Mattia Albiero,1,2 Nicol Poncina,1,2 Stefano Ciciliot,1,2 Roberta Cappellari,1

Lisa Menegazzo,1,2 Francesca Ferraro,3,4 Chiara Bolego,5 Andrea Cignarella,1

Angelo Avogaro,1,2 and Gian Paolo Fadini1,2

Bone Marrow Macrophages Contribute toDiabetic Stem Cell Mobilopathy byProducing Oncostatin MDiabetes 2015;64:2957–2968 | DOI: 10.2337/db14-1473

Diabetes affects bone marrow (BM) structure and impairsmobilization of stem cells (SCs) into peripheral blood (PB).This amplifies multiorgan complications because BMSCspromote vascular repair. Because diabetes skews mac-rophage phenotypes and BM macrophages (BMMF) pre-vent SCmobilization, we hypothesized that excess BMMF

contribute to diabetic SC mobilopathy. We show thatpatients with diabetes have increased M1 macrophages,whereas diabetic mice have increased CD169+ BMMF

with SC-retaining activity. Depletion of BMMF restoredSC mobilization in diabetic mice. We found that CD169labels M1 macrophages and that conditioned medium(CM) from M1 macrophages, but not from M0 and M2macrophages, induced chemokine (C-X-C motif) ligand12 (CXCL12) expression by mesenchymal stem/stromalcells. In silico data mining and in vitro validation identifiedoncostatin M (OSM) as the soluble mediator contained inM1 CM that induces CXCL12 expression via a mitogen-activated protein kinase kinase-p38-signal transducerand activator of a transcription 3–dependent pathway.In diabetic mice, OSM neutralization prevented CXCL12induction and improved granulocyte-colony stimulatingfactor and ischemia-induced mobilization, SC homing toischemic muscles, and vascular recovery. In patients withdiabetes, BM plasma OSM levels were higher and corre-lated with the BM-to-PB SC ratio. In conclusion, BMMF

prevent SC mobilization by OSM secretion, and OSM an-tagonism is a strategy to restore BM function in diabetes,which can translate into protection mediated by BMSCs.

Diabetes leads to multiorgan pathology, which ultimatelyreduces life expectancy (1). A series of consistent studies

in animals and humans indicate that diabetes affects thestructure and function of the bone marrow (BM) (2).Extensive remodeling of the BM microvasculature hasbeen demonstrated in mice (3) and patients (4) with di-abetes. Along with autonomic neuropathy (5), such pro-found alterations in the stem cell (SC) niche cause BMdysfunction, evidenced by an impaired SC mobilization inresponse to ischemia (6,7) and granulocyte-colony stimu-lation factor (G-CSF) (8,9). This novel type of chronicdiabetes complication, termed “stem cell mobilopathy”(10), has implications for the care of patients with diabe-tes with hematological disorders. In addition, because theBM is a reservoir of vascular regenerative cells, BM alter-ations may pave the way to multiorgan damage (2).

On a molecular level, diabetes prevents the chemokine(C-X-C motif) ligand 12 (CXCL12) switch (11,12) (i.e., thesuppression of intramarrow levels of the chemokine CXCL12that normally allows stem cell mobilization) (13). Althougha maladaptive response of the CXCL12-cleaving enzymedipeptidyl peptidase-4 has been hypothesized (14), the exactmechanism perturbing a coordinated CXCL12 regulation indiabetes is unclear. We have previously shown that bypass-ing BM neuronal control, through sympathetic nervoussystem–independent stimuli, restores SC mobilization indiabetes (5). Indeed, diabetic mice can be effectively mobi-lized by the clinical-grade chemokine C-X-C receptor type 4(CXCR4) antagonist AMD3100/Plerixafor, which desensi-tizes SCs to CXCL12, thereby letting them leave the BMand reach the systemic circulation (15).

Although the mobilizing activity of G-CSF is partlymediated by the sympathetic nervous system (16), G-CSFsignals primarily via a receptor expressed on CD68+

1Department of Medicine, University of Padova, Padova, Italy2Venetian Institute of Molecular Medicine, Padova, Italy3Pennsylvania Hospital, University of Pennsylvania Health System, Philadelphia, PA4Fox Chase Cancer Center, Philadelphia, PA5Department of Pharmaceutical Sciences, University of Padova, Italy

Corresponding author: Gian Paolo Fadini, gianpaolofadini@hotmail.com or gianpaolo.fadini@unipd.it.

Received 27 September 2014 and accepted 16 March 2015.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-1473/-/DC1.

© 2015 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

Diabetes Volume 64, August 2015 2957

PATHOPHYSIO

LOGY

macrophages (17), and macrophage suppression is essen-tial to induce SC mobilization (18). Intramarrow macro-phages expressing CD169 (Siglec-1) have been shown tosecrete a hitherto unknown soluble protein that increasesthe expression and release of CXCL12 by mesenchymalstem/stromal cells (MSCs), providing a strong retentionsignal for SCs in the marrow (19). Identification of sucha macrophage-derived factor is a primary challenge in thisfield, because it will eventually turn into a therapeutictarget in “poor mobilizer” conditions, such as diabetes.On the basis of observations that hyperglycemia promotesmyelopoiesis (20) and that diabetes alters macrophagepopulations (21) and is associated with a defectiveCXCL12 switch (11,12), in the study reported here weexamined the role of BM macrophages (BMMF) in diabeticSC mobilopathy. We found an excess of proinflammatorymacrophages in the diabetic BM and that macrophage de-pletion restores mobilization. We also describe the discov-ery that oncostatin M (OSM) is the long-sought solublefactor released by macrophages that sustains CXCL12 ex-pression by MSCs. Neutralization of OSM is thereforea candidate therapy to restore SC mobilization and vascularrepair in diabetes.

