Research Monitor

What they found

Autosomal recessive deafness 9, caused by OTOF gene mutations, is characterized by severe-to-complete congenital deafness1. Although gene therapy has shown benefits in a small number of patients2-5, its safety and efficacy across broader age ranges and longer follow-up periods, as well as predictors of treatment outcomes, remain unclear. In this single-arm, multicentre trial conducted at eight centres, 42 participants (aged 0.8-32.3 years) received adeno-associated virus (AAV) serotype 1 carrying a human OTOF coding transgene (AAV1-hOTOF) at three vector dose groups, with up to 2.5-year follow-up. The primary end point was dose-limiting toxicity within 6 weeks. The secondary end point assessed efficacy and adverse events. No dose-limiting toxicities were observed. Grade 3 adverse events included decreased neutrophil count. Hearing was recovered in 90% of participants treated with AAV1-hOTOF, with gradual and stable improvement in auditory brainstem response threshold from greater than 97 ± 1 dB normalized hearing level at baseline to 54 ± 3, 51 ± 3, 50 ± 3 and 42 ± 5 dB normalized hearing level at 1, 1.5, 2 and 2.5 years, respectively, and behavioural audiometry improving from greater than 96 ± 3 dB hearing level at baseline to 37 ± 5 dB hearing level at 2.5 years. Participants aged 0.5-18 years showed greater hearing improvement than adults. A higher number of present distortion product otoacoustic emissions at baseline or biallelic non-truncated OTOF variants was associated with better hearing recovery. Participants with hearing recovery demonstrated gradual improvement in speech perception. AAV1-hOTOF is well-tolerated and efficacious across a broader patient population, with sustained therapeutic benefits for up to 2.5 years. Chinese Clinical Trial Regi

Links

• pubmed_id: https://pubmed.ncbi.nlm.nih.gov/42020731/
• DOI: https://doi.org/10.1038/s41586-026-10393-y

Connections

• [source] auto-indexed 2026-04-23 by [[strc-lit-watch]]

What they found

Comprehensive clinical review of RADA16 (PuraMatrix, PuraStat, PuraSinus) scaffold in vivo performance across tissue types. Best available PK/degradation data: bulk gel dissolution observed within 3–7 days in rat liver injury model, complete by 2 weeks; rat brain TBI model shows scaffold present at 1 week, absent at 3 weeks. Modified/crosslinked variants: 61% degraded at 35 days, 82% non-crosslinked. Degradation products are L-amino acids — no inflammation. Fibril dimensions: individual monomer ~6 nm; nanofibers smaller than wavelength of light; pores 50–200 nm. Modified RADA16 (RADA16-RGD, RADA16-IKVAV etc.): adding ≤12 residues preserves gelation; longer sequences progressively reduce beta-sheet signal and viscoelasticity.

How this applies to h09 — CRITICAL PK FLAG

This is the primary source for in vivo RADA16 degradation data. The scaffold persists 1–3 weeks in soft CNS/liver tissue. This is INCONSISTENT with the model’s K_PROTEOLYSIS = 1.4/h (t½ = 30 min). That rate constant is likely applicable to free monomeric peptide in serum, not the assembled hydrogel scaffold. The assembled gel is sterically protected: literature shows days-to-weeks persistence, implying effective gel t½ on the order of ~1–7 days (k ~0.004–0.029/h for gel, not 1.4/h).

The 118 aa tail (h09 construct) exceeds the 12-residue functional appendage limit established here. Literature consensus: the longer the appended motif, the lower the beta-sheet signal and viscoelasticity. A 118 aa WH2 domain appended to RADA16 is untested — no paper has characterized assembly for such long fusions. This is a critical open question.

Key numbers — CRITICAL DEGRADATION DATA

• Rat liver: scaffold visible 3–7 days, complete dissolution by 14 days
• Rat brain TBI: scaffold present at 7 days, absent at ~21 days
• Crosslinked RADA16: 61% degraded at day 35; uncrosslinked: 82% degraded at day 35
• Maximum appended sequence preserving assembly: ~12 residues (empirical rule)
• Nanofiber individual monomer: ~6 nm length
• Pore size: 50–200 nm

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/34164387/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC8216384/
• DOI: https://doi.org/10.3389/fbioe.2021.679525

Connections

• [[STRC Phase 4d F-actin Bundling Model]] — K_PROTEOLYSIS flag; gel t½ days not minutes
• [[STRC Horizontal Top Connector Hydrogel Hypothesis]] — clinical PK; 118 aa tail assembly concern
• [see-also] [[Yokoi 2005 RADA16 Reassembly]] — fibril geometry
• [see-also] [[Modification Strategies RADA16 PMC9739689]] — tail length limits

What they found

RADA16-I nanofiber scaffold sonicated into 20–100 nm fragments reassembles to original length (hundreds of nm to few microns) within 2 hours; this cycles reproducibly. Key geometry data from AFM: nanofiber height 1.3–1.5 nm (single peptide thickness); individual peptide monomer ~5 nm long × 1.3 nm wide × 0.8 nm thick. Assembled fibers reach 615 ± 104 nm mean length. Hydrogel forms at 0.6–3 mM (1–5 mg/ml), >99.5% water. Storage modulus G′ ~50 Pa at full reassembly. pH 7.5, 20 mM Tris.

How this applies to h09

AFM length distribution directly supports the model’s fibril_length_nm = 1000 — fibers reach several hundred nm to low microns, so 1 μm is a plausible upper-range estimate. The 1.3–1.5 nm AFM height matches a single-peptide-layer tape, consistent with the bilayer (double-sheet) model that underlies the 4.35 peptides/nm packing estimate. No direct statement of peptides/nm, but the monomer dimensions are consistent with H-bond spacing ~0.47 nm.

Key numbers — CRITICAL GEOMETRY

• Peptide monomer: 5 nm long × 1.3 nm wide × 0.8 nm thick
• Nanofiber AFM height: 1.3–1.5 nm (single peptide layer)
• Nanofiber length (intact): several hundred nm to few μm; mean ~615 ± 104 nm
• Post-sonication fragments: 20–100 nm
• Reassembly complete: ~2 hours
• Hydrogel concentration: 0.6–3 mM (1–5 mg/ml)
• G′ (storage modulus): ~50 Pa at full reassembly (1 Hz, 25°C)
• Assembly conditions: 20 mM Tris pH 7.5, >99.5% water

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/15939888/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC1150805/
• DOI: https://doi.org/10.1073/pnas.0407843102

Connections

• [[STRC Phase 4d F-actin Bundling Model]] — fibril_length_nm = 1000 justified; peptide monomer dims
• [[STRC Horizontal Top Connector Hydrogel Hypothesis]] — AFM geometry, reassembly kinetics
• [see-also] [[Zhang 1993 EAK16 RADA16 Original Paper]] — original discovery
• [see-also] [[Paravastu 2014 RADA16 Molecular Structure]] — atomic model of packing

What they found

Comprehensive review of molecular self-assembly strategies for novel biomaterials, covering both top-down and bottom-up approaches. Covers RADA16 and EAK16 as the canonical ionic self-complementary SAPs, lipid-like peptides, and rotaxanes/dendrimers. Describes chemical complementarity (charge pairing, hydrophobic matching) and structural compatibility as the two design rules. Review of 3D scaffold applications in cell culture, drug delivery, tissue engineering. Nat Biotechnol 21:1171–1178.

How this applies to h09

Most useful as a citation anchor for RADA16’s design principles and the claim that the scaffold is a biomaterial platform. Does not provide precise fibril geometry numbers beyond what the original papers report. The two design rules (chemical complementarity + structural compatibility) are directly relevant to asking whether attaching a 118 aa WH2-containing tail violates the structural compatibility requirement — large disordered appendages increase configurational entropy at the beta-sheet surface and are likely to disrupt packing at high molar fraction.

Key numbers

• RADA16: ionic self-complementary, 50% charged residues, alternating hydrophobic/hydrophilic
• 3D scaffold pore size consistent with earlier papers: 50–200 nm
• Design rule: >50% sequence must be the self-assembling core to preserve assembly

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/14520402/
• DOI: https://doi.org/10.1038/nbt874

Connections

• [[STRC Horizontal Top Connector Hydrogel Hypothesis]] — RADA16 as biomaterial backbone
• [[STRC Phase 4d F-actin Bundling Model]] — design rules for tail-appended RADA16
• [see-also] [[Zhang 1993 EAK16 RADA16 Original Paper]] — original discovery
• [see-also] [[Yokoi 2005 RADA16 Reassembly]] — structural follow-up

What they found

Landmark Northwestern paper introducing pH-triggered self-assembly of peptide amphiphiles (PAs) into nanofibers that direct hydroxyapatite mineralization aligned with the fibril long axis — mimicking bone collagen. PA design: alkyl tail + beta-sheet segment + charged head. Reversibly crosslinked with cysteine disulfides. Fibril geometry: cylindrical nanofibers, diameter ~7–8 nm by TEM. pH-responsive: assembles below pH 7, disperses above pH 7. Science 294:1684, DOI 10.1126/science.1063187.

How this applies to h09

PA nanofibers are a structural comparator to RADA16 SAP fibrils. Key difference: PAs are cylindrical (not flat tapes), rely on hydrophobic core + beta-sheet. The ~7–8 nm diameter is consistent with the 3–10 nm range seen in beta-sheet-forming SAPs. PA geometry is less directly applicable to RADA16 packing density, but supports the general principle of sub-10 nm fibril cross-sections for beta-sheet SAPs. Stupp group in vivo persistence: not quantified in this paper.

Key numbers

• Fibril diameter (TEM): ~7–8 nm (cylindrical)
• Assembly condition: pH < 7 (protonation of head groups)
• Crosslinking: cysteine disulfides (reversible with reducing agents)
• Mineralization: HA c-axis aligned with fibril long axis

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/11721046/
• DOI: https://doi.org/10.1126/science.1063187

Connections

• [see-also] [[STRC Horizontal Top Connector Hydrogel Hypothesis]] — PA as comparator scaffold class
• [see-also] [[Yokoi 2005 RADA16 Reassembly]] — RADA16 geometry comparison
• [see-also] [[Zhang 2003 RADA16 Biomaterials Review]] — SAP review contextualizes PAs

What they found

RADA16/EAK16 scaffolds support neuronal attachment, extensive neurite outgrowth (PC12, hippocampal, cerebellar, DRG neurons), and functional synapse formation comparable to Matrigel. First in vivo biocompatibility data: intramuscular injection in Fisher 344 rats (1 mg/ml, 20 μl) showed no inflammation, necrosis, motor impairment, or tissue abnormalities at 9 days to 5 weeks. Injection site was visually undetectable. No immune responses detected (rabbit/goat antibody assay). Scaffold pore size 50–200 nm, >99% water, macroscopic filament diameter 10–20 nm (consistent with Zhang 1993).

How this applies to h09

Biocompatibility data is directly relevant to the ototopical delivery route. The five stabilization mechanisms listed (H-bonds, ionic bonds, hydrophobic interactions, overlap, salt coordination) confirm that assembly depends on all four RADA units — truncation or charge disruption anywhere is likely to impair gelation, which is relevant to the 118 aa tail question. The “no immune response” data supports the scaffold as a safe delivery vehicle.

Key numbers

• Scaffold filament diameter (EM): 10–20 nm (bundle; consistent with Zhang 1993)
• Scaffold pore size: 50–200 nm
• In vivo biocompatibility: no inflammatory response, 9 days–5 weeks, rat IM injection
• Peptide assembly concentration: 1–10 mg/ml
• Cell types supporting growth: PC12, hippocampal, cerebellar granule, DRG, sympathetic neurons

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/10841558/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC18719/
• DOI: https://doi.org/10.1073/pnas.97.12.6728

Connections

• [[STRC Horizontal Top Connector Hydrogel Hypothesis]] — in vivo biocompatibility; no immune response
• [[STRC Phase 4d F-actin Bundling Model]] — supports safe delivery assumption
• [see-also] [[Zhang 1993 EAK16 RADA16 Original Paper]] — structural basis
• [see-also] [[Yokoi 2005 RADA16 Reassembly]] — geometry follow-up

What they found

Original description of ionic self-complementary oligopeptides EAK16 and RADA16 (RAD16). A 16-residue peptide with alternating hydrophilic charged and hydrophobic alanine residues forms a characteristic beta-sheet CD spectrum in water; upon addition of millimolar monovalent salts (physiological conditions), spontaneously assembles into a stable macroscopic membrane. Membrane resists heat, extremes of pH, guanidinium HCl, SDS/urea, and proteolytic enzymes. Filament diameter by scanning EM: approximately 10–20 nm. No atomic-resolution structure — primary contribution is discovery and macroscopic characterization.

How this applies to h09

Establishes the RADA16 scaffold as the backbone for h09’s synthetic horizontal top connector. The 10–20 nm filament diameter (EM) is the original size estimate; later refined to 3–8 nm individual fibrils (Paravastu 2014 TEM) — the larger value likely reflects bundled fibers. The observation that the assembled membrane resists proteolytic enzymes is relevant to the K_PROTEOLYSIS assumption, though no kinetic quantification is given.

