DNA nanoflower Oligo-PROTAC for targeted degradation of FUS to treat neurodegenerative diseases

Abstract

Oligonucleotide-based medicine faces challenges in efficiently crossing the blood-brain barrier and rapidly reducing toxic proteins. To address these challenges, here we establish an integrated modality, brain-penetrant DNA nanoflowers incorporated with oligonucleotide-based proteolysis targeting chimeras. Using FUS as a proof-of-concept, mutations of which cause frontotemporal dementia and amyotrophic lateral sclerosis, we demonstrate that a FUS-engaging RNA oligonucleotide crosslinked to a ligand for Cereblon efficiently degrade FUS and its cytoplasmic disease-causing mutants through a ubiquitin-proteasomal pathway. The DNA nanoflower contains hundreds of oligonucleotide binding sites and transferrin receptor-engaging aptamers, allowing efficient loading of the oligonucleotide-based degrader and engaging transferrin receptors for brain delivery. A single dose intravenous injection of this modality reaches brain parenchyma within 2 h and degrades 80% FUS protein there, sustained for two weeks without noticeable toxicity. DNA nanoflower oligonucleotide-based degrader is a therapeutic strategy for neurodegenerative diseases that leverages the advantages of designer oligonucleotides and targeted protein degradation.

Fig. 1. Design, synthesis, and profiling of Oligo-PROTACs.

a Mechanism of action of Oligo-PROTAC-mediated FUS degradation. b The RNA Oligos selected for Oligo-PROTAC synthesis and profiling. The FUS-recognition GGUG motif is marked in red. c, d Immunoblots (c) and quantitative analysis (d) of FUS in PC12 cells transfected with CRBN-based degrader (dFUS-…C) or VHL-based degrader (dFUS-…V) for 24 h. The samples derive from the same experiment, and the blots were processed in parallel. e, f Images (e) and quantitative analysis (f) of PC12 cells treated with dFUS-E9C (125 nM), followed by immunostaining with anti-FUS antibodies and DAPI. Scale bar, 20 μm. g Streptavidin pulldown assays to examine the binding of dFUS-E9C with FUS and CRBN. h MST assays to determine the affinity of dFUS-E9C to GFP and GFP-FUS. i Quantitative PCR to determine the FUS mRNA levels in PC12 cells treated with dFUS-E9C (125 nM) for 24 h. j Immunoblot analysis of FUS and GAPDH in SH-SY5Y cells treated with dFUS-E9C (125 nM) for 24 h with or without MG132 treatment. k Immunoblots of SH-SY5Y cells treated with dFUS-E9C (125 nM) together with excessive CRBNL. l Immunoblot analyses of FUS, CRBN, and GAPDH in SH-SY5Y cells treated with dFUS-E9C (125 nM) and siRNAs targeting CRBN. m HEK-293T cells were transfected with His-Myc-Ub followed by treatment with dFUS-E9C (125 nM) for 6 h. Cell lysates were subjected to immunoprecipitation with the FUS antibody, followed by immunoblotting with the indicated antibodies. The averages of n = 3 (d, f, h, i) biologically independent samples are shown. Data are shown as the mean ± SEM. Statistical significance in (d, f, i) was assessed using one-way ANOVA with multiple comparisons and two-tailed t tests, respectively. The data presented in (c, e, g, j–m) are representative of three independent experiments. Source data are provided as a Source Data file.

