Resolution metabolomes activated by hypoxic environment

See allHide authors and affiliations

Science Advances  23 Oct 2019:
Vol. 5, no. 10, eaax4895
DOI: 10.1126/sciadv.aax4895


Targeting hypoxia-sensitive pathways in immune cells is of interest in treating diseases. Here, we demonstrate that physiologic hypoxia (1% O2), as encountered in bone marrow and spleen, accelerates human M2 macrophage efferocytosis of apoptotic-neutrophils and senescent erythrocytes via lipolysis-dependent biosynthesis of specialized pro-resolving mediators (SPMs), i.e. resolvins, protectins, maresins and lipoxin. SPM-production was enhanced via hypoxia in M2 macrophages interacting with neutrophils and erythrocytes enabling structural elucidation of a novel eicosapentaenoic acid (EPA)–derived resolvin, resolvin E4 (RvE4) that stimulates efferocytosis of senescent erythrocytes and more potently than aspirin in mouse hemorrhagic exudates. In hypoxia, glycolysis inhibition enhanced neutrophil RvE4-SPM biosynthesis. Human macrophage-erythrocyte co-incubations in physiologic hypoxia produced RvE4-SPM from erythrocyte stores of omega-3 fatty acids. These results indicate that hypoxic environments, including bone marrow and spleen as well as sites of inflammation, activate SPM-biosynthetic circuits that in turn stimulate resolution and clearance of senescent erythrocytes and apoptotic neutrophils.


Hypoxia accompanies inflammation in peripheral tissues and drives cell maturation in immunologic niches including bone marrow and lymphoid tissues (1, 2). Pathologic hypoxic tissues as in ischemic tissues and sites of chronic inflammation experience severe and disrupted oxygen gradients (1) that can arise from loss of blood supply and increased oxygen demand from infiltrating neutrophils to sustain their oxidative burst required for microbial killing (1, 2). Proinflammatory lipid mediators (LMs) including leukotrienes and prostaglandins (PGs) are produced in pathologic hypoxic tissues (3, 4), yet the connection between physiologic hypoxic local environments and the resolution of inflammation via proresolving mediators is currently unknown. Physiologic hypoxia, as encountered in local niches, is defined as 1 to 5% molecular oxygen (5, 6) and has been observed in this range and as low as 0.5% O2 in spleen in vivo (6). During the resolution phase of inflammation, specialized proresolving mediators (SPMs) are temporally biosynthesized to limit further neutrophil infiltration and stimulate macrophage efferocytosis of apoptotic neutrophils and cellular debris. SPMs include lipoxins (LX) from arachidonic acid (AA) and resolvins (Rv), protectins (PD), and maresins (MaR) from the n-3 essential fatty acids, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) (7). Each of the potent bioactive mediators in the SPM superfamily stimulates macrophage efferocytosis (7, 8).

Tissues that experience physiologic hypoxic niches such as in bone marrow and lymphoid tissues display consistent and sustained oxygen gradients (1) and are enriched in SPM (9, 10) through undefined mechanisms. Human macrophages and apoptotic neutrophils are major cellular sources of SPMs in the innate immune system (11, 12). Thus, physiologic mechanisms that control SPM production and signaling are of interest in understanding the events that activate the termination of acute inflammation. We therefore questioned whether hypoxia-activated SPM production is linked to the clearance of apoptotic and senescent cells in physiologic niches, including bone marrow and spleen, as well as pathophysiologic niches encountered during inflammation. Here, we report that physiologic hypoxia activates lipolysis-dependent SPM biosynthesis along with production of a novel bioactive resolvin, termed resolvin E4 (RvE4), from endogenous sources of EPA in human macrophages as well as in macrophage-erythrocyte and macrophage-neutrophil coincubations. RvE4 demonstrated greater potency for stimulating efferocytosis than DHA-derived resolvins, namely, RvD5 and RvD6, as well as enhanced resolution of hemorrhagic exudates in vivo, thus establishing a new EPA-derived LM biosynthetic circuit as well as the novel concept that physiologic hypoxia activates production of SPM and thus their proresolving actions.


Physiologic hypoxia activates SPM production in human leukocytes

To investigate the relationship between physiologic microenvironment hypoxia (namely, 1% molecular oxygen) and resolution phase mediators of inflammation, we interrogated liquid chromatography–tandem mass spectrometry (LC-MS/MS) LM profiles obtained from human leukocytes involved in first-line innate immune responses (neutrophils) as well as the cells that clear neutrophils from inflammatory sites (M2 macrophages) following maintenance in either normoxia (18.6% O2) or physiologic hypoxia (1% O2) for 24 hours. Physiologic hypoxia (5) enhanced total SPM production as well as expression of hypoxia-inducible factor–1α (HIF-1α) in M2 macrophages in a statistically significant manner (Fig. 1, A and B). We assessed M2 macrophage phenotype and confirmed increased surface expression of CD163 and CD206 relative to macrophages incubated without interleukin-4 (IL-4) before maintenance in physiologic hypoxia (fig. S1A). Total PGs were not statistically significantly increased with physiologic hypoxia from M2 cells (1.7-fold; P = 0.36). This consequently increased the SPM:PG ratio ~20-fold; P < 0.01 [also known as LM class switching (7)] in M2 macrophages. LM-SPM pathway analysis (fig. S1B) indicated that specific SPM from each pathway family increased in human M2 macrophages with hypoxia >10-fold, including D-series resolvin, protectin, maresin, lipoxin, and E-series resolvin pathways.

Fig. 1 Physiologic hypoxia activates SPM production in human M2 macrophages and neutrophils.

SPM production, HIF expression, and SPM pathway lipoxygenase expression in M2 macrophages and neutrophils incubated in normoxia (18.6% O2) or physiologic hypoxia (1% O2) at 37°C, 24 hours. (A) Increased amounts of total SPM in human M2 macrophages with physiologic hypoxia versus normoxia determined by LC-MS/MS. (B) Increased HIF-1α expression in M2 macrophages with physiologic hypoxia versus normoxia determined by flow cytometry. (C) Increased expression of 15-LOX and 5-LOX in M2 macrophages with physiologic hypoxia versus normoxia measured by flow cytometry. (D) Hypoxia stimulates nuclear colocalization of 15-LOX and 5-LOX compared to normoxia as assessed by confocal microscopy. DAPI, 4′,6-diamidino-2-phenylindole. (E) Increased amounts of total SPM in human neutrophils with physiologic hypoxia versus normoxia. (F) Increased HIF-1α expression in neutrophils with physiologic hypoxia versus normoxia. (G) Coincubation of human M2 macrophages (M2) with human neutrophils (PMN) enhanced SPM production versus summation of SPM from M2 and PMN incubated separately during physiologic hypoxia, i.e., costimulation index = [M2 + PMN] versus [M2] + [PMN]; coincubations 1:10 M2:PMN. (H) Cytoscape pathway analysis of LC-MS/MS results demonstrating changes in LMs in M2 macrophage–neutrophil coincubations with hypoxia versus normoxia represented by heat map. Circle size represents picogram (pg) quantities with physiologic hypoxia, dashed boxes indicate LM families, and asterisk denotes EPA-derived product with unknown functions (see text for details). Results are the means ± SEM of three healthy human neutrophil donors and three healthy human peripheral blood mononuclear cell donors for M2 macrophages; *P < 0.05 for physiologic hypoxia versus normoxia using paired ratio t test. MFI, mean fluorescence intensity.

