# Short RNA chaperones promote aggregation-resistant TDP-43 conformers to mitigate neurodegeneration.

- **DOI:** [10.1126/science.adv3301](https://doi.org/10.1126/science.adv3301)
- **Rating:** Strong (4.1 out of 5)
- **Verdict:** Defines how short RNAs allosterically remodel TDP-43 into aggregation-resistant conformers and reverses pathology in mice, constrained by a single acute preclinical model.
- **Framework version:** v5.9
- **Pipeline version:** v0.4
- **Evaluation date:** 2026-10-08
- **Evaluation page:** https://www.nabu.science/paper/10.1126%2Fscience.adv3301

## Key claims

1. These short RNAs engage and stabilize the TDP-43 RNA recognition motifs, which allosterically destabilizes a conserved helical region in the prion-like domain, thereby promoting aggregation-resistant conformers.
   - Support: Supported
   - Type: Primary finding
   - Basis: Converging hydrogen-exchange, NMR, binding and simulation evidence; causal sufficiency of helix destabilisation inferred rather than directly tested.
   - Location: Abstract; Results, Figs. 1–2
2. In mice with cytoplasmic TDP-43 aggregation and motor neuron loss, an enhanced short RNA chaperone reduced pathological aggregation, restored TDP-43 function, and conferred neuroprotection.
   - Support: Tentative
   - Type: Secondary
   - Basis: One acute overexpression model, 3–5 day follow-up, saline-only comparator, function read at a single splicing target, no behavioural endpoints.
   - Location: Abstract; Results, Fig. 6
3. Enhanced short RNA chaperones mitigated aberrant TDP-43 phenotypes in optogenetic models and in ALS patient-derived and control motor neurons.
   - Support: Supported
   - Type: Secondary
   - Basis: Consistent rescue across two systems, but few patient lines, one control line, and some imaging sample sizes not stated.
   - Location: Abstract; Results, Fig. 5 and figs. S21–S25
4. Malat1_start emerged as the most potent natural RNA chaperone, with the lowest IC50 values against WT TDP-43 (IC50~0.36 μM) and disease-linked variants (IC50~0.17 μM–0.44 μM).
   - Support: Supported
   - Type: Secondary
   - Basis: Dose-response IC50 panels across many variants on one assay platform; potency margin over the prior lead is modest.
   - Location: Results, Mining section; Fig. 4B

## Strengths

- **Mechanism supported by four orthogonal methods** The allosteric link from RNA recognition motif engagement to destabilisation of the prion-like domain's conserved region is built from domain-deletion biochemistry, hydrogen/deuterium-exchange mass spectrometry with at least 87.9% sequence coverage across a 1 s to 14.5 h timescale, NMR chemical shift perturbations, and all-atom molecular dynamics. Bimodal mass spectra at early timepoints further show that RNA binding depopulates the more stabilised conserved-region conformer.
  - Dimension: methodological rigour
  - Location: Results, Figs. 1–2 and figs. S5–S14
- **Functional counter-screen disqualifies a potent candidate** Rather than advancing the tightest binder, the work uses a TDP-43 loss-of-function biosensor to show that (UG)17 blocks aggregation while impairing TDP-43 function, and removes it from development. This is a substantive use of a negative result to constrain the therapeutic hypothesis and cautions against affinity-only screening.
  - Dimension: contribution
  - Location: Results, CUTS biosensor section, fig. S21G; Discussion
- **Randomised, blinded, powered in vivo experiment** The mouse study reports a prospective power calculation, randomisation of TDP-43ΔNLS animals to saline or RNA, motor-neuron counting by a blinded second individual, stated inclusion criteria based on YFP confirmation, and no attrition. Target engagement is independently evidenced by partial colocalisation of Cy5-labelled RNA with TDP-43 puncta.
  - Dimension: methodological rigour
  - Location: Materials and Methods, Animals and Animal Data Analysis Considerations; Fig. 6; fig. S26
- **Breadth across disease-relevant TDP-43 forms** Chaperone activity is quantified by IC50 against eight missense variants and five post-translational modification mimetics, including the K145/K192Q acetylation mimetic that defeats the prior lead. Identifying an RNA that retains activity against that variant addresses a concrete translational concern about patient heterogeneity.
  - Dimension: contribution
  - Location: Results, Figs. 3–4 and figs. S15–S17

## Limitations

- **No matched control oligonucleotide in vivo** The mouse experiment compares RNA treatment against saline only. Without a sequence-matched, non-binding oligonucleotide of the same 2'OMe and phosphorothioate chemistry, non-specific effects of the modified backbone on the observed neuroprotection and aggregate reduction are not excluded.
  - Dimension: methodological rigour
  - Location: Materials and Methods, Spinal Delivery of RNA; Fig. 6
- **Single acute, single-sex in vivo model** In vivo efficacy rests on one acute AAV9 TDP-43ΔNLS overexpression paradigm in female C57BL/6J mice, with a single dose delivered by laminectomy and gelfoam application and outcomes read 3–5 days later. No behavioural, motor or survival endpoints are reported, and function is assessed at a single splicing target.
  - Dimension: contribution
  - Location: Materials and Methods, Animals and Intraspinal Delivery; Fig. 6
- **Hydrogen-exchange differences binned without statistics** Consensus exchange difference maps are produced by manually scoring peptides into ±10%, ±20% and ±30% bins and rounding the per-residue average, with no statistical test and with the stated caveat that the displayed values do not report the actual percentage difference or proximity to each cutoff.
  - Dimension: methodological rigour
  - Location: Materials and Methods, HXMS Data Visualization
- **Translational limits absent from limitations section** The discussion names additional mouse models, broader patient-line diversity and further RNA optimisation as limitations, but does not address the invasive surgical delivery route, the female-only cohort, or the short follow-up when asserting translational implications for symptomatic patients.
  - Dimension: positioning
  - Location: Discussion, limitations paragraph and preceding in vivo paragraph
- **Semi-quantitative and technical-replicate cellular readouts** Cryptic splicing of STMN2 and KCNQ2 is quantified from band intensity after 40 PCR cycles, which saturates, and is shown only in control neurons under arsenite stress. Neuritic granule measurements are averaged over 20–25 neurites in four technical replicates of the same lines rather than independent biological replicates.
  - Dimension: methodological rigour
  - Location: Materials and Methods, RT-PCR and iPSC-derived neuron treatment and immunostaining; figs. S23–S25

## Reliability checks

Caveats (4 of 4 checks).

- **Results check:** Clean. The reported numbers and statistics are consistent with the stated results.
- **Reporting and transparency:** Noted. Animal allocation leaves 10 virus-injected animals unassigned in the description, sham treatment is described differently across sections, and the exclusion rule, power calculation, and fluorescence terminology require clarification.
  - Conflict of interest: not stated · Ethics approval: declared · Data availability: declared
- **Reference integrity:** Clean
  - 100 references in manuscript
  - 100 of 100 checkable references found in an index
  - References not verified automatically are resolved manually within 24 hours.

## Post-publication record

- **Status:** Clean
- **Summary:** No retraction or expression of concern found in Crossref (includes Retraction Watch).
- **Last checked:** 2026-10-17
