Resolvin D1 promotes efferocytosis in aging by limiting senescent cell-induced MerTK cleavage

Abstract

Inflammation-resolution is mediated by the balance between specialized pro-resolving mediators (SPMs) like resolvin D1 (RvD1) and pro-inflammatory factors, like leukotriene B₄ (LTB₄). A key cellular process of inflammation-resolution is efferocytosis. Aging is associated with defective inflammation-resolution and the accumulation of pro-inflammatory senescent cells (SCs). Therefore, understanding mechanism(s) that underpin this impairment is a critical gap. Here, using a model of hind-limb ischemia reperfusion remote lung injury (I/R), we present evidence that aging is associated with heightened inflammation, impaired SPM:LT ratio, defective efferocytosis and a decrease in MerTK levels in injured lungs. Treatment with RvD1 mitigated I/R lung injury in aging, promoted efferocytosis, and prevented the decrease of MerTK in injured lungs from old mice. Old MerTK cleavage resistant mice (MerTK^CR) exhibited less PMN infiltration and had improved efferocytosis compared with old WT controls. Mechanistically, macrophages that were treated with conditioned media from senescent cells (CM) had increased MerTK cleavage, impaired efferocytosis and a defective RvD1:LTB₄ ratio. Macrophages from MerTK^CR mice were resistant to CM-induced efferocytosis defects and had an improved RvD1:LTB₄ ratio. RvD1-stimulated macrophages prevented CM-induced MerTK cleavage and promoted efferocytosis. Together, these data suggest a new mechanism and a potential therapy to promote inflammation-resolution and efferocytosis in aging.

Introduction

Aging is associated with persistent, non-resolving inflammation, or inflammaging, which results in functional impairment at the tissue, cellular and molecular levels(1). Inflammaging drives tissue damage and ultimate physical decline(2) so identifying mechanisms to appropriately control this response addresses a critical gap in medicine. The resolution of inflammation (or inflammation-resolution) is governed by several factors, including specialized pro-resolving mediators (SPMs). SPMs comprise a family of arachidonic acid (AA)-derived lipoxins, docosahexanoic acid (DHA)-derived D-series resolvins, protectins and maresins, and eicosapentanoic acid (EPA)-derived E-Series resolvins(3). SPMs temper inflammation and promote tissue repair and regeneration, without causing immunosuppression (3). A key process of inflammation-resolution includes the clearance of apoptotic cells, or efferocytosis(4). Importantly, when apoptotic cells are not efficiently cleared, they undergo secondary necrosis and release harmful mediators that exacerbate inflammation and tissue damage(5). Efferocytosis is required for tissue repair and homeostasis and although aging is associated with impaired efferocytosis(6, 7), mechanisms underlying this defect are currently underexplored.

Macrophages are key cellular players in resolution because of their ability to generate SPMs and to carry out efferocytosis(4, 5). MerTK is a well-known efferocytosis receptor on macrophages and in certain contexts, MerTK signaling has been shown to increase the synthesis of SPMs over pro-inflammatory leukotrienes (LTs)(8–10). Furthermore, MerTK undergoes an ADAM17-mediated cleavage event during inflammation or oxidative stress (OS) in which the extracellular portion of the receptor is shed, resulting in decreased functional MerTK on the surface of macrophages(11). Whether MerTK cleavage plays a role in inflammaging or impaired efferocytosis in aging is currently not known. Moreover, inflammaging is associated with an accumulation of senescent cells in tissues. Senescent cells undergo a phenotypic switch called the senescence associated secretory phenotype (SASP), which consists of proteolytic and pro-inflammatory factors(12). Because MerTK is cleaved by inflammation and ROS, we also questioned whether the SASP cleaves MerTK to diminish efferocytosis and promote inflammaging.

Here we found that aging is associated with increased hind-limb ischemia reperfusion (I/R) lung injury, a defect in the SPM:LT ratio and local MerTK cleavage. Administration of the SPM Resolvin D1 (RvD1) to old mice decreased I/R-lung injury, promoted efferocytosis and restored MerTK levels. Old MerTK cleavage resistant mice (MerTK^CR) demonstrated less I/R-lung injury and improved efferocytosis compared with old wild type controls. Mechanistically, we found that conditioned media from SCs (CM) promoted MerTK cleavage, limited efferocytosis and reduced the RvD1:LTB₄ ratio. Macrophages from MerTK^CR mice were resistant to CM-induced efferocytosis defects and had an improved RvD1:LTB₄ ratio. Lastly, RvD1 prevented CM-induced MerTK cleavage and promoted efferocytosis. Together, these results shed light on a new mechanism as to why efferocytosis is defective in aging and suggest that RvD1 may be a therapeutic approach to promote efferocytosis and limit inflammation in aging.

Materials and Methods

Hind-limb injury model of ischemia/reperfusion.

Vehicle or RvD1 treatment:

Male C57/BL6 mice were purchased from Taconic at 52 weeks (12 months old). These mice were then housed in the AMC animal facility for an additional 5 months and experiments were carried out between 16 and 17 months of age. Young (10-week-old) male mice were purchased from Taconic and housed in the AMC animal facility for 2 weeks and experiments were carried out on 3-month-old mice. Old or young mice were placed in an isoflurane chamber (O₂ flow rate of 2L/min, 2% isoflurane) for the duration of the hind-limb ischemia reperfusion injury experiments. Hind-limb ischemia reperfusion injury was carried out as previously described(13). Briefly, both hind limbs were ligated with rubber bands and after 1 hr., under continued anesthesia, the rubber bands were removed and Veh or RvD1 (500 ng/mouse) were intravenously injected. Reperfusion was carried out for 150 mins, after which the mice were sacrificed, the lungs were perfused, removed and placed in 4% PFA for tissue sections and H&E analyses.

