CASP8 intronic expansion identified by poly-glycine-arginine pathology increases Alzheimer's disease risk

Abstract

Alzheimer's disease (AD) affects more than 10% of the population ≥65 y of age, but the underlying biological risks of most AD cases are unclear. We show anti-poly-glycine-arginine (a-polyGR) positive aggregates frequently accumulate in sporadic AD autopsy brains (45/80 cases). We hypothesize that these aggregates are caused by one or more polyGR-encoding repeat expansion mutations. We developed a CRISPR/deactivated-Cas9 enrichment strategy to identify candidate GR-encoding repeat expansion mutations directly from genomic DNA isolated from a-polyGR(+) AD cases. Using this approach, we isolated an interrupted (GGGAGA)_n intronic expansion within a SINE-VNTR-Alu element in CASP8 (CASP8-GGGAGA^EXP). Immunostaining using a-polyGR and locus-specific C-terminal antibodies demonstrate that the CASP8-GGGAGA^EXP expresses hybrid poly(GR)n(GE)n(RE)n proteins that accumulate in CASP8-GGGAGA^EXP(+) AD brains. In cells, expression of CASP8-GGGAGA^EXP minigenes leads to increased p-Tau (Ser202/Thr205) levels. Consistent with other types of repeat-associated non-AUG (RAN) proteins, poly(GR)n(GE)n(RE)n protein levels are increased by stress. Additionally, levels of these stress-induced proteins are reduced by metformin. Association studies show specific aggregate promoting interrupted CASP8-GGGAGA^EXP sequence variants found in ~3.6% of controls and 7.5% AD cases increase AD risk [CASP8-GGGAGA-AD-R1; OR 2.2, 95% CI (1.5185 to 3.1896), P = 3.1 × 10^-5]. Cells transfected with a high-risk CASP8-GGGAGA-AD-R1 variant show increased toxicity and increased levels of poly(GR)n(GE)n(RE)n aggregates. Taken together, these data identify polyGR(+) aggregates as a frequent and unexpected type of brain pathology in AD and CASP8-GGGAGA-AD-R1 alleles as a relatively common AD risk factor. Taken together, these data support a model in which CASP8-GGGAGA^EXP alleles combined with stress increase AD risk.

Keywords: Alzheimer’s disease; RAN proteins; microsatellite expansion mutations; protein aggregates; repeat associated non-AUG (RAN) translation.

Fig. 1.

polyGR aggregates accumulate in AD autopsy brains. (A) Example polyGR IHC staining (red) detected by a rabbit polyclonal a-polyGR antibody in hippocampal sections (HC) from AD and control cases. Each panel represents a different AD or control case. (B) Summary diagram showing location of polyGR aggregates (red dots) in AD HC sections; HC = hippocampus, ERC = entorhinal cortex. (C) Quantification of polyGR aggregates in HC region indicated by red square in (B) from AD (n = 80) and control (n = 18) cases. (D) Example dot blot of a-polyGR signal using a rat monoclonal a-polyGR antibody and protein extracts of the frozen frontal cortex (FCX) from AD (n = 65) and control (n = 20) cases. (E) Relative polyGR levels in FCX protein lysates from AD and controls by dot blot. (F) Example IHC staining showing distinct a-polyGR staining compared to other AD pathological hallmarks including p-Tau (S202 and T205), Aβ plaques, and p-TDP43 in Cornu Ammonis (CA) and dentate gyrus (DG) of the HC. Positive staining shown in red, (Scale bar, 20 μm.) (G) Examples of double IHC staining showing a-polyGR staining (pink) detected in brain regions with both high and low p-Tau (brown) subregions of the same AD brain section. Black arrows: cells with both polyGR and p-tau signal, open arrows: cells with polyGR staining, blue arrows: cells with p-Tau staining. (H) Plot of a-polyGR and p-Tau (S202 and T205) staining detected in sequential slides from 21 randomly selected AD cases shows a positive correlation. Data represent mean ± SEM. Two-tailed, unpaired t test. ****P < 0.0001.

Fig. 2.

Identification of interrupted GGGAGA repeat expansion in CASP8 (CASP8-GGGAGA^EXP) using CRISPR deactivated Cas9-based repeat enrichment and detection (dCas9READ). (A) Schematic diagram showing dCas9READ strategy for enrichment of repeat expansion mutations encoding specific RAN protein motifs. (B) Experimental flow using dCas9READ to identify novel polyGR-encoding repeat expansions and genetic association studies to test the association of candidate mutations with AD. (C) Example data showing fold enrichment of ten loci after dCas9READ. Genomic DNA from five polyGR(+) AD cases, a polyGR(-) control (Cntl), a polyGR(−) AD (AD#4) was used. CASP8 was the top hit in two polyGR(+) AD cases. (D) Diagram showing insertion of the CASP8-GGGAGA^EXP within an SVA-E retrotransposon element which is inserted in the opposite orientation with reference repeat sequence and repeat-prime PCR (RP-PCR) primers used to characterize the repeat expansion. (E and F) RP-PCR data showing five different repeat expansion patterns of CASP8-GGGAGA^EXP in 5 AD cases using (GGGAGA)₃GG (E) or interrupted primer intP (GGGAGA)₃CG (F). (G) Percentage of individuals carrying CASP8-SVA in initial AD and control cohorts. (H) Frequency of alleles containing CASP8-SVA insertion in AD (n = 266) and controls (n =400). (I) Distribution of estimated GGGAGA repeat lengths measured by LR-PCR in initial AD and control cohorts. (G and H) Chi-square test, **P < 0.01, ***P < 0.001.

