RNA biosensor and gene therapy controller using disease-specific cryptic EXON retention

Abstract

Provided herein is a composition including a first nucleic acid, an excisable second nucleic acid, a third nucleic acid, and one or more cleavable fourth nucleic acids encoding a self-cleavable peptide. The composition further includes a recombinant transducing particle or viral vector and a pharmaceutically acceptable carrier. The composition may be applied for method of detecting aberrant expression or dysfunction of a protein in a cell.

Description

RNA BIOSENSOR AND GENE THERAPY CONTROLLER USING DISEASESPECIFIC CRYPTIC EXON RETENTIONCROSS REFERENCE TO RELATED APPLICATION

[0001] The present application claims priority to U.S. Provisional Patent Application No. 63 / 662,566, filed June 21 , 2024, the disclosure of which is hereby incorporated by reference in its entirety.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0002] This invention was made with government support under Grant Nos. NS105756 and NS127187 awarded by the National Institutes of Health. The government has certain rights in the invention.REFERENCE TO A SEQUENCE LISTING

[0003] The Sequence Listing associated with this application is filed in electronic format via Patent Center and is hereby incorporated by reference into the specification in its entirety. The name of the file containing the Sequence Listing is 2501390. xml. The size of the file is 28,577 bytes, and the file was created on June 18, 2025.BACKGROUND OF THE INVENTIONField of the Invention

[0004] Provided herein are sensors and therapeutic compositions and methods for detecting or treating a condition with aberrant mRNA splicing mechanisms, such as loss of TDP-43 function, for example, associated with neurodegenerative diseases such as Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Degeneration (FTD), and Alzheimer’s Disease (AD).Description of Related Art

[0005] Amyotrophic lateral sclerosis (ALS) is a progressive and fatal neurodegenerative disease (NND) characterized by a persistent degeneration of the neurons of the spinal cord and motor cortex. The dysregulation of the RNA binding protein (RBP), TAR DNA-binding protein 43 (TDP-43) is a hallmark pathobiology observed in ~97% of all ALS patients that, ~45% of Frontotemporal lobar degeneration (FTLD) patients, and 40%-60% of Limbic Associated TDP-43 Encephalopathy (LATE) patients. Under physiological conditions, TDP-43 orchestrates many cellular processes critical for neuronal health and homeostasis, including regulating RNA metabolism, splicing, and stress response pathways. In disease, TDP-43 is depletedfrom the nucleus and mislocalizes to the cytoplasmic compartment, losing the ability to perform its canonical functions and transitioning into insoluble aggregates.

[0006] Current gene therapy approaches, for example those targeting TDP-43, lack specificity and control, leading to potential risks such as side effects from delivery to non-pathological cells, overexpression, and an inability to deactivate the therapy once the pathology is resolved. Accordingly, there is a need in the art to provide improved methods for identifying and treating dysregulation, including in TDP-43.SUMMARY OF THE INVENTION

[0007] Provided herein is a composition including a first nucleic acid, an excisable second nucleic acid, a third nucleic acid, and one or more cleavable fourth nucleic acids encoding a self-cleavable peptide.

[0008] Also provided herein is a composition including a recombinant transducing particle or viral vector, a composition as described herein, and a pharmaceutically- acceptable carrier.

[0009] Also provided herein is a method of detecting aberrant expression or dysfunction of a protein in a cell, including: delivering to the cell a as described herein, wherein the third nucleic acid is translated or expressed when the protein is aberrantly expressed or is dysfunctional and the third nucleic acid is not expressed when the protein is not aberrantly expressed or is not dysfunctional.

[0010] Also provided herein is a method of treating a patient, including introducing a composition as described herein into a cell of the patient, wherein the third nucleic acid encodes a therapeutic composition, such that the third nucleic acid is expressed in the cell of the patient.

[0011] Also provided herein is a nucleic acid including, from a 5’ to a 3’ direction, SEQ ID NO: 1 , SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, and SEQ ID NO: 7.

[0012] Also provided herein is a nucleic acid including, from a 5’ to a 3’ direction, SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 16, and SEQ ID NO: 7.

[0013] Also provided herein is a nucleic acid including, from a 5’ direction to a 3’ direction, SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, and SEQ ID NO: 7.

[0014] Also provided herein is a method of detecting aberrant expression or dysfunction of TDP-43 in a cell of a patient, including introducing a composition as described herein into a cell of the patient.

[0015] Further non-limiting embodiments are provided in the following numbered clauses:

[0016] 1. A composition comprising: a first nucleic acid; an excisable second nucleic acid; a third nucleic acid; and one or more cleavable fourth nucleic acids encoding a self-cleavable peptide.

[0017] 2. The composition of clause 1 , wherein the first nucleic acid encodes a first reporter peptide.

[0018] 3. The composition of clause 1 or clause 2, wherein the excisable second nucleic acid comprises a cryptic exon.

[0019] 4. The composition of clause 3, wherein the cryptic exon comprises an exon from a cystic fibrosis transmembrane conductance regulator (CFTR) gene.

[0020] 5. The composition of clause 3 or clause 4, wherein the cryptic exon comprises an intron from an llnc-13 homolog A (UNC13A) gene.

[0021] 6. The composition of clause 3, wherein the cryptic exon comprises exon 9 from CFTR and intron 20 from UNC13A.

[0022] 7. The composition of any of clauses 1 -6, wherein the second nucleic acid comprises one or more binding sites for transactive response DNA-binding protein 43 (TDP-43).

[0023] 8. The composition of clause 7, wherein the one or more binding sites forTDP-43 comprise one or more UG-rich regions.

[0024] 9. The composition of clause 1 , wherein the nucleic acids are ribonucleic acids.

[0025] 10. The composition of any of clauses 1 -9, wherein the first and third nucleic acids encode a fluorescent protein.

[0026] 11. The composition of any of clauses 1-10, wherein the first nucleic acid encodes mCherry and the third nucleic acid encodes EGFP.

[0027] 12. The composition of any of clauses 1-11 , wherein the third nucleic acid encodes a therapeutic peptide.

[0028] 13. The composition of clause 12, wherein the therapeutic peptide is a peptide useful for treating a neurodegenerative disease.

[0029] 14. The composition of clause 13, wherein the neurodegenerative disease is Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Degeneration (FTD), and / or Alzheimer’s Disease (AD).

[0030] 15. The composition of any of clauses 1 -14, wherein the fourth nucleic acid comprises a T2A self-cleavage sequence.

[0031] 16. The composition of any of clauses 1-15, wherein the nucleic acid comprises, from a 5’ end to a 3’ end: the first nucleic acid; the fourth nucleic acid; the excisable second nucleic acid; a stop codon, wherein the stop codon is out of frame when the excisable second nucleic acid is present in the composition and the stop codon is in frame when the excisable second nucleic acid is excised from the composition; the fourth nucleic acid; and the third nucleic acid, the third nucleic acid encoding a reporter peptide or a therapeutic peptide; wherein, when introduced into a cell, the cell translates or expresses the third nucleic acid, thereby producing a reporter or therapeutic peptide, when the excisable second nucleic acid is excised.

[0032] 17. The composition of any of clauses 1 -16, wherein the composition comprises a recombinant transducing particle or a viral expression vector comprising the first, second, third, and fourth nucleic acids.

[0033] 18. The composition of clause 17, wherein the viral expression vector is a lentiviral, an adenoviral, an adeno-associated virus (AAV), or a herpesvirus recombinant genome, or a plasmid DNA vector.

[0034] 19. A composition comprising the recombinant transducing particle or viral vector of clause 17 or clause 18 in a pharmaceutically-acceptable carrier.

[0035] 20. A method of detecting aberrant expression or dysfunction of a protein in a cell, comprising: delivering to the cell the composition of any of clauses 1 -19, wherein the third nucleic acid is translated or expressed when the protein is aberrantly expressed or is dysfunctional and the third nucleic acid is not expressed when the protein is not aberrantly expressed or is not dysfunctional.

[0036] 21. The method of clause 20, wherein translation or expression of the third nucleic acid correlates to the level of aberrance of expression of dysfunction of the protein.

[0037] 22. The method of clause 20 or clause 21 , wherein the protein is TDP-43.

[0038] 23. A method of treating a patient, comprising introducing the composition of any of clauses 1 -19 into a cell of the patient, wherein the third nucleic acid encodesa therapeutic peptide, such that the third nucleic acid is expressed in the cell of the patient.

[0039] 24. The method of clause 23, wherein the patient has a neurodegenerative disease.

[0040] 25. The method of clause 24, wherein the neurodegenerative disease is Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Degeneration (FTD), or Alzheimer’s Disease (AD).

[0041] 26. A nucleic acid comprising, from a 5’ to a 3’ direction, SEQ ID NO: 1 , SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, and SEQ ID NO: 7.

[0042] 27. The nucleic acid of clause 26 further comprising, at the 3’ end, SEQ ID NO: 9.

[0043] 28. The nucleic acid of clause 26 or clause 27, further comprising an additional nucleic acid encoding a reporter peptide or a therapeutic peptide, the additional nucleic acid arranged at a 3’ end of SEQ ID NO: 7.

[0044] 29. The nucleic acid of clause 28, wherein the nucleic acid encoding the reporter peptide has the sequence of SEQ ID NO: 8.

[0045] 30. A nucleic acid comprising, from a 5’ to a 3’ direction, SEQ ID NO: 10,SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 16, and SEQ ID NO: 7.

[0046] 31. The nucleic acid of clause 30, further comprising, at a 3’ end of SEQ IDNO: 7, SEQ ID NO: 8.

[0047] 32. The nucleic acid of clause 30 or clause 31 , further comprising, at the 3’ end, SEQ ID NO: 9.

[0048] 33. A nucleic acid comprising, from a 5’ direction to a 3’ direction, SEQ IDNO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, and SEQ ID NO: 7.

[0049] 34. A method of detecting aberrant expression or dysfunction of TDP-43 in a cell of a patient, comprising introducing the nucleic acid of any of clauses 26-33 into a cell of the patient.

[0050] 35. The method of clause 34, wherein the cell is obtained from a patient sample.

[0051] 36. The method of clause 34 or clause 35, wherein the patient sample is a blood sample.BRIEF DESCRIPTION OF THE DRAWINGS

[0052] FIGS. 1A-1 J show comparisons of stable polyclonal HEK cells expressing UNC13A-TS, CFTR-TS, or CUTS following treatment with siRNA control (siControl) (20nM) or TDP-43 (siTDP-43) (0.6nM - 20nM). Cells were reverse transfected with siRNA treatment in complete media supplemented with doxycycline (1000 ng / mL). After 72 h, cells were analyzed by live imagining and protein lysate was harvested for western blot analysis. (A) Schematic of the TDP-43 loss of function Sensor (TS) system design. (B) Overview of the UNC13A-TS, CFTR-TS, and CUTS cassette design. (C) Representative live imaging of TS comparison (10X). (D) Mean intensity quantification of GFP signal intensity as shown in (C). (E-G) Western blot analysis of (E) UNC13A-TS, (F) CFTR-TS, and (G) CUTS developing again GFP and TDP-43 proteins. (H-J) Relative pixel quantification of GFP and TDP-43 band normalized to total protein (Ponceau S) for the indicated TS from E-G. Statistical significance was determined by one-way ANOVA and Tukey's multiple comparison test (* = P < 0.03; ** = P < 0.002; *** = P < 0.0002; **** = P < 0.0001 ). Green = GFP signal; red = mCherry signal. Scale bar = 100 pm. N=3 biological replicates.

[0053] FIG. 2 shows Comparison of stable polyclonal HEK cells expressing UNC13A-TS, CFTR-TS, or CUTS following treatment with siRNA control (siControl) (20nM) or TDP-43 (siTDP-43) (0.6nM - 20nM). Cells were reverse transfected with siRNA treatment in complete media supplemented with doxycycline (1000 ng / mL). After 72 h, cells were analyzed by live imagining and protein lysate was harvested for western blot analysis. (A) Mean intensity quantification of GFP signal intensity from live imaging, presented as fold change from mock. (B) Relative pixel quantification of GFP normalized to total protein (Ponceau S), presented as fold change from mock. Statistical significance was determined by two-way ANOVA and Tukey's multiple comparison test (* = P < 0.03; ** = P < 0.002; *** = P < 0.0002; **** = P < 0.0001 ). N=3 biological replicates.

[0054] FIGS. 3A-3G show Low-dose siRNA TDP-43 (siTDP-43) treatment was performed in stable polyclonal HEK cells expressing CUTS. CUTS-expressing cells were reverse transfected with siRNA control (siControl) or siTDP43 in a doseresponse curve (38 to 1200pM) in doxycycline supplemented media (1000ng / ml) for 72hr. (A) Representative immunofluorescence images of CUTS-expressing HEK cells under low doses of siRNA TDP-43 treatment. (60X). (B) Mean intensity quantification of GFP signal from (A) with normalization to the number mCherry positive cells. (C)Western blot of GFP and TDP-43 proteins from HEK cell lysate expressing CUTS under low doses of siRNA TDP-43. Ponceau S is shown as a loading control. (D) Pixel intensity quantification of the GFP and TDP-43 bands shown in (C), presented as foldchange from the mock-treated sample. (E) Schematic showing the position of qPCR primers, developed to detect CUTS cryptic exon inclusion (referred to as 'CUTS-CE' and 'CUTS-J'). (F) Representative agarose gel showing qPCR product from melting curve detecting CUTS cryptic exon inclusion using the primers shown in (E). (G) qPCR quantification of the siTDP-43 dose curve presented as fold-change from the mock- treated sample. Purple text indicates the GFP fold change from the mock-treated sample. Red text indicates the percentage of total detectable TDP-43 knockdown. Linear regression analysis shown in (D) and (G) was performed on Log values. Fitting method = least squares regression. Green = GFP; red = mCherry. Scale bar = 50 pm. N=3 biological replicates.

[0055] FIGS. 4A-4C show Stable HEK cells expressing CUTS were induced with doxycycline (1000 ng / mL) for 24 hours before transfection with the following plasmids: pCMV backbone, TDP-43WT, TDP-43ANLS, TDP-435FL, TDP-43ANLS 5FL, or nontransfected. Following transfection, plasmids were expressed for 72 h, followed by live imaging and protein analysis. (A) Live-imaging of CUTS HEK cells expressing WT or mutant TDP-43 gene cassettes. (B) Representative western blot of exogenous and endogenous GFP and TDP-43. Ponceau S is shown as a loading control. (C) Relative GFP pixel intensity quantification of the band is shown in (B). Statistical significance was determined by one-way ANOVA and Tukey's multiple comparison test (* = P < 0.03; ** = P < 0.002; *** = P < 0.0002; **** = P < 0.0001 ). Green = GFP; red = mCherry. Scale bar = 100 pm. N=3 biological replicates.

[0056] FIGS. 5A-5C show Transient CUTS expression in Hela TDP-43 KO we induced with doxocycline (1000n / ml) for 24 hours before transfection of pCM backbone or TDP-43WT plasmids. Following transfection, plasmids were expressed for 72 h, followed by live imaging and protein analysis. (A) Live-imaging of Hela TDP- 43 KO expressing CUTS in combination with TDP-43WT or pCMV backbone control. (B) Representative WB of exogenous and endogenous GFP and TDP-43. Ponceau S is shown as a loading control. (C) Relative GFP pixel intensity quantification of the protein bands shown in (B). Statistical significance was determined by one-way ANOVA and Tukey's multiple comparison test (* = P < 0.03; ** = P < 0.002; *** = P <0.0002; **** = P < 0.0001 ). Green = GFP; red = mCherry. Scale bar = 100 pm. N=3 biological replicates.

[0057] FIGS. 6A-6E show (A) Schematic of CUTS as an autoregulatory controller of TDP-43 expression (CUTS-TDP43). (B-C) TDP-43 siRNA (siTDP43) doseresponse curve in stable polyclonal HEK cells expressing CUTS, CUTS-TDP43, or CUTS-TDP43 (codon optimized). The codon-optimized variation allows for continued expression during siTDP-43 treatment. HEK cells expressing the CUTS, CUTS-TDP- 43 and CUTS-TDP-43 codon optimize system were reverse transfected with control siRNA (siControl) or siTDP43 in a dose-response curve (0.6nM-20nM) in a doxycycline (1000ng / ml) supplement media for 72hr. Cells were then used for live imaging or protein analysis. (B) Live imaging of the CUTS variants. (C) Immunoblot assay of GFP and TDP-43. Ponceau S is shown as a loading control. (D-E) CFTR minigene assay in stable CUTS or CUTS-TDP43 (codon optimized) expressing HEK cells. Cells were induced with doxycycline (1000 ng / mL) for 24 h before transfection with the CFTR minigene. Following an additional 24h of expression, cells were transfected with 20nM siControl or siTDP-43. Cells were harvested 48 h following siRNA transfection for RNA extraction and RT-PCR analysis. (D) PCR agarose gel of CFTR minigene. (E) PCR analysis of the ratio between the CFTR cryptic exon inclusion and the correctly spliced product from CFTR as shown in (D). Statistical significance was determined by student t-test (* = P < 0.03; ** = P < 0.002; *** = P < 0.0002; **** = P < 0.0001 ). Green = GFP; red = mCherry. Scale bar = 100 pm. N=3 biological replicates.

[0058] FIGS. 7A-7O show (A) Schematic diagram of the experimental design of normal CUTS with siRNA knockdown of TDP-43’s interactors. (B) Representative live confocal images of the stable HEK293 cells expressing doxycycline-inducible (72 h of 1 mg / mL doxycycline) CUTS with reverse transfected siRNA control (siControl) or the top hits that cause LOF in the preliminary screening in (A) (20 nM, 96 h). (C) Relative GFP fluorescence intensity quantification normalized to siControl from live confocal imaging in (C) (N=3 biological replicates). (D) Schematic diagram of the experimental design of CUTS under partial TDP-43 knockdown (siTDP-43) with siRNA knockdown of TDP-43’s interactors. (E) Representative live confocal images of the stable HEK293 cells expressing doxycycline-inducible (72 h of 1 mg / mL doxycycline) CUTS with reverse co-transfected siTDP-43 (5 nM, 96 h) with siControl or the top hits that cause or rescue LOF in the preliminary screening in (A) (20 nM, 96 h). (F) Western blotanalysis of the stable HEK293 cells expressing doxycycline-inducible (72 h of 1 mg / mL doxycycline) CUTS with reverse co-transfected siTDP-43 (5 nM, 96 h) with siControl or siCDC40, siHNRNPA3, and siPRPF31 (20 nM, 96 h) that exacerbate the LOF in (E). (G) Western blot analysis of the stable HEK293 cells expressing doxycycline- inducible (72 h of 1 mg / mL doxycycline) CUTS with reverse co-transfected siTDP-43 (5 nM, 96 h) with siControl or siHNRNPAO, siHNRNPC, siMATR3 (20 nM, 96 h) that rescue the LOF in (E). (H) Relative GFP fluorescence intensity quantification normalized to siControl (+siTDP-43) from live confocal imaging in (E) (N=3 biological replicates). (I) Schematic diagram of the experimental design of normal CUTS with cDNA overexpression of TDP-43’s interactors. (J) Representative live confocal images of the stable HEK293 cells expressing doxycycline-inducible (72 h of 1 mg / mL doxycycline) CUTS with control (mock) or transfection of CMV-driven NUFIP2 cDNA in a plasmid (500 ng, 72 h). For (B), (E), and (J), Green = GFP signal; Red = mCherry signal. Scale bar = 100 pm. (K) Western blot analysis of the cells in (J). (L) Relative GFP protein intensity quantification normalized to control from western blot analysis of the stable HEK293 cells expressing doxycycline-inducible (72 h of 1 mg / mL doxycycline) CUTS with control (mock) or transfection of five available cDNA plasmids of the hits driven by CMV promoter (500 ng, 72 h). (N=3 biological replicates). (M) Schematic diagram of the experimental design of CUTS under partial siTDP-43 with cDNA overexpression of TDP-43’s interactors. (N) Western blot analysis of the stable HEK293 cells expressing doxycycline-inducible (72 h of 1 mg / mL doxycycline) CUTS with reverse transfected siTDP-43 (5 nM, 96 h) and control (mock) or transfection of five available cDNA plasmids of the hits driven by CMV promoter (500 ng, 72 h). (O) Relative GFP protein intensity quantification normalized to control from western blot analysis in (N) (N=3 biological replicates).

[0059] For (C), (H), (L), and (O), statistical significance was determined by one- sample two-sided t-test (* = P < 0.05; ** = P < 0.01 ; *** = P < 0.001 ; **** = P < 0.0001 ). Data are the mean ± sd.

