TGF-beta1 promotes microglial amyloid-beta clearance and reduces plaque burden in transgenic mice.

Abnormal accumulation of the amyloid-beta peptide (Abeta) in the brain appears crucial to pathogenesis in all forms of Alzheimer disease (AD), but the underlying mechanisms in the sporadic forms of AD remain unknown. Transforming growth factor beta1 (TGF-beta1), a key regulator of the brain's responses to injury and inflammation, has been implicated in Abeta deposition in vivo. Here we demonstrate that a modest increase in astroglial TGF-beta1 production in aged transgenic mice expressing the human beta-amyloid precursor protein (hAPP) results in a three-fold reduction in the number of parenchymal amyloid plaques, a 50% reduction in the overall Abeta load in the hippocampus and neocortex, and a decrease in the number of dystrophic neurites. In mice expressing hAPP and TGF-beta1, Abeta accumulated substantially in cerebral blood vessels, but not in parenchymal plaques. In human cases of AD, Abeta immunoreactivity associated with parenchymal plaques was inversely correlated with Abeta in blood vessels and cortical TGF-beta1 mRNA levels. The reduction of parenchymal plaques in hAPP/TGF-beta1 mice was associated with a strong activation of microglia and an increase in inflammatory mediators. Recombinant TGF-beta1 stimulated Abeta clearance in microglial cell cultures. These results demonstrate that TGF-beta1 is an important modifier of amyloid deposition in vivo and indicate that TGF-beta1 might promote microglial processes that inhibit the accumulation of Abeta in the brain parenchyma.

Growth regulatory effects of cyclic AMP and polyamine depletion are dissociable in cultured mouse lymphoma cells.

Treatment of mouse lymphoma S49 cells with D,L-alpha-difluoromethylornithine (DFMO), an inhibitor of ornithine decarboxylase, depleted cellular polyamine levels and stopped cell growth. The cells were arrested predominantly in G1. Thus, polyamine depletion may lead to a regulatory growth arrest in S49 cells. We tested two hypotheses regarding the relationship of growth arrest mediated by polyamine limitation to that mediated by cyclic AMP (cAMP). The hypothesis that cAMP-induced arrest results from polyamine depletion is not tenable, because the arrest could not be reversed by addition of exogenous polyamines, and because cellular polyamine levels do not drop in dibuturyl cyclic AMP (Bt2cAMP)-arrested cells. The hypothesis that polyamine-mediated growth arrest is effected via modulation of cAMP levels or cAMP-dependent protein kinase activity was also shown to be incorrect, because a S49 variant deficient in cAMP-dependent protein kinase was arrested by DFMO. The activities of the polyamine-synthesizing enzymes ornithine decarboxylase (ODC) and S-adenosyl methionine decarboxylase (SAMD) are both reduced in Bt2cAMP-treated cells to about 10% of that in control populations, as shown previously. DFMO diminishes ODC activity and augments SAMD activity in both untreated and Bt2cAMP-treated cells, leading to polyamine depletion in both cases.

Overexpression of tau in a nonneuronal cell induces long cellular processes.

The ways in which the various microtubule-associated proteins (MAPs) contribute to cellular function are unknown beyond the ability of these proteins to modify microtubule dynamics. One member of the MAP family, tau protein, is restricted in its distribution to the axonal compartment of neurons, and has therefore prompted studies that attempt to relate tau function to the generation or maintenance of this structure. Sf9 cells from a moth ovary, when infected with a baculovirus containing a tau cDNA insert, elaborate very long processes. This single gene product expressed in a foreign host cell grossly alters the normal rounded morphology of these cells. The slender, relatively nonbranched appearance of these processes as well as their uniform caliber resembles the light-microscopic appearance of axons observed in several neuronal culture systems. Immunolabeling of the tau-expressing Sf9 cells demonstrated tau reactivity in the induced processes, and EM that microtubule bundles were present in the processes. Microtubule stabilization alone was insufficient to generate processes, since taxol treatment did not alter the overall cell shape, despite the induction of microtubule bundling within the cell body.

Amplification and expression of heterologous ornithine decarboxylase in Chinese hamster cells.

We have developed an amplifiable mammalian expression vector based on the enzyme ornithine decarboxylase (ODC). We show greater than 700-fold amplification of this vector in ODC-deficient Chinese hamster ovary cells. A passive coamplified marker, dihydrofolate reductase (dhfr), was amplified and overexpressed 1,000-fold. This ODC vector was a dominant marker in a variety of cell types and displayed at least 300-fold amplification in wild-type Chinese hamster ovary cells.