WT and MerTK^CR experiments:

WT and MerTK^CR mice were aged in the AMC animal facility for 16 months. Young controls were 3 months old at the time of experiments. Both hind limbs were ligated with rubber bands as above. After 1 h, under continued anesthesia, the rubber bands were removed, and reperfusion was carried out for 150 mins. The lungs were perfused with PBS, and quickly removed and either snap-frozen in liquid nitrogen for ELISA analysis or placed in 4% PFA for tissue sections and H&E analyses.

Quantitation of lung myeloperoxidase (MPO) or IL-6.

The left lung from each mouse was homogenized in 450 μL of ice-cold RIPA buffer. The tissue was homogenized with a bullet blender and then centrifuged at 12,000 rpm for 10 minutes. The supernatants were transferred to clean tubes, and myeloperoxidase (MPO) or IL-6 was quantified using R&D ELISA kits as per manufacturer’s instructions. The ELISA values were normalized to total lung protein.

Identification and quantification of lipid mediators by LC-MS/MS

Murine lungs were harvested and immediately placed in 1 mL of ice-cold methanol, followed by mincing. Samples were then stored at −80°C until lipid mediator extraction was performed. Sample preparation and metabololipidomic analysis was then conducted using methods and instrumentation as described recently in detail (14). Briefly, deuterium-labeled internal standards (e.g. d₅-RvD2, d₄-LXA₄, d₄-LTB₄, d₄-PGE₄) were added to the samples to assess extraction recovery. Samples were then centrifuged, and supernatants were collected and subjected to solid phase extraction using C18 cartridges. Lipid mediators were eluted from the column with methyl formate and the collected fraction was concentrated using a steady stream of N₂ gas. The samples were then resuspended in methanol:water (50:50) prior to injection and analysis by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Lipid mediators were identified using criteria including retention time, specific multiple reaction monitoring (MRM) transitions and diagnostic MS/MS fragmentation spectra. Quantification of mediators was accomplished using standard curves determined for each mediator with authentic standards following normalization of extraction recovery based on internal deuterium-labeled standards.

Senescence-associated secretory phenotype - conditioned media preparation

Conditioned media from senescent IMR-90 fibroblast cells was collected according to the method described in (15). Briefly, sub-confluent IMR-90 cells (passage 6–15, 0.5×10⁶ cells/10 cm tissue culture dish) were subjected to 10 gray of ionizing radiation and allowed to senesce for 10 days. The cells were refreshed with 10% FBS containing DMEM on day 2, day 5 and day 8. On the 10^th day, senescent cells were incubated with serum-free DMEM culture media. After 24h, the serum-free SASP conditioned media was filtered using 0.45μm syringe filter and used for functional assays.

Soluble MerTK Western blotting

Bone marrow derived macrophages (BMDMs) from C57/BL6 mice (10–12 weeks of age) were cultured for 7 days in DMEM containing 10% FBS (vol/vol), 20% L cell media (vol/vol). The BMDMs were then plated (0.5 ×10⁶ cells/well) and cultured in 12-well tissue culture-treated plates and stimulated with Veh, 10 nM RvD1, or 10 μM NAC for 20 mins in serum free containing media. Conditioned media from SCs or control media (500 uL/well) were added for an additional 2 hrs. The media were collected, concentrated 10-fold with ultracentrifuge filters (10,000 molecular weight cut-off, Amicon), lysed with 4x Laemmli sample buffer, and subjected to SDS-polyacrylamide gel electrophoresis and immunoblot analysis as in (8).

In vivo efferocytosis

Male C57BL/6-Tg(UBC-GFP)30Scha/J mice were intraperitoneally injected with 200 μg of zymosan to elicit PMN. PMN were collected by lavage 6 hrs. post zymosan injection. PMN were resuspended in DMEM and placed in an incubator (37°C, 5% CO₂) overnight to stimulate apoptosis. Apoptotic PMN (3 × 10⁶ cells/mouse) were intravenously injected and mice were euthanized 1 hr. post i.v. injection. Spleens were harvested and single cell preparations were generated to evaluate efferocytosis of transferred GFP+ apoptotic cells using flow cytometry. Splenocytes were evaluated for CD11b (clone M1/70, PE-Cy7 conjugated) and Ly6G (clone 1A8, PE-conjugated) using an LSR II equipped with Diva software (BD Biosciences). CD11b⁺ Ly6G⁻ negative cells were evaluated for GFP expression to enumerate macrophages that engulfed apoptotic donor PMN. Analysis was performed using FlowJo (Treestar).

