Microvesicles carrying LRP5 induce macrophage polarization to an anti‐inflammatory phenotype

Abstract Microvesicles (MV) contribute to cell‐to‐cell communication through their transported proteins and nucleic acids. MV, released into the extracellular space, exert paracrine regulation by modulating cellular responses after interaction with near and far target cells. MV are released at high concentrations by activated inflammatory cells. Different subtypes of human macrophages have been characterized based on surface epitopes being CD16+ macrophages associated with anti‐inflammatory phenotypes. We have previously shown that low‐density lipoprotein receptor‐related protein 5 (LRP5), a member of the LDLR family that participates in lipid homeostasis, is expressed in macrophage CD16+ with repair and survival functions. The goal of our study was to characterize the cargo and tentative function of macrophage‐derived MV, whether LRP5 is delivered into MV and whether these MV are able to induce inflammatory cell differentiation to a specific CD16− or CD16+ phenotype. We show, for the first time, that lipid‐loaded macrophages release MV containing LRP5. LDL loading induces increased expression of macrophage pro‐inflammatory markers and increased release of MV containing pro‐inflammatory markers. Conditioning of fresh macrophages with MV released by Lrp5‐silenced macrophages induced the transcription of inflammatory genes and reduced the transcription of anti‐inflammatory genes. Thus, MV containing LRP5 induce anti‐inflammatory phenotypes in macrophages.

interaction of MV with target cells and the release of their content modulate cell responses. 17 MV have been described in inflammatory processes and associated with several cardiovascular risk factors 18 contributing to the initiation and progression of cardiovascular diseases, including atherosclerosis.
Atherosclerosis is characterized by chronic inflammation induced by increasing accumulation of low-density lipoproteins (LDL) and apoptotic cells in the intima layer of the arteries. 19 The low-density lipoprotein receptor-related protein 5 (LRP5) is a multifunctional receptor involved in both endocytosis of lipids and the canonical Wnt signalling pathway. 20 LRP5 is a single-pass transmembrane receptor that participates in the Wnt/β-catenin signalling pathway. LRP5 activation causes the stabilization of β-catenin that translocates into the nucleus, binds to the transcription factor TCF/LEF1 and starts the transcription of Wnt target genes that regulate fundamental aspects of embryonic cell development 21 and adult cell function. 20,22,23 LDL loading induces high LRP5 expression in human macrophages. 20 Macrophages can be classified into classical activated CD14 + CD16 − , pro-inflammatory macrophages and alternatively activated CD14 − CD16 + , anti-inflammatory macrophages. 24,25 LRP5 participates in inflammation and macrophage polarization by association with the anti-inflammatory macrophage subtype CD16 + derived from CD14 + CD16 + patrolling circulating monocytes. 26 LRP5 confers the motile function to CD16 + macrophages by triggering the canonical Wnt signalling. Furthermore, CD16 + LRP5 + macrophages, found in advanced atherosclerotic human plaques, trigger an antiinflammatory, defensive and repair response. 26 The in-depth understanding of the formation, cargo and function of MV is an ongoing task in the field. The objectives of this study were (a) to characterize the cargo and function of macrophage-derived MV and their ability to induce inflammatory cell differentiation to a CD16 − or a CD16 + phenotype, and (b) to investigate whether LRP5 is delivered into MV and whether it can exert paracrine functions.
We show that LDL-loaded macrophages release MV carrying LRP5 and exert paracrine and/or autocrine regulation. LDL loading induces increased expression of macrophage cellular proinflammatory markers and increased release of MV. Interestingly, LRP5 is released in MV that contain both pro-and anti-inflammatory markers. Conditioning of recipient macrophages with MV released by Lrp5-silenced macrophages induced pro-inflammatory gene transcription and a reduced expression of anti-inflammatory genes indicating that LRP5 induces macrophage differentiation into the anti-inflammatory phenotype.

