Modulation of cardiac macrophages byphosphatidylserine-presenting liposomesimproves infarct repairTamar Harel-Adara, Tamar Ben Mordechaib, Yoram Amsalemb, Micha S. Feinbergb, Jonathan Leorb, and Smadar Cohena,1

aAvram and Stella Goldstein-Goren Department of Biotechnology Engineering, Ben-Gurion University of the Negev, Beer-Sheva, 84105, Israel; andbNeufeld Cardiac Research Institute, Tel Aviv University, Sheba Medical Center, Tel-Hashomer, 52621, Israel

Edited* by Robert Langer, Massachusetts Institute of Technology, Cambridge, MA, and approved December 17, 2010 (received for review October 20, 2010)

Herein we investigated a new strategy for the modulation of car-diac macrophages to a reparative state, at a predetermined timeafter myocardial infarction (MI), in aim to promote resolution ofinflammation and elicit infarct repair. The strategy employed intra-venous injections of phosphatidylserine (PS)-presenting liposomes,mimicking the anti-inflammatory effects of apoptotic cells. Follow-ing PS-liposome uptake by macrophages in vitro and in vivo, thecells secreted high levels of anti-inflammatory cytokines [trans-forming growth factor β (TGFβ) and interleukin 10 (IL-10)] andupregulated the expression of the mannose receptor—CD206, con-comitant with downregulation of proinflammatory markers, suchas tumor necrosis factor α (TNFα) and the surface marker CD86. In arat model of acute MI, targeting of PS-presenting liposomes to in-farct macrophages after injection via the femoral vein was demon-strated by magnetic resonance imaging (MRI). The treatmentpromoted angiogenesis, the preservation of small scars, and pre-vented ventricular dilatation and remodeling. This strategy repre-sents a unique and accessible approach for myocardial infarctrepair.

An important goal in cardiology is to minimize infarct size andimprove healing after myocardial infarction (MI). Following

MI, resident and recruited macrophages remove necrotic andapoptotic cells, secrete cytokines, and modulate angiogenesisat the infarct site (1). As such, the macrophage is a primary re-sponder cell involved in the regulation of post-MI infarct woundhealing at multiple levels (2). According to recent studies, differ-ent subsets of macrophages are responsible for these differentactivities; during the early inflammatory phase (phase 1), proin-flammatory macrophages dominate the injury site and phagocy-tose apoptotic/necrotic myocytes and other debris, whereasduring inflammation resolution (phase 2), the dominant subsetsare the reparative macrophages, which propagate infarct repair(3, 4). The duration and extent of the early inflammatory phasehave major implications on infarct size and ventricle remodel-ing (5).

Herein, we conceived a previously undescribed strategy forcontrolling the duration and extent of the inflammatory phasefollowing MI, in aim to reduce myocardium damage, preserve in-farct size, and prevent ventricle remodeling. Our strategy exploitsthe principle underlying the anti-inflammatory effects of apopto-tic cells, which are known to actively suppress inflammation byinhibiting the release of proinflammatory cytokines from macro-phages while augmenting the secretion of anti-inflammatorycytokines, such as transforming growth factor β (TGFβ) andInterleukin 10 (IL-10) as well as the expression of the mannosereceptor, CD206 (6, 7). Macrophages recognize the apoptoticcells via surface ligands, among them phosphatidylserine (PS),and “silently” clear the cells (8–10). In humans, apoptosis afterMI occurs mainly in the border zones and in the remote areasof ischemia (11), thus its effects on inflammation resolution areconsidered to be minor.

Our strategy is based on exogenous administration of PS-pre-senting liposomes, designed to mimic the apoptotic cells in terms

of PS presentation on their surface and their anti-inflammatoryeffects. We examined liposome uptake by peritoneal and cardiacmacrophages, in vitro and in vivo, and verified the consequentupregulation in anti-inflammatory responses, by measuring theprofile of cytokine secretion and surface marker expressions.The effects of i.v. injections of PS-presenting liposomes on angio-genesis, infarct size, and ventricle remodeling were examined in arat model of acute MI.