Multiple mechanisms are responsible for altered expression of ornithine decarboxylase in overproducing variant cells.

We selected and characterized a series of mouse S49 cell variants that overproduce ornithine decarboxylase (ODC). Previously, we described variants that have an amplified ODC gene and produce about 500-fold more ODC than the wild-type cells of origin (L. McConlogue and P. Coffino, J. Biol. Chem. 258:12083-12086, 1983). We examined a series of independent variants that overproduce ODC to a lesser degree and found that a number of mechanisms other than gene amplification are responsible for the increased ODC activity. Variants were selected for resistance to 0.1 mM difluoromethylornithine, an inhibitor of ODC, by either a single or a multistep process. All showed increased ODC activity and increased ODC mRNA steady-state levels. The half-life of the enzyme was not increased in any of the variants. In one class of variant the increase of ODC mRNA was sufficient to account for ODC overproduction. In a second class, the rate of synthesis of ODC polypeptide per ODC mRNA was at least four- to eightfold higher than that in wild-type cells. Therefore, these variants were altered in the translatability of ODC mRNA. Southern analysis showed that gene amplification does not account for the increased ODC mRNA levels in any of the variants. In both variant and wild-type cells, ODC activity was responsive to changes in polyamine pools; activity was reduced following augmentation of pool size. This change in activity was associated with modification of the rate of synthesis and degradation of ODC but no change in the level of ODC mRNA.

High-level neuronal expression of abeta 1-42 in wild-type human amyloid protein precursor transgenic mice: synaptotoxicity without plaque formation.

Amyloid plaques are a neuropathological hallmark of Alzheimer's disease (AD), but their relationship to neurodegeneration and dementia remains controversial. In contrast, there is a good correlation in AD between cognitive decline and loss of synaptophysin-immunoreactive (SYN-IR) presynaptic terminals in specific brain regions. We used expression-matched transgenic mouse lines to compare the effects of different human amyloid protein precursors (hAPP) and their products on plaque formation and SYN-IR presynaptic terminals. Four distinct minigenes were generated encoding wild-type hAPP or hAPP carrying mutations that alter the production of amyloidogenic Abeta peptides. The platelet-derived growth factor beta chain promoter was used to express these constructs in neurons. hAPP mutations associated with familial AD (FAD) increased cerebral Abeta(1-42) levels, whereas an experimental mutation of the beta-secretase cleavage site (671(M-->I)) eliminated production of human Abeta. High levels of Abeta(1-42) resulted in age-dependent formation of amyloid plaques in FAD-mutant hAPP mice but not in expression-matched wild-type hAPP mice. Yet, significant decreases in the density of SYN-IR presynaptic terminals were found in both groups of mice. Across mice from different transgenic lines, the density of SYN-IR presynaptic terminals correlated inversely with Abeta levels but not with hAPP levels or plaque load. We conclude that Abeta is synaptotoxic even in the absence of plaques and that high levels of Abeta(1-42) are insufficient to induce plaque formation in mice expressing wild-type hAPP. Our results support the emerging view that plaque-independent Abeta toxicity plays an important role in the development of synaptic deficits in AD and related conditions.

Presenilin-1 mutations associated with familial Alzheimer's disease do not disrupt protein transport from the endoplasmic reticulum to the Golgi apparatus.

Mutations in genes encoding presenilin-1 (PS1) and presenilin-2 (PS2) have been linked to familial forms of Alzheimer's disease (AD). Cells expressing mutant presenilins produce elevated levels of Abeta42, the major amyloid peptide found in AD plaques. The mechanism whereby this occurs remains unknown, but the localization of presenilins to endoplasmic reticulum (ER) and Golgi compartments has suggested that they may function in intracellular trafficking pathways involved in processing beta-amyloid precursor proteins (APP). To test this possibility, we coexpressed PS1(wt), PS1(M146L), or PS1(L286V) in HEK293 cells together with the LDL receptor, a classic glycoprotein marker that undergoes post-translational O-glycosylation in the Golgi compartment. Pulse-chase analysis of the receptor indicated that mutant presenilins had no effect on ER-->Golgi transport. Similar results were obtained when the studies were carried out with cells expressing the Swedish variant of APP (SWAPP751) instead of the LDL receptor. Moreover, secretion of the soluble exodomain polypeptide fragments of SWAPP751 that arise from alpha-secretase and beta-secretase cleavage was not markedly affected by the PS1 mutants. Despite the lack of discernible effect of the PS1 mutants on trafficking of proteins through the Golgi apparatus, they caused a substantial increase in the proportion of Abeta42 relative to total Abeta in the culture medium. The results suggest that mutant forms of PS1 cause elevated production of Abeta42 by a mechanism that is independent of a major disruption of exocytic trafficking of APP.

