Published on 22/12/2025
How to Evaluate T-cell Responses in Vaccine Trials (Step-by-Step)
Why T-cell Readouts Matter and Where They Fit in Vaccine Decisions
Antibody titers are critical, but they don’t tell the whole story. CD4+ and CD8+ T-cell responses contribute to viral clearance, breadth against variants, and durability when neutralization wanes. Regulators frequently ask for T-cell data to contextualize humoral findings, de-risk vulnerable populations (older adults, immunocompromised), or support immunobridging when clinical endpoints are scarce. A well-designed T-cell plan answers three questions: what is being measured (e.g., IFN-γ/IL-2 TNF-α polyfunctionality, cytotoxic readouts like granzyme B), how it is measured (ELISpot, ICS/flow, activation-induced markers [AIM], or proliferation), and how results influence dose/schedule or labeling decisions.
In early phase studies, T-cell assays help prioritize regimens with Th1-skewed immunity (desired for many viral vaccines). In Phase II/III, they provide mechanistic context and can enable bridging across age groups by showing comparable cellular profiles. The Statistical Analysis Plan (SAP) should define timepoints (e.g., Day 0, post-dose Day 14/28/35, durability Day 180), target cell populations (CD4+ vs CD8+), and estimands for intercurrent events (breakthrough infection or receipt of a non-study vaccine). Governance matters: an immunology lead signs
Picking the Right Assay: ELISpot vs ICS/Flow vs AIM (and When to Combine)
ELISpot (IFN-γ, IL-2): Highly sensitive for frequency of cytokine-secreting cells. Output is spots per 106 PBMC. Typical validation targets include LOD≈5 spots, LLOQ≈10 spots, ULOQ≈800 spots, with intra-assay CV≤20%. Strengths: sensitivity, relative simplicity. Limitations: limited multiplexing; no direct polyfunctionality.
Intracellular Cytokine Staining (ICS) with flow cytometry: Quantifies polyfunctional T cells producing combinations (e.g., IFN-γ/IL-2/TNF-α) and distinguishes CD4+/CD8+ phenotypes. Report as % of parent (e.g., %CD4+IFN-γ+). Define reportable range (e.g., 0.01–20%), LOD≈0.005%, LLOQ≈0.01%, and acceptance criteria for compensation residuals <2%. Requires rigorous panel design, single-stain controls, FMO (fluorescence minus one), and stability of fluorochromes.
Activation-Induced Marker (AIM): Uses markers (e.g., CD69, CD40L [CD154], OX40, 4-1BB) to identify antigen-specific T cells without relying on intracellular cytokine capture. Useful for breadth and helper subsets (Tfh). Report as %AIM+ of CD4+/CD8+. LOD≈0.005%, LLOQ≈0.01% similar to ICS.
Programs often pair ELISpot (for sensitivity) with ICS (for polyfunctionality) or AIM (for breadth). Each method’s Lab Manual must lock stimulation conditions (peptide pools spanning overlapping 15-mers at 1–2 µg/mL per peptide), incubation times (e.g., 16–20 h ELISpot; 6 h ICS with brefeldin A), and positive controls (SEB or CEFX peptide megapools). Include plate acceptance criteria, instrument QC, and replicate rules. Below is an illustrative comparison.
| Assay | Primary Readout | LOD | LLOQ | Strength | Limitation |
|---|---|---|---|---|---|
| ELISpot (IFN-γ) | Spots/106 PBMC | 5 spots | 10 spots | High sensitivity | No polyfunctionality |
| ICS/Flow | % cytokine+ of CD4/CD8 | 0.005% | 0.01% | Polyfunctionality, phenotype | Complex, instrument heavy |
| AIM | % AIM+ T cells | 0.005% | 0.01% | Broad antigen-specificity | Indirect functional readout |
Assay choice should align with your decision questions: if you must differentiate Th1/Th2 skew, include ICS (IFN-γ vs IL-4/IL-5). If durability is key, run ELISpot longitudinally to track memory. Where manufacturing changes occur, include comparability panels to ensure no assay-induced shifts mask biology.
PBMC Handling, QC, and Acceptance Criteria: Getting Pre-Analytical Controls Right
Pre-analytical variability can drown a true biological signal. Standardize phlebotomy tubes, processing time (e.g., isolate PBMC within 6 h; 2–4 h preferred), Ficoll gradient parameters (e.g., brake off, 400–500 g for 30 min), and cryopreservation (10% DMSO in serum-containing media; controlled-rate freeze ~1 °C/min to −80 °C, then liquid nitrogen). Predefine acceptance criteria: viability at thaw ≥85% (target ≥90%), recovery ≥70%, and ≤2 freeze-thaw cycles. Track shipment on dry ice with continuous temperature logging; excursions trigger quarantine and re-test rules.
Positive controls (SEB, PHA, or CEFX) ensure cells are competent; set laboratory cutoffs (e.g., ELISpot positive control >500 spots/106; ICS positive control %IFN-γ+ CD4 ≥0.3%). Negative control wells (DMSO vehicle) define background for subtraction. Instrument QC: daily cytometer performance tracking (e.g., CS&T beads), target MFI windows for each channel, and compensation matrix residuals <2%. Document panel lot numbers, cytometer configurations, and any service events.
| Parameter | Threshold | Action if Out |
|---|---|---|
| Post-thaw viability | ≥85% | Repeat thaw if aliquot available; flag for sensitivity |
| Recovery | ≥70% | Note in LIMS; interpret cautiously |
| ELISpot PC (SEB) | >500 spots/106 | Repeat plate; investigate cells/reagents |
| ICS compensation residuals | <2% | Re-run compensation; check panel |
Finally, transparency matters for ethics and inspectors. While clinical teams don’t compute manufacturing PDE or cleaning MACO, referencing example limits (e.g., PDE 3 mg/day for a residual; MACO 1.0–1.2 µg/25 cm2 surface swab) in your quality narrative demonstrates end-to-end control of risks across product and testing—useful context when T-cell data are used for immunobridging or accelerated filings.
