The rise of IL-2 therapy — a picture beyond T_reg cells

We thank Dr Jens Y. Humrich and Dr Gabriela Riemekasten for their interest in our study¹, which they discussed in a News & Views commentary (Humrich, J. Y. & Riemekasten, G. The rise of IL-2 therapy — a novel biologic treatment for SLE. Nat. Rev. Rheumatol. 12, 695-696; 2016)². We are particularly excited that more clinical and basic scientists are working to understand the mechanisms of action underpinning the promising 'experimental medicine' observations of low-dose IL-2 therapy in systemic lupus erythematosus (SLE). After reading the commentary, however, we noted that there was a misunderstanding of the research design of our study. We hereby provide further clarification.

The inspiration for this study of low-dose IL-2 therapy in autoimmune diseases came from our previous publication showing that hyperactivation of follicular helper T (T_FH) cells correlated with disease activity in patients with SLE and rheumatoid arthritis³; this is a different focus from the studies of low-dose IL-2 therapy, which centred on regulatory T (T_reg) cells^4,5,6. Although IL-2 had been shown to suppress T_FH and type 17 T helper (T_H17) cells in mouse models^7,8, its effects in human cells were unknown. We started our study in 2013, with clinical responses as the primary end point and an emphasis on immunological responses, including changes in T_FH, T_H17 and T_reg cells, as secondary end points⁹. For the first time, we demonstrated that low-dose IL-2 therapy could suppress T_FH and T_H17 cells in humans¹. In addition, using a mouse model, we revealed that suppression of T_FH and T_H17 cells was as sensitive to low-dose IL-2 as the promotion of T_reg cells. During the peer review process, we were asked to exclude the possibility that the increase in T_reg cells and decreases in T_FH and T_H17 cells accompanying low-dose IL-2 administration were a consequence of significantly lowered disease activity. Therefore, we analysed the immunological phenotype in another, separate cohort of patients who underwent a comparable response under conventional immunosuppressive treatments, and found no such changes in T_reg, T_FH and T_H17 cells. This separate cohort was not a placebo-control group and we agree with the proposal of Drs Humrich and Riemekasten that a parallel study would be ideal.

Our study characterized T_reg cells based on a CD4⁺CD25^highCD127^low phenotype, a method reported in 2006 (Refs 10,11) and cited in thousands of publications. Drs Humrich and Riemekasten suggest CD4⁺FOXP3⁺CD127^low, a less well-characterized phenotype, could be better. Notably, their own study indicated a substantial proportion of CD4⁺FOXP3⁺CD127^low cells did not express CD25 but secreted effector cytokines such as IFNγ, suggesting possible contamination with conventional effector T cells¹². This agrees with the results of many studies on human samples showing that FOXP3 is expressed in activated effector CD4⁺ T cells in addition to T_reg cells¹³. Owing to the limitations of the available phenotypical markers used to characterise T_reg cells, we also performed a standard suppressive assay showing that low-dose IL-2 treatment in mice and humans promoted the suppressive function of T_reg cells¹.

SLE is a very complex and poorly treated disease. As low-dose IL-2 emerges as a new therapy for this disease, we agree that more studies, including well-controlled randomized trials, will be needed to understand the underlying mechanisms, optimise treatment regimens and select the most suitable patients. These important follow-up studies are currently underway.