| Isolation of human monocytes and human macrophages primary cultures and LDL loading
Human monocytes were obtained by standard protocols from buffy coats of healthy blood donors. 20,[26][27][28] All procedures were approved by the Institutional Review and Ethics Committee, and the investigation conformed to the principles outlined in the Declaration of Helsinki with informed consent given by donors. Briefly, blood was applied on 15 mL of Ficoll-Hypaque and centrifuged at 300 g for 1 hour at 22°C, with no brake. Mononuclear cells were obtained from the central white band of the gradient, exhaustively washed in Dulbecco's phosphate buffer saline, and suspended in RPMI medium (Gibco) supplemented with 10% human serum AB (Sigma). Isolated monocytes (Mo) were left overnight in culture, washed and treated with 100 μg/mL nLDL (native LDL) or agLDL (aggregated LDL) for the described times. A second set of isolated Mo were left 7 days in culture and allowed to differentiate into macrophages (Mac) by changing the cell culture media (RPMI supplemented with 10% human serum AB, 100 units/mL penicillin and 100 µg/mL streptomycin) every 3 days. After several washings with PBS to completely remove serum, human macrophages were incubated with 100 μg/mL nLDL or 100 µg/mL agLDL in serum-free medium. 20,[26][27][28] At the end of the experiments, human Mo and Mac were exhaustively washed (twice with PBS, twice with PBS/1% BSA, once with PBS/1%BSA/heparin 100 U/mL, twice with PBS/1% BSA and twice with PBS) and prepared for the collection of mRNA and protein detection as described below.

| LDL isolation and modification
Human LDL (d1.019-d1.063 g/mL) were obtained as previously described. 28 Briefly, human LDLs were obtained from pooled sera of normocholesterolemic volunteers and isolated by sequential ultracentrifugation. LDLs were dialyzed three times against 200 volumes of 150 mmol/L NaCl, 1 mmol/L EDTA, and 20 mmol/L Tris-HCl, pH 7.4, overnight and once against 150 mmol/L NaCl. LDL protein concentration was determined by the bicinchoninic acid, and vortexing was monitored by measuring the turbidity (absorbance at 680 nm).
The model system of agLDL was generated by vortexing LDL (1 mg/ mL) for 4 minutes at room temperature at maximal speed. The percentage of LDL in aggregated form was calculated by measuring the fraction of protein recovered in the pellet obtained after centrifugation at 10 000 g for 10 minutes. The different fractions were analysed by agarose electrophoresis, and the precipitated fraction composed of 100% agLDL was added to cell cultures.

| MV isolation and quantification
LDL-loaded or non-loaded human Mo and Mac were cultured for 24 or 48 hours and the MV released into the supernatants collected.
MV were isolated by five-step high-speed centrifugations. Briefly, 2 mL of fresh supernatant aliquots were centrifuged at 3200 g for 20 minutes to guarantee complete cell and debris removal. The recovered supernatants were centrifuged at room temperature at 300 g, (10 minutes); at 1200 g, (20 minutes); and at 12 500 g (5 minutes) in two repeated processes to ensure the elimination of nLDL or agLDL. The cleared supernatants were transferred to another vial and centrifuged at 20 500 g for 150 minutes at RT to pellet the MV.  Table S1 shows the different antibodies and the concentrations used for microvesicle identification and characterization. Samples were incubated 20 minutes at room temperature in the dark and diluted with annexin binding buffer before being immediately analysed and counted on a FACSCanto II flow cytometer. The number of monocytes or macrophage per well were counted using Neubauer chambers, and the number of MV/cell type was obtained.
AV binding level was corrected for autofluorescence using fluorescence signals obtained with MV in a calcium-free buffer PBS.
MV were identified and quantified based on their forward scatter/ side scatter characteristics according to their size, binding or not to AV and reactivity to specific monoclonal antibodies. Figure Table S2. Figure S2 shows the gating strategy for live macrophages by flow cytometry analysis. Samples were diluted with 400 μL flow cytometry buffer prior to being immediately analysed. For each sample, at least 10 000 events were acquired on a FACSCantoII (Beckton Dickinson). Data was analysed with the FACSDiva 6.1.3 software.

