Protein retention in the endoplasmic reticulum rescues Aβ toxicity in Drosophila

Abstract

Amyloid β (Aβ) accumulation is a hallmark of Alzheimer's disease. In adult Drosophila brains, human Aβ overexpression harms climbing and lifespan. It's uncertain whether Aβ is intrinsically toxic or activates downstream neurodegeneration pathways. Our study uncovers a novel protective role against Aβ toxicity: intra-endoplasmic reticulum (ER) protein accumulation with a focus on laminin and collagen subunits. Despite high Aβ, laminin B1 (LanB1) overexpression robustly counters toxicity, suggesting a potential Aβ resistance mechanism. Other laminin subunits and collagen IV also alleviate Aβ toxicity; combining them with LanB1 augments the effect. Imaging reveals ER retention of LanB1 without altering Aβ secretion. LanB1's rescue function operates independently of the IRE1α/XBP1 ER stress response. ER-targeted GFP overexpression also mitigates Aβ toxicity, highlighting broader ER protein retention advantages. Proof-of-principle tests in murine hippocampal slices using mouse Lamb1 demonstrate ER retention in transduced cells, indicating a conserved mechanism. Though ER protein retention generally harms, it could paradoxically counter neuronal Aβ toxicity, offering a new therapeutic avenue for Alzheimer's disease.

Keywords: Alzheimer’s disease; Aβ toxicity; Drosophila melanogaster; ER retention; Endoplasmic reticulum; Laminin.

Graphical abstract

Fig. 1

Co-expression of LanB1 rescued the toxic effect of Aβ expression. (A) Survival curves of female flies expressing Aβ^Arc in adult neurons. Induction of Aβ^Arc with Elav-GeneSwitch significantly (p = 1.31 × 10⁻⁷⁰; log-rank test) shortened lifespan compared to uninduced controls. LanB1 and Aβ^Arc co-expression resulted in a significant rescue (p = 1.38 × 10⁻⁵⁰; log-rank test) of the short-lived phenotype. (B) Repeat of experiment in (A). LanB1 significantly rescued Aβ^Arc toxicity (p = 8.31 × 10⁻⁵⁶; log-rank test). (C) LanB1 significantly rescued Aβ^Arc toxicity in male flies (p = 8.31 × 10⁻⁵⁶; log-rank test). (D) Induction of Aβ^X2 significantly (p = 5.99 × 10⁻⁶⁹; log-rank test) shortened lifespan compared to uninduced controls. LanB1 and Aβ^X2 co-expression resulted in a significant rescue (p = 8.04 × 10⁻⁴⁰; log-rank test). (E) Induction of Aβ^Arc with NsybGS significantly (p = 3.38 × 10⁻⁶⁷; log-rank test) shortened lifespan compared to uninduced controls. LanB1 significantly rescued Aβ^Arc toxicity (p = 6.03 × 10⁻⁴⁰; log-rank test). Dashed lines represent uninduced “RU−” controls, and solid lines represent induced “RU+” conditions. For all lifespan experiments, n = 150 flies per condition. (F) Climbing ability was measured until day 51. Climbing ability of Aβ^Arc flies was lower than that of all other groups (p < 0.0001, 2-way analysis of variances with Tukey’s post-hoc test). LanB1 significantly rescued the Aβ^Arc-induced decline in climbing ability, though not to the same level as uninduced controls (p < 0.0001, 2-way analysis of variances with Tukey’s post-hoc test). Data are shown as mean ± SEM (n = 37–70 flies measured per time point). (G) LanB1 suppressed Aβ toxicity in the eyes of flies raised at 29 °C. Control flies displayed a highly ordered ommatidia lattice (left). Expression of Aβ^X2 using the eye-specific GMR-GAL4 driver resulted in small, glassy eyes that accumulated necrotic spots (middle). Co-expression of Aβ^X2 with LanB1 rescued most of the size and organization defects (right). (H) Drosophila laminins and their assembly in the ER. Created using BioRender. Abbreviations: Aβ, amyloid β; ER, endoplasmic reticulum; ElavGS, Elav-GeneSwitch; LanB1, Laminin B1.

