STRC Research

ARBITER Workflow for Cochlear Enhancer Discovery

2 min read

ARBITER Workflow for Cochlear Enhancer Discovery

Finding the regulatory elements that drive OHC-specific gene expression is a hard problem. The cochlea is tiny, has very few accessible cells, and can’t be transfected in culture in any way that replicates in vivo biology. ARBITER solves this by making the cochlea itself the assay system.

The workflow:

  1. Pick a hearing loss gene you want to regulate.
  2. Search its genomic locus for conserved non-coding elements (CNEs) — sequences >100 bp that are conserved across 9 species. These are likely regulatory.
  3. Also pull ATAC-seq and ChIP-seq data from published neonatal HC datasets to find open chromatin near the TSS.
  4. Clone each candidate element upstream of a minimal promoter + fluorescent reporter in an AAV plasmid.
  5. Package in AAV-ie. Inject into neonatal mouse cochleae via round window (P1-P3).
  6. Collect at P7-P16. Image by immunofluorescence. Score transduction efficiency and cell-type specificity.
  7. Iterate: combine elements, swap modules, discard non-performers.

What Zhao et al. 2025 showed: For Slc26a5, two CNEs (E1: 247 bp, E2: 525 bp in introns 1 and 2) together drive OHC-specific expression. Neither alone was sufficient. E1 acts as a “facilitator” — no activity alone, but dramatically boosts E2. The combination E1+E2 gave OHC-specific but low-level expression.

The engineering step: They then dissected each CNE into conserved sub-modules, scored each module’s contribution, and reassembled the best modules with repetition. Final synthetic enhancer B8 = E1P3×2 + E2P2×2 + E2P3×2. Result: ~100% OHC transduction, no off-target, expression at wild-type levels.

Direct relevance to STRC: The STRC gene locus (DFNB16, chromosome 15) almost certainly has analogous OHC-specific CNEs waiting to be found. Running ARBITER on STRC would identify enhancers that could replace the generic Myo15 promoter currently used in STRC gene therapy, improving specificity and reducing off-target AAV dose. This is a concrete next-step experiment that doesn’t require new animal models — just bioinformatics + existing AAV-ie infrastructure.

Connections

Graph View

Backlinks