RESEARCH DESIGN AND METHODS

PatientsAll protocols involving patients were approved by thelocal ethics committee and done in accordance with theprinciples of the Declaration of Helsinki as revised in2008. All subjects provided written informed consent.Patients with type 1 diabetes (T1D) were recruited fromthe outpatient diabetes clinic of the University Hospitalof Padova, whereas subjects without T1D were selectedamong individuals presenting for a cardiometabolic screen-ing. Details on inclusion and exclusion criteria and clinicalcharacterization of patients are provided in the Supplemen-tary Data. Enrollment in the BM stimulation protocol andtreatment with G-CSF in the trial NCT01102699 are de-scribed elsewhere (9). Coupled peripheral blood (PB) andBM samples were collected from patients undergoing hipreplacement surgery.

AnimalsAll procedures were approved by the local ethics commit-tee and from the Italian Ministry of Health. Experimentswere conducted according to the Principles of LaboratoryAnimal Care (National Institutes of Health publication85-23.4, revised 1985). All animals were on a C57BL/6background. Diabetes was induced with one injection ofstreptozotocin. Additional details are provided in the Sup-plementary Data.

Mobilization AssaysThe following mobilization assays were used: 1) 4-day courseof subcutaneous G-CSF; 2) clodronate liposome injection,followed by a course of G-CSF; and 3) antibody-mediatedOSM neutralization, followed by a course of G-CSF. Detailsare provided in the Supplementary Data.

FACS AnalysisHuman circulating monocyte-macrophages were identi-fied and quantified as previously described (21). M1 weredefined as CD68+CCR2+ cells and M2 were defined asCX3CR1+CD163+/CD206+ cells. Baseline and post–G-CSFlevels of circulating CD34+ SCs were quantified as previ-ously described (9). We used the protocol described byChow et al. (19) to identify murine BMMF phenotypes.Macrophages were identified in the Gr-12/CD1152 gateas cells expressing the macrophage marker F4/80 withlow side scatter. In parallel experiments, macrophageswere also identified as cells that coexpressed F4/80 andMHC-II in the CD45+Gr-12 gate. Mouse progenitor celllevels were quantified in PB before and after mobilization:cells were stained with APC-lineage cocktail, phycoerythrinanti–Sca-1, and fluorescein isothiocyanate anti-cKit to quan-tify Lin2/c-Kit+/Sca-1+ (LKS) cells or with Alexa 647 anti-CD34 and Alexa 488 anti–Flk-1 to quantify endothelialprogenitor cells. Additional details are provided in theSupplementary Data.

Cell CulturesHuman monocyte-derived macrophages were obtained aspreviously described (21): resting M0 cells were polarizedinto M1 or M2 macrophages by 48-h incubation withlipopolysaccharide (LPS) and interferon-g (IFN-g) or in-terleukin (IL)-4 and IL-13, respectively. After 48 h, themedium was removed, macrophages were kept in serum-free RPMI for a further 72 h, and then conditioned media(CM) were harvested. Human MSCs were obtained fromthe BM of patients undergoing orthopedic surgery at theUniversity Hospital of Padova. BM aspirate pellets wereplated on TC Petri dishes with mesenchymal medium.Murine MSCs were obtained by from C56Bl6/J mice: fe-murs and tibia were flushed with ice-cold PBS, and cellswere cultured in minimum essential medium-a (MEM-a).Murine macrophages were obtained from BM cells cul-tured in RPMI 1640 supplemented with macrophage-CSF for 7 days and then polarized for 48 h with LPSand IFN-g (M1) or IL-4 and IL-13 (M2).

Gene Expression AnalysesTotal RNA was extracted using TRIzol reagent followingthe manufacturer’s protocol. RNA was reverse transcribedusing the First Strand cDNA Synthesis Kit, and duplicatessample cDNA were amplified on the 7900HT Fast Real-Time PCR System. Expression data were normalized tothe mean of housekeeping gene ubiquitin C. Additionaldata can be found in the Supplementary Data.

In Vitro CXCL12 AssaysHuman and mouse MSCs at 90% confluence were incubatedwith cytokines or CM. CM were incubated with anti-OSMantibodies, mouse anti-human, and rabbit anti-mouse(MAB295 and AF-495-NA, respectively; R&D Systems). Allexperiments were conducted for 48 h, and then cells werelysed with TRIzol for gene expression analysis. Additionaldetails are provided in the Supplementary Data.

2958 Macrophages Prevent Stem Cell Mobilization Diabetes Volume 64, August 2015

In Silico AnalysesIn silico analyses were performed on gene expressiondata of human and mouse macrophages and MSCsretrieved from Gene Expression Omnibus (GEO) data-base series. Murine and human expression data weredivided in two different groups [M(2) or M0 versusM(IFNg+LPS) or M1 and M(IL4) or M2 versus M(IFNg+LPS)or M1] and then analyzed using the GEO2R tool. Wesubsequently filtered data according to these stringentcriteria: being upregulated in M(IFNg+LPS) versus M(–)and versus M(IL4) macrophages at least fivefold (log-foldchange .2.32) and with an adjusted P value of ,0.001in both groups. The two groups were then crossed to geta list of genes commonly upregulated in M(IFNg+LPS)macrophages. To select genes encoding for secreted pro-teins, data were further filtered by comparing the humanand murine gene lists obtained as above with a species-specific secreted protein list from the Metazoa Secre-tome and Subcellular Proteome Knowledgebase. Wethen manually checked for proteins with a receptorexpressed in MSCs according to relevant GEO series.Technical details on this method can be found in theSupplementary Data.

Statistical AnalysisData are expressed as mean 6 SE or as percentage. Nor-mality was checked using the Kolmogorov-Smirnov test,and nonnormal data were log-transformed before analy-sis. Comparison between two or more groups was per-formed using the Student t test and ANOVA for normalvariables or the Mann-Whitney U test and Kruskal-Wallistest for nonnormal variables. Linear correlations werechecked using the Pearson r coefficient. Statistical analysiswas accepted at P , 0.05.