ResultsPhosphatidylserine (PS)-Presenting Liposomes. The liposomes wereconstructed to present the “death signal” PS on their surfaceavailable for ligation by the PS receptor (PSR) on macrophages.The presence of PS on liposome surface was validated by Fluor-escence Activated Cell Sorter (FACS) analysis. Fluoresceinisothiocyanate (FITC)-annexin V was bound to 98% of thePS-presenting liposomes (Fig. 1A), while staining was minor inPS-lacking liposomes (11.21%, probably nonspecific staining)(Fig. 1B). Additionally, zeta potential measurements revealeda net surface charge of (−98.6� 11.3 mV) on PS-presenting lipo-somes, while that on the PS-lacking liposomes was less negative(−21.43� 0.46 mV), due to the greater fraction of phosphatidyl-choline (PC) on the surface of the latter liposomes. In PS-lackingliposomes, the PC fraction constituted 74% by mass, while inPS-presenting liposomes it was 36.5%. Thus, the PS-presentingliposomes dispersed uniformly in physiological medium due torepulsion forces (Fig. 1C), while the PS-lacking liposomes formedaggregates (Fig. 1D). By cryotransmission electron microscope(cryo-TEM) analysis (Fig. 1 E and F), most liposomes had fewlamellas. No differences in outer shape or in size of liposomesconstructed with or without PS were observed. The liposomeshad a particle size of 1.2� 0.3 μm for efficient uptake by macro-phages (Fig. S1) (12).

Uptake of PS-Presenting Liposomes by Peritoneal Macrophages andImmunomodulation.The in vitro uptake mechanism of PS-present-ing liposomes was initially studied in cultures of peritonealmacrophages because they represent an accessible cell type com-pared to cardiac macrophages; their isolation protocol is simplerand reproducible (see SI Methods). The peritoneal macrophageswere treated with cytochalasin B (CY), which inhibits particlephagocytosis but not the PS-receptor-mediated endocytosis(13), prior to administration of FITC-bovine serum albumin(BSA) encapsulating liposomes. The treatment with CY had no

Author contributions: T.H.-A. and S.C. designed research; T.H.-A., T.B.M., and Y.A.performed research; T.H.-A., M.S.F., J.L., and S.C. analyzed data; and T.H.-A. and S.C. wrotethe paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.1To whom correspondence should be addressed. E-mail: scohen@bgu.ac.il.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1015623108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1015623108 PNAS ∣ February 1, 2011 ∣ vol. 108 ∣ no. 5 ∣ 1827–1832

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

May

20,

202

0

effect on the uptake of PS-presenting liposomes by macrophages,while it reduced by a factor of almost 20 the uptake of PS-lackingliposomes (Fig. 2A), indicating that the uptake of PS-presentingliposomes occurs via the PS-receptor-mediated endocytosispathway.

The in vivo uptake of PS-presenting liposomes encapsulatingFITC-(BSA) 3 h after i.p. injection into mice peritoneum wasconfirmed and quantified by FACS analysis of the lavaged macro-phages. The macrophages (identified by their CD11b marker)uptook threefold more of the PS-presenting liposomes thanPS-lacking ones (Fig. 2B). The difference in liposome uptakeis attributed to the specific engulfment of the PS-presenting lipo-somes through the PS receptor on macrophages.

Next, we examined whether the uptake of PS-presenting lipo-somes by macrophages can mimic the modulating effect of apop-totic cell uptake and elicit anti-inflammatory responses inactivated macrophages. For that, a proinflammatory responsewas induced in macrophages by injecting a strong inducer ofproinflammatory responses—lipopolysaccharide (LPS), into theperitoneum of mice. Fig. 2 C and D shows that following the invivo uptake of PS-presenting liposomes, the state of the lavagedmacrophages changed from proinflammatory (treated with LPS)to anti-inflammatory, as judged by the significant decrease inCD86 expression—a surface marker for proinflammatory macro-phages (Fig. 2C, p < 0.05) and the increase in CD206 mannosereceptor expression—a surface marker for reparative macro-phages (Fig. 2D, p < 0.05) (14, 15). By contrast, the uptake ofPS-lacking liposomes did not affect the macrophage state, andLPS-treated macrophages maintained their proinflammatoryphenotype induced by the LPS.

The results were further supported by the ELISAs performedon the peritoneal lavage fluids for determining cytokine secretion

levels. Fig. 2 E and F shows significantly greater secretion levelsof the anti-inflammatory cytokines, TGFβ and IL-10, from LPS-activated macrophages following the uptake of PS-presentingliposomes, and lower levels after the uptake of PS-lacking lipo-somes, compared with the nontreated mice. These findings vali-date that PS-presenting liposomes can mimic apoptotic cells andtherefore promote a “silent” anti-inflammatory clearance.

Effect of PS-Presenting Liposomes on Macrophage Population at theInfarct. Because the treatment time with PS-presenting liposomesafter MI is an important parameter in the proposed strategy, weperformed an analysis to identify and quantify the macrophagenumbers and their subpopulation ratio, proinflammatory vs.reparative, at the infarct at different times after MI inductionin mice. The percentage of cardiac macrophages (F4∕80 positive

Fig. 1. Liposome features. (A and B) FACS analysis of PS-presenting (A) andPS-lacking (B) liposomes; the numbers on pictures represent the percentageliposomes with bound Annexin V on their surface. (C and D) Light microscopeimages of PS-presenting (C) and PS-lacking (D) liposomes. Bar ¼ 32 μm. (E andF) Cryo-TEM pictures of PS-presenting (E) and PS-lacking (F) liposomes.Bar ¼ 200 nm.

