High levels of circulating Abeta42 are sequestered by plasma proteins in Alzheimer's disease.

A previously unrecognized large pool of Abeta was discovered in freshly drawn plasma of patients diagnosed with Alzheimer's disease (AD) and non-demented control subjects. This Abeta pool was revealed after acid denaturation and chromatographic separation of plasma proteins followed by Abeta quantitation in the 4.5 kDa fractions by europium immunoassay. The mean values of Abeta42 in the AD and control individuals amounted to 236 ng/ml and 38 ng/ml, respectively. These Abeta values are on the average far higher than previously measured. Surprisingly, the circulating Abeta42 is about 16 times more abundant than Abeta40 in the AD population. Addition of Abeta to freshly drawn plasma demonstrated the rapid disappearance of Abeta epitopes, as detected by immunochemical techniques, suggesting either proteolytic degradation or Abeta sequestration. Incubation of Abeta with purified plasma proteins and lipoproteins rapidly decreases detectable levels of free Abeta suggesting epitope masking as the likely mechanism. The free and protein-bound Abetab in the circulation may represent a potential source for deposition in the cerebrovasculature and brain parenchyma of AD.

Elevated low-density lipoprotein in Alzheimer's disease correlates with brain abeta 1-42 levels.

Sera obtained in the immediate postmortem from 100 individuals, 64 neuropathologically diagnosed Alzheimer's disease (AD) cases and 36 nondemented controls, were analyzed for cholesterol, lipoproteins, apolipoproteins (Apo), and triglycerides. All individuals were ApoE genotyped, and the amounts of Abeta (N-40 and N-42) in cerebral cortex of AD and control subjects were determined. When compared to controls, AD individuals had significantly higher LDL cholesterol (P = 0.006), ApoB (P = 0.018), Abeta N-40 (P = 0.024) and Abeta N-42 (P < 0.001), and significantly lower HDL cholesterol (P = 0.040). There were positive correlations between the levels of serum total cholesterol (r = 0.359, P = 0.004), LDL cholesterol (r = 0.328, P = 0.008), and ApoB (r = 0.395, P = 0.001) to the amount of Abeta N-42 in AD brains, but not to Abeta N-40. These correlations were independent of ApoE genotype and were not seen in the control group. The present results suggest for the first time that elevated serum cholesterol, especially in the form of LDL, influences the expression of AD-related pathology.

RAGE-Abeta interactions in the pathophysiology of Alzheimer's disease.

RAGE is a cell surface molecule primarily identified for its capacity to bind advanced glycation end-products and amphoterin. Immunocytochemical studies demonstrated that in Alzheimer's Disease (AD) the expression of RAGE is elevated in neurons close to neuritic plaque beta-amyloid (Abeta) deposits and in the cells of Abeta containing vessels. Cross-linking of surface bound Abeta 1-40 to endothelial cells, yielded a band of 50 kDa identified as RAGE. Using the soluble extracellular domain of recombinant human RAGE, we found that Abeta binds to RAGE with a Kd = 57 +/- 14 nM, a value close to those found for mouse brain endothelial cells and rat cortical neurons. The interaction of Abeta with RAGE in neuronal, endothelial, and RAGE-transfected COS-1 cells induced oxidative stress, as assessed by the TBARS and MTT assays. ELISA demonstrated a 2.5 times increase of RAGE in AD over control brains. Activated microglia also showed elevated expression of RAGE. In the BV-2 microglial cell line, RAGE bound Abeta in dose dependent manner with a Kd of 25 +/- 9 nM. Soluble Abeta induced the migration of microglia along a concentration gradient, while immobilized Abeta arrested this migration. Abeta-RAGE interaction also activated NF-kappaB, resulting in neuronal up-regulation of macrophage-colony stimulating factor (M-CSF) which also induced microglial migration. Taken together, our data suggest that RAGE-Abeta interactions play an important role in the pathophysiology of Alzheimer's Disease.

Carbohydrate binding specificity of the neutrophil-activating protein of Helicobacter pylori.

The possible interaction of the neutrophil-activating protein of Helicobacter pylori with target cell glycoconjugates was investigated by the binding of 125I-labeled recombinant protein to glycosphingolipids from human neutrophils in solid phase assays. Thereby, a distinct binding of the neutrophil-activating protein to four bands in the acid glycosphingolipid fraction from human neutrophils was detected, whereas no binding to the non-acid glycosphingolipids or polyglycosyl ceramides from these cells was obtained. When using glycosphingolipids not present in the cell membrane of human neutrophils, it was found that the neutrophil-activating protein also bound to sulfated glycosphingolipids as sulfatide and sulfated gangliotetraosyl ceramide. Comparison of the binding preferences of the protein to reference glycosphingolipids from other sources suggested that in human granulocytes, the neutrophil-activating protein of H. pylori preferentially recognizes glycoconjugates with a terminally unsubstituted NeuAcalpha3Galbeta4GlcNAcbeta3Galbeta4GlcNAcbeta sequence.

Genetic evidence for host specificity in the adhesin-encoding genes hxaA of Helicobacter acinonyx, hnaA of H. nemestrinae and hpaA of H. pylori.