RESULTS

Macrophage Phenotypes and SC Mobilization inPatients With T1DWe previously showed that prediabetes and type 2diabetes are associated with imbalances in circulatingmonocyte-macrophage phenotypes, reflecting disequilib-rium in BM populations (21,22). We herein report thatpatients with T1D have a significant increase in circulat-ing CD68+CCR2+ M1-like (proinflammatory) macrophagescompared with matched control subjects (clinical charac-teristics in Table 1). This was attributable to a significantincrease in CD68 and, to a lesser extent, CCR2 expression

Table 1—Characteristics of the study population

Variable No diabetes (n = 21) T1D (n = 21) P value

Age, years 47.9 6 2.2 35.4 6 2.5 ,0.001

Male sex, % 52.4 57.2 0.763

BMI, kg/m2 25.3 6 0.9 24.7 6 0.6 0.589

Waist circumference, cm 94.2 6 2.6 89.7 6 2.0 0.189

Fasting plasma glucose, mg/dL 90.4 6 1.6 149.8 6 10.6 ,0.001

HbA1c, % 5.5 6 0.1 7.9 6 0.2 ,0.001

Hypertension, % 23.8 19.0 0.715

Blood pressure, mmHgSystolic 126.4 6 4.1 121.0 6 3.5 0.320Diastolic 81.8 6 2.3 74.3 6 1.6 0.009

Smoking habit, % 14.3 9.5 0.643

Cholesterol, mg/dLTotal 196.3 6 9.6 176.0 6 7.6 0.102HDL 58.7 6 3.6 60.2 6 2.3 0.707LDL 112.7 6 7.8 99.4 6 5.7 0.171

Triglycerides, mg/dL 124.6 6 24.5 82.1 6 9.7 0.101

Retinopathy, % 0.0 33.3 —

Nephropathy, % 0.0 9.5 —

Neuropathy, % 0.0 9.5 —

Atherosclerotic CVD, % 4.7 9.5 0.560

MedicationsInsulin, % 0.0 100.0 —

Metformin, % 0.0 14.2 —

ACE inhibitors, % 19.0 19.0 1.000Other antihypertensives, % 4.8 4.8 1.000Aspirin, % 23.8 4.8 0.081Statin, % 61.9 28.6 0.03

ACE, angiotensin-converting enzyme; CVD, cardiovascular disease.

diabetes.diabetesjournals.org Albiero and Associates 2959

on monocytes of patients with T1D (Fig. 1A and B). Fur-thermore, in patients with T1D undergoing BM stimulationwith G-CSF in the NCT01102699 study (9), there was astrong negative correlation between the degree of CD34+ SCmobilization and the change in CX3CR1+CD163+/CD206+

M2-like monocyte-macrophages, which was not observedin control subjects without diabetes (Fig. 1C). These obser-vations prompted us to explore the role of BMMF in thediabetic SC mobilopathy.

BMMF in T1D MiceWe first examined the percentages of macrophages in theBM of mice with streptozotocin-induced T1D and innondiabetic mice. The Gr-12CD1152F4/80+SSClow and

CD45+Gr-12MHC-II+F4/80+ macrophage phenotypes wereboth increased more than twofold by T1D (Fig. 2A andB). This increase in BMMF appears to be independentfrom sympathetic innervation and from Sirt1 down-regulation, which have been previously shown to mediatediabetic BM dysfunction (5): indeed, chemical sympathec-tomy with 6-hydroxydopamine or hematopoietic Sirt1knockout did not affect the percentages of BMMF (Sup-plementary Fig. 1). Gr-12CD1152F4/80+SSClow cells werefurther characterized by FACS for the expression of clas-sical M1 (CD86) and M2 (scavenger receptors CD301 andCD206) macrophage markers. Although CD301 wasslightly more expressed on BMMF from diabetic mice,there was no obvious prevalence of M1 versus M2 markers

Figure 1—Circulating monocyte-macrophages in patients with T1D and control (CTRL) subjects. A: Representative FACS plots illustratethe gates used to identify circulating M1- and M2-like monocyte-macrophages in a CTRL subject and patient with T1D. B: Quantification ofthe expression of single M1 and M2 markers (left panel) and of M1 and M2 cells (right panel) in CTRL subjects and patients with T1D. *P <0.05 T1D vs. CTRL. C: Correlations between percent change in M2 cells and the change in CD34+ cells after G-CSF stimulation in CTRLsubjects and patients with T1D from the NCT01102699 study. FSC, forward scatter.

2960 Macrophages Prevent Stem Cell Mobilization Diabetes Volume 64, August 2015

(Fig. 2C). Gene expression analysis also did not allow us tounequivocally define Gr-12CD1152F4/80+SSClow cells as M1or M2 (Fig. 2D). BM Gr-12CD1152F4/80+SSClow macro-phages of diabetic mice showed significantly higher surfaceand gene expression of CD169 (Fig. 2E), which preferen-tially labels BMMF compared with other cell populations(Supplementary Fig. 2) and identifies macrophages pro-vided with SC-retaining activity in mice and humans(19,23).

