Newly identified adipose tissue macrophage populations in obesity with distinct chemokine and chemokine receptor expression



Infiltration by macrophages is a hallmark of obesity-related adipose tissue (AT) inflammation that is tightly linked to insulin resistance. Although CD11c+ AT macrophages (ATMs) have recently been shown to promote inflammation in obese mice, the knowledge on phenotype and function of different ATM populations is still very limited. This study aimed at identifying and characterizing ATM populations in obesity.


Isolation of ATM populations defined by CD11c and mannose receptor (MR) expression and analysis of gene expression in high-fat diet-induced obese mice.


Obesity provoked a shift from a predominant MR+CD11c population (‘MR-ATM’) to two MR populations, namely MRCD11c+ (‘CD11c-ATM’) and MRCD11c (double negative, ‘DN-ATM’). Although CD11c-ATMs were of a clear inflammatory M1 phenotype, DN-ATMs expressed few inflammatory mediators and highly expressed genes for alternative activation (M2) markers involved in tissue repair, such as arginase and YM1. In contrast, MR-ATMs marginally expressed M1 and M2 markers but highly expressed chemokines, including Mcp-1 (Ccl2) and Mcp-3 (Ccl7). Both CD11c-ATMs and DN-ATMs, but not MR-ATM, highly expressed a panel of chemokine receptors (namely Ccr2, Ccr5, Ccr3 and Cx3cr1), whereas the expression of Ccr7 and Ccr9 was selective for CD11c-ATMs and DN-ATMs, respectively. Notably, stressed adipocytes upregulated various chemokines capable of attracting CD11c-ATM and DN-ATM.


This study identifies a novel ATM population with a putatively beneficial role in AT inflammation. This DN-ATM population could be attracted to the obese AT by similar chemokines such as inflammatory CD11c-ATM, on which only Ccr7 is uniquely expressed.


Obesity is associated with adipose tissue (AT) inflammation that is causally involved in the development of insulin resistance.1 AT macrophages (ATMs) are the main source of inflammatory mediators in the obese AT.2, 3 ATMs derive from the bone marrow and infiltrate the AT upon obesity in mice4, 5 and humans.6 To understand the link of obesity to AT inflammation and insulin resistance mechanisms by which monocytes are attracted to the obese AT and their differentiation to tissue macrophages have to be elucidated.

Obesity promotes endoplasmic reticulum (ER) stress in adipocytes7 and subsequent necrosis-like cell death that is probably based on adipocyte hypertrophy.8 Dead adipocytes are often found surrounded by ATMs in so-called ‘crown-like structures’ supposed to scavenge cell debris and free lipid droplets.8 Scavenging dead adipocytes may be a necessary process for tissue homeostasis, and macrophages are also involved in tissue repair. ATMs express a series of markers such as YM1 and arginase typical for alternatively activated, anti-inflammatory macrophages, designated as M2.9, 10, 11 Hence, macrophages attracted to the AT could also exert distinct beneficial functions. A recent study12 showed that in obesity, crown-like structures are enriched with ATM negative for the M2 marker MGL-1 but highly positive for CD11c. This surface molecule is regarded as a marker of inflammatory M1 macrophages and its expression in ATM is considerably increased upon high-fat diet feeding.9, 13 Macrophage mannose receptor (MR; also designated CD206), another M2 marker molecule, has been previously shown by us and others to be highly expressed on human ATMs.14, 15 In previous unpublished experiments, we found that MR is also expressed on the surface of murine ATMs but further characteristics of MR-expressing macrophages remained elusive.