We next questioned whether physiologic hypoxia activates SPM production via modulation of pathway-specific enzyme expression and/or their translocation in M2 macrophages compared to cells maintained in normoxia. Expression of 15-lipoxygenase (15-LOX) and 5-LOX increased in this hypoxic environment (Fig. 1C), and each also translocated from cytoplasmic regions with normoxia to nuclear regions with hypoxia (Fig. 1D).

In human neutrophils, physiologic hypoxia (1% O2) also statistically significantly increased total SPM production (Fig. 1E) as well as HIF-1α expression (Fig. 1F). Specific SPM from each pathway family were increased >5-fold (fig. S1C). Physiologic hypoxia decreased total PGs in neutrophils (fig. S1C) without statistical significance (1.8-fold; P = 0.07) and increased the SPM:PG ratio ~7-fold (P < 0.05).

Macrophage-neutrophil interactions are essential for resolving inflammation via efferocytosis (7, 8). We therefore assessed the role of hypoxia on interaction of human M2 macrophages and neutrophils to costimulate production of SPMs. Coincubations of M2 macrophages and neutrophils maintained in physiologic hypoxia gave increased total SPM production ~6-fold above the summation of total SPM produced in each cell type alone (i.e., costimulation), as shown in Fig. 1G. In these coincubations, physiologic hypoxia increased production of specific mediators from each SPM pathway family >10-fold (Fig. 1H). Together, these results demonstrate that physiologic hypoxia drives leukocyte SPM production by increasing 5-LOX and 15-LOX expression as well as their colocalization that enables LM class switching in macrophages and neutrophils.

Metabolic targeting: Inhibition of glycolysis enhances human neutrophil SPM production

M2 macrophages function with high rates of fatty acid oxidation and oxidative phosphorylation and a low rate of glycolysis that is associated with their role in wound healing and resolution of inflammation and infections (13, 14). Neutrophils, on the other hand, are almost entirely glycolytic (15). Along these lines, inhibition of glycolysis with 2-deoxyglucose (2-DG) mitigates neutrophilic inflammation and lung injury initiated by nebulized lipopolysaccharide (LPS) (16). Given the opposing metabolic requirements and distinct roles of M2 macrophages and neutrophils in inflammation resolution, we assessed the relationships between glycolysis, SPM production, and hypoxia (1%) in these human phagocytes. Inhibiting glycolysis in neutrophils with 2-DG in a hypoxic environment statistically significantly increased total SPM production ~5-fold (fig. S2A). In M2 macrophages, inhibition of glycolysis did not alter SPM production. Pathway analysis indicated that RvD1, RvD4, LXA4, and 18S-RvE1 were each increased >10-fold with 2-DG in human neutrophils (fig. S2B). To determine whether hypoxic neutrophil SPM production is enhanced via glycolysis without blocking exokinase-dependent glycosylation of newly synthesized proteins, we carried out experiments with neutrophils in the presence of hypoxia and galactose. Inhibition of glycolysis with galactose statistically significantly increased DHA-derived SPM production in neutrophils in hypoxia (fig. S2C). These results demonstrate that leukocyte SPM production is negatively regulated by glycolysis in human neutrophils in a hypoxic environment.

Structural elucidation of a novel resolvin in physiologic hypoxic human leukocytes

In hypoxic macrophage-neutrophil coincubations, we identified a new product of unknown structure that increased along with E-series resolvins, namely, 18S-RvE1, 18S-RvE2, and 18S-RvE3 (Fig. 2A). The endogenous material gave a 12.5-min LC retention time consistent with chromatographic behavior of a dihydroxy polyunsaturated fatty acid on a reversed-phase LC column, an ultraviolet (UV) absorption spectrum with λmax = 244 nm, as well as negative-mode MS/MS fragmentation ions including a mass/charge ratio (m/z) 333 (M-H; parent ion) shown in Fig. 2A, and fragmentation ions including: m/z 315 (M—H—H2O), 271 (M—H—H2O—CO2), and 253 (M—H—2H2O—CO2) that indicated an EPA backbone containing two alcohol groups; m/z 115, 217, and 199 (217-H2O) fragment ions that indicated an alcohol group at carbon 5; as well as m/z 235, 191 (235-CO2), and 173 (235-H2O—CO2) fragment ions that indicated an alcohol group at carbon 15 (Fig. 2A). Structural assignments of the endogenous material were confirmed by matching LC retention time, UV λmax, and MS/MS fragmentation with 5S,15S-dihydroxyeicosapentaenoic acid and the positions of the double bond prepared via “one-pot” biogenic synthesis with substrate EPA and 15-LOX (see Materials and Methods for conditions) followed by LC purification (Fig. 2A). This new structure was termed RvE4 since it is produced from EPA, carries bioactivity (vide infra), and is increased along with E-series resolvins in physiologic hypoxia by phagocytes (Figs. 1D and 2A). RvE4 was biosynthesized via two consecutive lipoxygenations of EPA and contains two conjugated dienes, as evidenced by its 244-nm UV λmax, and alcohols at carbons 5 and 15 with deduced structure: 5S,15S-dihydroxyeicosa-6E,8Z,11Z,13E,17Z-pentaenoic acid (Fig. 2B).

Fig. 2 Novel RvE4 biosynthesis in hypoxic human leukocytes: Structural elucidation.

(A) Enhancement of E-series resolvins along with uncharacterized product in human M2 macrophage–neutrophil (M2-PMN) coincubations with hypoxia; structural characterization of biogenic RvE4 via LC-MS/MS and UV spectroscopy and MS/MS matching with endogenous material from M2-PMN coincubation during hypoxia. (B) RvE4 biosynthetic pathway in relation to the SPM superfamily of resolution mediators; scheme shows the production of SPM initiated by PLA2 hydrolysis of phospholipids and lipase hydrolysis of cholesteryl esters and triglycerides for mobilization of EPA, DHA, and AA in physiologic hypoxia (as shown in fig. S3, A and B). (C) Increased production of RvE4 in human M2 (left) and PMN (middle) with physiologic hypoxia versus normoxia. Enhancement of RvE4 amounts with coincubations of M2 and PMN versus summation of RvE4 amounts from M2 and PMN incubated separately during hypoxia, i.e., costimulation = [M2 + PMN] versus [M2] + [PMN]. Results are the means ± SEM of three healthy human neutrophil donors and three healthy human peripheral blood mononuclear cell donors for M2 macrophages; *P < 0.05 and **P < 0.01 for physiologic hypoxia versus normoxia using paired ratio t test.