In situ efferocytosis

Lungs were collected from IR-induced C57BL/6J mice and embedded in paraffin as described previously (13). Sections were cut and mounted on microscopic slides at the histology core at Albany Medical College. Before staining, sections were deparaffinized by incubation in a 60°C oven for 1hr followed by immersion in xylene and rehydration in graded series of ethanol. For detecting apoptotic cells, sections were incubated with perm wash (BD Biosciences, 554723) for 8 mins. Sections were then washed 2x in 1XPBS. TUNEL staining was performed for labeling apoptotic cells using the In Situ Cell Death Detection Kit, TMR red according to manufacturer’s instructions (Roche, 12156792910). For detecting macrophages, sections were blocked with 1% BSA in 1x PBS for 20 mins. Sections were further incubated with CD107b or Mac3 (BD Pharmingen, 553322, clone M3/84) in 1% BSA at 1:100 overnight at 4°C. Excess antibody was removed by washing with 1x PBS. Alexa 488 anti-rat antibody was then added to the sections at 1:200 and incubated for 2hrs at room temperature. Nuclei were stained with Hoechst for 10 mins followed by mounting the slides. Images were taken using the Leica SPE microscope. Six images were taken per mice per condition.

To determine efferocytosis, free and associated TUNEL-positive cells (apoptotic) were counted using FIJI software. Apoptotic cells that were surrounded or in close contact with the macrophages were defined as macrophage-associated and those not associated with macrophages were defined as free. The ratio of associated to free apoptotic cells was quantified(16–18).

In vitro efferocytosis

Murine BMDMs (0.25 × 10⁶ cells/well, 24-well plates) were plated in DMEM containing 10% FBS, and 20% L cell media (vol/vol) 24 hrs prior to efferocytosis assays. Parallel studies used human monocyte-derived macrophages that were plated (0.2 × 10⁶ cells/well in a 24-well plate) in RPMI containing 10% FBS. The following day Jurkats were collected, enumerated and with stained with PKH26 (Sigma) according to the manufacturer’s instructions. Excess dye was removed and the Jurkats were resuspend in RPMI containing 10% FBS. To induce apoptosis, Jurkats were then exposed to UV irradiation (0.16 Amps, 115 Volts, 254nm wavelength) for 15 mins in room temperature and then placed in an incubator (37°C, 5% CO₂) for 3 hrs. Macrophages were either stimulated with 10 nM RvD1 or treated with 10 μM NAC 20 mins prior to the addition of the SASP. After 20 mins, the SASP was added for an additional 2 hrs, after which apoptotic Jurkats were added in a 3:1 ratio. Efferocytosis was carried out for 30 mins, after which non-engulfed apoptotic Jurkats were washed off. Human macrophage efferocytosis experiments were conducted as above, but efferocytosis was carried out for 1 hr. (37°C, 5% CO₂). Macrophages were imaged using a ZOE BioRad fluorescence imager and quantified with ImageJ.

Quantification of RvD1 and LTB₄ by ELISA:

Murine BMDMs (0.5 × 10⁶ cells/well in a 12-well plate) were plated in DMEM containing 10% FBS, and 20% L cell media (vol/vol) and efferocytosis was carried out as above. Supernatants were collected and excess Jurkats were removed by centrifugation (5,000 rpm, 5 mins, 4°C). Supernatants were then collected and subjected to RvD1 or LTB₄ ELISA (Cayman Chemical).

Human monocyte derived macrophage differentiation.

Buffy coats from de-identified healthy human volunteers were purchased from the New York Blood Center. Human peripheral blood mononuclear cells (PBMCs) were isolated by density-gradient centrifugation using Ficoll-Histopaque-1077 (Sigma, #10771) as in(10). Briefly, PBMCs were collected and plated in a petri dish for 1hr (37°C, 5% CO₂) to allow for adhesion of monocytes. After 1 hr., floating cells were removed and the remaining attached cells were cultured in RPMI media containing 10% FBS, 10 ng/mL recombinant human GM-CSF, for 10 days. Media containing 10 ng/mL of GM-CSF was replenished on day 3.

Determination of p16^INK4A expression in lungs

Perfused lungs were homogenized in RLT buffer and RNA was isolated with a Qiagen RNeasy isolation kit (Qiagen). qPCR was carried out with a BioRad CFX Connect Real-Time qPCR system. The sequences of the murine p16^INK4A and 18S are as follows: Murine p16^INK4A primer sequence is forward: 5’-AATCTCCGCGAGGAAAGC-3’, Reverse: 5’-GTCTGCAGCGGACTCCAT-3’. Murine 18S was used as a house keeping control and the sequence is forward 5’-ATGCGGCGGCGTTATTCC, reverse 5’-GCTATCAATCTGTCAATCCTGTCC-3’.

Statistical analysis:

Mice were randomly assigned into groups for all in vivo experiments. All in vitro experiments were repeated at least 3 times and the statistical significance was evaluated. The human macrophage experiments represented at least three separate human donors. Results are shown as mean ± SEM, and statistical differences were determined using the two-tailed Student’s t-test, one-way ANOVA or two-way ANOVA depending on the appropriate contexts. Prism (GraphPad Inc., La Jolla, CA) software was used and a p < 0.05 was considered statistically significant. Details regarding statistical tests can be found in the figure legends.