The expression of CC chemokines Mcp-1 (gene: CCL2), Mcp-2 (CCL8), Mcp-3 (CCL7), RANTES (CCL5), MIP-1α (CCL3) and CCL11 is increased in the AT of morbidly obese compared with lean subjects.16 Furthermore, gene expression of CC chemokine receptors CCR1, CCR2, CCR3 and CCR5 is elevated in omental and subcutaneous AT of obese patients.16 Moreover, in the AT of obese mice, the expression of chemotactic proteins such as Mcp-1, Mcp-2 and RANTES, as well as that of chemokine receptors such as Ccr2 and Ccr5 is increased.17, 18 Studies in Mcp-1(Ccl2)-deficient mice and mice overexpressing Mcp-1 support a role of this chemokine in attracting macrophages to the AT and reduced insulin sensitivity in obesity,19, 20 but these data have been repeatedly challenged by others.21, 22, 23 ATMs that are newly recruited to the AT during high-fat diet feeding exhibit higher expression of the Mcp-1 receptor Ccr2 compared with resident ATMs,24 and Ccr2-deficient mice show reduced macrophage infiltration into the AT and improved insulin sensitivity after high-fat diet feeding.12, 25, 26 These studies indicate that Ccr2 is important even though not exclusively responsible for ATM recruitment, but the chemoattractants provoking ATM recruitment to the AT and their respective receptors still have to be investigated.

As little is known about the biological role of defined macrophage populations and the mechanisms underlying their attraction to the AT, this study aimed at identification and characterization of ATM populations in murine obesity to identify putative mechanisms and consequences of their recruitment. We found that cell-surface markers MR and CD11c define three distinct populations in the murine gonadal AT, namely MRCD11c+ (‘CD11c-ATM’), MR+CD11c (‘MR-ATM’) and MRCD11c (double negative, ‘DN-ATM’). Diet-induced obesity shifted the distribution of these populations from MR-ATMs to CD11c-ATMs, which were of a classically activated M1 type, and DN-ATMs, which resembled a presumably beneficial, M2-like macrophage type. ATMs form Ccl2-knockout mice showed comparable results. Analysis of chemokine receptors identified Ccr2, Ccr5, Ccr3 and Cx3cr1 to be expressed on both CD11c-ATM and DN-ATM, whereas Ccr7 and Ccr9 were selectively expressed in CD11c-ATM and DN-ATM, respectively. In vitro analyses showed that ER stress in adipocytes induced ligands for these chemokine receptors, indicating a potential mode of ATM recruitment to the obese AT.

Materials and methods


Male C57BL/6J mice (wild type, wt) and Ccl2 knockouts (Charles River Laboratories, Sulzfeld, Germany) aged 7 weeks were placed for 20 weeks on high-fat (60% fat calories, D12492, Research Diets Inc., New Brunswick, NJ, USA) and low-fat (10% fat calories, D12450B, Research Diets Inc.) diet to induce obesity and to serve as lean controls, respectively. Mice had free access to food and water. After 20 weeks on diet, mice were killed by cervical dislocation. After killing, the gonadal white AT was removed and immediately put on ice-cold phosphate-buffered saline for cell preparation. All mice were maintained on a 12-h light/dark cycle. The study protocols were approved by the local ethics committee for animal experiments and followed the guidelines on accommodation and care of animals formulated by the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes.

Cell preparation and flow cytometry

As a well-accepted source for intra-abdominal fat in quantities sufficient for cell sorting, we used the gonadal AT depot for our analyses. Visible vessels and connective tissue were carefully removed and the AT was cut into pieces, extensively washed in phosphate-buffered saline and digested for 1 h with 35 μg ml−1 liberase 3 (Roche, Mannheim, Germany) and 50 Units ml−1 DNAse I (Sigma, St Louis, MO, USA). The digested tissue was passed through 70-μm mesh filters and centrifuged to obtain the stromal-vascular cell fraction. Cells were stained for three-color immunofluorescence analysis by dye-labeled antibodies according to standard procedures using CD16/CD32 blocking Abs (BD Biosciences, San Jose, CA, USA) and F/480-PE, CD206(MR)-AlexaFluor 488 (both from AbD Serotec, Düsseldorf, Germany) and CD11c-APC (BD Biosciences). Flow cytometric analysis was performed on a FACSCanto (BD Biosciences). For isolation of ATM populations, stromal-vascular cells were stained in identical manner. Subsequently, F4/80+MRCD11c+ (CD11c-ATM), F4/80+MRCD11c (DN-ATM) and F4/80+MRCD11c (MR-ATM) were sorted using FACSAria (BD Biosciences) and immediately lysed in TRIzol reagent (Invitrogen, Carlsbad, CA, USA) for RNA extraction.