In M2 macrophages, lipase activity was required for mobilization of polyunsaturated fatty acid precursors AA, EPA, and DHA for physiologic hypoxia-activated biosynthesis of SPM, including the new RvE4, as determined by addition of an inhibitor of hormone-sensitive lipase and adipose triglyceride lipase (17) (fig. S3A). In neutrophils, phospholipase A2 hydrolysis of phospholipids was required for liberation of precursors, namely, EPA and AA, to biosynthesize lipoxin and E-series resolvins, including RvE4, whereas lipase activity was required for precursor DHA mobilization to biosynthesize maresins, protectins, and D-series resolvins (fig. S3B). Physiologic hypoxia increased production of RvE4 in both M2 macrophages (P < 0.05) and neutrophils (P < 0.05), and its production was further increased ~8-fold by costimulation of M2 macrophages together with neutrophils (Fig. 2C). These results identify a novel resolvin, namely RvE4, derived from EPA that is produced in human macrophages and neutrophils activated by hypoxia.

RvE4 and SPM production via macrophage-erythrocyte interactions in hypoxia

We next questioned whether hypoxia (1%) enhances SPM biosynthesis via macrophage-erythrocyte interactions given that neutrophil-macrophage interactions costimulated SPM production in this setting (Fig. 1, B and D). Human M2 macrophages were incubated with erythrocytes that contained a portion of membrane phospholipids incorporated with esterified penta-deuterium (d5)–labeled DHA and EPA (cells referred to as RBCd5 throughout; Fig. 3A) to track transcellular SPM biosynthesis from red blood cell (RBC) to macrophages.

Fig. 3 Human erythrocyte membrane phospholipids are a source of RvE4 and SPM production in physiologic hypoxia-activated macrophages.

(A) RBC-derived SPM production from coincubations of RBC containing d5-labeled DHA and EPA (d5-DHA and d5-EPA) esterified in phospholipid membranes (RBCd5) with M2 macrophages (100:1 RBC:M2) in physiologic hypoxia versus normoxia; Cytoscape pathway analysis of LC-MS/MS results demonstrating changes in LMs in M2-RBCd5 coincubations with physiologic hypoxia versus normoxia represented by a heat map; circle size represents picogram (pg) quantities with physiologic hypoxia. (B) LC-MS/MS chromatograms of RBC d5-DHA– and d5-EPA–derived SPMs and pathway markers. (C) MS/MS spectra of RBC d5-DHA–derived RvD5 (d5-RvD5) and RBC d5-EPA–derived RvE4 (d5-RvE4) with d5-containing ion fragments indicated in blue. (D) Total d5-DHA–derived SPMs represent the summation of d5-RvD5, d5-RvD6, d5-MaR2, and d5-10S,17S-diHDHA; total d5-EPA–derived SPMs represent the sum of d5-RvE3 and d5-RvE4. Results are the means ± SEM of three healthy human RBC donors and three healthy human peripheral blood mononuclear cell donors for M2 macrophages. *P < 0.05 for RBCd5 + M2 with physiologic hypoxia versus normoxia using paired ratio t test.

Deuterium-labeled SPMs derived from d5-EPA and d5-DHA were identified in RBCd5-macrophage coincubations and were increased with physiologic hypoxia versus normoxia (Fig. 3A). Identification of these deuterium-labeled SPMs, including d5-RvD5 and d5-RvE4 (Fig. 3B), was confirmed by matching multiple reaction monitoring (MRM) chromatogram retention time and MS/MS matching of diagnostic ions, including +5 m/z shifts (increase of 5 u) in parent mass and specific fragment ions that also contained the penta-deuterated omega carbon of the molecule (Fig. 3C). Deuterium-labeled SPMs were statistically significantly increased with RBCd5-macrophage coincubations with physiologic hypoxia versus RBCd5 alone with hypoxia or RBCd5-macrophages or RBCd5 alone with normoxia (P < 0.05; Fig. 3D). These results demonstrate that physiologic hypoxia promotes delivery of n-3 fatty acids from erythrocytes to macrophages for SPM transcellular biosynthesis.

Hypoxic environment accelerates efferocytosis of senescent RBC and apoptotic neutrophils via endogenous RvE4-SPM

Given that macrophage erythrophagocytosis promotes resolution of intracerebral hemorrhage and neurological recovery (18), we carried out experiments to determine the role of endogenous RvE4 and SPM production on macrophage efferocytosis of apoptotic polymorphonuclear neutrophils (aPMN) and senescent RBC (sRBC). RBC senescence was activated by incubation of cells at 4°C for 24 hours before efferocytosis experiments, and senescence was measured using flow cytometry analysis of CD47 surface expression, which was statistically significantly reduced versus freshly obtained RBCs (fig. S4). Hypoxia statistically significantly increased M2 macrophage efferocytosis of aPMN (Fig. 4A) and sRBC (Fig. 4B). Hypoxia enhanced macrophage efferocytosis as high as 106% with aPMN and 85% with sRBC from individual healthy human donors.

Fig. 4 Physiologic hypoxia enhances macrophage efferocytosis of sRBC and apoptotic neutrophils via endogenous RvE4-SPM production.

Human M2 macrophages (M2) were preincubated in physiologic hypoxia or normoxia for 24 hours, 37°C before addition of carboxyfluorescein succinimidyl ester (CFSE)–labeled human sRBC (1:20 M2:sRBC) or CFSE-labeled human aPMN (1:3 M2:aPMN) for 60 min, 37°C. (A) Increased M2 macrophage efferocytosis of apoptotic neutrophils with physiologic hypoxia versus normoxia assessed by flow cytometry (left) and confocal microscopy (right); arrows indicate apoptotic neutrophils; n = 5 healthy human donors. (B) Increased M2 macrophage efferocytosis of sRBC with physiologic hypoxia versus normoxia; *P < 0.05 for physiologic hypoxia versus normoxia; n = 6 healthy human donors. (C) Inhibition of M2 macrophage efferocytosis of apoptotic neutrophils (left) and apoptotic RBC (right) after incubation in physiologic hypoxia and addition of a lipoxygenase inhibitor (baicalein; 20 μM) and rescue of efferocytosis with addition of SPMs produced by M2 macrophages in physiologic hypoxia as shown in fig. S1B (hypoxia SPMs = RvD2, RvD5, RvD6, MaR1, PD1, and RvE4; each together at 10 nM); *P < 0.05 for +baicalein versus physiologic hypoxia; n = 6 healthy human donors. (D) Enhancement of M2 macrophage efferocytosis of aPMN (left) and sRBC (right) with 15-min preincubation of RvD5 (10 nM), RvD6 (10 nM), RvE4 (10 nM), or the hypoxia SPM panel [as in (C); each at 10 nM] assessed by flow cytometry. Results are the means ± SEM of six healthy human donors; *P < 0.05 and **P < 0.01 for individual SPM and hypoxia SPM panel versus vehicle.