Key numbers

• Filament diameter (scanning EM): 10–20 nm
• Assembly trigger: millimolar monovalent salt (NaCl, KCl), physiological conditions
• Stability: resists protease, denaturants, heat, extreme pH after assembly
• Peptide: 16 residues, (RADA)4 or (EAKA)4 motif

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/7682699/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC46294/
• DOI: https://doi.org/10.1073/pnas.90.8.3334

Connections

• [[STRC Horizontal Top Connector Hydrogel Hypothesis]] — foundational scaffold paper; fibril dimensions
• [[STRC Phase 4d F-actin Bundling Model]] — supports proteolytic stability claim but provides no rate data
• [see-also] [[Yokoi 2005 RADA16 Reassembly]] — fibril geometry AFM follow-up
• [see-also] [[Paravastu 2014 RADA16 Molecular Structure]] — atomic structure NMR follow-up

What they found

4-Phenylbutyric acid (4PBA), a chemical chaperone, rescued cell surface localization of a missense-mutated membrane protein (DSCAML1 A2105T) that was trapped intracellularly. In a knock-in mouse model, oral 4PBA administration corrected dendritic retraction, hippocampal interneuron clustering deficits, and reduced epileptic spike-and-wave discharges. This is an in vivo proof-of-concept that a chemical chaperone can rescue trafficking of a missense-mutated extracellular protein.

Lateral connection

Stereocilin (STRC) is an extracellular protein that must traffic to the hair bundle surface. Misha’s maternal E1659A missense mutation likely causes misfolding or mistrafficking analogous to DSCAML1 A2105T. This paper demonstrates that a generic chemical chaperone (4PBA) can rescue surface localization of a patient-specific missense mutation in vivo — the exact therapeutic strategy of the STRC Pharmacochaperone campaign. The fact that oral administration was sufficient is encouraging for clinical translation feasibility.

Hypothesis suggested

A pharmacological chaperone (potentially 4PBA itself, or a more specific compound from our virtual screen) could rescue STRC E1659A trafficking to the stereocilia surface. Maps directly to STRC Pharmacochaperone Virtual Screen E1659A (S-tier). Testable: does 4PBA restore STRC E1659A surface expression in a heterologous cell system?

What could be computed

Molecular dynamics comparison of STRC E1659A misfolding trajectory vs. DSCAML1 A2105T. Docking of 4PBA to the STRC E1659A predicted misfolded intermediate. Free energy perturbation of 4PBA binding to STRC vs. our current salicylic acid / indole-3-acetic acid lead compounds.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42003794/
• DOI: https://doi.org/10.1111/gtc.70115

Connections

• [part-of] [[h01 hub]]
• [source] auto-indexed 2026-04-21 by [[strc-lit-watch]]

What they found

A structure-based virtual screen of >11,000 compounds against the α-synuclein fibril structure identified IP-045 (2-fluorophenyl 3-(1H-indol-3-yl)propanoate) as a lead chemical chaperone. IP-045 strongly inhibited α-synuclein aggregation in vitro, reduced ROS and ER stress markers in cell models, and improved motor/cognitive performance in a rotenone-induced PD rat model. The compound is an indole derivative — notably, indole-3-acetic acid is one of our pharmacochaperone lead compounds.

Lateral connection

The virtual screening pipeline (11K compounds → 4 candidates → lead optimization → in vivo) parallels our STRC pharmacochaperone campaign, which has completed phases 0-3 with a top-5 leads shortlist including an indole-3-acetic acid derivative. IP-045 is structurally related (indol-3-yl propanoate vs. indole-3-acetic acid). Their demonstration that an indole-class chaperone works in vivo for protein misfolding provides confidence in our indole lead chemotype.

Hypothesis suggested

Indole-scaffold chemical chaperones may have broad utility for rescuing protein misfolding/aggregation. Our indole-3-acetic acid lead for STRC E1659A is from the same chemical family. Maps to STRC Pharmacochaperone Virtual Screen E1659A (S-tier). Testable: does IP-045 itself show any STRC E1659A binding activity?

What could be computed

Pharmacophore comparison between IP-045 and our indole-3-acetic acid lead. Dock IP-045 against STRC E1659A binding pocket. SAR analysis of the indole scaffold — which substituents drive chaperone vs. aggregation-inhibition activity?

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42003184/
• DOI: https://doi.org/10.1002/ddr.70288

Connections

• [part-of] [[h01 hub]]
• [source] auto-indexed 2026-04-21 by [[strc-lit-watch]]

What they found

Low-intensity focused ultrasound (FUS) combined with IV microbubbles achieved non-invasive, spatially precise AAV9 delivery to the macaque cerebellum by transiently opening the blood-brain barrier. ssAAV9-CMV-mCherry achieved robust transduction of virtually all neurons in the targeted deep cerebellar nuclei. Two vector types were tested (scAAV9-CBA-GFP and ssAAV9-CMV-mCherry), both with high spatial precision. This is among the first demonstrations of FUS-mediated AAV gene therapy in NHP deep brain structures.

Lateral connection

The cochlea has analogous barrier challenges: the blood-labyrinth barrier limits systemic delivery of gene therapy vectors. FUS-mediated transient barrier opening could enable non-invasive cochlear gene delivery without surgical round window injection. The NHP success with AAV9 is relevant because AAV serotypes with inner ear tropism (Anc80L65, AAV9) could potentially be delivered systemically with FUS targeting the cochlea. This would eliminate the surgical delivery bottleneck that currently limits all inner ear gene therapies.

Hypothesis suggested

FUS + IV microbubbles could enable non-invasive cochlear gene delivery by transiently opening the blood-labyrinth barrier, allowing systemically administered AAV or LNP vectors to transduce hair cells. Maps to Alternative STRC Delivery Hypotheses (reference tier) and indirectly supports STRC Mini-STRC Single-Vector (S-tier) and STRC mRNA-LNP Strategy B (S-tier) by expanding delivery options. Testable: does FUS targeting the cochlea in a rodent model open the blood-labyrinth barrier sufficiently for AAV9 transduction of OHCs?

What could be computed

Acoustic simulation of FUS focal zone size achievable in the temporal bone at cochlear depth. Comparison of blood-labyrinth barrier vs. BBB permeability thresholds for AAV9 passage. Literature meta-analysis of FUS parameters (frequency, pressure, microbubble dose) that achieved barrier opening without tissue damage.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42003738/
• DOI: https://doi.org/10.1002/advs.75307

Connections

• [source] auto-indexed 2026-04-21 by [[strc-lit-watch]]

What they found

The CV-FEST framework uses Harmonic Linear Discriminant Analysis (HLDA) collective variables from short MD trajectories to predict how point mutations affect protein folding/unfolding kinetics — without exhaustive sampling. Validated on Chignolin miniprotein mutants, the HLDA-derived residue-level scores predict whether mutations accelerate or slow conformational transitions. The leading HLDA eigenvalue correlates significantly with transition rates across mutations. This enables kinetic prediction from minimal local sampling.

Lateral connection

The STRC E1659A mutation likely alters the folding kinetics of stereocilin’s local domain, potentially trapping it in a misfolded intermediate that prevents trafficking. CV-FEST could predict whether E1659A accelerates or slows the folding transition compared to wild-type, and quantify the kinetic barrier that a pharmacochaperone must overcome. This is complementary to our existing electrostatic analysis of E1659A and could inform the pharmacochaperone binding energy threshold needed for therapeutic rescue.

Hypothesis suggested

CV-FEST analysis of the STRC region around position 1659 could quantify the folding kinetic penalty of E→A substitution and predict whether our pharmacochaperone leads have sufficient binding energy to compensate. Maps to STRC Pharmacochaperone Virtual Screen E1659A (S-tier) and STRC Electrostatic Analysis E1659A (reference). Testable: apply HLDA to short MD trajectories of WT vs. E1659A STRC domain to predict kinetic barrier difference.

What could be computed

HLDA collective variables from WT and E1659A STRC domain MD. Free energy barrier estimation for local folding transition. Correlation with pharmacochaperone binding free energies from our virtual screen. Residue-level mutation sensitivity scores for positions adjacent to E1659 (could identify synergistic stabilizing mutations for protein engineering).

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42007551/
• DOI: https://doi.org/10.1021/acs.jctc.6c00418

Connections

• [source] auto-indexed 2026-04-21 by [[strc-lit-watch]]

What they found

LinkCraft is a computational tool for rational design of intrinsically disordered linkers (IDLs) in multi-domain proteins. It suggests optimal IDL lengths as a function of inter-domain distance, supports custom or auto-generated linker sequences based on desired physicochemical properties, and enables ensemble-based structural modeling of the complete multi-modular protein. The tool treats linkers as active, tunable determinants of molecular function rather than passive connectors.

Lateral connection

Mini-STRC (S-tier) involves truncating stereocilin domains while preserving function. The remaining domains must maintain proper spatial relationships, and the truncation junctions are effectively new linker regions. LinkCraft could help design optimal junction sequences that maintain the inter-domain geometry of full-length STRC after domain deletion. Additionally, the STRC In Situ SpyCatcher Assembly hypothesis (B-tier) requires two STRC fragments to assemble with proper inter-domain spacing — linker design is critical for both the SpyCatcher/SpyTag junction and the overall protein architecture.

Hypothesis suggested

Rational linker design at Mini-STRC truncation junctions could improve folding efficiency and functional preservation of the minimized protein. Maps to STRC Mini-STRC Single-Vector Hypothesis (S-tier) and STRC In Situ SpyCatcher Assembly (B-tier). Testable: use LinkCraft to model the truncation junction in Ultra-Mini-STRC and compare predicted ensemble dynamics against full-length domain-domain distance distributions from MD.

What could be computed

LinkCraft modeling of the Ultra-Mini-STRC truncation junctions. Ensemble conformational sampling of Mini-STRC with optimized vs. native junction sequences. Inter-domain distance distributions compared to full-length STRC. For SpyCatcher assembly: optimal linker length between SpyCatcher/SpyTag and the STRC fragment termini.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42001978/
• DOI: https://doi.org/10.1016/j.jmb.2026.169814

Connections

• [source] auto-indexed 2026-04-21 by [[strc-lit-watch]]

What they found

A synergistic templating method using PEDOT:PSS and wood fiber paper dramatically enhances the out-of-plane piezoelectric coefficient of P(VDF-TrFE) without requiring electric poling. The mechanism involves supramolecular interactions between PSS and paper that enrich linear PEDOT at the interface, which promotes β-phase crystallite growth over α-phase. Computational modeling confirms linear PEDOT provides more hydrogen bonds favoring β-phase PVDF nucleation. This eliminates the poling step that is typically required for PVDF-TrFE piezoelectric activation.

Lateral connection

PVDF-TrFE is the exact piezoelectric material in the STRC Piezoelectric TM Bioelectronic Amplifier hypothesis (S-tier). The central fabrication challenge for that hypothesis is depositing a piezoelectric film on or near the tectorial membrane. Eliminating the electric poling step is a major practical advance — poling requires high voltage across the film, which is incompatible with biological tissue. If a templating approach can achieve β-phase orientation without poling, it could enable in situ or ex vivo PVDF-TrFE deposition for cochlear applications.

Hypothesis suggested

PVDF-TrFE could be deposited on a biocompatible substrate using a conducting polymer template to achieve piezoelectric activity without poling. Maps directly to STRC Piezoelectric TM Bioelectronic Amplifier (S-tier). Testable: can PEDOT:PSS or a similar conductive template be applied to a collagen or protein substrate (mimicking TM composition) to template β-phase PVDF-TrFE growth?

What could be computed

MD simulation of PVDF-TrFE nucleation on collagen vs. PEDOT:PSS vs. bare substrate. DFT calculation of hydrogen bonding between PVDF-TrFE and candidate biological templates. Finite element model of piezoelectric output from a templated (no-poling) vs. poled PVDF-TrFE film at cochlear mechanical frequencies (1-20 kHz).

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42004736/
• DOI: https://doi.org/10.1021/acsaelm.6c00088

Connections

• [source] auto-indexed 2026-04-21 by [[strc-lit-watch]]

What they found

A programmable Janus bilayer scaffold integrates a piezoelectric PVDF/barium titanate (BT) electrospun membrane with a thermoresponsive hydrogel. Three switchable modes: continuous NIR for antibacterial therapy, intermittent NIR for immunomodulation, and no irradiation for mechanical support with Ca²⁺ release promoting mineralization. In vivo rat models achieved >60% bone healing at 4 weeks. The key advance is demonstrating that a piezoelectric polymer membrane can be programmatically activated in a living system to deliver different bioelectric signals.

Lateral connection

The programmable multi-mode concept is relevant to the STRC Piezoelectric TM Bioelectronic Amplifier (S-tier). A cochlear piezoelectric implant would similarly need to operate in different modes: passive mechanical coupling at rest, active piezoelectric transduction during sound stimulation. The PVDF/BT composite achieving switchable bioelectric output in vivo demonstrates material feasibility for implantable piezoelectric devices, though the NIR trigger mechanism is not applicable to deep cochlear tissue.

Hypothesis suggested

Piezoelectric PVDF composites with barium titanate could increase charge output per mechanical deflection, potentially improving the signal-to-noise ratio of a TM-coupled piezoelectric amplifier. Maps to STRC Piezoelectric TM Bioelectronic Amplifier (S-tier). Testable: does BT doping of PVDF-TrFE improve piezoelectric output at the low force levels relevant to TM deflection (~nN range)?

