UCSF

For thousands of years, therapeutics consisted almost exclusively of molecules that up- or down-regulated various pathways. Even as methods were developed for complex and systematic reprogramming of biological systems, such tools were long employed only for research purposes in the laboratory. However, recent advances in synthetic biology and immunology have given rise to the era of therapeutic cell engineering, where complex diseases are now being treated through artificial reprogramming of human cells. Clinical-stage technologies such as chimeric antigen receptor (CAR) T cells have opened the door to this new paradigm in medicine, which holds the promise of safer, more effective therapeutics with capabilities far beyond those of molecules. However, extant cell-based therapies have only scratched the surface of the sophistication that is possible: natural cellular programs exemplify phenomenal complexity and underpin all of life and its diseases. Solutions to many of the unsolved challenges in biology and medicine may be within reach if we can learn to reprogram our own cells with the right capabilities. Extending the programmability of human cells is the fundamental motivation for this work.Here, we developed synthetic circuits that rewire how immune cells communicate with one another. We used these circuits both to explore the design principles underlying the natural architecture of the immune system, asking why evolution converged on the native solutions, and further to control immune responses in therapeutically useful ways in the context of cancer and autoimmune diseases. First presented is a systematic study of strategies for synthetically delivering cytokines to enhance antitumor T cell responses against solid tumors. Synthetic Notch receptor (synNotch) based circuits that detect a tumor antigen and induce local production of the cytokine IL-2 enabled engineered T cells to infiltrate and clear immune-excluded pancreatic and melanoma tumors that were otherwise refractory to treatment, without the systemic toxicity associated with IL-2 administration. We identified design principles that were essential for antitumor efficacy: only an autocrine configuration, in which cytokine production and cytotoxic function are housed in the same cell, was effective in the presence of an intact host immune system that competes for the shared cytokine. Second presented is an exploration of the tradeoffs of selfish versus shared T cell cytokine signaling, in which IL-2 was engineered with non-native signaling properties and employed both to probe the role of shared signaling and to build therapeutic T cell control circuits to treat cancer and autoimmune disease. We engineered autoIL-2, a variant of IL-2 restricted to cell-autonomous (selfish) signaling, and found that replacing native IL-2 with autoIL-2 yields T cells that proliferate normally but are entirely refractory to regulatory T cell suppression, driving unrestrained autoimmunity in vivo. This is a direct demonstration that the shared character of IL-2 is what renders T cell amplification governable. Placed instead under the control of an orthogonal synNotch receptor receptor, the same molecule confers capabilities no natural T cell possesses: tissue-targeted expansion and persistence without collateral inflammation. SynNotch to autoIL2 circuit T cells exhibited amplification that regulatory T cells cannot intercept, and markedly improved efficacy of anti-inflammatory payload-delivering T cells in a model of neuroinflammatory disease, a setting where wild-type IL-2 substantially worsened disease. Together, these studies expand our understanding of how and why immune cells communicate through cytokine signaling, and extend the repertoire of synthetic tools for reprogramming immune cells with therapeutically useful properties.