Gene expression analysis

Cells were lysed in TRIzol reagent (Invitrogen), and RNA was isolated according to the manufacturer's protocol. A total of 0.5 μg of total RNA was reverse transcribed into cDNA using Superscript II and random hexamer primers (all from Invitrogen). The gene expression of was determined using commercial Assays-on-Demand kits for quantitative real-time reverse transcriptase-PCR on an ABI Prism 7000 cycler (all from Applied Biosystems, Foster City, CA, USA). For quantification, the ΔΔCt method was applied using ubiquitin C as the housekeeping gene. In general, the mean value of wt MR-ATM was considered the base for ‘fold induction’.


Frozen sections from the gonadal AT were stained with rat anti-mouse CD206 IgG (AbD Serotec) detected by goat anti-rat IgG AlexaFluor 488 and 594 conjugates (both from Invitrogen); rabbit anti-mouse YM1 polyclonal antibody (StemCell Technologies Inc., Vancouver, Canada) detected by a goat anti-rabbit IgG AlexaFluor 594 conjugate (Invitrogen); and directly AlexaFluor 488-labeled hamster anti-mouse CD11c (BioLegend, San Diego, CA, USA). The nuclei were visualized by 4′,6-diamidino-2′-phenylindole staining. Slides were mounted in Vectashield mounting medium (Vector Laboratories Inc., Burlingame, CA, USA) and examined under a Leica fluorescence microscope (Leica, Microsystems, Wetzlar, Germany) using × 100/0.70 objective magnification.

In vitro adipocyte differentiation and ER stress

3T3-L1 cells were differentiated into adipocytes by incubation in DMEM (Dulbecco's modified Eagle's medium; Invitrogen) supplemented with 4500 mg l−1 glucose, 110 mg l−1 sodium pyruvate, 2 mM L-glutamine, 50 μg ml−1 streptomycin, 50 Units ml−1 penicillin (all from Invitrogen), 4 μg ml−1 pantothenate and 8 μg ml−1 biotin (both from Sigma; this supplemented medium is referred to as sDMEM) containing 10% fetal bovine serum (both from Invitrogen), 500 nM dexamethasone, 500 μM 3-isobutyl-L-methylxanthine and 1 μg ml−1 insulin from bovine pancreas (all from Sigma) for 2 days. Subsequently, cells were incubated in sDMEM containing 10% fetal bovine serum and 1 μg ml−1 insulin for 4 days before cultivation in sDMEM containing 10% fetal bovine serum for another 4 days and induction of ER stress by incubation with 1 μM calcimycin (A-23187 free acid, Invitrogen), 2.5 μg ml−1 tunicamycin and 2.5 μM thapsigargin (both from Sigma) or solvent (dimethyl sulfoxide, 0.025% (v/v)) for 8 and 16 h, as indicated.


For comparison of two groups (Figure 1), Student's t-test, for comparison of three groups (Figures 2, 4, 5 and 6 and Supplementary Figure 1), one-way ANOVA (analysis of variance), followed by post hoc Student's t-test and for comparison of several groups with one control group (Figure 7 and Supplementary Figure 2), one-way ANOVA, followed by post hoc Dunnett's test were performed. A P-value 0.05 was generally regarded as statistically significant.