Results

Aging enhances remote lung injury and is associated with a low SPM:LT ratio in I/R lungs

Ischemia reperfusion (I/R) injury occurs when blood flow is restored to tissues that have sustained a period of interrupted blood supply(19). Reperfusion causes local and distant (remote) organ injury and is associated with various surgical procedures including cross-clamping, cardiopulmonary bypass, coronary artery bypass graft (CABG), and organ transplantation (20). In these settings, reperfusion can lead to neutrophil-mediated tissue damage in various tissues like the lung and heart as examples (19). We used a model of hind limb I/R to mimic temporary vascular occlusion, reperfusion and subsequent remote organ (e.g. lung) damage(21). For these experiments, ischemia was carried out for 60 minutes, after which tourniquets were removed and reperfusion occurred for 150 minutes. Mice were then sacrificed, lungs were perfused and harvested for analysis. We compared I/R-challenged young mice (3 months of age) with old mice (16 months of age) and found that old mice had significantly more PMN (which was assessed by lung MPO levels) compared with young mice (Fig. 1A). These results suggest that aging exacerbates I/R-induced remote lung injury.

Figure 1. Aging enhances ischemia reperfusion (I/R) remote lung injury and impairs the SPM:LT ratio.

Young (3 months) or old (16 months) male C567Bl6 mice were subjected to bilateral hind limb ischemia for 60 minutes, followed by 150-minute reperfusion. (A) To quantify PMN, left perfused lungs were homogenized and were assayed for MPO by ELISA. MPO values were normalized to total lung protein. Results are mean ± SEM, n = 5 per group, *p < 0.05., t-test. (B-D) Right lungs were frozen in liquid nitrogen and subjected to LC-MS/MS analysis. (B) Representative MRM chromatograms (C) and representative MS-MS spectra from old lungs subjected to I/R are shown. (D) SPMs and LTs were quantified and are represented as a ratio of SPM:LT. Results are mean ± SEM, n = 6 per group *p < 0.05 t-test.

The accumulation of senescent cells is associated with aging(22). Senescent cells are maladaptive in aging because they undergo a highly pro-inflammatory phenotypic switch called the senescence associated secretory phenotype (SASP)(12). Therefore, we first questioned whether there was an increase in senescence markers in lungs from aged mice. As expected, we observed significantly higher expression of p16^INK4A (i.e. a marker for senescent cells) in lungs from old mice, compared with lungs from young mice (Fig. S1A). Moreover, we found that levels of a major SASP cytokine called IL-6 were significantly increased in aged lungs subjected to I/R compared with young controls (Fig. S1B).

Because imbalances in SPMs and LTs are associated with aging and age-related diseases(6, 17), we next questioned whether old mice in the context of I/R injury also had a low SPM:LT ratio. Ischemia reperfusion was carried out as above and lungs were harvested for LC-MS-MS analysis. Several SPMs were identified in both the young and old I/R lungs, including D-series resolvins, protectins, lipoxins and RvE1 (Table 1). MRM chromatograms of the identified SPMs and representative mass spectra of LTB₄ and RvD1 are shown in Fig. 1B and C, respectively. SPMs (i.e. 15R-LXA₄, LXB₄, 15R-LXB₄, RvD1–6, 10S,17S-diHDHA, and RvE1) and LTs (LTB₄, 6-trans LTB₄, and 6-trans, 12epi-LTB₄) were quantified and we found a significant imbalance in the SPM:LT ratio compared with young controls (Fig. 1D). Overall, these data suggest that there is heighted inflammation, increased SASP, and an imbalance in the SPM:LT ratio in injured lungs from old mice.

Table 1. Identified lipid mediators from I/R-induced injured lungs in young and old mice.

Experiments were carried out as described in the methods. Results are mean ± SEM, n = 6 per group.

	Young I/R			Old I/R
	AVG	SEM		AVG	SEM
AA	21065.61	1640.08		30662.50	1950.64
15R-LXA4	47.73	11.40		127.73	33.84
LXB4	1396.32	280.11		1655.82	285.15
15R-LXB4	184.09	37.15		333.11	54.83
LTB4	71.32	33.88		239.38	118.66
6-trans LTB4	10.28	3.67		47.01	23.40
6-trans, 12-epi LTB4	11.93	4.17		42.78	17.64
PGE2	7201.86	1724.54		8818.19	1714.92
PGD2	3538.37	963.55		5129.06	1072.45
PGF2a	1793.54	562.04		2381.35	478.41
TXB2	1714.14	569.56		2723.00	653.70
15-HETE	2166.09	467.16		4676.95	805.00
12-HETE	2845.86	896.47		7298.29	852.28
5-HETE	346.42	139.90		1227.94	631.71
12-HHT	3308.46	1002.64		6386.44	1063.45
DHA	16903.17	1497.20		21356.70	1538.86
RvD1	11.96	3.72		12.21	3.48
RvD4	3.04	0.84		4.09	1.29
RvD5	3.62	1.30		12.60	7.05
RvD6	2.71	0.58		5.61	2.21
PD1	11.66	1.44		14.01	4.28
17R-PD1	7.26	1.17		13.17	7.68
10S,17S-diHDHA	65.47	18.14		153.42	96.22
17-HDHA	1288.67	286.24		2294.44	949.65
14-HDHA	2145.04	601.33		4180.16	1564.28
7-HDHA	58.00	14.84		87.09	22.79
4-HDHA	145.39	37.95		377.02	185.28
EPA	13371.97	1132.80		19113.53	572.05
RvE1	5.77	0.81		21.48	6.02
18-HEPE	161.08	44.13		338.45	133.66
15-HEPE	376.38	110.13		1159.80	695.17
12-HEPE	1376.73	639.53		3255.16	783.74
5-HEPE	110.57	30.88		280.56	130.95

