transforming growth factor-beta inhibits apoptosis induced by beta-amyloid peptide fragment 25-35 in cultured neuronal cells.

Previously, we demonstrated that transforming growth factor-beta (TGF-beta) pretreatment protects neuroblastoma cell lines, human hNT neurons, and primary rat embryo hippocampal neurons (REHIPs) from degeneration caused by incubation with beta-amyloid peptide (Abeta). Here we present evidence suggesting that TGF-beta interferes with an apoptotic pathway induced by Abeta. TGF-beta preteatment decreases the amount of DNA laddering seen following Abeta treatment in neuroblastoma cells, while in REHIPs, TGF-beta decreases the number of positive cells detected in situ by Klenow labelling following Abeta treatment. RT-PCR shows that in REHIPs, Abeta decreases mRNA expression of Bcl-2, as well as the ratio of Bcl-xL/Bcl-xS, with little effect on Bax expression. These changes are expected to promote apoptosis. When REHIPs are incubated with TGF-beta before addition of Abeta, the Bcl-xL/Bcl-xS ratio and Bcl-2 levels are increased compared to cells treated with Abeta alone. Again there is little effect on Bax expression. Western blotting and immunohistochemistry experiments also show that TGF-beta maintains increased levels of Bcl-2 and Bcl-xL protein in REHIPs even in the presence of Abeta. This pattern of gene expression should function to decrease apoptosis. Similarly, RT-PCR analysis of mRNA prepared from hNT cells shows that TGF-beta pretreatment before addition of Abeta maintains a higher level of Bcl-2 expression and an increased Bcl-xL/Bcl-xS ratio as compared to cells treated with Abeta alone. In neuronal cell types treated with Abeta, TGF-beta appears to regulate expression of genes in the Bcl-2 family to favor an anti-apoptotic pathway.

Physiological levels of beta-amyloid peptide stimulate protein kinase C in PC12 cells.

Alzheimer's beta-amyloid peptide (A beta) is normally present at nanomolar concentrations in body fluids and in the medium of cultured cells. In vitro experiments have shown that A beta has neurotrophic effects and can promote neuronal adhesion and elongation of axon-like processes. In an attempt to understand the molecular mechanisms underlying such effects, we have recently reported that nanomolar doses of A beta can stimulate protein tyrosine phosphorylation and activate phosphatidylinositol-3-kinase in neuronal cells. Here we show evidence that A beta can also activate protein kinase C, a serine/threonine kinase, in PC12 cells. First, using a serine-containing S6 peptide as an exogenous substrate, we found that nanomolar levels of A beta peptides 1-40 or 1-42 significantly stimulated an S6 phosphorylating kinase activity, whereas the A beta40-1 reverse sequence peptide had no effect. Down-regulation of PKC by prolonged (18 h) treatment with 1 microM PMA prevented the A beta-induced S6 phosphorylation. Using a more specific PKC substrate, N-terminal acetylated peptide (4-14) from myelin basic protein, we then demonstrated that A beta indeed increased PKC activity and that this activity could be blocked by the PKC inhibitor, staurosporine. Finally, immunoblotting experiments showed that A beta induced translocation of PKCgamma from cytosol to membrane and also significantly reduced cytosolic PKCalpha levels. Taken together, these data suggest that physiological levels of A beta can regulate PKC activity.

Transforming growth factor-beta protects human hNT cells from degeneration induced by beta-amyloid peptide: involvement of the TGF-beta type II receptor.

Post-mitotic, human neurons (hNT cells) which have a phenotype similar to that of terminally differentiated neurons of the central nervous system were generated by treating the NT2/D1 human teratocarcinoma cell line with retinoic acid. Treatment of both hNT and NT2/D1 cells with 10(-5) M beta-amyloid peptide fragment 25-35 (A beta P) for 24 h resulted in a decrease in cell viability as determined by MTT incorporation and Trypan blue exclusion, and also induced an apoptotic morphology in hNT cells. Pre-treatment of cells for 24 h with 10 ng/ml TGF-beta 1 or 2 before addition of A beta P reduced the apoptotic morphology of hNT cells and increased cell viability in hNT cells, but not in NT2/D1 cells. Results of RT-PCR, immunohistochemistry and analysis of receptor cross-linking of [125I]TGF-beta 1 to the cell membrane, all showed that the TGF-beta type II receptor is expressed by hNT cells, but not NT2/D1 cells. These results suggest that TGF-beta can protect human, terminally differentiated, TGF-beta type II receptor-positive neurons from A beta P toxicity. We propose that the increased expression of TGF-beta in brains of patients with Alzheimer's disease may offer some degree of neuroprotection if neurons also express a functional TGF-beta type II receptor.