[0060] FIGS. 8A-8J show (A) Schematic diagram of TDP-43 knockdownreplacement (TKR) experimental design. Three HEK293 stable cell lines built by Piggybac were leveraged. TKR1 : expressing only CUTS, as the baseline control with normal TDP-43 function; TKR2: expressing CUTS together with shRNA targeting endogenous TDP-43, as the positive control with TDP-43 loss of function (LOF); TKR3: expressing shRNA targeting endogenous TDP-43 and CUTS-controlledexpressing of codon-optimized EGFP-TDP-43, as the TKR therapy group. All the three constructs were activated by doxycycline-induced expression for 4 days with N=3 biological replicates. (B) Schematic diagram of CUTS-controlled HNRNPC knockdown (CCK) therapy’s experiment design. Two HEK293 stable cell lines with four different siRNA treatment groups were used. CCK1 : CUTS-shControl cell line with siControl, as the baseline control with normal TDP-43 function; CCK2: CUTS-shControl cell line with siTDP-43, as the positive control with TDP-43 LOF; CCK3: CUTS-shHNRNPC cell line with siControl, as the inactive control for CUTS-shHNRNPC with normal TDP- 43 function; CCK4: CUTS-shHNRNPC cell line with siTDP-43, as the CCK therapy group. All the groups were activated by siRNA treatment for 7 days with N=3 biological replicates. (C) Schematic diagram of the mRNA sequencing samples design and the rescue ratio for both TKR and CCK groups. The rescue ratio for each group was calculated from the changing in the percent spliced-in (APS I, calculated by MAJ IQ) of specific splicing junctions, showing as the ratio between rescued APSI (positive control vs therapy group) and LOF APSI (positive control vs baseline control). (D) Venn diagram of the significantly changed (|APSI|>0.1 , probability_of_changing>0.95, calculated by MAJIQ) splicing junctions among the positive controls from TKR, CCK experiments and a previous study in iPSC-derived model. Splicing junctions with overlapping of at least 2 groups with the same direction of APSIs were selected as the evaluation indexes for measuring TDP-43 LOF. (E) Co-relationship analysis of the rescue ratio between TKR and CCK therapies over the selected junctions from F. Cryptic junctions were labeled as they indicated the cryptic exon (CE) retention caused by TDP-43 LOF. (F) Bar plot of the rescue ratio from TKR (left) and CCK (right) of 28 cryptic junctions in (G). The junctions were ranked by TKR rescue ratio, from high to low. (G) Bar plot analysis of the PSI values of the cryptic junction in HDGFL2 (492152-4493703) in TKR groups (left) and CCK groups (right), Bar plot analysis of the PSI values of the cryptic junction in ACBD3 (226156780-226159184) in TKR groups (left) and CCK groups (right), and Bar plot analysis of the PSI values of the cryptic junction in EPB41 L4A (112267284-112275326) in TKR groups (left) and CCK groups (right). (H) Bar plot analysis of the PSI values of the cryptic junction in STMN2 (79611214-79616822) in CCK groups. For G-H, the grouped (bars) and individual (points) PSI values samples were calculated separately by MAJIQ. (J) Representative splicing plots of the cryptic exon (Exon 9) in HDGFL2 in TKR and CCK groups by Voila. (K) Two possible working models of HNRNPC regulating TDP-43’s splicing CE targets.HNRNPC can be a competitive antagonist of TDP-43 to its binding site on pre-mRNA (left): Under normal TDP-43 function, HNRNPC is blocked from TDP-43’s binding sites the splicing of CE is repressed; Without TDP-43’s binding (LOF), HNRNPC binds to these sites and promotes splicing, resulting in CE retention; If HNRNPC is further knocked-down, the CE retention is partially rescued. HNRNPC can also be a noncompetitive antagonist (right) with binding to the CE sites next to TDP-43 and promotes splicing. However, under normal condition HNRNPC is repressed by TDP- 43; Without TDP-43, this inhibition is released and HNRNPC promotes CE retention.

[0061] FIGS. 9A-9I show (A) Schematic diagram of two novel rescue approaches rescuing TDP-43 LOF by cis-targeting TDP-43 or trans-regulating TDP-43’s interactors. (B) Schematic diagram of the plasmids design to generate stable HEK293 cell lines with TDP-43 knockdown-replacement (TKR) strategy. Endogenous TDP-43 was knocked down by shRNA and replaced with doxycycline-inducible CUTS- controlled expression of codon-optimized EGFP-TDP-43. (C) Representative live confocal images of wildtype HEK293 or the stable cell lines expressing TKR or control systems in (B) with or without doxycycline (96 h of 1 mg / mL doxycycline). Green = GFP signal; Red = mCherry signal. Scale bar = 100 pm. (D) Western blot analysis of the cells in (C). (E) Percentage of endogenous or exogenous TDP-43 protein levels normalized to wildtype control from western blot analysis in (D). (F) Schematic diagram of the plasmid design to generate stable HEK293 cell lines with CUTS-controlled HNRNPC knockdown (CCK) rescue. EGFP-NLS-coupled shRNA targeting HNRNPC is controlled by CUTS through the Tet-On 3G system (TRE3g) to enable LOF-specific activation. (G) Representative live confocal images of wildtype HEK293 or the stable cell lines expressing CUTS-shHNRNPC or CUTS-shControl systems in (F) with reverse transfected siControl or siTDP-43 (20 nM, 7 days). Green = GFP signal; Red = mCherry signal. Scale bar = 100 pm. (H) Western blot analysis of the cells in (G). (I) Representative image of RT-PCR products from TDP-43-regulated endogenous cryptic exon targets (ATG4B, DNAJC5, and HDGFL2) in stable HEK293 cells from (H). Images represent similar results in two independent experiments.DESCRIPTION OF THE INVENTION

[0062] The use of numerical values in the various ranges specified in this application, unless expressly indicated otherwise, are stated as approximations as though the minimum and maximum values within the stated ranges are both precededby the word "about". In this manner, slight variations above and below the stated ranges can be used to achieve substantially the same results as values within the ranges. Also, unless indicated otherwise, the disclosure of ranges is intended as a continuous range including every value between the minimum and maximum values. As used herein, "a" and "an" refer to one or more.

[0063] As used herein, the term "comprising" is open-ended and may be synonymous with 'including', 'containing', or 'characterized by'. The term "consisting essentially of' limits the scope of a claim to the specified materials or steps, and those that do not materially affect basic and novel characteristic(s). The term "consisting of" excludes any element, step, or ingredient not specified in the claim. As used herein, embodiments "comprising" one or more stated elements or steps also include but are not limited to embodiments "consisting essentially of" and "consisting of" these stated elements or steps.

[0064] By "expression" or “gene expression,” it is meant the overall flow of information from a gene. A “gene” is a functional genetic unit for producing a gene product, such as RNA or a protein in a cell, or other expression system encoded on a nucleic acid and generally comprising: a transcriptional control sequence, such as a promoter and other cis-acting elements, such as transcriptional response elements (TREs) and / or enhancers; an expressed sequence that typically encodes a protein (referred to as an open-reading frame or ORF) or functional / structural RNA; and a polyadenylation sequence). A gene produces a gene product (typically a protein, optionally post-translationally modified or a functional / structural RNA) when transcribed. By "expression of genes under transcriptional control of," or alternately "subject to control by," a designated sequence such as a promotor, it is meant gene expression from a gene containing the designated sequence operably linked (functionally attached, typically in cis) to the gene. A gene that is “under transcriptional control” of a promotor or transcription control element, is a gene that is transcribed at detectably different levels in the presence of a transcription factor, e.g., in the presence of a suitable chemical compound, such as doxycycline in the case of a dox-responsive promoter, such as a tet-inducible promoter. Another promoter example is a Pol III promoter, which are useful for production of small RNA molecules (see, e.g., Ma H, et al. Wu H. Pol III Promoters to Express Small RNAs: Delineation of Transcription Initiation. Mol Ther Nucleic Acids. 2014 May 6;3(5):e161 ).

[0065] A "gene for expression of" a stated gene product, such as an sgRNA or a dCas transcription activator is a gene capable of expressing that stated gene product when placed in a suitable environment, that is, for example, when transformed, transfected, transduced, etc. into a cell, and subjected to suitable conditions for expression. In the case of a constitutive promoter "suitable conditions" means that the gene typically need only be introduced into a host cell. In the case of an inducible promoter, such as a tissue-specific promoter, "suitable conditions" means when factors that regulate transcription, such as DNA-binding proteins, are present or absent, for example, an amount of the respective inducer is available to the expression system (e.g., cell), or factors causing suppression of a gene are unavailable or displaced - effective to cause expression of the gene.

[0066] Transcriptional control elements include promoters, enhancers, transcription factor-responsive elements (TREs, e.g., transcription factor binding sequences), suppressors, introns, etc., as are broadly-known. Additional transcription control elements, such as a WPRE (woodchuck hepatitis virus post-transcriptional regulatory element) can be included in a gene.

[0067] A nucleic acid molecule (a nucleic acid) refers to a polymeric form of nucleotides, which may include both sense and anti-sense strands of RNA, cDNA, genomic DNA, gRNA, plasmid DNA, viral DNA, and synthetic forms and mixed polymers of the above. As used herein, “nucleic acid” may mean both DNA and RNA, with the understanding that a DNA product may be delivered to a cell, where it may be transcribed to an RNA. Accordingly, references herein to a composition that includes a nucleic acid should be understood to include, within the scope thereof, a composition that is DNA, a composition that is RNA, and / or a composition that is a DNA and that, once delivered to a cell, is transcribed to an RNA. Thus, nucleic acid sequences disclosed in this application and in the appended sequence listing, while referred to or categorized as DNA, should be understood to include RNA sequences (e.g., uracil in place of thymine) and RNA transcribed from a DNA sequence (e.g., a disclosed sequence that includes the nucleotides ACGGCA encompasses the transcribed sequence UGCCGU). A nucleotide refers to a ribonucleotide, deoxynucleotide or a modified form of either type of nucleotide. The term "nucleic acid molecule" as used herein is synonymous with "nucleic acid" and "polynucleotide." The term includes single- and double-stranded forms of DNA. A polynucleotide may include either or bothnaturally occurring and modified nucleotides linked together by naturally occurring and / or non-naturally occurring nucleotide linkages.

[0068] A first nucleic acid is said to be operably linked to a second nucleic acid when the first nucleic acid is placed in a functional relationship with the second nucleic acid. Generally, operably linked DNA sequences are contiguous (e.g., in cis) and, where the sequences act to join two protein coding regions, in the same reading frame (e.g., open reading frame or ORF), for example to produce a fusion protein. Operably linked nucleic acids include a first nucleic acid contiguous with the 5’ or 3’ end of a second nucleic acid. In examples, a second nucleic acid may be considered operably linked to a first nucleic acid when it is embedded within the first nucleic acid, for example, where the nucleic acid construct includes (in order) a portion of the first nucleic acid, the second nucleic acid, and the remainder of the first nucleic acid.

[0069] A "codon-optimized" nucleic acid refers to a nucleic acid sequence that has been altered such that the codons are optimal for expression in a particular system (such as a particular species of group of species). For example, a nucleic acid sequence can be optimized for expression in human cells. Codon optimization does not alter the amino acid sequence of the encoded protein.

[0070] Complementary refers to the ability of polynucleotides (nucleic acids) to hybridize to one another, forming inter-strand base pairs. Base pairs are formed by hydrogen bonding between nucleotide units in antiparallel polynucleotide strands. Complementary polynucleotide strands can base pair (hybridize) in the Watson-Crick manner (e.g., A to T, A to U, C to G), or in any other manner that allows for the formation of duplexes. When using RNA as opposed to DNA, uracil rather than thymine is the base that is considered to be complementary to adenosine. Two sequences comprising complementary sequences can hybridize if they form duplexes under specified conditions, such as in water, saline (e.g., normal saline, or 0.9% w / v saline) or phosphate-buffered saline), or under other stringency conditions, such as, for example and without limitation, 0.1X SSC (saline sodium citrate) to 10X SSC, where 1X SSC is 0.15M NaCI and 0.015M sodium citrate in water. Hybridization of complementary sequences is dictated, e.g., by salt concentration and temperature, with the melting temperature (Tm) lowering with increased mismatches and increased stringency. Perfectly matched sequences are said to be fully complementary, or have 100% sequence identity (gaps are not counted and the measurement is in relation to the shorter of the two sequences). In one example, a sequence that "specificallyhybridizes" to another sequence, does so in a hybridization solution containing 0.5M sodium phosphate buffer, pH 7.2, containing 7% SDS, 1 mM EDTA, and 100 mg / ml of salmon sperm DNA at 65° C for 16 hours and washing twice at 65° C for twenty minutes in a washing solution containing 0.5xSSC and 0.1 % SDS, or does so under conditions more stringent than 2X SSC at 65°C, for example, in 0.2X SSC at 55°C. A sequence that specifically hybridizes to another typically has at least 80%, 85%, 90%, 95%, OR 99% sequence identity with the other sequence.

[0071] A recombinant nucleic acid refers to a nucleic acid molecule (or protein or virus) that is not naturally occurring or has a sequence that is made by an artificial combination of two otherwise separated segments of sequence. This artificial combination is accomplished by chemical synthesis or, more commonly, by the artificial manipulation of isolated segments of nucleic acids. The term recombinant includes nucleic acids and proteins that have been altered solely by addition, substitution, or deletion of a portion of a natural nucleic acid molecule or protein.

[0072] "Sequence identity" refers to the similarity between nucleic acid or amino acid sequences is expressed in terms of the similarity between the sequences, otherwise referred to as sequence identity. Sequence identity may be measured in terms of percentage identity (or similarity or homology); the higher the percentage, the more similar the two sequences are. Homologs, orthologs, or variants of a polypeptide will possess a relatively high degree of sequence identity when aligned using standard methods. Methods of alignment of sequences for comparison are well-known in the art. As used herein, reference to "at least 80% identity" (or similar language) refers to "at least 80%, at least 85%, at least 90%, at least 91 %, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or even 100% identity" to a specified reference sequence. As used herein, reference to "at least 90% identity" (or similar language) refers to "at least 90%, at least 91 %, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or even 100% identity" to a specified reference sequence. Sequence identity may be measured by any acceptable algorithm.

[0073] The term "therapeutically effective amount" as used herein means a dosage which is sufficient to be effective for the treatment of the patient compared with no treatment. As used herein, the term "patient" or "subject" refers to members of the animal kingdom including but not limited to human beings, and "mammal" refers to all mammals, including, but not limited to human beings.

[0074] As is known in the art, a therapeutically effective dose of an active agent can vary from patient to patient based on many different factors, including, but not limited to, age, weight, gender, genotype, other medical conditions, etc. A doctor or medical provider overseeing the treatment of a patient is best suited to determine the therapeutically effective dose based on their knowledge and experience working with that patient. A therapeutically-effective amount may be an amount of a therapeutic agent effective to improve one or more symptoms of the disease, or normalize one or more markers of a disease in a patient. By normalize, it is meant to bring values of a marker in a patient towards or into a range considered as normal for a patient.

[0075] The term "treatment" or "treat" or "treating" with respect to a disease or medical condition as used herein means the management and care of a patient having developed a disease, condition, or disorder. The purpose of treatment is to combat the disease, condition, or disorder. Treatment includes, but is not limited to, the administration of a pharmaceutical composition to alleviate one or more symptoms associated with the disease, medical condition, or disorder. Treatment may result in the partial or full alleviation of all symptoms, or curing of said disease, medical condition, or disorder. A therapeutically effective amount may result in an improvement in a biomarker associated with a condition, for example TDP-43 dysfunction and conditions, diseases, and / or disorders associated with TDP-43 dysfunction, such as Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Degeneration (FTD), and / or Alzheimer’s Disease (AD). In non-limiting embodiments, an improvement in such a condition may be identified in terms of overall survival, improvement in ALS Functional Rating (ALSFRS-R), improved cognitive function, decreased CSF or plasma Neurofilament light chains (NfL) levels, and / or the like.

[0076] As used herein, the term "pharmaceutical composition" describes a composition that comprises one or more active agents and one or more pharmaceutically acceptable excipients. The excipient(s) must be acceptable in the sense of being compatible with the other ingredients of the composition and not deleterious to the recipient thereof. The pharmaceutical compositions can be for use in the treatment of any of the conditions described herein. The excipient(s) may be suitable for use in a parenteral formulation and administration of the active agent(s). The excipient(s) may be suitable for use in an intravenous formulation for administration of the active agent(s). The excipient(s) may be suitable for use in asubcutaneous, intramuscular, intratumoral, or other suitable formulation for administration of the active agent(s).

[0077] In one example, pharmaceutical compositions adapted for parental administration include aqueous and non-aqueous sterile injection solutions which may contain antioxidants, buffers, bacteriostats, and solutes which render the composition isotonic with the blood of the intended recipient; and aqueous and non-aqueous sterile suspensions which may include suspending agents and thickening agents. The compositions may be presented in unit-dose or multi-dose containers, for example sealed ampoules and vials, and may be stored in a freeze-dried (lyophilized) condition requiring only the addition of the sterile liquid carrier, for example water for injection or sterile lipid in oil solution, immediately prior to use. The pharmaceutical composition may comprise non-aqueous liquid excipients that may optionally be combined with water to form an emulsion.

[0078] In one example, a pharmaceutical composition may comprise a composition, such as a polynucleotide, as described herein and at least one pharmaceutically acceptable excipient.

[0079] A patient may receive from 0.1 ng to 100 mg of the composition per dose, including any increment therebetween. The patient may receive, per dose, 0.1 mg, 1 mg, 10 mg, 20 mg, 30 mg, 40 mg, 50 mg, 60 mg, 70 mg, 80 mg, 90 mg, or 100 mg of the miRNA, or mimic thereof. The patient may receive, per dose, approximately or about 0.1 mg, 1 mg, 10 mg, 20 mg, 30 mg, 40 mg, 50 mg, 60 mg, 70 mg, 80 mg, 90 mg, or 100 mg of the composition.

[0080] The patient may receive, per dose, from 0.001 to 1.00 mg / kg of the composition (based on patient body weight) including any increment therebetween. The patient may receive, per dose, 0.01 mg / kg, 0.1 mg / kg, 0.2 mg / kg, 0.3 mg / kg, 0.4 mg / kg, 0.5 mg / kg, 0.6 mg / kg, 0.7 mg / kg, 0.8 mg / kg, 0.9 mg / kg, or 1.0 mg / kg of the composition. The patient may receive, per dose, approximately or about 0.01 mg / kg, 0.1 mg / kg, 0.2 mg / kg, 0.3 mg / kg, 0.4 mg / kg, 0.5 mg / kg, 0.6 mg / kg, 0.7 mg / kg, 0.8 mg / kg, 0.9 mg / kg, or 1.0 mg / kg of the composition.

[0081] A pharmaceutical composition comprises a carrier or vehicle that includes at least one pharmaceutically acceptable excipient that is compatible for administration to a human patient. "Excipient" can refer to a pharmaceutically-acceptable material or composition, such as a liquid or solid filler, diluent, manufacturing aid (e.g., lubricant, talc magnesium, calcium or zinc stearate, or steric acid), or solvent encapsulatingmaterial, involved in carrying or transporting a therapeutic agent to a patient. Each excipient can be "acceptable" in the sense of being compatible with the other ingredients of the formulation and not injurious to the subject being treated. Some nonlimiting examples of materials which can serve as pharmaceutically-acceptable carriers or excipients include: (1 ) sugars, such as lactose, glucose and sucrose; (2) starches, such as corn starch and potato starch; (3) cellulose, and its derivatives, such as sodium carboxymethyl cellulose, ethyl cellulose and cellulose acetate; (4) powdered tragacanth; (5) malt; (6) gelatin; (7) lubricating agents, such as magnesium state, sodium lauryl sulfate and talc; (8) cocoa butter and suppository waxes; (9) oils, such as peanut oil, cottonseed oil, safflower oil, sesame oil, olive oil, corn oil and soybean oil; (10) glycols, such as propylene glycol; (11 ) polyols, such as glycerin, sorbitol, mannitol and polyethylene glycol; (12) esters, such as ethyl oleate and ethyl laurate; (13) agar; (14) buffering agents, such as magnesium hydroxide and aluminum hydroxide; (15) alginic acid; (16) pyrogen-free water; (17) isotonic saline; (18) Ringer's solution; (19) ethyl alcohol; (20) pH buffered solutions; (21 ) polyesters, polycarbonates and / or polyanhydrides; (22) bulking agents; (23) serum component, such as serum albumin, HDL and LDL; (24) rheology modifiers; and (25) other non-toxic compatible substances employed in pharmaceutical formulations.

