Alzheimer Disease–Associated Peptide, Amyloid β40, Inhibits Vascular Regeneration With Induction of Endothelial Autophagy

We performed all procedures according to protocols approved by Animal Care Committees of Gifu University Graduate School of Medicine and Osaka University Graduate School of Medicine.

Aβ40 (Aβ1-40), Aβ40-1, and Aβ42 were obtained from Peptide Institute. Rapamycin and 3-methyladenin (3-MA) were purchased from Sigma. Please see methods in detail regarding Aβ peptides in the supplemental materials (available online at http://atvb.ahajournals.org).

HBECs, human aortic ECs, human umbilical vein ECs, and VSMCs were obtained from Applied Cell Biology Research Institute. We cultured ECs in Modified MCDB131 medium supplemented with 2% FBS and an endothelial growth supplement kit (Clonetics), so-called EGM, and used them for experiments between passages 3 to 5. Vascular smooth muscle cells were cultured in DMEM supplemented with 10% FBS. Cells that had reached 80% confluence were used for the following experiments.

ECs or VSMCs cultured on 12-well culture plates were preincubated in Modified MCDB131 medium with 5% FBS for 24 hours, and stimulated with Aβ peptides for 72 hours. The cell number in each group was manually counted with a hemocytometer. Proliferative activity of vascular cells was also confirmed by BrdU-ELISA according to the manufacturer’s instructions (Roche).

HBECs were seeded on 6-well culture plates coated with Matrigel Basement Membrane Matrix (BD Bioscience), and treated with Aβ40 or Aβ40-1 for 18 hours. Tube-forming ECs were photographed, and their total lengths were quantified with imaging software (Kurabo). All groups were studied in at least 4 independent experiments.

A migration assay was performed using a modified Boyden chamber (BD Bioscience). Briefly, serum-free medium with vascular endothelial growth factor (VEGF; 50 ng/mL, Peprotech Inc), Aβ40 (0.5 μmol/L), Aβ40-1 (0.5 μmol/L), or 10% FBS was put in the lower compartment of the chamber, seeded fluorescence (Calcein-AM, Molecular Probes)-labeled HBECs in the upper chamber (5×10⁴ cells/50 μL), and they were incubated for 6 hours. Transmigrated cells were counted in 4 random high-power fields (×400). All groups were studied in at least 4 independent experiments.

HBECs cultured on 12-well culture plates were stimulated with or without Aβ peptides (0.5 μmol/L) for 18 hours, and fixed with 1% paraformaldehyde (PFA) in 0.1 mol/L PBS (pH 7.4) for 20 minutes. Cells were stained with a primary antibody against cleaved caspase-3 (Cell Signaling) and a species-specific Alexa Fluor secondary antibody (Molecular Probes Inc). Advanced stages apoptotic ECs were stained with Hoechst 33342 (Sigma) after 48 hours of treatment with Aβ peptides. The number of apoptotic cells was counted and normalized with reference to the total number of cells in each field.

HBECs cultured on 12-well culture plates were stimulated with or without Aβ peptides (0.5 μmol/L) for 24 hours. Released lactate dehydrogenase (LDH) in the cell culture supernatant was measured using the enzymatic reaction with an LDH cytotoxicity detection kit (Takara).

Please see methods in detail in the supplemental materials.

Please see methods in detail in the supplemental materials.

Please see methods in detail in the supplemental materials.

HBECs were cultured with or without Aβ peptides (0.5 μmol/L) for 18 hours. Cell sections in epon were counterstained with uranyl acetate and lead citrate and examined with a Hitachi H7600 transmission electron microscopy (TEM). Autophagic HBECs were also visualized by microtubule-associated protein light-chain3 (LC3) staining of the transduced GFP-LC3 construct as described previously.⁹ Please see methods in detail in the supplemental materials.