What could be computed

FEM comparison of charge output from pure PVDF-TrFE vs. PVDF/BT composite under TM-scale deflections. Biocompatibility literature review of BT nanoparticles in neural/sensory tissue contexts.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42006006/
• DOI: https://doi.org/10.1016/j.bioactmat.2026.03.046

Connections

• [source] auto-indexed 2026-04-21 by [[strc-lit-watch]]

What they found

Kif19A, a kinesin-8 family motor, controls the molecular organization at the tip of fly mechanosensory cilia by enriching mechanosensory molecules without affecting overall ciliary structure. Individual Kif19A motors show transient microtubule binding and non-processive movement, but multiple motors coordinate for effective cargo transport. A theoretical model explains how local transport and binding sites counteract rapid diffusion to establish nanoscale molecular organization at the tip.

Lateral connection

Stereocilia tips are the critical functional zone where stereocilin must localize to form the top connectors between adjacent stereocilia and the tectorial membrane attachment crowns. The question of HOW proteins are concentrated and maintained at stereocilia tips — against diffusion — is directly analogous to this paper’s finding. While stereocilia use actin (not microtubules), the principle of localized active transport + binding sites counteracting diffusion could explain how stereocilin is concentrated at tip regions. If mini-STRC has altered diffusion kinetics due to truncation, understanding tip-enrichment mechanisms becomes critical for predicting its localization.

Hypothesis suggested

Mini-STRC localization efficiency at stereocilia tips depends on a balance between active transport, local binding affinity, and diffusion rate. Truncation may alter this balance by changing the protein’s diffusion coefficient (smaller = faster diffusion) or removing binding domains that anchor it at the tip. Computational modeling of this transport-binding-diffusion balance could predict whether mini-STRC will achieve sufficient tip concentration.

What could be computed

Brownian dynamics simulation of stereocilin vs. mini-stereocilin diffusion along stereocilia, incorporating: (1) protein hydrodynamic radius from AlphaFold-predicted structures, (2) estimated binding affinities at tip sites, (3) transport rates. This would predict steady-state tip concentration ratios between full-length and truncated forms.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41941305/
• DOI: https://doi.org/10.1083/jcb.202412213

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Integrated seed-based framework design with ribosome display to generate novel nanobody-like protein scaffolds. Four conserved 10-residue seed frameworks were embedded in random sequences creating a library >10^13 variants. Iterative ribosome display selection plus AlphaFold 3 structural prediction identified candidates with canonical nanobody-like folds. Two candidates demonstrated stable expression with thermal stability ~63.2C, validating that minimal conserved “seed” sequences can nucleate correct protein folding even in randomized contexts.

Lateral connection

The seed-based design principle — keeping minimal conserved framework residues while randomizing the rest — is conceptually the inverse of our mini-STRC truncation approach. For mini-STRC, we’re removing internal domains while keeping flanking regions. The question is: what are the minimal “seed” sequences in stereocilin that must be preserved for correct folding and function? This paper’s methodology of identifying conserved seed frameworks through sequence alignment across homologs could be applied to stereocilin to identify the irreducible core sequences. AlphaFold 3 validation of fold predictions further supports our computational approach to evaluating truncation designs.

Hypothesis suggested

Stereocilin contains conserved “seed framework” residues that are necessary and sufficient for correct extracellular folding. These seed residues define the minimum viable truncation boundaries for mini-STRC. AlphaFold 3-guided identification of these seeds could produce a smaller, more stable mini-STRC than the current truncation 700-1775 design.

What could be computed

Multiple sequence alignment of stereocilin orthologs across species to identify conserved framework residues. Map these onto the AlphaFold-predicted structure. Test whether removing non-conserved regions while preserving framework residues yields stable predicted folds. Compare predicted stability of seed-based mini-STRC designs vs. the current contiguous truncation.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41995290/
• DOI: https://doi.org/10.1093/bbb/zbag057

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Combined combinatorial mutagenesis with ML to engineer base editors with motif-specific activity. Profiled 160,000 evoAPOBEC1 and 64 million TadA variants in human cells, using only 0.004% of the mutational landscape for predictions. Identified variants that eliminated residual adenine editing in cytosine base editors. Achieved undetectable bystander edits in 50% of >800 disease-associated mutations tested. A structure-based deep learning model predicted functional TadA variants with 63% success across 20^26 variants spanning 26 amino acid sites without experimental data.

Lateral connection

STRC mutations causing DFNB16 include point mutations that could theoretically be corrected by base editing. The challenge has been bystander edits in the surrounding sequence context. This paper’s ML-powered approach to customizing base editors for specific motifs means that for any given STRC point mutation, a custom editor with minimal off-target activity could be computationally designed. The 0.004% sampling efficiency makes this experimentally tractable. Additionally, the principle of ML-guided enzyme engineering applies directly to our prime editing hypothesis — the same approach could optimize prime editors for STRC correction efficiency.

Hypothesis suggested

For prevalent STRC point mutations (e.g., missense mutations in conserved domains), ML-customized base editors designed for the specific sequence context could achieve precise correction with undetectable bystander editing, potentially delivered via AAV to outer hair cells.

What could be computed

Catalog all known pathogenic STRC point mutations, determine which fall within base editor windows (C>T or A>G), and computationally predict optimal editor variants using the structure-based deep learning approach described. Estimate what fraction of DFNB16 patients could be addressed by customized base editors vs. requiring the mini-gene approach.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41997156/
• DOI: https://doi.org/10.1016/j.molcel.2026.03.030

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

CenSpark is a cell-permeable dual-ligand fluorescent probe that selectively labels centriolar and axonemal microtubule structures without genetic manipulation. Achieves high selectivity in live and fixed specimens for centrioles, cilia, and flagella. Enables tracking primary cilium formation rates and monitoring centriole dynamics in real time.

Lateral connection

While stereocilia are actin-based (not microtubule-based) and would not be labeled by CenSpark directly, the kinocilium of developing hair cells IS a true cilium with axonemal microtubules. CenSpark could enable live imaging of kinocilium dynamics during hair cell development — a period when stereocilia bundle polarity is being established. Understanding kinocilium-stereocilia coordination during development informs when mini-STRC must be present to establish proper bundle architecture.

Hypothesis suggested

Live imaging of kinocilium dynamics with CenSpark in organoid or explant hair cell cultures could reveal the developmental time window during which stereocilin must be present for proper bundle formation, informing the therapeutic window for mini-STRC gene therapy delivery.

What could be computed

N/A — this is primarily an experimental tool. However, image analysis pipelines for quantifying cilium formation rates from CenSpark imaging data could be adapted for automated stereocilia bundle morphology quantification in mini-STRC efficacy studies.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41986560/
• DOI: https://doi.org/10.1038/s41589-026-02186-1

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Developed ProteinAligner, a multimodal pretraining framework integrating protein sequences, 3D structures, and scientific literature text. Uses sequence as an anchor to align other modalities through contrastive learning. Outperforms existing protein foundation models in predicting protein functions and properties across diverse downstream tasks by capturing richer, more holistic protein representations than sequence-only models.

Lateral connection

Stereocilin is a poorly characterized protein with no solved experimental structure — predictions of its function rely heavily on computational approaches. A tri-modal model incorporating sequence, AlphaFold-predicted structure, AND published literature about stereocilin and homologs could provide better functional predictions for truncated regions than any single modality. This is especially relevant for predicting which domains of stereocilin are functionally dispensable (safe to truncate) vs. essential.

Hypothesis suggested

Multimodal protein representation models that incorporate literature context alongside sequence and structure could predict functional impact of stereocilin truncations more accurately than structure-only approaches, because literature encodes experimental knowledge about domain functions that pure structural models miss.

What could be computed

Apply ProteinAligner (or similar multimodal model) to full-length stereocilin and candidate mini-STRC truncations. Compare predicted functional properties between full-length and truncated forms. Benchmark against AlphaFold-only structural predictions to assess whether literature-informed representations add predictive value for this specific protein.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41990738/
• DOI: https://doi.org/10.1016/j.crmeth.2026.101407

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Generated two transgene-free iPSC lines from a DFNB9 patient with compound heterozygous OTOF mutations (c.2122C>T and c.5197G>A) and a heterozygous carrier relative. Both lines showed normal pluripotency, karyotype, and trilineage differentiation. The carrier line enables comparative analysis of heterozygous mutation effects on auditory pathology.

How this applies to mini-STRC

Establishes a disease modeling platform for another autosomal recessive hearing loss gene (OTOF/DFNB9) that parallels our STRC/DFNB16 work. The iPSC-derived approach for mechanistic studies and gene therapy optimization is directly transferable — similar compound heterozygous patient lines could be generated for STRC mutations to test mini-STRC constructs in human-derived cells. The carrier line concept is particularly useful for dose-response studies of partial gene function.

Key numbers

• Patient mutations: c.2122C>T and c.5197G>A (compound heterozygous)
• Method: transgene-free episomal reprogramming
• Lines generated: 2 (patient + carrier)

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/42000544/
• DOI: https://doi.org/10.1016/j.scr.2026.103988

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Developed a multicolor mechanoluminescence platform that converts focused ultrasound into tunable light emission (461-592 nm) using ROS-responsive chemiluminescent donors coupled to fluorescent acceptors via FRET. Established a design principle: electronic energy gap-dependent ROS generation predicts material performance. Successfully activated three different optogenetic actuators (ChR2, eOPN3, ChRmine) under focused ultrasound for noninvasive deep-tissue neuromodulation.

Lateral connection

This directly bridges ultrasound energy to optogenetic control in deep tissue — the exact capability needed for our sonogenetics hypothesis applied to inner ear hair cells. The cochlea is acoustically accessible by definition. If mechanoluminescent nanoparticles could be delivered to the organ of Corti, focused ultrasound could activate light-sensitive ion channels in hair cells, providing a non-genetic alternative or complement to gene therapy. The tunable emission wavelength (blue to red) means compatibility with multiple channelrhodopsin variants, and red-shifted emission would minimize phototoxicity in the sensitive cochlear environment.

Hypothesis suggested

Mechanoluminescent nanoparticles delivered to the perilymph could convert cochlear sound-frequency vibrations or externally applied FUS into local light emission, activating optogenetic channels in transduced hair cells. This creates a hybrid sono-optogenetic cochlear prosthesis: AAV delivers the channelrhodopsin, nanoparticles provide the light source triggered by sound itself.

What could be computed

Acoustic field modeling of the cochlea to determine: (1) whether cochlear mechanics generate sufficient local pressure to trigger mechanoluminescence, (2) optimal nanoparticle positioning relative to outer hair cells, (3) required nanoparticle concentration to achieve threshold photon flux for channelrhodopsin activation at relevant frequencies.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41974592/
• DOI: https://doi.org/10.1021/jacs.5c22630

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Combined FUS-mediated blood-brain barrier disruption with docetaxel-loaded perfluorocarbon nanodroplets (62 nm diameter, >90% encapsulation efficiency) for glioblastoma treatment. Achieved 28-fold reduction in systemic clearance versus free drug and extended median survival to 36 days (vs. 20 days control) with 33% long-term survival. The dual approach of barrier opening plus targeted nanocarrier delivery proved synergistic.

Lateral connection

The blood-labyrinth barrier (BLB) protecting the inner ear is structurally analogous to the BBB. FUS-mediated transient BLB opening combined with drug-loaded nanocarriers could enable non-invasive delivery of gene therapy cargo (AAV or mRNA-loaded nanoparticles) to the cochlea without surgical round window access. The 62 nm nanodroplet size is within the range that could penetrate cochlear tissues. The 28-fold reduction in systemic clearance suggests this approach could dramatically reduce the AAV dose needed for cochlear transduction, addressing immunogenicity concerns.

Hypothesis suggested

FUS-mediated transient BLB opening combined with AAV-mini-STRC loaded in perfluorocarbon nanodroplets could achieve cochlear transduction without invasive surgical delivery, potentially enabling bilateral treatment and repeat dosing.

What could be computed

Finite element modeling of FUS propagation through temporal bone to the cochlea, predicting: (1) achievable focal pressures at the organ of Corti, (2) safety margins relative to hair cell damage thresholds, (3) optimal transducer geometry for cochlear targeting.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41993780/
• DOI: https://doi.org/10.2147/IJN.S571560

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Systematic review and meta-analysis of 9 preclinical studies on FUS-mediated gene therapy for orthotopic glioblastoma. Tested viral, non-viral, nanoparticle, and exosome-based vectors delivering therapeutic genes (MDA-7/IL-24, HSV-TK, CRISPR/Cas9). FUS significantly enhanced gene expression (pooled effect size 6.34, 95% CI 2.21-18.18), reduced tumor volume (pooled effect size 4.03, 95% CI 1.46-11.12), and improved survival (HR 1.33, 95% CI 1.13-1.56).

Lateral connection

The 6.34x enhancement of gene expression via FUS is a quantitative benchmark applicable to inner ear gene therapy. If FUS could similarly enhance AAV transduction of outer hair cells through the blood-labyrinth barrier, the required AAV dose for mini-STRC delivery would drop proportionally — potentially solving both the dosing and immunogenicity challenges. The meta-analysis across multiple vector types (viral, non-viral, exosome) suggests the FUS enhancement is vector-agnostic, meaning it would likely work with AAV-mini-STRC.

