Alzheimer's-related peptide amyloid-β plays a conserved role in angiogenesis

Abstract

Alzheimer's disease research has been at an impasse in recent years with lingering questions about the involvement of Amyloid-β (Aβ). Early versions of the amyloid hypothesis considered Aβ something of an undesirable byproduct of APP processing that wreaks havoc on the human neocortex, yet evolutionary conservation--over three hundred million years--indicates this peptide plays an important biological role in survival and reproductive fitness. Here we describe how Aβ regulates blood vessel branching in tissues as varied as human umbilical vein and zebrafish hindbrain. High physiological concentrations of Aβ monomer induced angiogenesis by a conserved mechanism that blocks γ-secretase processing of a Notch intermediate, NEXT, and reduces the expression of downstream Notch target genes. Our findings allude to an integration of signaling pathways that utilize γ-secretase activity, which may have significant implications for our understanding of Alzheimer's pathogenesis vis-à-vis vascular changes that set the stage for ensuing neurodegeneration.

Figure 1. Aβ monomer induces tip cell formation in HUVEC.

A) HUVEC plated on a 3-D matrix spontaneously formed vessel-like tube structures after 4 h. Tip cells are indicated with arrows. B) Tip cell frequency was significantly higher in cultures treated with GSI (p = 4.6e−29). C) Cultures treated with 55 nM Aβ did not have significantly more tips than controls, but cultures treated with 225 nM Aβ (D) had significantly more tip cells (p = 2.3e−29). E) Treatment with reverse Aβ42-1 did not increase tip cell frequency. F) Histogram of mean tip cells in all cultures (*** = p<0.001 vs. control, one-way ANOVA with Bonferroni correction). Scale bar  = 250 µm, for A–E.

Figure 2. Aβ monomer causes HUVEC to accumulate NEXT and reduces mRNA levels of the downstream Notch targets Hes-1 and Hey-1.

A) Western blot of whole cell lysates from HUVEC cells treated with GSI (1 µM), Aβ (55 nM), Aβ (225 nM), reverse Aβ_42-1 (225 nM), or PBS (control). Note the levels of NEXT (∼90 kD) are significantly higher in cells treated with GSI and both concentrations of Aβ. The highest level of NEXT was detected in HUVEC treated with 225 nM Aβ. B) Histograms showing mRNA transcript levels for Hes-1 (B) and Hey-1 (C) in treated HUVEC cells, as detected by qPCR. Messenger RNA levels for GSI-treated HUVEC were significantly lower than PBS-treated control cells. A high (225 nM) concentration of Aβ monomer reduced both Hes-1 and Hey-1 mRNA levels; whereas, moderate (55 nM) Aβ and revAβ (225 nM) did not (n = 9, triplicate wells in 3 separate experiments, *** = p<0.001 vs. control, one-way ANOVA with Bonferroni correction).

Figure 3. Morphology and vascular imaging of GFP-expressing zebrafish embryos.

A) Projection image of a 3 dpf embryo using images captured by confocal microscopy. B, C) Comparable projection of a 3 dpf embryo treated with Aβ (25 µg/mL) and Rev Aβ, respectively. Boxes indicate the region analyzed for CtA branching, shown in detail in figure 4. Scale bar  = 100 µm, for A–C.

Figure 4. Analysis of CtA blood vessels in 3 dpf embryos.

A) Digitized image of the regions highlighted in figure 3 in a control (untreated) 3 dpf embryo. B) CtA branching in GSI-treated (10 µg/mL) embryo at 3 dpf. C) CtA branching in an embryo treated with 15 µg/mL Aβ. D) CtA branching in an embryo treated with 25 µg/mL Aβ at 3 dpf. E) RevAβ did not significantly increase CtA branching. F) Scatter plot of CtA branches in all embryos in each condition (open circle). In the case of multiple values at given value for each condition, the circles were slightly offset. The mean for each group is shown as a filled circle and error bars represent standard errors of the mean. A dashed line indicates the mean value of the controls for comparison across the other treatments. Asterisks indicate significant differences with the untreated control group (*** = p<0.001; * = p<0.05 one-way ANOVA with Bonferroni correction). Scale bar  = 25 µm, for A–E. (CtA – cerebral artery, CCtA – common CtA, PCeV – posterior cerebral vein, PHBC - posterior hindbrain channel).

Figure 5. Proposed mechanism for hypervascularization in response to high concentrations of Aβ.

High levels of Aβ (green helix) develop in the brain prior to the onset of Alzheimer’s pathology (top left). Longer forms of Aβ have more hydrophobic amino acid residues (gold) on remnant transmembrane domain, thereby increasing transient insertion (downward arrow) into the plasma membranes of nearby cells including endothelial cells (red). Elevated tissue levels of Aβ cause an accumulation of Aβ on the surface of the cell (several helices). Some peptide enters the endosome pathway (left). The amino termini of inserted Aβ peptides are identical to the Nicastrin-binding region of the APP intermediate C99 (not shown) and transiently bind to the γ-secretase complex (two sided arrow). Competition for Nicastrin binding and γ-secretase access reduces the processing of other substrates such as NEXT (shown). Interference with NEXT processing inhibits the production of NICD, which de-represses transcription factors such as Hes-1 and Hey-1. GSI has a similar effect. De-repression of NICD targets causes endothelial shaft cells to adopt tip cell morphology (upper right). Tip cells lead to formation of new blood vessels, and cycling of this process (looped arrow) over years and decades leads to hypervascularization (top right). This process may also occur with other extracellular cleavage (ex) products of γ-secretase (shown) creating feedback that tempers γ-secretase activity.