RvD1 limits I/R-induced remote lung injury and promotes efferocytosis in aging

To test causation for the low SPM:LT ratio, we sought to tip the balance in favor of SPMs by treating mice with RvD1, which was one of the identified SPM from above (Fig.1 C,D). For these experiments, ischemia was carried out for 60 minutes, after which tourniquets were removed and Veh or 500 ng/mouse of RvD1 were intravenously injected. Reperfusion occurred for 150 minutes and mice were then sacrificed, lungs were perfused and harvested for analysis. Representative H&E images of lungs from sham young and old mice are shown and reveal no inflammation or injury prior to I/R-induced injury (Fig. S2A). As expected, there was a significant increase in neutrophils (PMN) in lungs from old mice, compared with young controls (Fig. 2A). Representative H&E images are shown on the left and clearly depict increased cells and injury in lungs subjected to I/R from old mice (Fig. 2A, top right panel). First RvD1 significantly decreased PMN infiltration into the lungs from young mice, which is consistent with the literature(13). RvD1 also significantly decreased PMN infiltration into the lungs from old mice and diminished PMN to levels commensurate to young I/R challenged mice (Fig. 2A). These results suggest that RvD1 dampened age-associated I/R-induced lung injury (Fig. 2A).

Figure 2. RvD1 limits excessive inflammation, promotes efferocytosis and prevents to loss of MerTK levels in aged lungs subjected to hind limb I/R.

Young (3 months) or old (17 months) male C57Bl6 mice were subjected hind limb I/R as in Fig. 1 and Veh or 500 ng/mouse of RvD1 were injected i.v. just after reperfusion. (A) Representative H&E images of perfused lungs are shown on the left. Scale bar is 20 μm. PMN were enumerated based on H&E staining. At least 4 lung sections per mouse per group were quantified. (B) Efferocytosis was assessed by calculating the ratio of TUNEL-associated to TUNEL-free macrophages in lung sections. Lung sections were immunostained with anti-mac3 antibody for macrophages, TUNEL for apoptotic cells and counterstained with Hoechst for nuclei. Images were acquired on a Leica SPE confocal microscope and were analyzed with ImageJ. (C) Lungs were immunostained with anti-MerTK antibody and counterstained with Hoechst. Images were acquired on a Leica SPE confocal microscope and analyzed with Imaris software. (A-C) Results are mean ± SEM, n = 4 per group *p < 0.05 young Veh versus other groups, ^p < 0.05 old Veh versus old RvD1, Two-way ANOVA with Tukey’s multiple comparisons.

Local removal of dead cells, or efferocytosis, is critical for tissue repair. Therefore, we next sought to explore the role of efferocytosis in this context. Lung sections were immunostained with Mac3 for macrophages and TUNEL for apoptotic cells and a ratio of TUNEL-associated macrophages to TUNEL-free macrophages were quantified. We observed that lungs from old mice had significantly less efferocytosis compared with I/R challenged lungs from young mice (Fig. 2B). These results are in agreement with the literature with regard to age-related defects in efferocytosis (6, 7, 23–25). Importantly, RvD1 promoted efferocytosis in the lungs from young and old mice (Fig. 2B). Together, these results suggest that tipping the balance in favor of SPMs by the addition of RvD1 during the reperfusion phase was beneficial toward limiting excessive inflammation and impaired efferocytosis associated with aging.

MerTK, a critical efferocytosis receptor on macrophages, undergoes a cleavage event in the presence of certain pro-inflammatory factors or ROS(11). MerTK cleavage results in lower surface levels of MerTK, which renders the receptor less active and results in impaired efferocytosis(11). Because I/R challenged lungs from old mice had increased inflammation and reduced efferocytosis, we next questioned if MerTK levels were lower in lungs from old mice and if so, whether RvD1 prevented the loss of MerTK. Lung sections were immunostained with an anti-MerTK antibody and levels of MerTK in the lungs were assessed by confocal microscopy and analyzed with Imaris software. Indeed, there was a significant decrease in MerTK levels in lungs from old mice subjected to I/R compared with young controls (Fig. 2C) and RvD1 prevented the decrease of MerTK in old mice (Fig. 2C). Overall these results suggest that RvD1 limits PMN infiltration, promotes efferocytosis and retains macrophage MerTK levels in lungs from old mice subjected to I/R.

MerTK cleavage promotes tissue injury and impairs efferocytosis in aging

To prove causation for the decreased levels of MerTK in lungs from old mice subjected to I/R, we used a mouse in which full length MerTK was replaced with a functional yet cleavage resistant MerTK(8). Here, we aged wild-type (WT) or MerTK cleavage resistant (MerTK^CR) mice and performed hind limb I/R remote lung injury in old (16 month old) and young (3 month old) mice. I/R injury was carried out as above and representative H&E images of lungs from sham young and old WT and MerTK^CR are shown in Fig. S2B and reveal no injury prior to I/R. Representative H&E images of lungs subjected to I/R are shown in Fig. 3A (left panels) and there is a clear increase of cellular infiltrates in the lungs from old WT mice compared with other groups. PMN were quantified by MPO ELISA, which was normalized to total lung protein (Fig. 3A). Similar results were obtained when PMN were quantified by H&E staining and were enumerated as number of PMN per 100x visual field (Fig. S3A). First, lungs from old WT mice that were subjected to I/R-induced injury had significantly more PMN compared with young controls (Fig. 3A, right panel). Lungs from old MerTK^CR mice that were subjected to I/R had significantly less PMN compared with old WT controls and had MPO levels that were commensurate to young WT injured lungs (Fig. 3A). MerTK^CR mice also had decreased PMN in young mice, which is consistent with the literature(8). Lungs from old MerTK^CR mice subjected to I/R also had significantly less lung IL-6 compared with aged WT controls (Fig. 3B), which suggests that MerTK cleavage drives excessive inflammation associated with hind limb I/R injury in aging.

