A team led by Professor Min Peng at Tsinghua University has engineered CAR-T cells that hunt down and destroy tumor necrosis factor, the inflammatory protein at the center of rheumatoid arthritis. Writing in hLife, the researchers describe cells they call TNFR1TIF, which capture soluble TNF and pull it inside for degradation through receptor-mediated endocytosis. In mice carrying a human TNF transgene, a single infusion prevented and treated the disease at all stages, dropping serum human TNF close to normal levels and performing on par with repeated high-dose adalimumab.
The distinction the researchers draw matters. Today's first-line biologics, including adalimumab (Humira), neutralize TNF rather than eliminate it, and they require injections every two weeks for what can be decades. Many patients eventually develop anti-drug antibodies that blunt the benefit and lead to secondary treatment failure, and long-term TNF blockade raises infection risk, including reactivation of latent tuberculosis. Newer targeted protein degradation tools such as LYTACs and KineTACs also need repeat dosing and depend on the body's own degradation machinery. By outsourcing that job to an engineered cell population, the Tsinghua group aims at a one-time intervention with years of effect.
Getting there required solving a practical problem: conventional CAR-T cells usually fail to expand without lymphodepleting preconditioning, a harsh step that would be difficult to justify in a chronic inflammatory disease. The team used CRISPR-Cas9 to delete Bcor and Zc3h12a, and the resulting cells expanded and held steady in the blood of fully immunocompetent mice for more than a year with no conditioning at all. Their survival tracked directly with their target - in TNF-knockout animals, expansion did not happen. The CAR itself fuses the mouse TNFR1 ectodomain to CD28 and CD3 zeta signaling domains.
Twelve months of safety monitoring turned up no disruption of normal T cell balance, organ integrity, or the ability to fight bacterial infection, and an anti-Thy1.1 depleting antibody served as a kill switch to clear the cells on demand. That fail-safe would be a prerequisite for any move toward human testing. For now the work sits squarely in mouse models, and the roughly 18 million people living with rheumatoid arthritis will not see a living-drug alternative to injections until the approach clears far larger hurdles.