To quantify the specific involvement of endogenous SPM signaling on enhancement of M2 macrophage efferocytosis in physiologic hypoxia, SPM production was inhibited with a lipoxygenase inhibitor before addition of aPMN and sRBC. SPM production in physiologic hypoxia was statistically significantly reduced by >99% in the presence of a lipoxygenase inhibitor, which did not statistically significantly reduce total PG production (fig. S5). Inhibition of endogenous macrophage SPM production statistically significantly reduced efferocytosis of aPMN and sRBC by ~50%, which was restored by addition of the hypoxia SPM cluster (RvE4, RvD2, RvD5, RvD6, MaR1, and PD1; each together at 10 nM) in the presence of the lipoxygenase inhibitor, as shown in Fig. 4C.

We next assessed structure-function relationships between the novel RvE4, RvD5, and RvD6, all of which are components of the hypoxia-activated macrophage SPM cluster (Fig. 1D and fig. S1B). M2 macrophage efferocytosis of aPMN was statistically significantly increased with the addition of RvE4 (10 nM) or a cluster of SPM produced by physiologic hypoxic M2 (RvE4, RvD2, RvD5, RvD6, MaR1, and PD1, each together at 10 nM) but not with RvD5 or RvD6 (10 nM) (Fig. 4D). Efferocytosis of sRBC was statistically significantly increased with addition of RvE4, RvD5, and the hypoxia SPM cluster but not with RvD6 (Fig. 4D). These results demonstrate that hypoxia enhances human macrophage efferocytosis of apoptotic neutrophils and senescent erythrocytes that is RvE4-SPM dependent.

RvE4 stimulates resolution of hemorrhagic exudates

The acute inflammatory response shapes a hypoxic niche that develops, in part, via increased oxygen consumption by infiltrating neutrophils and monocytes (1, 19). LM-SPM profiles were therefore interrogated with inflammatory exudates using an established murine model of hemorrhagic peritonitis that is defined by increased abundance of neutrophils, RBC, and microthrombi (20, 21). Aspirin, which carries both proresolving and anti-inflammatory actions, was also given to assess whether SPM production can be enhanced in pathologic hypoxia via acetylation of cyclooxygenase-2 enzyme (COX-2). Mice were given zymosan (1 mg) and thrombin (5 U) with or without aspirin [acetylsalicylic acid (ASA); 100 mg/kg] via intraperitoneal injection. Hemorrhagic exudates were collected 12 hours after zymosan/thrombin, and LMs were extracted and taken to LC-MS/MS for LM-SPM metabololipidomics. Principal components analysis (PCA) indicated distinct clustering of hemorrhagic exudates and hemorrhagic exudates with aspirin (fig. S6A). Hemorrhagic exudates were associated with PGs and leukotrienes, and hemorrhagic exudates with aspirin were associated with SPMs, including aspirin-triggered (AT)–LXA4, AT-RvD1, RvE1, RvE2, RvE3, RvD1, RvD4, RvD5, PD1, LXA4, and RvE4 (fig. S6A). Aspirin statistically significantly increased total amounts of aspirin-triggered SPMs, namely, AT-LXA4, AT-RvD1, RvE1, RvE2, and RvE3, and reduced amounts of PGs (PGE2, PGD2, PGF2a, and TxB2), which statistically significantly increased the SPM:PG ratio (fig. S6B). RvE4 was identified (with and without aspirin) in these hemorrhagic exudates (fig. S6, A and C).

Here, we interrogated RvE4 proresolving bioactions in vivo in hemorrhagic exudates with or without aspirin. RvE4 and aspirin each statistically significantly reduced peak neutrophil numbers at 12 hours (Tmax) by >40% (P < 0.05) and reduced the time needed to reach half-maximal neutrophil numbers (T50) from Tmax [resolution interval (Ri)] by ~8.5 hours. RvE4 also statistically significantly reduced neutrophil numbers versus aspirin at 24 hours (Fig. 5, A and B). Macrophage numbers were statistically significantly increased with RvE4 versus hemorrhagic exudates with and without aspirin at 24 hours (Fig. 5, A and C). In addition to increasing macrophage numbers, RvE4 statistically significantly increased total numbers of macrophages containing neutrophils and RBC versus hemorrhagic exudates with or without aspirin at 24 hours (Fig. 5, A and D). Together, these results demonstrate that RvE4 stimulates resolution of inflammation by reducing neutrophilic infiltration and enhancing macrophage efferocytosis of neutrophils and erythrocytes in hemorrhagic exudates in vivo.

Fig. 5 RvE4 is a proresolving mediator that stops neutrophilic infiltration and enhances macrophage efferocytosis of cellular debris in hemorrhagic exudates.

Mice were given zymosan A (1 mg) and thrombin (5 U) with or without aspirin (4 mg) or RvE4 (100 ng) by intraperitoneal injection. Peritoneal lavages were collected at 0, 12, and 24 hours and taken to flow cytometry for leukocyte enumeration and to LC-MS/MS for LM-SPM profiling. (A) Representative flow cytometry dot blots indicating neutrophils and monocytes at 24 hours (top) and corresponding reductions in neutrophil numbers with aspirin and RvE4 (right); macrophages at 24 hours (middle) with corresponding increases in macrophage numbers with RvE4 (top right); representative flow cytometry dot blots macrophages containing neutrophils and RBC at 24 hours (bottom) with corresponding total macrophage efferocytosis of neutrophils and RBC at 24 hours (bottom right). (B) Time course of neutrophil infiltration indicating reduced neutrophil numbers with RvE4 versus ASA or hemorrhagic exudates at 24 hours. (C) Time course of macrophage egress followed by infiltration indicating increased macrophage numbers with RvE4 versus ASA or hemorrhagic exudates at 24 hours. (D) Macrophage numbers containing engulfed neutrophils (PMN efferocytosis; left), engulfed RBC (RBC efferocytosis; middle), and both neutrophils and PMN (PMN-RBC efferocytosis; right). Results are the means ± SEM; n = 3 to 6 mice per group; *P < 0.05 and ***P < 0.001 for +RvE4 versus hemorrhagic exudates; #P < 0.05 and ##P < 0.01, ###P < 0.001, and ####P < 0.0001 for +RvE4 versus +Aspirin.