Figure 3. Old MerTK mice have decreased remote lung injury, limited IL-6 levels and enhanced efferocytosis compared with old controls.

Young (3 months) or old (16 month) males were subjected to hind limb ischemia reperfusion injury as in Fig 1. (A) Perfused lungs were sectioned and stained with H&E and representative images are shown on the left. To quantify PMN, left perfused lungs were homogenized and were assayed for MPO by ELISA (right panel). MPO values were normalized to total lung protein and values are represented as “arbitrary units” (AU). (B) Left perfused lungs were homogenized and were assayed for IL-6 by ELISA. IL-6 values were normalized to total lung protein and are represented as arbitrary units (AU). (A-B) Results are mean ± SEM, n = 4 per group *p < 0.05 young WT versus other groups, ^p < 0.05 old WT versus old MerTK^CR, Two-way ANOVA with Tukey’s multiple comparison test. (C) Efferocytosis was carried out as in Fig. 2B Results are mean ± SEM, n = 4 per group, **p < 0.01, t-test.

Next, we examined in situ efferocytosis in I/R lungs and found that the lungs subjected to I/R from old MerTK^CR mice had significantly more efferocytosis compared with old WT controls (Fig. 3C). To further explore efferocytosis in aging, we next performed an in vivo efferocytosis assay in which 3 million GFP-labeled apoptotic PMN from C57/BL6-Tg(UBC-GFP)30Scha/J mice we intravenously injected into young WT control, aged WT or old MerTK^CR mice. After 1 hr., spleens were removed and processed for flow cytometry. Splenic macrophages were subjected to flow cytometry and CD11b⁺LyG6⁻ macrophages that were also positive for GFP were quantified as an efferocytic event. Indeed, old MerTK^CR mice had improved efferocytosis compared with old WT controls (Fig. S3B), further suggesting that MerTK cleavage, is a mechanism that deranges efferocytosis in aging.

The SASP limits efferocytosis through MerTK cleavage

Because aging is associated with an accumulation of senescent cells and SASP(22), we next questioned whether the SASP promoted MerTK cleavage and impaired efferocytosis. Conditioned media from senescent cells (CM) was generated from γ-irradiated IMR-90 cells (15). CM or control media was incubated with macrophages for 2 hrs, after which PKH26-labeled apoptotic Jurkats, were then co-cultured with these macrophages in a 3:1 ratio for an additional 30 mins. at 37°C. Images were acquired on a fluorescence microscope and efferocytosis was quantified as the percent of efferocytosis per total macrophages within a 40x visual field. We observed that SASP-treated macrophages had significantly less efferocytosis compared with Veh-treated cells (Fig. 4A). To determine whether conditioned media from senescent cells (CM) promoted MerTK cleavage we stimulated macrophages with CM or Veh media for 2hr, after which we collected the supernatant and immunoblotted for soluble-Mer (sol-Mer), a readout for MerTK cleavage. Indeed, we found that the CM led to an increase in sol-Mer (Fig. 4A, inset). To prove causation for the role of MerTK cleavage in promoting CM-induced defects in efferocytosis, we next assessed efferocytosis in macrophages from WT and MerTK^CR mice. Again, we found that the CM significantly decreased efferocytosis in WT macrophages (Fig. 4B). Macrophages from MerTK^CR were protected from the CM-induced defect in efferocytosis (Fig. 4B). Efferocytosis promotes the biosynthesis of SPMs over LTs (26) and MerTK was recently shown to play a critical role in this process (8, 10). Therefore, we next questioned whether the CM impacted the SPM:LT ratio and if so, was this due MerTK cleavage events. Indeed, CM-treated macrophages had a significant imbalance in the RvD1:LTB₄ ratio during efferocytosis and MerTK^CR macrophages rescued this impairment (Fig. 4C). Together, these results suggest that conditioned media from senescent cells limits efferocytosis and diminishes the SPM:LT ratio through MerTK cleavage events.

Figure 4. Conditioned media from senescent cells (CM) impairs efferocytosis and the RvD1:LTB₄ ratio through MerTK cleavage events.