Fig. 2

LanB1 did not affect Aβ RNA or protein levels. (A) Aβ mRNA was significantly upregulated upon RU induction (p < 0.0001; 1-way analysis of variances (ANOVA) with Tukey’s post-hoc test) and was not affected by co-expression of LanB1. Data indicate the mean (n = 5 biological replicates per condition). (B) Total Aβ₄₂ protein levels increased significantly over time with RU induction compared to uninduced controls (p < 0.0001, 2-way ANOVA with Tukey’s post-hoc test), and were unaffected by LanB1 co-expression. Data are shown as mean ± SEM (n = 4 biological replicates per time point). (C) Western blot with quantification of soluble Aβ. The green arrow indicates the distinctive crescent shape of Aβ used for quantification. Soluble Aβ levels increased in induced conditions and were unaffected by LanB1 co-expression. Data indicate the mean (n = 4 biological replicates per condition). (D) After 3 weeks of induction, neither soluble nor insoluble Aβ levels were affected by LanB1 co-expression. Data are shown as mean ± SEM (n = 7 biological replicates per condition). (E) After 3 weeks of RU, induced Aβ^Arc flies consumed a significantly smaller amount of food over 7 days compared to uninduced controls (p < 0.0001, 2-way ANOVA with Tukey’s post-hoc test). The amount of food ingested by flies with a history of LanB1 and Aβ^Arc induction was not significantly different from uninduced controls. Food consumption was measured for 10 flies per condition every 24 hours. Data are shown as mean ± SEM. (F) Expression of Aβ for the first 3 weeks of adulthood significantly reduced lifespan compared to uninduced controls (p = 4.07 × 10⁻¹²; log-rank test). Three-week induction of Aβ and LanB1 together resulted in a small but significant lifespan extension compared to uninduced controls (p = 5.04 × 10⁻⁰⁶; log-rank test). Abbreviations: Aβ, amyloid β; ElavGS, Elav-GeneSwitch; LanB1, Laminin B1.

Fig. 3

Intracellular accumulation of Aβ and LanB1. (A) Aβ accumulated intracellularly and did not colocalize with the nuclear marker histone H3:GFP. The arrow indicates the non-nuclear expression of Aβ. (B) Similarly, LanB1 accumulated intracellularly, not in the nucleus, and partially overlapped with Aβ expression. The arrow indicates the colocalization of Aβ and LanB1. (A and B) Representative confocal fluorescence z projections taken at 20× magnification of whole brains from 21-day-old female flies stained with Aβ (6E10-green) and DAPI (blue). The yellow box inset shows a single section of the same brain taken at 63× magnification. Endogenous fluorescence (i.e., without staining) of H3:GFP and LanB1:RFP is shown. H3:GFP has been false-colored red to aid comparison. (C) Z-projection and (D) single section showing intracellular accumulation of Aβ in Dilp2 neurons. (E) Z-projection and (F and G) single sections showing intracellular accumulation of Aβ and LanB1 in Dilp2 neurons. LanB1 expression was not uniformly distributed through the cytoplasm of these cell bodies and appeared to accumulate in discrete intracellular compartments. Arrows highlight the colocalization of Aβ and LanB1, while arrowheads highlight areas with no overlap, often in the same cell. (C-G) Representative confocal images taken at 63x magnification of Dilp2 neurons from 21-day-old female flies stained with Aβ (6E10-green) and DAPI (blue). Endogenous fluorescence (i.e., without staining) of LanB1:RFP is shown. Genotypes: (A) ElavGS>Aβ^X2+ H3:GFP; (B) ElavGS>Aβ^X2+ LanB1:RFP; (C and D) Dilp2-GS>Aβ^Arc; (E-G) Dilp2-GS>Aβ^Arc + LanB1:RFP. Scale bar, 100 µm for (A and B), 10 µm for (C-G). Abbreviations: Aβ, amyloid β; Dilp2, Drosophila insulin-like peptide 2; ElavGS, Elav-GeneSwitch; LanB1, Laminin B1.

Fig. 4

LanB1 accumulates in the ER. (A) Z-projection and (B) single section showing neuronal expression of LanB1 with the ER marker, KDEL:GFP, using the ElavGS driver. (C) Magnified view of the dashed yellow box inset in (B). LanB1 colocalizes with KDEL:GFP. (D) Without RU486 induction, there was very little induction of LanB1 and/or KDEL:GFP expression. Representative confocal fluorescence images were taken at 63× magnification from 7-day-old female flies. Images show the antennal lobe and surrounding brain regions. Endogenous fluorescence (i.e., without staining) of KDEL:GFP and LanB1:RFP is shown. Genotype: ElavGS>KDEL:GFP + LanB1:RFP. Scale bar, 50 µm for (A, B, and D), 10 µm for (C). Abbreviations: Dilp2, Drosophila insulin-like peptide 2; ElavGS, Elav-GeneSwitch; LanB1, Laminin B1.