Excess BMMF Contribute to SC Mobilopathy inDiabetic MiceIn nondiabetic mice, the ability of G-CSF to suppressBMMF content is supposed to mediate its mobilizing

activity by lifting the inhibitory signal provided byCXCL12 of mesenchymal origin (17,18). We found thatalthough G-CSF reduced BMMF by ;80% in nondiabeticmice, suppression of BMMF content in the diabetic micewas ,40% and that the percentage of post–G-CSFBMMF was more than fivefold higher in diabetic com-pared with nondiabetic mice (Fig. 3A). We therefore hy-pothesized that SC mobilization failure in response toG-CSF in diabetic mice is at least partly attributable toexcess BMMF content. To clarify this point, we depletedBMMF using clodronate liposomes, which kill phagocyticcells by delivering toxic intracellular concentrations ofclodronate (24). This approach effectively and equally de-pleted BMMF and abated BM CXCL12 gene expression in

Figure 2—Murine BMMF quantification and characterization. A: Representative FACS plots illustrate the gates used to identify BMGr-12CD1152F4/80+SSClow macrophages. The box plot on the right shows quantification of this phenotype in nondiabetic (control [CTRL])and T1D mice. B: Representative FACS plots illustrate the gates used to identify BM CD45+Gr-12MHC-II+F4/80+ macrophages. The boxplot on the right shows quantification of this phenotype in CTRL and T1D mice. FSC, forward scatter. Horizontal line is the median, bordersof the box are the interquartile range, and whiskers are the minimal to maximal range. C: Representative FACS histograms illustrate theexpression of M1 (CD86) and M2 (CD301 and CD206) markers on Gr-12CD1152F4/80+SSClow BMMF. The graph on the right showsquantification in macrophages from CTRL and T1D mice. D: Gene expression analysis of Gr-12CD1152F4/80+SSClow BMMF sorted fromCTRL and T1D mice. E: FACS histogram (left) and graph (middle) illustrate surface expression of CD169 and gene expression (right) onGr-12CD1152F4/80+SSClow BMMF of CTRL and T1D mice. *P < 0.05.

diabetes.diabetesjournals.org Albiero and Associates 2961

nondiabetic and diabetic mice (Fig. 3B and C). PB macro-phages and Gr-1high monocytes were also depleted, butneutrophils were unaffected. Among other niche genes,clodronate liposome treatment also reduced Vcam1 ex-pression (Supplementary Fig. 3A and B). As a result, spon-taneous and G-CSF induced mobilization of LKSs, andCD34+Flk-1+ SCs were restored toward normal levels indiabetic mice (Fig. 3D and E). These data confirm that ex-cess BMMF contribute to SC mobilopathy in diabetes. How-ever, an unrestricted targeting of BMMF with clodronateliposomes is unlikely to be a suitable therapeutic approach torestore BM function because it can have negative off-targeteffects. We therefore sought to identify the hitherto un-known soluble factor released by macrophages that preventsSC mobilization.

M1 Macrophages Provide SC Retention SignalsTo better understand the meaning of CD169 overexpres-sion in diabetic macrophages, we performed an in silicoanalysis of the expression of CD169 probes in the publicGEO data set GDS2429, reporting gene expressionprofiles during typical M1 (IFN-g+LPS) and M2 (IL-4)

macrophage polarization from human monocytes (Fig. 4A).CD169 was confirmed as a macrophage marker becauseits expression markedly increased during monocytes-macrophage differentiation and was much more expressedin M1 than in M0 and M2 macrophages (Fig. 4B). Tovalidate this finding, we obtained human M0, M1, andM2 macrophages using the same in vitro polarization pro-tocol. Polarization efficiency was verified by upregulation ofM1 genes Il1b, Tnfa, and iNos and downregulation ofMrc1/CD206 in cells treated with IFN-g+LPS comparedwith cells treated with IL-4 (Supplementary Fig. 4).CD169 was more expressed in M1 than in M0 and M2by FACS and quantitative PCR (Fig. 4C and D). Thesedata indicate that CD169 is a marker of the proinflamma-tory M1 macrophage phenotype.

The macrophage CM is known to increase the expres-sion and release of the retention chemokine CXCL12by BM MSCs, thereby preventing SC mobilization (19).We therefore obtained CM from M0, M1, and M2 humanmacrophages, and cultured human BM-derived MSCs,most of which expressed Nestin (Fig. 4E), a niche-supporting cell marker (25). After incubating Nestin+ MSCs

Figure 3—Effects of BMMF depletion. A: Percent BMMF (fold change vs. baseline in controls) in nondiabetic (control [CTRL]) and T1Dmice before (baseline) and after a full course of G-CSF stimulation (post–G-CSF). *P < 0.05 vs. baseline or in the comparison indicated bythe line. B: Percent BMMF (fold change vs. baseline in controls) in CTRL and T1D mice before and after macrophage depletion withclodronate liposomes. *P < 0.05 vs. baseline or in the comparison indicated by the line; n.s., not significant. C: CXCL12 expression in thewhole BM in CTRL and T1D mice before and after macrophage depletion with clodronate liposomes. *P < 0.05 vs. baseline. LKS cell (D) andCD34+Flk-1+ cell (E ) (endothelial progenitor cells) mobilization in CTRL and T1D mice in response to clodronate and clodronate+G-CSFadministration in nondiabetic and diabetic mice. *P < 0.05 vs. baseline. #P < 0.05 vs. clodronate alone.

2962 Macrophages Prevent Stem Cell Mobilization Diabetes Volume 64, August 2015

with macrophage CM for 48 h, we found that CM from Mmacrophages, but not from M0 and M2 macrophages,induced CXCL12 expression in MSCs (Fig. 4E). Amongother niche genes, we found that M1 CM increasedAngpt1 and Kitl expression (Supplementary Fig. 5). Thisfinding, which was consistently reproducible using differ-ent batches of CM and different MSC donors, suggeststhat only proinflammatory macrophages (M1) promoteretention versus mobilization of BMSCs through the se-cretion of a soluble mediator. In further support, by usingmouse BM-derived macrophages polarized into M1 andM2 as well as mouse BM MSCs, we confirmed that M1macrophages express higher CD169 levels and that M1CM, but not M0 and M2 CM, induces CXCL12 expression(Supplementary Fig. 6A–C).