Increase in AP-1 transcription factor DNA binding activity by valproic acid.

Valproic acid (VPA), a simple branched fatty acid anticonvulsant, has been demonstrated to have clinical efficacy in the treatment of manic-depressive illness (Bowden et al., 1994), but the mechanism(s) by which VPA produces its therapeutic effects remain to be elucidated. VPA's clinical antimanic action require a lag period for onset and are not immediately reversed upon discontinuation of treatment, effects that suggest alterations at the genomic level; we therefore investigated the effects of VPA on the modulation of the DNA binding activity of key transcription factors. DNA binding activities of activator protein 1 (AP-1) and cAMP responsive element binding protein (CREB) were studied in acute (hours) and chronic (days) VPA-treated rat C6 glioma cells. VPA did not affect CREB DNA binding activity, but concentration- and time-dependently increased AP-1 DNA binding activity. The activity was raised at 2 hours (the shortest time examined) and remained high after 6 days (the longest time used) of continuing VPA treatment. VPA also enhanced AP-1 DNA binding activity in human neuroblastoma (SH-SY5Y) cells. Because the effects of VPA were markedly inhibited by cycloheximide, they appear to require new protein synthesis. Taken together, the data suggest that antimanic agents may affect gene expression by modulation of the activity of major transcription factors; in view of the key roles of these nuclear transcription regulatory factors in long-term neuronal plasticity and cellular responsiveness, these effects may play a major role in VPA's therapeutic efficacy and are worthy of further study.

Attenuation of cyclic AMP production by carbamazepine.

The anticonvulsant carbamazepine is an effective treatment both for epilepsy and for bipolar affective disorder, but the molecular mechanism(s) underlying its therapeutic effects have not been identified. We have found that carbamazepine exerts significant inhibitory effects on the cyclic AMP (cAMP) generating system. Within the clinical therapeutic range (approximately 50 microM), carbamazepine inhibited both basal and forskolin-stimulated cAMP production, without having any significant effects on phosphodiesterase activity. Carbamazepine also exerted its inhibitory effects on the cAMP generating system in pertussis toxin-treated cells, suggesting that the action of carbamazepine was likely mediated through an inhibitory guanine nucleotide binding protein-independent mechanism. A forskolin affinity purification column was used to purify adenylyl cyclases from rat cerebral cortex, and we found that carbamazepine inhibited both basal and forskolin-stimulated activity of purified adenylyl cyclase. We also investigated the effects of carbamazepine on the levels of the transcription factor, cAMP response element binding protein in the phosphorylated (active) state, and found that carbamazepine significantly inhibited forskolin-induced phosphorylation of the cAMP response element binding protein. The data indicate that carbamazepine inhibits adenylyl cyclase activity as well as the downstream effects of activation of adenylyl cyclase.

Effects of valproic acid on beta-adrenergic receptors, G-proteins, and adenylyl cyclase in rat C6 glioma cells.

Valproic acid (VPA) is an anticonvulsant drug with demonstrated efficacy in the treatment of mania. In the present study, we found that chronic exposure of rat C6 glioma cells to VPA induces a coordinate decrease in multiple components of the beta-adrenergic receptor- (beta-AR) coupled cyclic adenosine 3'-5'monophosphate (cAMP) generating system. Chronic VPA decreased the number of beta-ARs in a time- and concentration-dependent manner; the decrease of beta-ARs was largely beta 1-AR selective and affected beta-ARs in both the high- and low-affinity states. Chronic VPA also significantly attenuated receptor- and postreceptor-stimulated cAMP production, [3H]forskolin binding sites, immunolabeling of G alpha s 45, and cholera toxin catalyzed ADP-ribosylation of G alpha s 52 and 45. Although the precise underlying mechanisms remain to be elucidated, such profound long-term changes in the functioning of this key signaling pathway may help explain the antimanic effects of chronic VPA treatment and are worthy of further study.

