Ascidians: an invertebrate chordate model to study Alzheimer's disease pathogenesis

Abstract

Here we present the ascidian Ciona intestinalis as an alternative invertebrate system to study Alzheimer's disease (AD) pathogenesis. Through the use of AD animal models, researchers often attempt to reproduce various aspects of the disease, particularly the coordinated processing of the amyloid precursor protein (APP) by alpha-, beta- and gamma-secretases to generate amyloid beta (Abeta)-containing plaques. Recently, Drosophila and C. elegans AD models have been developed, exploiting the relative simplicity of these invertebrate systems, but they lack a functional Abeta sequence and a beta-secretase ortholog, thus complicating efforts to examine APP processing in vivo. We propose that the ascidian is a more appropriate invertebrate AD model owing to their phylogenetic relationship with humans. This is supported by bioinformatic analyses, which indicate that the ascidian genome contains orthologs of all AD-relevant genes. We report that transgenic ascidian larvae can properly process human APP(695) to generate Abeta peptides. Furthermore, Abeta can rapidly aggregate to form amyloid-like plaques, and plaque deposition is significantly increased in larvae expressing a human APP(695) variant associated with familial Alzheimer's disease. We also demonstrate that nervous system-specific Abeta expression alters normal larval behavior during attachment. Importantly, plaque formation and alterations in behavior are not only observed within 24 hours post-fertilization, but anti-amyloid drug treatment improves these AD-like pathologies. This ascidian model for AD provides a powerful and rapid system to study APP processing, Abeta plaque formation and behavioral alterations, and could aid in identifying factors that modulate amyloid deposition and the associated disruption of normal cellular function and behaviors.

Fig. 1.

APP processing and Aβ plaque formation. Amyloidogenic processing of APP occurs as a result of the β-secretase activity of BACE. This results in the production of the sAPPβ and the β-CTF. Intramembranous cleavage of the β-CTF by the γ-secretase enzymatic complex results in the production of Aβ which can subsequently aggregate to form plaques. Non-amyloidogenic processing is a result of APP cleavage by α-secretase which takes place within the Aβ sequence, precluding its formation, and is considered to be neuroprotective.

Fig. 2.

Human APP₆₉₅ is properly processed in transgenic ascidian embryos. (A) Representative transgenic ascidian embryo ubiquitously expressing hAPP₆₉₅ fused, by the C-terminal end, to CFP. The cis-regulatory region of the ascidian ef1a gene was used to drive ubiquitous expression of hAPP₆₉₅. Images for DIC (differential interference contrast, top) and CFP (bottom) are shown. (B) Immunoblot of whole transgenic embryo lysates expressing hAPP₆₉₅ or CFP alone. Full-length hAPP₆₉₅ and fragments of hAPP₆₉₅ processed by α-secretase (α), β-secretase (β), or β- and γ-secretase (β→γ) were detected using the Aβ-specific monoclonal antibody 6E10. APP-specific bands are only observed in transgenic embryos expressing hAPP₆₉₅, suggesting that ascidians possess functional APP processing enzyme complexes. The bands that are common to both lanes are ascidian proteins cross-reacting to the human Aβ antibody.

Fig. 3.

APP mutations synergize to greatly increase Aβ plaque formation in transgenic ascidian embryos. (A) Partial amino acid sequences of the wild-type (wt) and mutant hAPP₆₉₅ constructs. The specific point mutations introduced by PCR are indicated in red. (B–E) Representative transgenic embryos (∼23 hpf) ubiquitously expressing CFP or hAPP. All transgenes were driven by the ef1a enhancer. Images for DIC, thioflavin S staining, mAb 6E10 staining and fluorescent overlays (OL) are shown. (B) Control transgenic embryos expressing CFP. Note the lack of thioflavin S and mAb 6E10 staining. (C) Transgenic embryo expressing hAPPwt. Note that there are some thioflavin S positive plaques and there is some 6E10 staining. (D) Transgenic embryo expressing APPmtβ. There is an increase in both thioflavin S and mAb 6E10 staining compared with either CFP-expressing or hAPP-expressing embryos. (E) Transgenic embryo expressing APPmtβγ. These transgenic embryos show the most thioflavin S and mAb 6E10 staining. (F) Quantitation of the thioflavin S-positive plaques of the embryos shown in B–E. Data are plotted as the average number of plaques per embryo ± standard error (S.E.) versus the transgene, n=50 embryos per construct. A one-way ANOVA analysis was performed on the four categories shown; F (3,196)=72.87, P=0.000. A Tukey post-hoc comparison of the four groups indicates that APPmtβ-expressing embryos had a significantly greater number of plaques compared with larvae expressing APPwt or APPmtγ, P<0.05 (*). In addition, the number of plaques in APPmtβγ-expressing larvae is much greater than for all other transgenic embryos, suggesting that mutations at both cleavage sites synergize to produce a significant increase in Aβ plaques, P<0.001 (***). Data are representative of three separate experiments. Bars, 50 μm (B–D); 100 μm (E).

