Hippocampal delivery of neurotrophic factor-α1/carboxypeptidase E gene prevents neurodegeneration, amyloidosis, memory loss in Alzheimer's Disease male mice

Abstract

Alzheimer's Disease (AD) is a prevalent neurodegenerative disease characterized by tau hyperphosphorylation, Aβ1-42 aggregation and cognitive dysfunction. Therapeutic agents directed at mitigating tau aggregation and clearing Aβ1-42, and delivery of growth factor genes (BDNF, FGF2), have ameliorated cognitive deficits, but these approaches did not prevent or stop AD progression. Here we report that viral-(AAV) delivery of Neurotrophic Factor-α1/Carboxypeptidase E (NF-α1/CPE) gene in hippocampus at an early age prevented later development of cognitive deficits as assessed by Morris water maze and novel object recognition assays, neurodegeneration, and tau hyperphosphorylation in male 3xTg-AD mice. Additionally, amyloid precursor protein (APP) expression was reduced to near non-AD levels, and insoluble Aβ1-42 was reduced significantly. Pro-survival proteins: mitochondrial Bcl2 and Serpina3g were increased; and mitophagy inhibitor Plin4 and pro-inflammatory protein Card14 were decreased in AAV-NF-α1/CPE treated versus untreated AD mice. Thus NF-α1/CPE gene therapy targets many regulatory components to prevent cognitive deficits in 3xTg-AD mice and has implications as a new therapy to prevent AD progression by promoting cell survival, inhibiting APP overexpression and tau hyperphosphorylation.

Fig. 1. CPE expression in hippocampus of 3xTg-AD mice after AAV-NF-α1/CPE injection.

A Hippocampal CPE expression in WT and 3xTg-AD mice at age of 3, 4.5 and 6.5 months. B Hippocampal CPE expression 1wk, 8wk, 16wk, and ~6 months after stereotaxic injection of AAV-NF-α1/CPE, or AAV-GFP into hippocampus of 2 months old 3xTg-AD mice. (Left panel) CPE expression was significantly increased in 3xTg+CPE in comparison with 3xTg+GFP group. n = 4–5 mice, t-test, **p < 0.01 compared with GFP control, values are mean ± SEM. (Right panel) Hippocampal CPE expression at ~8 months old after stereotaxic injection in nonTg+GFP, 3xTg+GFP and 3xTg+CPE at 2 months of age. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 16.47, p = 0.0004]. *p = 0.0006 for 3xTg+CPE compared with 3xTg+GFP, n = 5 per each genotype. Values are mean ± SEM. C Representative immunohistochemistry images showing CPE expression after stereotaxic injection of AAV-NF-α1/CPE or AAV-GFP in the hippocampus of nonTg+GFP, 3xTg+GFP, and 3xTg+CPE mice at ~8 months of age. Scale bar = 1 mm. Arrows show areas of increased CPE immunoreactivity in 3xTg-CPE mice. Inset: scale bar = 50 μm.

Fig. 2. AAV-NF-α1/CPE gene delivery in hippocampus prevents memory loss in 3xTg-AD mice.

A Experimental design for AAV injection followed by behavioral and pathological analyses of 3xTg-AD mice. 3xTg-AD mice received bilateral hippocampal injections of AAV-GFP or NF-α1/CPE at age 2 months and were evaluated by a series of behavioral tests at age of 7 months. Pathology of the hippocampus of the mice was examined at age ~8 months. B The effect of overexpression of NF-α1/CPE on spatial learning in Morris water maze. 3xTg+GFP mice displayed longer latency compared with nonTg+GFP on day 2, day 3, and day 4, p = 0.0048 for day 2, p = 0.0061 for day 3, p < 0.0001 for day 4, p = 0.0547 for day 5. 3xTg+CPE mice showed no significant difference in latency compared to 3xTg+GFP mice, except on day 4, p = 0.0103 and a trend on day 5. Two-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, for factor day [F_(4,170) = 27.86, p < 0.0001]. For factor genotype [F_(2,170) = 21.11, p < 0.0001]. Day × Genotype: [F_(8,170) = 0.7515, p = 0.6459]. n = 12 for nonTg+GFP, n = 11 for 3xTg+GFP, n = 14 for 3xTg+CPE. Values are mean ± SEM. C Overexpression of NF-α1/CPE prevented memory deficit of 3xTg-AD mice in Morris water maze test. 3xTg+GFP mice spent less time in the target area NE, and more time in non-target areas. Both nonTg+GFP and 3xTg+CPE mice displayed similar pattern of time in non-target quadrants and target quadrant. In NE target quadrant, 3xTg+CPE mice spent more time, similar to nonTg+GFP mice, than 3xTg+GFP mice. *p = 0.0025 for 3xTg+CPE compared with 3xTg+GFP. Two-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, for factor quadrant [F_(3,136) = 8.691, p < 0.0001]. For factor genotype [F_(2,136) = 5.942e−006, p > 0.9999]. Quadrant × Genotype: [F_(6,136) = 4.475, p = 0.0004]. n = 12 for nonTg+GFP, n = 11 for 3xTg+GFP, n = 14 for 3xTg+CPE. Values are mean ± SEM. D Effect of overexpression of NF-α1/CPE on cognitive function of 3xTg AD mice in novel object recognition test. ANOVA analysis followed by Tukey’s post-hoc multiple comparison test showed no significant differences across groups, but a trend showing memory improvement in 3xTg-AD mice. n = 10 for nonTg+GFP, n = 8 for 3xTg+GFP, n = 12 for 3xTg+CPE. Values are mean ± SEM.