Chronic sodium valproate selectively decreases protein kinase C alpha and epsilon in vitro.

Valproic acid (VPA) is a fatty acid antiepileptic with demonstrated antimanic properties, but the molecular mechanism or mechanisms underlying its therapeutic efficacy remain to be elucidated. In view of the increasing evidence demonstrating effects of the first-line antimanic drug, lithium, on protein kinase C (PKC), we investigated the effects of VPA on various aspects of this enzyme. Chronic exposure (6-7 days) of rat C6 glioma cells to "therapeutic" concentrations (0.6 mM) of VPA resulted in decreased PKC activity in both membrane and cytosolic fractions and increased the cytosol/membrane ratio of PKC activity. Western blot analysis revealed isozyme-selective decreases in the levels of PKC alpha and epsilon (but not delta or zeta) in both the membrane and cytosolic fractions after chronic VPA exposure; VPA added to reaction mixtures did not alter PKC activity or 3H-phorbol ester binding. Together, these data suggest that chronic VPA indirectly lowers the levels of specific isozymes of PKC in C6 cells. Given the pivotal role of PKC in regulating neuronal signal transduction and modulating intracellular cross-talk between neurotransmitter systems, the specific decreases in PKC alpha and epsilon may play a role in the antimanic effects of VPA.

Selective fasciculation as a mechanism for the formation of specific chemical connections between Aplysia neurons in vitro.

Selective fasciculation of growth cones along preestablished axon pathways expressing matching or complementary adhesion molecules is thought to be an important strategy in axon guidance. Growth cone inhibiting factors also appear to influence pathfinding decisions. We have used identified Aplysia neurons in vitro to explore the hypothesis that similar mechanisms could be involved in target selection. Co-cultures of L10 neurons with RB neuron targets or R2 neurons with RUQ neuron targets reliably formed chemical connections. In contrast, co-cultures of L10 with RUQ targets usually failed to form detectable chemical connections unless cell-cell contact was forced during plating by intertwining the major axons. These data suggested that differences in the ability to form cell-cell contacts might underlie the observed synaptic specificity. This notion was supported when fluorescent dye fills of L10 and R2 revealed a positive correlation between the amount of target contact and the frequency of synapse formation: L10-RUQ cultures showed much less target contact than L10-RB or R2-RUQ cultures. To examine the cellular mechanisms of these differences in target contact, presynaptic growth cones were observed as they interacted with target processes. L10-RUQ cultures showed much less fasciculation and more avoidance behavior compared to L10-RB and R2-RUQ cultures. This initial specificity suggested that the differences in amount of target contact arose through selective fasciculation and avoidance rather than through selective elimination after indiscriminate fasciculation. Selective fasciculation and avoidance might, therefore, aid in target selection by regulating the amount of contact between presynaptic processes and potential target cells.

GM1 ganglioside treatment after global ischemia protects changes in membrane fatty acids and properties of Na+, K+-ATPase and Mg2+-ATPase.

An examination was made of the effects of ganglioside GM1 (i.m.) on the losses of membrane fatty acids (palmitic, stearic, oleic, linoleic, and arachidonic), the plasma membrane enzyme Na+, K+-ATPase, and the mitochondrial membrane enzyme Mg2+-ATPase, associated with global ischemia 24 hr after permanent unilateral occlusion of the carotid artery in Mongolian gerbils. While there was a significant loss of fatty acids in saline controls, no loss was detected in membranes from GM1-injected gerbils. Rather, we found an increase in membrane fatty acid content, indicative of altered turnover. A 38% loss of Na+, K+-ATPase and a 36% loss of mitochondrial Mg2+-ATPase observed in membranes from saline controls was reduced in membranes from GM1-injected animals to losses of 15% and 8% respectively. These effects are further described by analyses of enzyme kinetics (apparent Vmax and apparent Km). After 1 week of storage, the activities of both membrane ATPases from saline controls decreased substantially more than from GM1-injected animals, suggesting that the GM1 membranes were better "preserved." Since there was a minimal loss in protein content after 24 hr of ischemia, these results indicate that systemically injected GM1 may protect structure and function of plama membranes during the acute phases of ischemic injury.