In the present report, we identified a new cluster of SPMs (select resolvins, protectins, lipoxins, and maresins; see Fig. 1H) and a novel E-series resolvin, termed RvE4, that is produced by M2 macrophage–neutrophil interactions and is triggered by hypoxia. These endogenous mediators enhance human M2 macrophage efferocytosis of both apoptotic neutrophils and senescent erythrocytes in local hypoxic environments (Figs. 1, 4, and 5 and figs. S1, S6, and S7). SPMs are abundant and functional in both human and murine tissues that experience physiologic hypoxia, namely, spleen, lymph nodes, bone marrow, intestinal mucosa, retina, and placenta (9, 10, 2225), and therein promote protection from excessive inflammation. In bone marrow, for example, RvD1 and RvD2 are produced and enhance macrophage efferocytosis of apoptotic osteoclasts (9). In spleen, SPM-producing monocytes and macrophages mobilize and then migrate to the left ventricle in response to myocardial infarction to promote resolution and tissue recovery (10). In the retina, RvE1, RvD1, and PD1 promote vessel regrowth to inhibit pathological neovascularization following O2-induced retinopathy (25). All the above mentioned organs and tissues experience physiologic hypoxia (1, 5).

Here, we elucidated the novel RvE4 produced by M2 macrophage–neutrophil interactions in a hypoxic environment and deduced its structure using biogenic synthesis, fragmentation with MS/MS and matching physical properties with endogenous material (Fig. 2A). We used an in vitro biogenic synthesis (Fig. 2, A and B) as well as cell-based inhibition of lipoxygenases (fig. S4) to establish a 15-LOX–dependent double dioxygenation mechanism in the biosynthetic route for RvE4 biosynthesis. Enhancement of RvE4 production in hypoxic M2 macrophage–neutrophil coincubations suggests that transcellular routes involving 15-LOX/5-LOX are also operative (Fig. 2C). The new resolvin RvE4 stimulates resolution in pathologic hypoxia as evidenced in hemorrhagic exudates in mice (Fig. 5). Aspirin, a known inhibitor of cyclooxygenases and PGs that triggers production of epimer SPMs [reviewed in (7)], also enhanced resolution in hemorrhagic exudates by increasing the production of aspirin-triggered resolvins (AT-RvE1, RvE2, RvE3, AT-RvD1, and AT-LXA4) and other SPM in vivo in mice (Fig. 5 and fig. S6, A and B). Of interest, RvE4 was more potent than aspirin on a dose basis (100 ng of RvE4 versus 4 mg of aspirin per mouse) in stimulating macrophage recruitment as well as macrophage efferocytosis of both apoptotic neutrophils and senescent erythrocytes in the resolution phase (Fig. 5B). These results indicate that stimulating SPM production, as in physiologic hypoxia (Figs. 1 to 3 and fig. S2), accelerates cell clearance during the resolution of inflammation. Since aspirin and hypoxia activate SPM production via independent mechanisms, the intersection of these pathways poses a unique environment to explore therapeutic strategies to augment resolution.

Removal of neutrophils from sites of infection and injury protects the host from excessive exposure to extracellular reactive oxygen species and tissue-destructive enzymes that enables wound healing and catabasis (21). Hence, rapid removal of erythrocytes from these sites is also key for protection from hemolysis-derived, iron-containing hemoglobin and its degradation products (26), as well as hemin, which delays resolution as determined by resolution indices (27). In addition, macrophage efferocytosis of erythrocytes promotes resolution of intracerebral hemorrhage for neurological recovery (18). These events can also contribute to the vascular occlusion and pathology in sickle cell disease that were recently shown to be mitigated by aspirin-triggered resolvin D1 in mice (28).

LM SPMs, including RvE4, were shown here to accelerate clearance of neutrophils and erythrocytes as well as shorten the resolution interval. Resolution indices, introduced to quantitate resolution (7), are used to identify protein mediators [e.g., annexin-A1 (29), erythropoietin (30), CXCL-13 (31), and recently DEL-1 (32)], gaseous mediators [e.g., hydrogen sulfide (33) and nitric oxide (7, 34)] as well as pharmacologic agents [e.g., aspirin and statins (reviewed in 7, 34, 35)] that activate resolution by shortening the resolution interval (7). SPMs, by definition, stimulate efferocytosis of apoptotic neutrophils and inhibit neutrophil infiltration at sites of infection and tissue injury (7). The abundance of distinct SPM structures that exert these actions underscores the robust resolution response that is required to counter-regulate the many cytokines and chemokines, which carry overlapping proinflammatory functions (36) as well as endogenous factors that delay resolution, including the neuronal axon guidance protein neogenin [Neo1 (37)], phosphodiesterase 4 (38), and hemin (27). MicroRNAs, including miR-146b and miR-219, are up-regulated in mouse peritonitis and human macrophages in response to RvD1 and therein counter-regulate nuclear factor κB and 5-LOX (39). In addition to these microRNAs that are produced in response to resolvin-mediated signaling, neutrophils can transfer miR-223 to lung epithelial cells that reduces acute lung injury via repression of poly[adenosine 5′-diphosphate (ADP)–ribose] polymerase–1 (40). Hence, understanding the relationships between hypoxic local environments and production of the growing number of proresolution mediators identified (i.e., LMs, protein, and gaseous mediators) (7, 29, 33) is of interest and may be useful for developing novel therapeutics to treat inflammatory diseases.

Hypoxia and the acute inflammatory response are linked to cellular metabolism that controls innate immune cell functions (1, 13, 14). Our results here demonstrate that inhibition of neutrophil glycolysis enhances RvE4-SPM production and with greater magnitude in physiologic hypoxia versus normoxia (fig. S2, A and B) that supports enhanced signaling for the removal of apoptotic neutrophils by macrophages and resolution. In contrast, M2 macrophage SPM production in hypoxic environment was not affected by glycolysis inhibition (fig. S2A). HIF is known to increase glycolytic gene expression (41) that is associated with M1 macrophage polarization (1, 14) as well as exaggerated neutrophil tissue infiltration and neutrophil survival (16). Hence, metabolic targeting of glycolysis is proposed as a strategy to prevent chronic or excessive neutrophilic inflammation (16) that is likely linked to enhancement of neutrophil RvE4-SPM production demonstrated in the present report. It is important to emphasize that HIF-1α stabilization in other cell types promotes anti-inflammatory responses; for example, in alveolar epithelial cells, HIF-1α attenuates acute lung injury (42), and in regulatory T cells, HIF-1α induces FoxP3 to limit tissue damage in colitis (43). Moreover, purinergic signaling directed by hypoxia promotes mucosal barrier restitution in inflammatory bowel disease (44). Another major HIF isoform, HIF-2α, promotes cardioprotection by inducing the epithelial growth factor amphiregulin (45). Hypoxia also plays HIF-independent anti-inflammatory roles in humans by increasing adenosine levels, which, in turn, dampens endotoxin-induced cytokine production (46).

