A conformation-specific antibody against oligomeric β-amyloid restores neuronal integrity in a mouse model of Alzheimer's disease

Abstract

Conformationally distinct aggregates of the amyloid β (Aβ) peptide accumulate in brains of patients with Alzheimer's disease (AD), but the roles of the different aggregates in disease progression are not clear. We previously isolated two single-chain variable domain antibody fragments (scFvs), C6T and A4, that selectively bind different toxic conformational variants of oligomeric Aβ. Here, we utilize these scFvs to localize the presence of these Aβ variants in human AD brain and to demonstrate their potential as therapeutic agents for treating AD. Both A4 and C6T label oligomeric Aβ in extracellular amyloid plaques, whereas C6T also labels intracellular oligomeric Aβ in human AD brain tissue and in an AD mouse model. For therapeutic studies, the A4 and C6T scFvs were expressed in the AD mice by viral infection of liver cells. The scFvs were administered at 2 months of age, and mice sacrificed at 9 months. The scFvs contained a peptide tag to facilitate transport across the blood brain barrier. While treatment with C6T only slightly decreased Aβ deposits and plaque-associated inflammation, it restored neuronal integrity to WT levels, significantly promoted growth of new neurons, and impressively rescued survival rates to WT levels. Treatment with A4 on the other hand significantly decreased Aβ deposits but did not significantly decrease neuroinflammation or promote neuronal integrity, neurogenesis, or survival rate. These results suggest that the specific Aβ conformation targeted in therapeutic applications greatly affects the outcome, and the location of the targeted Aβ variants may also play a critical factor.

Keywords: Alzheimer's disease; neuron; oligomeric beta amyloid; single-chain antibody; transgenic mice.

Figure 1

A4 and C6T immunopositive structures in postmortem human brain tissue. Aβ aggregates were stained using an antibody against Aβ_1–17 (clone 6E10, red). Little staining of A4 and C6T as well as 6E10 was observed in the human brains with nondementia (ND). Counter staining by 4',6-diamidino-2-phenylindole. A, the oligomeric Aβ variants recognized by A4 scFv (green), together with Aβ aggregates stained by antibody 6E10 (red, arrowhead), in AD brain tissue. B, the oligomeric Aβ variants recognized by C6T scFv (green), together with Aβ plaques shown by 6E10 (red, arrowhead), as well as C6T-positive staining around the nuclei (arrow), in AD brain tissue. C, the extracellular and intracellular aggregates that C6T recognized were observed (green) in AD but not ND brain tissue. A control using secondary antibody against cMyc without the primary scFv is also shown. The neuronal cells were stained by specific antibody against MAP2 (red, arrow). The bars represent 20 μm. MAP2, microtubule-associated protein 2.

Figure 2

Localization and levels of A4- and C6T-recognized Aβ structures in brain tissue from APP/PS1 mice. Aβ accumulation was stained by a specific antibody against Aβ_1–17 (6E10, red). Counter staining by 4',6-diamidino-2-phenylindole. A, staining with purified A4 scFv in the APP/PS1 mice brain tissue. B, the staining of C6T-recognized oligomeric Aβ (green) and Aβ aggregates (6E10, red) in WT-GFP, Tg-GFP, Tg-A4, and Tg-C6T mouse brain tissue and negative staining of A4 and C6T, as well as 6E10 in the brains of WT mice. The bars represent 50 μm. C, levels of oligomeric Aβ reactive with A4 scFv following treatment with rAAV-GFP, rAAV-A4, and rAAV-C6T. D, levels of oligomeric Aβ reactive with C6T scFv following treatment with rAAV-GFP, rAAV-A4, and rAAV-C6T. All data were expressed as fold increase relative to WT-GFP mice in the cortex and hippocampus. The quantitative assay was performed with the mice of WT-GFP (n = 3), Tg-GFP (n = 10), Tg-A4 (n = 7), and Tg-C6T (n = 9). ^###p < 0.001 to age-matched WT-GFP mice. ∗p < 0.05 and ∗∗∗p < 0.001 to littermate Tg-GFP vehicle mice, and ^&p < 0.05 to littermate Tg-A4 mice.