| Macrophages isolation by cell sorter
Lipid loaded macrophages were gently detached from culture dishes and stained with CD11b, CD14 and CD206 or CD80 antibodies (Table S2)  Cells were sorted using a FACSAria-I (BD Biosciences) operated using a 100 µm nozzle with the 488 nm and 633 nm laser lines.

| Supernatant collection
Cell sorted macrophages were cultured in serum-free RPMI GlutaMax medium for 2 days when supernatants were collected and centrifuged at 15 000 g, 15 minutes, 4°C. Pellets were discharged and supernatants were precipitated using a methanol/chloroform protocol. Briefly, one volume of supernatant was mixed with three volumes of cold methanol and one volume of chloroform, vortexed vigorously for 30 seconds, and then, three volumes of H 2 O were added to the sample to induce phase separation. The mix was centrifuged at 10 000 g for 5 minutes, and the upper phase was eliminated without disturbing the interphase. Three volumes of methanol were added to the mix, and samples were centrifuged at 10 000 g for 5 minutes. Supernatants were discharged and the precipitated proteins (pellet) were let to air-dry. Finally, samples were suspended in 100 µL of lysis protein buffer solution and frozen at −20°C until western blots were performed.

| Western blot
Protein extracts (50 µL) were resolved by SDS-PAGE and transferred to nitrocellulose membranes, blocked with 5% bovine serum albumin and probed for monoclonal primary antibodies against IL-1β, TNFα and TGFβ from Cell Signalling. Membranes were then washed and blotted with antimouse secondary antibodies (Dako). Band densities were determined with the ChemiDoc XRS system (Bio-Rad) in chemiluminescence detection modus and Quantity-One software (Bio-Rad).

| RNA isolation and Real time PCR
Total RNA was isolated from cultured human monocytes and macrophages using the total RNA extraction kit (Qiagen). Total RNA concentration was determined by NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Inc), and purity was checked by the A260/A280 ratio (ratios between 1.8 and 2.1 were considered acceptable), in addition, an agarose gel was run to assess quality. cDNA was synthesized from 1 μg RNA with cDNA reverse transcription kit (Qiagen) The resulting cDNA samples were amplified by polymerase chain reaction (PCR) using a DNA thermal cycler (MJ Research) and the following specific human probes from Applied Biotechnologies: LRP5, iNOS, CD80, CD163 and IL1Ra. Normalization was performed against r18S.

| Statistical analysis
A StatView statistical package was used for all the analysis. Results are expressed as mean ± SD or n (%) when indicated. When possible, comparisons among groups were performed by parametric (one factor ANOVA) analysis. Statistical significance was considered when P < .05. All the experiments were performed at least three times.

| Lipid loading increases LRP5 + MV secretion
We have previously shown that lipid loading with modified lipoproteins (agLDL) increases LRP5 expression in macrophages. 20,26 We hypothesized that the LDL loading would increase the generation of MV carrying LRP5. Primary cultures of human monocytes and macrophages were treated with 100 μg/mL nLDL or agLDL for 24 hours or 48 hours and, indeed, agLDL loading induced a massive generation of AV + MV from monocytes while a modest amount of AV + MV were released by macrophages ( Figure 1D). LDL loading induced the release of AV + LRP5 + MV in larger quantities in monocytes than in macrophages ( Figure 1E). However, the relative release of AV + LRP5 + MV (normalized by total AV + MV) was significantly induced by LDL loading in macrophages both after 24 hours and 48 hours incubation ( Figure 1F).
We then estimated the amount of MV produced by each monocyte or macrophage (MV/Mo and MV/Mac). Lipid-loaded Mo release more AV + MV than Mac after 24 hours (175 ± 21 AV + MV/Mo vs 6 ± 0.8 AV + MV/Mac, Figure S3A) and 48 hours agLDL incubation (108 ± 15 AV + MV/Mo vs 6 ± 0.4 AV + MV/Mac, Figure S3A). AgLDL treatments induced more LRP5 + MV release in individual monocytes than in macrophages both after 24 hours and 48 hours incubation ( Figure S3B). Finally, the relative amount of AV + LRP5 + MV/cell type (normalized by AV + MV/cell type) released by macrophages was higher than that released by monocytes after 24 hours and 48 hours agLDL incubation ( Figure S3C).