Fig. 2. dFUS-E9C interacts with RGG2 and RRM domains of FUS.

a A summary of tested FUS truncation mutants and their affinity to dFUS-E9C. b Streptavidin pulldown assays to determine the binding capability between Biotin-dFUS-E9C and FUS truncations. c, d Predicted interactions of RRG2 (c) or RRM (d) domain with dFUS-E9C by molecular simulation. Red bases are predicted to be critical for the interaction between the Oligo and FUS. e The critical bases of FUS-RNA#E9 responsible for FUS binding were mutated and marked in red. f MST assays to determine the binding affinity of GFP-FUS to dFUS-E9C and its mutants. g Streptavidin pulldown assays to assess the interaction between FUS and Biotin-dFUS-E9 and its mutants. h, i Immunoblots (h) and quantitative analysis (i) of PC12, SH-SY5Y, and Neuro-2a cells transfected with increasing concentrations of dFUS-E9C for 24 h. The averages of n = 3 (f, i) biologically independent samples are shown. Data are shown as the mean ± SEM. The data presented in (b, g, h) are representative of three independent experiments. Source data are provided as a Source Data file.

Fig. 3. dFUS-E9C specifically degrades FUS and prefers cytoplasmic FUS aggregates.

a, b Comparison of proteomic changes after treating neuron-like Neuro-2a cells with dFUS-E9C (a) or siFUS (b) for 24 h. Data are the mean of three biological replicates. c Heatmap showing proteins with significant changes between control and dFUS-E9C treatment. The overlapped proteins with significant changes in both dFUS-E9C and siFUS treatment are marked in red. d, e HEK-293T cells were transfected with GFP-FUS-WT, GFP-FUS-P525L, GFP-FUS-R521C, GFP-FUS-R521G, and GFP-FUS-R521H plasmids followed by transfecting with increasing concentrations of dFUS-E9C. Cell lysates were then subjected to immunoblotting (d), and the relative FUS levels were quantified by densitometry (e). f, g GFP-FUS-P525L-expressing HEK-293T cells were transfected with the indicated concentrations of dFUS-E9C, and the cytosol and nucleus compartments were isolated and subjected to immunoblotting (f). The relative FUS levels in the cytosol and nucleus were quantified by densitometry (g). h Immunofluorescence images of GFP-tagged wild-type FUS and disease-causing mutants in HEK-293T cells transfected with the indicated plasmids, followed by treating with dFUS-E9C (2 μM) for 24 h. Scale bar, 20 μm. i Quantification of the percentage of cells with GFP-FUS aggregates. j Streptavidin pulldown assays to determine the interactions between CRBN and FUS mutants in the presence or absence of dFUS-E9C. k HEK-293T cells were transfected with His-Myc-Ub plasmids together with GFP-FUS-WT, GFP-FUS-P525L, GFP-FUS-R521C, GFP-FUS-R521G, and GFP-FUS-R521H plasmids, followed by treatment with dFUS-E9C (2 μM) and MG132 (10 μM) for 24 h. Cell lysates were subjected to immunoprecipitation with the GFP antibody, followed by immunoblotting with the Myc and GFP antibodies. The averages of n = 3 (e, g, i) biologically independent samples are shown. Data are shown as the mean ± SEM. Statistical significance in (g, i) was assessed using two-tailed t tests. The data presented in (d, f, h, j, k) are representative of three independent experiments. Source data are provided as a Source Data file.

Fig. 4. DNA nanoflowers efficiently deliver dFUS-PS9C into the brain.