[0082] A nucleic acid may be delivered using any effective carrier. In one example, the carrier is aqueous, such as water, saline, or PBS. The carrier may be a lipid-based vehicle (See, e.g., Baumann V, Winkler J. miRNA-based therapies: strategies and delivery platforms for oligonucleotide and non-oligonucleotide agents. Future Med Chem. 2014;6(17): 1967-84 and Bulcha JT, et al. Viral vector platforms within the gene therapy landscape. Signal Transduct Target Then 2021 Feb 8;6(1 ):53, and see below), for example, where at least one excipient may be a non-aqueous liquid. Examples of vehicles employing non-aqueous lipid excipients include, but are not limited to, the MaxSuppressor™ In Vivo RNA-LANCEr II (Lucerna-Chem AG, Luzern, Switzerland) that comprises a neutral lipid, a non-ionic detergent, an oil, and a proprietary mixture of small molecules. The lipid-based carrier may be mixed with water to form an emulsion. The specific excipients can be selected based on the mode of administration of the composition and compatibility with the one or more miRNAs, or mimic thereof, present therein. The composition may comprise a lipid-based carrier. The lipid-based carrier may be suitable for intravenous or subcutaneous administration to the patient. The lipid-based carrier may be a lipid nanoparticle (LNP) compositionand / or compositions, e.g., as described in U.S. Patent Nos. 10,844,028, 10,189,802, 9,872,911 , 9,556,110, 9,439,968, 9,227,917, 8,969,353, and 8,450,298, as well as in U.S. Patent Application Publication Nos. 2017 / 0204075, 2019 / 0177289, 2017 / 0152213, 2016 / 0114042, 2015 / 0203439, 2014 / 0322309, 2014 / 0161830, 2011 / 0293703, and 2010 / 0331234, each of which incorporated herein by reference for its technical disclosure relating to compounds and compositions useful in delivery of nucleic acid cargoes, and to the extent it is consistent with the present disclosure. Additional examples of LNPs are described in U.S. Patent Nos. 9,404,127, 9,364,435, and US 8,058,069, each of which incorporated herein by reference for its technical disclosure relating to compounds and compositions useful in delivery of nucleic acid cargoes, and to the extent it is consistent with the present disclosure (see, also, e.g., Sabnis S, et al., A Novel Amino Lipid Series for mRNA Delivery: Improved Endosomal Escape and Sustained Pharmacology and Safety in Non-human Primates. Mol Then 2018; 26(6): 1509-1519 and Yonezawa S, et al., Recent advances in siRNA delivery mediated by lipid-based nanoparticles. Adv Drug Deliv Rev. 2020; 154-155:64-78). Examples of lipid nanoparticles and methods of making lipid nanoparticles are described in Whitehead KA, et al., Degradable lipid nanoparticles with predictable in vivo siRNA delivery activity. Nat Commun. 2014 Jun 27; 5:4277. Virus-like particles also are useful in delivering nucleic acids and proteins, such as CRISPR / Cas proteins or gRNAs, and / or nucleic acid(s) comprising gene(s) for expression of a gRNA or a dCas transcriptional activator, as described herein (see, e.g., Hamilton JR, et al. Targeted delivery of CRISPR-Cas9 and transgenes enables complex immune cell engineering. Cell Rep. 2021 Jun 1 ;35(9): 109207 and Banskota S, et al. Engineered virus-like particles for efficient in vivo delivery of therapeutic proteins. Cell. 2022 Jan 20;185(2):250-265.e16).

[0083] Production of useful nucleic acid constructs, such as recombinant viral vectors for production of nucleic acids, such as the genetic constructs and recombinant viral genomes described herein, is routine, in that molecular cloning and gene assembly methods are routine. Further, a number of companies can custom- synthesize and verify multi-kilobase genes, making the production of genes or genomes as described herein, such as rAAV or scAAV genomes, routine (See, e.g., Gene Synthesis Handbook, 2d Edition, 2014, GenScript USA, Inc.).

[0084] A vector is a nucleic acid molecule allowing insertion of foreign nucleic acid without disrupting the ability of the vector to replicate and / or integrate in a host cell. Avector can include nucleic acid sequences that permit it to replicate in a host cell, such as an origin of replication. An insertional vector is capable of inserting itself into a host nucleic acid. A vector can also include one or more selectable marker genes and other genetic elements. An expression vector is a vector that contains the necessary regulatory sequences to allow transcription and translation of inserted gene or genes.

[0085] A variety of viral vectors have found use in the delivery of genes for expression of a protein in a cell, such as a cancer cell as described herein. See, e.g., Butt MH, Zaman M, Ahmad A, Khan R, Mallhi TH, Hasan MM, Khan YH, Hafeez S, Massoud EES, Rahman MH, Cavalu S. Appraisal for the Potential of Viral and Nonviral Vectors in Gene Therapy: A Review. Genes (Basel). 2022 Jul 30; 13(8): 1370. The following paragraphs describe use of Adeno-associated virus particles for delivery of genes. AAV, described in further detail below, is but one of many robust and well- characterized viral vectors suited for gene therapy, which also includes, without limitation, gammaretroviruses, lentiviruses, adenovirus, and herpes simplex virus. While AAV may be preferred in instances, other safe and effective viral transducing particles can be developed based on the genes described herein for use in the devices, systems and methods described herein.

[0086] AAV (adeno-associated virus), is a virus belonging to the genus Dependoparvovirus, and family Parvoviridae. The virus is a small replicationdefective, non-enveloped virus. AAV is not currently known to cause any disease by itself. AAV requires a helper virus, such as adenovirus or herpes simplex virus, to facilitate productive infection and replication. In the absence of helper virus, AAVs establish a latent infection within the cell, either by site-specific integration into the host genome or by persisting in episomal forms. Gene therapy vectors using AAV can infect both dividing and quiescent cells. Furthermore, AAV serotypes have different tropism and can infect cells of multiple diverse tissue types. While eleven serotypes of AAV have been identified to date, AAV2 was among the first to be identified and has been consistently used for the generation of recombinant AAV vectors. Further certain natural or modified AAVs transduce specific organs or cell populations.

[0087] The AAV virion shell is approximately 25 nanometers (nm) in diameter and encapsulates a single-stranded DNA genome that consists of two large open reading frames (ORFs) flanked by inverted terminal repeats (ITR). The ITRs are the only cisacting elements required for genome replication and packaging. In wild-type AAV, the left ORF encodes four replication proteins responsible for site-specific integration,nicking, and helicase activity, as well as regulation of promoters within the AAV genome. AAV possesses a 4.7 kb genome, and as such, efficient packaging of recombinant AAV (rAAV) vectors can be performed with constructs ranging from 4.1 kb to 4.9 kb in size.

[0088] Helper-free production of the rAAV requires transfection of the following components into host cells, typically 293 cells (HEK293 cells), which are broadly available, or similar cell lines: (1 ) an rAAV vector containing the transgene expression cassette flanked by the two ITRs; (2) expression of Rep and Cap proteins, typically provided by a helper plasmid in trans; and (3) adenovirus genes encoding E1 , E2A, E4, and virus-associated RNA, also provided, at least in part by another helper plasmid in trans (293 cells produce the Ad E1 gene in trans). Rep and Cap proteins, which are necessary for viral packaging, are replication proteins and capsid proteins, respectively. Rep proteins consist of rep 78, 68, 52, and 40. They specifically are involved with the replication of AAV. Cap proteins are comprised of three proteins, VP1 , VP2 and VP3, with molecular weight of 87, 72 and 62 kDa, respectively. These capsid proteins assemble into a near-spherical protein shell of 60 subunits. Helper- free AAV packaging systems are broadly available commercially, and see, e.g., U.S. Patent Nos. 6,093,570, 6,458,587, 6,951 ,758, and 7,439,065. In scAAV (self- complementary AAV), the right ITR contains a deletion of D-sequence (the packaging signal) and a terminal resolution site mutation (?trs), which prevent Rep-mediated nicking and force packaging of dimer or self-complementary genomes. Making dsAAV from scAAV vector renders much improved transduction both in vitro and in vivo.

[0089] Preparation of rAAV transducing particles, such as scAAV transducing particles is routine. Since the transfection method is often considered unsuitable for large-scale production, the infection of cell lines stably expressing Rep and Cap with adenovirus carrying a vector genome has afforded the ability to scale-up. Another option includes infection of proviral cell lines with adenovirus or herpes simplex virus vector carrying an AAV Rep and Cap expression cassette. These methods still require the complete elimination of adenovirus (or herpesvirus) during the production process. However, in baculovirus expression vector systems for rAAV vector production in insect SF9 cells, the components of AAV production, including Rep and Cap proteins, as well as vector genomes are provided by separate recombinant baculoviruses. Numerous robust rAAV production methods are available (see, e.g., Merten OW. Development of Stable Packaging and Producer Cell Lines for the Production of AAVVectors. Microorganisms. 2024 Feb 13;12(2):384; Gray S J, et al. (2011 ) Production of recombinant adeno-associated viral vectors and use in in vitro and in vivo administration. Curr Protoc Neurosci. doi:10.1002 / 0471142301 ,ns0417s57; and Guo P, et al. (2012) Rapid and simplified purification of recombinant adeno-associated virus. J Virol Methods 183(2): 139-146).

[0090] Once the virus has been produced in the, e.g., 293 cells, the cells are collected, lysed, and the resultant virus is purified. Density gradient ultracentrifugation, e.g., in cesium chloride or nonionic iodixanol (VISIPAQTM) gradients and column chromatography, such as ion-exchange, heparin-affinity, or mucin-affinity column chromatography, depending on the AAV serotype. Once the rAAV has been purified and concentrated to a suitable concentration, the virus can be used for in vitro cell transduction or for in vivo animal injection at an appropriate MOI (Multiplicity of Infection).

[0091] Numerous rAAV vectors have been made containing genes for expressing proteins, and are commercially available. Due to size limitations, genes for use in rAAV vectors typically do not include introns. rAAV vectors also include the 5' ITR and 3' ITR flanking the gene, which is referred to as a transgene. Thus, a typical rAAV genome has the following structure, in order from 5' to 3' on the sense strand: ITR - promoter - transgene ORF - pA - ITR. Methods of molecular cloning of rAAV transgene constructs, preparation of rAAV particles, and storage and use thereof are broadly- known and further technical details are unnecessary for one of ordinary skill in the art to be able to construct useful rAAV vectors, and produce and use rAAV particles as described herein. As indicated above, so long as the gene sequence is less than the packaging limit of rAAV or scAAV, it is useful for production of a transduction particle as described herein. As stated above, AAV is merely an example of the many methods and compositions useful for delivering nucleic acids to a patient.

[0092] Provided herein is a composition, useful as a biosensor and / or a therapeutic delivery system, based on RNA technology. As a biosensor, the composition allows for variable expression of a marker, such as fluorescent protein, to aid in identifying and / or quantifying the extent of a disorder or condition (for example as shown in FIG. 1 A). As a delivery system, the composition allows for targeted delivery, such that the therapeutic is only delivered to cells possessing a target disorder or condition, for example a dysfunction in a gene and / or protein, for example transactive response DNA-binding protein 43 (TDP-43).

[0093] In a non-limiting embodiment, a composition is disclosed. In non-limiting embodiments, the composition may be a polynucleotide including a plurality of nucleic acids. In non-limiting embodiments, the composition includes a first nucleic acid, a second nucleic acid, a third nucleic acid, and a fourth nucleic acid.

[0094] In non-limiting embodiments, the first nucleic acid is a control sequence. By control sequence it is meant that the first nucleic acid does not encode a functioning protein, and, instead, is merely a spacer. In non-limiting embodiments, the first nucleic acid encodes a marker, for example a first fluorescent protein. Suitable markers, such as fluorescent proteins, are known to those of skill in the art, and may include, without limitation, mCherry. Exemplary sequences for a first nucleic acid are provided in Table 1 , below.

[0095] In non-limiting embodiments, the composition includes a second nucleic acid. In non-limiting embodiments, the second nucleic acid is an excisable nucleic acid. As will be appreciated by those of skill in the art, an excisable nucleic acid refers to a nucleic acid molecule that can be removed or excised from its original location. In non-limiting embodiments, the second nucleic acid is excisable, such that in the presence of a compound, such as a protein, for example TDP-43, that binds to the second nucleic acid, the second nucleic acid is excised from the polynucleotide and one or more downstream nucleic acids remain in frame. Accordingly, in non-limiting embodiments, the second nucleic acid includes one or more binding sites for the compound (e.g., TDP-43). In non-limiting embodiments, the one or more binding sites comprise UG-rich regions. In non-limiting embodiments, one or more of the one or more binding cites must include a UG-rich region. By “UG rich region” it is meant that the region has a higher percentage of uracil and guanine than the remaining portion of the composition. While TDP-43 is exemplified herein as a binding partner for the second nucleic acid, those of skill in the art will appreciate that the system described herein is not so limited, and that other nucleic acids, which may be binding targets for other proteins, may be utilized. For example, in non-limiting embodiments, a fused in sarcoma (FUS) protein may be a binding partner for the second nucleic acid (in nonlimiting embodiments, the second nucleic acid may include one or more GGUG, CGCGC, and / or GUGGU regions, to which FUS may bind).

[0096] Conversely, if no binding compound is present, the second nucleic acid is not excised from the polynucleotide, and one or more downstream nucleic acids are frame shifted and thus are out of frame. In non-limiting embodiments, the secondnucleic acid is a cryptic exon. As used herein, the term “cryptic exon” means an intronic sequence (a non-coding region of DNA) that is mistakenly included as an exon in the mature RNA transcript instead of being removed during splicing. In non-limiting embodiments, the cryptic exon may be from or may be derived from any exon that binds TDP-43. In non-limiting embodiments, the cryptic exon is from or is derived from a cystic fibrosis transmembrane conductance regulator (CFTR) gene. In non-limiting embodiments, the second nucleic acid includes an intron from an llnc-13 homolog A (UNC13A) gene. In non-limiting embodiments, the second nucleic acid includes an exon (e.g., exon 9) from CFTR and an intron (e.g., intron 20) from UNC13A. Exemplary sequences for a second nucleic acid are provided in Table 1 , below. In non-limiting embodiments, the cryptic exon includes a UG-rich region, in non-limiting embodiments a UG-rich region is arranged on a 5’ end and / or a 3’ of the cryptic exon. In non-limiting embodiments, a UG-rich region is arranged on a 5’ and / or a 3’ end of the intron from UNC13A. In non-limiting embodiments, the cryptic exon is from or is derived from a stathmin-2 (STMN2) gene. In non-limiting embodiments, the cryptic exon is from or is derived from an acyl-CoA binding domain-3 (ACBD3) gene. In nonlimiting embodiments, the cryptic exon is from or is derived from a hepatoma-derived growth factor-related protein 2 (HDGFL2) gene.

[0097] In non-limiting embodiments, the composition includes a third nucleic acid. In non-limiting embodiments, the third nucleic acid encodes a marker, for example a second fluorescent protein. In non-limiting embodiments, the second fluorescent protein is the same as or is different than the first fluorescent protein. In non-limiting embodiments, the second fluorescent protein is enhanced green fluorescent protein (EGFP), though those of skill will appreciate that other markers, such as fluorescent proteins, may be used.

[0098] In non-limiting embodiments, the third nucleic acid is and / or encodes a therapeutic composition. In non-limiting embodiments, the third nucleic acid may be a codon-optimized nucleic acid encoding a specific protein, for example a protein that is dysfunctional and / or is present in only low levels or not at all in a cell to which the composition is introduced, thereby supplementing, and providing, a normal level of the protein, for example TDP-43. As used herein, “dysfunctional” means a protein that is altered from wild type (e.g., includes one or more deletions, insertions, substitutions and / or mutations), is present in lower levels than in a wild-type comparison, and / or exhibits lower activity than in a wild-type comparison. The terms “dysfunctional” and“aberrantly expressed” in reference to a protein of interest are used interchangeably herein. In non-limiting embodiments, the third nucleic acid encodes a different protein (e.g., a protein that is different than a protein that is dysfunctional and / or present at only low levels in a cell), such that the different protein is overexpressed in the cell. In non-limiting embodiments the different protein may be one or more of a heterogenous nuclear ribonucleoprotein L protein (HNRNPL), a protein capable of rescuing splicing events caused by TDP-43 dysfunction / loss of function, an optimized / mutant / transacted / reconstructed version of TDP-43 that is functional but less aggregation-prone, a proteolysis targeting chimeric (PROTAC) molecule, an intrabody, a RING-bait that targets TDP-43 aggregation for clearance (described below), a Cas9 system (e.g., a guide RNA) to perform gene editing (e.g., KO, KD, activation, base edition, and the like).

[0099] In non-limiting embodiments, the different protein may modulate levels and / or activity of the dysfunctional protein (e.g., by modulating expression of a gene encoding the dysfunctional protein), thereby increasing levels and / or activity of the dysfunctional protein. In non-limiting embodiments, the therapeutic composition may be a nucleic acid, optionally with a promoter component, or protein encoded thereby, for knocking down and / or silencing a gene, such as a non-coding RNA and / or a shRNA. Exemplary sequences for a third nucleic acid are provided in Table 1 , below. In non-limiting embodiments, any combination of the foregoing may be included as a third nucleic acid, in non-limiting embodiments with a separator (such as a 2A peptide) therebetween.

[0100] In non-limiting embodiments, the third nucleic acid is or encodes a TDP-43 Knockdown-Replacement (TKR) gene. In non-limiting embodiments, the composition includes a gene (e.g., as the third nucleic acid) that express a wild-type, codon- optimized TDP-43. In non-limiting embodiments, the composition includes, either separately or together with the gene encoding the wild-type, codon-optimized TDP- 43, a vehicle for knocking down endogenous TDP-43. The term “knocking down” is used interchangeably with "reducing," "silencing," "downregulating," "suppressing", “inhibiting”, and other similar terms, and includes any level of inhibition. Knockdown technologies (e.g., RNAi and the like) are broadly known (including in at least United States Patent No. 7,737,265 and International Patent Publication No. WO 2016 / 209862, the contents of which are incorporated herein by reference in their entirety) and those of skill in the art will appreciate that any suitable nucleic acid ortherapeutic may be used to knock down TDP-43. Further non-limiting examples include shRNAs targeting heterogenous nuclear ribonucleoprotein C protein (HNRNPC), for example through the TRE3G system, knockdown genes that could rescue splicing events caused by TDP-43 loss of function, shRNAs targeting endogenous / mutant TDP-43, shRNAs targeting C9ORF72 (GGGGCC) repeat expansions, and the like. In non-limiting embodiments, because of the sensitivity of the composition and the (in non-limiting embodiments) ability for the composition to only express the third nucleic acid in the presence of dysfunctional TDP-43, the composition may act as an accurately-controlled TDP-43 replacement that never overexpress beyond endogenous TDP-43.

[0101] In non-limiting embodiments, the third nucleic acid is or encodes HNRNPC, HNRNPL, MATR3, RBM14, YTHDF3, WAC, TIAL1 , HNRNPUL1 , G3BP2, HNRNPA1 , RC3H1 , HNRNPM, and / or HNRNPAO. Those of ordinary skill in the art will appreciate that any combination of any of the foregoing markers or therapeutic compositions (or a nucleic acid encoding the foregoing) may be included in a single or multiple compositions that may be delivered to a patient’s cell(s).

[0102] In non-limiting embodiments, the composition may be included in a viral vector as described herein, for example, a TKR cassette may be packaged into an AAV vector under a strong neuron-specific promoter (e.g. hSynl ) and delivered to human neurons (e.g., human iPSC-derived motor neurons) that exhibit TDP-43 pathology.

[0103] In non-limiting embodiment, the composition may include an effector that is capable of counteracting TDP-43 gain-of-function toxicity. For example, a composition as described herein may include a TDP-43 proteolysis-targeting chimera (PROTAC)285 or other protein-degradation triggered to be activated specifically when TDP-43 aggregates are detected. In non-limiting embodiments, the composition may include (e.g., as a third nucleic acid) a sequence that encodes a TDP-43 aggregate sensor. In non-limiting embodiments, such a sensor would encode a TDP-43 fragment or a conformation-sensitive peptide fused to a ligase (e.g., an E3 ligase recruitment domain). In such embodiments, when TDP-43 begins aggregating in the cytoplasm, the chimeric protein expressed by the composition described herein would coaggregate and induce ubiquitination of the entire aggregate, marking it for proteasomal or autophagic clearance. In non-limiting embodiments, such an approach, delivered via a viral vector (such as an AAV), clears existing TDP-43 aggregates and preventsformation of new aggregates. As with other embodiments, the arrangement of the third nucleic acid relative to the other components of the composition allows for such an aggregate clearance mechanism to only be activated in cells where TDP-43 LOF / aggregation is occurring (thus limiting potential damage in healthy cells).

[0104] In non-limiting embodiments, a composition as described herein may target different proteins, e.g., FUS and / or a poly-GR dipeptide repeat protein (DPR) by substituting the second nucleic acid (e.g., the binding site(s)) and including a region that binds FUS and / or poly-GR DPR.

[0105] In non-limiting embodiments, the composition may include (e.g., as a third nucleic acid), one or more antisense oligonucleotides (ASOs) to suppress the production of toxic / dysfunctional TDP-43 or to block cryptic exons in transcripts. In non-limiting embodiment a composition as described herein could include one or more ASOs to reduce acute toxicity from aggregates, while the same (or a second) composition could include a nucleic acid that restores long-term neuronal functions (e.g., wild-type and / or normal functioning TDP-43).

[0106] In non-limiting embodiments, a composition as described herein may include (for example, as a third nucleic acid) a nucleic acid encoding an anti-TDP-43 nanobody (e.g., when dysfunctional TDP-43 is detected through the second nucleic acid (e.g, the binding site).

[0107] In non-limiting embodiments, for example as described above, in the presence of a binding compound, the second nucleic acid is excised from the composition, resulting in a downstream stop codon being in frame, and preventing expression of the third nucleic acid. In non-limiting embodiments, for example as also described above, the second nucleic acid is not excised from the composition (for example, because the binding compound is not present and / or is dysfunctional), resulting in a downstream stop codon being in out of frame, and allowing expression of the third nucleic acid. In non-limiting embodiments, the one or more fourth nucleic acids encode a cleavable peptide, such as a self-cleavable peptide. In non-limiting embodiments, the one or more fourth nucleic acids encode a 2A peptide, such as a E2A, F2A peptide, P2A peptide, and / or a T2A peptide. In non-limiting embodiments, the fourth nucleic may be an inducible promoter. In non-limiting embodiments, the fourth nucleic acid may be or may be a part of a transcriptional activator system (such as, for example and without limitation a tet-on activator system, such as tet-on 3G (TREG3)). Tet-on systems and inducible promoters (including those sensitive to light,temperature, and / or chemical stimuli) are broadly known to those of skill in the art. Exemplary sequences for a fourth nucleic acid are provided in Table 1 , below.