Figure 1

CD11c- and MR-defined ATM populations in lean and obese mice. (a) Flow cytometric analysis of stromal-vascular cells from the gonadal adipose tissue of C57BL/6J mice. Typical dot blots of F4/80-positive cells from lean or diet-induced obese mice additionally stained with anti-CD11c and anti-MR mAbs or the respective isotype controls, as indicated, are shown. (b) Summary of FACS analysis. The diagrams show the percentage of ATM (F4/80-positive cells) positive for MR or CD11c or double-negative (DN) in mean±s.e.m. of six lean and obese animals, respectively. ***P<0.001 for lean vs obese.

Figure 2

M1 and M2 marker gene expression in CD11c- and MR-defined ATM populations. (ac) F4/80+MR+CD11c (MR-ATM), F4/80+MR+CD11c (DN-ATM) and F4/80+CD11c+MR (CD11c-ATM) cells were isolated from the gonadal adipose tissue of high-fat diet-induced obese C57BL/6J mice by fluorescence-activated cell sorting. Gene expression was determined by quantitative real-time RT-PCR and quantified by the ΔΔCt method using ubiquitin C as the housekeeping gene. Diagrams show mRNA expression levels of indicated genes in mean±s.e.m. related to MR-ATM (n=6). Statistical significances are indicated by letters above columns: identical letters indicate no statistically significant differences between respective groups (P>0.05).


Surface MR and CD11c define three ATM populations of distinct types of differentiation

The reported emergence of CD11c+ macrophages upon high-fat diet feeding9 in parallel with a decrease in the proportion of MR-expressing cells (own unpublished observation) prompted us to investigate populations of ATM from the gonadal AT as determined by these markers. Flow cytometric analysis showed that CD11c and MR were rarely coexpressed on ATM, which we defined as F4/80+ stromal-vascular cells(<2.5% of ATM; Figure 1a). Thus, by CD11c and MR surface expression, three ATM populations can be distinguished: single CD11c+ (CD11c-ATM), double-negative (DN-ATM) and single MR+ ATM (MR-ATM; Figure 1a). Comparing lean animals on low-fat diet with high-fat diet-induced obese animals, we found the proportion of the CD11c-ATM population to be increased but still comprising only 25±1.5% of total ATM. However, a marked increase in DN-ATMs to >50% of ATMs was detected in parallel with a decreased percentage of MR-ATM cells (Figure 1b).

We isolated ATMs subdivided into three populations as determined by CD11c and MR surface expression from gonadal fat pads of obese mice using fluorescence-activated cell sorting. Expression of genes for CD11c (Itgax) and MR (Mrc1) confirmed proper sorting of the populations (Figure 2a). Whereas Itgax was almost exclusively expressed in CD11c-ATM, the expression of Mrc1 was enhanced by only twofold in MR-ATM compared with DN- and CD11c-ATM. These data indicate that MR surface expression in these ATM populations is subjected to extensive posttranscriptional regulation.

Next, we analyzed the three ATM populations with respect to expression of a series of prototypic classical (M1) and alternative activation (M2) marker genes. Figure 2b shows that CD11c-ATMs preferentially express genes for typical M1 markers and cytokines supposed to be involved in the onset of obesity-induced insulin resistance. iNOS (gene: Nos2), IL (interleukin)-6 (Il6) and OPN (Spp1)27, 28 were markedly higher expressed in CD11c-ATMs compared with DN-ATMs and MR-ATMs (Figure 2b). However, as exceptions, IL-1β (Il1b) expression was upregulated to a similar extent in CD11c-ATMs and DN-ATMs compared with MR-ATMs, whereas tumor necrosis factor-α (Tnf) expression did not significantly differ between all three populations (Figure 2b). Strikingly, genes for typical M2 markers and proteins involved in tissue repair YM1 (Chi3l3) and Arginase (Arg1), were preferentially expressed in DN-ATMs, whereas anti-inflammatory IL-10 (Il10) was uniformly expressed in all investigated populations (Figure 2c). Summarizing, these data show that CD11c-ATMs are of a M1 phenotype as expected, whereas DN-ATMs, the number of which markedly increased in obesity (Figure 1b), are of an alternatively activated (M2) phenotype and MR-ATMs appear to be relatively quiescent.

