Cancer Immunotherapy · asia
Providing T Cells With Two Activation Signals: Dual-Protein Engagers Attack Colorectal Cancer Organoids
Researchers used mesothelin and NECTIN2, respectively, to recognize cancer cells and deliver the CD3 activation signal and CD28 costimulatory signal to T cells. In vitro results showed that double-positive tumors were more susceptible to killing, but animal safety validation and human trials remain far off.
Solid tumors are not only difficult for immune cells to penetrate but also often lack sufficient stimulatory signals to sustain an attack. A research team in Thailand is now testing two complementary bispecific protein engagers that separately deliver the “activation” and “costimulatory” signals required by T cells, with both signals dependent on markers on the surface of colorectal cancer cells.
One end of the first engager recognizes the tumor-associated antigen mesothelin (MSLN), while the other binds CD3 on T cells. The second connects NECTIN2 on the surface of cancer cells with CD28 on T cells. This design effectively uses two molecular bridges to draw T cells closer to cancer cells while providing both the primary CD3 signal and the CD28 costimulatory signal, in an attempt to improve selectivity for tumors that express both antigens.
The researchers first analyzed MSLN and NECTIN2 expression in colorectal cancer and then established cell lines capable of stably producing the two engagers. In vitro binding assays showed that the two proteins could separately recognize their intended tumor antigens and T-cell receptors. Compared with T cells without added engagers, T cells stimulated with both engagers showed stronger cytotoxic activity against cancer cells with high MSLN and NECTIN2 expression.
The research team observed results in the same direction in patient-derived colorectal cancer organoids: killing of double-positive organoids increased. After coculture, T cells released effector molecules including TNF-α, IFN-γ, granzymes, perforin, and granulysin, while IL-6 did not increase. This finding supports that the dual engagers can indeed drive an immune response in a three-dimensional model that more closely reflects the characteristics of patients’ tumors, but it cannot be used to infer that cytokine-related toxicity will not occur in humans.
Organoids still cannot fully reproduce blood vessels, stroma, immunosuppressive cells, or drug distribution in the body. The research also does not currently provide data on efficacy in animals, cross-reactivity with normal tissues, long-term toxicity, or human dosing. In particular, CD28 costimulation may cause strong immune activation, and neither MSLN nor NECTIN2 is a marker exclusive to cancer cells. The safety margin will therefore be a central issue in subsequent development.
This work was presented at the WIC-APAC meeting held in Taipei from March 28 to 29, 2026, by Suyanee Thongchot of Mahidol University during the SITC-WIC Young Investigator Award session, and the full paper was published on July 29. The journal also noted that the version currently available has not yet undergone final copyediting. The next step will still require animal models to assess antitumor effects, tissue selectivity, and systemic immune risks before it can be determined whether this off-the-shelf protein therapy meets the conditions for entering clinical development.