[0108] In non-limiting embodiments, one or more of the first, second, third, and / or fourth nucleic acid is a ribonucleic acid. As described above, the nucleic acids may be deoxyribonucleic acids prior to delivery, and, once delivered to a cell, may be transcribed to ribonucleic acids, at which point splicing as described herein may occur.

[0109] In non-limiting embodiments, one or more additional nucleic acids may be included in the composition. For example, in non-limiting embodiments a localization sequence may be included. In non-limiting embodiments, the localization sequence is a nuclear localization sequence (NLS), and in non-limiting embodiments is a 3X NLS. In non-limiting embodiments, the NLS is or is derived from a simian virus 40 (SV40), though those of skill in the art will appreciate that NLSs are broadly known and others, including from human histone proteins, may be used.

[0110] In non-limiting embodiments the composition is a polynucleotide that includes, from a 5’ to a 3’ direction, SEQ ID NO: 1 (mCherry), SEQ ID NO: 10 (T2A), SEQ ID NO: 11 (CFTR exon 8), SEQ ID NO: 14 (CFTR exon 9), SEQ ID NO: 15 (UNO CE), SEQ ID NO: 16 (UNO intron 20), and SEQ ID NO: 7 (T2A). In non-limiting embodiments, the polynucleotide further includes, at the 3’ end, SEQ ID NO: 9 (3X NLS). In non-limiting embodiments, the composition is a polynucleotide that includes, from a 5’ to a 3’ direction, SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, and SEQ ID NO: 7. In non-limiting embodiments, the composition is a polynucleotide that includes, from a 5’ to a 3’ direction, SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 16, and SEQ ID NO: 7. While the sequences disclosed herein and in the attached sequence listing are DNA sequences, those of skill in the art will appreciate that such nucleic acids may, once delivered to a cell, be transcribed to RNA, and thus the specific sequences disclosed herein, while including thymine residues, should be understood to also encompass RNA sequences, including with uracil residues as appropriate and / or transcriptions of the sequences.

[0111] In non-limiting embodiments, the polynucleotide may include an additional nucleic acid encoding a reporter peptide or a therapeutic peptide arranged at a 3’ end of SEQ ID NO: 7. In non-limiting embodiments, the nucleic acid encoding the reporter peptide has the sequence of SEQ ID NO: 8.

[0112] In non-limiting embodiments, the composition is a polynucleotide including, from a 5’ end to a 3’ end: the first nucleic acid; the fourth nucleic acid; the excisable second nucleic acid; a stop codon that is out of frame when the excisable second nucleic acid is present in the composition and is in frame when the excisable second nucleic acid is excised from the composition; the fourth nucleic acid; and the third nucleic acid, the third nucleic acid is or encodes a therapeutic substance, or encodes a reporter, such as a fluorescent protein. In non-limiting embodiments, when introduced into a cell, the cell translates or expresses the third nucleic acid, thereby producing a reporter or therapeutic composition when the excisable second nucleic acid is excised.

[0113] In non-limiting embodiments, the composition is a recombinant transducing particle or a viral expression vector that includes the first, second, third, and fourth nucleic acids. In non-limiting embodiments, the viral expression vector is a lentiviral, an adenoviral, an adeno-associated virus (AAV), or a herpesvirus recombinant genome, or a plasmid DNA vector. In non-limiting embodiments, the composition is provided in a pharmaceutically-acceptable excipient and / or carrier (for example those disclosed herein), for example in a kit.

[0114] Also provided herein are methods of using a composition as described herein. In non-limiting embodiments, a composition as described herein may be used to detect aberrant expression or dysfunction of a protein in a cell. The cell may be obtained from a patient who has been diagnosed, who is suspected of having, or is at risk of developing a condition associated with aberrant expression and / or dysfunction of the protein. Such a method may include delivering a composition as described herein to the cell. As described herein, the third nucleic acid may be translated or expressed when the protein is aberrantly expressed or is dysfunctional (e.g., when the protein does not bind to and / or excise the second nucleic acid, and thus the stop codon is out of frame), and the third nucleic acid is not expressed when the protein is not aberrantly expressed or is not dysfunctional (e.g., when the protein binds to and / or excises the second nucleic acid, and thus the stop codon is in frame). Those of skill will appreciate that, due to the selectivity of the composition (e.g., the third nucleic acid is only expressed in the presence of a dysfunctional or aberrant protein of interest), the composition need not be selectively delivered to a patient, or in a targeted manner, such that any suitable delivery system for a polynucleotide may be useful here.

[0115] As described above, in non-limiting embodiments in which the third nucleic acid is expressed, for example, because the protein of interest (e.g., TDP-43) is dysfunctional and / or is aberrantly expressed, translation and / or expression of the third nucleic acid correlates to the level of aberrance of expression of dysfunction of the protein. That is, in non-limiting embodiments, the lower the level of the protein and / or the greater the dysfunction (e.g., the greater the inability of the protein to bind to and / or excise the second nucleic acid), the greater the expression of the third nucleic acid (e.g., the greater the expression of a marker and / or fluoresce protein, such as EGFP).

[0116] Also provided herein are methods of using a composition as described herein to treat a condition, disorder, and / or dysfunction associated with dysfunction and / or aberrant expression of a protein in a cell. In such embodiments, the third nucleic acid may be and / or encode, in addition to or instead of a reporter (such as a fluorescent protein), a therapeutic composition. As described herein, the third nucleic acid, in non-limiting embodiments where the third nucleic acid is and / or encodes a therapeutic composition, may be translated or expressed when the protein is aberrantly expressed or is dysfunctional (e.g., when the protein does not bind to and / or excise the second nucleic acid, and thus the stop codon is out of frame), and the third nucleic acid is not expressed when the protein is not aberrantly expressed or is not dysfunctional (e.g., when the protein binds to and / or excises the second nucleic acid, and thus the stop codon is in frame). As noted above, those of skill will appreciate that, due to the selectivity of the composition (e.g., the third nucleic acid is only expressed in the presence of a dysfunctional or aberrant protein of interest), the composition need not be selectively delivered to a patient, or in a targeted manner, such that any suitable delivery system for a polynucleotide may be useful here.

[0117] In the methods disclosed herein, the patient may have or may be suspected of having and / or may be at risk of developing a disease, such as a neurodegenerative disease. In non-limiting embodiments, the disease may be Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Degeneration (FTD), limbic-predominant age-related TDP-43 encephalopathy neuropathological change (LATE-NC), inclusion body myositis, chronic traumatic encephalopathy (CTE), and / or Alzheimer’s Disease (AD).Example 1Methods:

[0118] Generation of plasm ids

[0119] The CUTS sequence, plasmid, and map were originally generated in this study. The CFTR-TS, UNC13A-TS and CUTS DNA sequences were designed in silico and de novo synthesized by Genewiz. These sequences were and assembled into Tet3G vector between EcoRI and Notl with NEBuilder HiFi DNA Assembly Master Mix (NEB, E2621 L) following the manufacturer's protocol. The full DNA sequence for CUTS, CFTR-TS, and UNC13A-TS can be found in Table 1.

[0120] The CFTR minigene assay plasmid (pTB-CFTR-A455E) was a kind gift from Dr. Yuna Ayala.

[0121] The exogenous TDP-43 plasmids were constructed in a pCMV backbone by linking a 3xFlag-APEX2 protein (Addgene #164622) to TDP-43 coding sequences with WT, cyto, 5FL, or cyto 5FL modifications.

[0122] The codon-optimized TARDBP coding sequence (Table 1 ) was synthesized by IDT and assembled downstream of the GFP sequence of CUTS with NEBuilder HiFi DNA Assembly Master Mix. Those of skill will appreciate that variations from the below sequences are possible and are within the scope of the present disclosure. For example, and without limitation, sequences having 70% or more, 75% or more, 80% or more, 85% or more, 90% or more, 95% or more, and / or 99% or more, all values and subranges therebetween inclusive, sequence identity to any of the below sequences are considered to be within the scope of the present disclosure.Table 1

[0123] All the primers were synthesized by IDT. All plasmids were verified using whole-plasmid sequencing via Oxford Nanopore, provided by Plasmidsaurus.

[0124] Cell culture and transfection

[0125] Human Embryonic Kidney 293 (HEK293) cells (female genotype, acquired from the American Type Culture Collection (ATCC)) and Hela TDP-43 knock-out (KO) cells (a kind gift from Dr. Shawn M Ferguson) (Roczniak-Ferguson and Ferguson 2019) were cultivated in Dulbecco's Modified Eagle Medium high glucose, pyruvate(DMEM, Thermo Fisher Scientific, 10-313-039) supplemented with 10% HyClone Bovine Growth Serum (Cytiva HyClon, SH3054103HI) and 1X GlutaMAX (Thermo Fisher Scientific, 10-313-039). Cells were incubated at 37°C in a 5% CO2 atmosphere with high humidity. For transfection assays, cells were plated on collagen-coated coverslips or dishes (50 pg / mL, GIBCO) and transfected with designated DNA quantities using Lipofectamine 3000 (Thermo Scientific, L3000015) following the provider's protocol.

[0126] HEK293 stable cell line generation via Piggybac transposition

[0127] For stable cell line creation, HEK293 cells were pre-plated on 6-well plates and transfected at approximately 70% confluence with 2.5 pg of Piggybac plasmids encoding CUTS, CFTR-TS, UNC13A-TS, and CUTS-TDP43 alongside 0.5 pg of the Super PiggyBac Transposase Expressing plasmid (PB200PA-1 ) using Lipofectamine 3000, according to the manufacturer's guidelines. A transfection control without transposase was included. Following a 48-hour post-transfection period, cells were selected with puromycin (Sigma, P8833) at 5 pg / mL, with media changes every two days. Selection resulted in control cell death within approximately 5 days, while surviving populations were expanded and maintained in reduced puromycin concentrations (2.5 pg / mL) to establish stable lines. Expression of the transgenes was confirmed by immunofluorescence staining and Western blot analysis.

[0128] SDS-PAGE and Western blot

[0129] For protein analysis, cells were lysed directly on the plate using fresh and pre-chilled Urea-RIPA buffer: 2M fresh urea in 1XRIPA buffer (Boston Bioproducts, BP-115X), supplemented with 1 % protease inhibitor cocktail (Sigma, P8340) and sonicated. Protein concentrations were quantified using the Pierce BCA Protein Assay Kit (Thermo Scientific, 23227). Proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes for WB analysis. Membranes were blocked and probed with primary antibodies: mouse-anti-GFP (Santa Cruz, sc-9996, 1 :200), mouse-anti- a-tubulin (Sigma, T5168, 1 :1000), rabbit-anti-TDP-43 (Proteintech, 10782-2-AP, 1 :2500), and rabbit-anti-m Cherry (Cell Signaling, 43590, 1 :1000), followed by HRP- conjugated secondary antibodies: donkey-anti-mouse (JacksonlmmunoResearch 715035151 , 1 :5000) or donkey-anti-rabbit (JacksonlmmunoResearch 711035152, 1 :5000). Detection was achieved using Western Lightning ECL Pro (Revvity, NEL1201001 EA) or Supersignal West Femto Maximum Sensitivity ChemiluminescentSubstrate (Thermo Scientific, 34095) in an Amersham ImageQuant 800 GxP biomolecular imager system (Amersham, 29653452).

[0130] Confocal microscopy

[0131] Confocal imaging was performed on a Nikon A1 laser-scanning microscope using either a 60X oil immersion or a 10X / 20X objective for live-cell observations. A Tokai HIT stage-top incubator maintained the required environmental conditions. Nikon Elements software facilitated image acquisition and analysis. Representative images were chosen from at least two independent experiments with a minimum of three biological replicates each.

[0132] siRNA reverse transfection

[0133] siRNA reverse transfections were conducted using Lipofectamine RNAiMAX reagent (Thermo Scientific, 13778150), adhering to the supplier's protocol. To knockdown TDP-43, the following siRNAs were used: ON-TARGETplus SMART pool siRNA against TARDBP (Dharmaco, L-012394-00-0005) and siGENOME non-Targeting siRNA for control (Dharmaco, D-001206-13-05).

[0134] RNA extraction, RT-PCR, and qPCR

[0135] RNA extraction was performed using the RNeasy Mini Kit (Qiagen, 74106) with concentration determinations via Nanodrop Spectrophotometer (Nanodrop, ND- 1000). Reverse transcription of extracted RNA (0.2 pg) to cDNA utilized iScript Reverse Transcription Supermix (Bio-Rad, 1708841 ) in accordance with the manufacturer's guidelines.

[0136] The RT-PCR or RT-qPCR was conducted with cDNA diluted 10-fold. For RT-PCR assay, CFTR cryptic exon region was amplified with the following primer pair: P690-F (5’-CAACTTCAAGCTCCTAAGCCACTGCCTGC) (SEQ ID NO: 19) and P691 -R (5’-TAGGATCCGGTCACCAGGAAGTTGGTTAAATCA) (SEQ ID NO: 20). CUTS' cryptic exon region was amplified with the following primer pair: CUTS-CE-F (5’-ATCCCGGCCCTGGATCCG) (SEQ ID NO: 21 ) and CUTS-CE-R (5’- GTCAGCTTGCCGTAGGTGGC) (SEQ ID NO: 22). PCR products were separated by agarose gel electrophoresis, and the bands were visualized with Amersham ImageQuant 800 GxP biomolecular imager system.

[0137] For RT-qPCR assay, SsoAdvanced™ Universal SYBR Green Supermix (Biorad, 1725272) was used following the supplier's protocol on a CFX96 Touch Real- Time PCR Detection System (Biorad). Three technical replicates were included for each sample with the following program: 95°C for 30 s, 40 cycles of 95°C for 15 s and60°C for 20 s. CUTS-CE-F (5’-ATCCCGGCCCTGGATCCG) (SEQ ID NO: 23) and CUTS-CE-R (5’-GTCAGCTTGCCGTAGGTGGC) (SEQ ID NO: 24) were used to quantify normal CUTS transcript. CUTS-J-F (5’-TCCGGCGAGGGATTTGGG) (SEQ ID NO: 25) and CUTS-J-R (5’-CCCCACCTAGACCCATCTCTCC) (SEQ ID NO: 26) were primers targeting the cryptic exon junctions to quantify cryptic exon-specific CUTS transcript. Relative quantification of CUTS cryptic exon was determined by the ACt value of CUTS-J normalized to CUTS-CE.

[0138] Statistical analysis

[0139] Statistical significance was evaluated using GraphPad Prism 9 software, and specific tests used for each experiment are outlined in the respective figure legends.Results

[0140] CUTS TDP-43 LOF sensor design utilizing known Cryptic Exons

[0141] To improve the detection of TDP-43 LOF, we designed a novel TDP-43 LOF sensor (TS) for real-time screening-based detection using previously reported genes known to undergo TDP-43 regulated splicing (UNC13A and CFTR). The TS cassettes are constructed with a constitutively expressed mCherry, followed by a TDP- 43 regulated CE and a GFP linked to a 3x nuclear localization signal (NLS), each separated by a T2A self-cleavage sequence (Figure 1A). We positioned the GFP reporter outside the mCherry open reading frame (ORF), introducing an early stop codon upstream to the GFP (Figure 1A). This strategic design achieved three key outcomes: (1 ) Under physiological TDP-43 levels, the binding of TDP-43 to the CE and UG-rich sequence should promote complete intronic splicing, maintaining the inframe stop codon and allowing only mCherry expression. (2) TDP-43 loss and / or failure to bind the CE sites, will result in CE retention and a subsequent frameshift, causing resulting in an out-of-frame stop-codon GFP codon inclusion. (3) The GFP output of the TS should be proportional to the level of TDP-43's LOF (Figure 1 A).

[0142] To assess the functionality of TS, we generated CFTR-TS and UNC1 SATS, which utilized the TDP-43 regulated CEs from CFTR and UNC13A, respectively (Figure 1 B). Additionally, we engineered a combined construct termed CUTS (CFTR- UNC13A TS), integrating both CE sequences (Figure 1B). Additional base modifications were incorporated into the cassette designs to prevent unexpected stop codons within the CE regions (see Table S1 for full sequence details). Each construct was coupled with a Tet3g promotor, cloned into a Piggybac vector, and stablyexpressed in HEK293 cells. To evaluate the accuracy and reliability of the three TS constructs, we performed a TDP-43 LOF assay using increasing siRNA concentrations, followed by live confocal imaging and western blot (WB) analysis (Figures 1A-1J and 2). Across all cell lines, we observed constitutive mCherry expression and an increased trend in GFP signals with higher siTDP43 concentrations. The CFTR-TS construct exhibited a notably high baseline, with detectable GFP leakage in control groups without TDP-43 loss (Figures 1C-1 D, 1F, 11, 2). In contrast, the UNC13A-TS construct demonstrated superior accuracy compared to CFTR-TS, showing no detectable GFP expression under control conditions in both imaging-based and WB analyses (Figures 1C-1 E, 1 H, 2). However, the UNC13A-TS showed limited sensitivity, with only a modest GFP signal (17% of CUTS) under high quantities of siTDP43 treatment. Interestingly, cells expressing the CUTS construct exhibited a synergistic effect from both the CFTR and UNC13A CE sequences, achieving high sensitivity evidenced by a clear dose-responsive GFP expression with increasing siTDP43 concentrations, while maintaining high accuracy with minimal leakage via imaging and GFP immunoblotting (Figures 1C-1 D, 1G, 1 J, 2). Given the promising accuracy of the CUTS sensor, we proceeded to further characterize the CUTS RNA biosensor.

[0143] To challenge the stability and sensitivity of CUTS, we next conducted an ultra-low dose TDP-43 siRNA transfection, ranging from 38 - 1200 pM. Immunofluorescence staining (IF) revealed a consistent increase in both GFP intensity and GFP-positive cell ratios in CUTS-expressing cells in response to elevated siTDP43, with minimal baseline expression observed (Figures 3A-3B). We confirmed the ultra-sensitivity and accuracy of CUTS using WB analysis, demonstrating measurable GFP even at the lowest doses of siTDP43 assessed. While changes in TDP-43 levels were undetectable by WB at siTDP43 doses of 37.5 - 75 pM (measured as 2% TDP-43 KD by WB), the CUTS system demonstrated a 7 to 55-fold increase in GFP expression at these doses compared to baseline (Figures 3C-3D). This increase in GFP expression continued consistently up to the highest dose of siTDP43 (1 ,200 pM; 98% TDP-43 KD), showcasing a 118,224-fold increase in GFP compared to baseline. Pearson's correlation analysis confirmed a highly significant relationship between GFP expression and siTDP43 dose (P = 0.0011 ), while the correlation between measurable TDP-43 and siTDP43 concentration was less significant (P = 0.0429). Linear regression analysis between the logarithmic GFP fold increase andsiRNA doses demonstrated an exceptional linear relationship (R2 = 0.9998). These results indicate that GFP expression produced by CUTS is a more sensitive method for detecting TDP-43 KD (and therefore LOF) than TDP-43 WB detection. The linear relationship between GFP and siTDP-43 dose also demonstrates the potential of CUTS to be used as a predictive model for TDP-43 LOF.

[0144] We assessed CUTS' sensitivity at the transcript level using the siTDP43 dose curve and RT-qPCR assessment in Figures 3C-3D. To determine the relative amount of CUTS' CE retention, we designed primers targeting either the entire transcript or the specific junction sites of the CE (Figures 3E-3F). The RT-qPCR quantification demonstrated increased sensitivity at detecting changes in TDP-43 levels compared to WB analysis, with the capability of detecting changes in TDP-43 between each siTDP43 dose (1 - 70% TDP-43 KD) in a linear manner (R2 = 0.9303). Using CUTS detection, we observed a clear linear logarithmic relationship between the amount of CUTS CE-retention and the increasing siTDP43 doses (R2=0.9994). Even at 1 % KD in TDP-43, the CUTS system detected a 3-fold increase in CE retention compared to baseline, which increased to a 1 ,488-fold increase at 70% TDP-43 KD (Figure 3G). As with our WB analysis (Figure 3D), there was a significant correlation between both GFP expression and TDP-43 when correlated to siTDP43 dose (P < 0.0001 and P = 0.0279, respectively). Thus, these data indicate that CUTS is a reliable approach to quantifying TDP-43's loss across an extensive range, highlighting its ability as a TDP-43 LOF biosensor. Additionally, CUTS demonstrated ultra-sensitivity under low-level TDP-43 KD, beyond the detection limit of both WB and RT-qPCR.

[0145] In ALS / FTLD, the absolute TDP-43 level remains largely unaffected. Instead, TDP-43 undergoes pathological mislocalization and / or phase transitions likely due to a reduction in RNA binding, which reduces the functional cellular TDP-43. To evaluate whether these events contribute to TDP-43 loss-of-function, we tested CUTS's ability to detect TDP-43 LOF caused by TDP-43 mislocalization or aggregation via aberrant phase transitions. We transfected CUTS HEK293 cells with four tagged TDP-43 isoforms: (1 ) TDP-43WT, (2) TDP-43cyto, (3) TDP-435FL, and (4) TDP-43cyto 5FL. The TDP-43cyto variants contain point mutations located within the nuclear localization signal (NLS) of TDP-43, resulting in cytoplasmic mislocalization. The 5FL form contains five phenylalanine-to-leucine mutations within the two RNA recognition motif (RRM) domains of TDP-43 that greatly impaired TDP- 43's RNA binding ability and were previously reported to form aggregated "anisomes"inside the nucleus. The TDP-43cyto 5FL combines both modifications, leading to insoluble cytoplasmic inclusions. Excluding TDP-43WT, all three modified versions have proven to sequestrate endogenous TDP-43 into mislocalized or aggregated inclusions. Therefore, our objective was to utilize CUTS to determine whether the expression of these aggregation-prone TDP-43 variants elicits TDP-43 LOF.