Fig. 3. AAV-NF-α1/CPE gene delivery prevents hippocampal neurodegeneration in 3xTg-AD mice.

A Representative sections of MAP2 immunofluorescence staining of hippocampal CA1 region of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. Magnification 10x, scale bar =50 μm., Inset: magnification 20x, scale bar =50 μm. n = 6 mice per genotype. B Quantification of MAP2 intensity in CA1 region of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. MAP2 intensity was decreased in 3xTg+GFP mice in comparison with nonTg+GFP, *p = 0.0002; 3xTg+CPE increased MAP2 intensity significantly in comparison with 3xTg+GFP, ^#p = 0.0129. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,15) = 15.45, p = 0.0002]. Values are mean ± SEM. 4 sections per mouse, n = 6 mice per genotype. nonTg+GFP made = 100% as control. C Representative sections of GFAP immunofluorescence staining of hippocampal CA1 of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. Magnification 10x, scale bar = 50 μm. n = 6 mice per genotype. D Quantification of GFAP positive cells in CA1 region of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test did not show any significant differences. Values are mean ± SEM. 4 sections per mouse, n = 6 mice per genotype. E Representative Western blot and quantification of GFAP expression in hippocampus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test did not show any significant differences. n = 5 mice per genotype. Values are mean ± SEM. F Representative Western blot and quantification of CD11b/c expression in hippocampus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. CD11b/c was increased in 3xTg+GFP in comparison with nonTg+GFP, *p = 0.0175; but not significantly different from 3xTg+CPE. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 5.401, p = 0.0212]. n = 5 mice per genotype. Values are mean ± SEM. G Representative sections of CD68 immunostaining for activated microglia in hippocampal CA1 of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. Magnification 20x, scale bar = 50 μm. n = 6 mice per genotype. H Quantification of CD68 positive cells in hippocampal CA1 of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. CD68 positive cells were significantly increased in 3xTg+GFP in comparison with nonTg+GFP, *p = 0.0016. Overexpression of NF-α1/CPE in 3xTg mice reduced activated microglia ^#p = 0.0323. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,15) = 9.728, p = 0.002]. 4 sections per mouse, n = 6 mice per genotype. Values are mean ± SEM. I Representative sections of doublecortin (DCX) stained immature neurons in subgranular zone of dentate gyrus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. Magnification 20x, scale bar = 50 μm. n = 6 mice per genotype. J Quantification of DCX positive cells in subgranular zone of dentate gyrus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. DCX positive cells were significantly decreased in 3xTg+GFP in comparison with nonTg+GFP, *p = 0.015. Overexpression of NF-α1/CPE in 3xTg+CPE did not mitigate this deficit in neurogenesis p = 0.9947. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,15) = 7.063, p = 0.0069]. 4 sections per mouse, n = 6 mice per genotype. Values are mean ± SEM.

Fig. 4. AAV-NF-α1/CPE reduces tau phosphorylation and APP and β-amyloid42 expression.

A Representative Western blot and quantification of phosphorylated Tau expression in hippocampus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. Phosphorylated tau at Ser396 was increased in hippocampus of 3xTg+GFP in comparison with nonTg+GFP, *p = 0.0075; overexpression of NF-α1/CPE in 3xTg+CPE mice significantly reduced phosphorylated tau, ^#p = 0.0425. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 7.497, p = 0.0077]. n = 5 mice per genotype. Values are mean ± SEM. B (i) Representative Western blot and quantification of mouse + human β-amyloid precursor protein (m + h APP) expression in hippocampus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age ~8 months. β-amyloid precursor protein was significantly increased in 3xTg+GFP, *p = 0.0039; but decreased with overexpression of NF-α1/CPE in 3xTg+CPE mice, ^#p = 0.0137. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 9.622, p = 0.0032]. n = 5 mice per genotype. Values are mean ± SEM. (ii) Representative Western blot of human β-amyloid precursor protein (hAPP) expression in hippocampus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice and quantification of hAPP of 3xTg+GFP and 3xTg+CPE at age ~8 months. β-amyloid precursor was decreased with overexpression of NF-α1/CPE in 3xTg+CPE mice, *p = 0.0001. t test, n = 3 mice per genotype. Values are mean ± SEM. C Representative image of β-amyloid precursor expression in hippocampus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice at age of ~8 months. Magnification 2x, scale bar = 1 mm. Inset: β-amyloid precursor expression in CA1 area, magnification 20x, scale bar = 50 μm. D Human β-amyloid40 in hippocampus of 3xTg+GFP and 3xTg+CPE mice at age ~8 months. Both soluble and insoluble β-amyloid40 showed no significant difference between 3xTg +GFP versus 3xTg+CPE mice. Student t test. n = 5 mice per genotype. Values are the mean ± SEM. E Human β-amyloid42 in hippocampus of 3xTg+GFP and 3xTg+CPE mice at age ~8 months. Soluble β-amyloid42 showed no significant difference between 3xTg+GFP and 3xTg+CPE. Insoluble β-amyloid42 was decreased in 3xTg+CPE compared to 3xTg+GFP mice, Student t test, *p = 0.0194, n = 5 mice per genotype. Values are the mean ± SEM.