Fig. 2. In vitro/in vivo uptake of PS-presenting liposomes by peritonealmacrophages (F4∕80 positive) (A and B) and immunomodulation (C–F).(A) In vitro uptake of (FITC-BSA)-liposomes by adhered macrophages, pre-treated for 2 h with 10 μM of cytochalasin B (CY) prior to incubation withliposomes for 1 h. After uptake, the cells were detached and analyzed byFACS for F4∕80þ cells that have uptaken fluorescent (FITC) liposomes (n ¼ 3

wells per each group). (B) In vivo macrophage uptake of PS-presenting lipo-somes (n ¼ 3) and PS-lacking liposomes (n ¼ 3) containing FITC-BSA; the lipo-somes were injected i.p. to mice and 3 h later the peritoneal cells werelavaged and analyzed by FACS. (C–F) Immunomodulation of macrophagesby PS-presenting liposomes. A proinflammatory response was induced inperitoneal macrophages by i.p. injection of lipopolysacharide (LPS, 250 μg)into mice. Three h later, the animals were randomly divided into threegroups: those injected with only LPS (LPS group, n ¼ 3); LPS followed byi.p. injection of PS-presenting liposomes (LPS+PS group, n ¼ 3); or LPS fol-lowed by i.p. injection of PS-lacking liposomes (LPS+PC, n ¼ 3). An additionalgroup consisted of nontreated (NT, n ¼ 3) animals (with no LPS activation).Three h after treatment with liposomes, peritoneum lavage was analyzed byFACS (cells) and ELISA (lavage fluid). (C and D) FACS analyses for the relativeexpression of CD86 (C) and CD206 (D) out of the F4∕80 positive macrophages.(E and F) ELISA of lavage fluids for secretion levels of (E) IL-10 and (F) TGFβ.

1828 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1015623108 Harel-Adar et al.

Dow

nloa

ded

by g

uest

on

May

20,

202

0

cells out of total cardiac cells, following enzymatic digestion ofthe heart) showed a trend of increasing level starting 2 d afterMI (Fig. S2A). Analysis of the proinflammatory macrophage sub-set (CD86-positive out of F4∕80 positive cells) and reparativemacrophage subset (CD206 out of F4∕80 positive cells) revealeda trend of change in the initial ratio of reparative/proinflamma-tory subpopulations (0.85) in favor of the reparative macrophagesubset (1.39) 4 d after MI induction (Fig. S2B).

An intramyocardial injection of PS-presenting liposomes,PS-lacking liposomes, or saline immediately after MI had nosignificant effect on the total macrophage number at the infarct(Fig. S2A). However, the macrophage subset ratio at the infarctchanged after treatment with PS-presenting liposomes as re-flected by the anti-inflammatory cytokine secretion profiles ofIL-10 (Fig. 3A) and TGFβ (Fig. 3B); there was an increase in theirsecretion level from macrophages isolated as early as 3 d after MIand treatment with PS-presenting liposomes compared to theirlevels in animals treated with PS-lacking liposomes or saline.The same trend was seen in cardiacmacrophages isolated 4 d afterMI with the respective treatments (Fig. S2C and D). Concomi-tantly with the increase in anti-inflammatory responses, a trendof decrease in the secretion level of the proinflammatory cytokineTNFα was noticed 3 d after MI, in macrophages isolated fromPS-presenting liposomes treated mice compared with PS-lackingliposomes or saline-treated mice (Fig. 3C). Taken together, theseresults indicate that the treatment with PS-presenting liposomescan induce cardiac macrophages to secrete anti-inflammatorycytokines as early as 3 d after treatment, 1 d earlier than withouttreatment.

Tracking Liposomes in Infarct after i.v. Administration. The ability ofPS-presenting liposomes to target the infarct after i.v. injectionand be engulfed by cardiac macrophages was investigated withliposomes containing iron-oxide (see SI Methods). The liposomeswere injected through the femoral vein, 48 h after MI inductionin rats. MRI scans of the hearts 4 d later (Fig. 4A) reveal the pre-

sence of resident and/or infiltrating macrophages at the infarctthat had taken up PS-presenting liposomes and accumulatedin the infarct (detected by the dark dots–arrows).