Hypothesis suggested

FUS applied to the cochlea during systemic or local AAV-mini-STRC administration could enhance outer hair cell transduction 3-6x compared to passive delivery, achieving therapeutic stereocilin levels with lower vector doses.

What could be computed

Dose-response modeling: given known AAV transduction efficiencies in cochlear studies, what minimum FUS-enhancement factor is needed to achieve >50% OHC transduction with clinically feasible AAV doses? Sensitivity analysis across reported enhancement ranges (2.2x to 18.2x from the meta-analysis CI).

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41955864/
• DOI: https://doi.org/10.1016/j.biopha.2026.119325

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Comprehensive review of how endothelial cells sense hemodynamic shear stress through multiple mechanosensitive proteins: ion channels (Piezo1, TRPV4), GPCRs, integrins, primary cilia, the glycocalyx, and cytoskeletal elements. Each sensor responds to different aspects of the mechanical environment (magnitude, direction, frequency). Pathological mechanosensing drives atherosclerosis and vascular anomalies. Key insight: cells use a distributed network of mechanosensors, not a single receptor.

Lateral connection

Hair cells similarly use a distributed mechanosensing apparatus — tip links, stereocilin-mediated lateral and top connectors, and the tectorial membrane attachment. Stereocilin’s role may be analogous to the endothelial glycocalyx: an extracellular mechanical coupling layer that shapes HOW forces are transmitted to the primary mechanotransduction apparatus (tip links/MET channels). Loss of stereocilin may not eliminate mechanosensation but alter its spectral characteristics — changing which frequencies and amplitudes are transduced, analogous to how glycocalyx degradation alters endothelial shear sensing without eliminating it.

Hypothesis suggested

Stereocilin functions as a mechanical frequency filter analogous to the endothelial glycocalyx — it doesn’t transduce forces directly but shapes the frequency-dependent mechanical coupling between the tectorial membrane and the MET channel apparatus. Mini-STRC must preserve this filtering function, not just physical attachment.

What could be computed

Lumped-parameter mechanical model of the OHC stereocilia bundle with stereocilin-mediated connectors modeled as viscoelastic elements. Compute the frequency response (transfer function) from tectorial membrane displacement to MET channel gating with: (1) full-length stereocilin, (2) mini-stereocilin (altered viscoelastic properties), (3) no stereocilin. Predict whether mini-STRC preserves the frequency filtering function.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41969033/
• DOI: https://doi.org/10.1242/jcs.264456

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Used ProteinMPNN deep learning framework to redesign non-epitope scaffold regions of a fusion protein for enhanced thermodynamic stability. Molecular dynamics simulations confirmed the redesigned construct achieved a rigid, compact native state. The optimized protein achieved high-yield soluble expression in E. coli without inclusion bodies, demonstrating that computational scaffold redesign can dramatically improve protein stability and expression while preserving functional regions.

Lateral connection

Mini-STRC truncation creates new domain junctions and exposed surfaces that may destabilize the protein. The ProteinMPNN approach of redesigning non-functional scaffold regions while preserving functional epitopes is directly applicable: the truncation junction in mini-STRC (where amino acids 699 and 1776 are joined) could be computationally redesigned using ProteinMPNN to optimize stability of the new interface without altering the functional N-terminal and C-terminal domains. The MD validation step provides a framework for computationally vetting designs before expensive in vivo testing.

Hypothesis suggested

ProteinMPNN-guided redesign of the truncation junction and exposed surfaces in mini-STRC could produce a more thermodynamically stable protein than simple truncation, improving folding efficiency, secretion, and ultimately stereocilia localization.

What could be computed

(1) Generate AlphaFold structure of mini-STRC with the 700-1775 truncation. (2) Use ProteinMPNN to redesign 5-10 residues flanking the junction while constraining the N-terminal and C-terminal functional domains. (3) Run MD simulations comparing stability of wild-type junction vs. ProteinMPNN-optimized junction. (4) Predict aggregation propensity using tools like CamSol.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41988819/
• DOI: https://doi.org/10.1111/1751-7915.70348

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Used high-speed video microscopy to biophysically characterize cilia from PCD patients homozygous for an RSPH4A founder variant. Measured significantly reduced angular excursion and beat amplitude compared to healthy controls. Parameters including bent angle, net angle, and amplitude per second provided an objective framework for quantifying the biophysical consequences of a specific genetic variant on ciliary mechanics.

Lateral connection

The quantitative biophysical phenotyping approach — measuring angular excursion, beat amplitude, and dynamic parameters via high-speed video — could be adapted for stereocilia bundle mechanics in STRC models. While stereocilia don’t beat like motile cilia, their deflection amplitude, stiffness, and recovery dynamics in response to mechanical stimulation are analogous measurements. The methodology of linking a specific genetic variant to quantitative biophysical parameters is exactly what’s needed to compare wild-type stereocilin vs. mini-STRC bundle mechanics.

Hypothesis suggested

High-speed video microscopy of stereocilia bundle deflection in response to calibrated mechanical stimulation could quantitatively distinguish the biophysical phenotype of mini-STRC-expressing OHCs from wild-type and knockout controls, providing a functional readout beyond simple morphology.

What could be computed

Develop image analysis algorithms for automated quantification of stereocilia bundle deflection parameters from high-speed video: maximum deflection angle, stiffness (force/deflection), viscoelastic relaxation time constant. Apply to existing or planned explant culture recordings comparing WT, STRC-KO, and mini-STRC-rescued hair cells.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41972697/
• DOI: https://doi.org/10.3390/cells15070607

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

Tunable stress-relaxation hydrogels revealed that mechanical confinement regulates chondrocyte ECM production through MAPK and Hedgehog signaling pathways. Low confinement promoted matrix deposition; high confinement increased catabolysis. Hedgehog activation enhanced matrix deposition and restored chondrogenic morphology. Primary cilia length inversely correlated with ECM production, with cilia acting as the mechanosensory antenna integrating confinement signals.

Lateral connection

Outer hair cells exist in a mechanically confined environment between the basilar membrane and tectorial membrane. Stereocilin forms the physical connections (top connectors, TM attachment crowns) that define this confinement geometry. The finding that mechanical confinement level directly regulates ECM remodeling via cilia-dependent signaling raises the question: does loss of stereocilin (and consequent loss of TM attachment) alter the mechanical confinement experienced by OHCs, triggering aberrant signaling that accelerates degeneration? This would mean DFNB16 pathology involves not just loss of mechanical coupling but also downstream signaling dysregulation.

Hypothesis suggested

Loss of stereocilin-mediated TM attachment in DFNB16 alters the mechanical confinement of OHCs, activating MAPK-mediated catabolic pathways that accelerate hair cell degeneration beyond what simple mechanical decoupling would cause. Mini-STRC must restore sufficient TM attachment to normalize mechanical confinement signaling, not just acoustic coupling.

What could be computed

Finite element modeling of OHC mechanical confinement with and without TM attachment (stereocilin present vs. absent). Predict local stress/strain changes at the hair cell apical surface. Map these to known MAPK/Hedgehog activation thresholds to predict whether confinement changes are sufficient to trigger degenerative signaling.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41985400/
• DOI: https://doi.org/10.1016/j.biomaterials.2026.124171

Connections

• [source] auto-indexed 2026-04-20 by [[strc-lit-watch]]

What they found

RBM24 is an RNA-binding protein that controls alternative splicing in cochlear hair cells. Exon 4 of mouse Rbm24 undergoes alternative splicing to produce two isoforms with distinct functions. Deletion of this exon causes stereocilia disorganization and progressive hair cell loss with severe hearing impairment. Critically: STRC is identified as a direct RBM24 splicing target — RBM24 regulates STRC mRNA processing, and appropriate expression levels are required for stereocilia structural integrity.

How this applies to mini-STRC

STRC expression is under splicing-level control by RBM24. This opens a hypothesis we didn’t have: instead of delivering a gene, could we modulate endogenous STRC expression by targeting RBM24? If a patient has a hypomorphic STRC mutation (partial function) rather than a null mutation, upregulating RBM24 activity or correcting its splicing targets could boost endogenous STRC to therapeutic levels without AAV delivery at all. Separate from mini-STRC, this paper also tells us that STRC mRNA structure matters — the transcript has splicing-sensitive regions that any gene therapy construct must account for.

Key numbers

• RBM24 targets: STRC among others (direct regulatory target confirmed)
• Phenotype of Rbm24 exon 4 deletion: stereocilia disorganization + hair cell loss + severe HL
• Mechanism: splicing regulation, not transcription level

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41973913/
• DOI: https://doi.org/10.1073/pnas.2531564123

Connections

• [source] [[STRC RBM24 Regulatory Hypothesis]] — primary basis for the regulatory hypothesis
• [see-also] [[STRC Mini-STRC Single-Vector Hypothesis]] — RBM24 acts at RNA level, mini-STRC at protein level — additive tracks

What they found

Organic nanoparticles that convert focused ultrasound into tunable light emission (blue to red) via mechanoluminescence — no implanted light source needed. The system couples reactive oxygen species-responsive chemiluminescent donors with fluorescent acceptors through energy transfer. In vitro: ultrasound triggered these nanoparticles to activate optogenetic proteins. The design principle (electronic energy gap → ROS generation → light output) is predictive and tunable, meaning specific wavelengths can be engineered for specific opsins.

Lateral connection

Our sonogenetics hypothesis requires getting light into OHCs deep in the cochlea to activate light-sensitive ion channels — without surgical implantation of an LED or fiber optic. This paper eliminates that barrier. Mechanoluminescent nanoparticles could be delivered through the round window membrane (same route as AAV), concentrate near OHCs, and then be activated externally via transcranial focused ultrasound. No implant. No surgery beyond what gene therapy already requires. The cochlea is acoustically accessible by design — it’s built to receive sound pressure.

Hypothesis suggested

Mechanoluminescent nanoparticles delivered via round window injection, combined with OHC-expressed channelrhodopsin (via AAV), could restore sound-driven OHC electromotility in DFNB16 patients without fixing STRC. External FUS transducer converts acoustic signal → nanoparticle light → opsin activation → OHC depolarization → amplification.

What could be computed

1. Acoustic modeling: FEM simulation of ultrasound propagation through skull + temporal bone to estimate intensity at OHC depth — does enough energy reach the cochlea to trigger mechanoluminescence?
2. Wavelength matching: Map available mechanoluminescent emission spectra against channelrhodopsin action spectra (ChR2 peaks at 470nm, ReaChR at 590nm) to identify optimal nanoparticle-opsin pairs.
3. Nanoparticle diffusion model: Simulate round window injection → endolymph diffusion → OHC surface concentration over time.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41974592/
• DOI: https://doi.org/10.1021/jacs.5c22630

Connections

• [[Sonogenetic STRC Computational Proof]] — resolves the implant-free light-delivery problem
• [see-also] [[STRC Mini-STRC Single-Vector Hypothesis]] — enables sonogenetic + mini-STRC convergence path
• [see-also] [[Alternative STRC Delivery Hypotheses]] — adds a sound→light→biology pathway alongside sonoporation

What they found

TMEM145 is a newly identified transmembrane protein that forms the core of TM-ACs (tectorial membrane attachment crowns) on outer hair cell stereocilia. Without TMEM145, stereocilin is lost from OHC stereocilia, the hair bundle physically disconnects from the tectorial membrane, and profound hearing loss results. TMEM145 appears to act as an anchor or scaffold that retains stereocilin at the stereocilia tip — not just a passenger in the complex.

How this applies to mini-STRC

This changes the structural model for mini-STRC design. Stereocilin doesn’t sit at the TM-AC alone — it depends on TMEM145 for localization. Mini-STRC (residues 700–1775) must preserve the TMEM145-binding interface, or the truncated protein will express but fail to localize correctly. This is the critical test: does the truncated region contain the TMEM145 interaction domain? Computational priority: AlphaFold-Multimer model of STRC-TMEM145 complex, then map which STRC residues form the interface, then verify those residues fall within our 700–1775 window.

Key numbers

• Without TMEM145: stereocilin lost from OHC stereocilia entirely
• Result: hair bundle disconnects from tectorial membrane → profound HL
• TMEM145 is transmembrane (OHC-expressed), stereocilin is extracellular

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41923617/
• DOI: https://doi.org/10.1016/j.neuron.2026.03.007

Connections

• [[STRC Mini-STRC Single-Vector Hypothesis]] — TMEM145 as anchor; mini-STRC must preserve binding interface
• [[STRC Stereocilia Bundle Mechanics Model]] — TM attachment depends on TMEM145–stereocilin link
• [see-also] [[Derstroff et al 2026 TMEM145 Paper]] — full analysis note

Numbers that matter (2026-04-23 audit)

SPR/BLI Kd for STRC × TMEM145 — NOT found in Derstroff 2026 Neuron.

Full text and PubMed abstract checked. Methods used for STRC–TMEM145 interaction:

• Co-immunoprecipitation (CoIP): qualitative, detected interaction, no Kd
• NanoSPD pull-down assay: qualitative, detected TMEM145–tubby interaction, no Kd
• AlphaFold Multimer: ipTM 0.79 and 0.71 (computational confidence, not binding affinity)

The model’s claim of Kd = 10 nM has no experimental SPR/BLI source in this paper. Supplementary contains only Fig S11 (band quantification from CoIP) — no biophysical binding measurements.

