Our first-generation MUC1-specific CAR, termed S28z, contained an SM3 scFv fused to a CD28 hinge. Surprisingly, however, S28z grafted T cells were poorly activated by a MUC1 + IgG fusion protein or MUC1 expressing T47D breast cancer cells. By contrast, S28z enabled T-cell activation when the MUC1 epitope was presented as a crosslinked peptide. Together, these findings suggested that steric hindrance and/or poor access to the epitope are limiting factors in CAR-based targeting of MUC1. To overcome this, a flexible monomeric hinge derived from IgD was introduced, thereby creating SD28z. Despite reduced stability, the SD28z CAR enabled T cells to proliferate in response to MUC1 glycoforms found in breast cancer. Stability of SD28z was further improved by inclusion of IgG1 Fc sequences in the extracellular domain (giving SDF28z). SDF28z exhibited greater functional activity, enabling T cells to kill T47D tumour cells. In a second approach to optimize function, a scFv was cloned from the MUC1-specific HMFG2 hybridoma. HMFG2 binds breast tumour cells with greater intensity than SM3. In keeping with this, all HMFG2-derived CAR exhibited greater functional activity than their SM3 counterparts. In the MUC1-specific CAR that exhibits greatest activity (HDF28z), an HMFG2 scFv has been fused to the IgD hinge and IgG1 Fc (HDF28z). HDF28z grafted human T cells exhibit potent cytolytic activity against MUC1 expressing breast cancer cells, associated with cytokine production and subsequent T-cell clonal expansion.