Macrophage-Derived OSM Promotes CXCL12ExpressionThe effect of M1 CM on CXCL12 expression by MSCs wascompletely abolished by proteinase K but not by a proteaseinhibitor (Fig. 4F and G), suggesting that the M1 CM, butnot the M0 and M2 CM, contains a protein that signalsin MSCs and stimulates CXCL12 expression. To discoverthis soluble factor, we performed an in silico analysisof the public gene expression profiles of human andmouse polarized macrophages and MSCs (SupplementaryFig. 7A). Genes that were significantly upregulated in M1

compared with M0 and M2 were screened for thoseencoding secreted factors that had a receptor expressedon BM-derived MSCs. A list of candidate human andmouse genes/proteins was thus retrieved and scored bya literature search, looking at soluble factors produced byinflammatory macrophages that might induce CXCL12expression by MSCs. Final candidates were validated invitro by a dose-response stimulation of MSCs (Supple-mentary Fig. 7B). After unsuccessful testing of severalcandidates (CXCL10, CXCL11, tumor necrosis factor-a,platelet-derived growth factor-A, IL-15, and endothelin-1;Supplementary Fig. 8), we found that OSM was ableto exponentially increase CXCL12 expression by MSCs(Fig. 5A). The other gp130 ligands, namely IL-6 andleukemia inhibitory factor, did not exert the sameCXCL12-inducing effect (Supplementary Fig. 9). As measuredby ELISA, OSM protein concentrations were markedlyhigher in the CM of mouse and human M1 than in M0and M2 (Fig. 5B and Supplementary Fig. 6D), and OSMgene expression was several times upregulated in M1verus M0 and M2 macrophages (Fig. 5C). Gene expressionlevels of OSM in the BM were significantly reduced aftertreatment with clodronate liposomes in diabetic andnondiabetic mice (Fig. 5D).

Thus, we focused on OSM as the most likely candidateSC retention factor produced by BMMF. Incubation ofhuman MSCs with CM from human M1 in the presence

Figure 4—CD169 expression and CM activity of M1 macrophages. A: Schematic representation of the strategy used for the in silico analysisof CD169 expression in monocytes and M0, M1, and M2 macrophages (GEO data set GDS2429). B: Average CD169 probe expression inmonocytes (Monos) and M0, M1, and M2 macrophages from GEO data set GDS2429. *P < 0.05. C: Representative FACS histograms ofsurface CD169 expression on in vitro polarization of M0, M1, and M2 macrophages. D: Gene expression analysis of CD169 expression oncultured M0, M1, and M2 macrophages. *P < 0.05 vs. M0. E: Immunofluorescence image indicates Nestin expression on human BM-derivedMSCs (left) and effects of M0, M1, and M2 macrophage CM on CXCL12 gene expression by MSCs. *P< 0.05 vs. control (CTRL). F: Effects ofM0, M1, and M2 macrophage CM, with or without proteinase K, on CXCL12 expression by MSCs. *P < 0.05 vs. CTRL. G: Effects of M0, M1,and M2 macrophage CM, with or without a protease inhibitor, on CXCL12 expression by MSCs. *P < 0.05 vs. CTRL.

diabetes.diabetesjournals.org Albiero and Associates 2963

of a neutralizing anti-human OSM monoclonal antibodycompletely abolished the effect on CXCL12 expression(Fig. 5E). This confirmed that OSM is the soluble factorcontained in M1 CM that stimulates CXCL12 expressionby MSCs. CXCL12 induction by OSM was confirmed inother cell types, such as endothelial cells (human umbil-ical vein endothelial cells and human aortic endothelialcells) and fibroblasts (Supplementary Fig. 10A). Al-though microvascular permeability affects SC mobili-zation (26), we did not find any effect of OSM onpermeability of a human umbilical vein endothelial cellsmonolayer (Supplementary Fig. 10B), suggesting thatCXCL12 regulation is the most likely candidate mecha-nism by which OSM regulates mobilization. To explore

the molecular mechanisms, we analyzed classical path-ways activated by the OSM receptor. Induction ofCXCL12 expression by OSM was abolished when MSCswere cotreated with inhibitors of mitogen-activatedprotein kinase kinase (MEK) (U0126), p38 (SB202190),and signal transducer and activator of transcription3 (STAT3) (Stattic) (Supplementary Fig. 11A). As shownby FACS, STAT3 phosphorylation by OSM was reducedby cotreatment with the p38 inhibitor (SupplementaryFig. 11B), suggesting a MEK-p38-STAT3 pathway. How-ever, simple p38 or STAT3 activation was insufficient toinduce CXCL12 in MSCs (Supplementary Fig. 11C), suggest-ing that recruitment of other cofactors by OSM signalingis required.

Figure 5—Macrophage-derived OSM prevents stem cell mobilization. A: CXCL12 gene expression by MSCs induced by increasingconcentrations of OSM. *P < 0.05 vs. 0. B: OSM protein concentration in M0, M1, and M2 macrophage CM. *P < 0.05 vs. M0. C:OSM gene expression in cultured M0, M1, and M2 macrophages. *P < 0.05 vs. M0. D: OSM gene expression in the BM as a whole atbaseline and after treatment with clodronate liposomes in nondiabetic and diabetic mice. *P < 0.05 vs. baseline. E: CXCL12 geneexpression by MSCs treated with M0, M1, and M2 CM in the presence or in the absence of an anti-OSM neutralizing antibody (aOSM).mAb, monoclonal antibody. *P < 0.05 vs. M0. LKS cell mobilization in response to G-CSF alone (F ) or G-CSF plus a neutralizing anti-OSMantibody (G) in nondiabetic and diabetic mice. *P < 0.05 vs. baseline. Endothelial progenitor cell mobilization in response to G-CSF alone(H) or G-CSF plus a neutralizing anti-OSM antibody (I) in nondiabetic and diabetic mice. *P < 0.05 vs. baseline. J: OSM protein concen-tration in BM plasma of patients without (n = 6) and with (n = 6) diabetes. *P < 0.05. K: BM-to-PB ratio of CD34+ cells, a surrogate ofsteady-state SC mobilization, in patients without and with diabetes. *P < 0.05. L: Linear correlation between OSM concentrations and theBM-to-PB CD34+ cell ratio.