HBECs cultured in 6-well culture plates were transiently transfected with 1 μg constitutive active Akt constructs or pUSE-Amp Vector (control constructs without the Akt gene, UBI), using TransIT-LT1 reagent 48 hours before Aβ treatment. These cells were reseeded on 12-well culture plates 24 hours before Aβ treatment, and incubated with EGM for additional 24 hours. Then, cells were treated with Aβ peptides for 48 hours. Proliferative activity was examined as described above. In some case, HBECs were transfected with both active Akt construct and pEGFP-LC3. Transfection efficiency (40% to 45%) and viability of cells (85% to 90%) were initially tested at 48 hours after transfection.

Eight-week-old male GFP-LC3 transgenic mice were used as marker of autophagy induction.¹⁰ Briefly, brain tissues from age-matched GFP-LC3 transgenic mice were vertically cut into 2-mm slices in ice-cold PBSG (PBS with 0.6% glucose). Tissue culture of these slices containing hippocampal lesions was performed with EGM and Aβ40 peptides for 48 hours. Frozen sections (6 μm) from these brain slices were postfixed with 4% PFA in 0.1 mol/L PBS for 20 minutes and stained with rat antimouse platelet endothelial cell adhesion molecule-1 (PECAM-1) primary antibody (Pharmingen), Alexa 488-conjugated goat anti-rat secondary antibody (Molecular Probe), and DAPI (Molecular Probe). Vascular density (PECAM-1) and autophagy induction (GFP-LC3) in hippocampal CA1 lesions were evaluated with a fluorescence microscope.

Please see methods in detail in the supplemental materials.

Each result was obtained in at least 4 separate experiments. All values are expressed as mean±SEM. ANOVA was used to determine the significance of differences in multiple comparisons. P<0.05 was considered statistically significant.

Microvascular formation initially requires several steps, including endothelial proliferation, capillary tube formation, and migration.⁷ As shown in Figure 1A, Aβ40 (Aβ1-40) significantly reduced proliferation of HBECs as compared to control, whereas Aβ40-1, as a control peptide with a reversed amino acid sequence, did not affect them. Aβ40 also inhibited the proliferation of human aortic ECs and human umbilical vein ECs, but did not affect proliferation of vascular smooth muscle cells (VSMCs; Figure 1A and data not shown). In contrast, Aβ42 inhibited proliferation of both human vascular ECs and VSMCs, suggesting the possibility that Aβ40 but not Aβ42 has endothelium specific effects. We also obtained consistent results that Aβ40 impaired endothelial proliferation by using a BrdU incorporation assay (data not shown). Notably, Aβ40 did not increase lactate dehydrogenase (LDH) release from HBECs after 24 hours of treatment, whereas Aβ42 markedly increased LDH release, suggesting that Aβ40 has low toxic effects on HBECs as compared to Aβ42 (Figure 1B and 1C). In addition, we showed that Aβ40 significantly reduced capillary tube formation of HBECs after 18 hours of treatment (Figure 1D). Moreover, vascular endothelial growth factor (VEGF) stimulated the migration of HBECs, whereas stimulation by VEGF was significantly reduced by Aβ40 (Figure 1E). These results indicate that Aβ40 impairs in vitro vascular regeneration through its effects on ECs.

We next addressed the question of whether Aβ40 might affect endothelial regeneration in the setting of AD. As about 50% of patients with hippocampal hypoperfusion converted to preclinical stage AD with mild cognitive impairment,^8,11 we first examined the effect of Aβ40 on endothelial regrowth in hippocampal lesions. Using our ex vivo model of hippocampal revascularization (supplemental Figure IA), we found newly formed vessels positive for PECAM-1 around hippocampal explants on day 3 (Figure 2A). In contrast, however, Aβ40 significantly impaired vessel sprouting from explants (Figure 2A and B).