Parallel TMEM145 paper (different group): Nature Communications 2025, DOI 10.1038/s41467-025-67011-0 — also no SPR/BLI Kd reported. Also uses pull-down and co-IP.

Bottom line: No published SPR/BLI Kd for STRC × TMEM145 exists in the literature as of 2026-04-23. The 10 nM figure in the model is unsourced — either inferred from AF3 ipTM or fabricated. This is the most critical gap in Phase 1 parameterization.

What they found

Cryo-EM structure of PCDH15 — the lower tip link protein that connects stereocilia and gates mechanotransduction channels. PCDH15 forms parallel cis-dimers arranged as a right-handed double helix. This specific oligomeric architecture is required for tip link function: mutations that disrupt dimerization impair mechanotransduction even when the protein is expressed and present. Architecture = function, not just presence.

How this applies to mini-STRC

Stereocilin and PCDH15 are both extracellular proteins that form load-bearing connections between stereocilia structures. The PCDH15 finding generalizes: in the hair bundle, extracellular connector proteins don’t just need to be present — they need to self-assemble into precise supramolecular structures. Mini-STRC must therefore preserve not only folding stability but oligomerization capacity. The truncation at residues 700–1775 may disrupt self-assembly domains that exist outside this window. Computationally: predict STRC oligomeric state with AlphaFold-Multimer, identify self-assembly interfaces, verify they fall within the mini-STRC window.

Key numbers

• PCDH15 architecture: right-handed double helix parallel cis-dimer
• Dimerization domain mutations: impair mechanotransduction even with normal expression levels
• Technique: cryo-EM

Links

• bioRxiv: https://biorxiv.org/content/10.64898/2026.03.02.709101
• DOI: https://doi.org/10.64898/2026.03.02.709101

Connections

• [[STRC Mini-STRC Single-Vector Hypothesis]] — generalizes oligomerization constraint; AF3-Multimer STRC homodimer becomes required step

What they found

This paper reviews the current state of otoferlin (OTOF) gene therapy for infants and children with auditory neuropathy spectrum disorder (ANSD). After decades of research on hair cell regeneration, otoferlin gene therapy has emerged as the first successful inner ear gene therapy, with early clinical results demonstrating hearing restoration. The authors discuss the clinical framework for treating pediatric patients and raise the critical question of whether this early success will translate to other forms of genetic hearing loss. The paper contextualizes otoferlin gene therapy within the broader landscape of hearing loss treatment.

How this applies to mini-STRC

This is highly relevant to our program: (1) Otoferlin gene therapy is the furthest-advanced hearing loss gene therapy and serves as the regulatory and clinical precedent for all subsequent inner ear gene therapies, including mini-STRC. (2) OTOF (~6kb CDS) is similarly oversized for a single AAV, and the dual-AAV approach used for otoferlin is an important reference — though our mini-gene strategy takes a different approach to the payload problem. (3) The clinical endpoints, outcome measures, and patient selection criteria established for OTOF trials will likely inform FDA expectations for STRC gene therapy. (4) The paper’s discussion of whether success with otoferlin will replicate in other forms of genetic HL directly frames the question our program must answer. (5) Understanding the pediatric clinical framework is essential since DFNB16 patients are typically identified in childhood.

Key numbers

• Gene: OTOF (otoferlin) — ~6kb CDS, dual-AAV delivery approach
• Target population: infants and children with ANSD
• Status: early clinical success demonstrated, ongoing trials

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41979424/
• DOI: https://doi.org/10.1097/AUD.0000000000001773

Connections

• [[STRC Dual-Vector vs Single-Vector Transduction]] — OTOF dual-AAV clinical data used for v3 model recalibration (2.2x)
• [[STRC Mini-STRC Single-Vector Hypothesis]] — clinical ABR projection calibration
• [[STRC Anti-AAV Immune Response Model]] — pediatric AAV gene therapy precedent

What they found

SLC26A4 mutations are the second most common cause of hereditary hearing loss in many Asian countries (DFNB4). The authors demonstrated that postnatal AAV-mediated Slc26a4 gene therapy can improve hearing and preserve structural integrity of the inner ear in an animal model. Critically, they identified a therapeutic window for intervention — establishing that timing of gene delivery matters for efficacy. The study targeted the endolymphatic sac and cochlear lateral wall, showing these are viable sites for effective gene therapy intervention in DFNB4.

How this applies to mini-STRC

This is directly informative for our mini-STRC program in several ways: (1) It demonstrates postnatal gene therapy feasibility for a different hereditary hearing loss gene, validating the general approach. (2) The identification of a critical therapeutic window is important — we should consider whether STRC/DFNB16 has a similar window where intervention must occur before irreversible hair cell damage. (3) Their targeting of specific cochlear structures (endolymphatic sac, lateral wall) contrasts with our need to target outer hair cells specifically, but the delivery methodology and surgical approach may be informative. (4) As another oversized-gene adjacent program (SLC26A4 is ~2.3kb CDS, fitting AAV), their regulatory/preclinical path provides a reference for DFNB16 IND strategy.

Key numbers

• Gene: SLC26A4 (pendrin) — second most common cause of hereditary HL in Asian populations
• Target structures: endolymphatic sac and cochlear lateral wall
• Finding: critical therapeutic window identified for postnatal intervention

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/41701544/
• DOI: https://doi.org/10.1172/JCI193812

Connections

• [[STRC Mini-STRC Single-Vector Hypothesis]] — postnatal therapeutic-window calibration for clinical projections
• [[Adult Treatment Window STRC]] — comparable structural-protein gene therapy

What they found

Foundational paper identifying Ser133 as the PKA phosphorylation site on CREB and demonstrating that pCREB activates the somatostatin CRE. Established the PKA→CREB→CRE transcriptional cascade. Does not directly measure dephosphorylation kinetics of pCREB in cochlear hair cells.

Numbers that matter

• PKA phosphorylates CREB at Ser133: confirmed
• Dephosphorylation half-life of pCREB: t½ ~5–10 min in HeLa/COS cells (later work by the Montminy lab and others; not directly measured in this 1989 paper)
• t½ 5–10 min → k_dephos = 0.0012–0.0023/s
• The h05 model uses K_CREB_DEPHOS_S = 0.005/s (t½ 2.3 min) and cites this paper — a 2–4× discrepancy. The model dephosphorylates CREB faster than Gonzalez & Montminy’s data implies.
• Hair cell-specific CREB-P kinetics not measured anywhere; the 2–4× discrepancy may be acceptable given tissue differences, but should be flagged.

Fit to h05

The citation is real and correct for the CREB phosphorylation mechanism. However, the specific rate constant K_CREB_DEPHOS_S = 0.005/s is faster than the cited t½ range implies. This should be explicitly noted as a deviation. A sensitivity analysis varying this parameter by 4× is warranted.

Connections

• [[STRC Calcium Oscillation Acoustic Therapy]] — CREB dephosphorylation rate
• [part-of] [[calcium-oscillation]] (literature-params topic)

What they found

Characterization of calcineurin’s dual Ca²⁺-sensing mechanism: calmodulin binding to the CaM-binding domain increases catalytic rate and modulates Ca²⁺ response; calcineurin B (regulatory subunit) sets baseline Ca²⁺ sensitivity. Together they produce highly cooperative Ca²⁺ activation (Hill n = 2.8–3). Half-maximum Ca²⁺ concentration for activation: 0.6–1.3 µM depending on CaM concentration.

Numbers that matter

• Ca²⁺ for half-max calcineurin activation: 0.6–1.3 µM (model Kd_CaN = 500 nM is slightly below range — a minor underestimate)
• Hill coefficient: 2.8–3 (model n_CaN_Hill = 4.0 is slightly steeper than measured)
• CaM increases catalytic turnover rate ~10-fold at saturating Ca²⁺
• Calcineurin B shifts Ca²⁺ sensitivity to lower concentrations

Fit to h05

The RBM24 ODE model uses Kd_CaN = 500 nM and n_CaN_Hill = 4.0. Stemmer & Klee 1994 gives 600–1300 nM and Hill n = 2.8–3. Both values deviate modestly: Kd is ~20% low (conservative — makes CaN activate earlier), Hill n is ~35% steep (makes CaN more switch-like). Neither is a fatal error for the qualitative CaMKII:CaN decoding hypothesis.

Connections

• [[STRC Calcium Oscillation Acoustic Therapy]] — CaN kinetics in RBM24 ODE
• [see-also] [[1998-dolmetsch-calcium-oscillations-gene-expression]] — downstream of CaN
• [part-of] [[calcium-oscillation]] (literature-params topic)

What they found

In Jurkat T cells, Ca²⁺ oscillations (rather than sustained Ca²⁺ elevations) activate NF-κB, NF-AT, and Oct/OAP transcription factors with high efficiency and specificity. The key finding: different transcription factors have different threshold frequencies — NF-AT is activated at low frequencies; NF-κB requires higher frequency. CaMKII autophosphorylation accumulates preferentially with high-frequency oscillations; calcineurin (which dephosphorylates and activates NF-AT) is activated by lower-frequency sustained signals. This establishes the concept of frequency decoding of Ca²⁺ signals.

Numbers that matter

• Optimal oscillation frequency for NF-κB: ~0.2–0.5 Hz
• Calcineurin (NF-AT pathway) activated by low-frequency/sustained Ca²⁺: EC50 ~ 200–500 nM
• CaMKII switches to sustained activity at higher frequencies due to autophosphorylation memory
• Rate constants k_on_CaN = 0.3/s and k_off_CaN = 0.05/s in the h05 model are not directly from this paper — they are fitted values consistent with the conceptual framework

Fit to h05

The RBM24 ODE Phase 1 model is built on the Dolmetsch frequency-decoding concept. Dolmetsch 1998 confirms the biological plausibility of the CaMKII:CaN ratio as a Ca²⁺ frequency readout. However, the specific rate constants used in the model are estimates, not direct measurements from this paper. The paper does not study cochlear hair cells.

Connections

• [[STRC Calcium Oscillation Acoustic Therapy]] — conceptual basis for RBM24 ODE frequency decoder
• [see-also] [[1994-stemmer-klee-calcineurin-dual-calcium]] — calcineurin kinetics
• [part-of] [[calcium-oscillation]] (literature-params topic)

What they found

Patch-clamp recording from apical and basal OHCs in neonatal rat cochlea. Single-channel conductance of the MET channel increases from 145 pS (apical, low CF) to 210 pS (basal, high CF). Channels are highly Ca²⁺-selective (PCa/PNa ≈ 5–7). Ca²⁺ fraction of the total MET current is ~15% under physiological ionic conditions.

Numbers that matter

• Single MET channel conductance: 145–210 pS (apical to basal gradient)
• Model value g_MET = 150 pS is correct for apical OHC
• Ca²⁺ fraction of MET current f_Ca = 0.15: confirmed (Beurg 2006 Table 2)
• Channel open probability at rest: ~0.1–0.15 (before stimulation)
• Activation time constant: sub-millisecond
• Channel is mechanically gated via tip-link tension

Fit to h05

Confirms both g_MET = 150e-12 S and f_Ca = 0.15 in the RBM24 ODE. These are the two load-bearing MET parameters in the Ca²⁺ influx term. Both values are well-supported.

Connections

• [[STRC Calcium Oscillation Acoustic Therapy]] — g_MET and f_Ca
• [see-also] [[2019-fettiplace-kim-met-channel-review]] — broader review
• [part-of] [[calcium-oscillation]] (literature-params topic)

What they found

Comprehensive review of all mammalian adenylyl cyclase isoforms, with emphasis on Ca²⁺-regulated isoforms (AC1, AC3, AC8 as Ca²⁺/CaM-stimulated; AC5, AC6 as Ca²⁺-inhibited). For AC1 specifically: Ca²⁺/CaM stimulation is cooperative, EC50 for Ca²⁺ is in the 100–500 nM range (depending on assay conditions), and Vmax in reconstituted systems is in the µM/min range. Review synthesizes data from multiple labs on cAMP microdomain signaling and the spatial organization of AC isoforms with PKA scaffolds (AKAPs).

Numbers that matter

• AC1 EC50 for Ca²⁺ (via CaM): ~100–500 nM (assay-dependent; K_Ca = 150 nM used in h05 model is within this range)
• AC1 Vmax: ~2–5 µM cAMP/min in membrane preparations; Vmax = 2000 nM/s ≈ 120 µM/min is likely an overestimate unless at saturation in a concentrated cell compartment
• AC1 Hill coefficient for Ca²⁺ activation: 1.5–2.5 (n=2 in model is consistent)
• AC1 mRNA/protein expression: highest in neurons, cochlea, olfactory bulb; expressed in cochlear IHC and OHC (cited separately as Visel 1997/Vorobiova 1997 series)
• This is the canonical reference that should replace the phantom “Wu 2011” citation

Fit to h05

Willoughby & Cooper 2007 is the correct source for the AC1 kinetic rationale in the h05 pivot model. The model’s K_CA_AC1_NM = 150 nM falls within the reported EC50 range; AC1_VMAX_NM_S = 2000 nM/s needs to be checked against cell-level measurements (membrane prep values converted to cytoplasmic concentrations depend on AC1 copy number, which is not specified in the model).