To validate the above, cross-sections from the treated heartswere immuno-stained for ED-1 (macrophage marker, brownstain) and for iron (blue stain). The overlapping picture (rightpanel) reveals colocalization of the two stains, indicating target-ing and uptake of liposomes by macrophages at the infarct(Fig. 4B). By contrast, in the control group, sporadic weak stain-ing for endogenous iron (red blood cells) and macrophages wasseen, but no colocalization of the stains.

Effect of i.v. Administration of PS-Presenting Liposomes on CardiacFunction and Structure. The finding that PS-presenting liposomescan be targeted to the infarct after i.v. administration and be ta-ken up by cardiac macrophages prompted us to test the efficacy ofthis strategy to improve infarct repair after MI. The treatmentwas performed 48 h after MI induction because at this time pointthe macrophages are found in greater numbers at the infarct thanimmediately after MI. The rats were randomly subjected to aninjection into the femoral vein of either PS-presenting liposomes(n ¼ 10), PS-lacking liposomes (n ¼ 10), or saline (n ¼ 10).Echocardiography studies were performed 1 d after MI, priorto treatment, to validate infarction and 1 mo later, to examinethe effect of treatment on infarct repair. In addition, the rathearts were examined by histochemistry for angiogenesis andby postmortem morphometric analysis for dimensions of LV re-modeling.

Treatment with PS-presenting liposomes by i.v. administrationenhanced angiogenesis at the infarct compared to infarcts treatedwith PS-lacking liposomes or saline. This is revealed when com-paring the vessel density at the infarct 4 weeks after MI, evaluated

Fig. 3. Immunomodulation of cardiac macrophages after MI. MI wasinduced in mice, followed by intramyocardial (into the infarct) injectionsof either PS-presenting liposomes (PS lip) (n ¼ 15), PS-lacking liposomes(PC lip) (n ¼ 15), saline (n ¼ 25), or no treatment (NT) (n ¼ 10). Three dayslater, macrophages were isolated from the infarcted hearts, cultured for24 h and the collected culture medium was analyzed by the respective ELISAsusing antibodies against IL-10 (A), TGFβ (B), and TNFα (C). * denotes statisti-cally significant difference.

Fig. 4. In vivo uptake and accumulation of PS-presenting liposomes incardiac macrophages after i.v. administration to rats. MI was induced in ratsand 48 h later, PS-presenting liposomes entrapping iron-oxide or saline wereinjected through the femoral vein. Four days later, the rats were examined byMRI (A) to evaluate macrophage accumulation at the infarct. The dark areasin the coronal sections represent macrophages, which have uptaken PS-presenting liposomes containing the iron oxide. (B) Histology/immuno-histo-chemistry of cross-sections from hearts excised from mice treated with i.v.injections of PS-presenting liposomes containing iron oxide, 48 h after MI.Four days after liposome injection, the animals were sacrificed and heartswere fixated and sliced for histology and immuno-staining for ED1 (a markerfor resident macrophages, brown color) and iron oxide (blue color).Bar ¼ 500 μm.

Harel-Adar et al. PNAS ∣ February 1, 2011 ∣ vol. 108 ∣ no. 5 ∣ 1829

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

May

20,

202

0

by immuno-staining of both lectin (Fig. S3A) marking the en-dothelial cells in blood vessels and α-smooth muscle actin (SMA,Fig. S3B), marking the pericytes and arterioles that surroundmatured blood vessels. Quantitatively, it is observed that themature blood vessel density is 1.5-fold greater in rats treated withPS-presenting liposomes compared to animals treated with PS-lacking liposomes or saline (Fig. 5A). It is also seen that PS-lack-ing liposomes did have some effect on angiogenesis compared tosaline; however, it was not as distinct as PS-presenting liposomes.

Morphometric analyses on Masson’s tri-chrome-stained cross-sections (Fig. S4) reveal the presence of thinner scars in ratstreated with saline or PS-lacking liposomes compared to thosetreated with PS-presenting liposomes. The expansion indexand scar thickness (Fig. 5B) showed significant improvementsafter treatment with PS-presenting liposomes compared to salinetreatment.

Left ventricle end systolic and diastolic areas (LVES area andLVED area) (Fig. 5C) were expanded in the groups treated withPS-lacking liposomes or saline compared to the group treatedwith PS-presenting liposomes. In the latter group, the left ventri-cle wall maintained its area size, indicating that the treatment

with PS-presenting liposomes is capable of preventing the LVremodeling associated with MI.

DiscussionHerein, we investigated a previously undescribed strategy formodulating macrophages to an anti-inflammatory state at a pre-determined time after MI, in aim to promote resolution of in-flammation and preserve small infarct size and LV dimensions.In conceiving the strategy, several criteria guided us: the useof natural, well-defined and safe materials; the treatment shouldbe minimally invasive; and the principles underlying the strategyto resolve inflammation should mimic the anti-inflammatoryeffects of apoptotic cell clearance by macrophages (16).

