Evidence Tracing

What the stance gets right — and where it can mislead

Prioritizing interventional over observational evidence is justified because depletion experiments directly test causality. The mouse allergic airway study demonstrates this principle: NK-depleted mice showed reduced airway inflammation, eosinophilia, IL-5, IL-13, and OVA-specific IgE compared to WT — revealing a pro-inflammatory NK role that correlation alone could not establish NK depletion reveals pro-inflammatory role in vivo. Similarly, granzyme M deficiency reduced proinflammatory cytokines and improved LPS-induced survival, showing that cytotoxic effector molecules can paradoxically drive inflammation — a finding only accessible through genetic intervention Granzyme M promotes TLR4 inflammation.

The blindspot: tissue-resident NK ≠ blood NK

The most consequential blindspot of a mechanistic-first stance is assuming that mechanistic insights from blood or murine NK studies transfer to mucosal tissue. The nasal-vs-blood comparison shows this assumption fails dramatically: blood NK cells express perforin at ~1210-fold higher levels than tissue-resident nasal NK cells, while nasal cells upregulate ITGA2 (~48-fold) and IL8 (~30-fold) — indicating tissue NK cells are oriented toward immune orchestration, not cytotoxicity Tissue-resident vs blood NK gene expression.

Blue = genes higher in blood NK cells; red = genes higher in tissue-resident nasal NK cells. Data from a 71-gene Nanostring panel (n=18 subjects) Gene panel methodology

This means: the rectal explant NK-depletion experiment is epistemically stronger than the original correlational study, but its interpretation must be grounded in rectal-resident NK biology, not extrapolated from blood NK cytotoxicity or from the allergic airway pro-inflammatory model. The NKp46 receptor also marks ILC1s — homozygous NKp46 deficiency nearly eliminates ILC1s without affecting other ILC subsets NKp46 required for ILC1 development — meaning NKp46-targeting antibodies in the explant experiment will co-deplete ILC1s, and any phenotype change cannot be attributed to conventional NK cells alone without ILC1-specific controls.

What would change this assessment

The stance would need revision if: (1) the paired depletion experiment shows no p24 change despite validated NK/ILC1 depletion — indicating the original correlation was confounded by a variable that co-varies with NK abundance rather than NK function; (2) cytokine changes (IL-17A, IFN-γ, IP-10) do not mediate the p24 effect within the paired design, breaking the assumed mechanistic chain; or (3) donor heterogeneity reveals that depletion produces opposite effects in different donors, suggesting that NK function in rectal mucosa is determined by a variable not captured in bulk depletion (e.g., microbial engagement via NKp46-RadD-like interactions that are strain-specific RadD-NKp46 context-dependent NK modulation).