STRC Research

STRC mRNA-LNP Strategy B Full-Length

9 min read

STRC mRNA-LNP Strategy B Full-Length

Per-OHC and cochlea-mean PK/PD for Strategy B (full-length STRC mRNA delivered directly by LNP, no RBM24 intermediate). Parallel to STRC mRNA-LNP PKPD Multi-Dose Schedule (Strategy A). Two findings specific to Strategy B: (1) per-OHC dose required is ~50× higher than Strategy A’s catalytic RBM24 route — direct stoichiometric translation has no amplification, so 10,000 mol/OHC intracellular m1ψ Q6W is the floor to reach s_exog ≥ 1× WT per transfected OHC; (2) Strategy B breaks the Hill ceiling that constrained Strategy A — cochlea-mean STRC is unbounded in dose, so at 5% cochlear-tropic LNP Strategy B hits cochlea-mean ≥ 2× WT for Misha (2.18×), which Strategy A could not exceed 1.10× at the same LNP efficiency. For Misha’s paternal 98 kb STRC deletion, Strategy B is the only viable mRNA option (Strategy A needs pre-mRNA substrate that the deletion eliminates). Untargeted LNP (0.8%) remains therapeutically dead at any dose.

Method

Two-compartment ODE (no RBM24 intermediate, no Hill saturation):

dm/dt = pulse(t; D, T) − k_m · m
ds/dt = K_TRANSLATE_STRC · m − k_s · s
  • m — exogenous STRC mRNA, mol/OHC intracellular
  • s — exogenous STRC protein, arbitrary units matching s_WT = 3021
  • Pulse: Δm = D at t = 0, T, 2T, ... (intracellular dose; extracellular = D / 0.02)

Calibration fixes K_TRANSLATE_STRC such that endogenous steady state is consistent with Strategy A’s baseline:

ParameterValueSource
s_WT3021 unitsStrategy A baseline (arbitrary units)
endogenous STRC mRNA SS50 copies/OHCbiological estimate for moderate-expression gene
K_TRANSLATE_STRCs_WT · k_s / 50 = 3.46×10⁻⁵derived
STRC protein t½14 d (k_s = ln2 / 14 d)inherited
mRNA t½ (unmod, OHC)4 hmrna_stability_cochlear_results.json
mRNA t½ (m1ψ, OHC)12 hidem
LNP endosomal escape2%literature

Population fold (cochlea-mean over s_WT):

tx OHC fold     = s_endo_frac + s_exog_tx / s_WT
non-tx OHC fold = s_endo_frac
cochlea_mean    = s_endo_frac + eff · s_exog_tx / s_WT

No hard ceiling — cochlea-mean grows linearly with dose (soft biological ceiling at ~5-10× from ER/UPR stress is a model limitation, not modelled).

Three clinical scenarios:

Scenarios_endo_fracNotes
WT reference1.0homozygous WT, therapy adds on top
DFNB16 null0.0biallelic STRC loss-of-function
Misha0.15paternal 98 kb Δ null + maternal E1659A hypomorphic residual

Sweep: 2 modifications × 3 targetings × 6 intervals × 8 doses = 288 regimens, 365-day horizon, LSODA with dose-time breakpoints. Runtime < 1 min.

Results — per-OHC dose floor (transfected OHC only)

Minimum intracellular dose for s_exog_trough / s_WT ≥ 1 (functional restoration in transfected OHC, m1ψ):

IntervalMin dose/pulseDoses/yrIntra/yrExtra/yr (÷0.02)s_exog trough
Q2W2,000 mol2652,0002.6×10⁶1.49
Q3W5,000 mol1785,0004.25×10⁶2.03
Q4W5,000 mol1365,0003.25×10⁶1.24
Q6W10,000 mol990,0004.5×10⁶1.06
Q8W30,000 mol7210,0001.05×10⁷1.48
Q12W100,000 mol4400,0002.0×10⁷1.18

Minimum for s_exog_trough ≥ 2 (2× WT, matches Strategy A threshold) m1ψ:

IntervalMin doseIntra/yrExtra/yr
Q2W5,000130,0006.5×10⁶
Q3W5,00085,0004.25×10⁶
Q4W10,000130,0006.5×10⁶
Q6W30,000270,0001.35×10⁷
Q8W100,000700,0003.5×10⁷
Q12W—infeasible—

Strategy B needs ~50× more extracellular mRNA per transfected OHC than Strategy A (4.5M vs 90K mol/yr at Q6W m1ψ for equivalent per-OHC rescue). Direct translation produces one protein per ribosome per mRNA per lifetime; RBM24 in Strategy A catalyses many STRC splicing events per RBM24 molecule.

