T helper type 2-polarized invariant natural killer T cells reduce disease severity in acute intra-abdominal sepsis

Abstract

Sepsis is characterized by a severe systemic inflammatory response to infection that is associated with high morbidity and mortality despite optimal care. Invariant natural killer T (iNK T) cells are potent regulatory lymphocytes that can produce pro- and/or anti-inflammatory cytokines, thus shaping the course and nature of immune responses; however, little is known about their role in sepsis. We demonstrate here that patients with sepsis/severe sepsis have significantly elevated proportions of iNK T cells in their peripheral blood (as a percentage of their circulating T cells) compared to non-septic patients. We therefore investigated the role of iNK T cells in a mouse model of intra-abdominal sepsis (IAS). Our data show that iNK T cells are pathogenic in IAS, and that T helper type 2 (Th2) polarization of iNK T cells using the synthetic glycolipid OCH significantly reduces mortality from IAS. This reduction in mortality is associated with the systemic elevation of the anti-inflammatory cytokine interleukin (IL)-13 and reduction of several proinflammatory cytokines within the spleen, notably interleukin (IL)-17. Finally, we show that treatment of sepsis with OCH in mice is accompanied by significantly reduced apoptosis of splenic T and B lymphocytes and macrophages, but not natural killer cells. We propose that modulation of iNK T cell responses towards a Th2 phenotype may be an effective therapeutic strategy in early sepsis.

Keywords: Th2 response; acute intra-abdominal sepsis; glycolipids; invariant natural killer T cells.

Fig. 1

Characterization of invariant natural killer T (iNK T) cell populations in the blood samples of critically ill patients. (a) Representative flow cytometry plots of peripheral blood sampled from a septic and non-septic patient. (b) Histograms (median ± standard error of the mean) comparing frequency of T cells, NK cells and iNK T:T cell ratios in septic and non-septic patients in the intensive care unit. *P < 0·05 by Mann–Whitney U-test. (c) Kaplan–Meier survival curves from time of blood collection to time of discharge.

Fig. 2

Invariant natural killer T (iNK T) cells are pathogenic in intra-abdominal sepsis (IAS). (a) B6 and iNK T cell-deficient Jα18^–/– mice were injected with faecal slurry (90 mg/ml) to induce IAS and monitored during the experimental time-line. Murine sepsis scores were significantly higher compared to sham-treated B6 and Jα18^–/– mice [injected intraperitoneally with normal saline (NS)] and Jα18^–/– mice with IAS (n = 6 for sham B6 mice, n = 6 for Jα18^–/– mice, n = 10 for septic B6 mice, n = 6 for septic Jα18^–/– mice). ***P < 0·001 by two-way analysis of variance test. (b) Mortality for B6 mice with IAS were significantly higher than sham B6 and Jα18^–/– mice, as well as septic Jα18^–/– mice (n = 6 for sham B6 mice, n = 6 for Jα18^–/– mice, n = 10 for septic B6 mice, n = 6 for septic Jα18^–/– mice). ***P < 0·001 by log-rank test.

Fig. 3

Tissue-specific distribution of invariant natural killer T (iNK T) cells is altered during intra-abdominal sepsis (IAS). (a) The distribution of CD3⁺ T and CD3⁺CD1d tetramer⁺ iNK T cells in the spleen and omentum is altered significantly in IAS, but remains unchanged in the liver (n = 7, n = 10 in sham and IAS groups, respectively). Percentages of cell populations are represented as means ± standard errors of the mean. **P < 0·001; * P < 0·05 by Mann–Whitney U-test (b) Quantitative reverse transcription–polymerase chain reaction (RT–PCR) using custom-designed probes designed for detecting the invariant T cell receptor (TCR) demonstrates a significantly elevated expression of TCR transcripts within the spleen, liver and omentum following IAS. Relative fold changes of the expression of the invariant TCR were calculated based on the ΔΔCt method after normalizing to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as an internal control (n = 3, n = 5 in sham and IAS groups, respectively). ***P < 0·0001; **P < 0·001; *P < 0·05 by Mann–Whitney U-test.