A mouse lymphoma cell mutant whose major protein product is ornithine decarboxylase.

Mutant mouse lymphoma cells that overproduce ornithine decarboxylase have been generated by selection for resistance to difluoromethylornithine, an inhibitor of the enzyme. Starting with wild type S49 mouse lymphoma cells, sensitive to growth inhibition by 10 microM difluoromethylornithine, we obtained the Z.12 line, which is approximately 100 times more resistant to that drug (McConlogue, L., and Coffino, P. (1983) J. Biol. Chem. 258, 8384-8388). Subsequent selection for still higher levels of resistance was applied to the Z.12 cells and resulted in the generation of the D4.1 line, resistant to 10 mM difluoromethylornithine. The relative synthesis of ornithine decarboxylase in wild type, Z.12, and D4.1 cells was assessed by pulse labeling these cells with [35S]methionine and analyzing the radiolabeled proteins directly, or after immunoprecipitation, on sodium dodecyl sulfate-polyacrylamide gels. As shown previously, the rate of ornithine decarboxylase synthesis is augmented in Z.12 as compared to wild type. In D4.1 cells, the rate of synthesis of ornithine decarboxylase exceeds that of any other single protein; about 15% of total protein synthesis is devoted to the enzyme. The relative amounts of translatable ornithine decarboxylase mRNA in each cell line was determined by in vitro translation of extracted RNA. These results showed that the relative rate of synthesis in each cell line is a reflection of the cell's relative content of translatable ornithine decarboxylase mRNA. Examination of the chromosomes of wild type and D4.1 cells revealed that the former are pseudodiploid and the latter tetraploid. Two of the four chromosomes 14 in D4.1 contain large homogeneously staining regions, a finding consistent with the presence of regions of gene amplification.

Ornithine decarboxylase in difluoromethylornithine-resistant mouse lymphoma cells. Two-dimensional gel analysis of synthesis and turnover.

Variant S49 mouse lymphoma cells with increased ornithine decarboxylase activity were obtained by selecting for resistance to alpha-difluoromethylornithine (DFMO), a specific inhibitor of the enzyme. Ornithine decarboxylase was identified as a specifically immunoprecipitable polypeptide that was made at an increased rate in the variant cells. Ornithine decarboxylase was also identified on a two-dimensional gel as a metabolically labeled polypeptide of Mr approximately 55,000 which was synthesized at an increased rate in two independently selected variants. Synthesis of this polypeptide was further augmented by treatment of cells with inhibitors of ornithine decarboxylase activity. The charge of the polypeptide was altered by treatment of either cells or cellular extracts with DFMO, a suicide substrate which binds covalently to the enzyme. This charge alteration and the inactivation of ornithine decarboxylase showed the same dependence on DFMO concentration and both effects were prevented by addition of either ornithine or putrescine. Pulse-chase experiments showed that the half-life of the ornithine decarboxylase polypeptide in these variant cells was 45 min. We conclude that ornithine decarboxylase is made at an increased rate in the resistant variants and that the polypeptide turns over rapidly.

Microtubule-associated protein function: lessons from expression in Spodoptera frugiperda cells.

The phenotypes induced by the expression of neuronal microtubule-associated proteins (MAPs) in Sf9 cells have provided data on the in situ function of these proteins. Both MAP2 and tau can induce long processes in Sf9 cells, and the processes contain bundles of microtubules. In both cases the microtubules are aligned with their plus ends distal. Tau expression usually induces a single process that is unbranched and of uniform caliber. Processes can form even when the cells are grown in suspension. Microtubules do not extend all the way to the tip; instead the terminal region contains an actin-rich meshwork. Taxol treatment of Sf9 cells also induces the assembly of microtubules into bundles but does not induce process formation in Sf9 cells. Therefore the in vitro properties of tau as a molecule capable of assembling, stabilizing, and bundling microtubules do not fully account for the in vivo ability of tau alone to transduce microtubule assembly into a change in cell shape. The morphological features of the processes induced by MAP2 differ in highly informative ways.

Molecular cloning and expression of the mouse ornithine decarboxylase gene.