Results — cochlea-mean by LNP targeting × scenario (the key finding)

Minimum dose to reach cochlea-mean ≥ 2× WT (m1ψ):

LNP targetingWT refDFNB16 nullMisha
Untargeted (0.8%)infeasible (max 1.59×)infeasible (max 0.59×)infeasible (max 0.74×)
Cochlear-tropic (5%)Q2W × 30K (39M/yr) → 2.11×Q3W × 100K (85M/yr) → 2.03×Q3W × 100K (85M/yr) → 2.18×
OHC-targeted (20%)Q2W × 10K (13M/yr) → 2.49×Q3W × 30K (25.5M/yr) → 2.43×Q3W × 30K (25.5M/yr) → 2.58×

Minimum dose to reach cochlea-mean ≥ 1× WT (functional restoration, m1ψ):

LNP targetingWT refDFNB16 nullMisha
Untargeted (0.8%)at baselineinfeasible (max 0.59×)infeasible (max 0.74×)
Cochlear-tropic (5%)at baselineQ2W × 30K (39M/yr) → 1.11×Q2W × 30K (39M/yr) → 1.26×
OHC-targeted (20%)at baselineQ2W × 10K (13M/yr) → 1.49×Q2W × 10K (13M/yr) → 1.64×

The Strategy A → B delivery inversion. At 5% cochlear-tropic LNP:

  • Strategy A ceiling: cochlea-mean ≤ 0.05 · 3 + 0.95 = 1.10× (Hill caps per-OHC fold at 3)
  • Strategy B achievable: cochlea-mean 2.03–2.18× (DFNB16 null / Misha, Q3W × 100K m1ψ)

Strategy B can compensate for modest LNP tropism with high dose, because linear dose-response has no pharmacological ceiling — delivery and dose become substitutable knobs. Strategy A has no such substitution: once per-OHC trough saturates at the Hill max_boost, extra dose produces no extra STRC, and cochlea-wide coverage is upper-bounded by eff_ohc alone.

Interpretation

  1. The mechanism inversion: direct replacement scales with dose, catalytic regulation saturates. Strategy A has a pharmacology ceiling (Hill max_boost = 3) but is dose-frugal (1800 mol/yr intra per OHC). Strategy B has no pharmacology ceiling but is dose-expensive (90–270K mol/yr intra per OHC). The two strategies trade off opposite failure modes.

  2. For LNP tropism in the 5-20% range, Strategy B dominates Strategy A on cochlea-mean. At 5%, Strategy A maxes at 1.10× while Strategy B reaches 2.18× (Misha) with Q3W dosing. Given published cochlear LNPs cluster in this range, Strategy B may be the right bet for the next-5-years window while SORT-LNP-class platforms mature.

  3. Strategy B is the only viable mRNA option for Misha’s paternal allele. The 98 kb deletion eliminates the entire STRC locus — no pre-mRNA, no substrate for RBM24 splicing correction. Strategy A upregulates only the maternal E1659A allele (which is hypomorphic, giving ≤ 1× WT total). Strategy B provides de novo stereocilin regardless of locus state.

  4. Untargeted LNP (0.8% OHC) is therapeutically dead for null/hypomorphic patients even at Strategy B’s full dose range. Max cochlea-mean for DFNB16 null = 0.59× WT at 100K mol/OHC Q2W (= 5M extracellular/OHC/pulse, a million times physiologic transcript loading). 99.2% of OHCs never see LNP and stay at 0% endogenous STRC; the 0.8% transfected cannot numerically dilute up to cochlea-mean ≥ 1.

  5. Misha-specific path forward (compound het 98 kb Δ paternal + E1659A maternal):

    • OHC-targeted 20% LNP + Strategy B Q3W × 30,000 mol/OHC m1ψ: cochlea-mean 2.58× (therapeutic), extracellular burden 25.5 M mol/yr per OHC = ~0.5 mg mRNA total per cochlea per year. Practically manufacturable.
    • Complementary to STRC Pharmacochaperone Virtual Screen E1659A on the maternal allele (affinity repair × quantity restoration multiplicatively combine).
    • Complementary to STRC Piezoelectric TM Bioelectronic Amplifier as non-overlapping mechanism (molecular coupling vs bioelectronic amplification).

  6. Safety envelope at high dose: the linear ODE ignores ER-stress / UPR limits on secreted-protein overexpression. Mammalian cells typically tolerate ~5-10× overexpression of secretory proteins before UPR activation. Strategy B’s per-OHC troughs at high dose can exceed 70× WT (at 100K mol/OHC Q2W m1ψ) — beyond any physiological tolerance. Practical dose cap is probably around s_exog_tx ≈ 5-10× WT. This makes 20% OHC-targeted LNP the realistic target eff_ohc, because it hits cochlea-mean 2× with only ~10× per-OHC excess.