Fig. 4.

Treating transgenic embryos that express mutant human APP with an anti-amyloid drug reduces the number of thioflavin S-positive plaques in a dose-dependent manner. At the two-cell stage (1 hpf), transgenic embryos expressing APPmtβγ were grown in the presence of varying concentrations of the anti-amyloid drug 3-APS. At 23 hpf, embryos were fixed and stained for plaque formation with thioflavin S. The number of plaques per embryo ± S.E. is plotted versus drug treatment. Embryos grown in 10 mM 3-APS had significantly reduced numbers of plaques compared with other concentrations of the drug [P=0.001, 10 mM vs 0 mM (***); P=0.01, 10 mM vs 0.1 mM; P=0.01, 10 mM vs 1 mM].

Fig. 5.

Reduction of plaque formation in transgenic ascidian embryos expressing Aβ_1-42-CFP or following 3-APS treatment. (A,B) Representative transgenic ascidian tadpoles (23 hpf) expressing Aβ_1–42 (A) or Aβ_1–42-CFP (B) in the nervous system. In both cases, transgene expression is driven by cis-regulatory elements of the pan-neural synaptotagmin gene. Larvae were stained using both 6E10 and thioflavin S to reveal plaque deposits. Bars, 50 μm. (C) Relative quantification of thioflavin S-reactive, Aβ-containing amyloid plaques in larvae expressing Aβ_1–42, Aβ_1–42-CFP, or Aβ_1–42 in the presence of 1 mM 3-APS. The plaque load was normalized to the number of plaques in the Aβ_1–42-expressing larvae. Data are graphed as the relative plaque load per embryo ± S.E.; n=3 independent experiments, and at least 20 embryos per condition were analyzed per trial. A one-way ANOVA analysis was performed on the three separate conditions; F (2,6)=52.75, P=0.000. A Tukey post-hoc comparison indicates that fusion of CFP to Aβ_1–42 and 1 mM 3-APS treatment both significantly reduced plaque deposition in vivo, P<0.001 (***).

Fig. 6.

Transgenic ascidian embryos expressing human Aβ_1–42 fail to undergo normal settlement behaviors. (A) Schematic of the attachment assay. At 23 hpf, non-transgenic (NT) or transgenic larvae expressing Aβ_1–42, Aβ_1–42-CFP, or CFP in the nervous system, were transferred to an apparatus that monitors attachment to the underside of a floating 100×15 mm Petri plate. Embryos were scored for attachment at 24 hours following the transfer. (B) Relative quantification of attachment, normalized to dechorionated nontransgenic larvae. Embryos expressing Aβ_1–42 exhibit reduced attachment compared with controls (NT and CFP). To assess the effectiveness of an anti-amyloid inhibitor, transgenic embryos expressing Aβ_1–42 were grown in the presence of 1 mM 3-APS. Data are plotted as the average percentage attachment ± S.E.; n=3 and at least 50 embryos per construct were analyzed per trial. A one-way ANOVA analysis was performed; F (4,10)=32.09, P=0.000. A Tukey post-hoc comparison of the five groups indicates that larvae expressing Aβ_1–42 displayed a significant reduction in the levels of attachment compared with larvae expressing CFP alone, P<0.001 (***). The attachment rate is significantly improved with larvae expressing Aβ_1–42-CFP or upon 3-APS treatment, P<0.01 (**).