Fig. 4

T helper type 2 (Th2)-polarizing glycolipid OCH reduces disease severity in intra-abdominal sepsis (IAS). C57BL/6 (B6) mice were injected concomitantly intraperitoneally with faecal slurry (FS; 500 μl of 90 mg/ml solution) to induce intra-abdominal sepsis (IAS), and vehicle, OCH or KRN7000. (a) OCH-treated mice had prolonged survival significantly compared to vehicle- and KRN7000-treated mice (n = 19, n = 15, n = 8 for OCH, vehicle and KRN7000 groups, respectively). ***P < 0·001 by log-rank test. (b) OCH-treated mice demonstrated significantly reduced disease severity compared to vehicle-treated and KRN7000-treated mice (n = 19, n = 15 and n = 8 mice, respectively, for OCH, KRN7000, and vehicle groups). ***P < 0·001 by two-way analysis of variance (ANOVA) with Bonferroni post-test. (c) iNK T-deficient Jα18^–/– mice were given FS (500 μl of a 90 mg/ml solution) to induce IAS and treated concomitantly with OCH or vehicle. Murine sepsis scores were similar between vehicle and OCH-treated mice (n = 3 per group). (d,e) Administration of OCH and KRN7000 resulted in significantly reduced detection of CD3⁺CD1d tetramer⁺ invariant natural killer T (iNK T) cells among septic B6 mice compared to vehicle treatments. The percentages of CD3⁺ T cells remained unchanged with administration of iNK T-specific glycolipid agonists (n = 6, n = 4, n = 6, and n = 3 for vehicle, OCH, vehicle (KRN7000) and KRN7000 groups, respectively). *P < 0·05; **P < 0·01 by Mann–Whitney U-test. (f) Bacterial counts in blood and multiple organs were similar between vehicle-, OCH- and KRN7000-treated mice with sepsis (n = 7–9 per group). Data are representative of at least three independent experiments.

Fig. 5

Cytokine levels at 24 h in the sera and spleens of septic B6 mice. (a) Sera and spleen homogenates from vehicle-, OCH- and KRN7000-treated B6 mice with intra-abdominal sepsis (IAS) were analysed for 32 inflammatory cytokines by multiplex array, and displayed as a heat map (n = 4 mice per group). Concentrations of invariant natural killer T (iNK T) cell-specific cytokines are shown from sera (b) and spleen homogenates (c) of septic mice treated with vehicle, OCH or KRN7000 (n = 4–8 per group). Concentrations of cytokines are shown in pg/ml. *P < 0·05; **P < 0·01; ***P < 0·001 by one-way analysis of variance with post-hoc Tukey's multiple comparison test. Data are representative of at least three independent experiments.

Fig. 6

Histopathology of septic B6 mice treated with glycolipid agonists of invariant natural killer T (iNK T) cells. (a) Treatment with OCH significantly reduced apoptosis within the spleen compared to vehicle- and KRN7000-treated mice with intra-abdominal sepsis (IAS), both by haematoxylin and eosin (H&E) staining as well as terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) staining. Lymphocyte migration to the omentum is also ameliorated in OCH-treated mice compared to vehicle- and KRN7000-treated mice. There were no histopathological differences in the liver. Images are representative of four animals per treatment group (size bar, 50 μm). (b) Histopathological scoring of the degree of apoptosis observed within the spleens of sham and septic B6 mice treated with vehicle, KRN7000 or OCH (n = 4 animals per treatment group). Apoptosis was defined histologically by the presence of cell clusters with nuclear shrinkage (karyorrhexis), dark eosinophilic cytoplasm, intact plasma membrane and relative paucity of surrounding inflammatory cells within the splenic follicles on H&E staining. Scores assigned to each animal by a blinded independent pathologist were as follows: 0 for complete absence of apoptosis; 1 for mild presence of apoptosis (0–15% per follicle); 2 for moderate apoptosis (16–30% per follicle); and 3 for severe apoptosis (31–45% per follicle). ***P < 0·0001 by two-tailed Mann–Whitney U-test.

Fig. 7

Analysis of apoptotic cell populations in the spleens of septic B6 mice. Splenocytes from sham and septic B6 mice treated with OCH, KRN7000 or vehicle were stained for T, B and natural killer (NK) cells and macrophages, and stained further for annexin V (a marker for early apoptosis) and 7-aminoactinomycin D (7-AAD) viability dye. Early and late apoptotic cells (annexin V⁺ 7AAD^– and annexin V⁺ 7AAD⁺ cells, respectively) were quantified and compared between treatments. OCH treatment significantly reduced apoptosis among T and B cells, as well as macrophages, but not NK cells (n = 3–6 mice per group). *P < 0·05; **P < 0·01; ***P < 0·001 by one-way analysis of variance with Tukey's post-hoc multiple comparison test. Data are representative of three independent experiments.