| agLDL loading induces macrophage polarization
We previously observed that LRP5 is mainly expressed in CD16 + macrophages and lipid loading induces LRP5 expression in these cells 20,26 ; therefore we investigated whether macrophage polariza-  Macrophage pro-inflammatory phenotype after lipid loading was confirmed by pro and anti-inflammatory protein secretion analyses. Cell sorting was performed on lipid loaded macrophages to obtain CD11b + CD14 + CD206 − and CD11b + CD14 + CD206 + or CD11b + CD14 + CD80 − and CD11b + CD14 + CD80 + macrophage subpopulations. The different macrophage subpopulations were seeded in culture dishes and supernatants were collected after 48 hours. Increased release of the pro-inflammatory proteins TNFα and IL1β was observed in the pro-inflammatory CD80 + subpopulation while the levels remained low in the anti-inflammatory CD206 + subpopulation ( Figure 2D). Conversely, the release of the anti-inflammatory protein TGFβ was higher in the CD206 + macrophage subpopulation than in the CD80 + macrophage subpopulation ( Figure 2D). Therefore, inflammatory protein release confirms the pro-inflammatory polarized phenotype in macrophages observed by cell surface markers expression after lipid loading.

| Lipid loading induces LRP5 expression in macrophages
We next examined the expression levels of LRP5 in the different macrophage subpopulations by staining macrophages with a F I G U R E 1 AgLDL treatments in macrophages induce LRP5 + MV secretion. 24 hours or 48 hours supernatants from undifferentiated monocytes (Mo) or from 7 to 10 days fully differentiated macrophages (Mac) were collected and the amount of (A) microvesicles/mL; (B) Annexin V + microvesicles/mL and (C) Annexin V + LRP5 + microvesicles/mL were analysed. (D) Monocytes (Mo) or macrophages (Mac) were treated with 100 μg/mL nLDL or 100 μg/mL agLDL for 24 hours or 48 hours and the amount of AV + MV/mL and of (E) AV + LRP5 + MV/mL was analysed. (F) The ratio between MV that are AV + LRP5 + /AV + in control and lipid-loaded monocytes (Mo) and macrophages (Mac). All experiments were performed at least four times in duplicates or triplicates. *P < .05, **P < .01, ***P < .005  Figure 2E).

| LRP5 + MV contain pro-inflammatory and antiinflammatory proteins
MV released from lipid-loaded and non-loaded macrophages were isolated and stained for pro-inflammatory and anti-inflammatory markers and for LRP5. Interestingly, LRP5 was delivered into MV containing both pro-inflammatory and anti-inflammatory proteins, indicating that the delivery of LRP5 into MV is independent of the inflammatory proteins delivered into the MV ( Figure 3E).

| Characterization of donor macrophages and their released MV
Macrophage specific inhibition of LRP5 expression (with siRNA) was used to identify whether LRP5 was playing a role in macrophage differentiation towards a CD16 − or a CD16 + phenotype.
Macrophages were silenced or not for LRP5 and agLDL-loaded or not ( Figure 4A). Analysis of donor macrophages mRNA expression by RT-PCR showed a 92 ± 3% LRP5 reduction in siRNA-LRP5 control cells and a 90 ± 2% LRP5 reduction in siRNA-LRP5 lipidloaded macrophages. LRP5 mRNA expression was increased in lipid-loaded macrophages ( Figure 4B). LRP5 silencing did not modify the number of AV + MV/mL released by macrophages neither in control nor in lipid-loaded conditions ( Figure 4C). However, a consistent reduction in LRP5 + AV + MV release by siRNA-LRP5-treated macrophages was observed both in untreated and agLDL-loaded macrophages ( Figure 4D).