(A) Murine bone marrow-derived macrophages (BMDM) were incubated with Veh or conditioned media from senescent cells (CM) for 2 hours. In parallel, Jurkats were stained with PKH26 (red) and subjected to UV-radiation to initiate apoptosis. The red apoptotic Jurkat cells were then co-cultured with macrophages in a 3:1 ratio for 30 minutes and images were acquired with a BioRad Zoe Fluorescence Cell Imager. Results are mean ± SEM, n = 3 separate experiments, ***p< 0.001, two-tailed unpaired Student’s t-test. (A, inset) MerTK cleavage was monitored by the presence of sol-Mer in supernatants. Briefly, Veh or CM media was incubated with macrophages for 2 hours. Media was collected, concentrated and subjected to W. Blot analysis. Results are representative of 3 separate experiments each done in triplicate. (B) Efferocytosis was carried out as above with BMDMs from with WT or MerTK^CR mice. Results are mean ± SEM, n = 3 separate experiments, ***p< 0.001, two-way ANOVA, Tukey’s multiple comparisons test. (C) Efferocytosis was carried out as above and supernatant were assayed for RvD1 and LTB₄ by ELISA. Results are mean ± SEM, n = 3 separate experiments, **p< 0.01, one-way ANOVA, Kruskal-Wallis test. (D) Macrophages were treated with Veh or 10 μM NAC for 20 minutes prior to the addition of CM. Inset, Media was collected 2 hours after the addition of Veh or CM and was subjected to W. Blot as above. Results are representative of 3 separate experiments each done in triplicate. Efferocytosis was carried out as above. Results are mean ± SEM, n = 4 separate experiments **p< 0.01, one-way ANOVA, Tukey’s multiple comparisons test.

Because ROS is known to stimulate MerTK cleavage we next questioned whether SASP-induced MerTK cleavage and impaired efferocytosis was due to an increase in ROS in macrophages. We incubated macrophages with 10 μM N-acetylcysteine (NAC), which is an inhibitor of ROS. Indeed, NAC limited MerTK cleavage (Fig. 4D, inset) and promoted efferocytosis (Fig. 4D) in the presence of CM. These results suggest that ROS plays a critical role in promoting CM-induced MerTK cleavage and defective efferocytosis. The remaining question is whether SPM treatment can limit CM-induced MerTK cleavage and efferocytosis defects.

RvD1 limits senescent cell conditioned media-induced defects in efferocytosis

SPMs like RvD1 are biosynthesized during efferocytosis and in a feed forward manner, also act on macrophages to enhance efferocytosis(3, 8). Because the CM from senescent cells impaired the RvD1:LTB₄ ratio (Fig. 4C), we next questioned whether RvD1 could rescue CM-induced defects in efferocytosis and limit MerTK cleavage. Macrophages were stimulated with 10nM of RvD1 or Veh for 20 mins prior to the addition of CM. CM was then incubated with the macrophages for an additional 2 hrs and sol-Mer and efferocytosis were assessed as above. First, we observed that RvD1 decreased CM-induced sol-Mer (Fig. 5A) and promoted efferocytosis in the presence of CM (Fig. 5B). Because NAC limited CM-induced MerTK cleavage and increased efferocytosis, we next questioned whether RvD1’s actions were through limiting ROS in macrophages. For these experiments, 10 nM RvD1, 10 μM NAC or RvD1 and NAC were incubated with macrophages, after which CM was added for an additional 2 hours. NAC or RvD1 on their own rescued CM-induced defects in efferocytosis and the co-treatment was not additive, which suggests that RvD1 may promote efferocytosis through limiting ROS in macrophages (Fig. 5B). We also found that RvD1 did not further enhance efferocytosis in MerTK^CR macrophages, again suggesting that RvD1 promotes efferocytosis by preserving surface levels of MerTK on macrophages (Fig. 5C). Lastly, we also found that RvD1 significantly increased efferocytosis in human macrophages stimulated with CM (Fig. 5D). Together these results suggest that RvD1 limits CM-induced MerTK cleavage to promote efferocytosis.

Figure 5. RvD1 restores CM-induced defects in efferocytosis and MerTK cleavage.

(A) Macrophages were treated with Veh or 10 nM RvD1 for 20 minutes prior to the addition of conditioned media from senescent cells (CM). Media was collected 2 hours after the addition of Veh or CM media and then subjected to W. Blot as in Fig 4. Results are representative of 3 separate experiments each done in triplicate. (B) BMDMs were stimulated with 10 nM RvD1, 10 μM NAC or RvD1 and NAC for 20 minutes prior to CM addition. Efferocytosis was then carried out as above. Results are mean ± SEM, n = 3 separate experiments, **p< 0.01, one-way ANOVA, Tukey’s multiple comparisons test. (C) Efferocytosis assays were carried out as above with macrophages from WT or MerTK^CR mice. Results are mean ± SEM, n = 3 separate experiments, **p< 0.01, two-way ANOVA, Tukey’s multiple comparisons test. (D) Efferocytosis was carried out as in Fig. 4. with human macrophages. Results are mean ± SEM, n = 3 separate experiments, ***p< 0.001, one-way ANOVA, Tukey’s multiple comparisons test.

Discussion

The findings of this study provide a new mechanism associated with defective efferocytosis in aging and further evidence that the critical SPM:LT ratio is impaired in inflammaging. Herein we also uncovered a new maladaptive role for the SASP in which the conditioned media from senescent cells deranged efferocytosis through MerTK cleavage. Importantly RvD1 promoted efferocytosis in aging and prevented CM-induced MerTK cleavage. These results offer a new framework as to how senescent cells derange inflammation-resolution programs, which may lead to the development of new therapeutics to thwart inflammaging.