[0146] The introduction of TDP-43cyto, TDP-435FL, and TDP-43cyto 5FL induced nuclear GFP signal when assessed by immunofluorescence analysis, while neither the tagged plasmid backbone nor TDP-43WT caused any detectable GFP (Figure 4A). We confirmed the expression of exogenous and endogenous TDP-43 levels by WB and quantified the relative GFP level in each condition (Figures 4B-4C). As we have previously shown that CUTS demonstrates a proportional response to TDP-43's LOF (Figures 3A-3G), we were able to directly interpret the relative ability of the different TDP-43 mutants to trigger TDP-43 LOF by comparison of their GFP levels. All three TDP-43 mutants' expression triggered significant LOF compared to the control conditions, albeit at varying significance levels. The most modest LOF effect was achieved by TDP-43cyto, followed by TDP-43cyto 5FL, and TDP-435FL (Figure 4C). Interestingly, although both TDP-435FL and TDP-43cyto 5FL caused significantly elevated LOF, the TDP-435FL mutant alone mediated greater LOF than when combined with the NLS mutations highlighting the potential role of nuclear homotypic TDP-43 interactions potentially contributing to TDP-43 LOF in disease absent its cytoplasmic mislocalization. To further validate the functionality of exogenous TDP- 43, we transfected the same tagged TDP-43WT into a HeLa TDP-43 knock-out cell line expressing CUTS. We detected a significantly decreased GFP signal compared to the backbone or non-transfected controls, which confirmed the full splicing function of TDP-43WT (Figures 5A-5C). Owing to CUTS' performance, we show the functional consequence caused by TDP-43’s mislocalization and / or aberrant phase transitions, demonstrating that aggregation-prone TDP-43 variants mediated direct LOF toxicity in addition to any gain-of-function (GOF) toxic events. Furthermore, these results strongly support CUTS's capability in measuring functional TDP-43 levels under broader contexts beyond TDP-43 KD.

[0147] Given the growing recognition of the role TDP-43 LOF is believed to play in disease progression, numerous efforts have been committed to developing rescue methods aimed at re-delivering TDP-43 or other gene payloads to restore its physiological splicing function. However, a significant challenge in LOF therapies liesin maintaining precise TDP-43 levels within neurons, as even slight overexpression can lead to GOF toxicity. Consequently, a generalized TDP-43 gene-replacement therapy without genome integration carries a substantial risk of overexpression toxicity. Therefore, a CE biosensor such as CUTS may be used to control cell- and temporal-specific regulation of a gene payload. To test this, we generated a CUTS- controlled TDP-43 (CUTS-TDP43) transgene (Figure 6A). Considering the ultrasensitivity to TDP-43 LOF and minimal leakage under physiological TDP-43 levels, CUTS-TDP43 may have the potential to autonomously negatively regulate its expression, ensuring levels will not surpass physiological levels.

[0148] To test this, we created a new polyclonal stable line in HEK293 cells (CUTS-TDP43) by replacing the 3xNLS in the original CUTS cassette with the TARDBP ORF fused to a GFP reporter (Figure 6A). However, as our siTDP43 targets the sequence within the coding region, CUTS-TDP43 was also knocked down upon siRNA transfection, shown by a generalized decreased mCherry signal (Figure 6B). Therefore, we designed a codon-optimized CUTS-TDP43CO that is not targeted by siTDP43 (Figure 6B). Live imaging analyses showed that CUTS GFP signal demonstrated a steady increase in expression in response to increasing doses of siTDP43; however, the GFP signal from CUTS-TDP43CO remained undetectable (Figure 6B). WB analysis further demonstrated successful TDP-43 rescue under endogenous TDP-43 KD, as shown by the increasing exogenous TDP-43 observed following decreases in endogenous TDP-43 (Figure 6C). The amount of total TDP-43 appeared to remain consistent throughout the increasing siTDP43 doses, indicating tight regulation of the rescue parameters. To further confirm whether CUTS-TDP43CO could rescue TDP-43 splicing functionality, we performed a CFTR minigene assay. The expression of CUTS-TDP43SO demonstrated partial, yet significant rescue of cryptic exon 9 splicing in CFTR minigene, supporting its controlled efficacy in rescuing splicing LOF (Figures 6D-6E). Taken together, these data indicate that CUTS can autoregulate a TDP-43 payload to physiological levels in response to TDP-43 knockdown.Discussion

[0149] We developed and characterized the CUTS system, a novel approach to detect TDP-43 LOF. The CUTS system utilizes TDP-43-dependent CE events to correlate the level of TDP-43 LOF directly with the expression of a reporting gene. By combining the CFTR-TS and UNC13A-TS, our findings demonstrate that the CUTSsystem provides an optimal balance of sensitivity and accuracy. This was evidenced by the ability of CUTS to detect modest levels of TDP-43 LOF, as shown by the dose- responsive increase in the expression of the reporter gene, GFP, under various siTDP- 43 concentrations (Figures 3A-3G). Furthermore, our results suggest that CUTS can effectively discern TDP-43 LOF induced by pathological phase transitions or mislocalization, a critical aspect in the context of neurodegenerative diseases containing TDP-43 pathology, such as ALS and FTLD. The CUTS system's potential for application in gene-replacement therapies was also highlighted, offering a promising avenue for autoregulated rescue of TDP-43 function, which is critical for avoiding the deleterious effects of TDP-43 overexpression.

[0150] This tool findings enables for the significant advancement in the capacity to detect TDP-43 LOF using biosensor assays across diverse experimental settings and through multiple analytical methods. Previously, the CFTR minigene assay has been the predominant approach for detecting TDP-43 LOF. However, this approach is associated with several limitations, all of which are effectively addressed by utilizing the CUTS system. The first advantage of the CUTS system is its ability to detect TDP- 43 LOF in real-time through live-imaging analysis. Unlike CFTR minigene assays, which typically necessitate endpoint experimental analysis, CUTS facilitates continuous monitoring, eliminating the need for multiple fixed time points. Additionally, CUTS can be seamlessly integrated with various analytical methods, including RT- qPCR (at the RNA level), WB analysis (at the protein level), live imaging (for real-time assessment), and immunofluorescence imaging (to correlate with relevant markers). While not evaluated in this study, it is conceivable that CUTS could be adapted for use with flow cytometry-based techniques, leveraging GFP-positive cells as an output for analysis, as previously demonstrated with a CFTR-modified sensor.

[0151] In addition to the expanded array of analytical methods offered by CUTS compared to CFTR minigene assays, we anticipate that the CUTS system will exhibit superior sensitivity and accuracy. This is supported by the comparison of CUTS with the CFTR-TS or UNC13A-TS cassette (Figures 1A-1J), underscoring the potential of CUTS to outperform single minigene-based approaches in TDP-43 LOF detection. While recent work in two recent preprints suggests other CE biosensors are in development, CUTS appears to exhibit enhanced sensitivity. The expression of the GFP reporter in CUTS achieved up to 118,224-fold increase upon TDP-43 knockdown, compared to ADNP2 (< 5-fold); TDP-REGv1 (<20-fold); and TDP-REGv2 (<300-fold).We also show that CUTS can detect ultra-low levels of TDP-43 knockdown (increasing > 7-fold), below the WB or RT-qPCR detection limit. Furthermore, CUTS exhibits a robust log-linear relationship to siRNA doses, making it suitable for quantitative purposes.

[0152] Due to the high flexibility of the CUTS system, its application can be expanded in vitro and in vivo when coupled with disease models. Integrating the CUTS system with disease models enables the evaluation of the model's fidelity in recapitulating TDP-43 LOF phenotypes. Such assessments are crucial for selecting appropriate models that faithfully replicate the desired study context. Furthermore, coupling the CUTS system with TDP-43 models presents a valuable approach for diverse screening studies. For example, CUTS can be leveraged for high-throughput drug screening and CRISPR screening methodologies. Such approaches hold promise for uncovering critical insights into cell-specific disease mechanisms, identifying pivotal disease modifiers, and delineating potential therapeutic genetic targets.

[0153] A significant advantage of the CUTS system lies in its capacity to deliver precisely regulated gene therapy for rescuing TDP-43 LOF. This study illustrates this capability by placing a functional TDP-43 transcript downstream of the CUTS regulatory elements. The system's self-regulating ability enhances its safety profile as a gene therapy approach, ensuring gene expression occurs only when necessary and exclusively in cells lacking TDP-43 function. Furthermore, this system can be expanded by substituting the TDP-43 transcript with other genetic modifiers of disease, such as antibodies and PROTACs, or genes with established therapeutic potential, including heat shock proteins (HSPs) or heterogeneous nuclear ribonucleoproteins (hnRNPs). This adaptability holds promise for achieving safe therapeutic outcomes without the need for direct TDP-43 expression.Example 2Methods

[0154] Plasmid construction

[0155] All the vectors, unless specified from other resources, were generated by NEBuilder HiFi DNA Assembly Master Mix (NEB, E2621 L) following the manufacturer's protocol.

[0156] All the plasmids were verified using whole-plasmid sequencing via Oxford Nanopore provided by Plasmidsaurus.

[0157] Cell culture and plasmid transfection

[0158] HEK293 cells (female, purchased from ATCC), Hela TDP-43 knock-out (KO) cells (a kind gift from Dr. Shawn M Ferguson58 were maintained in DMEM (Thermo Fisher Scientific) supplemented with 10% HyClone Bovine Growth Serum (GE Healthcare Life Sciences) and 1x GlutaMAX (Thermo Fisher Scientific) at 37°C and 5% CO2, with a humidified atmosphere. Cells were seeded onto coverslips or plates coated with collagen (50 mg / mL, GIBCO) and allowed to incubate overnight prior to transfections using Lipofectamine 3000 (Invitrogen, L3000015) with a proper amount of DNA performed according to manufacturer's instructions.

[0159] Piggybac stable cell line generation

[0160] HEK293 cells were seeded onto 6-well plates 24h before transfection and were grown to about 70% confluence. 2.5ug of plasmid with Piggybac backbone was co-transfected with 0.5ug Super PiggyBac Transposase (PB200PA-1 ) plasmid respectively to each well using Lipofectamine 3000 (Invitrogen, L3000015) according to manufacturer's instructions. One well without transposase was also conducted as the negative control. The media was changed 48h after transfection into selection media with puromycin (Sigma, P8833) to a final concentration of 5 pg / mL, or hygromycin B (Gibco, 10687010) to a final concentration of 500 pg / mL. The selection media was then refreshed every 2 days. The control cells were all died in about 5 days. The surviving cell lines were continuously cultured after selection and passaged for 2- 3 generations with a reduced concentration of puromycin (2.5 pg / mL) or hygromycin B (250 pg / mL) to obtain stable cell lines. The induced expression of desired constructs was verified via Immunofluorescence staining or Western blotting.

[0161] APEX2-medicated proximity labeling

[0162] Stable HEK293 cell lines expressing 3xFlag-APEX2-TDP-43 (WT, ANLS, 5FL, 5FL+ANLS), 3xFlag-APEX2-3XNES, and 3xFlag-APEX2-3XNLS were split into 24-well plates containing pre-coated coverslips (for Immunofluorescence staining), 6- well plates (for Western blotting), or 10-cm dishes (for Mass Spectrum) 24h before induction. Doxycycline (Sigma, D9891 ) was added into the media to a final concentration of 1 pg / mL to induce the expression of APEX2 constructs. For transient transfected experiments, wildtype HEK293 cells were transfected with pCMV-3xFlag- APEX2-TDP-43 (WT, ANLS, 5FL, 5FL+ANLS) plasmids using Lipofectamine 3000. After 48h of expression (for transfection, 24h), the plates were replaced with fresh media containing 500 pM of biotin phenol (ApexBio, 41994-02-9) and were incubatedfor 30 min. The labeling was initiated by adding H2O2 (Sigma, 216763) to reach a final concentration of 1 mM for precisely 1 min. The reaction was quenched by immediately adding freshly prepared 2X quenching buffer (10 mM Trolox (Sigma, 238813), 20 mM sodium ascorbate (Sigma, A4034), and 20 mM sodium azide (Fisher Scientific, BP922I) in pre-chilled DPBS (Thermo Scientific, 14190235)) of the same volume as the media, and rinsing two more times with 1X quenching solution (50% 2X quenching buffer diluted with pre-chilled DPBS). Negative controls without biotin phenol (BP), H2O2, or doxycycline (Dox) were also included in the validation tests. The cells were further proceeded according to different experimental purposes.

[0163] Stressor treatment

[0164] All the stressors were mixed with the media to desired final concentration and added to cells via media changing. For recovery, the cells were washed one time with fresh media and then changed into recovery media.

[0165] RNA Oligonucleotides

[0166] RNA oligonucleotides ([AC]i7 and [UG]i?) were ordered from Horizon Discovery, HPLC purified, fully 2’0Me modified.

[0167] SDS-PAGE and Western blotting

[0168] Cells were lysed on-plate with pre-chilled RIPA buffer (Boston Bioproducts, BP-115X) with 1 % protease inhibitor (PI) cocktail (Sigma, P8340) on ice for 10 min and then sonicated at 20% power by a sonicator (Brason) for 3 times, with each time 15 s in an ice bath and 2 min of cooling between each sonication. Protein concentrations were measured using the Pierce BCA Protein Assay Kit (Thermo Scientific, 23227). The samples were then denatured by adding 4X Laemmli Sample Buffer (Biorad, 1610747) and incubating at 95 °C for 10 min before Sodium Dodecyl Sulfate PolyAcrylamide Gel Electrophoresis (SDS-PAGE). The denatured samples were separated by SDS-PAGE using 4%-20% Mini-PROTEAN TGX Precast Gels (Biorad, 4561096) and transferred to 0.45 pm Nitrocellulose Membranes (Biorad, 1620146) using Mini Gel Tank and Blot Module Set (Invitrogen, NW2000). Total protein levels were measured by Ponceau S dye (ThermoFisher A40000279). Following water and TBS washes, membranes were blocked in 5% non-fat milk (Thermo Scientific, 50-488-786) in TBS-T (0.1 % Tween 20 (Fisher Scientific, BP337) in TBS) and incubated with primary antibodies in the blocking buffer at 4°C overnight with mild shaking. The primary antibodies and their dilution ratios were as follows: mouse-anti-FLAG (Sigma, F1804), mouse-anti-GAPDH (Cell Signaling, 2118),mouse-anti-GFP (Santa Cruz, sc-9996), mouse-anti-TDP-43 (Proteintech, 60019-2- Ig), rabbit-anti-TDP-43 (Proteintech, 10782-2-AP), rabbit-anti-mCherry (Cell Signaling, 43590), rabbit-anti-DBR1 (Proteintech, 16019-1 -AP), rabbit anti-NUFIP2 (Proteintech, 17752-1 -AP), rabbit-anti-HNRNPC (Proteintech, 11760-1 -AP), rabbit-anti-HNRNPL (Proteintech, 18354-1 -AP). After incubation, membranes were washed thrice with TBS-T and incubated with Donkey-anti-mouse (JacksonlmmunoResearch, 715035151 ) or Donkey-anti-rabbit (JacksonlmmunoResearch, 711035152) secondary antibodies conjugated with HRP, or Streptavidin-HRP (EMD Millipore, OR03L) at room temperature for 1 h. After washing again with TBS-T for three times, the membranes were visualized using Western Lightning ECL Pro (Revvity, NEL1201001 EA) or Supersignal West Femto Maximum Sensitivity Chemiluminescent Substrate (Thermo Scientific, 34095) in an Amersham ImageQuant 800 GxP biomolecular imager system (Amersham, 29653452). Western blotting results were quantified using Fiji ImageJ (V2.3.1 ) software.

[0169] RIPA soluble / insoluble (sol / insol) fractionation

[0170] HEK293 cells were processed for fractionation directly from culture. Briefly, cell pellets scraped with pre-chilled PBS were transferred into 1.5 mL microcentrifuge tubes. The samples were centrifuged at 500* g for 5 min at 4°C to pellet the cells. The supernatant was then carefully removed, and each pellet was resuspended in prechilled RIPA buffer supplemented with PI. The suspensions were incubated on ice for 10 min to facilitate complete lysis. Following lysis, the samples were centrifuged at 17,000 x g for 10 min at 4°C. The supernatants, which constitute the RIPA-soluble fraction, were collected and stored at -20°C until further analysis. The remaining insoluble material was further processed by washing. Each pellet was resuspended in the same RIPA buffer (with PI) and centrifuged at 17,000 x g for 5 min at 4°C. This wash was repeated for a total of two washes, with the supernatant discarded after each centrifugation. Thereafter, the insoluble pellets were resuspended in a modified RIPA buffer containing 4 M urea, PI, and Benzonase (Sigma, E1014-25KU). The resuspended pellets were then agitated at 1350 rpm for 10 min on a thermomixer (Eppendorf) at 25°C to promote extraction of the remaining proteins. A single 10- second sonication pulse at 20% power was applied to fully dissolve the pellets and acquire the final insoluble fraction. Both soluble and insoluble fractions were subsequently analyzed by SDS-PAGE and western blotting, as described above.

[0171] Immunofluorescence staining

[0172] Cells on coverslips were fixed with 4% PFA (Electron Microscopy Sciences, 15714-S) at room temperature for 20 min. The cells were then washed with PBS (Fisher Scientific, 20012050) and blocked with 5% Normal Donkey Serum (Jackson ImmunoResearch, 017-000-121 ) in PBS-T (0.1 % Tween 20 in PBS). The cells were incubated with primary antibodies in the blocking buffer at 4°C overnight. The primary antibodies and their dilution ratios were as follows: mouse-anti-FLAG (Sigma, F1804), mouse-anti-G3BP1 (Santa Cruz, sc-365338), mouse-anti-TDP-43 (Proteintech, 60019-2-lg), mouse-anti-SC35(SRRM2) (Abeam, ab11826), rabbit-anti-TDP-43 (Proteintech, 10782 -2 -AP), rabbit anti-ATXN2 (Proteintech, 21776-1 -AP), rabbit anti- NUFIP2 (Proteintech, 17752-1 -AP). After incubation, cells were washed thrice with PBS-T and incubated with secondary antibodies, Streptavidin, or Hoechst at room temperature for 1 h as follows: DyLight 405 AffiniPure Donkey Anti-Mouse IgG (H+L) (Jackson ImmunoResearch, 715-475-151 ), Alexa Fluor 488 Donkey anti-mouse IgG (H+L) (Jackson ImmunoResearch, 715-545-150), Alexa Fluor 488 Donkey anti-rabbit IgG (H+L) (Jackson ImmunoResearch, 711 -545-152), Alexa Fluor 594 Donkey antimouse IgG (H+L) (Jackson ImmunoResearch, 715-585-151 ), Alexa Fluor 594 Donkey anti-rabbit IgG (H+L) (Jackson ImmunoResearch, 711 -585-152), Alexa Fluor 647 Donkey anti-mouse IgG (H+L) (Jackson ImmunoResearch, 715-605-150), Alexa Fluor 647 Donkey anti-rabbit IgG (H+L) (Jackson ImmunoResearch, 711-605-152), Cy3- Streptavidin (Jackson ImmunoResearch, 016-160-084), Cy5-Streptavidin (Jackson ImmunoResearch, 016-170-084), and Hoechst 33258 solution (Sigma, 94403). Cells were washed two times in PBS-T and one time in PBS. The coverslips were mounted onto slides with Vibrance Antifade Mounting Medium (Vector Laboratories, H-1700) for 24 h away from light. Then they were proceeded to be visualized by confocal microscopy.

[0173] Confocal microscopy

[0174] All confocal experiments were performed on a Nikon A1 laser-scanning confocal microscope system. For regular confocal microscopy, CFI Plan Apo Lambda 60X Oil immersion objective (Nikon) was used. For live-cell imaging, a Tokai HIT stage-top incubator preheated (37°C and 5% CO2) for 10 min prior to imaging, and CFI Plan Apo Lambda 10X or 60X (Oil) objectives (Nikon) were used. Nikon Elements imaging software was used to control the microscope and perform visualization. The images presented were representatives of at least two independent experiments with three or more biological replicates per experiment. The data presented arerepresentatives of at least two independent experiments utilizing three or more replicates per experiment.

[0175] Fluorescence recovery after photobleaching (FRAP)

[0176] FRAP imaging was performed on HEK293 cell lines cultured in DMEM containing 10% FBS on 24-well glass-bottom plates. Where indicated, cells expressing EGFP-TDP-43 variants were induced for 48 hours by adding doxycycline (1 pg / mL) to the medium. Using a Nikon A1 laser-scanning confocal microscope with a 60* oilimmersion objective, granules or inclusions were located, and circular regions of interest (ROIs) with a diameter of roughly 2 pm were photobleached for 500 ms at 50% laser power (488 nm line). For each experiment, 2-5 baseline images were examined before bleaching, and fluorescence recovery was monitored for up to 3 minutes at 5-second intervals. Nikon NIS-Elements software was used to quantify recovery curves, normalizing the fluorescence intensity in the bleached region to its pre-bleach level and setting the intensity immediately after bleaching to zero. At least 6-10 bleached structures per condition were analyzed to calculate averages and standard deviations. The data presented are representatives of at least 5 independent images, from at least two biological replicates.

