A new study from University College London reframes PTPN22, a long known autoimmune risk gene, as a cytoskeletal architect that physically restrains T cell activation through actin remodeling at the immunological synapse.
T cells do not switch on. They build. When a T cell encounters an antigen-presenting cell, it assembles a structure called the immunological synapse: a precise, ring-shaped zone of receptor clustering and actin scaffolding that gathers the molecular players needed to commit the cell to action. The actin network that holds that structure together is not decoration. It is the architecture of restraint. A new paper from University College London shows that a familiar autoimmune-risk gene, PTPN22, is part of that architecture, and that losing it changes the shape of the synapse itself.
The work, published in Science Signaling and summarized in Genetic Engineering and Biotechnology News, comes from Megan Joseph and colleagues at University College London. The team reports that PTPN22 partners with a cytoskeletal adaptor called PSTPIP1 to keep actin remodeling in check at the synapse. Strip PTPN22 out, and the geometry of activation tilts. The cell begins to respond to weaker inputs that a healthy T cell would ignore.
That last point is the part that matters for autoimmunity. PTPN22 has been on the autoimmune shortlist for years. A common variant, R620W, has been tied to higher risk of lupus, rheumatoid arthritis, type 1 diabetes, and several other conditions. The standard explanation has been biochemical: PTPN22 is a phosphatase, so it dampens the kinases that drive T-cell activation. The new work adds a second layer. PTPN22 also shapes the physical scaffold on which activation happens. When that scaffold is altered, the cell's response to weak antigens, including self-antigens that should be tolerated, becomes disproportionate.
The team tested this by engineering a clean system. They started with Jurkat cells, a T-cell line, and deleted the endogenous T-cell receptor, then reintroduced a single transgenic TCR tuned to recognize two peptide antigens with different affinities: pTax, a strong stimulus, and pHuD, a weaker one. Into that controlled background they knocked out PTPN22. With PTPN22 present, the cells responded to pTax but not pHuD. With PTPN22 gone, the same cells began to fire on pHuD as well. The threshold for activation had dropped.
To see why, the researchers turned to super-resolution DNA-PAINT imaging, a method that can resolve the actin cytoskeleton at the scale relevant to a single synapse. On glass surfaces coated with the activating antibodies anti-CD3 and anti-CD28, they mapped the location of filamentous actin, the actin-nucleation factor WASp, the related WASH complex, and PSTPIP1 itself.
In wild-type cells, the actin network at the synapse was distributed in a pattern consistent with a working, well-tuned signaling platform. In PTPN22-knockout cells, PSTPIP1 piled up at T-cell receptor clusters, where it would normally cycle through. Arp2/3, the molecular machine that branches actin filaments and pushes the membrane outward to meet the antigen-presenting cell, was disrupted. Dense central foci of F-actin appeared, an architecture the wild-type cells did not build. Calcium signaling, the cell's downstream readout of productive TCR engagement, ran hotter in the knockout, especially under low-affinity stimulation.
The picture that emerges is a structural one. PTPN22 is not only a phosphatase that turns off kinases. It is also a regulator of the actin scaffold that holds the synapse in its proper shape. PSTPIP1, better known from studies of the autoinflammatory disease PAPA syndrome, is the adapter through which PTPN22 does that work. When the brake is lost, the scaffold does not fail dramatically. It changes shape. The actin network becomes stiffer and more central. The cell becomes more sensitive to inputs that a normal T cell would not bother responding to.
For autoimmune biology, that distinction matters. A signaling defect and a structural defect can produce the same downstream consequence, more self-reactive T cells in the periphery, but they point to different therapeutic levers. If PTPN22 variants tilt the cytoskeleton before the kinase cascade even runs, then a drug that only blocks downstream kinases will not restore the original architecture. It will compensate for it.
The authors are careful about the distance between a Jurkat-based structural result and any clinical use. The data are cellular and imaging-based. Primary human T cells, in vivo autoimmunity, and patient-level consequences are not part of this paper. Joseph and colleagues frame the work as opening a cytoskeletal layer of immune control that can be mapped, not as a treatment lead. The translational note in the GEN summary, that the axis could eventually matter for cancer immunotherapy, where T cells are deliberately turned up rather than restrained, is a forward-looking comment by the authors, not a demonstrated effect.
What is now a single research group, one journal paper, and a small set of well-controlled experiments is the kind of finding that tends to attract follow-up. The questions are clear. Does the same structural defect show up in primary human T cells, especially those carrying the R620W variant? Do PSTPIP1 and PTPN22 interact at the synapse the same way in resting and activating T cells? And if the architecture can be restored, does autoimmunity cool down with it?
For now, the image that should stick is not a switch. It is a T cell reaching across a synapse, building a scaffold to feel a single peptide, and a phosphatase that makes sure the scaffold does not overbuild when the peptide is the cell's own.