In humans, apoptosis after MI is not as significant as necroticcell death, which may explain why the inflammation responsesafter MI are frequently so excessive, causing massive tissue da-mage (17). Recognizing the anti-inflammatory effects of apoptoticcells, we hypothesized that the exogenous application of PS-presenting liposomes as apoptotic-mimicking particles wouldact to resolve the inflammation after MI while providing a safea-cellular, reproducible, and accessible approach. Thus, we de-signed liposomes carrying the apoptotic signal PS on their surface,sized 1-μm, suitable for i.v. administration as well as for uptakeby macrophages. An initial proof-of-concept study to prove thecapability of PS-presenting liposomes to modulate macrophageactivity was performed with peritoneal macrophages due to theirgreater accessibility. Peritoneal and cardiac macrophages share asimilar machinery of apoptotic cell recognition; they both respondto the “eat me” signal presented as phosphatidylserine on apop-totic cell surface, and as a result they secrete similar patterns ofcytokines. Thus, the studies with peritoneal macrophages mayrepresent faithfully the behavior of cardiac macrophages towardapoptotic cells. We confirmed that the uptake of PS-presentingliposomes by macrophages occurs via PS-receptor endocytosissince inhibition of the regular nonspecific phagocytosis by cyto-chalasin B did not affect their uptake.

The uptake of PS-presenting liposomes by LPS-activatedperitoneal macrophages upregulated the anti-inflammatoryphenotype, as reflected by the enhanced expression of CD206(Mannose receptor) and the greater secretion levels of theanti-inflammatory cytokines TGFβ and IL-10. Concomitantly,we noted downregulation in the expression of the proinflamma-tory marker CD86. The anti-inflammatory macrophage pheno-type after uptake of PS-presenting liposomes mimicked that ofactivated macrophages after the uptake of apoptotic cells (14,15). Furthermore, our studies show that the uptake of PS-present-ing liposomes was effective in reversing the strong proinflamma-tory effects of LPS on macrophages by upregulating theexpression of anti-inflammatory markers and cytokine secretion.These factors have been previously shown to contribute to theimmunosuppressive effects of apoptotic cells in vivo, as shownin IL-10-deficient mice (18).

Importantly, the PS-presenting liposomes were capable ofmodulating the macrophage activation state at the infarct site.An intramyocardial injection of PS-presenting liposomes afterMI induced the anti-inflammatory phenotype in macrophageson site already at day 3 after treatment, while with no treatmentor when treated with PS-lacking liposomes this event occurredlater after MI, from day 4 and onwards. The induced anti-inflam-matory phenotype was accompanied by elevated secretion levelsof IL-10 and TGFβ and decreased secretion of the proinflamma-tory cytokine TNFα. These results prompted us to test whetherthis strategy would be valuable for infarct therapy.

The choice to administer the liposomes for infarct therapy byi.v. injection was made because this represents a more clinicallyrelevant treatment modality. The liposomes were injected 2 dafter MI, at a time point where the number of macrophages atthe infarct begins to increase and the cells become influential.

Fig. 5. Effect of i.v. injected PS-presenting liposomes on angiogenesis and LVremodeling after MI. Rats were subjected to MI and 48 h later were injectedi.v. with either PS-presenting liposomes (n ¼ 10) (PS lip), PS-lacking liposomes(n ¼ 10) (PC lip), or saline (n ¼ 10). Four weeks later, the harvested heartswere sectioned and stained with antibodies against α-SMA and lectin for de-tecting mature blood vessels (Fig. S3A) and Masson’s trichrome to allow mor-phometric measurements of scar thickness and expansion index (Fig. S3B). (A)The vessel density (#∕mm2) was determined from five different fields in eachslide, randomly selected from the α-SMA and lectin-immuno-stained cross-sectioned slides, using Cell* software (Olympus Soft imaging SolutionsGmbH). (B) Expansion index and relative infarct size determined from mor-phometric analysis. (C) Results of echocardiography study for measured LVEDand LVES areas. Bonferroni post test was performed.

1830 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1015623108 Harel-Adar et al.

Dow

nloa

ded

by g

uest

on

May

20,

202

0

By MRI and immuno-histochemistry, we confirmed that the PS-presenting liposomes targeted the cardiac macrophages at the in-farcted heart after i.v. injection. The precise uptake mechanism isyet unclear; either the liposomes were targeted to the infarctwhere they were uptaken by resident/recruited macrophages,or while in circulation, the liposomes were uptaken by circulatingmacrophages, which subsequently migrated to the infarct site.