2964 Macrophages Prevent Stem Cell Mobilization Diabetes Volume 64, August 2015

OSM Neutralization Restores SC Mobilization, Homing,and Vascular Recovery in T1DFor an in vivo validation of the role of OSM as a retentionfactor, we treated T1D mice and nondiabetic mice witha neutralizing anti-mouse OSM antibody the day beforeG-CSF stimulation was started. This protocol abatedcirculating OSM concentrations and restored a significantCXCL12 gradient switch by G-CSF in diabetic mice byraising its PB-to-BM concentration ratio (SupplementaryFig. 12). Although the diabetic mice failed to mobilize LKSand CD34+Flk-1+ cells in response to G-CSF alone, OSMneutralization was able to restore G-CSF–induced LKS andCD34+Flk-1+ cell mobilization in diabetic mice to the levelseen in nondiabetic mice and beyond (Fig. 5F–I). In view ofclinical translation, we examined OSM concentrations inrelation to PB and BM CD34+ cell distribution in six indi-viduals with diabetes and six matched individuals withoutdiabetes (clinical characteristics in Supplementary Table 2).The BM plasma OSM concentration and the BM-to-PBCD34+ cell ratio were higher in the patients with diabetes,and a close direct correlation between these parameterswas found, supporting the notion that OSM regulatesBM-to-PB SC mobilization (Fig. 5J–L).

In diabetes, G-CSF and ischemia-induced mobilizationare both defective. We therefore sought to verify whetherOSM inhibition affects the response to ischemia indiabetic mice undergoing hind limb ischemia with orwithout treatment with the neutralizing anti-OSM anti-body. By FACS analysis, we found that OSM neutraliza-tion increased the number of circulating LKS cells afterischemia and restored the number of LKS cells homed tothe ischemic muscles toward levels seen in nondiabeticmice (Fig. 6A and B). In addition, OSM neutralizationrestored hind limb perfusion 14 days after ischemia, asshown by laser Doppler imaging (Fig. 6C).

DISCUSSION

In this study, we show that excess proinflammatorymacrophages in the diabetic BM provide a retentionsignal for SCs by secreting OSM and inducing CXCL12in MSCs. Macrophage depletion and OSM neutralizationwere indeed able to restore SC mobilization towardnormal levels in diabetic mice, which translated intoimproved vascular recovery after ischemia.

Diabetes induces a profound remodeling in the BMSCniche in mice and humans (3–5,27), which impairs SCmobilization in response to ischemia and growth factors(6–9). Multiple molecular pathways may mediate this dys-function, but the exact mechanisms are unknown, thuslimiting the possibility to pursue targeted therapies.Based on the notion that diabetes affects the monocyte-macrophage compartment (21,22,28) and that macro-phages prevent SC mobilization (19), we assessed whetherBMMF contribute to diabetic SC mobilopathy. Usingstandardized FACS protocols (19), we found two- tothreefold increased levels of macrophages in the diabeticBM that were not adequately suppressed by G-CSF. This is

in line with the previous finding that hyperglycemiaskews SC differentiation to myeloid phenotypes throughadvanced glycosylation end product/receptor for ad-vanced glycosylation end products interactions (20).Chow et al. (19) clearly demonstrated that macrophagesequipped with SC retention activity are labeled by theadhesion molecule CD169 because selective depletion ofCD169+ cells allows SC mobilization. Macrophages in thediabetic BM displayed increased expression of CD169,and we therefore hypothesized that such an excess inBMMF blocks SC mobilization in diabetes. To test thishypothesis, we performed macrophage depletion usingclodronate liposomes, which selectively kill specializedphagocytic cells in the reticuloendothelial system, in-cluding the BM. Effective suppression of intramarrowmacrophages was followed by the release of SCs in di-abetic mice and restoration of response to G-CSF towardnormal levels.

These data suggest that macrophage targeting cantherapeutically restore BM function in diabetes. Wetherefore focused on the hitherto unrecognized signal(s)whereby BMMF affect function of the SC niche (29). By anin silico approach, we discovered that OSM is the long-sought soluble factor released by macrophages that inducesthe retention signal CXCL12 in MSCs, thereby preventingSC mobilization. The process that led us to this discoverywas primed by the finding that the adhesion moleculeCD169, which labels macrophages provided with retentionactivity, is markedly upregulated in proinflammatory,so-called classically activated M1 or M(LPS+IFNg) macro-phages, compared with resting M0 and M2 or M(IL-4)macrophages. The segregation of CD169 expression andCM activity enabled us to mine the widely available geneexpression profiles of human and mouse M0, M1, and M2cells in search of candidate secreted factors that signalthrough receptors expressed on MSCs. Potential candi-dates, scored by literature searches, were validated bya simple and rapid in vitro assay of CXCL12 induction.Only OSM, but not the other gp130 ligands LIF and IL-6,fulfilled all such requisites, and OSM blockade completelyabolished the effects of M1 CM on CXCL12 expression inMSCs. A preliminary study of signaling pathways identi-fied the MEK-p38-STAT3 axis as likely mediating theeffect of OSM on CXCL12 expression. Upon an accurateliterature review, it appeared that OSM was alreadyknown to be induced in classically activated M1 macro-phages (30), able to stimulate CXCL12 production byMSCs (31), and possibly involved in SC retention withinthe BM (32). In addition, gp130 ligands have been shownto be upregulated in the BM of long-standing murine di-abetes, whereas genetic deletion of gp130 reverses somepathologic hematopoietic features associated with diabetes(33). Similarly to what we show in the BM, macrophage-derived OSM seems to play a role also in adipose tissueinflammation (34,35), where SCs, the microvasculature,macrophages, and adipocytes form a structure resemblingthe BM niche.

