Choosing CD3 Protein Formats for T-Cell Engager Development

> DRAFT — for scientific review before publication. All KD values and PK numbers are reported/typical; confirm against primary literature and vendor COAs before program decisions.

The anti-CD3 arm is the engine and the liability of every T-cell engager (TCE): it drives potency, cytokine release, and — through its format — the entire reagent and assay stack. Four decisions upstream of any BsAb campaign determine which CD3 proteins you must buy, screen against, and functionalize. This guide walks them in order.

Decision 1 — CD3ε vs CD3δ/γ arm

Nearly all clinically validated TCEs target CD3ε (blinatumomab, teclistamab, tarlatamab are all anti-CD3ε arms), because CD3ε is the subunit common to both TCR-CD3 heterodimers (CD3εδ and CD3εγ), so an ε-binder engages both αβ and γδ T cells and sees ~2 sites per TCR. The immunodominant epitope recognized by the reference clones (OKT3, UCHT1, SP34 class) is conformational and only expressed when CD3ε associates with CD3δ or CD3γ ([Salmerón et al., J Immunol 1991](https://europepmc.org/article/MED/1717585)) — which is why this decision is inseparable from the antigen-format decision below.

CD3δ- or CD3γ-directed, and εδ/εγ-heterodimer-selective arms are an emerging space aimed at T-cell subset selectivity (e.g., γδ engagement or reduced broad activation); they trade away decades of clinical precedent and most off-the-shelf reagents ([heterodimer-selective binder patent literature](https://patents.google.com/patent/US20240117050A1/en)).

Choose a CD3ε arm when you want pan-T-cell engagement and maximal comparator data; choose a δ/γ- or heterodimer-selective arm only when you have a concrete subset-selective hypothesis and budget for custom reagents and assays.

Decision 2 — Affinity-attenuated vs full-affinity anti-CD3

Full-affinity parental clones (OKT3/SP34-derived) typically bind CD3 in the low-nM to tens-of-nM range on T cells (reported; confirm with primary literature). Clinically advanced TCEs frequently deploy affinity-attenuated arms — reported KD values for engineered clinical anti-CD3 scFvs fall in the tens-to-hundreds of nM range, and the arms used across marketed/clinical BsAbs differ measurably in binding properties ([characterization of anti-CD3 arms in clinical BsAbs](https://snu.elsevierpure.com/en/publications/characterization-of-anti-cd3-antibodies-in-clinically-available-b/); [CD3 affinity attenuation review](https://www.tandfonline.com/doi/pdf/10.1080/17460441.2025.2522088)).

The logic: full-affinity arms activate T cells at maximal signaling even against low-density tumor antigen and are potently cytokine-inducing; attenuation lowers potency at low target density, reduces activation of circulating T cells at off-tumor sites, and is the standard lever for mitigating cytokine release syndrome (CRS) — at the cost of efficacy if over-attenuated. The affinity–CRS relationship is real but complex, and no universal optimum exists ([Expert Opin Drug Discov 2025](https://www.tandfonline.com/doi/pdf/10.1080/17460441.2025.2522088)).

Choose attenuation when your target has any normal-tissue/low-density exposure or CRS is the dominant program risk; choose a full-affinity arm only when potency at low antigen density is the binding constraint and you have a CRS management plan.

Decision 3 — CD3ε ECD vs TCR-CD3 complex antigen for screening

Because the dominant anti-CD3 epitope is conformational and requires the εδ or εγ heterodimer ([Salmerón 1991](https://europepmc.org/article/MED/1717585)), monomeric CD3ε ECD is a poor screening antigen for most arms: it only reports binders to linear/domain epitopes, not the functional OKT3/UCHT1-class epitope. Minimal faithful antigens are recombinant CD3εδ and CD3εγ heterodimers (vendor-supplied, MALS-verified, e.g., [ACROBiosystems CD3 protein line](https://www.acrobiosystems.com/category/recombinant-proteins/cd3-proteins)). For TCR-proximal epitopes, avidity effects, or clones raised against cells, screen on the full TCR-CD3 complex (recombinant or cell-surface, e.g., Jurkat or primary T cells).