Connections

• [[STRC Calcium Oscillation Acoustic Therapy]] — replaces phantom “Wu 2011”
• [see-also] [[2012-masada-ac1-ac8-calmodulin-kinetics]] — mechanistic follow-up
• [part-of] [[calcium-oscillation]] (literature-params topic)

What they found

Masada et al. used GST pull-down, mass spectrometry, and stopped-flow fluorescence to dissect how calmodulin (CaM) binds and activates adenylyl cyclase 1 (AC1) vs AC8. Key finding: AC1 is activated by an initial encounter complex between CaM and the IQ-like domain in the N-terminus, forming a 1:1 AC1:CaM complex at physiological Ca²⁺ concentrations. Activation is cooperative with Ca²⁺ occupancy of CaM’s C-lobe driving the stimulatory interaction. This establishes the mechanistic basis for AC1’s Ca²⁺/CaM sensitivity.

Numbers that matter

• AC1 activation by Ca²⁺/CaM: cooperative, driven by C-lobe Ca²⁺ binding
• Calmodulin binding stoichiometry: 1:1 (AC1:CaM)
• Mechanism: IQ-like domain in AC1 N-terminus is the primary CaM docking site
• AC8 uses a different (IQ-motif independent) mechanism — not applicable to h05
• K_Ca for AC1 activation not given as a single Kd number — determined by CaM affinity constants (~nM–µM range depending on Ca²⁺ saturation state)
• Note: the model’s K_CA_AC1_NM = 150 nM and AC1_VMAX_NM_S = 2000 nM/s are not directly tabulated here; these likely derive from earlier biochemical work (Willoughby & Cooper 2007 review synthesis).

Fit to h05

This is the AC1 kinetics reference for the pivot model. The “Wu 2011” citation in the scripts is a phantom — Masada 2012 is the real paper. However, Masada 2012 gives mechanistic insight and relative rates, not the exact nM/s Vmax values used in the ODE. Those values require supplemental justification from Willoughby & Cooper 2007 or explicit labeling as estimates.

Connections

• [[STRC Calcium Oscillation Acoustic Therapy]] — AC1 kinetics underpins pivot model
• [see-also] [[2007-willoughby-cooper-adenylyl-cyclase-review]] — companion canonical review
• [part-of] [[calcium-oscillation]] (literature-params topic)

What they found

Comprehensive review of MET channel biology: biophysics, gating, Ca²⁺ permeation, adaptation, molecular identity (TMC1/TMC2), and tonotopy. Synthesizes data on channel numbers per stereocilium (~2), bundle conductance (~1–6 nA peak), single-channel conductance gradients (apical ~100 pS to basal ~300 pS in mouse), and the role of Ca²⁺ in fast and slow adaptation.

Numbers that matter

• MET channels per stereocilium: ~1–2 (i.e., ~50–100 per OHC for a bundle of 50–70 stereocilia)
• n_channels = 134 in the h05 RBM24 model cites “Fettiplace 2017” — the correct citation is Fettiplace & Kim 2014 (this paper). The year “2017” is wrong. 134 channels is plausible for a mid-frequency OHC (67 stereocilia × 2 channels each).
• Ca²⁺ permeation fraction f_Ca: ~15% at physiological ionic conditions (consistent with Beurg 2006)
• Single-channel conductance: 145–210 pS apical-to-basal in rat (matches Beurg 2006)

Fit to h05

This review is the correct source for n_channels = 134 (replacing the phantom “Fettiplace 2017”). The value itself is plausible per the review’s channel density data. Note: the review itself notes variability in channel number across cochlear position and species; 134 should be treated as an order-of-magnitude estimate, not a precise measurement.

Connections

• [[STRC Calcium Oscillation Acoustic Therapy]] — MET channel count (n_channels)
• [see-also] [[2006-beurg-met-channel-conductance]] — g_MET value
• [part-of] [[calcium-oscillation]] (literature-params topic)

What they found

Crystal structures of WH2 domains from WASP, WAVE2, and WIP in complex with G-actin (ternary complex with DNase I as crystallization cofactor). Parallel ITC biochemistry comparing WH2 vs thymosin-β (Tβ) binding to actin. First structural evidence that short WH2 domains (~17 aa) can coexist with intersubunit contacts in F-actin; the canonical G-actin binding site (barbed-end groove between subdomains 1 and 3) is partly accessible on the filament surface.

Numbers that matter

WH2 G-actin binding affinities — ITC, 25°C, G-buffer (2 mM Tris pH 7.5, 0.2 mM CaCl₂, 0.2 mM ATP):

• Exact Kd values are in Table 2 (SI, not open-access). Narrative in text only gives relatives:
  • WAVE WH2: strongest binder (5× higher affinity than WASP WH2)
  • WASP WH2: weakest binder in series
  • All WH2 domains: ~10× higher affinity than Tβ domains
• Since Tβ4 Kd = 1 μM (Husson 2010, Xue 2014), this implies WH2 Kd range ~50–200 nM
• WASP WA domain (WH2 + C region together): Kd = 0.6 μM from Rohatgi 2000 (cited)
• Isolated WASP WH2 (V domain alone, from Padrick/Kim 2011 citing this paper): Kd ≈ 3.1 μM

CRITICAL NOTE on “100 nM” in model: The ~10× boost vs Tβ4 is for construct-dependent WH2 variants. The isolated WASP WH2 V-domain alone is ~3 μM, not 100 nM. The 100 nM figure likely applies to long WH2 constructs (WIP-type) or in-context domains with favorable electrostatics (WAVE with Arg-Arg in LKKT gives extra salt bridges). Model parameter WH2_KD_GACTIN_M = 200 nM is near the favorable end; range is 50 nM–3 μM depending on construct.

WH2 × F-actin side-binding — NOT MEASURED in this paper:

• The paper shows WH2 binding site (barbed-end groove, subdomains 1/3) is ALSO the site of longitudinal actin–actin contacts in the filament long-pitch helix
• Short WH2 domains (~17 aa) can “coexist with intersubunit contacts in F-actin” structurally, but this means nucleation (WH2 templating monomers along a strand), not side-binding of preformed filaments
• No co-sedimentation or F-actin binding Kd measured

Tβ4 × F-actin: weak cooperative binding, Kd = 5–10 mM (cited from Husson 2010) This is 3 orders of magnitude weaker than G-actin binding — the C-terminal α-helix in Tβ4 interferes with filament contacts, but the same region being absent in WH2 doesn’t mean WH2 side-binds better; it just means WH2 doesn’t lock monomer as completely.

Why WH2 × F-actin side-binding is structurally disfavored

The N-terminal amphipathic helix of WH2 binds in the cleft between actin subdomains 1 and 3. In the filament (Holmes model), this same cleft is occupied by longitudinal actin–actin contacts along the long-pitch helix — it is partially buried. Residual surface accessibility does not constitute a high-affinity independent binding site. No evidence that WH2 can bind to the side of a preformed filament with affinity < 1 mM. The model’s WH2_KD_FACTIN_M = 5 μM is optimistic; the real value is likely 1–100 mM or unmeasurable.

Connections

• [source] [[STRC H09 WH2 F-actin Bundling Hypothesis]] — provides structural rationale and Kd reference range for G-actin binding; establishes F-actin side-binding as unmeasured and structurally disfavored
• [[Hydrogel Phase 4d F-actin Bundling Model]] — WH2_KD_GACTIN_M defensible at 50–200 nM for favorable WH2 constructs; WH2_KD_FACTIN_M = 5 μM is speculative/optimistic
• [[WH2 Actin Binding Mechanism]] — barbed-end groove binding

What they found

Dominguez’s authoritative review arguing that tandem WH2 domain repeats (Spire, Cobl, VopL/VopF, Sca2) are insufficient to drive efficient actin nucleation on their own — other domains or cooperating proteins are always required. Mechanistic review of WH2 family across all WH2-containing nucleators.

Numbers that matter

WH2 × G-actin Kd (from text):

• “Most WH2 domains bind ATP-actin with low micromolar and, in some cases, nanomolar affinity”
• No specific numbers in text; references primary papers (Chereau 2005, Rebowski 2010, etc.)
• The “nanomolar” cases likely refer to full-context domains (WAVE with Arg-Arg in LKKT, or long WH2 constructs with favorable electrostatics)

WH2 × F-actin side-binding:

• NOT measured in this review; no Kd reported
• Review discusses that tandem WH2 domains can coexist with F-actin contacts during nucleation (i.e., they template filament-like arrangements of monomers) but this is mechanistically distinct from side-binding a preformed filament

Key mechanistic point:

• WH2 binds the barbed-end groove of G-actin (subdomains 1/3 cleft)
• In F-actin, this same surface is involved in longitudinal actin–actin contacts along the long-pitch helix — partially inaccessible for independent high-affinity side-binding
• The remaining accessible surface of F-actin does not constitute a structurally coherent independent WH2 binding site

Significance for h09

This review directly supports the concern flagged in the model: WH2 is canonically a G-actin binding motif; its use as an F-actin cross-linker/bundling agent lacks structural precedent. The review explicitly states WH2 is “necessary but insufficient” for nucleation even in tandem — making it further implausible as a standalone F-actin bundler at μM affinity.

Connections

• [[STRC H09 WH2 F-actin Bundling Hypothesis]] — Dominguez is the domain expert; this review is the most comprehensive source on WH2 biophysics
• [[Hydrogel Phase 4d F-actin Bundling Model]] — supports flagging WH2_KD_FACTIN_M = 5 μM as speculative/unsupported
• [[Chereau 2005 WH2 Actin PNAS]] — same author’s original structural paper

What they found

Comprehensive review/research article from the Carlier lab (MF Carlier is the authoritative actin dynamics biochemist) on the β-thymosin/WH2 module. Used chimeric protein engineering, actin polymerization assays, ITC, NMR, and SAXS to dissect how a single βT/WH2 domain can function as G-actin sequesterer, filament barbed-end deliverer, nucleator, or severer depending on C-terminal sequence context.

Numbers that matter

Thymosin-β4 (Tβ4) × G-actin:

• Kd = 1 μM for ATP-actin (moderate affinity, the reference baseline)
• Kd = 80–100 μM for ADP-actin (50–100× weaker — nucleotide-state dependent)

Tβ4 × F-actin (side-binding):

• Kd = 5–10 mM (weak cooperative binding at [Tβ4] > 20 μM)
• This is 5,000–10,000× weaker than G-actin binding
• Mechanism: C-terminal α-helix of Tβ4 interferes with filament contacts

Implication for WH2 × F-actin:

• WH2 lacks the C-terminal α-helix that interferes with F-actin in Tβ4
• But WH2 canonical binding site (barbed-end groove) is also involved in longitudinal actin–actin contacts in filament — still not a clean side-binding site
• No direct measurement of WH2 × F-actin Kd reported in this paper
• The paper does not support the existence of a low-μM WH2 × F-actin interaction

Context for model parameter WH2_KD_FACTIN_M = 5 μM:

• Based on all available literature, 5 μM would be VERY optimistic for a WH2 × F-actin side-binding event. It is possible if constructs are specifically engineered, but not for a canonical isolated WH2 peptide. True value is likely ≥ 1 mM or unmeasurable.

Connections

• [source] [[STRC H09 WH2 F-actin Bundling Hypothesis]] — Tβ4/WH2 Kd values and F-actin binding data; shows F-actin side-binding is mM-range or absent
• [[Hydrogel Phase 4d F-actin Bundling Model]] — key constraint on WH2_KD_FACTIN_M: the 5 μM value in model is 1000× more optimistic than Tβ4 literature
• [[Actin Treadmilling Stereocilia]] — Carlier lab is primary authority on actin dynamics constants

What they found

Highly sensitive single-cell proteomics (SCoPE-MS) applied to E15 chick utricle hair cells. Quantified relative molar fractions of ~1500 proteins per cell type. Provides the most precise molecular-level census of hair cell protein composition, including actin copy numbers that can be used to estimate local [F-actin] in stereocilia.

Numbers that matter

Actin copy numbers:

• E15 chick utricle hair cell: ~15,000,000 total actin molecules per cell
• Per stereocilium: ~200,000 actin molecules (72 stereocilia counted per utricle hair cell)
• E20 (later development): ~400,000 actin molecules per stereocilium (estimate)
• ACTG1 group (total actin) = 0.043 ± 0.001 molar fraction in FM1-43high cells

Derived local [F-actin] in stereocilia (calculation, not in paper):

• OHC stereocilium: ~200 nm diameter, ~3 μm long → volume ≈ 9.4 × 10⁻¹⁷ L
• 200,000 molecules / (6.022×10²³ × 9.4×10⁻¹⁷) ≈ 3,500 μM total actin
• Assuming 90% filamentous → [F-actin] ≈ 3,150 μM ≈ 3.15 mM
• Cross-check with ~100 nm diameter: [actin] ≈ 14,000 μM (extreme upper bound, likely for developing stereocilia with denser core)
• Practical working estimate: [F-actin] in mature stereocilium core = 1–5 mM

Critical implication for WH2 × F-actin bundling model:

• [F-actin] is 1–5 mM — this is far above any plausible WH2_KD_FACTIN_M
• If WH2_KD_FACTIN_M = 5 μM (as in model), fractional occupancy at 1 mM [F-actin] would approach 100% — but this is irrelevant if the binding event doesn’t exist in the first place
• The high local [F-actin] is not a bottleneck; the question is purely whether WH2 can side-bind at all

STRC/stereocilin: Not quantified in this study (chick utricle dataset; STRC is mammalian cochlear-specific in its critical role).

