Current treatments for autoimmune diseases mostly rely on general immunosuppression, leaving patients prone to infections. Approaches that specifically target pathogenic autoimmune cells in autoimmune diseases would be ideal; however, strategies to achieve this have so far been elusive. Now, reporting in Science, Ellebrecht et al. show that chimeric antigen receptor T cells (CAR-T cells) — which are in clinical trials as novel anticancer agents — can be tweaked to seek out and eliminate self-reactive B cells.

Credit: Ikon Images/Alamy Stock Photo

The authors investigated the life-threatening autoimmune blistering disease pemphigus vulgaris (PV), which is caused by autoreactive antibodies targeted at the keratinocyte adhesion protein desmoglein-3 (DSG3). Deletion of B cells with CD20-targeted antibodies can induce short-term remission, but most patients relapse and the treatment leaves them vulnerable to fatal infections. In relapsed patients, the disease-causing DSG3-reactive B cell clones re-emerge, whereas disease remission is associated with the complete disappearance of these clones.

This observation led the authors to investigate whether CAR-T cells, which have engineered receptors that enable the cells to specifically detect and kill cells bearing a particular surface antigen, could be used to eradicate DSG3-reactive B cell clones. They engineered a CAR using several antigenic epitopes of DSG3 in the extracellular domain, fused to an intracellular signalling domain that has been used in CAR-T cell constructs that have shown clinical efficacy in cancer. The resulting chimeric autoantibody receptor (CAAR) was cloned into T cells, creating DSG3 CAAR-T cells. In vitro experiments showed that these CAAR-T cells selectively kill DSG3-reactive B cells while sparing B cells with other specificities.

As PV patients have polyclonal circulating anti-DSG3 antibodies that might either neutralize or stimulate DSG3 CAAR-T cell cytotoxicity, experiments were repeated in the presence of anti-DSG3 antibodies. High-affinity antibodies appeared to mildly decrease CAAR-T cytotoxicity, whereas lower-affinity antibodies had no effect or even increased cytotoxicity, leading the authors to conclude that circulating anti-DSG3 antibodies will not prevent and may even promote CAAR-T cell efficacy.

The treatment decreased autoreactive antibody titres, abrogated autoantibody binding and prevented disease manifestation

Next, the DSG3 CAAR-T cells were investigated in a model of PV. This was created by injecting a polyclonal mixture of bioluminescent anti-DSG3 B cell hybridomas into NOD-scid-gamma mice, followed by CAAR-T cells 5 days later. The treatment decreased autoreactive antibody titres, abrogated autoantibody binding and prevented disease manifestation. In a PV mouse model created with human CD19+ B cells that expressed anti-DSG3 B cell receptors cloned from PV patients, the CAAR-T cells showed similar efficacy to anti-CD19 CAR-T cells that have been used for clinical studies in B cell malignancies.

Importantly, further in vitro and in vivo experiments showed that DSG3 CAAR-T cells did not affect keratinocytes; these cells express desmocolins and desmogleins, the physiological binding partners of DSG3. Moreover, in mice that received human skin xenografts, DSG3 CAAR-T cells had no skin toxicity or any other off-target toxicity against other tissues, and did not react against immune cells that might bind to CAAR-bound anti-DSG3 antibodies via their Fc receptors.

Overall, these studies suggest that CAAR-T cells could be developed as a targeted approach to treat autoantibody-mediated immune diseases while avoiding general immunosuppression.