diabetes.diabetesjournals.org Albiero and Associates 2965

In further support of the role of OSM in macrophage-mediated regulation of SC trafficking, we found thatBM OSM expression was significantly suppressed aftermacrophage depletion. In view of clinical translation, wealso found that OSM concentrations were higher in theBM plasma of patients with diabetes compared withpatients without diabetes and were correlated with a sur-rogate index of steady-state mobilization, such as theratio of CD34+ SCs between the BM and PB. As a final proofof concept, we show that in vivo OSM neutralization

restored the SC mobilization response to G-CSF in diabeticmice. Ischemia-induced mobilization and homing of LKScells was also improved after treating diabetic mice withthe OSM neutralizing antibody, which was associatedwith restoration of perfusion. This indicates that anyeventual peripheral effect of OSM blocking did not pre-vent SC from homing to ischemic tissues.

These findings have relevant implications for ourknowledge of how macrophages regulate the niche andadd an important piece to the complicated puzzle of the

Figure 6—OSM neutralization improves mobilization and response to ischemia. A: Circulating LKS SCs were determined at baseline and 3days after ischemia in nondiabetic mice, diabetic mice, and diabetic mice pretreated with a neutralizing anti-OSM antibody (aOSM). *P <0.05 vs. baseline. B: LKS cells were identified by FACS in the cell suspension of ischemic gastrocnemius and adductor muscles ofnondiabetic mice, diabetic mice, and diabetic mice pretreated with aOSM. *P < 0.05 as indicated. Panels on the right show representativeFACS plots of c-kit and Sca-1 staining after gating live (7AAD2) Lin2 events. C: Laser Doppler imaging was used to determine perfusionrecovery after ischemia as the ischemic-to-nonischemic ratio in the three groups of animals. *P < 0.05 as indicated. Right panels showrepresentative laser Doppler images of data quantified in left panel C.

2966 Macrophages Prevent Stem Cell Mobilization Diabetes Volume 64, August 2015

diabetic BM pathology. The pathological pathway gener-ated by excess BMMF seems independent from othertypical features of the diabetic BM, such as neuropathy,oxidative stress, and Sirt1 downregulation (5). However,in view of the severe BM remodeling taking place in di-abetes, it is not surprising that both macrophage deple-tion and OSM neutralization, although effective inrestoring response to G-CSF, often elicited a blunted mo-bilization in diabetic compared with nondiabetic mice.

On the background of the well-known hyperglycemia-driven myelopoiesis (20), our data indicate that the gen-eration of proinflammatory macrophages represents a keyevent in BM dysfunction. Identification of OSM as themediator of macrophage-retaining activity has therapeuticimplications to revert diabetic SC mobilopathy and, po-tentially, in other “poor mobilizer” conditions. BecauseBM-derived cells play a major role in diabetes complica-tions (36), restoration of BMSC mobilization with a tar-geted molecular approach may restore endogenousvascular regenerative capacity and improve the outcomeof patients with diabetes.

Funding. This study was supported by grant GR-2010-2301676 of the ItalianMinistry of Health to G.P.F. and by a European Foundation for the Study ofDiabetes (EFSD)/Novartis program grant to G.P.F. M.A. is supported by the ItalianSociety of Diabetology.Duality of Interest. M.A., S.C., and G.P.F. are the inventors of a patentpending, held by the University of Padova, on the use of OSM inhibition for theinduction of stem cell mobilization in diabetes. No other potential conflicts ofinterest relevant to this article were reported.Author Contributions. M.A., N.P., S.C., R.C., L.M., and F.F. researcheddata. C.B. and A.C. researched data and contributed to discussion. A.A. contrib-uted to discussion and reviewed and edited the manuscript. G.P.F. researcheddata and wrote the manuscript. G.P.F. is the guarantor of this work and, as such,had full access to all the data in the study and takes responsibility for the integrityof the data and the accuracy of the data analysis.

References1. Seshasai SR, Kaptoge S, Thompson A, et al.; Emerging Risk Factors Col-laboration. Diabetes mellitus, fasting glucose, and risk of cause-specific death.N Engl J Med 2011;364:829–8412. Fadini GP, Ferraro F, Quaini F, Asahara T, Madeddu P. Concise review:diabetes, the bone marrow niche, and impaired vascular regeneration. Stem CellsTransl Med 2014;3:949–9573. Oikawa A, Siragusa M, Quaini F, et al. Diabetes mellitus induces bonemarrow microangiopathy. Arterioscler Thromb Vasc Biol 2010;30:498–5084. Spinetti G, Cordella D, Fortunato O, et al. Global remodeling of thevascular stem cell niche in bone marrow of diabetic patients: implication ofthe microRNA-155/FOXO3a signaling pathway. Circ Res 2013;112:510–5225. Albiero M, Poncina N, Tjwa M, et al. Diabetes causes bone marrow auto-nomic neuropathy and impairs stem cell mobilization via dysregulated p66Shcand Sirt1. Diabetes 2014;63:1353–13656. Fadini GP, Sartore S, Schiavon M, et al. Diabetes impairs progenitor cellmobilisation after hindlimb ischaemia-reperfusion injury in rats. Diabetologia2006;49:3075–30847. Ling L, Shen Y, Wang K, et al. Worse clinical outcomes in acute myocardialinfarction patients with type 2 diabetes mellitus: relevance to impaired endo-thelial progenitor cells mobilization. PLoS One 2012;7:e50739

