Comment on “Interplay between TNF and Regulatory T Cells in a TNF-Driven Murine Model of Arthritis”

We read with great interest the article by Biton and colleagues (1) in the April 1, 2011 issue of The Journal of Immunology. Their study examined the interplay between TNF and regulatory T cells (Tregs), and concluded that anti-TNF treatment increased the number and function of Tregs in a mouse arthritis model that was driven by overexpression of human TNF (1). However, it is well established that human TNF can bind only to mouse TNFR1, but not TNFR2 (2). We and others have reported that TNF promotes Treg function (3, 4), and the opposite effect of TNF on Tregs has also been reported (5, 6). Nevertheless, all these studies agree that the effect of TNF on Tregs is mediated by TNFR2, which is predominantly expressed by human and mouse Tregs. Therefore, the mouse arthritis model based on overexpression of human TNF (1) cannot reflect the requisite interaction between TNF and Tregs through TNFR2. Furthermore, in the article by Biton et al, only CD4⁺Foxp3⁺CD25⁺ cells were identified as Tregs (1), which is not accurate because highly suppressive CD4⁺ Foxp3⁺CD25^- cells also exist in normal mice (7) and are increased at inflammatory sites (8, 9). CD25 as a surrogate surface marker of Tregs is also upregulated on activated CD4⁺Foxp3^- effector T cells. Therefore, isolated CD4⁺ CD25⁺ cells in an inflammation model, as in the study by Biton et al. (1), may not accurately reflect the function of bona fide Tregs. Taken together, the inappropriate mouse model and inaccurate methods in the identification of Tregs used in this study do not support the authors’ conclusion that anti-TNF treatment augments the number and function of Tregs.