Engulfment of activated apoptotic cells abolishes TGF-β-mediated immunoregulation via the induction of IL-6

Abstract

Phagocytosis of apoptotic cells (ACs) is usually a potent immunoregulatory signal but can also promote inflammation. In this article, we show that administration of apoptotic dendritic cells (DCs) inhibited inflammation in vivo through increasing production of TGF-β from intrinsic DCs and B cells. However, ACs derived from LPS-activated DCs failed to restrain inflammation because of a short-lived but marked IL-6 response, which abolished the increase in TGF-β. Inhibition of IL-6 restored the protective anti-inflammatory properties of aACs and the TGF-β response. DCs isolated from mice that had received resting but not activated ACs could transfer the suppression of inflammation to recipient mice. These transferred DCs stimulated B cell TGF-β production and relied on an intact B cell compartment to limit inflammation. These results highlight how the activation state of AC governs their ability to control inflammation through reciprocal regulation of IL-6 and TGF-β.

FIGURE 1.

aACs induce production of the proinflammatory cytokines IL-6 and TNF-α in vitro. (A) Peritoneal macrophages were cultured for 6 h alone (No AC) or with apoptotic DCs (AC) or LPS-activated apoptotic DCs (aAC) and then stained for IL-10, TNF-α (red), and DAPI (blue) and imaged using a Leica TSC SPE confocal microscope (original magnification ×40). Arrows show cytokine production. Bar charts show combined data of the mean ± SEM percentage of cytokine-producing cells per field from six independent experiments. (B) Splenic DCs, B cells, and macrophages were cultured for 24 h alone (Nil) or with ACs or aACs. Supernatants were collected, and IL-10, TNF-α, and IL-6 levels were determined. Graphs show mean ± SEM of pooled data from three independent experiments. (C) Bone marrow–derived macrophages were left untreated (M) or pretreated with a TLR4 inhibitor (i). Macrophages were subsequently cultured with 0.01 μg/ml LPS, ACs, or aACs for 72 h, and IL-6 concentration was determined by ELISA. Graph shows mean ± SEM from six independent experiments.

FIGURE 2.

aACs induce production of the proinflammatory cytokines IL-6 and TNF-α in vivo. (A) C57BL/6 mice were left untreated or injected with ACs or aACs and spleens snap-frozen 1, 2, or 4 h later, and the fold increase in IL-10, TGF-β, TNF-α, and IL-6 mRNA production was determined in mice treated with ACs or aACs compared with untreated mice. All data were normalized to HPRT. Graphs show mean ± SEM of pooled data from three independent experiments. (B) PKH-26⁺ ACs or aACs were injected into mice, and splenic cells responsible for the engulfment of ACs and aACs were determined by gating on the PKH-26⁺ cells and analyzed for coexpression of F4/80 to identify macrophages, CD11c to identify DCs, and CD19 to identify B cells. FACS plots show representative data, and graphs show the mean ± SEM percentage of ACs and aACs engulfed by B cells, DCs, or macrophages. Data are pooled from four independent experiments. (C) Cells were further analyzed using ImageStream to determine the percentage of PKH-26⁺ cells engulfed in the spleen. Graph shows data from two independent experiments from six individual mice. (D) Images showing engulfment by DCs (top panel) and macrophages (bottom panel) are representative of the data (original magnification ×40). The markers/stains used in (D) are indicated above the lanes in the figure. Details of the fluorescent labels attached to CD11c and F4/80 are given in the Materials and Methods (CD11c-FITC and F4/80-APC).

FIGURE 3.

Activation of ACs abolishes the ability of ACs to induce TGF-β. (A) Arthritis was monitored in mice left untreated (No AC) or injected with AC or aAC on the day of immunization and for a further 2 consecutive days before intra-articular injection of mBSA to induce inflammation in the knee. Lymph nodes and spleens were harvested on day 7 after knee injection. LN cells were analyzed for IL-17 production by (B) ELISA after no stimulation (Nil) or stimulation with anti-CD3 mAb (aCD3) and (C) flow cytometry. IFN-γ production in the LNs was also determined by flow cytometry after gating on lymphocytes (C). IL-10 production by the spleen was assessed by (B) ELISA and (D) flow cytometry, after gating on lymphocytes. TGF-β production by splenic B cells (E) and DCs (F) was analyzed by flow cytometry. Live lymphocytes were gated; then CD11c⁺ or CD19⁺ populations were gated and assessed for TGF-β production. n = 12 pooled from 4 independent experiments. (G) Mice were left untreated (No AC) or injected with AC or aAC, then immunized and spleens harvested 48 h later. The production of TGF-β by DCs and B cells was determined by flow cytometry, gated as mentioned earlier. Histograms show pooled data from 4 independent experiments; n = 12. FACS plots are representative data.

FIGURE 4.

Activation of ACs abolishes their tolerogenic properties through an IL-6–dependent inhibition of TGF-β. A total of 20 × 10⁶ ACs or aACs were given i.v. to C57BL/6 mice; in some groups, aACs were given in combination with 100 μg anti–IL-6 Ab (aAC + aIL-6). Control groups were i.p. injected with 100 μg anti–IL-6 (aIL-6) or isotype control (Ig control) Abs. All groups were immunized with mBSA and CFA, and aAC, aAC + aIL-6, and AC groups received 20 × 10⁶ cells i.v. for 2 further consecutive days (A). Six days after knee injection, LNs were analyzed for production of IL-17 by ELISA (B), and splenic B cells (C) and DCs (D) were analyzed for production of TGF-β by flow cytometry. Live lymphocytes were gated; then CD11c⁺ or CD19⁺ populations were gated and assessed for TGF-β production. Data are pooled from two independent experiments; n = 6–8. Values in plots show the percentage of cells making the cytokine.

FIGURE 5.

TGF-β–producing splenic DCs mediate the protective effects of ACs. CD11c⁺ DCs were isolated from the spleens of untreated (No AC) or AC- (AC DC) or aAC (aAC DC)-treated mice after 48 h; 1.5 × 10⁶ cells were transferred i.v. into naive WT mice and AIA induced as illustrated schematically in (A). (B) Clinical scores of untreated mice (UT) or mice adoptively transferred with Nil DCs, AC DCs, or aAC DCs over 7 d after arthritis induction. Graph shows pooled data from four independent experiments; n = 12. (C) AC DCs were transferred on the day of immunization in combination with TGF- β blockade using 400 μg TGF-βRI kinase inhibitor VI (DC + anti-TGF), or control mice received TGF-β blockade alone (anti-TGF). Arthritis was induced and clinical scores were determined for 3 d. Data are pooled from two independent experiments; n = 8. (D) The percentages of B cells making TGF-β from untreated mice (UT) or mice adoptively transferred with Nil DCs, AC DCs, or aAC DCs were determined 7 d after arthritis induction. FACS plots show representative data; histograms show pooled data from three independent experiments. Cells were gated for live lymphocytes, then for CD19⁺ cells. CD19⁺ cells were assessed for TGF-β production. (E) AC DCs were transferred into wild type (WT) or B cell–deficient mice (μMT) at the time of immunization, and clinical score was monitored for 6 d after knee injection. Data are pooled from two independent experiments; n = 10.