Connections

• [[Hydrogel Phase 4d F-actin Bundling Model]] — provides 200,000 actin/stereocilium baseline for [F-actin] calculation
• [[STRC Normal OHC Concentration Parameter]] — STRC_NORMAL_OHC_M needs own source; this paper doesn’t cover STRC
• [[Barr-Gillespie Lab Stereocilia Atlas]] — anchor paper for quantitative hair cell proteomics

What they found

The benchmark actin polymerization kinetics paper. Measured rate constants for ATP-actin and ADP-actin association/dissociation at both barbed and pointed ends using Limulus sperm acrosomal processes as nuclei and electron microscopy. Defines critical concentrations for both ends and both nucleotide states.

Numbers that matter

ATP-actin:

• Barbed end k_on = 11.6 μM⁻¹s⁻¹, k_off = 1.4 s⁻¹, Kd (critical conc.) = 0.12 μM
• Pointed end k_on = 1.3 μM⁻¹s⁻¹, k_off = 0.16 s⁻¹, Kd = 0.12 μM

ADP-actin:

• Barbed end k_on = 3.8 μM⁻¹s⁻¹, k_off = 7.2 s⁻¹, Kd = 1.9 μM
• Pointed end k_on = 0.16 μM⁻¹s⁻¹, k_off = 0.27 s⁻¹, Kd = 1.7 μM
• Overall critical concentration for ADP-actin polymerization = 1.8 μM

Physiological context:

• Cells maintain ~50–200 μM total actin; ~25–100 μM unpolymerized (monomer pool)
• Critical concentration at barbed end (ATP-actin) = ~0.1 μM
• Treadmilling occurs because pointed-end Kd (1.7 μM, ADP) >> barbed-end Kd (0.12 μM, ATP)

Relevance to stereocilia model:

• Stereocilia actin is largely stable (slow turnover along shaft, active only at tips)
• The actin treadmilling model applies at stereocilia tip region (~0.5 μm zone)
• Local [F-actin] in OHC stereocilium: ~3,500 μM (200 nm dia, 3 μm long, 200,000 molecules) — far above critical concentration; essentially all actin in shaft is filamentous

Connections

• [[Hydrogel Phase 4d F-actin Bundling Model]] — barbed/pointed end rate constants for actin assembly in stereocilia context
• [[Actin Treadmilling Stereocilia]] — sets rate constants used in tip-turnover estimates
• [[STRC H09 WH2 F-actin Bundling Hypothesis]] — benchmark for comparison with WH2 on/off rates

What they found

Comprehensive authoritative review of actin biochemistry and all major actin-binding protein families. Covers sequestering proteins (profilin, thymosin-β4), nucleators (Arp2/3, formins, WH2-domain proteins), elongators, cappers, severers, and cross-linkers. Full-text open access on PMC.

Numbers that matter

Profilin × G-actin:

• Kd = 0.1 μM for ATP-actin — sets benchmark for high-affinity sequestration

Thymosin-β4 × G-actin:

• Micromolar affinity (Kd ~1 μM); cellular concentration > 100 μM → sequesters large monomer pool

Actin-binding domains (ABDs, calponin-homology type) × F-actin:

• “ABDs typically have relatively low affinity for actin filaments (Kd ~10 μM)”
• This is the Pollard consensus for ABD-class F-actin binders
• WH2 is NOT an ABD (different fold, different binding site) but the comparison is instructive: even canonical F-actin cross-linkers have Kd ~10 μM

WH2 × actin (from this review):

• WH2 binds in the barbed-end groove (actin subdomains 1/3 cleft)
• Functions: deliver actin to barbed end, template nucleation via tandem repeats
• No Kd measurements given for WH2 directly; references Chereau 2005

Barbed end kinetics (ATP-actin):

• k_on ≈ 10 μM⁻¹s⁻¹, k_off ≈ 1 s⁻¹, critical concentration ≈ 0.1 μM

Total cellular actin:

• 50–200 μM in eukaryotic cells
• ~half unpolymerized at any moment (maintained by profilin + thymosin-β4 pool)

Critical concentration for ADP-actin = 1.8 μM (Pollard 1986; consistent)

Relevance to hydrogel model

The “Kd ~10 μM for ABD-class cross-linkers” is the best available proxy for WH2 × F-actin side-binding in the absence of direct measurement. The model’s WH2_KD_FACTIN_M = 5 μM is therefore in the right ballpark as an optimistic estimate but has no primary measurement backing. A more conservative estimate would be 10–100 μM or higher.

Connections

• [[Hydrogel Phase 4d F-actin Bundling Model]] — Kd ~10 μM for ABD-class F-actin binders is best proxy for WH2_KD_FACTIN_M
• [[STRC H09 WH2 F-actin Bundling Hypothesis]] — WH2 mechanism, profilin competition, physiological actin monomer concentrations
• [[Pollard 1986 Actin Rate Constants]] — contains same kinetics plus broader context

What they found

Pioneering electron microscopy and optical diffraction study of actin filament organization in stereocilia of alligator lizard cochlea. Established that stereocilia contain paracrystalline arrays of actin filaments with the long-axis crossover points in near-perfect register. The primary source for inter-filament spacing in mature cochlear stereocilia.

Numbers that matter

Inter-filament spacing:

• Generally cited as “~10 nm” inter-filament distance in stereocilia from this paper
• Filaments packed in scalloped rows (“festooned profile” in transverse section) — NOT simple hexagonal packing; instead a liquid-to-hexagonal hybrid
• Later work refines: Krey 2016 (Pls1 paper) measures 7.9–9.7 nm center-to-center by FFT analysis of freeze-substituted EM; espins give ~12 nm spacing, fascin ~6–8 nm
• Model’s STEREOCILIA_INTER_FILAMENT_NM = 12.0 corresponds to espin-crosslinked arrays specifically; native OHC core (plastin-1 + fimbrin) is ~9–10 nm

What the spacing tells us about bundling geometry:

• Filament center-to-center ~9–12 nm → outer surface distance ~2–5 nm (actin diameter ~7 nm)
• For a cross-linker to span two filaments it needs to bridge ~2–5 nm — this is geometrically plausible for a peptide (~3–5 nm extended)
• WH2 motif (~17–32 aa) = ~6–12 nm extended chain; could geometrically span two filaments IF it binds both with sufficient affinity (the key unproven assumption)

Source quality note:

• Full text behind Rockefeller University Press paywall (not PMC open access)
• Spacing value ~10 nm consistently cited across review literature as “Tilney et al. 1980”
• Krey 2016 (PMC5119939) is the modern measurement with EM + FFT, cites Tilney as historical reference; use Krey 2016 as primary quantitative source for spacing

Connections

• [[Stereocilia Actin Architecture]] — first measurement of inter-filament spacing in cochlear hair cells
• [[Krey 2016 Plastin Stereocilia Spacing]] — modern FFT-EM measurement; Tilney 1980 = ~10 nm historical, Krey 2016 = 7.9–9.7 nm
• [[Hydrogel Phase 4d F-actin Bundling Model]] — STEREOCILIA_INTER_FILAMENT_NM parameter should use 9–10 nm for OHC native, 12 nm if espin-dominated

What they found

First proof-of-concept STRC gene therapy in Strc-/- mice. Dual AAV (PHP.B serotype) delivering split STRC injected at P1. Confirmed STRC protein at OHC stereocilia tips, structural restoration, and partial hearing recovery.

Key numbers

• Injection age: P1 (postnatal day 1) — NEONATAL ONLY
• OHC transduction: 59% of OHC bundles showed distinct STRC localization
• Hearing recovery: 50-60 dB improvement in best responders; thresholds as low as 30 dB
• Bundle morphology: 61-65% of bundles showed normal organization and top connectors
• TM attachment: Restored — attachment crowns reformed after STRC delivery

Critical finding on TM re-attachment

Even in full STRC knockout mice, where TM imprints were ABSENT (Verpy 2011), gene therapy at P1 successfully restored TM-OHC attachment crowns.

This means: TM contact can reform when STRC is provided, even starting from zero. The TM retains the structural capacity to embed stereocilia tips if STRC is present.

For adult treatment, this is cautiously positive — but untested.

Age limitation (explicit from paper)

“Because injection in adult mouse cochleas is technically challenging, we focused on vector delivery during the first postnatal week, and whether dual vector delivery of STRC into mature cochleas is a viable treatment option remains to be determined.”

This is the critical open question. No adult data exists.

Comparison to Iranfar 2026

Iranfar (CTM 2026) is a more recent replication using AAV9-PHP.eB with similar results. Both papers demonstrate neonatal STRC restoration. Neither tests adult.

Relevance to Misha

The TM re-attachment finding is encouraging: even full KO cochlea can reform TM contact. Misha’s partial STRC likely maintained more TM contact than KO. His cochlea is probably in a BETTER state for rescue than the KO mice that were treated.

The gap: no one has tried delivery at P30+, P60+, or adult in STRC-null mice. This is the experiment the field needs — and the data Misha’s program needs.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/34910522/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC8673757/
• DOI: https://doi.org/10.1126/sciadv.abi7629

Connections

• [[STRC Stereocilia Bundle Mechanics Model]] — TM crown reformation from full-KO baseline at P1
• [[STRC Dual-Vector vs Single-Vector Transduction]] — published baseline (59% OHC, 50–60 dB gain) for the dual-vector arm
• [[Adult Treatment Window STRC]] — only existing STRC gene therapy data, P1 only

What they found

Measured human scala tympani (ST) volume from 30 micro-CT cadaveric temporal bone datasets using 3D segmentation (Slicer 4.10.2 or Stradwin 6.1, 24–30 μm isotropic voxels). Primary purpose was cochlear implant surgical planning, but yields the best direct human ST volume measurement.

Numbers that matter

Implication for model: The model uses PERILYMPH_VOL_UL = 70 μL for “human cochlea.” This is inconsistent with direct anatomical measurement:

• ST alone: 34.2 μL mean
• Total perilymph (ST + SV), from Ekdale 2016 μCT: ~93 μL
• 70 μL is not directly supported by either measurement; it may be a rough midpoint or from an older anatomical estimate not traced in the model

The drug-accessible compartment for RWM-delivered agents is primarily ST (drug enters basally, distributes apically by diffusion). Using 93 μL would underestimate concentration; 34 μL would overestimate. 70 μL has no traceable primary citation.

Access

PMC open access: https://pmc.ncbi.nlm.nih.gov/articles/PMC8514755/

Connections

• [source] [[hydrogel_phase4e_cochlear_pkpd]] — PERILYMPH_VOL_UL discrepancy; 34 μL (ST only) vs 93 μL (total) vs 70 μL (model, unsourced)
• [[STRC Hypothesis Ranking]] — h09 PK volume parameter

What they found

Tonotopic gradients of hair-bundle mechanics in rat cochlea (apical region, 1–4 kHz, P7–P10). Measured total bundle stiffness K_HB, gating-spring contribution K_GS, and pivot contribution K_SP. This is a tip-link/gating-spring paper — it does NOT measure horizontal top connector (HTC) stiffness.

Key numbers (directly from paper)

OHC total bundle stiffness K_HB (rat, apical half, P7–P10, fluid-jet stimulation):

IHC total bundle stiffness K_HB (same preparation):

Gating-spring stiffness K_GS (single spring per tip-link complex):

• OHC at 1 kHz: 1.3 ± 0.4 mN/m
• OHC at 4 kHz: 3.7 ± 0.7 mN/m
• Bundle stiffness increased ~240% over 2 octaves (1–4 kHz) for OHC

Tip-link tension at rest:

• 1 kHz: ~5 pN
• 4 kHz OHC: 34 ± 8 pN

Species/age: Sprague-Dawley rat, P7–P10 (pre-hearing, juvenile) Method: calibrated fluid-jet stimulation

Critical correction for model

The model cites K_HTC_PN_PER_NM = 7.5 (approx, Tobin 2019) — this citation is WRONG.

Tobin 2019 measures tip-link complex stiffness (K_GS + K_SP). It contains no data on horizontal top connectors (HTCs). The paper’s formula is K_HB = K_GS + K_SP — no HTC term appears at all. HTC stiffness is measured in Dulon 2019 (Science Advances) and modeled in Kozlov 2011 (Nature).

Correct HTC stiffness sources:

• Dulon 2019 net contribution: ~3.07 pN/nm per bundle, ~0.16 pN/nm per link (mouse apical, P13–P15)
• Kozlov 2011 (Nature, bullfrog sacculus model): 20 mN/m = 20 pN/nm aggregate for whole bundle

The model’s WT_BUNDLE_STIFFNESS_PN_PER_NM = 1400 has no basis in this paper. Tobin 2019 reports 2.5–8.6 pN/nm for juvenile rat OHC. Even adult basal-turn OHC would not reach 1400 pN/nm (literature max is ~50–100 mN/m = 50–100 pN/nm for highest-frequency locations). The 1400 value is also dead code — never referenced in gate3_bundle_stiffness().