Endpoints, Positivity Criteria, and Statistics: From Events to Decisions
T-cell endpoints should be predefined and clinically interpretable. Common ELISpot endpoints include median (or mean) spot count per 106 PBMC (background-subtracted) at Day 14/28/35 and fold-rise from baseline; ICS endpoints include %CD4+IFN-γ+, %CD8+IFN-γ+, and polyfunctional % (e.g., IFN-γ/IL-2/TNF-α triple-positive). AIM endpoints capture %AIM+ CD4 or CD8. Positivity should be defined with dual criteria: (1) a minimum magnitude above LLOQ (e.g., ELISpot ≥30 spots/106 PBMC after background subtraction; ICS ≥0.03% cytokine+ of parent), and (2) a fold-over-background (e.g., ≥3× vehicle control) or fold-rise from baseline.
State analytical limits: for ICS/AIM, LOD≈0.005%, LLOQ≈0.01%, ULOQ≈20%; for ELISpot, LOD 5 spots, LLOQ 10 spots, ULOQ 800 spots with intra-assay CV≤20% and inter-assay CV≤25%. Handle values below LLOQ explicitly (e.g., set to half-LLOQ for geometric means) and define replicate rules (duplicate wells for ELISpot; technical duplicates or pooled replicates for ICS). Use ANCOVA on log-transformed readouts (add a small constant if zeros after background subtraction) with baseline and site as covariates, report geometric mean ratios (GMRs) and 95% CIs, and manage multiplicity via gatekeeping (e.g., CD4 endpoints first, then CD8, then polyfunctionality) or Hochberg. When bridging age cohorts, require non-inferiority margins (e.g., GMR lower bound ≥0.67).
| Assay | Magnitude Criterion | Fold Criterion | Decision |
|---|---|---|---|
| ELISpot | ≥30 spots/106 (post-BG) | ≥3× negative control | Responder |
| ICS (CD4) | ≥0.03% | ≥3× negative control | Responder |
| AIM (CD4) | ≥0.03% | ≥3× negative control | Responder |
For exploratory correlates, model clinical risk reduction per 2× increase in polyfunctional % using Cox or Poisson models within immune substudies; prespecify that these are supportive, not confirmatory, unless powered accordingly. Ensure your SAP includes sensitivity analyses (e.g., excluding samples with viability <85% or out-of-window collections) and spells out how missing data and outliers are handled.
Case Study: Hypothetical mRNA Vaccine—Polyfunctionality Drives the Dose Decision
Design: Adults receive 10 µg, 30 µg, or 100 µg doses (Day 0/28). ELISpot IFN-γ and ICS polyfunctionality (%CD4+IFN-γ/IL-2/TNF-α) are measured at Day 35; safety captures Grade 3 systemic AEs within 7 days. Assay parameters: ELISpot LLOQ 10 spots; ICS LLOQ 0.01% with compensation residuals <2% and CV≤20% for controls. Results (dummy):
| Arm | ELISpot IFN-γ (spots/106) | %CD4 Triple-Positive | %CD8 IFN-γ+ | Grade 3 Sys AEs (%) |
|---|---|---|---|---|
| 10 µg | 180 (95% CI 150–210) | 0.045% | 0.030% | 2.1% |
| 30 µg | 260 (220–300) | 0.085% | 0.055% | 3.8% |
| 100 µg | 290 (240–340) | 0.090% | 0.060% | 7.1% |
Interpretation: Moving from 30→100 µg yields marginal T-cell gains but doubles Grade 3 systemic AEs. The SAP’s decision rule favors the lowest dose achieving non-inferior polyfunctionality versus the next higher dose (GMR lower bound ≥0.67) and acceptable safety (Grade 3 AEs ≤5%). RP2D: 30 µg. Durability at Day 180 shows maintained ELISpot (≥120 spots) and preserved %CD4 triple-positives (≥0.04%), supporting schedule selection. These cellular data, paired with neutralization, underpin immunobridging to adolescents with predefined non-inferiority margins.
Documentation, TMF Readiness, and Regulatory Alignment
Inspection-ready T-cell packages are built on documentation discipline. The Lab Manual must fix peptide pool composition, stimulation conditions, gating strategy, positivity thresholds, and acceptance criteria. Store panel designs, compensation matrices, bead lots, and cytometer configurations under change control; include traceable curve-fitting or gate-applying scripts with checksums. In the TMF, file raw FCS/ELISpot images, annotated gates, QC trend charts, and deviation/CAPA logs; match analysis datasets (ADaM) to table shells in the SAP. For accelerated or conditional approvals, clarify that T-cell endpoints are supportive unless prospectively powered and alpha-controlled as primary. When ethics committees ask about end-to-end quality, reference representative CMC control examples (e.g., residual solvent PDE 3 mg/day; cleaning MACO 1.0–1.2 µg/25 cm2) to show product and assay are controlled across the lifecycle. For harmonized expectations on quality and statistics, consult the ICH Quality Guidelines.
Bottom line: T-cell evaluations complement serology by revealing breadth, quality, and durability of immunity. With fit-for-purpose assays, clear responder definitions, and GxP-tight documentation, your vaccine program can use cellular data to sharpen dose/schedule decisions, accelerate bridging, and build a more resilient benefit–risk case.