Heme Induces NLRP3 Inflammasome Formation in Primary Human Macrophages and May Propagate Hemolytic Inflammatory Processes By Inducing S100A8 Expression

Intravascular hemolysis results in the release of damaging hemoglobin and free heme into the circulation. A role for heme as a danger associated molecular pattern (DAMP), with a function in sterile inflammatory responses, is becoming increasingly recognized. Whilst heme has known effects on leukocytes, activating their migration, adhesion molecule expression and cytokine expression, more recent data demonstrate that this molecule can induce NLRP3 inflammasome formation in murine bone marrow macrophages, with consequent interleukin (IL)-1β processing and neutrophil recruitment (Dutra et al., Proc. Natl Acad Sci. 111: E4110, 2014). We aimed to investigate whether heme can also induce inflammasome activation in primary human macrophages (hMACs) and to further characterize the pathways by which heme-induced inflammatory responses may be amplified under sterile conditions. CD14⁺ cells were separated from human peripheral blood (using anti-CD14 magnetic beads) and differentiated into hMACs under M-CSF media supplementation and in the presence of 10% fetal bovine serum. In vitro results are expressed as means ± SEM for triplicate cultures and are representative of three independent experiments. Priming of hMACs with lipopolysaccharide (LPS; 100 ng/mL; 3h) alone induced low level secretion of IL-1β (14.11±9.2 pg/10⁶ cells, as measured by ELISA), while heme (50 µM), in the absence of pre-stimulation with LPS, was unable to induce significant IL-1β secretion within 3h (2.46±1.4 pg/10⁶ cells). In contrast, co-incubation of hMACs with both LPS and heme for 3h significantly enhanced hMAC IL-1β release (490.3±36.3 pg/10⁶ cells; P<0.05 compared to LPS alone). The inflamassome pathway inhibitors, MCC950 (5 µM; a specific inhibitor of NLRP3) and YVAD (40 µM; a caspase-1 inhibitor) significantly inhibited IL-1β secretion in LPS-primed hMACs stimulated with heme (reduced to 35.12±3.9; 184±30.4 pg/10⁶ cells, respectively; 3h; P<0.05 compared to LPS/heme). Co-incubation of the LPS-primed cells with varying concentrations of heme, under the conditions employed, did not induce TNF-α production (data not shown), consistent with the hypothesis that IL-1β processing in heme-induced LPS-primed hMAC was mediated by inflammasome formation. Interestingly, qPCR showed that incubation of hMACs (1x10⁶ cells/mL) with heme (50 µM) for 24h stimulated an approximately 10-fold increase (P<0.01) in the expression of the gene encoding, S100A8, another DAMP known to act as a TLR-4 agonist and to contribute to ischemia/reperfusion injury. Priming of hMACs with 1 µg/ml recombinant S100A8 for 3h and subsequent activation with heme (50 or 100 µM, 14h) significantly augmented the release of IL-1β (42.1±0.4 and 89.4±32.4 pg/10⁶ cells for 50 and 100 µM heme, respectively; P<0.05), compared with S100A8 alone (20.6±3.5 pg/10⁶ cells), without any modulation in TNF-α secretion (P>0.05). Using a model of acute intravascular hemolysis, we confirmed an association between heme release and S100A8 secretion, in vivo. Plasma heme levels increased significantly from 26.3±5 µM (i.v. saline control; N=4) to 87±18 µM in C57BL/6 mice at 1h after receiving i.v. water (150 µl; N=4, P=0.04). A concomitant increase in plasma S100A8 levels was also observed within 1h of the hemolytic stimulus (986±102 pg/mL, compared to 694.2±102 pg/ml in control mice; N=4, P=0.05), which was maintained for 3h (P<0.05). Thus, we present data to demonstrate that heme can induce IL-1β processing in LPS-primed human macrophages under in vitro conditions, probably via formation of the NLRP3/caspase-1 inflammasome machinery. In the absence of LPS, heme-stimulated hMACs can express the S100A8 DAMP; furthermore, a hemolytic stimulus induced mouse S100A8 production in vivo. As such, S100A8 may amplify heme-dependent inflammasome formation in an autocrine fashion, even under sterile conditions. Data provide new insights into the mechanisms by which heme may induce and potentiate inflammatory responses in hemolytic diseases, such as sickle cell disease, and suggest S100A8, together with heme, as potential therapeutic targets for reducing inflammation in these diseases.