Head-to-head: Strategy A vs Strategy B (m1ψ Q6W baseline)

MetricStrategy A (RBM24)Strategy B (full-length)
Per-OHC dose200 mol/OHC intra10,000 mol/OHC intra
Annual intra per OHC1,800 mol90,000 mol
Annual extra per OHC90,000 mol4.5×10⁶ mol
Per-OHC trough2.17× STRC folds_exog ≥ 1× WT
Pharmacology ceiling3× Hill capunbounded in dose
Cochlea-mean at 5% eff1.10× (capped)2.18× Misha (Q3W × 100K)
Cochlea-mean at 20% eff1.40× (capped)2.58× Misha (Q3W × 30K)
Works on null alleleno (needs pre-mRNA)yes (de novo STRC)
Payload~1.5 kb RBM24 CDS~6 kb STRC CDS
Off-target risk469 genes splicedminimal (direct replacement)

Limitations

  • Linear dose-response is an idealization. Real cells have ribosome loading saturation (~5 ribosomes per 5 kb mRNA at max) and ER/UPR stress at ~5-10× secretory overexpression. Model overstates feasibility above ~10× WT per-OHC trough.
  • K_TRANSLATE_STRC calibrated from an assumed 50-copy endogenous STRC mRNA steady state. Actual OHC transcriptomics would tighten this.
  • Shared limitation with Strategy A: no extracellular compartment model for LNP transit through perilymph, no anti-PEG immune response for repeat dosing, no stochastic transfection distribution across 12,000 OHCs.
  • Cochlea-mean is still the wrong clinical endpoint for a structural protein. The tonotopic audiogram projection (per STRC mRNA-LNP Audiogram Rescue) should be repeated with Strategy B per-OHC s_exog profile, not the Strategy A fold metric.
  • Misha’s s_endo_frac = 0.15 is estimated (maternal E1659A × 0.5 hemizygous dosage × ~0.3 residual activity). Direct measurement from patient cochlear sample is impossible; best estimate remains ~10–20% of WT.
  • No dose-sparing from non-cell-autonomous STRC secretion. Stereocilin is secreted extracellularly and may restore adjacent untransfected hair bundles partially; modelling this effect could lower the effective dose floor for Strategy B specifically.

Next steps

  1. Audiogram rescue for Strategy B: compose this per-OHC profile (s_exog ≥ 1× WT at transfected OHC; 0 at non-transfected for DFNB16 null / 0.15 for Misha) with the tonotopic × ABR transfer function from STRC mRNA-LNP Audiogram Rescue. Expectation: Strategy B at 5% cochlear-tropic should match Strategy A at 20% OHC-targeted on weighted audiogram improvement.
  2. ER/UPR safety model: soft ceiling on s_exog per OHC (Michaelis-Menten saturation at ~5× WT with dose) to bound the feasible regimen space.
  3. Non-cell-autonomous secretion: treat exogenous STRC as a diffusible pool that reaches neighboring untransfected OHCs. If diffusion-limited to adjacent cells (r ~ 10 μm), effective eff_ohc increases by ~3-fold with no extra dose.
  4. Strategy A+B combination: RBM24 mRNA + full-length STRC mRNA co-dosed in the same LNP (quantity-from-B × quality-from-A) for compound het patients like Misha where A alone undershoots (maxes at ~1× WT from single-allele upregulation).
  5. LNP payload-size penalty: 6 kb STRC mRNA-LNP loading efficiency vs 1.5 kb RBM24 mRNA-LNP. Larger mRNA loads fewer copies per LNP particle; may reduce effective per-particle dose by ~2-4×. Model with explicit LNP particle dynamics.

Replication

cd ~/STRC/models
/opt/miniconda3/bin/python3 strc_mrna_strategy_b_pkpd.py
# outputs: strc_mrna_strategy_b_pkpd.json

Files / Models

  • ~/STRC/models/strc_mrna_strategy_b_pkpd.py — 2-compartment ODE + sweep
  • ~/STRC/models/strc_mrna_strategy_b_pkpd.json — 288 regimens, per-OHC minimums (1× and 2× WT), cochlea-mean across scenarios × targetings
  • ~/STRC/models/rbm24_mrna_dose_response_results.json — s_WT calibration input
  • ~/STRC/models/mrna_stability_cochlear_results.json — mRNA half-life input

Ranking delta

Applied against STRC Hypothesis Ranking (2026-04-21):

Connections

Graph View

Backlinks