Unexpectedly, human M2 macrophages in physiologic hypoxia used lipolytic enzymes, which act on cholesteryl esters and triglycerides for mobilization of omega-3 fatty acids to produce RvE4-SPM (fig. S3), thus indicating that the proresolving phenotype of these cells is directly linked to their metabolic program (i.e., fatty acid β-oxidation and oxidative phosphorylation) via SPM production. Lipolysis in M2 macrophages is linked to fatty acid oxidation pathways for generation of adenosine 5′-triphosphate (14), as well as the lipoxygenase pathways involved in the biosynthesis of RvE4-SPM production, uncovered here (Fig. 1 and figs. S1B and S3). Of note, mast cells contain triglyceride-rich lipid droplets and also use lipase-dependent hydrolysis of fatty acids to produce eicosanoids including PGD2 and leukotriene C4 (LTC4) via immunoglobulin E–dependent activation (47). Thus, both hypoxia-dependent and independent activations of neutral lipases contribute to LM-SPM production. Further study of tissue and cell-specific regulation as well as HIF-dependent regulation of these mechanisms is therefore of interest. On the other hand, inflammatory stimuli including Toll-like receptor and purinergic receptor agonists mobilize arachidonic acid from phospholipids via cytosolic phospholipase (cPLA2) Group IVA to produce PGs, leukotrienes, and lipoxins (48). In addition, EPA and DHA are mobilized from phospholipids by secretory PLA2 (sPLA2) Group IID in lymph nodes during contact dermatitis (49). Together, these results elucidate a novel neutral lipid-derived SPM biosynthetic pathway activated by hypoxic environment that would be encountered in physiologic niches (1) that regulate efferocytosis in human M2 macrophages and is separate from phospholipid-derived LM biosynthetic pathways activated by inflammatory stimuli. Given that molecular oxygen is required for lipoxygenase-dependent SPM production as well as COX-dependent production of PGs, the hypoxic tone required to promote pathway enzyme activation, translocation, and expression that supports LM production is likely to be highly specific. Along these lines, hypoxia increases PG production, while extreme hypoxia limits biosynthesis in arteries of cattle (50). Establishing specific oxygen concentrations that activate or limit LM production in cells in controlled environment is therefore of interest.

Long-term omega-3 supplementation increases EPA and DHA content in erythrocyte phospholipids that is used as a marker of omega-3 intake, namely, the omega-3 index (51). Increasing the omega-3 index positively correlates with plasma SPM precursors in patients with peripheral arterial disease (52). Erythrocytes are known to enhance the formation of human neutrophil 5-LOX products (53). In Fig. 3, we demonstrated that EPA and DHA from erythrocyte membranes were a source of hypoxia-activated M2 macrophage production of RvE4 and SPM. These results provide mechanistic evidence for proresolving signaling that is directly linked to SPM biosynthesis initiated from erythrocyte-derived substrates via transcellular biosynthesis.

Along these lines, E-series resolvins exert potent host-protective actions in target cells and tissues. For example, RvE1 inhibits ADP-activated mobilization of P-selectin to minimize platelet aggregation (54), enhances monocyte and neutrophil phagocytosis and killing of Escherichia coli in human whole blood (20), and restores mitochondrial function in tumor necrosis factor–α–activated peripheral blood mononuclear cells (55). RvE2 is produced by human neutrophils activated with zymosan and is increased in hypoxia (56) and exerts shared actions with RvE1 on neutrophils (57). RvE3, unlike RvE1 and RvE2, is preferentially produced in human eosinophils and stops neutrophil chemotaxis and infiltration in zymosan peritonitis (58). In the brain, RvE1, RvE2, and RvE3 have all recently been shown to attenuate LPS-induced depression-like behavior in mice (59). Once produced by leukocytes, SPMs such as resolvin D1 also serve as autocrine signals (6062) that reduce the biosynthesis of proinflammatory mediators relevant in atherosclerosis. E-series resolvins, including the newly identified RvE4 from the present results, may therefore be of interest as mediators potentially contributing to beneficial actions of EPA supplementation in certain organs (35, 63).

In summation, the present results identify the hypoxic local environment as encountered in the acute inflammatory response, and organs such as bone marrow and lymphoid tissue (1), as an activator of proresolving mediator biosynthesis along with a new E-series resolvin. RvE4, and its structure and function characterized here, accelerates resolution of hemorrhagic exudates and inflammation, as well as the removal of apoptotic neutrophils and senescent erythrocytes. In addition, we uncovered mechanisms involving lipolysis of cholesteryl esters and triglycerides for SPM production by M2 macrophages, as well as metabolic targeting of neutrophil glycolysis that enhance RvE4-SPM specifically in hypoxic environment. Together, these findings demonstrate new links between physiologic hypoxia and SPM biosynthesis with production of a new RvE4 that, in turn, mediates proresolving functions.


Human peripheral blood isolation and neutrophil purification

Fresh human blood was collected with or without heparin (10 U/ml) from healthy volunteers. Each volunteer gave informed consent under protocol no. 1999-P-001297 approved by the Partners Human Research Committee (Institutional Review Board). All volunteers denied taking nonsteroidal anti-inflammatory drugs for ~2 weeks before donation. Human PMN were isolated by density-gradient Ficoll-Histopaque (Sigma-Aldrich).

Human M2 macrophages

M2 macrophages were obtained using validated procedures reported in (12). Human peripheral blood mononuclear cells from deidentified healthy human volunteers from the Children’s Hospital Boston Blood Bank were isolated by density-gradient, Ficoll-Histopaque isolation, which was followed by monocyte purification. The monocytes were then incubated for 7 days in RPMI 1640 medium (Lonza) containing 10% fetal calf serum, 2 mM l-glutamine, and 2 mM penicillin-streptomycin (Lonza) at 37°C and differentiated into macrophages through culturing with granulocyte-macrophage colony-stimulating factor (20 ng/ml) (PeproTech), followed by polarization for 48 hours with IL-4 (20 ng/ml) (PeproTech).