Gastric and non-gastric species of Helicobacter were examined for the presence of the adhesin-encoding gene, hpaA, from the human-associated gastric Helicobacter H. pylori (Hp), and for adhesin subunit protein HpaA. Amplification of a 375-bp internal DNA fragment of hpaA by PCR demonstrated the presence of the gene in Hp and in two closely related gastric Helicobacters, H. nemestrinae (Hn) and H. acinonyx (Hx), but not in the more distantly related H. felis (Hf) and H. mustelae (Hm). The non-gastric Helicobacters, H. canis (Hc), H. muridarum (Hr), H. fennelliae (He) and H. cinaedi (Hi), were all negative for hpaA. An immunoblot assay of water extracts with adhesin-specific antibody confirmed these results. The deduced amino acid (aa) sequences of Hp HpaA and Hn adhesin A (hereafter termed HnaA) are very similar, having identical receptor-binding motifs (rbm); also, the hemagglutination (HA) properties of Hn and Hp cells were indistinguishable. In contrast, the rbm of Hx adhesin A (hereafter termed HxaA), compared to that of Hp, contained a non-conservative aa substitution (Ile to Thr); also, there was variance in five consecutive aa from 10 to 14 residues upstream from the rbm. We conclude that these aa substitutions in HxaA are probably responsible for the difference in receptor recognition of this adhesin, as evidenced by the resistance of Hx HA to inhibition with N-acetylneuraminyl-alpha(2,3)-lactose. These results are consistent with the biological similarity between the natural host(s) of Hp and Hn; i.e., human and non-human primates, and the dissimilarity between these hosts and the feline host, the cheetah.

Restriction fragment length polymorphism in the adhesin gene hpaA of Helicobacter pylori.

OBJECTIVES: To assess the degree of restriction fragment length polymorphism (RFLP) in the Helicobacter pylori adhesin gene hpaA and to determine the molecular basis of RFLP in this gene. METHODS: A 375-bp, polymerase chain reaction-amplified internal sequence of hpaA, obtained from 50 different H. pylori isolates, was restricted with Sau3A and HinfI, individually. Polymerase chain reaction products representing different RFLP types were sequenced. RESULTS: Seven different polymorphic types were found in hpaA. Base substitutions at only four positions, two in Sau3A and two in HinfI sites, account for all of the RFLP types, including the size of the restriction fragments determined by gel electrophoresis. Most, 90%, of the base substitutions are very conservative, i.e., either do not change the encoded amino acid or substitute a homologous amino acid, and cause no detectable antigenic or functional effect on hpaA. The region of hpaA encoding the receptor-binding motif was particularly well conserved. CONCLUSIONS: RFLP typing of hpaA using Sau3A and HinfI provides an additional tool for comparing the genetic relatedness of H. pylori isolates collected during epidemiological and/or treatment studies.

Characterization of a Helicobacter pylori neutrophil-activating protein.

Helicobacter pylori-associated gastritis is mainly an inflammatory cell response. In earlier work we showed that activation of human neutrophils by a cell-free water extract of H. pylori is characterized by increased expression of neutrophil CD11b/CD18 and increased adhesiveness to endothelial cells. The work reported here indicates that the neutrophil-activating factor is a 150,000-molecular-weight protein (150K protein). Neutrophil proadhesive activity copurified with this protein, which is a polymer of identical 15K subunits. Specific antibody, prepared against the purified 15K subunit, neutralized the proadhesive activity of the pure protein and of water extracts obtained from different strains of H. pylori. The gene (napA) for this protein (termed HP-NAP, for H. pylori neutrophil-activating protein) was detected, by PCR amplification, in all of the H. pylori isolates tested; however, there was considerable strain variation in the level of expression of HP-NAP activity in vitro. HP-NAP could play an important role in the gastric inflammatory response to H. pylori infection.

Identification of four new prokaryotic bacterioferritins, from Helicobacter pylori, Anabaena variabilis, Bacillus subtilis and Treponema pallidum, by analysis of gene sequences.

The nucleotide (nt) sequence of the Helicobacter pylori (Hp) napA gene, encoding neutrophil-activating protein A (HPNAP) was determined. Alignment of this sequence with those of known bacterioferritins (Bfr) revealed sequence homology and conservation of a 7-amino-acid (aa) motif constituting the ferroxidase (Frx) center of Bfr in the HPNAP. The N-terminal aa sequence deduced from the iron-regulated mrgC gene of Bacillus subtilis [Chen et al., J. Bacteriol. 175 (1993) 5428-5437] is highly similar to that of HPNAP and contains five Frx center aa residues. The deduced aa sequences for proteins of unknown function in Treponema pallidum [Walfield et al., Infect. Immun. 57 (1989) 633-635] and in the cyanobacterium Anabaena variabilis [Sato, GenBank accession No. JU0384 (1991)] identify these two proteins as Bfr. Although the DNA-binding protein from starved cells of Escherichia coli [Almiron et al., Genes Dev. 6 (1992) 2646-2654] is clearly a HPNAP/Bfr homologue, a significant part of its Frx center is missing. It is unlikely that the intracellular function of HPNAP is related to its ability to activate neutrophils.