New Bispecific Antibody Redirects T Cells to Attack Melanoma Cells

Melanoma is one of the most aggressive forms of skin cancer, and although immune checkpoint inhibitors have transformed outcomes for many patients, a substantial fraction of those treated never achieve durable benefit. These antibodies work by releasing the brakes on the immune system, but they are not tumor-specific, and they do not directly physically connect immune cells to cancer cells. The result can be serious inflammatory side effects alongside incomplete tumor control. A research team at University Hospital Tübingen in Germany has now reported the generation and detailed characterization of a new class of T-cell redirecting molecules designed to address precisely this gap, targeting a surface antigen called neuron glial antigen 2, or NG2, which is expressed largely in a tumor-restricted manner on melanoma cells.

The work, published in the Journal of Translational Medicine, describes the construction of NG2-targeting bispecific antibodies, engineered molecules that carry two different binding arms: one that latches onto NG2 on the surface of melanoma cells and another that binds CD3, a component of the T-cell receptor complex on immune cells. By bridging these two cell types, the antibodies force cytotoxic T cells into intimate contact with their tumor targets, triggering formation of an immune synapse and directing the full killing machinery of the T cell against the cancer cell. The strategy belongs to a growing family of T-cell engaging therapeutics that have already produced dramatic results in hematologic malignancies, but which have proven more difficult to deploy successfully against solid tumors like melanoma.

A central question in the design of such molecules is how strongly the CD3-binding arm should grip its target. Very high CD3 affinity can potently activate T cells, but it may also cause off-target activation, cytokine release, and exhaustion of the immune cells, whereas too little affinity may fail to engage the T cell at all. To explore this trade-off systematically, the Tübingen team generated two constructs that differ only in this parameter. The first, named NG2xCD3high, contains a CD3 binder based on the UCHT-1 antibody clone with high affinity for CD3. The second, NG2xCD3low, carries a UCHT-1 variant with approximately one hundred-fold reduced CD3 affinity. Both molecules share the same NG2-recognition domain and the same IgG-scFv antibody format, making them an unusually clean experimental pair for isolating the effect of CD3 affinity alone.

The researchers then put both constructs through a battery of functional tests using melanoma cell lines engineered or selected to express either high or low levels of NG2, thereby covering several key aspects of the T-cell mediated immune response. In co-cultures with NG2-expressing melanoma cells, both antibodies mediated the formation of immune synapses between T cells and tumor cells, induced T-cell activation, and drove T-cell proliferation. Crucially, tumor cell killing was strictly NG2-dependent: when the target cells lacked the antigen, the antibodies did not direct T cells against them, a property that underscores the safety rationale for choosing NG2 as a target in the first place.

The differences between the two constructs emerged most clearly when the researchers examined the quality and durability of the T-cell response. NG2xCD3high maintained robust cytokine production and sustained cytotoxicity across both NG2-high and NG2-low tumor models, meaning it remained effective even against melanoma cells that display comparatively modest amounts of the target antigen on their surface. NG2xCD3low, by contrast, induced less cytokine secretion and diminished functional activity overall, and this weakness became particularly pronounced against tumor cells with lower NG2 expression. In other words, within this particular antibody format, reducing CD3 affinity came at a real cost in anti-tumor potency that was not offset by gains in selectivity under the conditions tested.

Interestingly, both constructs supported the differentiation of stimulated T cells into memory T-cell subsets, a feature considered valuable for long-term anti-tumor immunity because memory cells can persist in the body and respond rapidly if the cancer returns. This observation suggests that T cells engaged by the NG2-directed antibodies are not merely consumed in a single burst of activity but can acquire properties associated with durable immune protection. Nevertheless, the quantitative advantage of the high-affinity construct in maintaining effector function led the team to select NG2xCD3high as the lead candidate for further development in NG2-expressing malignancies.

The choice of NG2 as the tumor target is itself a scientifically significant decision. Also known as chondroitin sulfate proteoglycan 4, NG2 is found on melanoma cells but is largely absent from essential normal tissues, offering the tumor specificity that checkpoint inhibitors lack. Because bispecific antibodies physically tether T cells to antigen-expressing tumor cells, the safety and efficacy of the approach depend heavily on how selectively that antigen is expressed. An antigen with tumor-restricted expression reduces the risk that redirected T cells will attack healthy tissue, while sufficient and consistent expression across tumor cells is needed for the therapy to reach every malignant cell. The Tübingen group’s demonstration that the high-affinity construct kills targets across a range of NG2 expression levels speaks directly to this second requirement.

The study also contributes to a broader and still unresolved debate in the bispecific antibody field about optimal CD3 affinity. Some engineers have argued that lower-affinity CD3 binders should preferentially activate T cells at the tumor site, where high antigen density brings the molecules into close proximity, thereby sparing circulating T cells and reducing systemic toxicity. The new data complicate that picture for solid tumor targets expressed at variable levels: in the IgG-scFv format used here, the one hundred-fold lower CD3 affinity translated into weaker activation, reduced cytokine output, and notably impaired killing of NG2-low tumor cells, without any reported gain in functional selectivity. The authors conclude that, within this format, high CD3 affinity is required to achieve potent effector function across differing NG2 expression levels.

For patients, the findings represent an early but concrete step toward a new therapeutic option for melanoma, a disease in which many individuals still fail to benefit durably from existing immunotherapies. The work was conducted at the Clinical Collaboration Unit Translational Immunology at University Hospital Tübingen, within the German Cancer Consortium partner site Tübingen and the Cluster of Excellence iFIT, which focuses on image-guided and functionally instructed tumor therapies. The study was approved by the ethics committee of the Faculty of Medicine of the Eberhard Karls Universität Tübingen and conducted in accordance with the Declaration of Helsinki, with human material collected after informed consent. Funding came from Deutsche Krebshilfe, the Wilhelm Sander-Stiftung, and the German Research Foundation, including support under Germany’s Excellence Strategy.

Much work remains before an NG2-directed bispecific antibody can be tested in patients, including preclinical safety studies, pharmacokinetic optimization, and the inevitable challenges of translating laboratory co-culture results into the complex immune environment of human tumors. Solid tumors present barriers that blood cancers do not: heterogeneous antigen expression, immunosuppressive microenvironments, and physical obstacles to T-cell infiltration. Yet the Tübingen study provides two important building blocks. It validates NG2 as a promising target for T-cell redirecting therapy in melanoma, and it delivers a lead candidate, NG2xCD3high, whose functional profile suggests it can handle the variability in antigen expression that real tumors present. As bispecific antibody platforms continue to mature, studies of this kind, which carefully dissect how engineering choices shape immune function, will guide the rational design of the next generation of cancer immunotherapies.

Subject of Research: NG2-targeting bispecific antibodies for T-cell redirecting immunotherapy of melanoma

Article Title: Generation of an NG2 targeting bispecific antibody for the induction of T-cell immunity against melanoma

Article References: Prakash, N., Hörner, S., Hagelstein, I., Keller, J., Wang, K., Jung, G., Salih, H. R., & Lutz, M. S. (2026). Generation of an NG2 targeting bispecific antibody for the induction of T-cell immunity against melanoma. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-09000-5

Image Credits: AI Generated

DOI: 10.1186/s12967-026-09000-5

Keywords: bispecific antibodies, NG2, CD3 affinity, melanoma, T-cell redirection, cancer immunotherapy, immune synapse, cytokines, memory T cells, tumor immunology, Journal of Translational Medicine, University of Tübingen