The i.v. administration of PS-presenting liposomes affected theextent of angiogenesis, infarct size, and LV remodeling after MI.Angiogenesis leading to the formation of mature blood vesselswas greater after treatment with PS-presenting liposomes com-pared to PS-lacking liposomes (1.5-fold higher vessel density)although it did not reach a statistical significance. The echocar-diography study revealed that LV end systolic and diastolic areas(LVES area and LVED area) were preserved in the group treatedwith PS-presenting liposomes compared to the controls, PS-lack-ing liposomes, and saline-treated groups. All these effects are incorroboration with the reparative activities of the macrophagesubpopulation in phase 2. These cells have been shown to releaseproangiogenic factors such as vascular endothelial growth factor(VEGF), participate in the fusion of endothelial tip cells, andserve as bridge cells (7). Their effect on infarct size preservationcan be attributed to the salvation of the remaining myocardialtissue by the enhanced angiogenesis as well as by downregulatingthe excessive inflammation and secretion of the anti-inflamma-tory cytokines, such as IL-10 and TGF-β.

Collectively, our results are promising, indicating that themodulation of recruited or resident cardiac macrophages by ap-plying PS-presenting liposomes is feasible and with consequencesleading to attenuation in left ventricle remodeling and preventionof heart dilatation. With respect to translation into the clinics, thestrategy of using autologous apoptotic cells to treat chronic heartfailure has been clinically tested (ACCLAIM trial) showing somebenefit for the treatment (19). Yet, there are concerns using auto-logous apoptotic cells because this treatment can ameliorateautoimmune diseases, for example via the release of autoanti-gens. The use of well-defined PS-presenting liposomes abrogatesthe concerns associated with apoptotic cell treatment, whilestill benefiting from their anti-inflammatory effect and beneficialeffects on infarct therapy.

MethodsPreparation and Characterization of Liposomes. The liposomes were preparedas described (20) from a lipid mixture (Avanti Polar Lipids) of phosphatidyl-serine (PS), phosphatidylcholine (PC), and cholesterol (CH) at 1∶1∶1.33 molarratios, respectively (PS-presenting liposomes), or from PC and CH (1.5∶1molar ratio) (PS-lacking liposomes). The degree of PS exposure on liposomeswas assessed by (i) binding of FITC-annexin V to PS on liposome surface andanalysis by FACS and (ii) measuring the zeta potential of liposome prepara-tions using Zeta Plus particle size analyzer (Brookhaven InstrumentsCorporation).

SI Text provides the details for preparation method of liposomes encap-sulating fluorescein isothiocyanate-labeled bovine serum albumin (FITC-BSA,Sigma Chemical Co.) or iron-oxide nanoparticle solution (Endorem), as well asdetails of liposome evaluation by cryotransmission electron microscope(Cryo-TEM).

Uptake Studies by Peritoneal Macrophages. The online data supplement pro-vides details of the methods for isolation and cultivation of macrophagesfrom the peritoneum of Balb/c mice, as well as the in vitro and in vivo uptakeexperiments.

Induction of Myocardial Infarction (MI). The animal studies were performed inaccordance with the Animal Care and Use Committee guidelines of Tel AvivUniversity, which conforms to the policies of the American Heart Association.MI induction was attained by permanently occluding the left main coronaryartery (LAD) with an intramural suture (6-0 polypropylene), as previouslydescribed (21).

Evaluation of Macrophage Population after MI and Treatment. Twenty g Balb/cfemale mice (Harlan Lab) underwent MI induction and were randomly

divided into four groups (15 each); three groups received an intramyocaridal(into the infarct) injection of a solution of either PS-presenting liposomes(0.03 M), PS-lacking liposomes (0.03 M) or saline (35-μL). The fourth groupremained untreated. At various time points after MI induction and treat-ment, hearts were harvested and underwent three 10-min cycles of enzy-matic digestion using a cocktail of Dispase II and Trypsin-EDTA (22). Theisolated cells were analyzed by FACS for the percentage of total macrophages(F4∕80 positive), the fraction of proinflammatory macrophages (F4∕80 posi-tive cells expressing the proinflammatory marker CD86) and the fraction ofreparative macrophages (F4∕80 positive cells expressing the mannose recep-tor CD206) (14, 15).

The isolated macrophages were cultured in 24-well tissue culture plates,500;000 cells∕well, in medium composed of RPMI 1640 medium, supplemen-ted with 2 mM L-Glutamine, 10% ðv∕vÞ fetal bovine serum (FBS) and100 U∕mL Penicillin, 1 μg∕mL Streptomycin, 2.5 U∕mL Nystatin, for 24 h(all materials from Biological Industries). The collected culture mediumwas analyzed for secretion of TNFα, IL-10 and TGFβ by their respective ELISAs(R&D Systems Inc.) according to manufacturer’s instructions.