Microvesicles can stimulate targets cells by direct interaction with
target receptors and the transfer of the bioactive molecules they contain. 2,3,7,8,10,13,15 Here, we show, for the first time, that LRP5 is delivered into MV released by macrophages and monocytes.
However, the release of LRP5 + MV is only significantly increased in fully differentiated lipid-loaded macrophages.
In general, macrophages are classified into two main phenotypes, classical M1 CD16 − activated macrophages and alternative M2 CD16 + activated macrophages, which regulate pro-inflammatory and anti-inflammatory responses, respectively. 29 Regulation of lipid-induced macrophage polarization is a very new field of investigation. A recent study showed that satu-  38,39 For example, cytokine IL1β induces MV shedding from circulating monocytes. 40 Accordingly, here we show that lipid stimuli induce pro-inflammatory MV release. As lipid loaded macrophages show increased expression of LRP5 at the cell surface, it is plausible that this LRP5 will be delivered to their MV and released as LRP5 + MV. Human macrophages expressing LRP5 have been shown to provide survival and repair to damaged tissues. 26 It is our hypothesis that this is the function of LRP5 + MV but further work needs to be performed to prove it.
We also explored the function of macrophage-derived LRP5 + MV in the polarization fate of macrophages. Lipid-loaded macrophages that did not express LRP5 showed similar MV release than lipidloaded LRP5 + macrophages indicating that LRP5 does not participate in the MV release pathway. A reduction in LRP5 + MV release from macrophages without LRP5 was observed.
Classically activated CD16 − macrophages are characterized by the expression of several pro-inflammatory markers, including iNOS and CD80 41,42 while alternatively activated anti-inflammatory CD16 + macrophages express CD163 and IL1Ra. 42,43 Treatment with MV released by macrophages devoid of LRP5 induced iNOS and CD80 expression and reduced CD163 and IL1Ra expression in naive macrophages indicating that LRP5 + MV induce macrophages to differentiate towards an anti-inflammatory phenotype. A limitation of this study is that the size of LRP5 + MV was not assessed; therefore, we could not determine if LRP5 + MV have a different size than MV devoid of LRP5. However, always the same procedure was followed to prepare MV and only MV released by control macrophages were unable to induce high expression of pro-inflammatory genes. Also, MV released by lipid-loaded macrophages induced increased expression of pro-inflammatory genes independent of LRP5 expression further supporting that different stimulus in the cells of origin generates MV with different cargoes that will have different functions in the target cells.
In conclusion, here we demonstrate for the first time that a lipoprotein receptor, LRP5, is delivered into MV. MV released by lipid-loaded macrophages contain mainly pro-inflammatory proteins and LRP5. LRP5 + MV induce an anti-inflammatory genotype in naive macrophages. Therefore, a systematic blockade of monocyte/ macrophage infiltration in the prevention of atherosclerosis may be less effective than originally expected if the levels of macrophagederived LRP5 + MV are affected and reduced. Jesus Serra for their continuous support.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest. F I G U R E 5 Gene expression levels in receptor macrophages and monocytes. Receptor macrophages were treated with macrophagederived MV released by Control macrophages, siRNA-LRP5-treated macrophages, 100 μg/mL agLDL-treated macrophages or siRNA-LRP5+AgLDL-treated macrophages and mRNA expression levels of (A) LRP5, (B) iNOS, (C) CD80, (D) CD163 and (E) IL1Ra were analysed. Receptor monocytes were treated with macrophage-derived MV released by Control macrophages, siRNA-LRP5-treated macrophages, 100 μg/mL agLDL-treated macrophages or siRNA-LRP5+AgLDL-treated macrophages and mRNA expression levels of (F) LRP5, (G) iNOS, (H) CD80, (I) CD163 and (J) IL1Ra were analysed. Experiments were performed four times in triplicates. *P < .05, ***P < .005

DATA AVA I L A B I L I T Y S TAT E M E N T
The data underlying this article are available in the article and in its online supplementary material.