[0177] IHC and IF staining in post-mortem tissue

[0178] Primary antibodies included NUFIP2 (Proteintech 17751 -1 -AP), HNRNPC (Proteintech 11760-1 -AP), and 1 D3 (pTDP-43, a gift of Elisabeth Kremmer and Manuela Neumann). Sections were subject to microwave antigen retrieval and stained with primary antibody overnight. For immunohistochemistry, NUFIP2 was diluted 1 :250, HNRNPC was diluted 1 :2000, and 1 D3 was diluted 1 :400; the next day, biotinylated secondary antibodies were added, followed by an ABC kit (Vectorlabs) and visualized using immPACT DAB (Vectorlabs). For immunofluorescence, NUFIP2 was diluted 1 :200, HNRNPC was diluted 1 : 1000, and 1 D3 was diluted 1 :200; the next day, fluorescently labeled secondary antibodies (Invitrogen Alexa Fluor 488G and Alexa Fluor 594R) were added, followed by DAPI Fluoromount-G (SouthernBiotech). Images were captured on a Nikon Eclipse Ni microscope.

[0179] siRNA transfection

[0180] The siRNA (siGenome Smatpool siRNA, Dharmaco) information is published. siGENOME non-Targeting siRNA for control (Dharmaco, D-001206-13-05) was used as control (siControl). The indicated amount (usually to the final concentration of 20 nM) in each experiment was transfected into HEK 293 cells usingLipofectamine RNAiMAX reagent (Invitrogen, 13778150) according to the manufacturer's protocol. For TDP-43 knockdown, the following siRNA was used: ON- TARGETplus SMARTpool siRNA against TARDBP (Dharmaco, L-012394-00-0005).

[0181] cDNA and ORF plasmids

[0182] The cDNA and ORF-expressing plasmids information for CUTS screening and following experiments is published.

[0183] Streptavidin pull-down of biotinylated proteins

[0184] For the enrichment of biotinylated proteins, 15% of the cell lysates from one 10-cm dish were diluted five times with RIPA buffer and incubated with 50 pL NanoLINK Streptavidin Magnetic Beads (TriLink Biotechnologies, M-1002) overnight at 4? on a rotator. Beads were collected against a magnetic stand, and the supernatant was set aside for future analysis (termed flow-through). A total of eight washes were performed prior to on-beads digestion: (1 ) RIPA; (2) 50% RIPA, 50% UWB; (3) UWB (20 mM Tris-HCI pH7.6 (Rockland, MB-003), 150 mM NaCI (Invitrogen, AM9759), 2M Urea (Fisher Scientific, M-13269)); (4) 50% UWB, 50% RIPA; (5) RIPA; (6) ABC (50mM pH=8.0 ammonium bicarbonate (Sigma, A6141 )); (7) ABC; (8) ABC. For each time of the washes, the beads were completely resuspended by vortex and rotated end-over-end on a rotator for 5 min prior to magnetic separation. Before the last spin, 10% of the beads were saved for quality control.

[0185] On-beads digestion

[0186] Roughly 2.5 pg of Sequencing Grade Modified Trypsin (Promega, V5113) was added to the bead in ABC together with 1 mM DTT (Sigma, 43816) to increase the digesting efficiency165. The samples were then digested at 35°C overnight on a thermomixer (Eppendorf) shaking at 1000 rpm. After digestion the supernatant was removed and the beads were washed once with enough ABC to cover. After 10 minutes of gentle shaking, the wash was removed and combined with the initial supernatant. The peptide extracts were lyophilized, then resuspended in 30 uL 0.1 % trifluoroacetic acid. A small portion of the extract is used for fluorometric peptide quantification (Thermo Scientific Pierce).

[0187] Label-free DIA LC-MS

[0188] For each sample, 500 ng total peptide was loaded onto a disposable Evotip C18 trap column (Evosep Biosystems, Denmark) and subjected to nanoLC on an Evosep One instrument (Evosep Biosystems). Tips were eluted directly onto a PepSep analytical column, dimensions: 150umx25cm C18 column (PepSep, Denmark) with1.5 pm particle size (100 A pores) (Broker Daltronics). Mobile phases A and B were water with 0.1 % formic acid (v / v) and 80 / 20 / 0.1 % ACN / water / formic acid (v / v / vol), respectively. The standard pre-set method of 100 samples-per-day was used, which is a 14 min run. The mass Spectrometry was done on a hybrid trapped ion mobility spectrometry-quadrupole time of flight mass spectrometer (timsTOF HT, (Broker Daltonics, Bremen, Germany), operated in PASEF mode. The acquisition mode was Data-independent analysis (DIA). The acquisition scheme used for DIA consisted of 18 precursor windows at a width of 50m / z per cycle. The TIMS scans layer the doubly and triply charged peptides over an ion mobility, 1 / k0 range of 0.60-1.4 V*sec / cm2. Precursor windows began at 360 m / z and continued to 1200 m / z. The collision energy was ramped linearly as a function of the mobility from 63 eV at 1 / K0=1 .4 to 17 eV at 1 / K0=0.6.

[0189] MS data processing

[0190] Raw files were processed with Spectronaut version 18 (Biognosys, Zurich, Switzerland) using DirectDIA analysis mode. Mass tolerance / accuracy for precursor and fragment identification was set to default settings. The reviewed FASTA for Homo Sapiens, UP000005640 downloaded from Uniprot (on 11 Sep 2023) and a database of 112 common laboratory contaminants (https: / / www.thegpm.org / crap / ) were used. A maximum of two missing cleavages were allowed, the required minimum peptide sequence length was 7 amino acids, and the peptide mass was limited to a maximum of 4600 Da. Carbarn idomethylation of cysteine residues was set as a fixed modification, and methionine oxidation and acetylation of protein N termini as variable modifications. A decoy false discovery rate (FDR) at less than 1 % for peptide spectrum matches and protein group identifications was used for spectra filtering (Spectronaut default). Decoy database hits, proteins identified as potential contaminants, and proteins identified exclusively by one site modification were excluded from further analysis.

[0191] Method section for APEX-TDP project

[0192] Post-MS data processing

[0193] 7487 proteins were obtained from mass spectrometry and protein intensity values were log-transformed. Proteins with more than 50% missing values or mass spectrometry detection p value > 0.05 were removed, and a total of 6592 proteins were kept. We used impute. knn with default parameters from R package impute (DOI: 10.18129 / B9.bioc.impute) for missing data imputation.

[0194] Definition of TDP-43 interactome and differentially enriched TDP-43 interactors

[0195] To define the TDP-43 interactome under each experimental condition, we used nuclear export signal (NES) as a control for cytoplasm-localized TDP-43 and nuclear localization signal (NLS) as a control for nucleus-localized TDP-43. Differential interactors were identified by performing a t-test on log-transformed protein intensities between experimental groups and their respective controls. Proteins with Iog2-fold change (LFC) > 0.3 and p-value < 0.05 were considered part of the TDP-43 interactome. The resulting interactome lists were used as input for Gene Ontology (GO) term enrichment analysis using DAVID GO (Database for Annotation, Visualization, and Integrated Discovery), with default parameters and an FDR- corrected significance threshold of q < 0.05.

[0196] To define differential TDP-43 interactors between specific experimental conditions, we performed t-tests on log-transformed protein intensities between experimental groups and their matched controls. For comparisons within the same cellular compartment (both nuclear or both cytoplasmic), interactors were defined using LFC > 0.3 and p-value < 0.05. For comparisons across cellular compartments (nuclear vs. cytoplasmic), interactors were defined using LFC > 1 and q-value < 0.05 to account for larger expression differences. The lists of differential interactors were also analyzed for GO term enrichment using DAVID GO, with default parameters and an FDR-corrected significance threshold of q < 0.05.

[0197] To identify interactome alterations associated with RNA-binding impairment, TDP-43 mislocalization, and oxidative stress treatment respectively, we performed differential analysis using the limma package. Each experimental condition was modeled as an independent factor while adjusting for potential confounding effects from the other two conditions. Specifically, three separate linear models were fitted:1. RNA-binding impairment analysis: Differential expression was assessed between RNA-binding deficient and RNA-binding competent groups while controlling for mislocalization and oxidative stress treatment.2. TDP-43 mislocalization analysis: Cytoplasmic and nuclear TDP- 43 conditions were compared, adjusting for RNA-binding impairment and oxidative stress.3. Oxidative stress treatment analysis: Cells treated with sodium arsenite were compared to untreated controls, controlling for TDP-43 mislocalization and RNA-binding impairment.

[0198] For each contrast, a design matrix incorporating the focal variable and the additional covariates was constructed, and differential expression was computed using Limma's empirical Bayes framework. Significantly altered genes were identified using moderated t-statistics and ranked based on adjusted p-values.

[0199] Normalization of LFCs for cross-comparison analysis

[0200] To enable direct comparison across nine differential interactome analysis as well as differential expressed genes (DEG) from ALS / FLTD snRNA-seq, we adapted the normalization method from the snRNA-seq paper. We standardized the magnitude of LFCs by converting them into Z-scores. For each differential comparison, LFC values were transformed using the R function scale(x, center = 0, scale = sd(x)), where x represents the vector of LFCs for that specific comparison. This transformation ensures that the LFC values are centered at 0 while normalizing their magnitude relative to the variability within each dataset. By applying this Z-score transformation, we accounted for differences in dynamic range across conditions and facilitated a direct comparison of relative effect sizes across all nine differential analyses.

[0201] Identification and clustering of common significant genes

[0202] To identify common significant genes between our differential interactome analysis and patient-derived snRNA-seq data, we applied distinct selection criteria. RNA-seq differentially expressed genes (DEGs) were identified using thresholds from the original literature: |Z-score| > 1 and FDR-adjusted p-value < 0.05. For differential interactome analysis under mislocalized, oxidative stress (AS), and impaired RNA- binding conditions, a gene was considered significant if at least one of the three conditions showed p < 0.05 (t-test). The final common significant gene set was defined by the intersection of significant interactome genes across conditions with the RNA- seq DEGs.

[0203] To explore gene expression patterns, we performed hierarchical clustering using the pheatmap package in R. Both genes and conditions were clustered (cluster_rows = TRUE, cluster_cols = TRUE) with default settings. To refine functional classification, we applied cutree_rows = 15, partitioning the genes into 15 clusters. The resulting clusters were subsequently used as input for GO term enrichment and correlation analysis.

[0204] RNA extraction

[0205] Total RNA was extracted with RNeasy Mini Kit (Qiagen, 74106) following the manufacturer's protocol. The concentration of extracted RNA was measured by Nanodrop Spectrophotometer (Nanodrop, ND-1000). For mRNA sequencing, the RNA was accurately measured by Qubit RNA HS Assay Kit (Invitrogen, Q32852).

[0206] RT-PCR

[0207] The extracted total RNA was reverse-transcripted into cDNA using Iscript Reverse Transcription Supermix (Biorad, 1708841 ) following the manufacturer's instructions. Primer information is published. PCR products were separated by agarose gel electrophoresis, and the bands were visualized with Amersham ImageQuant 800 GxP biomolecular imager system.

[0208] mRNA sequencing and DEG analysis

[0209] mRNA-seq libraries were generated using the Illumina Stranded mRNA Prep Kit (Illumina, 20040534), following the protocol provided by the manufacturer. Library concentration and integrity were then evaluated with a Qubit fluorometer and an Agilent TapeStation. Subsequent sequencing was carried out on an Illumina NextSeq 2000 P3 platform, producing 100 bp paired-end reads. Each experiment was performed with three biological replicates, and each sample yielded approximately 50- 60 million raw reads.

[0210] For a quick DEG analysis, the RNA was processed using the Salmon pipeline166 with the default setting and quantified using the GRCh38.p13 human genome build with an Ensembl version of the Gencode V45 transcripts. The processed data were in the form of TPM (transcript per million) for mRNAs. For comparisons against the CUTS-only (TKR1 ) group, replicates #2 was excluded due to being an outlier in PCA. One-way ANOVA or two-way ANOVA followed by a Tukey's test was conducted when data underwent normality statistical comparisons between different groups.

[0211] Sequencing data processing and alignment for splicing analysis

[0212] Trimmed reads were aligned to the GRCh38.p13 human genome build with Gencode V38 gene models using STAR (v2.7.5a). Samples were aligned using ENCODE options as described in the STAR manual167, as well as two-pass mapping (--twoPassMode Basic). Gene-level counts were quantified with featureCounts (v2.0.1 ), and sample clustering was manually checked with DESeq2 (v1 .44.0) by using the vst() and plotPCAQ functions.

[0213] Splicing analysis with MAJIQ

[0214] Aligned reads were analyzed for group-level splicing alterations using MAJIQA / OILA (v2.4.dev102+g2cae150). Briefly, all samples were used to build a database of splicing events with majiq build, using the flags --min-intronic-cov 1 , -- simplify, and default settings otherwise. Splicing changes between groups were analyzed with majiq deltapsi with default settings. For comparisons against the CUTS- only (TKR1 ) group, replicates #2 was excluded due to being an outlier in PCA. Results tables were created with voila tsv and parsed using in-house R code. Any splice junction with |APSI|>0.1 and probability of changing >0.95 between conditions was considered a significant differential splicing event.

[0215] Statistical Analysis

[0216] Unless specified, quantitative data was analyzed using GraphPad Prism 10.0 software. The specific statistical significance analysis methods, p-value indicators, and number of replicates in each experiment are detailed in the respective figure legends.Results

[0217] Cellular mislocalization, RNA-binding deficiencies, and extracellular cell stressors induce TDP-43 loss of function

[0218] To identify interactors that modulate TDP-43 function, we utilized the CFTR-UNC13A TDP-43 Sensor (CUTS), a biosensor that monitors TDP-43 splicing activity through GFP expression upon the aberrant inclusion of a cryptic exon. We examined various triggers of TDP-43 loss of splicing function (LOF), including knockdown, localization mutants, cell stressors, and RNA-binding perturbations. CUTS-HEK293 cell line treated with siRNA-mediated TDP-43 depletion significantly increased GFP fluorescence compared to control siRNA, confirming the CUTS biosensor's sensitivity to TDP-43 LOF. Overexpression of mislocalization and aggregation-prone TDP-43 variants in CUTS-HEK293 cells, including cytoplasmic andRNA-binding-deficient mutants (ANLS, 5FL, 5FL / ANLS) also elevated CUTS-GFP fluorescence, suggesting impaired splicing activity of endogenous TDP-43. Similarly, transfection of these mutant variants (ANLS, 5FL, 5FL / ANLS) did not fully restore TDP-43 function) in TARDBP- / - HeLa cells, as assessed by comparing the GFP signal intensity to TDP-43 rescue. This demonstrates varying degrees of splicing impairment among these dysfunctional variants. CUTS-GFP fluorescence increased in a dosedependent manner upon treatment with cellular stressors known to disrupt TDP-43 localization, including oxidative stress (NaAsO2), hyperosmotic pressure (sorbitol), translation inhibition (puromycin)63, and proteasome inhibition (MG132), highlighting their pathological impact on TDP-43 splicing function. However, we did not detect increased LOF by membrane potential disruption (KCI) or ER stress (Tunicamycin) in the CUTS-HEK93 system. NaAsO2 (AS) triggered the highest level of LOF among all the stressors, which is consistent with recent studies likely due to the rapid formation of splicing-deficient AS-induced nuclear TDP-43 condensates. Additionally, consistent with recent findings, we observed significant TDP-43 LOF in the CUTS system by multivalent UG-rich RNA oligos ([UG]).

[0219] APEX2 Proximity Proteomics to Characterize the Context-Specific TDP-43 Interactome

[0220] To determine the protein interactome contributing to context-dependent TDP-43 LOF identified in, we employed APEX2-mediated proximity labeling. This system enables spatiotemporal biotinylation of protein interactors within ~10-20 nm over a 1 -minute window in live cells. Both transiently expressed cDNA constructs or inducible and stably expressing HEK293 cell lines were generated using the Piggybac system71 with a Tet3G doxycycline-inducible cassette expressing APEX2 -TDP-43 variants (WT, ANLS, 5FL, and 5FL / ANLS) were generated to perform proximity labelling of TDP-43 interacting proteins. While transient overexpression of ANLS, 5FL, and 5FL / ANLS variants formed nuclear or cytoplasmic aggregates as previously reported, the doxycycline-inducible stable cell lines did not exhibit noticeable morphological changes. Such differences might be caused by different protein concentrations, as these phase-separation structures were mostly found following overexpression. Consistent with this, APEX2-TDP-43 levels after transient transfection were higher than endogenous TDP-43, whereas expression levels in the tunable doxycycline-inducible stable cell lines remained comparable to endogenous TDP-43 levels. Therefore, we prioritized the doxycycline-inducible stable cell lines asthey better mimic physiological conditions since TDP-43 is not overexpressed in ALS / FTLD. This approach also allows for the identification of early TDP-43 protein interactors that may trigger insoluble condensate formation.

[0221] To systematically profile TDP-43 interactome changes under different pathological conditions, we combined APEX2-mediated proximity labeling with streptavidin pull-down of biotinylated proteins and label-free data-independent acquisition mass spectrometry (DIA-MS). DIA-MS enables accurate peptide quantification and deep proteome coverage. We selected three key pathological factors that induce TDP-43 LOF — mislocalization (ANLS), impaired RNA binding (5FL), and NaAsO2stress (+AS) — along with their double and triple combinations. To account for compartment-specific effects, we included doxycycline-inducible stable HEK293 cell lines expressing APEX2-3XNLS and APEX2-3XNES as nuclear and cytoplasmic controls, respectively.

[0222] A total of 6592 proteins were quantified across four biological replicates for all conditions. Principal component analysis (PCA) revealed distinct clustering of interactomes based on TDP-43's subcellular localization, RNA-binding ability, and NaAsO2induced stress, confirming a successful capture of context-dependent interactions. To assess dataset reliability, we compared our results with eight previously published TDP-43 interactome lists using various methods and cell lines, where five of them resembled TDP-43WT interactomes, two of them resembled TDP- 43ANLS interactomes, and one with combined datasets. Among 1539 reported interactors, 1134 (73.68%) were also identified by our APEX2 -TDP-43 system, showing strong reproducibility.

[0223] We next identified the TDP-43-associated proteins under each condition by comparing them to their compartment controls (3xNLS for the nucleus, 3xNES for the cytoplasm). TDP-43WT and TDP-43ANLS exhibited significant overlap with reported interactomes, given the relatively strict filter setting. Notably, while all conditions shared core interactions, each exhibited distinct interactors and overlapping features, suggesting TDP-43 resides in diverse cellular environments. Gene ontology (GO) analysis using DAVID GO further highlighted different functional classifications of TDP-43 interactors across various pathological conditions. While all seven groups showed significant enrichment for mRNA processing and splicing interactors, TDP- 43ANLS preferentially interacted with proteins related to P-body formation, mRNA decapping, and stress granule (SG) assembly, while TDP-435FL / ANLS and TDP-435FL / ANLS+AS interacted with proteins involved in DNA damage repair and cell cycle regulation.

[0224] Differential Interactomes Due to TDP-43 Mislocalization, Impaired RNA binding, and NaAsO2-induced Stress

[0225] We performed nine comparisons to determine the impact of TDP-43 mislocalization, impaired RNA binding, and NaAsO2-induced stress on its interactome. Using APEX2-TDP-43 variants, we identified specific interactors gained or lost under each condition, highlighting key functional shifts in TDP-43 protein networks. For TDP-43 mislocalization with ANLS as the single variable, interactome analysis revealed a significant loss of interactions with mRNA splicing factors as judged by GO term enrichment. Network analysis identified gene clusters that were lost in TDP-43's interactome under mislocalization. These included: hnRNPs and snRNPs; transcription elongation factors including POLR2A and CDK9; NURD complex & Histone deacetylase including RBBP4, MTA1 , MBD3, CHD3, and HDAC3; and BAF complex (SWI / SNF proteins). The significant reduction of associated hnRNPs was evident in the ranking plot of lost interactors, supporting TDP-43's critical function in the regulation of splicing in the nucleus together with different hnRNPs. In the cytoplasm, mislocalized TDP-43 gained interactions with translational initiation factors, SG and P-body components, as well as cytoskeleton proteins. These results underscore the critical role of TDP-43's localization in maintaining its association with RNA processing machinery in both the nucleus and cytoplasm, while also linking mislocalization to pathological aggregation-related proteins.

[0226] Impaired RNA binding, modeled by expressing the TDP-435FL and TDP- 435FL / ANLS variants, led to loss of interactions with the cytoplasmic RNA processing machinery, particularly with stress granules (SGs) and P-bodies - interactions that are enriched under mislocalization. This loss was accompanied by a gain in interactions with proteins associated with DNA damage repair pathways, Rho-GTPase signaling (including DSG2 and PKP4), and the nuclear pore complex. These findings suggest that the RNA-binding ability of TDP-43 is critical for its engagement with SG and P- body proteins, and its disruption may redirect TDP-43 toward DNA repair and transport processes.

