Loss of TDP-43 induces synaptic dysfunction that is rescued by UNC13A splice-switching ASOs

Abstract

TDP-43 loss of function induces multiple splicing changes, including a cryptic exon in the amyotrophic lateral sclerosis and fronto-temporal lobar degeneration risk gene UNC13A, leading to nonsense-mediated decay of UNC13A transcripts and loss of protein. UNC13A is an active zone protein with an integral role in coordinating pre-synaptic function. Here, we show TDP-43 depletion induces a severe reduction in synaptic transmission, leading to an asynchronous pattern of network activity. We demonstrate that these deficits are largely driven by a single cryptic exon in UNC13A. Antisense oligonucleotides targeting the UNC13A cryptic exon robustly rescue UNC13A protein levels and restore normal synaptic function, providing a potential new therapeutic approach for ALS and other TDP-43-related disorders.

Fig. 1 |. UNC13A and TDP-43 knockdown impair synaptic function.

a, Schematic representation of HaloTag edit at the beginning of exon 2 of TARDBP and PCR from genomic DNA indicating successful edit b, Immunofluorescence analysis of TDP-43 localization in HaloTDP iNeurons revealed nuclear localised TDP-43. Scale bar = 10 μm. c, RT-PCR analysis of UNC13A splicing after TDP-43 knockdown. d, Quantification of results in (c) n=12 biological replicates from 3 experiments. e, Western blot analysis of HaloTDP iNeurons treated with 30 nM HaloPROTAC-E for the indicated amount of time. f, Quantification of western blot in (e) n=2 biological replicates. g, Immunofluorescence labelling of synapsin and UNC13A presynaptic terminals in 4-week-old HaloTDP iNeurons grown on rat astrocytes following 2-weeks of TDP-43 depletion. Scale bar = 10 μm. h, Quantification of UNC13A intensity at synapsin positive synaptic terminals. Control n=116, TDP-43 knockdown n=115 synapses from 3 experiments. i, Miniature excitatory postsynaptic currents (mEPSCs) from control (n=12) and UNC13A knockdown (n=13) iNeurons pooled from 3 experiments. Representative traces and average traces shown. Quantification of mEPSC frequency and inter-event interval cumulative distributions. j, mEPSCs from control (n=20) and TDP-43 knockdown (n=22) HaloTDP iNeurons pooled from 3 experiments. Representative traces and average traces shown. Quantification of mEPSC frequency and inter-event interval cumulative distributions. k, Representative raster plots from D36 multi-electrode array recordings from control, UNC13A and TDP-43 knockdown conditions (30s window shown). Electrode bursts in blue and network bursts in pink. l, Time-course of network synchrony measured by the area under the normalised electrode cross correlation in control and UNC13A knockdown iNeurons. UNC13A normalised to control n=18 wells from 3 experiments. m, Time-course of network synchrony in control and TDP-43 knockdown iNeurons. TDP-43 normalised to control, n=18 wells from 3 experiments. Graphs for (d) (h) (i) and (j) represent mean ± s.e.m and independent experiments are denoted by circle, square or triangle markers. Statistics for (d) (h) (i) and (j) are two-sided Student’s t test. Statistics for L and M are ratio paired t tests. *P < 0.05; ** P <0.01; *** P <0.001; **** P < 0.0001; ns (not significant).

Fig. 2 |. Genomic deletion of UNC13A CE rescues protein levels following TDP-43 knockdown.

a, Schematic representation of genomic edit of UNC13A cryptic exon. b-f, Analysis of control and TDP-43 KD in WT and CE Del SH-SY5Y cells. (b) RT-PCR analysis of UNC13A splicing indicates CE deletion rescues splicing at the UNC13A exon 20–21 junction after TDP-43 KD. (c) Quantification of results in (B) n=8 biological replicates from 3 experiments. (d) RT-qPCR analysis of TDP-43, correctly spliced UNC13A shows CE deletion rescues UNC13A RNA levels after TDP-43 KD. n = 8 biological replicates from 3 experiments. (e) Western blot analysis of protein lysates shows CE deletion rescues UNC13A protein levels after TDP-43 KD. (f) Quantification of results in (e) n=5 biological replicates from 3 experiments. (g-k) Analysis of control and TDP-43 knockdown in WT and CE Del Halo-TDP iNeurons on DIV 28 following 14 days of HaloProtac treatment. (g) RT-PCR analysis of UNC13A splicing indicates CE deletion rescues splicing at UNC13A exon 20–21 junction after TDP-43 KD. (h) Quantification of results in (g) n=5 biological replicates from 2 experiments. (i) RT-qPCR analysis of correctly spliced UNC13A shows CE deletion rescues UNC13A RNA levels after TDP-43 KD. n=5 biological replicates from 2 experiments. (j) Western blot analysis of protein lysates indicates CE deletion rescues UNC13A protein levels after TDP-43 KD. (k) Quantification of results in (j) n=4 biological replicates from 2 experiments. Graphs for (c) (d) (f) (h) (i) and (k) represent mean ± s.e.m. One-way ANOVA with Tukey multiple comparison test. *P < 0.05; ** P <0.01; *** P <0.001; **** P < 0.0001; ns (not significant).