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/30932811/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC6464607/
• DOI: https://doi.org/10.7554/eLife.43473
• PDF: ~/BookLibrary/incoming/tobin-2019-elife-43473-tiplink-stiffness.pdf

Connections

• [[STRC Stereocilia Bundle Mechanics Model]] — K_HTC citation wrong; this paper measures tip-link/gating-spring stiffness, not HTC stiffness
• [[STRC Stereocilia Bundle Mechanics Model]] — real OHC bundle stiffness range: 2.5–8.6 pN/nm (juvenile rat 1–4 kHz), NOT 1400 pN/nm

What they found

First direct measurement of HTC spring contribution to OHC hair bundle stiffness using non-contact acoustic FM-AFM. Measured at P9–P15 in Strc-/-, Strc+/-, and Strc-/-/Tecta-/- mice, isolating the specific contribution of horizontal top connectors.

Key numbers (directly from paper)

• Reduction: ~60% in OHC bundle stiffness without HTCs
• IHC stiffness is UNCHANGED — HTCs are OHC-specific, IHC are unaffected
• Measurements at P13–P15 (post HTC maturation)

Unit conversion for model

• 5.12 pN/nm = 5.12 mN/m (per bundle, ~20 stereocilia)
• HTC contribution: 5.12 − 2.05 = 3.07 mN/m per bundle
• Per stereocilium: ~0.154 mN/m
• Per HTC link (19 links in 20-stereocilia chain): ~0.16 mN/m per link

This is ~5.7× larger than my original k_HTC estimate (0.028 mN/m). Model needs recalibration with these values.

Developmental timeline

Top connector maturation coincides with hearing onset (P12–P15). Significant increase in OHC (not IHC) stiffness from P12→P15 = HTC maturation window. No adult measurements — limitation of the study.

Relevance to model

Provides ground-truth calibration for k_HTC in stereocilia_bundle_mechanics.py. The 60% reduction figure is confirmed with direct measurement, not inferred. Critical: IHC stiffness unchanged = tip-link parameters remain valid for HTC-free state.

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/30801007/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC6382404/
• DOI: https://doi.org/10.1126/sciadv.aat9934

Connections

• [supports] [[STRC Stereocilia Bundle Mechanics Model]] — direct k_HTC measurement confirms 60% stiffness loss

Numbers that matter (2026-04-23 audit)

Direct AFM measurements confirmed (mouse, apical turn, P13–P15, Strc+/−/Tecta−/− vs Strc−/−/Tecta−/−):

Developmental series (apical, with HTCs): P9 = 0.92 pN/nm → P15 = 5.4 pN/nm (~6× increase).

Model parameter cross-check:

• WT_BUNDLE_STIFFNESS_PN_PER_NM = 1400 — NOT supported. Dulon 2019 reports 5.12 pN/nm (apical OHC, juvenile mouse). Even scaling to adult basal turn (×5–10×) gives ~25–50 pN/nm, still 28–56× below 1400. The 1400 value is ~150–700× above any published OHC bundle measurement and has no literature basis. The constant is also never referenced in gate3_bundle_stiffness() — it is dead code.
• K_HTC_PN_PER_NM = 7.5 — Dulon 2019 gives net HTC contribution of 3.07 pN/nm per bundle (~20 stereocilia, ~19 HTC links in a chain) → ~0.16 pN/nm per HTC link. The Kozlov 2011 Nature model uses 20 pN/nm aggregate for bullfrog sacculus. The 7.5 value sits between these two extremes but the Dulon paper does NOT report a per-link value; the model’s 7.5 is not directly derivable from this paper.
• HTC_PER_STEREOCILIUM = 6 — not reported in Dulon 2019. No TEM connector count data.
• HTC_SPACING_NM = 8.0 — not reported in Dulon 2019.

No SI data was parsed — PDF of Dulon 2019 not available (Science.org returned 403). Existing note may have been written from PMC full text (accessible). SI may contain connector counts — not yet checked.

What they found

Comprehensive review of pharmacokinetic principles for intratympanic drug delivery. Uses dexamethasone-phosphate and triamcinolone-acetonide as examples to illustrate how molecular properties (logP, TPSA) govern RWM crossing and perilymph distribution.

Key argument: dexamethasone-phosphate is poorly suited for local ear therapy because its polar, water-soluble form crosses lipid membranes poorly and is eliminated rapidly from basal perilymph without reaching apical regions. Triamcinolone-acetonide has more favorable pharmacokinetics.

Key findings — Middle ear clearance

Figure 4 shows time-course of drug concentration decline in the middle ear niche after a single drop application:

• Gentamicin: fell to 46% of initial at 83 min (t½ ≈ 83 × ln2 / ln(100/46) ≈ 70–75 min)
• Dexamethasone-phosphate: fell to 10% at 93 min (t½ ≈ 40 min, extremely fast)

No single mucociliary half-life is given for all substances — it is drug-dependent and dominated by Eustachian tube drainage, not just ciliary clearance. The model’s K_CLEAR_MIDDLE_EAR = 0.35/h (t½ = 2 h) is slower than all measured values in this paper.

Key findings — RWM permeability

No numerical permeability coefficients vs. MW table. Paper discusses lipophilicity (logP) and TPSA as drivers, not MW directly. Substances with high TPSA cross lipid membranes poorly. Does NOT provide a permeability value for 14 kDa peptides.

Key findings — Cochlear aqueduct / perilymph clearance

Mentions “cochlear aqueduct provides the outlet for fluid” and CSF-perilymph exchange is important in rodents. No quantitative human aqueduct clearance rate provided.

Critical gap for model

No explicit t½ for middle ear clearance of unprotected peptide. No RWM permeability value for 14 kDa peptide. Perilymph clearance half-time not given for humans.

Numbers that matter

Supplementary materials

None identified.

Access

PMC open access: https://pmc.ncbi.nlm.nih.gov/articles/PMC6133771/

Connections

• [source] [[hydrogel_phase4e_cochlear_pkpd]] — K_CLEAR_MIDDLE_EAR parameter; this paper does NOT support 2h half-life
• [[STRC Hypothesis Ranking]] — h09 cochlear PK basis

What they found

High-resolution μCT-based finite element model of the human cochlea using temporal bone specimen (6.7 and 3 μm resolution). Derived total cochlear perilymph volume from anatomical measurement.

Numbers that matter

Does not subdivide into ST vs SV individually. Clinical perilymph samples collected from patients range from 1–4 μL (a tiny fraction of total).

Access

PMC open access: https://pmc.ncbi.nlm.nih.gov/articles/PMC5292244/

Connections

• [source] [[hydrogel_phase4e_cochlear_pkpd]] — provides 93 μL total perilymph as upper bound for PERILYMPH_VOL_UL
• [[2021-dhanasingh-scala-tympani-volume]] — together bracket the range: 34 μL (ST only) to 93 μL (total)

What they found

Used FITC-dextran (large MW tracer — slower diffusion, less tissue uptake) to characterize perilymph homeostasis mechanisms. Found perilymph clearance is dominated by CSF exchange via cochlear aqueduct. All studies in guinea pig.

Numbers that matter

Implications for model

The K_PERILYMPH_CLEAR = 0.35/h (t½ = 2h) in the model is in the right ballpark for slow-clearing large molecules (230 min = 3.8h → 0.18/h) but faster than what this paper shows for high-MW tracers. For a 14 kDa peptide (smaller than FITC-dextran but larger than steroids), the actual clearance would be between these extremes.

CSF inflow rate of 30 nL/min corresponds to ~1.8 μL/h. In 70 μL perilymph, this gives a dilution half-life of ~27 h from CSF turnover alone — suggesting CSF exchange is NOT the dominant clearance route for most drugs within a 48h window; tissue uptake and aqueduct bulk flow dominate.

Access

PMC open access: https://pmc.ncbi.nlm.nih.gov/articles/PMC4417094/

Connections

• [source] [[hydrogel_phase4e_cochlear_pkpd]] — K_PERILYMPH_CLEAR; 2h model value is between guinea pig fast-MW (60 min) and dextran-slow (230 min) — plausible but not directly measured for 14 kDa peptide
• [[2001-salt-ma-quantification-rwm-permeability]] — context for ST clearance parameter

What they found

Extended characterization of Strc-/- mouse cochlea at multiple ages including late (5 months). Critical data on tectorial membrane morphology and OHC structural/survival status over time.

Key findings — OHC survival timeline

Critical distinction: basal (high-freq) OHCs degrade faster than apical (low-freq). At 5 months: apical OHCs largely preserved, basal OHCs significantly damaged.

Key findings — Tectorial membrane

TM imprints ABSENT throughout development and adulthood in Strc-/- mice:

“Stereocilia imprints on the lower surface of the tectorial membrane were not observed in Strc-/- mice”

This means: without STRC, the TM-OHC attachment crowns never form — not that they form and then degrade. The TM never embeds the OHC stereocilia tips.

However:

• Hensen’s stripe present and unaffected
• TM remains attached to spiral limbus normally
• TM gross structure maintained

Critical implication for Misha

Misha has PARTIAL STRC (VUS allele produces some protein). His TM likely DID form partial attachment crowns and imprints at birth. Unlike the full KO, he had some OHC-TM coupling. His question is not “did TM contact ever exist” (it probably did, partially) but “is it maintained at age 5?”

The apical vs basal gradient matters for Misha specifically:

• Left ear (milder, ~40-50 dB) → more apical involvement → better preserved OHCs
• Right ear (worse, ~50-65 dB) → more basal damage → more degradation

Links

• PubMed: https://pubmed.ncbi.nlm.nih.gov/21165971/
• PMC: https://pmc.ncbi.nlm.nih.gov/articles/PMC3375590/
• DOI: https://doi.org/10.1002/cne.22509

Connections

• [[STRC Stereocilia Bundle Mechanics Model]] — age-progression and TM-imprint absence; apical OHCs better preserved
• [[Adult Treatment Window STRC]] — bundle degeneration timeline supports neonatal intervention urgency

Numbers that matter (2026-04-23 audit)

Connector geometry — NOT reported in this paper. Verpy 2011 is a localization/developmental study only. No TEM connector counts, no HTC spacing measurements, no stiffness data.

• The paper confirms HTC (“horizontal top connectors”) are absent in Strc-/- and fail to form (not degraded) throughout development.
• “Resolution of immunogold labeling not sufficient to enable precise localization of stereocilin within the bands of top connectors.”
• Source for HTC_PER_STEREOCILIUM = 6 is NOT Verpy 2011. No count data here.
• No SPR/BLI, no stiffness numbers.

Verdict: Cannot validate any of the four target model parameters from this paper.

What they found

Proteomics of mouse perilymph vs CSF by LC-MS/MS. Identified 228+ proteins. Key finding for our purpose: perilymph is enriched in protease inhibitors, not active proteases.

Numbers that matter

Key insight for K_PROTEOLYSIS: Perilymph is protease-inhibitor-rich. The dominant serpins (serpin a1d, a1a, a1e) would strongly suppress serine protease-mediated degradation of an unprotected peptide. However:

• Cathepsin D was identified (lysosomal aspartyl protease — released during cell stress)
• No functional protease activity assays were performed
• No peptide half-life measured

The 30-min half-life for K_PROTEOLYSIS = 1.4/h in the model is conservative in the correct direction (perilymph is less proteolytic than serum) but the actual value for a 14 kDa peptide in perilymph is unmeasured. CSF stability data (extrinsic trypsin at 0.04 μM is fully inhibited by endogenous CSF inhibitors) suggests peptides may actually be stable for hours in perilymph.

Access

PMC open access: https://pmc.ncbi.nlm.nih.gov/articles/PMC2940114/

Supplementary materials

None noted.

Connections

• [source] [[hydrogel_phase4e_cochlear_pkpd]] — K_PROTEOLYSIS justification; paper supports that perilymph is protease-inhibitor-rich, but does NOT give t½ for a 14 kDa peptide
• [[STRC Hypothesis Ranking]] — h09 proteolytic stability in cochlea

What they found

Primary quantitative source for RWM permeability and scala tympani clearance. Used TMPA (trimethylphenylammonium — a low-MW ion) as marker in guinea pig. Applied to intact RWM for 90 min; measured distribution with ion-selective microelectrodes at turns 1 and 2.

After 90 min:

• Turn 1 (1.4 mm from base): 330 ± 147 μM (n=8)
• Turn 2 (7.5 mm from base): 15 ± 33 μM (n=5)

Simulation fitting yielded the canonical parameter set for the WUSTL Cochlear Fluids Simulator default (guinea pig).

Numbers that matter

Critical limitation: TMPA is a small ion (MW ~166 Da). The paper does NOT provide permeability for 14 kDa peptides. The model’s K_RWM = 0.02/h citing “Salt 2011” cannot be traced to this paper — and the 2001 value would be for a very different (small) molecule.

What this paper does NOT provide

• No MW scaling table for RWM permeability
• No human-specific parameters (all guinea pig)
• No middle ear clearance rate
• Guinea pig ST clearance t½ = 60 min ≠ human

Access status

PAYWALLED — ScienceDirect, Hearing Research Vol.154(1-2):88-97. Not on PMC. Needs institutional access or Sci-Hub.

Connections

• [source] [[hydrogel_phase4e_cochlear_pkpd]] — K_PERILYMPH_CLEAR baseline (guinea pig 60 min t½); K_RWM is not from this paper
• [[2018-salt-plontke-pharmacokinetic-principles-inner-ear]]