Cellular senescence is emerging as a major player in inflammaging(22). Senescent cells accumulate in tissues and undergo a phenotypic switch in which they release a barrage of proteolytic and pro-inflammatory factors (12). We observed that the CM promoted MerTK cleavage to limit efferocytosis. MerTK is cleaved by ADAM17, which is a metalloproteinase that is activated by pro-inflammatory ligands as well as ROS(11). Along these lines, we also found that NAC prevented CM-induced MerTK cleavage which suggests that the released factors from senescent cells leads to ROS within the ingesting macrophage to disable efficient efferocytosis of apoptotic cells. Earlier findings demonstrated that pro-inflammatory factors like TNFα, activated ROS within macrophages to limit efferocytosis (27) and so it is logical to suggest that components of the SASP (e.g. TNFα, IL-6, IL-8, IL-1α to name a few) stimulate ROS to cleave MerTK. Collectively, these data suggest that limiting excessive ROS within the macrophage may be an ideal mechanism to thwart MerTK cleavage and promote efficient efferocytosis. Interestingly we observed that RvD1 prevented CM-induced MerTK cleavage and efferocytosis defects in a similar manner as NAC. Previous findings indicate that RvD1 limits NADPH oxidase (and ROS) in macrophages to promote efficient efferocytosis(28). Therefore, our findings are in agreement with the literature and suggest that RvD1’s ability to control excessive ROS may be an important mechanism for efficient engulfment of apoptotic cells.

Defective efferocytosis associated with aging was previously observed (6, 7, 23–25), however mechanisms associated with this defect remain underexplored. We found that lung macrophages from old mice were defective in their ability to ingest dead cells (Fig. 2B). This is consistent with the literature in which lung macrophages from aged mice have previously been shown to exert defective phagocytosis (29, 30). Also, we observed that IL-6 is increased in lungs from old mice and while we do not know the cellular source, our data suggests that the increase may be due to activated macrophages in lungs. Collectively our results suggest that macrophages in lungs from old mice are more pro-inflammatory and less phagocytic.

Moreover, a recent study by Linehan et al found peritoneal macrophages from old mice had worsened efferocytosis compared with peritoneal macrophages from young mice(24). Interestingly, they transferred peritoneal macrophages from young mice into the peritoneum of old mice, and found that the young macrophages exhibited defective efferocytosis similar to that of macrophages from old mice(24). Linehan and others suggested that the aging milieu drives defective efferocytosis(7, 24). Our work adds important mechanism to these findings because we suggest that the senescent cells and their released factors promotes MerTK cleavage and deranges efferocytosis. Another recent study by Frisch et al found that Gas6, which is a ligand for MerTK (and other TAM receptors), was down regulated in the bone marrow from aged mice(25). Gas6 is a critical bridging molecule for MerTK signaling that facilitates efferocytosis and the SPM:LT ratio(8, 10). Therefore, the combined defects in Gas6 expression and MerTK cleavage could be detrimental in aging.

MerTK cleavage promotes necrotic core formation in advanced atherosclerosis and cardiac tissue injury in myocardial infarction (9, 31). With regard to atherosclerosis, Cai B. et al found that MerTK signaling stimulated the synthesis of SPMs over LTs both in isolated murine and human macrophages as well as in murine plaques. Defective MerTK signaling may be associated with the SPM:LT imbalance driven by conditioned media from senescent cells. (9, 10, 32).

Along these lines, imbalances in the SPM:LT ratio have been observed in mice and humans in the context of inflammaging (6, 33). For example, urinary lipoxins (LXs) in humans was decreased in the elderly resulting in a profound imbalance between pro-resolving LXs and LTs (33). Other relevant age-related diseases like atherosclerosis, peripheral vascular disease, periodontal disease and Alzheimer’s disease are also associated with imbalances in the SPM:LT ratio and restoration of defective SPMs to these pre-clinical models of disease result in protection(17, 34–36).

SPMs show promise as a potential new treatment strategy for several age related diseases because they have dual anti-inflammatory and pro-resolving functions(3). This is an important distinction because anti-inflammation refers to the process of blocking pro-inflammatory processes, whereas pro-resolution involves a series of programs that initiate repair and regeneration(37) and are likely necessary for recovery and normal tissue function. Cyclooxygenase 2 (COX2) and lipoxygenases (LOX) are required for SPM biosynthesis. Inhibition of either of these enzymes delays endogenous resolution in mice(38). Moreover, selective COX2 inhibitors have increased risk for cardiovascular events in humans (39) and the risks of selective COX2 inhibitors are even higher in the elderly (40). Therefore, new therapies that can promote vascular homeostasis without disruption of inflammation-resolution is ideal. SPMs are not immunosuppressive and promote vascular homeostasis (3, 41) and we demonstrate here that RvD1 promotes efferocytosis in aging, limits inflammation and reduces tissue injury.

Overall, our results offer an entirely new mechanism whereby senescence promotes defective efferocytosis, and this can be therapeutically targeted by RvD1. SPMs are particularly intriguing because they temper inflammation, which is already heightened in aging, and at the same time, activate host repair in a manner that does not cause immune suppression. Therapies that are not immunosuppressive are particularly important for the aging population, wherein susceptibility to infection is already increased(42). Our work supports the idea that pro-resolving therapies may be a key new strategy to limit excessive inflammation and promote tissue repair in aging.

Non-standard Abbreviations:

SPM

Specialized pro-resolving mediators

RvD1

Resolvin D1

LX

lipoxins

LT

Leukotrienes

SASP

senescence associated secretory phenotype

I/R

ischemia reperfusion injury

MerTK

Mer proto-oncogene tyrosine kinase