We also investigated whether Aβ40 might affect the healing process of the endothelium after vascular injury. We prepared rat aortic tissue on culture dishes, made a scratched wound on the endothelial surface, and evaluated ex vivo reendothelialization after 7 days of stimulation with endothelial growth medium (EGM) and Aβ40 (supplemental Figure IB). In control vessels without Aβ40, reendothelialization was observed in around 84% of scratched endothelium, which stained negatively with Evans blue dye. Similarly, 80% of the scratched lesions were reendothelialized in vessels with Aβ 40-1. In contrast, only about 24% of scratched lesions were reendothelialized in vessels with Aβ40, indicating that Aβ40 significantly inhibited the reendothelialization after vascular injury (Figure 2C and 2D). These results were confirmed by staining for endothelium specific uptake of DiI-labeled acetylated LDL (data not shown).

Recent reports indicate that endothelial progenitor cells (EPCs) in the adult peripheral circulation, in part, contribute to revascularization in hindlimb ischemia and ischemic brain disease.^12,13 As shown in Figure 2E, differentiated EPCs from isolated mononuclear cells were characterized by adherent cells double-positive for DiI acetylated-LDL uptake (red) and lectin binding (green). Interestingly, incubation of mononuclear cell with Aβ40 significantly decreased the number of EPC by about 65% compared to that in the control group, whereas Aβ40-1 did not affect the number of EPC (Figure 2E and 2F). Together, these results indicate the possibility that Aβ40 suppresses endothelial regeneration in the setting of AD.

To investigate whether Aβ40 induces phenotypic alterations in neurovascular ECs, we initially performed activated-caspase-3 staining after 18 hours of treatment with Aβ peptides (0.5 μmol/L) in cultured HBECs. Immunofluorescent staining (green) revealed that Aβ40 slightly increased the number of activated-caspase-3–positive cells as compared to control cells (8% versus 2% of total cells, respectively), indicating that Aβ40 has a weak effect on the induction of apoptosis, a well-known form of programmed cell death (PCD), in neurovascular ECs. In contrast, Aβ42 significantly increased the number of apoptotic cells (22% of total cells) (Figure 3A and 3B).

We next investigated whether Aβ40 induces another process of PCD with formation of autophagosomes and autolysosomes, so-called “self-digesting” autophagy.^14,15 As shown in Figure 3C, the expression of microtubule-associated protein light-chain3 (LC3), as a quantification marker of the early stage of autophagic formation,⁹ was significantly higher in HBECs treated with Aβ40 than in control, indicating the induction of molecular alterations for autophagy. Strikingly, the present study demonstrated by transmission electron microscopy (TEM) numerous double-membrane vesicles containing cytoplasmic organelles characteristic of autophagy in about 40% of HBECs after 18 hours of Aβ40 treatment, whereas few autophagic cells were observed in control cells without Aβ40 or with Aβ 40-1 (Figure 3D, upper and middle panels, and data not shown). Alternatively, we transduced a GFP-LC3 construct into HBECs,⁹ and monitored GFP-LC3 spots in the cell cytosol. Immunofluorescent images showed consistent results that Aβ40 induced endothelial autophagy with enhanced LC3 spots (Figure 3D, lower right panel), whereas fewer enhanced LC3 spots were observed in the cytoplasm of GFP-LC3–transfected control cells without Aβ40 and of GFP-transfected cells with Aβ40 (Figure 3D, lower left panel, and data not shown). Notably, the number of autophagic cells with enhanced LC3 spots was markedly lower in the group with Aβ42 than in group with Aβ40 (20% versus 38%, respectively; Figure 3E).

To further investigate whether inhibition of endothelial autophagy attenuates the impairment of endothelial proliferation by Aβ40, we evaluated proliferation of HBECs after treatment with Aβ40 and 3-methyladenin (3-MA), a widely used inhibitor of autophagy as well as an inhibitor of class3 PI3K. We found that treatment of HBECs with 3-MA attenuated the impairment of endothelial proliferation by Aβ40 (Figure 3F), suggesting the possibility that inhibition of endothelial autophagy restores neurovascular regrowth, and that class 3 PI3K is in part involved in Aβ40-induced endothelial autophagy. Together, these results indicate that neurovascular ECs exposed to Aβ40 undergo autophagy rather than apoptosis.