Practical consequence: KD measured on monomer vs heterodimer vs cells can differ by orders of magnitude; an attenuated arm may look weak on monomer yet bind cells well via avidity. Rank arms on heterodimer kinetics by SPR/BLI, confirm by cell binding (FACS), and always run the human *and* cynomolgus heterodimers if non-human-primate tox is planned ([cyno CD3εδ example](https://www.acrobiosystems.com/products/cd3e-cd3d-cdd-c52w4)).

Choose heterodimers as your primary screening antigen; escalate to full complex/cells only when epitope mapping, avidity ranking, or cross-reactivity demands it.

Decision 4 — Format consequences (BiTE vs IgG-like vs TriTAC) for reagents

• BiTE (tandem scFv, ~55 kDa, no Fc): short half-life — blinatumomab's is reported at ~2 h, historically requiring continuous infusion ([Frontiers DLL3/TCE review](https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2025.1592291/full)). No FcγR or FcRn considerations; renal clearance dominates. Reagent stack: heterodimer SPR + bridging ELISA only.
• IgG-like TCE (Fc-bearing): mAb-like half-life (reported days-to-weeks) via FcRn; requires Fc-silencing mutations (e.g., LALA, N297A) to prevent FcγR-mediated off-target T-cell activation. Reagent stack adds FcγR-binding panels, FcRn assays, and silenced-Fc controls in every functional assay.
• TriTAC / half-life-extended scFv (anti-albumin domain): no Fc but albumin binding extends half-life to roughly a week (reported for HPN217/HPN328; [genOway commentary on HPN328](https://www.genoway.com/science/publications/commentary/molloy-molecular-cancer-therapeutics-2024)). Reagent stack adds albumin-binding SPR and HSA-supplemented assay buffers.

Choose the format that matches your PK/toxicity budget, then build the reagent stack from the corresponding row — the CD3-arm screening itself is format-agnostic.

Tool set this implies (proteins/assays a program needs)

• CD3εδ and CD3εγ heterodimers, human + cynomolgus, tagged and biotinylated, MALS-verified — primary SPR/BLI analyte and capture reagent ([human εγ example](https://www.acrobiosystems.com/products/cd3e-cd3g-cdg-h52w9)).
• Monomeric CD3ε ECD — QC and linear-epitope binders only.
• Full TCR-CD3 complex / CD3+ cells (Jurkat, primary T cells) — FACS binding, avidity confirmation, functional readouts.
• Reference arm panel — OKT3, UCHT1, SP34 clones as KD and epitope-binning comparators.
• Functional panel — TDCC killing (target + effector co-culture), NFAT-luciferase Jurkat reporter, cytokine release (IL-2, IFN-γ, IL-6, TNF-α) for CRS prediction, CD25/CD69 activation markers.

References

• [Salmerón et al. — Conformational CD3ε epitope expressed upon association with CD3δ/γ (J Immunol 1991)](https://europepmc.org/article/MED/1717585)
• [Characterization of anti-CD3 antibodies in clinically available bispecific T-cell engagers](https://snu.elsevierpure.com/en/publications/characterization-of-anti-cd3-antibodies-in-clinically-available-b/)
• [CD3 affinity attenuation in BsAb design (Expert Opin Drug Discov, 2025)](https://www.tandfonline.com/doi/pdf/10.1080/17460441.2025.2522088)
• [Frontiers — DLL3/BiTE landscape and TCE formats (2025)](https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2025.1592291/full)
• [genOway commentary — HPN328 TriTAC (Mol Cancer Ther)](https://www.genoway.com/science/publications/commentary/molloy-molecular-cancer-therapeutics-2024)
• [ACROBiosystems — CD3 protein line for BsAb screening](https://www.acrobiosystems.com/category/recombinant-proteins/cd3-proteins)
• [ACROBiosystems — CD3ε/CD3γ heterodimer (MALS verified)](https://www.acrobiosystems.com/products/cd3e-cd3g-cdg-h52w9)
• [ACROBiosystems — cyno CD3ε/CD3δ heterodimer](https://www.acrobiosystems.com/products/cd3e-cd3d-cdd-c52w4)