Fig. 5. AAV-NF-α1/CPE gene delivery on hippocampal Bcl2, Bax, Plin4, Card14, BDNF and Serpina3g expression in 3xTg-AD mice.

A Representative Western blot and bar graph showing Bcl2 was increased in hippocampus of 3xTg+CPE in comparison with 3xTg+GFP, *p = 0.0003. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 19.22, p = 0.0002]. n = 5 per genotype. Values are the mean ± SEM. B Representative Western blot and bar graph showing Bax was decreased in the hippocampus of 3xTg+CPE in comparison with 3xTg+GFP, *p = 0.0179. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 5.275, p = 0.0227]. n = 5 per genotype. Values are the mean ± SEM. C Representative Western blot and bar graph showing Plin4 was decreased in the hippocampus of 3xTg+CPE in comparison with 3xTg+GFP mice, *p = 0.0219. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 5.655, p = 0.0186]. n = 5 per genotype. The values are the mean ± SEM. D Representative Western blot and bar graph showing Card14 was decreased in the hippocampus of 3xTg+CPE in comparison with 3xTg+GFP mice, *p = 0.045. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 6.391, p = 0.0129]. n = 5 per genotype. The values are the mean ± SEM. E Representative Western blot and bar graph showing quantification of pro-BDNF and mature (m) BDNF expression in the hippocampus of nonTg+GFP, 3xTg+GFP and 3xTg+CPE mice. No significant differences were observed in mBDNF and proBDNF among the three groups. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test. n = 4 per genotype. The values are the mean ± SEM. F Representative Western blot and bar graph showing Serpina3g protein was increased in the hippocampus of 3xTg+CPE in comparison with 3xTg+GFP mice. *p = 0.032. One-way ANOVA analysis followed by Tukey’s post-hoc multiple comparison test, [F_(2,12) = 5.315, p = 0.0222]. n = 5 per genotype. The values are the mean ± SEM.

Fig. 6. Factors regulating neuronal survival and APP mRNA transcription induced by NF-α1/CPE.

A Serpina3g protects HT22 hippocampal neurons against H₂O₂-induced cytotoxicity. Bar graphs (stippled bars) showing that cell cytotoxicity decreased in HT22 cells expressing serpina3g compared to cells expressing vector alone. Open bars showing cytotoxicity level (made = 1) in the control cells expressing serpina3g or vector, but not treated with H₂O₂. Values shown are the mean ± SEM, p < 0.05, t-test, N = 3. B Reduction of APP mRNA levels in 3xTg-AD mice and cells after NF-α1/CPE gene delivery. Left panel: Bar graphs showing relative human and mouse APP mRNA levels in 3xTg-AD mouse hippocampus after delivery of AAV-NF-α1/CPE or AAV-GFP. Values shown are the mean ± SD, n = 2 mice per genotype. Middle panel: HT22^cpe−/− cells were transiently transfected with CPE or vector for 48 h and then harvested for qRT- PCR assay. Bar graphs showing relative mAPP mRNA level. Values shown are the mean ± SEM, N = 3. Right panel: HEK293 cells were transiently transfected with CPE and vector for 48 h and then harvested for qRT- PCR assay. Bar graphs showing relative hAPP mRNA level. Values shown are the mean ± SEM, N = 3. C. Reduction of transcription factor mRNAs in HT22^cpe−/− cells after CPE transfection. HT22^cpe−/− cells were transiently transfected with CPE or vector for 48 h and then harvested for qRT- PCR assay. Bar graphs showing relative mRNA levels of 4 transcription factors were reduced with CPE transfection. Values shown are the mean ± SEM, N = 3. *p < 0.05, **p < 0.01, ***p < 0.001. D Schematic showing hippocampal injection of AAV- NF-α1/CPE in 3xTg-AD mice down-regulates expression of APP in pyramidal neurons via attenuating transcription factors Sp1 and Hsf-1 expression. These transcription factors bind to the promoter of APP gene to promote transcription. Red arrows indicate APP positive neurons.