Although Akt signaling is involved in cell survival or PCD of a wide range of cell types,¹⁶ recent reports indicated that signaling control of autophagy overlaps with class 1 PI3K-Akt signaling.^17,18 As Akt signaling is also involved in VEGF-induced vascular formation and endothelial cell survival,¹⁹ the effects of Aβ40 on the Akt signaling cascade were examined in HBECs. As shown in supplemental Figure II, Aβ40 significantly suppressed the phosphorylation of Akt (Ser473-phosphorylated) induced by VEGF (50 ng/mL). Aβ40 also suppressed the active level of endothelial nitric oxide synthase (eNOS; Ser1177-phosphorylated), whose activation led to NO production responsible for vascular remodeling and vascular formation.^20,21 These results suggest that Aβ40 inhibits the Akt signaling cascade in human neurovascular ECs.

To confirm whether enhancing the active Akt level may attenuate Aβ40-induced impairment of neurovascular regrowth, we transduced HBECs with the constitutive active Akt construct and evaluated their proliferative activity. Although treatment with Aβ40 (0.5 μmol/L) impaired proliferation of HBECs, transfection of the active Akt construct significantly attenuated the Aβ40-induced impairment of endothelial proliferation (Figure 4B). In vitro capillary tube formation assay showed consistent results that transfection of the active Akt construct enhanced tube formation of HBECs after treatment with Aβ40 (data no shown). To test whether Akt activation may regulate Aβ40-induced endothelial autophagy, HBECs transduced with the GFP-LC3 construct or active Akt construct were monitored after treatment with Aβ40 by fluorescence microscope. Notably, transfection of the active Akt construct significantly reduced the number of autophagic ECs with enhanced LC3 spots in their cytosol (Figure 4C and 4D). These data suggest that class 1 PI3K-Akt signaling in human neurovascular ECs is in part involved in endothelial autophagy and impairment of endothelial regrowth by Aβ40.

To further investigate whether Aβ40 in hippocampal lesions might induce endothelial autophagy, we performed tissue culture of murine hippocampal slices (2 mm) from GFP-LC3 transgenic mice, as a useful tool for detecting in vivo autophagy,¹⁰ and stimulated them with EGM and Aβ peptides for 48 hours. Frozen sections (6 μm) from these slices were used for fluorescence detection of LC3 (green) and PECAM-1 (red) in hippocampal CA1 lesions. Notably, both Aβ40 and Aβ42 significantly increased the number of autophagic cells with enhanced LC3 spots in CA1 lesions as compared to control with EGM stimulation alone (Figure 5A, GFP-LC3, and Figure 5B). In addition, as we expected from the results above (Figures 1 and 2A), both Aβ40 and Aβ42 also reduced vessel density in CA1 lesions (Figure 5A, PECAM-1). Strikingly, double-fluorescence images of LC3 and PECAM-1 revealed that Aβ40 increased the number of autophagic vessels in CA1 lesions (Figure 5A, merged, enlarged, and Figure 5C), indicating that Aβ40 induces endothelial autophagy in the hippocampus. Unexpectedly, Aβ42 induced fewer numbers of autophagic vessels in hippocampal lesions than did Aβ40, although Aβ42 showed effects on autophagy induction in nonvascular neural cells (Figure 5A, merged and enlarged, and Figure 5C). Moreover, by using tissue culture of aorta from GFP-LC3 mice, we obtained consistent results that Aβ40 rather than Aβ42 had stronger effects on autophagy induction in vascular tissue endothelium (supplemental Figure III). These results suggest that Aβ40 induces endothelial autophagy in the hippocampus in the setting of AD, and this induction associates with Aβ40-induced impairment of vascular regrowth.