Location of CD11c-, DN- and MR-ATM in the AT

To investigate the location of the described populations within the AT, we stained sections of the gonadal AT from obese mice using antibodies against CD11c, MR and YM1, which we found especially highly expressed in DN-ATMs (Figure 2c; Chi3l3) and was expressed exclusively in F4/80-positive cells (not shown). CD11c and MR expression overlapped only in a small proportion of cells (Figure 3a) confirming our flow cytometric analysis (Figure 1a). Whereas MR+ cells (MR-ATMs) were clearly negative for YM1 (Figure 3b), CD11c+ cells (CD11c-ATMs) expressed sufficient YM1 to stain positive in sections (Figure 3c), so that only YM1+CD11c cells in the immunofluorescence staining reliably correspond to DN-ATMs. However, these staining patterns confirm the existence of three different ATM populations in situ as defined by the expression of CD11c, MR or neither.

Figure 3

Distribution of CD11c-, YM1- and MR-positive cells in AT. Frozen sections of the gonadal adipose tissue from high-fat diet-induced obese C57BL/6J mice were double stained for (a) CD11c (green) and MR (red), (b) MR (green) and YM1 (red), and (c) CD11c (green) and YM1 (red). Nuclei are stained in blue (DAPI). Sections from at least four animals were analyzed for each staining, typical examples are shown.

MR-ATMs were found both as a single-laying interstitial ATM and clustered in the so-called crown-like structures (Figures 3a and b). In contrast, DN-ATM and CD11c-ATM cells were preferentially located in clusters constituting crown-like structures (Figures 3a–c). However, enrichment with CD11c and DN-ATMs differed considerably between individual crown-like structures (Figure 3c). Hence, variations in surface marker expression were associated with altered distribution of ATM populations in the AT.

CD11c-ATMs and DN-ATMs, but not MR-ATMs, express several chemokine receptors in common

Chemoattraction of ATM is a pivotal step in inflammation of the obese AT, and expression of chemokine receptors may provide hints on the chemoattractants involved. Hence, we analyzed whether the identified ATM populations from the obese gonadal AT differentially express specific chemokine receptors. Ccr2, Ccr3, Ccr5 and Cx3cr1 expression was considerably higher in CD11c-ATMs and in DN-ATMs compared with MR-ATMs (Figure 4 and Supplementary Figure 1a). In contrast, Ccr1 was expressed to a similar extent in all populations (Figure 4), and Ccr8 could not be detected at all (not shown). Strikingly, Ccr7 and Ccr9 were expressed specifically in CD11c-ATMs and DN-ATMs, respectively (Figure 4). Altogether, CD11c-ATMs and DN-ATMs share a panel of important chemokine receptors, whereas chemokine receptor expression is poor on MR-ATMs.

Figure 4

Chemokine receptor expression in CD11c- and MR-defined ATM populations. F4/80+MR+CD11c (MR-ATM), F4/80+MR+CD11c (DN-ATM) and F4/80+CD11c+MR (CD11c-ATM) cells were isolated from the gonadal adipose tissue of high-fat diet-induced obese C57BL/6J mice by fluorescence-activated cell sorting. Gene expression was determined by quantitative real-time RT-PCR and quantified by the ΔΔCt method using Ubiquitin C as the housekeeping gene. Diagrams show mRNA expression levels of indicated genes related to MR-ATM (n=6). Statistical significances are indicated by letters above columns: identical letters indicate no statistically significant differences between respective groups (P>0.05).