[0227] NaAsO2-induced oxidative stress introduced a unique shift in the TDP-43 interactome. Although TDP-43 is a well-established component of SGs and P-bodies, its interactions with these structures were unexpectedly lost under NaAsO2 stress,which induces the formation of SGs and P-bodies. Additionally, associations of TDP- 43 with ribosomal proteins (especially from the 43S complex), chromosome centromeric region (including CCNB1 , HJURP, and NLIF2), and lamellipodium- associated proteins (including ABI1 , BAIAP2, and CTTN) were also lost. Conversely, TDP-43 gained associations with tyrosine kinase signaling proteins (including Ras signaling), mitochondrial proteins (including TOMM34, PC, PCCB, and ACACA), and nucleocytoplasmic transport factors (IPO4, IPO5, and XPO5). Interestingly, under NaAsO2 stress, TDP-43WT exhibited bidirectional alterations in its interaction with mRNA splicing factors. Specifically, interactions with EFTUD2, RBM25, SON, FUS, DHX15, CRIPT, PPIG, HNRNPUL1 , and endogenous TDP-43 (suggesting selfaggregation or oligomerization) were gained, while interactions with SRSF3, SRSF10, and SRRM1 were lost. These observations highlight how environmental and extracellular stressors (e.g. NaAsO2) can modulate TDP-43 interactomes and functions, implicating additional pathways in stress-induced pathobiology.

[0228] In addition to identifying the interactome, we aimed to determine whether our approach could capture known interactors and reveal their pathological relationship with TDP-43. Of the 1134 reported proteins, we modeled each pathological factor independently, adjusting using the other two factors as confounders (see Methods), and examined the changes in TDP-43 interactions under mislocalization, impaired RNA binding, and NaAsO2 stress. We further investigated nine individual proteins previously reported as modifiers of TDP-43 pathobiology, including HSPB, ATXN2, DBR1 , POM121 , HSPA5, UPF1 , NUP214, XPO5, and TP53BP1 , and mapped their interactions with TDP-43 under different conditions. Notably, HSPB1 , a stress response protein, showed increased interaction with TDP- 43 under NaAsO2 stress in all three comparisons, consistent with recent findings. We show significantly decreased HSPB1 in 5FL variants, supporting that TDP-43 RNA recognition motifs (RRMs) are necessary for HSPB1-TDP-43's interaction, suggesting the potential involvement of RNA. Furthermore, we performed correlation analysis by calculating the Pearson correlation coefficient (PCC) matrix for the known modifiers and TDP-43 interactors across all nine pathological comparisons, which may identify TDP-43 regulators with similar effects. Together, these results support the robustness of our system in identifying conserved and novel mechanisms in TDP-43 dysfunction.

[0229] Systematic Mapping of TDP-43 Interactions with Biomolecular Condensate

[0230] Emerging evidence underscores the pathological significance of BMCs, membraneless subcellular compartments created through phase separation, in neurodegenerative diseases. As a hallmark in ALS / FTLD and a frequent co-pathology in AD, TDP-43 is a component of various BMCs under physiological and pathological conditions. To systematically characterize these interactions, we performed the global mapping of TDP-43 interactions with BMCs under mislocalization, impaired RNA binding, and NaAsO2-induced stress conditions. Unlike functional enrichment, where only significantly altered proteins are selected, this analysis considers all known components of specific BMCs to acquire an overall landscape, owing to our high- throughput quantitative coverage. These results reveal insights into the context- dependent altered interactions between TDP-43 and 19 previously annotated BMCs, including DNA damage foci, nuclear pore complex (NPC), nuclear speckle, SGs, and P-body.

[0231] TDP-43 displayed opposing interaction patterns with DNA damage foci under different pathological conditions. When mislocalized, TDP-43 greatly lost its interaction with most DNA damage foci proteins, indicating proper nuclear localization is essential for its involvement in DNA damage repair. However, with impaired RNA binding, TDP-43 gained interactions with DNA damage-associated proteins, especially TP53BP1 , consistent with its previously reported RNA-independent interaction with TDP-43. This suggests that TDP-43 is recruited to genomic repair components when RNA binding is disrupted, though the impact on repair pathways is unknown.

[0232] Similarly, TDP-43 significantly gained interactions with NPC components under impaired RNA binding conditions. This supports several lines of emerging evidence of NPC's involvement with TDP-43 pathobiology. Among them, interestingly, we detected a strikingly elevated interaction between POM121 (and its homolog, POM121 C) and TDP-43, which is impaired in ALS / FTLD. This adds to previous findings that RNA-deficient TDP-43 mislocalizes to the cytoplasm via passive diffuse through the NPC117, suggesting a possible crosstalk between POM121 and RNA- deficient TDP-43.

[0233] Nuclear speckles118 also showed altered context-dependent interactions with TDP-43. Mislocalized TDP-43 significantly lost interactions with nuclear speckle proteins, including both SRRM2 and SON, the scaffold components of nuclear speckles. Given the recent discovery of nuclear speckle's participation in regulating RNA splicing and its disruption evidenced in ALS / FTLD, this finding suggests thatproper nuclear localization is critical for TDP-43's association with nuclear speckle and its splicing-related functions.

[0234] In addition to nuclear BMCs, TDP-43 exhibited altered dynamic interactions with cytoplasmic SGs and P-bodies across the pathological conditions tested. This is consistent with our finding in functional enrichment. Under mislocalization conditions, TDP-43 gained robust interactions with SG and P-body proteins, consistent with their roles in cytoplasmic stress responses. However, these interactions were markedly diminished under impaired RNA binding, suggesting that TDP-43 RNA-binding capacity is critical for maintaining these associations. Notably, under NaAsO2 stress, TDP-43 lost interactions with SG and P-body proteins, contradicting prior expectations that stress universally promotes these associations. This pattern is consistent with most SG / P-body proteins, including several of their core components: DCP1A, DDX6, EDC3, G3BP1 / 2, UBAP2L and USP10.

[0235] Cytoplasmic TDP-43 is known to aggregate under stress but is not necessarily driven by SG condensation, which was also revealed by the increased labeling of endogenous TDP-43 under NaAsO2 stress. Our findings strongly suggest that under oxidative stress, cytoplasmic TDP-43 is not localized to SGs but selfaggregates, which is consistent with several recent studies. This differs from previously established paradigms and suggests a working model in which TDP-43 interactions with SGs and P-bodies are facultative and tightly regulated. Excessive or deficient interactions may disrupt cellular homeostasis, underscoring the fine-tuned balance required for TDP-43 functionality.

[0236] TDP-43 exhibited further context-dependent shifts among other important BMCs. To further evaluate the impact of the pathological factors on different BMCs, we ranked their associations with BMCs in all nine pathological comparisons. This provided detailed insights into TDP-43 dynamic interacting patterns among BMCs.

[0237] TDP-43 Interactors Converge with ALS / FTLD Patient-derived snRNA-seq Signatures

[0238] To determine whether the context-dependent TDP-43 interactors identified in our HEK293 cell models might uncover disease-relevant interactors, we performed an integrated analysis with a recently published snRNA-seq dataset55 of ALS / FTLD patients and extracted the data from excitatory neurons of the motor cortex (MCX). In support of our results, among all 6592 proteins detected in our mass spectrum dataset from HEK293 cells, 5443 (82.57%) of these also have their coding genes quantified inthe snRNA-seq data from patient-derived neurons. Correspondingly, differentially expressed (DE) genes from either C9-ALS (2248 genes) or sALS (2222 genes) patients are found in our proteomic data (56.27% and 52.34%). This high-overlapping ratio highlights the potential of HEK293 cells for high-throughput preliminary screenings before neuronal proteomics analysis, likely due to their unique background, which shares certain molecular characteristics with neurons. We also found 18 genome-wide association study (GWAS) annotated genes present in both snRNA-seq and our mass spectrum datasets.

[0239] We next cross-referenced lost or gained TDP-43 interactors from our APEX2 proximity labeling experiments under TDP-43 mislocalization, impaired RNA binding, or NaAsO2-induced stress conditions with genes that were significantly up- or down-regulated in C9-ALS and sALS patient neurons. Notably, we found HSPB1 to be one of the few proteins having gained interaction with TDP-43 under stress while it was significantly down-regulated in both C9-ALS and sALS patient neurons, which is supported by a recent study 39. We also identified some hnRNPs (especially HNRNPC and HNRNPR), SRRM2 (a scaffold protein of nuclear speckle), and NUFIP2 (a core protein of SG) that showed lost or gained interaction with TDP-43 under mislocalization while significantly altered in patients.

[0240] To further dissect these relationships, we performed hierarchical clustering on the combined dataset of TDP-43 proximity-labeled proteins (from nine distinct pathological comparisons) and the differentially expressed genes in ALS / FTLD patient neurons. This analysis yielded 10 primary gene clusters, of which two showed particularly intriguing features: Cluster 5 and Cluster 7. Gene Ontology (GO) enrichment of each cluster corroborated their distinct functional profiles.

[0241] Cluster 5 was enriched with mRNA splicing factors, which exhibited globally lost interaction with TDP-43 under mislocalization yet were up-regulated in C9-ALS and sALS patient neurons (but not in FTLD). This pattern may reflect feedback or compensatory up-regulation of splicing machinery in response to diminished TDP-43 functional engagement. In contrast, Cluster 7 encompassed many proteins implicated in stress granules (SGs), P-bodies, and nonsense-mediated decay (NMD). These factors displayed a globally gained interaction with mislocalized TDP-43 but lost interactions under impaired RNA binding and NaAsO2stress, consistent with our previous analysis. Notably, Cluster 7 genes showed a distinctive transcriptional response in patients — being generally down-regulated in C9-ALS and sALS, yetparadoxically up-regulated in C9-sFTLD. Such divergent expression patterns might reflect differential protective or pathogenic roles of RNA granule dynamics in TDP-43 pathology across ALS versus FTLD-TDP subtypes.

[0242] To expand our observations, we examined the broader concordance between all detected proteins in this study and differentially expressed genes in ALS or FTLD patient neurons, revealing that many of these context-dependent TDP-43 interactors exhibit robust, shared signatures of dysregulation in ALS / FTLD, underscoring their potential roles in disease-related mechanisms. Furthermore, we examined the behavior of known ALS / FTLD GWAS-associated genes across our TDP-43 proximity-labeled and snRNA-seq datasets for a "GTP" (Genomics- Transcriptomics-Proteomics) triple-omics analysis. Six out of 18 GWAS-reported ALS / FTLD genes (RPSA, KIF5A, TNIP1 , TBK1 , SPATA2, and ATRN) exhibited significant alterations in TDP-43 interactions across all three pathological factors and showed significant changes in at least one patient-derived snRNA-seq dataset (C9 or sporadic ALS / FTLD).

[0243] Together, these multiomics analyses demonstrate that TDP-43 pathological states in our APEX2 proximity labeling approach capture key molecular alterations also present in ALS / FTLD patient neurons. By pinpointing gene clusters with distinctive TDP-43 binding and transcriptom ic patterns — such as the splicing- factor-enriched Cluster 5 and the cyto-RNA-granule-associated Cluster 7 — our results suggest potential novel mechanistic avenues involving both upstream triggers (TDP-43 mislocalization or RNA-binding defects) and downstream compensatory or pathogenic responses (aberrant RNA splicing, granule assembly, or decay processes). This integrated view highlights how context-dependent TDP-43-protein interactions and disease-specific gene expression changes may converge to drive the pathogenesis of ALS / FTLD and related TDP-43-associated disorders.

[0244] A Functional Splicing Screen Reveals Bidirectional Modulation of TDP-43 Function by the Disease-Relevant Interactome

[0245] Building on the multiomics integration previously described, we next sought to determine whether specific context-dependent TDP-43-interacting proteins could influence the TDP-43 splicing function. From the 10 identified clusters, we selected 53 proteins enriched in mRNA splicing factors (primarily from Clusters 1 and 5), SG / P- body-related components (primarily from Cluster 7), additional key nuclear- cytoplasmic transport factors, and reported TDP-43 modifiers.

[0246] We employed our CUTS biosensor to functionally interrogate how each of these proteins impacts TDP-43's splicing role54 and performed a preliminary semi- quantitative screen by knocking down each of the 53 selected proteins using siRNA. Following siRNA-mediated knockdown, CUTS-HEK293 cells enable the measurement of GFP intensity as a proportional indicator of TDP-43 splicing LOF, relative to a control siRNA (siControl) (schematic in Figure 5A). The resulting heatmap (not shown) provides a ranked summary of the number of replicates in which each knockdown either increased (+1 ) or decreased (-1 ) in detectable GFP fluorescence compared to the siControl. This classification enabled rapid identification of putative enhancers (those that increased GFP levels) and potential suppressors of TDP-43 LOF (those that reduced GFP levels).

[0247] We validated the top enhancers by quantitative live-cell imaging of CUTS- HEK293 cells under single knockdown conditions ("CUTS Normal + KD"; schematic in Figure 7A). The knockdown of seven proteins — CDC40, DBR1 , HNRNPA3, HNRNPL136, NUFIP2, PRPF31 , and SRRM2 — caused modest but significant GFP upregulation, indicating that loss of these factors can simulate TDP-43 LOF (Figures 7B-7C).

[0248] To amplify and clarify their effects, we then tested a subset of hits, including the seven LOF triggers and top rescuers in a partial TDP-43 LOF background by cosilencing TDP-43 together with each candidate ("CUTS siTDP-43 + KD"; schematic in Figure 7D). While CDC40, HNRNPA3, and PRPF31 knockdown exacerbated the GFP increase with TDP-43 siRNA, indicating a further impaired TDP-43's splicing function, HNRNPA0, HNRNPC, and MATR3 knockdown partially rescued it, as shown by livecell confocal imaging and immunoblotting (Figures 7E-7H). Notably, HNRNPC, a hnRNP from Cluster 5 with significant up-regulation in ALS patients, displayed the most robust GFP reduction (~90%), underscoring its potential role as a buffering or compensatory factor against TDP-43 LOF.

[0249] In parallel, we assessed whether the overexpression (OE) of selected interactors could also modulate TDP-43 splicing function in the standard CUTS background ("CUTS Normal + OE"; schematic in Figure 7I). Surprisingly, among five candidates tested (DBR1 , HNRNPA0, HNRNPC, HNRNPL, and NUFIP2), NUFIP2 OE increased GFP signals, indicative of a LOF-like phenotype (Figures 7J-7L). Conversely, under partial TDP-43 knockdown ("CUTS siTDP-43 + OE"; schematic in Figure 7M), overexpressed HNRNPL mildly attenuated GFP signal (Figures 7N-7O).This observation aligns with a recent report136 that HNRNPL can partially rescue cryptic exon inclusion in UNC13A under TDP-43 LOF, since the CUTS biosensor partly contains the UNC13A-CE.

[0250] Altogether, our CUTS-based screening assays revealed both positive and negative regulators of TDP-43 splicing function among the context-dependent interactors identified by our proximity labeling, functional analysis, BMC mapping, and patient-derived snRNA-seq integration. The ability of specific knockdowns or overexpression events to drive or counteract TDP-43 LOF strongly supports our central hypothesis that distinct TDP-43-binding partners, modulated by pathological cues, critically impact TDP-43's role in mRNA splicing. Such context-dependent interactions may thus serve as potential mechanistic or therapeutic (Figure 8) targets in ALS / FTLD and related disorders associated with TDP-43.

[0251] SRRM2 and Nuclear Speckles Modulate Nuclear TDP-43 Retention and Splicing Function.

[0252] Building on our previous multiomics and CUTS screening analyses, we next investigated the mechanism by which SRRM2, a top hit that induces robust TDP- 43 loss of function (Figures 7B-7C), contributes to TDP-43 pathology. SRRM2 is a key scaffold protein of nuclear speckles alongside SON, and BMC mapping indicated TDP-43 interactions with nuclear speckle components are severely diminished under TDP-43 mislocalization. Compared to SRRM2, SON KD alone did not trigger a significant LOF in the CUTS biosensor, but co-knockdown of SRRM2 and SON, with the same total siRNA concentration, significantly exacerbated the GFP increase beyond that observed with SRRM2 KD alone, indicating a synergistic effect.

[0253] To visualize TDP-43 localization changes, we generated a stable HEK293 cell line via Piggybac with doxycycline-inducible expression of EGFP-TDP-43WT and treated it with either siRNA-mediated KD of SRRM2, SON, or a combination. In control and SON KD cells, EGFP-TDP-43WT primarily remained diffuse within the nucleus. However, SRRM2 KD or double KD (SRRM2 + SON) induced a distinct population of "abnormal" cells (~2% and ~5% of all the live cells, respectively) marked by nuclear TDP-43 foci and cytoplasmic mislocalization. Immunostaining for nuclear speckles with SC-35 (SRRM2) revealed that nearly all cells harboring nuclear TDP-43 inclusions similarly exhibited disrupted nuclear speckles (>96%), underscoring the link between nuclear speckle integrity and proper TDP-43 localization.

[0254] Fluorescence recovery after photobleaching (FRAP) was performed to assess whether TDP-43 nuclear inclusions in the abnormal cells exhibited altered biophysical properties. Compared with diffuse nuclear TDP-43, TDP-43 inclusions within SRRM2 KD or double KD cells showed substantially reduced dynamics, suggesting a more solid TDP-43 structure that is consistent with pathological aggregation.

[0255] Together, these findings suggest that SRRM2-containing nuclear speckles function as a crucial mechanism for retaining nuclear TDP-43. Dysregulated SRRM2, possibly due to reduced expression in patient-derived C9-ALS neurons, and disrupts nuclear speckles, can lead to TDP-43 mislocalization and aggregation, and ultimately resulting in TDP-43 loss of function (LOF). This mechanism highlights nuclear speckles as a vital regulatory structure that safeguards TDP-43 splicing function by maintaining its proper nuclear distribution and mobility.

[0256] NUFIP2 Promotes TDP-43 Mislocalization and Loss of Function

[0257] NUFIP2 overexpression was identified to regulate TDP-43 splicing function (Figures 7J-7L). NUFIP2 is a member of FMR1 -related proteins (others include FMR1 , FXR1 , and FXR2) but has not been reported to be linked to TDP-43. We showed that among these four FMR1 -related proteins, only NUFIP2 OE was sufficient to induce a robust increase in GFP signal, signifying TDP-43 LOF. This underscores the unique role of NUFIP2 in triggering TDP-43 dysfunction, aligning with its identification as a top promoter of LOF in our systematic screening.

[0258] Notably, NUFIP2-driven TDP-43 LOF is dose-dependent. Stable expression of NUFIP2 under the EF1A promoter increased NUFIP2 protein levels marginally above endogenous amounts, whereas transient transfection of NUFIP2 cDNA under a CMV promoter produced a higher expression. Removing the 3'UTR (A3'UTR) from the cDNA construct further elevated NUFIP2 beyond the full-length cDNA control. Consistent with expression levels, only CMV-NUFIP2(cDNA) and CMV- NUFIP2(A3'UTR) triggered increased GFP fluorescence in the CUTS reporter, with the A3'UTR construct causing a markedly more substantial LOF effect. These findings highlight a direct correlation between NUFIP2 concentration and its capacity to impair TDP-43 function.

[0259] The proximity labeling data highlights a significantly increased interaction between cytoplasmic TDP-43 and NUFIP2. We next examined whether this LOF phenotype was accompanied by TDP-43 mislocalization. Interestingly, cellsoverexpressing NUFIP2 displayed a cytoplasmic "shell-like" (ring-shaped on the cross-section) distribution for NUFIP2, within which endogenous TDP-43 localizes. The appearance of "shell-like" NUFIP2 condensates and TDP-43's mislocalization in individual cells corresponded with CUTS-GFP activation, underscoring a mechanistic link between the cytoplasmic sequestration of TDP-43 and its nuclear LOF. These "shell-like" structures broaden our understanding of TDP-43 aggregation patterns and suggest that NUFIP2 actively promotes TDP-43 to accumulate in the cytoplasmic compartments and may facilitate aberrant phase transitions.

[0260] The interactome data demonstrated decreased interactions between NUFIP2 and RNA-deficient TDP-43. To determine how NUFIP2-driven condensates influence TDP-43 under different RNA-binding statuses, we employed cell lines expressing cytoplasmic EGFP-TDP-43 with or without intact RNA-binding motifs (TDP-43ANLS vs. TDP-435FL / ANLS). In cells lacking NUFIP2 overexpression, EGFP-TDP-43ANLS and EGFP-TDP-435FL / ANLS appeared predominantly diffuse or occasionally formed small puncta. Following NUFIP2 overexpression, EGFP-TDP- 43ANLS formed the "shell-like" structures, whereas the EGFP-TDP-435FL / ANLS aggregated into "condensed" inclusions. FRAP analysis confirmed that the NUFIP2- induced "shell-like" TDP-43 structures had intermediate dynamics, whereas the "condensed" structures by RNA-deficient TDP-43 displayed minimal signal recovery, indicative of low molecular mobility.

[0261] To quantify the stability of these condensates, we examined RIPA-soluble and insoluble fractions of EGFP-TDP-43. NUFIP2 overexpression significantly raised the insoluble fraction of both TDP-43ANLS (~4.3-fold increase) and TDP-435FL / ANLS (~10.7-fold increase). The "condensed" structures formed by the 5FL / ANLS mutant showed particularly lower FRAP recovery and higher insoluble transitions, mirroring their more solid-state molecular architecture. Thus, although NUFIP2 induces broad cytoplasmic mislocalization of TDP-43 and the exact phase properties and biophysical outcomes depend strongly on TDP-43's RNA-binding status.

[0262] Collectively, our results support a working model in which NUFIP2, when elevated, sequesters TDP-43 in cytoplasmic "shell-like" compartments, thereby hindering TDP-43 nuclear localization and splicing function. If TDP-43 further loses the capacity to bind RNA, it transitions into "condensed" and low-dynamics structures. Notably, NUFIP2 was significantly up-regulated in C9-ALS and sFTLD patient tissues.These findings highlight NUFIP2's potency in reshaping TDP-43 phase separation and emphasize its underlying connection with TDP-43 pathology in disease.

