LAUSR

Regulatory T (Treg) cells dampen immune responses. Augmenting Treg cells — either numerically or functionally — could therefore be therapeutic in autoimmune diseases. Work from the Sakaguchi and Narumiya labs now shows that small-molecule inhibitors of CDK8 and CDK19 convert antigen-specific conventional T (Tconv) cells into Treg cells, which reduces autoimmunity in mouse models of multiple sclerosis, skin contact hypersensitivity and diabetes. Treg cells develop either in the thymus (tTreg cells) or in the periphery (pTreg cells). Previous attempts to develop an ex vivo method to generate induced Treg (iTreg) cells required administration of either transforming growth factor-β (TGFβ) or molecules that induce TGFβ expression, which would limit the clinical usefulness of such approaches because of toxicity. Furthermore, iTreg cells can only be generated from naive T cells, and proinflammatory cytokines, which are present at sites of autoimmune reactions, scupper this process. The authors sought an approach to generate Treg cells that could be therapeutically viable and so homed in on converting Tconv cells. First, Tconv cells are accessible and present both systemically and at sites of inflammation. Furthermore, some of these cells would have been exposed to autoantigens in patients with autoimmune conditions, so converting these antigenexposed cells could generate antigen-specific Treg cells. The authors first screened a 5,000-compound library to identify molecules that can convert Tconv cells into Treg cells, as determined by the expression of FOXP3, a defining characteristic of Treg cells. AS2863619 (AS) induced this conversion. In order for AS to be therapeutically useful, it would have to work in the presence of proinflammatory cytokines. Indeed, this compound could induce either naive T or effector or memory Tconv cells to become Treg cells, in a TGFβ-independent manner, in the presence of the proinflammatory cytokine IL-12, IL-4 or IL-6. Affinity purification followed by mass spectrometry pulled out CDK8 and CDK19 as two direct targets of AS; these targets were confirmed to be relevant as AS had no effect in cells deficient in these two proteins. Furthermore, overexpression of dominant-negative versions of CDK8 and CDK19 could induce FOXP3 expression. Molecular work pinpointed signal transducer and activator of transcription 5 (STAT5) as the intermediary protein: AS induced phosphorylation of STAT5 at the known CDK8 phosphorylation site and subsequent binding of STAT5 to its known transcriptional promoters, including the FOXP3 promoter. The investigators then moved to mouse models. Simultaneous administration of AS and immunization with the OVA peptide induced FOXP3+ Treg cells that suppressed OVA-specific Tconv cells. This occurred even in mice with no tTreg cells, suggesting that the induced Treg cells were derived from Tconv cells. In a model of skin contact hyper sensitivity, induced by 2,4-dinitrofluorobenzene, AS treatment after sensitization dampened the secondary response in a Treg celldependent manner. Notably, some of these Treg cells co-expressed KLRG1, a marker for antigen-experienced cells, suggesting that the Treg cells may be derived from self-reactive Tconv cells. Similarly, in nonobese diabetic mice, treatment with AS reduced the incidence and severity of spontaneous diabetes. Finally, AS treatment reduced the severity of experimental autoimmune encephalomyelitis, a model of multiple sclerosis. In all of these models, more Treg cells and fewer diseaserelevant types of Tconv cell were found in the affected tissues or regional lymph nodes. Inhibition of CDK8 and CDK19 could be an effective way to convert Tconv cells into Treg cells — a long sought-after goal for the treatment of autoimmune diseases.