The Evidence Layer for Scientific AI

 Quick Analysis Plan Copied

 Long Analysis Plan

No universal TEER threshold exists — values are device- and model-specific

Across the supplied evidence, "good barrier" TEER differs by more than two orders of magnitude depending on tissue, model, and instrument. Human renal proximal tubule cells in an ITO-electrode MPS plateaued at 37 Ω·cm² (vs 44 Ω·cm² on a commercial ERS-2), with day-10 coefficient of variation of 0.01 vs 0.22 Renal PT-MPS TEER plateau and CV. Caco-2 in a PET double-layer MPS peaked at 500–600 Ω·cm² (days 4–5) then declined to 150–200 Ω·cm² as villi-like structures matured — a reminder that a *falling* TEER can accompany *maturation*, not damage Caco-2 PET chip TEER time course. An NVU-on-chip with four-electrode TEER reached ~950 Ω·cm² after 14 days, with blank normalization and EGTA disruption validating sensitivity NVU-on-chip TEER maturation and EGTA validation. In ECIS-format endothelial comparisons, HDMECs reached 3837±78 Ω vs HUVEC 1721±47 Ω across 8 independent experiments HDMEC vs HUVEC ECIS TEER comparison — note these are electrode-normalized ohms, not Ω·cm².

Midpoints of reported ranges plotted; error bars span published ranges. Different units exist across studies (some report raw Ω on ECIS electrodes).

How misquotation happens — and how to avoid it

• Raw vs normalized: LATEER reports raw kΩ (blank 14.4 → 30.7 kΩ across MSC densities) that are meaningless outside a fixed electrode configuration; the authors explicitly warn values are not Ω·cm² and shift on electrode remounting LATEER raw kΩ density-dependence with n=1.
• Current-density uniformity: electrode geometry changes absolute values (PET chip Model 2 uniformity ratio 1.05 vs 1.20 for Model 1) Electrode geometry affects TEER values.
• Frequency and baseline: ex vivo colon TEER required AC excitation (~5 kHz target), baseline subtraction, and muscle removal shifted TEER ~39% — regional differences (distal ~19% lower) show even within one tissue a single "threshold" fails Ex vivo colon TEER methodology and regional variation.

Practical checklist: report (1) unit and normalization (Ω vs Ω·cm²), (2) device/electrode geometry and measurement frequency, (3) blank subtraction, (4) time course rather than a single threshold, (5) replicates (many studies here used n=1–4 per condition). A "human-relevant" TEER number quoted without these is not interpretable.

Blind spots

None of these studies establishes an in-vivo human gold-standard TEER for the corresponding tissue; all thresholds are internal comparisons within a device. Where TEER conflicted with other readouts (claudin-5 overexpression raised no TEER yet reduced EGTA-induced gaps; capacitance tracked detachment with r = −0.987 when TEER alone was ambiguous), single-threshold thinking failed — TEER should be paired with permeability or junctional markers TEER plus capacitance outperforms TEER alone. Confidence: moderate-high on the misquotation mechanics; low that any cross-study absolute threshold can be defended.