Human leukocyte and erythrocyte incubations in a physiologic hypoxic environment

Human M2 macrophages, neutrophils, and erythrocytes for physiologic hypoxia experiments were incubated in 1 ml Dulbecco’s Phosphate-Buffered Saline (DPBS) (Lonza) with Ca2+ and Mg2+ in six-well tissue culture–treated polystyrene plates (Thermo Fisher Scientific) and maintained in either normoxia (18.6% O2) or physiologic hypoxia (1% O2) (5, 6) according to procedures reported in (56). Cell incubations were stopped with two volumes of methanol in 1% O2 before LC-MS metabololipidomics to prevent reexposure of live cells to normoxia during sample workup. DPBS for physiologic hypoxia incubations were preconditioned for 24 hours in a hypoxic chamber (Thermo Fisher Scientific), with 1% O2, carbon dioxide (5%), balanced nitrogen, and water vapor. Cells were conditioned (37°C/5% CO2) for 24 hours, pH 7.45 in an incubator in normoxic conditions (18.6% O2) or in a hypoxia chamber (1% O2).

Human erythrocyte isolation and incorporation of deuterium-labeled EPA and DHA

Human peripheral blood erythrocytes from healthy human volunteers were isolated by centrifugation and aspiration of the platelet-rich plasma and buffy coat layer. Erythrocytes were purified by resuspension in RPMI 1640 (Lonza) [10% hematocrit (ht)] and centrifugation followed by aspiration of the top 10% of the erythrocyte layer (this purification procedure was carried out with five repetitions). Purified erythrocytes were then resuspended (20% ht) in RPMI 1640 containing deuterium-labeled EPA and DHA (d5-EPA, 10 μM and d5-DHA, 10 μM) with fatty acid–free human serum albumin (Sigma-Aldrich) and were incubated for 3 hours at 37°C to allow maximal incorporation of fatty acids into phospholipid membranes [see (64) for further details]. Cells were washed five times with 10 volumes of RPMI 1640 to remove extracellular free fatty acids.

LM metabololipidomics

Human macrophage incubations and mouse peritoneal lavages were immediately stopped by addition of two volumes of ice-cold methanol (Thermo Fisher Scientific) containing internal standards (500 pg each of d8-5-hydroxy-6E,8Z,11Z,14Z-eicosatetraenoic acid, d5-RvD2, d5-LXA4, d4-leukotriene B4 (LTB4), and d4-PGE2; Cayman Chemical Company) before solid-phase extraction and LC-MS/MS metabololipidomics. For experiments with d5-EPA– and d5-DHA–labeled erythrocytes, only d4-LTB4 was added to avoid cross-contamination with d5-EPA– and d5-DHA–derived products. Samples were placed on ice and protected from light for 45 min to allow for protein precipitation, followed by a centrifugation step (3000 rpm, 10 min, 4°C). Supernatants were collected from each sample, and solid-phase extraction was carried out according to optimized procedures reported in (22). Samples were loaded on C18 ISOLUTE 100-mg solid phase extraction (SPE) cartridges (Biotage) and washed with 3 ml of double distilled water followed by 3 ml of hexane. LMs were eluted in 3 ml of methyl formate (Sigma-Aldrich), evaporated under a gentle stream of nitrogen and resuspended in 50 μl of methanol/water (50:50) and analyzed by LC-MS/MS system, QTRAP 5500 (SCIEX) equipped with a Shimadzu LC-20AD high-performance LC (HPLC). The column implemented on this system was a Poroshell 120 EC-18 column (100 mm × 4.6 mm × 2.7 μm; Agilent), housed in a column oven regulated at 50°C, and LMs were eluted in a gradient of methanol/water/acetic acid from 55:45:0.01 (v/v/v) to 98:2:0.01 at a 0.5 ml/min flow rate. Targeted MRM and enhanced product ion scanning were used to quantify the mediator levels, with MS/MS matching of at least six diagnostic and signature ion fragments per molecule. A final analytical quantitation and recovery was performed using the deuterium-labeled internal standards, and an LM-SPM profile was created for sample.

RvE4 biogenic synthesis and isolation

Biogenic RvE4 was prepared by incubating EPA (Cayman Chemical) with soybean 15-LOX (Sigma-Aldrich). Briefly, EPA (50 μM) was suspended in 50 mM sodium borate buffer (pH 9.3; Sigma-Aldrich), and 15-LOX was added in 2-min increments until reaction was complete. The reaction mixture was quenched by adding one volume of cold methanol, and peroxide products were reduced with an excess amount of sodium borohydride. Reaction mixtures were taken to SPE for purification using the same SPE procedure for lipid mediator metabololipidomics (see above) before HPLC purification. RvE4 was separated from reaction mixture isomers on a chiral column (CHIRALCEL AD-RH, 150 mm × 2.1 mm) equipped with HPLC [Agilent HP 1100 Chemstation with diode array detector (DAD)] with gradient eluent, methanol/water (v/v) 95:5 to 100:0 for 15 min at the flow rate of 300 μl/min. Characteristic RvE4 chromophore of λmax = 244 nm was used for UV monitoring, and RvE4 was collected and stored at −80°C.

Human macrophage efferocytosis of senescent erythrocytes and neutrophils

Human erythrocytes were kept in DPBS containing Ca2+ and Mg2+ at 20% ht at 4°C for 24 hours to induce apoptosis. RBC senescence was assessed by diluting 1 μl of human whole blood or senescent cells into 100 μl of fluorescence-activated cell sorting (FACS) staining buffer followed by staining with anti-human CD235a allophycocyanin (APC)–Cy7 (clone HI264) and anti-human phycoerythrin (PE) CD47 (clone CC2C6) for phenotyping. RBCs were then suspended at 10% ht and stained with carboxyfluorescein succinimidyl ester (CFSE; 5 μM, 30 min, 37°C) and were washed three times with phosphate-buffered saline (PBS) via centrifugation and aspiration of supernatant. Human neutrophils were suspended in DPBS containing Ca2+ and Mg2+ (5 × 107 cells/10 ml) and placed in petri dishes and incubated for 24 hours at 37°C to induce apoptosis. Apoptotic neutrophils were removed from petri dishes with EDTA (5 mM) and stained with CFSE (5 μM, 30 min, 37°C). Human M2 macrophages were seeded in six-well plates (2 × 106 cells per well), treated with RvD5, RvD6, or RvE4 (10 nM), SPM panel (RvD2, RvD5, RvD6, MaR1, PD1, and RvE4; each together at 10 nM), or vehicle (0.01% ethanol) 15 min before addition of CFSE-labeled erythrocytes [1:50 (M2:RBC) ratio] or CFSE-labeled neutrophils [1:5 (M2:neutrophil) ratio]. After 1 hour of coincubation at 37°C, plates were washed thoroughly to remove noningested RBC, and cells were fixed with PBS containing 4% paraformaldehyde for 15 min on ice. Cells were washed with DPBS and removed from plates with a cell scraper in DPBS. Cells were then taken to flow cytometry to assess efferocytosis.

