Traceable Claims in Immunology

Visual Evidence Analysis — "Best Evidence: NY-ESO-1 TCR Safety"

VISUALIZE FIRST — Key quantitative signals (plots) from the supplied dataset

Data source: mouse xenograft outcomes reported in the DAC + dTCR-T study (n=5/group; 3/5 tumor‑free in dTCR‑T + DAC) DAC + dTCR-T xenograft outcomes

Interpretation: DAC induced robust NY‑ESO‑1 expression in U937, HL60, Kasumi‑1 but not THP‑1 or normal epithelial/kidney lines, indicating selective re‑expression in some AMLs and limited induction in tested normal cells NY-ESO-1 induction by DAC in cell lines

Key ARDitox insight: TAAs (like NY‑ESO‑1) can produce many OTEs and some low‑safety‑score candidates (safety score <3) that map to proteins expressed in critical tissues (e.g., TTN in heart) — motivates proteome‑wide computational pre‑screening and tissue expression filtering before clinical use ARDitox OTE enumeration and scoring

Evidence synthesis (explain second; concise)

• Preclinical efficacy: DAC re‑expresses NY‑ESO‑1 in several AML models, enabling NY‑ESO‑1–specific TCR‑T recognition, increased cytokine release, cytotoxicity, central‑memory phenotype, and improved xenograft outcomes (3/5 tumor‑free in best arm) — evidence from in vitro assays, bisulfite sequencing, and xenograft models DAC + TCR-T efficacy evidence
• Preclinical safety signals and gaps: the DAC study reported limited induction in tested normal cell lines, but it did not exhaustively test all human tissues or low‑level expression in critical organs (heart, brain, muscle). Engineered high‑affinity TCRs have known history of off‑target toxicities in clinical settings when cross‑reactive peptides exist in critical tissues (e.g., TTN) — this motivates rigorous off‑target discovery before first‑in‑human use ARDitox cross-reactivity risk evidence
• Quality & limitations of the best available evidence: The DAC+TCR-T study is high-quality preclinical (detailed molecular assays, primary blasts, multiple donors) but remains preclinical with small n in vivo and short follow-up, and limited normal tissue panels — thus evidence for human safety is weak-to-missing. ARDitox offers strong computational safety screening but depends on model thresholds, HLA coverage, and expression databases; both approaches are complementary but not sufficient alone Study limitations noted by authors

Practical safety roadmap (evidence‑driven recommendations)

1. Proteome‑wide in silico screening of candidate TCRs against the human proteome (ARDitox‑style): enumerate OTEs, rank by safety score, and annotate tissue expression (GTEx/Human Protein Atlas) — high priority before any clinical dosing ARDitox pipeline recommendation
2. Empirical validation of top-ranked OTEs: peptide pulsing, primary human cell co-cultures (cardiomyocytes, neurons, myocytes, lung, hepatocytes) and sensitive functional assays (cytokine release, cytotoxicity), plus a dose/affinity sweep to identify affinity windows minimizing off‑target activation.
3. Assess DAC-mediated NY‑ESO‑1 re‑expression across a broader panel of primary normal human tissues and single‑cell RNA/protein assays post‑DAC to detect low‑level re‑expression in critical organs.
4. Consider using TCRs with optimized moderate affinity (avoid supra-physiologic affinities that increase promiscuity), include safety switches (suicide genes), and staged first‑in‑human designs with sentinel dosing and tissue biomarker monitoring.

What would change the conclusion (falsification conditions)

• If comprehensive ARDitox-style screening + experimental validation identified zero high‑risk OTEs and broad normal tissue testing after DAC showed no re‑expression in critical tissues, confidence in safety would rise substantially.
• If, conversely, ARDitox or experiments reveal high‑risk OTEs in essential organs (e.g., TTN‑like peptides in heart) with functional activation at clinically relevant TCR affinities, the safety case would be falsified.

Key sources used in this analysis are cited inline. The DAC + TCR-T preclinical dataset supports feasibility but not clinical safety; ARDitox-style proteome screening is essential and complementary. Experimental validation of predicted OTEs and broader normal-tissue DAC-response profiling are the critical next steps.

Core citations

• Decitabine + NY-ESO-1 TCR-T preclinical study
• ARDitox proteome cross-reactivity pipeline
• NY-ESO-1 immunotherapy review

Limitations, blindspots, and bias caution

• Preclinical xenograft models do not recapitulate human antigen expression patterns, immune microenvironment, or long‑term toxicity.
• ARDitox is powerful but imperfect: false positives and negatives occur; thresholds and expression databases (GTEx/HPA) have limits — experimental confirmation of computational hits is mandatory.
• Small sample sizes, donor heterogeneity (three donors), limited normal tissue panels, and short follow‑up reduce translational certainty.
• Affinity enhancement of TCRs can increase both potency and promiscuity; optimal affinity window must be determined empirically.

Next-step experimental suggestions (concise)

1. Run ARDitox-style proteome scan for each candidate NY‑ESO‑1 TCR variant (including SNP variants) and select top 20 predicted OTEs for experimental testing.
2. Test top OTE peptides by peptide‑pulsing on primary human cardiomyocytes, skeletal myocytes, neurons, hepatocytes, and kidney cells with engineered T cells to detect functional activation/cytotoxicity at clinically relevant effector:target ratios.
3. Profile DAC-treated primary human organoids / ex vivo tissues to detect possible CTA re‑expression in normal tissues across DAC doses planned clinically.

If you want, I can launch an AI Scientist agent to run a proteome‑wide ARDitox-style in silico screen for a specified NY‑ESO‑1 TCR sequence (or the canonical NY‑ESO‑1 157–165 peptide) and produce ranked OTE lists with tissue‑expression annotations and prioritised experimental validation lists.