Fig. 3.

CASP8-GGGAGA^EXP produces polyGR(+) aggregates in CASP8-GGGAGA^EXP(+) AD brains. (A) Diagram showing putative expansion proteins translated from sense and antisense CASP8-GGGAGA^EXP transcripts. Amino acid sequences highlighted in red were used to generate frame-specific C-terminal (CT) antibodies. S, sense; AS, antisense; f1–3, reading frames 1 to 3; * stop codon. Analysis of the upstream flanking sequence of CASP8-GGGAGA^EXP showed an upstream ATG that could be used in one of the three frames; however, this ATG-containing frame shifts reading frame depending on the highly polymorphic interruptions within the repeat tract. (B) IHC detection of CASP8 RAN Sf3 protein aggregate staining (red) in the hippocampus of CASP8-GGGAGA^EXP(+) AD patients detected with a-CT-Sf3 antibody. Each photo panel represents a different AD or control case. (C) Double IF showing colocalization of a-polyGR (red) and a-CT-Sf3 (green) staining in the frontal cortex (FCX) of CASP8-GGGAGA^EXP(+) AD patients. Each row represents a different AD case.

Fig. 4.

CASP8-GGGAGA^EXP produces polyGR(+) proteins and increases p-Tau in transfected cells. (A) Diagram showing C8-GGGAGA^EXP-3T constructs containing 6xStop codons (two in each reading frame), a 100-bp of CASP8 flanking sequence upstream of the repeat, interrupted GGGAGA repeat expansions, and triple epitope tags in each of the three reading frames. (B) Protein plot showing C8-GGGAGA^EXP-3T expresses mutant proteins in FLAG and HA frames in transfected HEK293T cells but not empty vector controls. (C) Mutant proteins in all three reading frames (FLAG, HA, and Myc) were detected by IF in HEK293T cells transfected with C8-GGGAGA^EXP-3T but not empty vector plasmids. (D) Protein blotting showing effects of Thapsigargin (Tg) and metformin (M) on the levels of CASP8 RAN proteins expressed in the FLAG frame. (E) Quantification of CASP8 RAN proteins expressed from the FLAG and HA frames with Tg or Tg + metformin treatment (n = 7 to 12/condition). (F) IF images showing increased p-Tau levels in SH-SH5Y cells expressing FLAG-tagged-(GR)₆₀ protein encoded using alternative codon sequences. (G) Quantification of p-Tau levels in polyGR(+) (n = 46) and polyGR(−) SH-SY5Y cells (n = 287), Cntl: empty vector control. (H) IF images showing increased p-Tau levels in SH-SY5Y cells transfected with C8-GGGAGA^EXP-3T plasmids. Data represent mean ± SEM. One-way ANOVA Holm–Sidak’s multiple comparisons test (E) and unpaired two-tailed t test (G). ***P < 0.001, ****P < 0.0001.

Fig. 5.

CASP8-GGGAGA^EXP sequence variant increases AD risk. (A) Repeat legend showing repeat motifs and interruptions detected in CASP8 SVA-E GGGAGA^EXP alleles. (B) Repeat configurations of AD-associated CASP8 SVA-E GGGAGA^EXP variants (CASP8 GGGAGA-AD-R1) which contains a longer VNTR region with a 48 to 50 bp insertion (yellow box) and interrupted GGGAGA repeat expansion including 3 to 4 units of [(GGGAGA)₄CG] within the repeat, and 5 to 7 units of (GGGAGA) at the 3’ end. The red lines highlight unique features in the CASP8-GGGAGA-AD-R1 variant compared to other repeat variants in the CASP8 locus. (C) Repeat configurations of other CASP8 GGGAGA^EXP variants found in the general population. (D) A summary of the association of CASP8-GGGAGA-AD-R1 variants with AD in three independent AD and control cohorts. (E) Proposed model showing the various CASP8 alleles in the population and the order in which the CASP8 SVA-E insertion and tagging SNPs (rs1035142 and rs700635), the longer VNTR, and the interrupted hexanucleotide stretch on the AD-R1 alleles arose.

Fig. 6.

AD risk variant CASP8-GGGAGA-AD-R1 increases RNA and polyGR-containing RAN aggregate levels in cells. (A) Schematic diagram of constructs used to express CASP8-GGGAGA-AD-R1 (p-AD-R1) and a common CASP8-GGGAGA^EXP variant (p-C-Var) in transfected cells. (B) Example IF images showing a-polyGR staining in HEK293T cells transfected with p-AD-R1 or p-C-Var plasmids and the graph showing quantification of a-polyGR staining (n = 3/group), Cntl = empty vector control. (C) Example FISH images showing rGGGAGA^EXP RNA staining in HEK293T cells transfected with p-AD-R1, p-C-Var, or empty vector control (Cntl) plasmids and graph showing quantification of GGGAGA RNA inclusions (n = 3/group). (D) LDH toxicity assays showing effects of p-AD-R1, p-C-Var, and Cntl minigenes on survival of T98 cells (n = 3/group). (E) Model in which stress and allele configuration increase poly(GR)n(RE)n(GE)n and GGGAGA^EXP and p-Tau aggregates, which leads to increased AD risk. Data represent mean ± SEM. (B and D) One-way ANOVA Holm–Sidak’s multiple comparisons test, (C) unpaired two-tailed t test, *P < 0.05, **P < 0.01.