Human single-chain Fv intrabodies counteract in situ huntingtin aggregation in cellular models of Huntington's disease

Abstract

This investigation was pursued to test the use of intracellular antibodies (intrabodies) as a means of blocking the pathogenesis of Huntington's disease (HD). HD is characterized by abnormally elongated polyglutamine near the N terminus of the huntingtin protein, which induces pathological protein-protein interactions and aggregate formation by huntingtin or its exon 1-containing fragments. Selection from a large human phage display library yielded a single-chain Fv (sFv) antibody specific for the 17 N-terminal residues of huntingtin, adjacent to the polyglutamine in HD exon 1. This anti-huntingtin sFv intrabody was tested in a cellular model of the disease in which huntingtin exon 1 had been fused to green fluorescent protein (GFP). Expression of expanded repeat HD-polyQ-GFP in transfected cells shows perinuclear aggregation similar to human HD pathology, which worsens with increasing polyglutamine length; the number of aggregates in these transfected cells provided a quantifiable model of HD for this study. Coexpression of anti-huntingtin sFv intrabodies with the abnormal huntingtin-GFP fusion protein dramatically reduced the number of aggregates, compared with controls lacking the intrabody. Anti-huntingtin sFv fused with a nuclear localization signal retargeted huntingtin analogues to cell nuclei, providing further evidence of the anti-huntingtin sFv specificity and of its capacity to redirect the subcellular localization of exon 1. This study suggests that intrabody-mediated modulation of abnormal neuronal proteins may contribute to the treatment of neurodegenerative diseases such as HD, Alzheimer's, Parkinson's, prion disease, and the spinocerebellar ataxias.

Figure 1

Characterization of anti-HD C4 sFv. (A) Coomassie blue staining of affinity-purified GST-fusion proteins GST-DRPLA-Q35 (lane 2), GST-HD-Q25 (lane 3), and GST-HD-Q42 (lane 4), mixed with anti-HD C4 sFv. Molecular weight markers are in lane 1. (B) Direct binding of C4 sFv to immobilized GST-fusion proteins analyzed by ELISA. C4 sFv-myc-His₆ antibodies were quantitated in microtiter plates coated with 70 nM of GST-fusion proteins and blocked with 1% BSA in PBS. Bound sFv were detected with anti-myc antibodies (9E10, 1 μg/ml), followed by alkaline phosphatase-conjugated goat anti-mouse IgG. (C) Kinetic binding affinity of C4 sFv analyzed by BIAcore. A range of concentration (60–100 nM) of C4 sFv was used to measure the association rate (k_on) on 50 resonance units of biotinylated-HD peptide bound to a streptavidin sensor chip.

Figure 2

Retargeting of GFP-fusion HD fragments to the nucleus by anti-HD C4 sFv-HA-nls. Immunofluorescence of HD-Q103-myc-His₆ with C4 sFv-HA-nls (b, d, f) or a negative control sFv-HA-nls intrabody (a, c, e), after 48-h cotransfection in COS-7. HD-Q103-myc-His₆ was detected by 9E10 mAb, followed by FITC-labeled goat anti-mouse IgG antibodies (c, d). The intrabodies were detected with polyclonal anti-HA antibodies, followed by rhodamine-labeled goat anti-rabbit IgG antibodies (a, b). (e, f) Dual staining of the same fields.

Figure 3

Immunofluorescence of HD-polyQ-GFP in the presence of cotransfected anti-HD C4 sFv. (A) BHK-21 cells transfected with HD-polyQ-GFP (normal control, Q25; pathogenic, Q72 or Q103) alone or cotransfected with C4 intrabody or control (ML3–9 sFv-HA) at an sFv to antigen plasmid ratio of 5:1. Live cells are shown photographed at 48 h after transfection, showing aggregates as intense foci of fluorescence. Similar results were seen with HD-Q72-GFP cotransfections (not shown). (B) Double labeling of 5:1 C4/HD-polyQ-GFP cotransfections; HD-polyQ-GFP (green), C4 sFv (red, detected by anti-HA antibody and Alexa 568-conjugated secondary antibody) and merged image. Some aggregates are found in Q103 cotransfectants but never in Q25 transfectants; cells not expressing C4 intrabody (arrowheads) often display aggregates, and to a lesser degree, in some cotransfected cells (arrow).

Figure 4

Inhibition of aggregate formation of pathogenic HD-polyQ-GFP in the presence of cotransfected anti-HD C4 sFv. (A) Quantitation of aggregate numbers at 48 h after cotransfection (5:1 ratio) with anti-HD C4 sFv-HA, negative control ML3–9 sFv-HA, or parent vector (pcDNA) and HD-Q72-GFP in COS-7, BHK-21, or HEK 293 cells, as indicated. After lysis with 2% SDS/2% Triton X-100/50 mM Tris, the number of insoluble fluorescent aggregates in six or eight random fields was counted. Bars represent means of six cotransfections ± SEM. C4 sFv-HA significantly reduced the number of aggregates when compared with control ML3–9 sFv-HA or pcDNA (P < 0.005). Similar results seen with HD-Q103-GFP (not shown). (B) Western blot analysis of HD-Q72-GFP protein expression in cotransfections. Cell lysates were collected from 5:1 cotransfections (sFv/HD-72Q-GFP) in duplicate at 24 h, before prominent aggregate formation and loss of HD-72Q-GFP to the insoluble compartment, and subjected to SDS/PAGE. The membrane was probed with 1:1,000 anti-AFP primary antibody (Quantum Biotechnologies, Montreal, PQ, Canada), stripped and reprobed with 1:500 anti-actin antibody (Sigma). HD-Q72-GFP bands were of similar intensity after detection by HRP-conjugated secondary antibody and chemiluminescence. (C) Relative luminescence of cell lysates from cotransfections of pGL2-Control (a luciferase-expressing reporter construct) with C4 sFv, control intrabody or parental vector. Mean luminescence (±SEM) as determined by Wallac 1409 scintillation counter was not different.