We used mRNA from a mutant S49 mouse lymphoma cell line that produces ornithine decarboxylase (OrnDCase) as its major protein product to synthesize and clone cDNA. Plasmids containing OrnDCase cDNA were identified by hybrid selection of OrnDCase mRNA and in vitro translation. The two of these with the largest inserts together span 2.05 kilobases of cDNA. Southern blot analysis of DNA from wild-type or mutant S49 cells, cleaved with EcoRI or with BamHI, revealed multiple bands homologous to OrnD-Case cDNA, only one of which was amplified in the mutant cells. RNA transfer blot analysis showed that the major OrnD-Case mRNA in the mouse lymphoma cells is 2.0 kilobases long. A similar size mRNA was found in mouse kidney and was more abundant in the kidneys of mice treated with testosterone, an inducer of OrnDCase activity in that tissue.

Cell-type and amyloid precursor protein-type specific inhibition of A beta release by bafilomycin A1, a selective inhibitor of vacuolar ATPases.

Treatment of human 293 cells transfected with amyloid precursor protein (APP)K595N,M596L (the "Swedish" mutation) with a specific inhibitor of the vacuolar H(+)-ATPases, bafilomycin A1 (baf A), leads to a potent inhibition of the release of the A beta peptide. This is accompanied by a selective inhibition of beta-secretase activity. Surprisingly, baf A did not inhibit the production of A beta from either wild-type APP (WT APP) or from APPv7171 (the "Hardy" mutation), expressed in the same cell type. In contrast, the robust production of A beta from a human neuroglioma-derived cell line (HS683) transfected with WT APP, or from primary human mixed brain cultures (HMBC) expressing genomic WT APP, were also effectively inhibited by baf A. The inhibition of A beta production from the HMBC was also accompanied by the inhibition of beta-s-APP release. No inhibition of alpha-s-APP release was seen in any of the cell types tested. These results indicate that intracellular acidic processes are rate-limiting for beta-secretase cleavage and A beta production from SW APP, but not WT APP, in the peripheral 293 cell line. Furthermore, such acidic processes also play a rate-limiting role in A beta release from human central nervous system-derived cells, including HMBC. Differential trafficking of the SW APP into an acidic compartment conducive to beta-secretase cleavage and A beta release could be one explanation for the increased production of A beta observed on expression of this mutation.

Differential effects of a Rab6 mutant on secretory versus amyloidogenic processing of Alzheimer's beta-amyloid precursor protein.

The Ras-related GTP-binding protein, Rab6, is localized in late Golgi compartments where it mediates intra-Golgi vesicular trafficking. Herein we report that coexpression of Alzheimer's beta-amyloid precursor protein (beta APP751) with a dominant-negative Rab6 mutant (Rab6N126I) in human embryonal kidney 293 cells causes an increase in secretion of the soluble amino-terminal exodomain (s-APP alpha) derived from non-amyloidogenic processing of beta-APP751 by alpha-secretase. The effect was specific to Rab6N126I, since the corresponding mutation in Rab8 (i.e. Rab8N121I), which has been implicated in protein transport to the plasma membrane, caused a modest reduction in s-APP alpha secretion. While Rab6N126I stimulated secretion of APP alpha, the accumulation of amyloid beta peptide (A beta) in the medium was either moderately reduced or unaffected. Similar differential effects of Rab6N126I on secretion of s-APP alpha versus A beta were observed in cell cultures that were overproducing A beta after transfection with a plasmid encoding Swedish variant of beta APP751. Moreover, assays of medium from the latter cultures revealed a marked increase in secretion of s-APP alpha relative to s-APP beta (the immediate product derived from cleavage of beta APP by beta-secretase). The results indicate that vesicular transport events controlled by Rab6 occur at or near a critical juncture in the trans-Golgi network where beta APP is sorted into either the constitutive alpha-secretase pathway or the amyloidogenic beta-secretase pathway.

Isolation of baculovirus-derived secreted and full-length beta-amyloid precursor protein.