Tracking Liposomes at Infarct After i.v. Administration. The uptake of i.v. in-jected PS-presenting liposomes by cardiac macrophages was investigatedin a rat model of acute MI (23). Forty-eight h after MI, the SD rats wereinjected through the femoral vein with 0.03 M of PS-presenting liposomescontaining iron oxide (2 μg∕mL) (n ¼ 4) or saline as a control (n ¼ 4). Fourdays after injection, the rats were examined by MRI using a 0.5-T GE iMRImachine with a specially constructed animal probe. Image sequences in-cluded T1 spin echo and T2* gradient echo. Following imaging, the rats weresacrificed and their hearts were harvested and examined by immunohisto-chemical staining for ED1 in macrophages (Serotec) and for iron (Sigma-Aldrich).

I.v. Injections of PS-Presenting Liposomes and Echocardiography Studies. Forty-eight h after MI induction in female Sprague–Dawley rats (Harlan Labora-tories), the survivors were subjected to injection through the femoral veinof 150 μL of either a 0.06 M solution of PS-presenting liposomes (n ¼ 10)or PS-lacking liposomes (n ¼ 10). The control group was injected with saline(n ¼ 10).

Echocardiography studies were performed 24 h after MI induction (base-line echocardiogram) and after 1 mo. The rats were anesthetized, theirchests were shaved, and they were placed on their backs; images were ob-tained by placing the transducer against the chest on the left anterior side.Echocardiograms were performed with a commercially available echocardio-graphy system equipped with 12.5 MHz phased transducer (Hewlett Pack-ard). The heart was first imaged in the 2D mode in the parasternal long andshort axis views of the left ventricle. M-mode images were obtained at thelevel below the tip of the mitral valve leaflets at the level of the papillarymuscles. All measurements were averaged for three consecutive cardiaccycles and performed by an experienced technician blinded to the treat-ment group.

Immunohistochemical and Morphometric Analysis. One month after treat-ment, the hearts were arrested with 15% KCl, perfused with 4% formalde-hyde (15 mmHg) for 20 min, and then sectioned into 3–4 transverse slices andparallel to the atrioventricular ring. Each slice was fixed with 10% bufferedformalin, embedded in paraffin, and sectioned into 5-μm slices. Serial sec-tions were immuno-labeled with following antibodies: anti- α-smooth muscleactin (SMA; monoclonal from Sigma-Aldrich) against pericytes and arteriolesand lectin (Sigma-Aldrich) for detection of endothelial cells. The immuno-stained sections were viewed under Olympus light microscope (BX61, Motor-ized SystemMicroscope) connected to an Olympus (DD71) digital capture sys-tem. The vessel density (#∕mm2) was determined from three or five differentfields in each slide, randomly selected from the α-SMA and lectin-immuno-stained cross-sectioned slides, using Cell* software (Olympus Soft imagingSolutions GmbH).

Postmortem morphometric analysis was performed on heart slices ob-tained 5 mm from the heart’s apex. Five-μm thick cross-sections were sub-jected to Masson’s trichrome staining for nuclei, cytoplasm, and collagenvisualization. The slides were photographed and analyzed with planimetrysoftware (Sigma Scan Pro version 5). The following parameters were mea-sured: average wall thickness (mm) (averaged from three measurementsof septum thickness), average scar thickness (mm) (averaged from three mea-surements of scar thickness), LV cavity area (mm2), and whole LV area (mm2).Relative scar thickness was calculated as average scar thickness divided byaverage wall thickness. Infarct expansion index was calculated as follows:[LV cavity area/whole LV area]/ relative scar thickness.

Harel-Adar et al. PNAS ∣ February 1, 2011 ∣ vol. 108 ∣ no. 5 ∣ 1831

APP

LIED

BIOLO

GICAL

SCIENCE

S

Dow

nloa

ded

by g

uest

on

May

20,

202

0

Statistical Analysis. Statistical analysis was performed with GraphPad Prismversion 5.03 for Windows (GraphPad Software). All variables are expressedas mean� SEM from at least three independent experiments. One wayANOVA studies were performed in liposome uptake experiment, activationin vivo, macrophage activation mechanism after MI, vessel and macrophagedensity, and morphometric analysis. Two-way ANOVA studies were per-formed in the activation of macrophage in vitro experiment as well as inthe functional experiment. P < 0.05 was considered statistically significant.