Fig. 5

Quantification of Aβ secretion from Dilp2 neurons. (A) Aβ expression in Dilp2 neurons was highest in the cell body area, but there was also diffuse, punctate staining of Aβ outside the cell bodies. LanB1 co-expression had no effect on Aβ fluorescence outside Dilp2 neurons. Without RU induction, there was no Aβ found in Dilp2 neurons or the surrounding area. Representative confocal fluorescence z projections taken at 63× magnification of whole brains from 21-day-old female flies stained with Aβ (6E10-green) and DAPI (blue). The dashed yellow area indicates the area of fluorescence measurement. White asterisk in the top row of images indicates strong staining of esophageal muscle. Endogenous fluorescence (i.e., without staining) of LanB1:RFP is also shown. (B) Quantification of Aβ fluorescence outside Dilp2 neurons. There was a significant increase in diffuse, punctate Aβ staining outside Dilp2 neurons when induced (p = 0.048; 1-way analysis of variances). LanB1 co-expression had no effect on Aβ staining. Biological replicate numbers for each condition are labeled within the bars. (C) Diagram of the proposed secretion of Aβ from Dilp2 neurons into the extracellular space. Created with BioRender.com. Genotypes: Dilp2-GS>Aβ^Arc; and Dilp2-GS>Aβ^Arc + LanB1:RFP. Scale bar, 50 µm. Abbreviations: Aβ, amyloid β; Dilp2, Drosophila insulin-like peptide 2; LanB1, Laminin B1.

Fig. 6

Enhanced rescue of Aβ toxicity with combined laminin/collagen IV subunit overexpression. (A) LanB1 significantly rescued Aβ toxicity (p = 2.85 × 10⁻⁶²; log rank). Cg25C significantly rescued Aβ toxicity (p = 3.42 × 10⁻⁶³; log rank). There were small but significant extensions of lifespan in the uninduced controls (LanB1, p = 0.036; Cg25C, p = 7.51 × 10⁻⁰⁶; log rank vs. ElavGS>Aβ^Arc alone). (B) LanB1 significantly rescued Aβ toxicity (p = 8.04 × 10⁻⁴⁰; log rank). Cg25C significantly rescued Aβ toxicity (p = 1.25 × 10⁻³⁵; log rank). (C) LanB1 significantly rescued Aβ toxicity (p = 3.77 × 10⁻³³; log rank). Cg25C significantly rescued Aβ toxicity (p = 2.24 × 10⁻³⁰; log rank). LanB1 + Cg25C had a partially additive effect on the rescue of Aβ^Arc toxicity. Cox proportional hazard analysis showed a significant interaction between LanB1 and Cg25C (p < 0.001). (D) LanB1, Cg25C, and LanA rescued Aβ toxicity (LanB1, p = 3.34 × 10⁻⁶⁵; Cg25C, p = 2.11 × 10⁻⁵⁵; LanA, p = 1.10 × 10⁻³²; log rank vs. ElavGS>Aβ^Arc alone). LanB1 + Cg25C and LanB1 + LanA had a partially additive effect on the rescue of Aβ^Arc toxicity. Cox proportional hazard analysis showed a significant interaction with LanB1 for Cg25C (p < 0.001) and LanA (p = 0.022). There was a significant extension of lifespan in the uninduced controls (LanB1, p = 6.45 × 10⁻¹⁰; Cg25C, p = 1.57 × 10⁻¹¹; LanA, p = 1.40 × 10⁻¹⁰; LanA + LanB1, p = 1.66 × 10⁻³⁹; LanB1 + Cg25C, p = 0.026; log rank vs. ElavGS>Aβ^Arc alone). Dashed lines represent uninduced “RU−” controls, and solid lines represent induced “RU+” conditions. For all lifespan experiments, n = 150 flies per condition. Abbreviations: Aβ, amyloid β; ElavGS, Elav-GeneSwitch; LanB1, Laminin B1.