Flow cytometry

Human M2 macrophages and murine peritoneal exudate cells were suspended in FACS buffer (DPBS with 1% bovine serum albumin and 0.1% sodium azide) and incubated with Fc block (15 min, 4°C; BD PharMingen). Human M2 macrophages were incubated with anti-human fluorescein isothiocyanate (FITC) CD206 (clone 19.2) (BD Biosciences) and anti-human PerCP/Cy5.5 CD163 (clone RM3/1) (BioLegend) for phenotyping or intracellular staining with anti-human PE HIF-1α (clone 546-16, BioLegend). Murine peritoneal exudates were incubated with anti-mouse PerCP/Cy5.5 CD45 (clone 30-F11 BioLegend, CA, USA), anti-mouse PE/Cy7 CD11b (clone M1/70, eBioscience), anti-mouse APC F4/80 (clone BM8, eBioscience), anti-mouse FITC Ly6C (clone HK1.4, BioLegend), and anti-mouse PE Ly6G (clone 1A8, BioLegend) or appropriate isotype controls. Murine ex vivo efferocytosis was measured by surface labeling with anti-mouse APC F4/80 (clone BM8, eBioscience), followed by intracellular staining of neutrophils with anti-mouse FITC Ly6G (clone 1A8, BioLegend) and RBC with anti-mouse PE TER119 (clone TER-119, BioLegend) or isotypes. All flow cytometric samples were assessed using FACSDiva Canto II (BD Biosciences) and analyzed using FlowJo version X (TreeStar) and using procedures reported in (20).

Confocal microscopy

Subcellular localization of 5-LOX and 15-LOX. Immunofluorescence microscopy was carried out with slight modifications of procedures reported in (12). M2 macrophages (1 × 104 cells) were seeded on gelatin-coated glass coverslips in a 12-well plate and incubated in normoxia or hypoxia for 24 hours before fixation with 4% paraformaldehyde (15 min) and staining. Cells were then fixed with 4% paraformaldehyde (15 min) followed by incubation in 0.25% Triton X-100 (Sigma-Aldrich) containing Background Sniper (10 min, room temperature) to reduce background staining. Cells were then incubated with rhodamine phalloidin (Life Technologies) (30 min) for actin visualization, followed by incubation with mouse monoclonal anti-15-LOX (1:100) and rabbit monoclonal anti–5-LOX (1:100) (Abcam) overnight at 4°C. Coverslips were washed and labeled with the following secondary antibodies: Alexa Fluor 488 donkey anti-mouse (1:1000) and Alexa Fluor 647 donkey anti-rabbit (1:1000) (Life Technologies). To visualize the nuclei, coverslips were mounted with VECTASHIELD (Vector Laboratories) and examined by Zeiss LSM 800 with an Airyscan confocal system on a Zeiss Axio Observer Z1 inverted microscope equipped with an Axiocam 503 Monochrome Camera.

Phagocytosis. M2 macrophages (1 × 104 cells) were seeded on gelatin-coated glass coverslips in a 12-well plate and incubated in normoxia or hypoxia for 24 hours before addition of CFSE-labeled neutrophils (1:5 M2:neutrophil ratio) for 30 min. Cells were then fixed with 4% paraformaldehyde (15 min), followed by incubation in 0.25% Triton X-100 containing Background Sniper (10 min, room temperature) to reduce background staining. Cells were then incubated with rhodamine phalloidin (30 min). To visualize the nuclei, coverslips were mounted with VECTASHIELD (Vector Laboratories) and examined by Zeiss LSM 800 with Airyscan confocal system on a Zeiss Axio Observer Z1 inverted microscope equipped with an Axiocam 503 Monochrome Camera. At least 30 cells per condition were analyzed from each donor.

Murine hemorrhagic peritonitis

All experimental procedures were approved by the Brigham and Women’s Hospital Institutional Animal Care and Use Committee (protocol no. 2016N000145) and complied with institutional and U.S. National Institutes of Health guidelines. Male Friend leukemia virus B (FVB) mice (6- to 8-week old) were given zymosan A (1 mg/0.5 ml; Sigma-Aldrich), thrombin (5 U/0.5 ml; Sigma-Aldrich), or both for 0, 12, or 24 hours using procedures reported in (20). Mice were then euthanized with isoflurane before peritoneal lavage was performed with 5.0 ml of ice-cold PBS without divalent cations. Lavages were subjected to LC-MS/MS for metabololipidomics analysis and flow cytometric analysis of neutrophil numbers, macrophage numbers, and macrophage efferocytosis of both neutrophils and erythrocytes (see “Flow cytometry” section above for antibodies and staining protocol). Resolution indices were calculated as in (65): Tmax (the time interval when PMN numbers reach maximum), T50 (the time interval when PMN numbers reach 50% of maximum), and Ri (the interval between Tmax and T50).

Statistical analysis

Groups were compared by Student’s t test with Prism software version 6 (GraphPad). The criterion for statistical significance was P < 0.05. PCA was performed with SIMCA 13.0.3 software (MKS Data Analytics Solutions).


Supplementary material for this article is available at

Fig. S1. Hypoxia stimulates SPM biosynthesis in human M2 macrophages and neutrophils.

Fig. S2. Inhibition of human neutrophil glycolysis in physiologic hypoxia increases SPM production.

Fig. S3. SPM production in physiologic hypoxic human M2 macrophages and neutrophils utilizes lipase-dependent fatty acid mobilization.

Fig. S4. Senescent erythrocytes display reduced surface expression of CD47.

Fig. S5. M2 macrophage production of SPM in physiologic hypoxia is lipoxygenase dependent.

Fig. S6. RvE4 is produced in vivo in hemorrhagic exudates, and SPM epimers are increased with aspirin.

Fig. S7. Proposed model—hypoxia links lipolysis with RvE4 and SPM biosynthesis to drive efferocytosis in macrophages.

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We thank M. H. Small for expert assistance in manuscript preparation. Funding: This work was supported, in part, by the National Institutes of Health (grant nos. P01GM095467 and R01GM038765 to C.N.S. and F32HL142175 to S.L.). Author contributions: P.C.N. and S.L. designed, performed, and analyzed the experiments, as well as contributed to manuscript and figure preparations. P.C.N. performed LC-MS/MS lipidomics. S.L. performed flow cytometry and confocal microscopy analysis. C.N.S. designed the experiments and contributed to manuscript and figure preparation. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article