Post-sepsis immunosuppression depends on NKT cell regulation of mTOR/IFN-γ in NK cells
Edy Y. Kim, … , Tal Shay, Michael B. Brenner
Edy Y. Kim, … , Tal Shay, Michael B. Brenner
Published March 10, 2020
Citation Information: J Clin Invest. 2020;130(6):3238-3252. https://doi.org/10.1172/JCI128075.
View: Text | PDF
Research Article Immunology Infectious disease

Post-sepsis immunosuppression depends on NKT cell regulation of mTOR/IFN-γ in NK cells

Abstract

As treatment of the early, inflammatory phase of sepsis improves, post-sepsis immunosuppression and secondary infection have increased in importance. How early inflammation drives immunosuppression remains unclear. Although IFN-γ typically helps microbial clearance, we found that increased plasma IFN-γ in early clinical sepsis was associated with the later development of secondary Candida infection. Consistent with this observation, we found that exogenous IFN-γ suppressed macrophage phagocytosis of zymosan in vivo, and antibody blockade of IFN-γ after endotoxemia improved survival of secondary candidemia. Transcriptomic analysis of innate lymphocytes during endotoxemia suggested that NKT cells drove IFN-γ production by NK cells via mTORC1. Activation of invariant NKT (iNKT) cells with glycolipid antigen drove immunosuppression. Deletion of iNKT cells in Cd1d–/– mice or inhibition of mTOR by rapamycin reduced immunosuppression and susceptibility to secondary Candida infection. Thus, although rapamycin is typically an immunosuppressive medication, in the context of sepsis, rapamycin has the opposite effect. These results implicated an NKT cell/mTOR/IFN-γ axis in immunosuppression following endotoxemia or sepsis. In summary, in vivo iNKT cells activated mTORC1 in NK cells to produce IFN-γ, which worsened macrophage phagocytosis, clearance of secondary Candida infection, and mortality.

Authors

Edy Y. Kim, Hadas Ner-Gaon, Jack Varon, Aidan M. Cullen, Jingyu Guo, Jiyoung Choi, Diana Barragan-Bradford, Angelica Higuera, Mayra Pinilla-Vera, Samuel A.P. Short, Antonio Arciniegas-Rubio, Tomoyoshi Tamura, David E. Leaf, Rebecca M. Baron, Tal Shay, Michael B. Brenner

×

Figure 1

Early IFN-γ correlates with susceptibility to secondary Candida infection in clinical sepsis and experimental endotoxemia.

Options: View larger image (or click on image) Download as PowerPoint
Early IFN-γ correlates with susceptibility to secondary Candida infectio...
(A) Plasma IFN-γ levels in early clinical sepsis (within 48 hours of diagnosis). Patients are grouped by absence of secondary Candida infection (n = 19) or later development of secondary Candida infection (n = 12). (B) Survival curve for WT mice (n = 10–16 per group) after primary endotoxemia followed by secondary low-dose Candida i.v. at 1, 3, or 7 days after LPS. (C) Survival curve for WT mice (n = 10 per group) treated with higher- or lower-dose endotoxemia followed by secondary low-dose Candida i.v. (D) K-means clustering of transcriptomic (RNA-Seq) analysis of splenic Ly6c+ and Ly6c– macrophages from WT mice during endotoxemia (n = 2, not repeated). (E) Mean expression of transcripts from clusters from D with increased expression at 3 or 18 hours after endotoxemia. (F) Correlation of the transcriptomes of endotoxemia and IFN-γ response. Using RNA-Seq data from D for splenic Ly6c+ macrophages, the fold change in expression at 3 hours after LPS in vivo versus naive was calculated for each gene. The fold change in gene expression at 2 hours after IFN-γ treatment of macrophages in vitro versus untreated was calculated. The fold changes for 2 hours post–IFN-γ/naive versus 3 hours post-LPS/naive are plotted (left). A similar analysis for 18 hours is shown (right). Median ± IQR (A) and mean ± SEM (E) are shown. (A) Mann-Whitney test. (B and C) Log-rank test. *P < 0.05; ***P < 0.001.