Fig. 7

LanB1 rescue of Aβ toxicity is independent of the BiP/Xbp1 ER stress response pathway. (A) BiP mRNA was significantly upregulated upon Aβ induction (p < 0.01; 1-way analysis of variances with Tukey’s post-hoc test), and this was not changed by the overexpression of LanB1:RFP. Data are shown indicating the mean (n = 4–5 biological replicates per condition). (B) Western blot of BiP protein levels, and quantification in (C) confirm that BiP protein is also significantly upregulated with Aβ induction, and this is not changed by LanB1:RFP overexpression (p < 0.01; 1-way analysis of variances with Tukey’s post-hoc test). Data are shown indicating the mean (n = 3–4 biological replicates per condition). (D) Co-expression of Aβ^X2 with Xbp1^VDRC in the developing eye. Knockdown of Xbp1 significantly exacerbated Aβ toxicity, and these flies exhibited very small and depigmented eyes. LanB1:RFP expression further enhanced the combined toxicity of Aβ and Xbp1^VDRC. The combination of LanB1:RFP and Xbp1^VDRC without Aβ expression also resulted in a rough eye phenotype. (E) Knockdown of Xbp1 exacerbated Aβ toxicity and significantly shortened lifespan compared to controls (p = 3.03 × 10⁻³⁰; log rank). (F) Knockdown of Xbp1 exacerbated Aβ toxicity and significantly shortened lifespan compared to controls (p = 1.68 × 10⁻⁴⁴; log rank). Overexpression of LanB1 rescued the shorter lifespan of Xbp1^VDRC (Aβ+LanB1:RFP+Xbp1^VDRC vs. Aβ+Xbp1^VDRC, p = 9.13 × 10⁻²²; log rank). For display purposes, the control micrograph in (D) is the same as that in Supplementary Fig. 5. For all lifespan experiments, n = 150 flies per condition. “VDRC” indicates an RNAi transgene. Genotypes: (A) ElavGS>Aβ^Arc; ElavGS>Aβ^Arc+ LanB1:RFP; (B and C) NsybGS>Aβ^Arc; NsybGS>Aβ^Arc+ LanB1:RFP. Abbreviations: Aβ, amyloid β; ElavGS, Elav-GeneSwitch; LanB1, Laminin B1.

Fig. 8

Protein accumulation in the ER may be responsible for the rescue of Aβ toxicity. (A and B) Co-expression of Aβ^X2 with LanB2:GFP or KDEL:GFP rescued Aβ toxicity in the developing eye compared to mCD8:GFP controls, while secr:GFP exacerbated Aβ toxicity. (B) Quantification of eye sizes in (A). mCD8:GFP did not rescue the rough eye phenotype compared to Aβ^X2 alone (see Supplementary Fig. 5). LanB2:GFP and KDEL:GFP significantly rescued eye size while secr:GFP significantly reduced eye size compared to mCD8:GFP controls (p < 0.0001; 1-way analysis of variances with Dunnett’s post-hoc test). Data are shown as mean ± SEM (n = 34–46 eyes measured per condition). For display purposes, 2 micrographs in (A) are the same as those in Supplementary Fig. 5. (C) GFP mRNA was significantly upregulated in all GFP-tagged conditions compared to the ElavGS/w^Dah control (p < 0.001; 1-way analysis of variances on log-transformed data; p values were adjusted using the Tukey method for multiple comparisons); n = 5 biological replicates per condition. The Y-axis is the log scale. (D) Survival curves of female flies induced to express Aβ^Arc via NsybGS. LanB2:GFP and KDEL:GFP rescued Aβ toxicity while secr:GFP exacerbated Aβ toxicity compared to mCD8:GFP controls (LanB2:GFP, p = 2.37 × 10⁻²⁶; KDEL:GFP, p = 3.49 × 10⁻¹⁷; secr:GFP, p = 1.79 × 10⁻¹⁹; log rank). (E) One copy of KDEL:GFP significantly rescued Aβ^Arc toxicity (p = 1.15 × 10⁻⁴⁴; log-rank test), while 2 copies of KDEL:GFP led to an even greater rescue (p = 9.18 × 10⁻²⁷; log-rank test comparing 1 vs. 2 copies of KDEL:GFP). Genotypes: (A and B) GMR>Aβ^X2+ mCD8:GFP ; GMR>Aβ^X2+ LanB2:GFP; GMR>Aβ^X2+ KDEL:GFP; GMR>Aβ^X2+ secr:GFP, (C) ElavGS/w^Dah; ElavGS>mCD8:GFP; ElavGS>LanB2:GFP; ElavGS>KDEL:GFP; ElavGS>secr:GFP. For all lifespan experiments, n = 150 flies per condition.

Fig. 9

Intra-ER retention of overexpressed Lamb1 is conserved in mouse brain tissue. (A) Organotypic hippocampal slice cultures from 10-day-old mouse pups were incubated for 24 hours with lentivirus containing mLamb1+GFP or control lentivirus (GFP only), and then fixed 2-weeks later. Slices were then stained for Lamb1 and the ER marker, Calnexin. Endogenous GFP expression was used to identify successful transduction. DAPI labeled nuclei. The top 2 rows show examples of control lentivirus without mLamb1 induction. The bottom 2 rows show examples of lentiviral mLamb1 induction. Yellow arrowheads indicate areas of colocalization of Lamb1 and Calnexin. (B) Summary model of intra-ER laminin retention. Created with BioRender.com. Scale bar, 10 µm. Abbreviations: Aβ, Amyloid β; ER, endoplasmic reticulum.