ATM populations are unchanged in obese Ccl2-knockout mice

Mcp-1 (Ccl2) could be a key factor in attracting macrophages to the AT, but its role is discussed controversially in the literature.19, 20, 21, 22, 23 We did not find any differences in ATM numbers, and the percentage of F4/80-positve cells in high-fat diet-fed Ccl2−/− knockout compared with wt mice (not shown). Investigating the ATM populations defined in this study by CD11c and MR in Ccl2−/− mice showed a similar pattern in Ccl2−/− knockout compared with wt animals (Figure 5a cf. Figure 1a), with no statistically significant differences (data not shown). Notably, ATM populations in both genotypes exhibited comparable M1 and M2 markers (Figure 5b) and chemokine receptor expression (Figure 5c and Supplementary Figure 1b; all values in Figures 5b and c are related to wt MR-ATMs and depicted with scaling identical to Figures 2 and 4, respectively). The only major difference in Ccl2-deficient mice compared with wt mice was a marked reduction of Ccr2 expression in CD11c-ATMs and DN-ATMs, indicating reduced attraction by or induction of the cognate Ccl2 receptor (Figure 5c). However no compensatory overexpression of other C-C chemokine receptors (Figure 5c) or C-C chemokines (Supplementary Figure 1d) could be detected. These data corroborate reports that question a significant role of Ccl2 in murine obesity-induced AT inflammation.

Figure 5

CD11c- and MR-defined ATM populations and their gene expression Ccl2-knockout mice. (a) Flow cytometric analysis of stromal-vascular cells from the adipose tissue of in high-fat diet-induced obese Ccl2-knockout mice. The diagrams show the percentage of ATM (F4/80-positive cells) positive for MR or CD11c or double-negative (DN) in mean±s.e.m. of six obese animals, respectively. (b, c) F4/80+MR+CD11c (MR-ATM), F4/80+MR+CD11c (DN-ATM) and F4/80+CD11c+MR (CD11c-ATM) cells were isolated from the gonadal adipose tissue of high-fat diet-induced obese Ccl2-knockout mice by fluorescence-activated cell sorting. Gene expression was determined by quantitative real-time RT-PCR and quantified by the ΔΔCt method using Ubiquitin C as the housekeeping gene. Diagrams show mRNA expression levels of indicated genes in mean±s.e.m. related to wild-type MR-ATM (n=6). Axis scaling was chosen identical to Figures 2 and 4. Statistical significances are indicated by letters above columns: identical letters indicate no statistically significant differences between respective groups (P>0.05).

Chemokines are primarily expressed in MR-ATMs

It is known that ATMs express chemokines, thereby facilitating recruitment of more ATMs. The discovery of different ATM populations poses the question whether chemokine-expressing (attracting) ATMs and chemokine receptor-expressing (recruited) ATMs belong to the same or different populations. In addition to chemokine receptor data shown above, we therefore investigated the gene expression of a panel of C-C and C-X3-C chemokines, the expression of which has been shown to be elevated in murine and human obesity.16, 17, 18 Whereas chemokine receptor expression was low in the MR-ATM population sorted from the gonadal AT of obese mice (Figure 4), the majority of investigated chemokines (Ccl2, Ccl5, Ccl8, Ccl11 and Cx3cl1) was primarily expressed in MR-ATMs (Figure 6 and Supplementary Figure 1c). In contrast, Ccl3 expression was uniform in all ATM populations, whereas Ccl7 and Ccl9 were predominantly expressed in both DN-ATMs and CD11c-ATMs (Figure 6 and Supplementary Figure 1c. Ccl19 and Ccl25, the ligands for the CD11c-ATM- and DN-ATM-specific receptors Ccr7 and Ccr9, respectively, were both expressed at high levels in DN-ATMs, Cc119 additionally in MR-ATMs, although the differences between the groups are not in all cases statistically significant (Figure 6, bottom).

Figure 6

C-C chemokine expression in CD11c- and MR-defined ATM populations. F4/80+MR+CD11c (MR-ATM), F4/80+MR+CD11c (DN-ATM) and F4/80+CD11c+MR (CD11c-ATM) cells were isolated from the gonadal adipose tissue of high-fat diet-induced obese C57BL/6J mice by fluorescence-activated cell sorting. Gene expression was determined by quantitative real-time RT-PCR and quantified by the ΔΔCt method using Ubiquitin C as the housekeeping gene. Diagrams show mRNA expression levels of indicated genes in mean±s.e.m. related to MR-ATM (n=6). Statistical significances are indicated by letters above columns: identical letters indicate no statistically significant differences between respective groups (P>0.05).