Figure 3

Aβ staining in APP/PS1 mouse brain tissue.A, Aβ deposits were stained using anti-Aβ antibody 6E10 staining in the cortex (top) and the hippocampus (bottom) following the treatment with rAAV-GFP, rAAV-A4, and rAAV-C6T. B and C, the number of 6E10-positive deposits was counted and averaged per section in the cortex (B) and the hippocampus (C) of mice treated with rAAV-GFP, rAAV-A4, and rAAV-C6T. D, fibrillar deposits were verified by Congo red staining in the cortex (top) and the hippocampus (bottom). E and F, the number of fibrillar deposits was counted and averaged per section in the cortex (E) and the hippocampus (F). Nuclei were counterstained with hematoxylin. WT-GFP (n = 3), Tg-GFP (n = 15), Tg-A4 (n = 7), and Tg-C6T (n = 9). ∗p < 0.05 and ∗∗p < 0.01 to Tg-GFP mice and ^&p < 0.05 to littermate Tg-A4 mice.

Figure 4

Gliosis in response to Aβ in APP/PS1 mouse brain tissue.A, the immune cell microglia were immunostained by specific antibody against Iba1 (green). Aβ deposits were visualized by antibody against Aβ_1–17 (clone: 6E10, red). Counter stain by 4',6-diamidino-2-phenylindole (blue). B, positive staining of microglia was expressed as a percentage of total area. C, reactive astrocytes were visualized using an antibody against GFAP. The fibrillar deposits were visualized by Congo red staining. Cell nuclei were counterstained by hematoxylin. D, positive staining of reactive astrocytes was expressed as a percentage of total area. WT-GFP (n = 3), Tg-GFP (n = 5), Tg-A4 (n = 5), and Tg-C6T (n = 5). ^###p < 0.001 to age-matched WT-GFP mice, ∗p < 0.05 and ∗∗∗p < 0.001 to littermate Tg-GFP group, and ^&p < 0.05 to littermate Tg-A4 group. GFAP, glial fibrillary acidic protein.

Figure 5

Changes in dendritic spines and synapses in APP/PS1 mouse brain tissue.A, neuronal dendrites were immunostained using an antibody against MAP2 (green) in the cortex. B, the dendritic area of MAP2-positive staining was expressed as a percentage of total captured cortex regions. C, neuronal dendrites were visualized using an antibody against MAP2 (green) in CA3 region of the hippocampus. Aberrant dendritic arrangements are indicated (arrow). D, the dendritic area of MAP2-positive staining was expressed as a percentage of total captured hippocampal regions. E, the synapses were labeled with a specific antibody against synaptophysin (SYP, green). F, the area of SYP-positive staining was expressed as a percentage of total captured regions. In the images, Aβ deposits were labeled by antibody 6E10 (red). Counterstaining by 4',6-diamidino-2-phenylindole (blue). WT-GFP (n = 3), Tg-GFP (n = 5), Tg-A4 (n = 5), and Tg-C6T (n = 5). Statistical analysis was valued as ^###p < 0.001 to age-matched WT-GFP mice, ∗∗∗p < 0.001 to littermate Tg-GFP vehicle group, as well as ^&&p < 0.01 and ^&&&p < 0.001 to littermate Tg-A4 group. MAP2, microtubule-associated protein 2.

Figure 6

Changes in hippocampal neurogenesis and survival rate.A, the newborn immature neurons in the subgranular zone of hippocampus were visualized with an antibody against DCX (red). Cell nuclei were counterstained by 4',6-diamidino-2-phenylindole (blue). B, the number of DCX-positive cells was counted and averaged per sections. ^##p < 0.01 to age-matched WT-GFP mice, ∗p < 0.05 and ∗∗∗p < 0.001 to littermate Tg-GFP vehicle group, and ^&&&p < 0.001 to Tg-A4 mice. C, the survival rate of the APP/PS1 mice receiving rAAV-GFP, rAAV-C6T, and rAAV-A4 with time, compared with that of age-matched WT-GFP mice as controls. At age of 2 months (the time point of viral injection), the numbers of mice injected were as follows: WT-GFP (n = 15), Tg-GFP (n = 20), Tg-A4 (n = 12), and Tg-C6T (n = 10). Seven months later, the number of mice that survived were 15, 12, 7, and 9, respectively. DCX, doublecortin.