We have expressed two forms of the Alzheimer's beta-amyloid precursor protein (beta APP), the 695-amino acid form (695 beta APP), and the 751-amino acid form (751 beta APP) in a baculovirus system. Both forms were expressed as full-length precursor, and were subsequently processed in vivo to release extracellular secreted proteins. The secreted forms were cleaved from the full-length beta APP in a manner analogous to the cleavage of beta APP during constitutive secretion in mammalian cells (Weidemann, A., König, G., Bunke, D., Fischer, P., Salbaum, J. M., Masters, C. L., Beyreuther, K. (1989) Cell 57, 115-126; Oltersdorf, T., Ward, P. J., Henriksson, T., Beattie, E. C., Neve, R., Lieberburg, I., and Fritz, L. J. (1990) J. Biol. Chem. 265, 4492-4497). High levels of expression of 20-50 mg/liter were achieved. Both full-length and secreted forms of the beta-amyloid precursor proteins were purified using a combination of ion-exchange and immunoaffinity chromatography using a monoclonal antibody directed against beta APP. The 751 beta APP-derived full-length and secreted forms, which contain the Kunitz protease inhibitor domain, were shown to be as active in the inhibition of trypsin as is mammalian-derived secreted beta APP. The availability of purified full-length beta APP from the baculovirus system will be valuable for biochemical and cell biological analyses that may elucidate the mechanism of the inappropriate processing that leads to beta-amyloid formation in Alzheimer's disease.

Mouse ornithine decarboxylase gene: cloning, structure, and expression.

We used molecular cloning to isolate a functional gene for mouse ornithine decarboxylase (OrnDCase; L-ornithine carboxy-lyase, EC 4.1.1.17) from a cell line in which that gene had been selectively amplified. The position of the 5' terminus of the mRNA was identified, and the coding sequence was shown to be preceded by a 312- or 313-nucleotide (nt) untranslated leader. The latter is highly G + C rich, particularly in its 5'-most portion. The leader can be anticipated to have extensive and stable secondary structure. The transcription unit of the gene is of relatively small size, approximately equal to 6.2 kilobases (kb) from the start site to the proximal site of polyadenylylation. Sequence analysis of DNA near the transcription start position demonstrated the presence of a "TATA" box, but no "CAAT" box. Functional properties of the cloned gene were tested by transfecting it into cultured cells. Expression of the putative full-length gene efficiently conferred ornithine decarboxylase activity on recipient mutant cells deficient in that activity. To assess the function and strength of the OrnDCase promoter region and to delimit its boundaries, we used a transient expression assay. Upstream of a bacterial chloramphenicol acetyltransferase gene was placed a portion of the OrnDCase gene, including the presumed promoter region, spanning a region from approximately equal to 3.0 kb 5' of the site of transcription initiation to the first 250 nt of the transcript. When expressed in mouse NIH 3T3 cells, this OrnDCase genomic element was comparable in strength to the Rous sarcoma virus long terminal repeat promoter. A similar construct, truncated so as to retain only 264 base pairs of the OrnDCase gene 5' to the site of transcription start, yielded undiminished levels of expression.

Levels and alternative splicing of amyloid beta protein precursor (APP) transcripts in brains of APP transgenic mice and humans with Alzheimer's disease.

Abnormal expression of human amyloid precursor protein (hAPP) gene products may play a critical role in Alzheimer's disease (AD). Recently, a transgenic model was established in which platelet-derived growth factor (PDGF) promoter-driven neuronal expression of an alternatively spliced hAPP minigene resulted in prominent AD-type neuropathology (Games, D., Adams, D., Alessandrini, R., Barbour, R., Berthelette, P., Blackwell, C., Carr, T., Clemens, J., Donaldson, T., Gillespie, F., Guido, T., Hagopian, S., Johnson-Wood, K., Khan, K., Lee, M., Leibowitz, P., Lieberburg, I., Little, S., Masliah, E., McConlogue, L., Montoya-Zavala, M., Mucke, L., Paganini, L., and Penniman, E. (1995) Nature 373, 523-527). Here we compared the levels and alternative splicing of APP transcripts in brain tissue of hAPP transgenic and nontransgenic mice and of humans with and without AD. PDGF-hAPP mice showed severalfold higher levels of total APP mRNA than did nontransgenic mice or humans, whereas their endogenous mouse APP mRNA levels were decreased. This resulted in a high ratio of mRNAs encoding mutated hAPP versus wild-type mouse APP. Modifications of hAPP introns 6, 7, and 8 in the PDGF-hAPP construct resulted in a prominent change in alternative splice site selection with transcripts encoding hAPP770 or hAPP751 being expressed at substantially higher levels than hAPP695 mRNA. Frontal cortex of humans with AD showed a subtle increase in the relative abundance of hAPP751 mRNA compared with normal controls. These data identify specific intron sequences that may contribute to the normal neuronspecific alternative splicing of APP pre-mRNA in vivo and support a causal role of hAPP gene products in the development of AD-type brain alterations.