Fig. 3 |. Genomic deletion of UNC13A CE rescues synaptic function following TDP-43 knockdown.

a, Immunofluorescence labelling of synapsin and UNC13A pre-synaptic terminals in 4-week old HaloTDP and UNC13A CE Del iNeurons grown on rat astrocytes following 2-weeks of TDP-43 knockdown. Scale bar = 10 μm. b, Quantification of UNC13A intensity at synapsin positive synaptic terminals. Control n=71, TDP-43 knockdown n=71, CE Del n=66, CE Del TDP-43 knockdown n=75 synapses from 3 experiments. c, Representative mEPSC traces and average amplitude traces. d, Quantification of mEPSC frequency and inter-event interval cumulative distributions from control n=17, TDP-43 knockdown n=22, CE Del n=15, CE Del TDP-43 knockdown n=28 iNeurons pooled from 3 experiments. e, Representative raster plots from D39 multi-electrode array recordings from control, TDP-43 knockdown, CE Del, and CE Del TDP-43 knockdown conditions (30s window shown). Electrode bursts in blue and network bursts in pink. f, Time-course of network synchrony measured by the area under the normalised electrode cross correlation with TDP-43 knockdown conditions normalised to each control n=24 wells from 4 experiments. Graphs for (b) (d) (f) represent mean ± s.e.m. Statistics for (b) and (d) are one-way ANOVAs with Dunnet’s multiple comparisons test, statistics for (f) is a paired t test. *P < 0.05; ** P <0.01.

Fig. 4 |. ASO mediated correction of UNC13A CE rescues synaptic function following TDP-43 knockdown.

a, Schematic of UNC13A cryptic exon. b, RT-PCR analysis of UNC13A splicing in HaloTDP iNeurons shows 21nt ASOs prevent cryptic splicing after TDP-43 knockdown with quantification of results n=4 biological replicates from 2 experiments. c, Western blot analysis shows rescue of UNC13A protein following 21nt ASO treatment after TDP-43 KD with quantification. n=4 biological replicates from 2 experiments. d, Quantification of UNC13A intensity at synapsin positive synaptic terminals in TDP-43 knockdown iNeurons treated with MOE CTRL ASO (n=88) and M21–4 ASO (n=96). Scale bar = 10 μm. e, RT-PCR analysis of UNC13A splicing in HaloTDP iNeurons shows 18 nt ASOs prevent cryptic splicing after TDP-43 knockdown with quantification of results n=4 biological replicates from 2 experiments. f, Western blot analysis shows rescue of UNC13A protein following 18 nt ASO treatment after TDP-43 KD with quantification. n=4 biological replicates from 2 experiments. g, Representative mEPSCs traces and average mEPSCs traces from TDP-43 KD iNeurons treated with MOE CTRL ASO (n=17) and M21–4 ASO (n=18). h, Quantification of mEPSC frequency and inter-event interval cumulative distributions. i, Representative mEPSCs traces and average mEPSCs traces from TDP-43 KD iNeurons treated with CTRL ASO (n=19) and M18–17 ASO (n=19). j, Quantification of mEPSC frequency and inter-event interval cumulative distributions. Graphs for (b) (c) (d) (e) (f) (h) and (j) represent mean ± s.e.m. Statistics for (b) (c) (e) and (f) are One-way ANOVA with Tukey multiple comparison test. Statistics for (d) (h) and (j) are two-sided Student’s t test. *P < 0.05; ** P <0.01; *** P <0.001; **** P < 0.0001; ns (not significant).