ACKNOWLEDGMENTS. We thank Prof. Ron N. Apte and Dr. Alon Monsonegofor their critical reading of the paper and useful comments, Emil Ruvinovfor his help with the statistics, and Radka Holbova and Dr. Yael Mardorfor excellent technical assistance. The research was supported by grants fromthe Israel Science Foundation (1368∕08), the European Union FWP7 (INELPY),and the Israel Ministry of Science, Culture and Sport. This work was per-formed in partial fulfillment of the requirements of Tamar Harel-Adar fora PhD degree in the Department of Biotechnology Engineering, Ben-GurionUniversity of the Negev. Prof. Cohen holds the Claire and Harold OshryProfessor Chair in Biotechnology.

1. Frangogiannis NG, Smith CW, Entman ML (2002) The inflammatory response inmyocardial infarction. Cardiovasc Res 53:31–47.

2. Lambert J, Lopez E, Lindsey M (2008) Macrophage roles following myocardialinfarction. Int J Cardiol 130:147–158.

3. Nahrendorf M, et al. (2007) The healing myocardium sequentially mobilizes twomonocyte subsets with divergent and complementary functions. J Exp Med204:3037–3047.

4. Troidl C, et al. (2009) Classically and alternatively activated macrophages contribute totissue remodelling after myocardial infarction. J Cell Mol Med 13:3485–3496.

5. Henson PM (2005) Dampening inflammation. Nat Immunol 6:1179–1181.6. Furnrohr B, et al. (2005) Signals, receptors, and cytokines involved in the immunomo-

dulatory and anti-inflammatory properties of apoptotic cells. Signal Transduction6:356–365.

7. Schmidt T, Carmeliet P (2010) Bridges that guide and unite. Nature 465:697–699.8. Bose J, et al. (2004) The phosphatidylserine receptor has essential functions during

embryogenesis but not in apoptotic cell removal. J Biol 3:15.9. Fadok VA, et al. (1992) Exposure of phosphatidylserine on the surface of apoptotic

lymphocytes triggers specific recognition and removal by macrophages. J Immunol148:2207–2216.

10. Huynh ML, Fadok VA, Henson PM (2002) Phosphatidylserine-dependent ingestion ofapoptotic cells promotes TGF-beta1 secretion and the resolution of inflammation.J Clin Invest 109:41–50.

11. Thum T, Bauersachs J, Poole-Wilson PA, Volk HD, Anker SD (2005) The dying stem cellhypothesis: Immune modulation as a novel mechanism for progenitor cell therapy incardiac muscle. J Am Coll Cardiol 46:1799–1802.

12. Maderna P, Godson C (2003) Phagocytosis of apoptotic cells and the resolution ofinflammation. Biochim Biophys Acta 1639:141–151.

13. Malawista SE, Gee JB, Bensch KG (1971) Cytochalasin B reversibly inhibits phagocytosis:Functional, metabolic, and ultrastructural effects in human blood leukocytes andrabbit alveolar macrophages. Yale J Biol Med 44:286–300.

14. Wang Y, et al. (2007) Ex vivo programmed macrophages ameliorate experimentalchronic inflammatory renal disease. Kidney Int 72:290–299.

15. Gordon S (2003) Alternative activation of macrophages. Nat Rev Immunol 3:23–35.16. Mitchell JE, et al. (2006) The presumptive phosphatidylserine receptor is dispensable

for innate anti-inflammatory recognition and clearance of apoptotic cells. J Biol Chem281:5718–5725.

17. Garlichs CD, et al. (2004) Delay of neutrophil apoptosis in acute coronary syndromes.J Leukocyte Biol 75:828–835.

18. Ronchetti A, et al. (1999) Immunogenicity of apoptotic cells in vivo: Role of antigenload, antigen presenting cells, and cytokines. J Immunol 163:130–136.

19. Torre-Amione G, et al. (2008) Results of a non-specific immunomodulation therapy inchronic heart failure (ACCLAIM trial): A placebo-controlled randomised trial. Lancet371:228–236.

20. Cohen S, Baño MC, Chow M, Langer R (1991) Alginate-lipid interactions can renderchanges in lipid bilayer permeability. Biochim Biophys Acta 1063:95–102.

21. Leor J, et al. (2000) Bioengineered cardiac grafts: A new approach to repair theinfarcted myocardium? Circulation 102:III-56-III-61.

22. Itzhaki-Alfia A, et al. (2009) Patient characteristics and cell source determine thenumber of isolated human cardiac progenitor cells. Circulation 120:2559–2566.

23. Amsalem Y, et al. (2007) Iron-oxide labeling and outcome of transplanted mesench-ymal stem cells in the infarcted myocardium. Circulation 116:I-38–I-45.

1832 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1015623108 Harel-Adar et al.

Dow

nloa

ded

by g

uest

on

May

20,

202

0