8. Fadini GP, Avogaro A. Diabetes impairs mobilization of stem cells for thetreatment of cardiovascular disease: a meta-regression analysis. Int J Cardiol2013;168:892–8979. Fadini GP, Albiero M, Vigili de Kreutzenberg S, et al. Diabetes impairs stemcell and proangiogenic cell mobilization in humans. Diabetes Care 2013;36:943–94910. DiPersio JF. Diabetic stem-cell “mobilopathy”. N Engl J Med 2011;365:2536–253811. Tepper OM, Carr J, Allen RJ Jr, et al. Decreased circulating progenitor cellnumber and failed mechanisms of stromal cell-derived factor-1alpha mediated bonemarrow mobilization impair diabetic tissue repair. Diabetes 2010;59:1974–198312. Ferraro F, Lymperi S, Mendez-Ferrer S, et al. Diabetes impairs hemato-poietic stem cell mobilization by altering niche function. Sci Transl Med 2011;3:104ra10113. Petit I, Szyper-Kravitz M, Nagler A, et al. G-CSF induces stem cell mobili-zation by decreasing bone marrow SDF-1 and up-regulating CXCR4. Nat Immunol2002;3:687–69414. Fadini GP, Albiero M, Seeger F, et al. Stem cell compartmentalization indiabetes and high cardiovascular risk reveals the role of DPP-4 in diabetic stemcell mobilopathy. Basic Res Cardiol 2013;108:31315. Broxmeyer HE, Orschell CM, Clapp DW, et al. Rapid mobilization of murineand human hematopoietic stem and progenitor cells with AMD3100, a CXCR4antagonist. J Exp Med 2005;201:1307–131816. Katayama Y, Battista M, Kao WM, et al. Signals from the sympatheticnervous system regulate hematopoietic stem cell egress from bone marrow. Cell2006;124:407–42117. Christopher MJ, Rao M, Liu F, Woloszynek JR, Link DC. Expression ofthe G-CSF receptor in monocytic cells is sufficient to mediate hemato-poietic progenitor mobilization by G-CSF in mice. J Exp Med 2011;208:251–26018. Winkler IG, Sims NA, Pettit AR, et al. Bone marrow macrophages maintainhematopoietic stem cell (HSC) niches and their depletion mobilizes HSCs. Blood2010;116:4815–482819. Chow A, Lucas D, Hidalgo A, et al. Bone marrow CD169+ macrophagespromote the retention of hematopoietic stem and progenitor cells in the mes-enchymal stem cell niche. J Exp Med 2011;208:261–27120. Nagareddy PR, Murphy AJ, Stirzaker RA, et al. Hyperglycemia promotesmyelopoiesis and impairs the resolution of atherosclerosis. Cell Metab 2013;17:695–70821. Fadini GP, de Kreutzenberg SV, Boscaro E, et al. An unbalanced monocytepolarisation in peripheral blood and bone marrow of patients with type 2 diabeteshas an impact on microangiopathy. Diabetologia 2013;56:1856–186622. Fadini GP, Cappellari R, Mazzucato M, Agostini C, Vigili de Kreutzenberg S,Avogaro A. Monocyte-macrophage polarization balance in pre-diabetic in-dividuals. Acta Diabetol 2013;50:977–98223. Chow A, Huggins M, Ahmed J, et al. CD169+ macrophages provide a nichepromoting erythropoiesis under homeostasis and stress. Nat Med 2013;19:429–43624. Van Rooijen N, Sanders A. Liposome mediated depletion of macrophages:mechanism of action, preparation of liposomes and applications. J ImmunolMethods 1994;174:83–9325. Méndez-Ferrer S, Michurina TV, Ferraro F, et al. Mesenchymal and hae-matopoietic stem cells form a unique bone marrow niche. Nature 2010;466:829–83426. Mangialardi G, Katare R, Oikawa A, et al. Diabetes causes bone marrowendothelial barrier dysfunction by activation of the RhoA-Rho-associated kinasesignaling pathway. Arterioscler Thromb Vasc Biol 2013;33:555–56427. Teraa M, Fledderus JO, Rozbeh RI, Leguit RJ, Verhaar MC; Juventas StudyGroupdagger. Bone marrow microvascular and neuropathic alterations in patientswith critical limb ischemia. Circ Res 2014;114:311–31428. Olefsky JM, Glass CK. Macrophages, inflammation, and insulin resistance.Annu Rev Physiol 2010;72:219–246

diabetes.diabetesjournals.org Albiero and Associates 2967

29. Bordon Y. Macrophages: bloody regulators. Nat Rev Immunol 2013;13:30730. Guihard P, Danger Y, Brounais B, et al. Induction of osteogenesis in mes-enchymal stem cells by activated monocytes/macrophages depends on oncos-tatin M signaling. Stem Cells 2012;30:762–77231. Lee MJ, Song HY, Kim MR, Sung SM, Jung JS, Kim JH. Oncostatin Mstimulates expression of stromal-derived factor-1 in human mesenchymal stemcells. Int J Biochem Cell Biol 2007;39:650–65932. Minehata K, Takeuchi M, Hirabayashi Y, et al. Oncostatin m maintains thehematopoietic microenvironment and retains hematopoietic progenitors in thebone marrow. Int J Hematol 2006;84:319–327

33. Hazra S, Jarajapu YP, Stepps V, et al. Long-term type 1 diabetes influenceshaematopoietic stem cells by reducing vascular repair potential and increasing in-flammatory monocyte generation in a murine model. Diabetologia 2013;56:644–65334. Sanchez-Infantes D, White UA, Elks CM, et al. Oncostatin M is produced inadipose tissue and is regulated in conditions of obesity and type 2 diabetes. JClin Endocrinol Metab 2014;99:E217–E22535. Lapeire L, Hendrix A, Lambein K, et al. Cancer-associated adipose tissuepromotes breast cancer progression by paracrine oncostatin M and Jak/STAT3signaling. Cancer Res 2014;74:6806–681936. Fadini GP. A reappraisal of the role of circulating (progenitor) cells in thepathobiology of diabetic complications. Diabetologia 2014;57:4–15

2968 Macrophages Prevent Stem Cell Mobilization Diabetes Volume 64, August 2015