a, b Native PAGE analysis of the serum stability of FUS-RNA#E9, dFUS-E9C, and dFUS-PS9C (a), and the remaining Oligo levels at the indicated time points were quantified by densitometry (b). c, d Immunoblotting of FUS in SH-SY5Y cells transfected with increasing concentrations of dFUS-PS9C for 24 h (c), and the relative levels of FUS were quantified by densitometry (d). e Schematic showing the synthesis of DNA nanoflowers through rolling circle amplification and the generation of FRONTAC^FUS by functionalizing DNA nanoflowers with dFUS-PS9C. FRONTAC^FUS crosses the blood-brain barrier via TfR-mediated transcytosis. f Fluorescence spectral scanning of DNA nanoflowers (NF) and Cy5-labeled FRONTAC^FUS. g Dynamic light scattering to determine the particle size of NF and FRONTAC^FUS. h The loading efficiency of FRONTAC^FUS was determined by measuring the Cy5 signals. i Zeta potential analysis of dFUS-PS9C, NF, and FRONTAC^FUS. j Transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images of NF and FRONTAC^FUS. Scale bars, 200 nm for TEM and 300 nm for SEM. k Immunofluorescence images of SH-SY5Y cells treated with PBS, Cy5-dFUS-PS9C, and Cy5-FRONTAC^FUS, followed by staining with LysoTracker (green) and DAPI (blue). Scale bar, 20 μm. l Immunoblotting analysis of FUS in SH-SY5Y cells treated with increasing concentrations of FRONTAC^FUS for 24 h, and the relative FUS levels were quantified by densitometry. m Serum stability determination of Cy5-dFUS-PS9C and Cy5-FRONTAC^FUS. n In vivo and ex vivo images of mice intravenously injected with PBS, sc-FRONTAC^FUS, and FRONTAC^FUS. The averages of n = 3 (b, d, h, i, m) biologically independent samples are shown. Data are shown as the mean ± SEM. The data presented in (a, c, f–n) are representative of three independent experiments. Source data are provided as a Source Data file.

Fig. 5. Peripherally administered FRONTAC^FUS degrades FUS in the mouse brain.

a In vivo and ex vivo images of mice intravenously injected with PBS, Cy5-dFUS-PS9C, and Cy5-FRONTAC^FUS. b Real-time monitoring of brain Cy5 signals in mice administered (i.v.) with Cy5-FRONTAC^FUS. c Fluorescence images of brain sections from mice receiving Cy5-dFUS-PS9C or Cy5-FRONTAC^FUS through i.v. injection. Scale bar, 50 μm. d Immunoblots of the hippocampus and cortex area in mice receiving i.v. injection of different doses of FRONTAC^FUS. e, f Immunoblots (e) and quantification (f) of the hippocampus and cortex area in mice receiving i.v. injection of NF control and FRONTAC^FUS (10 mg/kg). g, h Immunohistochemistry images (g) and quantification (h) of brain sections in mice receiving i.v. injection of NF control and FRONTAC^FUS. Scale bar, 50 μm. i, j The number of red blood cells (RBCs) and reticulocytes (RETs) in mice receiving i.v. injection of NF control and FRONTAC^FUS were determined. The averages of n = 3 (f, h–j) biologically independent samples are shown. Data are shown as the mean ± SEM. Statistical significance in (f, h–j) was assessed using two-tailed t tests. The data presented in (a–e, g) are representative of three independent experiments. Source data are provided as a Source Data file.

Fig. 6. Sustained degradation of FUS by FRONTAC^FUS with no significant toxicity.

a Schematic of the administration of FRONTAC^FUS in BALB/c mice. Created in BioRender. Ge, R. (2025) https://BioRender.com/0t3564g. b, c The mice were intravenously administered with PBS control or FRONTAC^FUS (single dose, 10 mg/kg) for 7 days and 14 days, and the hippocampus and cortex were isolated for immunoblotting (b), and the levels of FUS were quantified by densitometry (c). d, e The representative hind-limb splay images (d) and statistical analysis (e) of mice after administration of PBS control or FRONTAC^FUS. f Measurement of body weight at the indicated intervals after administration of PBS or FRONTAC^FUS. g, h The serum was obtained at the indicated time points, and IL-6 (g) and TNF-α (h) concentrations were determined by ELISA. i, j Immunohistochemistry images (i) and quantification (j) of mouse brain sections stained with antibodies against GFAP and IBA1. Scale bar, 50 μm. The averages of n = 3 (c, e–h, j) biologically independent samples are shown. Data are shown as the mean ± SEM. Statistical significance in (c, f–h, j) was assessed using two-tailed t tests. The data presented in (b, d, i) are representative of three independent experiments. Source data are provided as a Source Data file.