Reply to “Is IL-7 from dendritic cells essential for the homeostasis of CD4⁺ T cells?”

Mackall and Guimond respond:

In our article¹, we concluded that interleukin 7 (IL-7) production is regulated by a simple feedback loop and that IL-7 signaling on antigen-presenting cells controls CD4⁺ T cell homeostatic expansion. Marilyn HY-reactive T cells proliferated more in IL-7-deficient (Il7^−/−) recipients of bone marrow from female mice deficient in recombination-activating gene 1 (Rag1^−/−) than in Rag1^−/− recipients of such bone marrow (Fig. 3d in ref. 1). Surh et al.² were surprised by the high degree of proliferation, could not reproduce our findings and assumed that contamination by bone marrow from male mice occurred in our experiments. However, we controlled these experiments by administering the same pool of marrow cells to Rag1^−/− recipients, which did not support proliferation, thus ruling out the possibility of contamination by bone marrow from male mice (Supplementary Fig. 1a). Furthermore, Marilyn cells administered to male hosts underwent greater population expansion and dilution of the cytosolic dye CFSE (Supplementary Fig. 1a) than did those given to Il7^−/− recipients. Our original manuscript contained additional data that supported the central model. Marilyn cells proliferated after being transferred into Rag1^−/− recipients that received bone marrow from female mice deficient in the common γ-chain but not after being transferred into Rag1^−/− recipients that received bone marrow from female Rag1^−/− mice (Supplementary Fig. 1b). If contamination by bone marrow from male mice were to explain this, the bone marrow deficient in the common γ-chain (but not the Rag1^−/− bone marrow) would have been consistently contaminated across three independent experiments, which seems implausible. Furthermore, Marilyn cells proliferated in female mice deficient in the IL-7 receptor α-chain (Il7r^−/−; Supplementary Fig. 1c), which were not chimeric, and we obtained similar results with chimeras generated by the transfer of bone marrow from female Il7r^−/− mice into female Il7r^−/− recipient mice (Supplementary Fig. 1d). In Il7r^−/− mice and female Il7r^−/− chimeras, Marilyn cells show a homeostatic pattern of proliferation, which differed from the rapid proliferation observed in the Il7^−/− mouse (Supplementary Fig. 1a). We do not yet understand why the patterns differ, but we observed similar differences with polyclonal T cells (Supplementary Fig. 1e). Together, the multiple models demonstrating enhanced Marilyn cell proliferation when dendritic cells lack IL-7 signaling, the absence of Marilyn cell proliferation in Rag1^−/− recipients and the homeostatic proliferation pattern in Il7r^−/− recipients refute the notion that contamination by bone marrow from male mice confounded our conclusions¹.

Furthermore, we respectfully disagree with the suggestion that results obtained with polyclonal T cells are to be discounted. Polyclonal T cell proliferation in this system was IL-7 dependent (Supplementary Fig. 1f), and most polyclonal T cells, not just the fraction responsive to environmental antigens, proliferated in Il7^−/− hosts (Supplementary Fig. 1e). Given that the goal of mouse experiments is to model the human condition, important insights are gleaned from studies of natural T cells. Depleted mouse and human immune systems reconstitute repertoires by responding to the wide array of antigens presented during lymphopenia, including self antigens, cognate antigens and cross-reactive antigens. We clearly demonstrated greater homeostatic expansion of naturally occurring CD4⁺ T cells when stroma-derived IL-7 was absent¹.

We cannot explain why Surh et al. cannot replicate our data. Possibilities include differences in the number of bone marrow cells transferred, which perhaps limited reconstitution of the plasmacytoid dendritic cell niche that supports CD4⁺ T cell population expansion; larger numbers of transferred Marilyn cells, which can compete for cross-reactive antigen; or variation in the T cell antigen receptor (TCR)-transgenic strains, as our Marilyn mice do not express Thy-1.1. However, potential explanations for discordant results in one small set of experiments do not undermine the global conclusions we presented¹. Together, our data shed light on an entirely new axis of CD4⁺ T cell regulation, were the first to our knowledge to explain why the regeneration of CD4⁺ T cells in humans is exquisitely thymus dependent and opened new possibilities for the therapeutic manipulation of the homeostatic expansion of CD4⁺ T cells in humans.