[0263] Cis- and Trans-Rescue Strategies for TDP-43 Loss of Function: Targeted TDP-43 Replacement and HNRNPC Knockdown

[0264] In the light of our previous identification of functional TDP-43 interactors, we next sought to determine whether targeting TDP-43 itself or its regulatory partners could restore TDP-43’s physiological splicing activity. We tested two distinct strategies to rescue TDP-43 splicing function (Figure 9A). The first involved a cis-targeting approach, in which endogenous TDP-43 is selectively knocked down while simultaneously replaced by a codon-optimized TDP-43 expressed under carefully controlled conditions (“TDP-43 knockdown-replacement,” TKR). This design addresses the detrimental effects of TDP-43 overexpression by controlling exogenous TDP-43 levels to not exceed the wild-type levels, thus avoiding excessive TDP-43 accumulation, as recently demonstrated. The second, a trans-regulatory approach, leverages our functional screening findings by modulating TDP-43’s protein partners activated by CUTS (only under TDP-43 LOF). We prioritized CUTS-controlled HNRNPC knockdown (CCK) for proof-of-concept based on its robust capacity to rescue TDP-43 LOF events in our CUTS assays (Figure 7).

[0265] For TKR, we generated stable HEK293 cell lines harboring an shRNA targeting TARDBP mRNA and a doxycycline-inducible, codon-optimized EGFP-TDP- 43 cassette (Figure 9B). The inducible expression cassette was further tuned with our CUTS biosensor to mirror the autoregulatory properties of wild-type TDP-43 to avoid pathological overexpression (Figures 9C-9D). We confirmed high-level silencing (~85%) of endogenous TDP-43 and partial but significant recovery (up to 70% total TDP-43 in the best case) from the inducible EGFP-TDP-43 (Figure 9E), achieving our goal of mimicking physiological TDP-43 levels.

[0266] In parallel, for the CCK strategy, we engineered stable HEK293 cells with an EGFP-NLS-shHNRNPC cassette, whose expression is activated by CUTS only upon TDP-43 LOF (Figure 9F). Under siRNA-induced TDP-43 depletion, robust HNRNPC knockdown was achieved, while baseline HNRNPC expression remained largely unaffected (Figures 9G-9H). Notably, CCK cells also partially restored the splicing profiles of several endogenous TDP-43 targets, including CE insertions in ATG4B, DNAJC5137, and HDGFL221 (Figure 9I). This data provided initial evidencethat the trans-regulation of certain TDP-43 interactors — here, lowering HNRNPC — can buffer or compensate for TDP-43 deficiency.

[0267] To more comprehensively evaluate these two rescue approaches, we performed mRNA sequencing on the TKR and CCK stable lines, each with appropriate baseline and positive (TDP-43 LOF) controls. The TKR system was modified to have doxycycline-inducible expression of shTARDBP (Figures 8A-8B). We defined a “rescue ratio” for each splicing junction — calculated as the fraction of LOF-induced splicing dysregulation in the positive controls that were reversed by either TKR or CCK (Figure 8C). We then curated a panel of 198 TDP-43-responsive splicing junctions based on intersections with previously published TDP-43 knockdown data in iPSC- derived neurons20 (Figure 8D). Across these events, TKR and CCK exhibited disparate rescue profiles, reflected by minimal correlation (R = -0.018, P = 0.81 ), suggesting distinct mechanisms (Figure 8E). Notably, CCK rescued 154 (77.78%) events, showing a broad-acting effect against TDP-43 LOF.

[0268] We further focused on the 28 CEs from these events as a hallmark downstream readout of TDP-43 LOF pathology (Figure 8F, cryptic_exon=TRUE). Ranking the CE splicing rescue by TKR (left panels) revealed diverse levels of responsiveness to exogenous TDP-43, wherein events such as HDGFL2-CE were nearly fully corrected (>80% rescue), but others (e.g. EPB41 L4A-CE) were far less responsive. Intriguingly , CCK (right panels) also restored most of these CEs — 24 out of 28 (85.71 %) — to some degree, with some matching or even surpassing TKR’s efficacy. For instance, HDGFL2-CE was strongly rescued (Figures 8G, 8I), ACBD3- CE was mildly rescued (Figure 8G) by both TKR and CCK, whereas EPB41 L4A-CE was poorly rescued by TKR but substantially corrected by CCK (Figure 8G). Although TKR generally could not rescue the neuron-specific STMN2-CE in our HEK293 models (likely due to low STMN2 expression), a partial rescue was detected with CCK, aligning with the knockdown of HNRNPC as a potentially broad-acting strategy (Figure 8H). Additionally, we found that STMN2 mRNA was significantly down- regulated under TDP-43 knockdown in CCK2 but was partially rescued (P = 0.05) in CCK4, which is consistent with the splicing outcome. Interestingly, we also detected a dramatic increase of NPTX2, a recently reported TDP-43-regulated gene highly overexpressed in both LOF cellular models and patients. However, this up-regulation was not rescued by HNRNPC knockdown, potentially due to a different regulatory mechanism (inhibition by TDP-43 binding to the 3’UTR) rather than splicing.

[0269] Taken together, these results reveal that (i) TKR partially restores TDP-43 function by directly replenishing TDP-43 under autoregulated control, thereby highlighting event-specific sensitivities to TDP-43 dosage, and (ii) CCK mitigates TDP- 43 LOF in splicing via a mechanistically distinct route. Based on these results as well as previous research showing HNRNPC and TDP-43 have many shared binding sites, we further propose that HNRNPC may act as a competitive or noncompetitive antagonist to TDP-43 at certain splicing sites (Figure 8J). Reducing HNRNPC can alleviate cryptic exon retention and other LOF-driven splicing aberrations — even when TDP-43 remains partially depleted — thus offering a “trans-regulatory” axis of intervention.

[0270] NUFIP2 and HNRNPC show Distinct TDP-43 Colocalization Patterns in Patient Tissue

[0271] SRRM2 pathology and nuclear speckles disruption were recently reported in C9-ALS patients, with post-mortem staining having a >60% co-occurrence with TDP-43 mislocalization in the soma of neurons. To evaluate the disease relevance of our findings on NUFIP2 and HNRNPC, we examined their expression and subcellular localization in a small cohort of post-mortem tissues from patients with TDP-43 proteinopathy. Both proteins were broadly distributed in the frontal cortex of sporadic FTLD-TDP cases, consistent with our prior cell-based and multiomics data showing their ubiquitous presence.

[0272] Strikingly, HNRNPC remained confined to the nucleus even in neurons exhibiting cytoplasmic phosphorylated TDP-43 (p-TDP-43) inclusions. This pattern was observed across distinct disease contexts, including the temporal cortex from a C9-FTLD patient, the motor cortex from an sALS patient, and the amygdala from a LATE patient. These findings confirm our APEX2 interactome observations — namely that HNRNPC loses its binding to mislocalized TDP-43 in the cytoplasm but is not redistributed. Moreover, they support the notion that HNRNPC could be leveraged as a “trans-regulatory” target to ameliorate TDP-43 LOF phenotypes without altering the total TDP-43 protein levels.

[0273] In contrast, NUFIP2 colocalized with some cytoplasmic p-TDP-43 inclusions in C9-ALS motor cortex, suggesting a NUFIP2-TDP-43 association in disease. This colocalization was not consistently observed in other TDP-43 proteinopathies tested, including sALS motor cortex, sFTLD hippocampus, or LATE amygdala. The preferential accumulation of NUFIP2 in C9-ALS aligns with snRNA-seq data showing elevated NUFIP2 mRNA levels most prominently in C9-ALS patient neurons.Discussion

[0274] TDP-43 pathology is a hallmark of ALS / FTLD, playing a central role in neurodegeneration through its cytoplasmic mislocalization (GOF) and nuclear loss of physiological functions (LOF)10. Understanding the TDP-43 interactome during these events is essential for deciphering the molecular mechanisms driving pathogenesis and identifying potential therapeutic targets. By employing APEX2-based proximity proteomics in HEK293 cells, our study provides a comprehensive landscape of the TDP-43 interactome across physiological and pathological contexts, revealing key molecular programs associated with TDP-43 mislocalization, impaired RNA-binding, and oxidative stress. We mapped the dynamic changes in TDP-43 interactions and found strong concordance with various recent findings from cellular models, iPSCs, and postmortem analyses of ALS / FTLD patients.

[0275] A major aspect of our findings is the context-dependent remodeling of TDP- 43 interactions with various biomolecular condensates (BMCs), or membraneless organelles (MLOs), which are often formed through liquid-liquid phase separation (LLPS) and are essential for cellular organization homeostasis. A recent study leveraging microscopic images of human neurons analyzed 24 organelles, including paraspeckles, nuclear speckles, stress granules (SGs), and P-bodies, and revealed their pathobiological relevance to TDP-43 pathology142, which is also identified in our study. Both studies reinforce that aberrant phase separation and organelle dysfunction represent fundamental mechanisms underlying TDP-43-associated neurodegeneration. Notably, consistent with recent works, we also found a reduction in TDP-43: SG interactions under oxidative stress and aggregate-prone conditions (RNA deficient TDP-43).

[0276] Beyond mapping interactome shifts, we functionally investigated the consequences of TDP-43 LOF by selectively perturbing key interactomes to assess their contribution to TDP-43 cellular roles via the CUTS splicing sensor. This approach led to the identification of proteins that play essential roles in TDP-43 LOF phenotypes, shedding light on potential therapeutic targets for TDP-43 proteinopathies. SRRM2, a key nuclear speckle scaffold, is essential for RNA processing. Recent studies have linked SRRM2 dysfunction to neurodegenerative diseases, showing that in C9-ALS, SRRM2 is sequestered into poly-GR cytoplasmic inclusions, leading to nuclearspeckle disruption and widespread RNA splicing defects. Our findings further demonstrate that SRRM2 or SRRM2 / S0N downregulation induces TDP-43 inclusions, mislocalization, and LOF, underscoring the interplay between TDP-43 and nuclear speckles. Notably, SRRM2 is downregulated in C9-ALS patients’ excitatory neurons in the motor cortex (MCX) and loses its interaction with TDP-43 under mislocalization, suggesting that SRRM2 acts as a critical nuclear “anchor” for functional TDP-43. This supports the idea that SRRM2 disruption is upstream of TDP-43 pathology, aligning with evidence that poly-GR triggers TDP-43 dysfunction in C9-ALS. Beyond ALS / FTLD, nuclear speckle dysfunction may contribute to other neurodegenerative diseases. Tau pathology was previously reported to be a driver of nuclear speckle component mislocalization, suggesting that SRRM2 dysregulation could also contribute to Alzheimer’s disease (AD) and related disorders. This broader impact of nuclear speckle dysfunction raises the intriguing possibility of shared pathogenic mechanisms, especially TDP-43 dysfunction, across neurodegenerative diseases.

[0277] Our study identifies NUFIP2 overexpression as a potent trigger of TDP-43 mislocalization in cell models, with colocalization between NUFIP2 and TDP-43 in 09- ALS patient MCX tissues exhibiting TDP-43 mislocalization. These findings suggest that NUFIP2 may sequester TDP-43, driving its cytoplasmic mislocalization and subsequent LOF. Notably, NUFIP2 RNA is elevated in patient tissue, and like other stress granule components, it shows increased TDP-43 interaction when TDP-43 is mislocalized, suggesting that NUFIP2 may act as a potential “driver” of TDP-43 pathology. Intriguingly, NUFIP2 knockdown in the CUTS biosensor system also triggers TDP-43 LOF (Figures 7B-7C), indicating the possibility of a dual effect on TDP-43 function, which raises the need for future investigation. In addition, NUFIP2’s role in post-transcriptional regulation of mRNA through the NMD pathway adds another layer to its involvement in cellular stress responses. Previous findings suggest that NUFIP2 is a cofactor for Roquin-mediated mRNA decayl 50, linking its role in RNA quality control to TDP-43 pathology. This pathway has recently been highlighted for its role in ALS / FTLD. Further exploration of NUFIP2 neuropathology across disease subtypes, and its interaction with components of the NMD machinery could reveal insights into TDP-43 mislocalization and dysfunction, which may advance our understanding of the ALS / FTLD pathogenesis.

[0278] HNRNPC is upregulated in sALS and C9-ALS patients’ MCX excitatory neurons alongside other hnRNPs and loses its interaction with TDP-43 when TDP-43is mislocalized, coinciding with impaired RNA binding of TDP-43. We thus propose HNRNPC as a potential “modulator” of TDP-43 splicing function. Using the CUTS sensor assay, we demonstrate that HNRNPC knockdown consistently rescues TDP- 43 LOF phenotypes. Furthermore, RNAseq and splicing event analyses indicate the rescue of TDP-43 cryptic exon inclusion upon HNRNPC KD. Notably, these rescued cryptic exon events were distinct from those observed in TDP-43 replacement experiments, suggesting that HNRNPC mediates TDP-43 LOF rescue through different but essential mechanisms. Previous studies on the hnRNP family have linked their various functions in neurodegenerative diseases, with the knockdown of several hnRNP orthologs, including HNRNPC, alleviating TDP-43 (TBPH) toxicity in Drosophila models, which aligns with our findings. However, to understand the specific impact of HNRNPC on splicing regulation, future studies should map HNRNPC and TDP-43 RNA binding sites under normal conditions or knocking down both proteins, as well as characterize HNRNPC-driven splicing events independent of TDP-43. This could provide insights into how alternative RNA processing pathways emerge in the context of TDP-43 pathology.

[0279] Recent studies identified HSPB1 as a molecular chaperone of TDP-43, with its downregulation leading to TDP-43 inclusions and mislocalization — suggesting a broader role for heat shock proteins in maintaining TDP-43 homeostasis. This matches with our multi-omics analysis that HSPB1 is downregulated in ALS / FTLD patient samples while gaining interaction with mislocalized TDP-43 under oxidative stress. Thus, by integrating the key examples of SRRM2, NUFIP2, HNRNPC, and HSPB1 , our study uncovers a complex interacting landscape governing TDP-43’s function. With robust evidence from proteomic shifts, patient-derived datasets, and functional screening, we present a comprehensive framework for understanding how diverse cellular machineries converge on TDP-43 pathology. This regulatory landscape deepens our insight into ALS / FTLD pathogenesis and opens new avenues for therapeutic intervention, targeting these distinct pathways to restore TDP-43 homeostasis.

[0280] Moreover, emerging evidence indicates that certain TDP-43 protein interactors, such as G3BP1 in SGs, are likewise regulated by TDP-43 at the mRNA level. This feedback mechanism, reminiscent of TDP-43’s own autoregulation, underline TDP-43’s multifaceted regulatory network spanning both RNA and protein interactions. In our datasets, we identified a subset of genes whose splicing isregulated by TDP-43 and that have the potential to interact with TDP-43 at the protein level in a context-dependent manner, including BBC3 (encoding the PUMA protein), PFKP, LENG8, and ACBD31 , all serving critical cellular functions related to neurodegenerative disorders. Further investigations to dissect these intricate interactions will likely yield deeper insights into TDP-43-mediated pathobiology.

[0281] The development for ALS / FTLD therapies remains challenging due to the lack of clear genetic markers and late-stage clinical presentation, highlighting the need for targeted therapeutic strategies. This study identifies key modulators of TDP-43 pathology and demonstrates two LOF rescue strategies: (i) finely tuning TDP-43 interactors, such as CUTS-controlled HNRNPC knockdown (CCK), and (ii) knocking down endogenous mutant TDP-43 while introducing a biosensor-controlled exogenous TDP-43 (TKR). Both approaches restored TDP-43-specific splicing events through distinct mechanisms. Our datasets, together with our findings, provide a valuable resource as well as a framework for identifying TDP-43 modulators, and lay the foundation for translational applications.

[0282] The present invention has been described with reference to certain exemplary embodiments, dispersible compositions and uses thereof. However, it will be recognized by those of ordinary skill in the art that various substitutions, modifications or combinations of any of the exemplary embodiments may be made without departing from the spirit and scope of the invention. Thus, the invention is not limited by the description of the exemplary embodiments.

Claims

THE INVENTION CLAIMED IS1 . A composition comprising: a first nucleic acid; an excisable second nucleic acid; a third nucleic acid; and one or more cleavable fourth nucleic acids encoding a self-cleavable peptide.

2. The composition of claim 1 , wherein the first nucleic acid encodes a first reporter peptide.

3. The composition of claim 1 , wherein the excisable second nucleic acid comprises a cryptic exon.

4. The composition of claim 3, wherein the cryptic exon comprises an exon from a cystic fibrosis transmembrane conductance regulator (CFTR) gene.

5. The composition of claim 3, wherein the cryptic exon comprises an intron from an llnc-13 homolog A (UNC13A) gene.

6. The composition of claim 3, wherein the cryptic exon comprises exon 9 from CFTR and intron 20 from UNC13A.

7. The composition of claim 1 , wherein the second nucleic acid comprises one or more binding sites for transactive response DNA-binding protein 43 (TDP-43).

8. The composition of claim 7, wherein the one or more binding sites for TDP-43 comprise one or more UG-rich regions.

9. The composition of claim 1 , wherein the nucleic acids are ribonucleic acids.

10. The composition of claim 1 , wherein the first and third nucleic acids encode a fluorescent protein.

11. The composition of claim 10, wherein the first nucleic acid encodes mCherry and the third nucleic acid encodes EGFP.

12. The composition of claim 1 , wherein the third nucleic acid encodes a therapeutic peptide.

13. The composition of claim 12, wherein the therapeutic peptide is a peptide useful for treating a neurodegenerative disease.

14. The composition of claim 13, wherein the neurodegenerative disease is Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Degeneration (FTD), and / or Alzheimer’s Disease (AD).

15. The composition of claim 1 , wherein the fourth nucleic acid comprises a T2A self-cleavage sequence.

16. The composition of claim 1 , wherein the nucleic acid comprises, from a 5’ end to a 3’ end: the first nucleic acid; the fourth nucleic acid; the excisable second nucleic acid; a stop codon, wherein the stop codon is out of frame when the excisable second nucleic acid is present in the composition and the stop codon is in frame when the excisable second nucleic acid is excised from the composition; the fourth nucleic acid; and the third nucleic acid, the third nucleic acid encoding a reporter peptide or a therapeutic peptide; wherein, when introduced into a cell, the cell translates or expresses the third nucleic acid, thereby producing a reporter or therapeutic peptide, when the excisable second nucleic acid is excised.

17. The composition of claim 1 , wherein the composition comprises a recombinant transducing particles or a viral expression vector comprising the first, second, third, and fourth nucleic acids.

18. The composition of claim 17, wherein the viral expression vector is a lentiviral, an adenoviral, an adeno-associated virus (AAV), or a herpesvirus recombinant genome, or a plasmid DNA vector.

19. A composition comprising the recombinant transducing particles or viral vector of claim 17 in a pharmaceutical ly-acceptable carrier.

20. A method of detecting aberrant expression or dysfunction of a protein in a cell, comprising: delivering to the cell the composition of claim 1 , wherein the third nucleic acid is translated or expressed when the protein is aberrantly expressed or is dysfunctional and the third nucleic acid is not expressed when the protein is not aberrantly expressed or is not dysfunctional.21 . The method of claim 20, wherein translation or expression of the third nucleic acid correlates to the level of aberrance of expression of dysfunction of the protein.

22. The method of claim 20, wherein the protein is TDP-43.

23. A method of treating a patient, comprising introducing the composition of claim 1 into a cell of the patient, wherein the third nucleic acid encodes a therapeutic peptide, such that the third nucleic acid is expressed in the cell of the patient.

24. The method of claim 23, wherein the patient has a neurodegenerative disease.

25. The method of claim 24, wherein the neurodegenerative disease is Amyotrophic Lateral Sclerosis (ALS), Frontotemporal Degeneration (FTD), or Alzheimer’s Disease (AD).

26. A nucleic acid comprising, from a 5’ to a 3’ direction, SEQ ID NO: 1 , SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, and SEQ ID NO: 7.

27. The nucleic acid of claim 26, further comprising, at the 3’ end, SEQ ID NO: 9.

28. The nucleic acid of claim 26, further comprising an additional nucleic acid encoding a reporter peptide or a therapeutic peptide, the additional nucleic acid arranged at a 3’ end of SEQ ID NO: 7.

29. The nucleic acid of claim 28, wherein the nucleic acid encoding the reporter peptide has the sequence of SEQ ID NO: 8.

30. A nucleic acid comprising, from a 5’ to a 3’ direction, SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 16, and SEQ ID NO: 7.

31. The nucleic acid of claim 30, further comprising, at a 3’ end of SEQ ID NO: 7, SEQ ID NO: 8.

32. The nucleic acid of claim 30, further comprising, at the 3’ end, SEQ ID NO: 9.

33. A nucleic acid comprising, from a 5’ direction to a 3’ direction, SEQ ID NO: 10, SEQ ID NO: 11 , SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, and SEQ ID NO: 7.

34. A method of detecting aberrant expression or dysfunction of TDP- 43 in a cell of a patient, comprising introducing the nucleic acid of claim 26 or claim 30 into a cell of the patient.

35. The method of claim 34, wherein the cell is obtained from a patient sample.

36. The method of claim 35, wherein the patient sample is a blood sample.

Images