Immunotherapy relies on a straightforward premise: program or unleash immune cells to recognize and destroy specific surface markers on target cells. However, a severe biological bottleneck arises when the chosen target marker is also present on the therapeutic cells themselves—either inherently or as a consequence of immune activation. The result is therapeutic fratricide, a destructive phenomenon where engineered immune cells or targeted antibodies inadvertently trigger the elimination of the very cells meant to fight the disease.
Understanding why immunotherapies turn on themselves requires looking at how these therapies interact at the cellular level, and how researchers are using precise genetic engineering to break these self-destructive loops.
The Shared Antigen Dilemma in Cellular Therapies
In cellular therapies like Chimeric Antigen Receptor (CAR) T or Natural Killer (NK) cell therapy, immune cells are engineered with a synthetic receptor designed to bind a specific tumor antigen. When treating hematological malignancies like T-cell leukemias or myeloid leukemias, a fundamental problem emerges: the cancer cells share the same lineage and surface markers as the healthy therapeutic cells.
[CAR-T Cell A] --(CAR Receptor)--> [CAR-T Cell B] (Expresses Target Antigen)
│ │
└─────────────── Kills ─────────────┘
During the manufacturing expansion phase or after infusion into a patient, a CAR-T cell's receptor will inevitably encounter its target antigen on a neighboring CAR-T cell. This spatial proximity triggers a cytotoxic release, causing mutual destruction. As reviewed by Zhang et al., this self-targeting severely depletes the population of viable therapeutic cells, limiting the potency and persistence of the treatment.
Hu notes that this constraint has historically made developing cellular therapies for T-cell malignancies far more challenging than for B-cell cancers (where the target, such as CD19, is absent on therapeutic T cells). For example, Liu et al. highlight how CD7-directed CAR-T cells aggressively attack one another during the in vitro expansion phase, crippling the manufacturing pipeline before the cells can even be delivered to a patient.
The Activation Trap and Paradoxical Feedback
Fratricide is not limited to baseline shared lineage markers; it can also be dynamically induced by the very process of immune activation. To overcome tumor-induced immunosuppression, some therapies target immune checkpoints or regulatory pathways. However, T cells naturally upregulate these exact markers when they become active.
Consider the CAR-T cells engineered by Qin et al. to target PD-L1, a ligand tumors use as a shield to evade immunity. While highly effective at eliminating tumor cells, these CAR-T cells face a delayed trap. As the therapeutic T cells successfully engage the tumor and become activated, they naturally upregulate PD-L1 on their own surfaces. The anti-PD-L1 CAR-T cells then perceive their active, non-malignant allies as threats, executing them and sabotaging the coexistence required for effective combination therapies.
A parallel form of self-undermining crosstalk occurs via cytokine signaling pathways. As demonstrated by Geels et al., the activation of effector CD8+ T cells can trigger the production of IL-2, which inadvertently feeds and expands immunosuppressive regulatory T cells (Tregs) via the ICOS pathway. This creates a paradoxical feedback loop where the therapy's success in activating effector cells simultaneously fuels the mechanisms of its own resistance.
Macrophage-Mediated Collateral Damage
Fratricide does not only occur cell-to-cell among engineered lymphocytes; it can also be mediated by the innate immune system via therapeutic antibodies. This occurs through two primary mechanisms: physical crosslinking and the removal of protective "don't eat me" signals.
Fc Receptor Crosslinking
Monoclonal antibodies are frequently designed to block inhibitory receptors on T cells (such as PD-1) to keep them active. However, an antibody consists of both a targeting "head" (variable region) and a structural "tail" (Fc region).
Song et al. demonstrated that standard IgG4 anti-PD-1 antibodies can inadvertently form a physical bridge between a T cell's PD-1 receptor and the FcγRI receptor on nearby macrophages. This crosslinking triggers Antibody-Dependent Cellular Phagocytosis (ADCP). Instead of protecting the T cell, the antibody acts as an opsonin, causing the macrophage to engulf and consume the very tumor-fighting T cell the drug was meant to unleash.
[T Cell] --(PD-1) <== [Anti-PD-1 Antibody] ==> (FcγRI) -- [Macrophage]
│
▼
Triggers Phagocytosis (ADCP) of T Cell
Stripping the "Don't Eat Me" Signal
A similar vulnerability involves CD47, a surface protein that acts as a universal "don't eat me" signal to protect cells from macrophage phagocytosis. Therapies targeting CD47 aim to block this signal on tumor cells to encourage their clearance.
However, Yamada‐Hunter et al. discovered that adoptively transferred therapeutic T cells also rely heavily on CD47 to survive macrophage surveillance. When anti-CD47 antibodies block this signal indiscriminately, macrophages rapidly clear the therapeutic T cells in vivo, destroying the synergy required for combination immunotherapies.
Engineering Solutions to Break the Loop
To make these therapies viable, researchers must decouple the therapeutic targeting mechanism from self-recognition. The most direct solution is the genetic ablation of the shared antigen within the therapeutic cells themselves, rendering them "invisible" to their own targeting systems.
CRISPR Gene Knockouts
By using CRISPR/Cas9 to knock out the gene encoding the target antigen in therapeutic cells, researchers can completely prevent direct fratricide:
- CD5 and CD7: Dai et al. engineered biepitopic CARs targeting CD5 but incorporated a CD5 knockout in the CAR-T cells, successfully mitigating self-fratricide. The authors expanded this strategy in a subsequent study (Dai et al.) by developing a bispecific CAR targeting both CD5 and CD7 combined with a double knockout of both antigens, preventing mutual destruction while simultaneously mitigating tumor antigen escape.
- CD33: Similarly, Wang et al. utilized CRISPR to knock out CD33 in induced natural killer (iNK) cells engineered with a "Loop CAR" targeting CD33 and mesothelin. This modification prevented fratricide and enabled a scalable, off-the-shelf therapy for acute myeloid leukemia.
Signal Engineering
For antibody-mediated clearance, the solution involves protecting therapeutic cells from macrophage recognition without compromising tumor targeting. To solve the CD47 clearance problem, Yamada‐Hunter et al. engineered T cells to express a modified, antibody-resistant version of CD47. This engineered shield maintains the vital "don't eat me" signal even in the presence of anti-CD47 therapies, protecting the T cells from premature macrophage clearance and unlocking potent anti-tumor synergy.
Key Foundational Work
Chimeric antigen receptor T cells targeting PD-L1 suppress tumor growth (2020) — contributed 3 concept elements to this direction.
The binding of an anti-PD-1 antibody to FcγRΙ has a profound impact on its biological functions (2018) — contributed 3 concept elements to this direction.