Abstract

Ulcerative colitis (UC) is driven in part by the IL-23 → JAK2/STAT3 axis. Approved anti-IL-23 monoclonal antibodies work but require injections, are costly, and carry systemic immunosuppression risks. We propose localized mucosal neutralization: three de-patented anti-IL-23p19 scFv prototypes (A/B/C) designed by motif-guided CDR engineering with AlphaFold checks, plus two mAb-derived scFv (risankizumab-p19, ustekinumab-p40) as performance baselines. Validation: recombinant-protein ELISA for binding and a Caco-2 inflammation model measuring p-JAK2^Y1007/1008 and p-STAT3^Y705 inhibition. Long-term, the best prototype will be secreted by engineered '''Lactobacillus acidophilus''' to reduce cost and systemic exposure.

Introduction

While systemic biologics against IL-23/Th17 improve UC, their cost, injection route, and systemic risks limit accessibility. Localized neutralization may reach similar efficacy with lower doses and reduced systemic exposure. We built a reproducible pipeline: motif-guided sequence design → AlphaFold structure checks → binding by recombinant-protein ELISA → signaling by Caco-2 p-JAK2/p-STAT3 → (long-term) probiotic delivery. This season, we aim to demonstrate at least one candidate that binds IL-23 and suppresses JAK2/STAT3 activation, paving the way for delivery.

Project Outline

Cycle-1 (Bacterial proof-of-concept, ustekinumab-derived scFv in Lactobacillus)

  • Construct:pLEM415-ldhL-SlpA-SP–VH–(G4S)₃–VL–6×His; intent: Sec-pathway secretion in Gram-positive host.
  • Outcome: Recurrent indels/point mutations during E. coli amplification (e.g., 1-bp deletion causing frameshift), so functional testing was halted; decision: pivot to mammalian expression for function first, while re-engineering bacterial cassette for stability.

Cycle-2 (Mammalian validation, candidate expansion)

  • Set: 3 de novo scFv (1/2/3) + risankizumab-p19 + ustekinumab-p40 on pTwist-CMV, Igκ leader for secretion, 6×His tag for QC.
  • Expression/Secretion (HEK293T): all five expressed; only p19/p40 secreted to supernatant; de novo scFv1/2/3 were intracellular in-house but secreted when produced externally in HEK293 (suspension), indicating a host/platform effect rather than intrinsic misfolding.
  • Binding (ELISA): robust for p19 (and p40); de novo scFv1/3 bind when purified externally; scFv2 lacked signal.
  • Function (Caco-2, 15–60 min window): IL-23 activates JAK2/STAT3; p19 lowers p-STAT3 close to baseline, p40 increases p-JAK2; scFv1 weak. This nominates p19-class for next steps.

Next

  • Consolidate p19-direction scFv (optimize sequence, secretion leader, and expression chassis); re-harden bacterial vector for return to Lactobacillus secretion and downstream application testing.

scFv

Molecule format & topology

All scFv use VH–(G4S)₃–VL with N-terminal secretion leader (Igκ for HEK293T; SlpA for Lactobacillus) and C-terminal 6×His for detection/purification.

Candidate panel

  • De novo (1/2/3): learned from common therapeutic CDR “motifs”; secretion was platform-dependent—intracellular in HEK293T, secreted when produced by HEK293 suspension vendor; purified scFv1/3 bind IL-23 at high concentration; scFv2 negative.
  • mAb-derived controls: risankizumab-p19 (anti-p19) and ustekinumab-p40 (anti-p40) serve as benchmarks for secretion/binding/function; p19 showed partial p-STAT3 inhibition, p40 did not inhibit and may enhance upstream signaling (↑p-JAK2).

QC & analytics

Anti-His Westerns (lysate/supernatant), ELISA against recombinant IL-23, and time-resolved Caco-2 p-JAK2/p-STAT3 readouts (15–60 min, 30-min focal point) with p/total normalization.

Application

Near-term (lab-validation)

  • Lock p19-direction scFv with consistent binding and p-STAT3 inhibition ; optimize secretion leader and expression conditions to avoid platform artifacts; establish reproducible SOPs for supernatant prep and phospho-assays.

Mid-term (bacterial delivery feasibility)

  • Re-engineer pLEM415-ldhL-SlpA cassette to eliminate unstable motifs, re-test secretion in Lactobacillus; maintain IP/ethics boundary (research-only for mAb-derived sequences).

Long-term (mucosal therapy concept)

  • Oral probiotic capsule with engineered Lactobacillus secreting anti-IL-23 scFv for local neutralization in the gut; integrate biosafety gates (auxotrophy/kill-switch, watermarking) and dose-tunable manufacturing.

Summary

Our data show: (i) expression of all five scFv in HEK293T; (ii) secretion in-house only for p19/p40, with the de novo trio secreting when produced by an external HEK293 (suspension) platform; (iii) binding confirmed for p19 (and externally for de novo scFv1/3); and (iv) functional signaling where p19 reduces p-STAT3 while p40 increases p-JAK2—guiding the p19 path forward. Therefore, the next rational steps are to lock a p19-class scFv, stabilize the bacterial expression cassette, and progress toward Lactobacillus secretion for localized neutralization—delivering on the project’s translational vision with documented, data-driven decisions.