Adipocytes upregulate chemokine expression upon ER stress

Chemokines can be produced by a wide variety of cells. Considering possible triggers of obesity-induced macrophage infiltration, stressed adipocytes may be a reasonable source of chemoattractants initiating AT inflammation. To test this hypothesis, we induced ER stress in differentiated 3T3-L1 adipocytes by incubation with calcimycin, tunicamycin and thapsigargin. Profound ER stress induction was confirmed by elevated expression of Grp78 (Hspa5; data not shown). ER-stressed adipocytes strongly upregulated expression of Ccl2, Ccl3, Ccl7 and Ccl19 after 8 h and that of Ccl5, Ccl9 and Ccl19 after 16 h of stress induction (Figure 7 and Supplementary Figure 2). Interestingly, Ccl8, Ccl11 and Cx3cl1 were downregulated by ER stress (Figure 7, Supplementary Figure 2). Ccl25 expression was not significantly altered by ER stress (Supplementary Figure 2). Thus, expression of ligands for Ccr2, Ccr3 and Ccr5 present on DN-ATM and CD11c-ATM, as well as Ccr7 present on CD11c-ATM were upregulated by stressed adipocytes. Altogether, chemokine expression data indicate that stressed adipocytes and macrophages, in particular MR-ATMs, may cooperate in monocyte/macrophage chemoattraction. Figure 8 provides a schematic overview on chemokine receptor and respective chemokine expression in ATMs and stressed adipocytes.

Figure 7

Regulation of C-C chemokine expression upon ER-stress in adipocytes. In 3T3-L1 adipocytes, ER stress was induced by incubation with 1 μM calcimycin (calci), 2.5 μg ml−1 tunicamycin (tunic) and 2.5 μM thapsigargin (thaps) for 8 and 16 h as indicated. Gene expression was determined by quantitative real-time RT-PCR and quantified by the ΔΔCt method using Ubiquitin C as the housekeeping gene. Diagrams show mRNA expression levels related to the solvent control (solv) of indicated genes in mean±s.e.m. (n=6). *P0.05; **P<0.01; ***P<0.001 compared with respective solvent control.

Figure 8

Overview on ATM populations and chemokine receptor/chemokine expression in obese murine adipose tissue. Round blue cells depict MR-ATMs, jagged blue cells DN-ATMs and jagged red cells CD11c-ATMs. Shared chemokine receptors on DN-ATMs and CD11c-ATMs, as well as respective ligands (chemokines) are depicted in green; Ccr7/Ccl19 and Ccr9/Ccl25 receptor/ligand pairs in orange and violet, respectively. For further explanation, see main text.


In this study, we describe the existence of three distinct ATM populations that exhibit entirely different profiles of M1 and M2 markers, chemokine and chemokine receptor expression. Although the selective expression of several M1 markers in CD11c-ATM corroborates a role in inflammatory and putatively detrimental responses as suggested previously,9, 13, 29 our results on two CD11c-negative populations (MR-ATM and DN-ATM) provide unexpected novel findings. First, the population discovered in this study that is double negative for CD11c and MR (DN-ATM) seems to represent a beneficial M2-like phenotype putatively involved in tissue repair according to its arginase (Arg1) and YM1 (Chi3l3) expression.10, 11 However, the expression of the IL-1β gene in these cells also suggests some proinflammatory potential. The large increase of their abundance in obesity together with profound chemokine receptor expression indicate that DN-ATMs are immigrated cells, as assumed for CD11c-ATMs.9 Second, although the MR is generally regarded as a marker for alternative (M2) activation of macrophages, our data suggest MR-ATMs to be regarded as quiescent rather than alternatively activated cells. MR-ATMs could represent resident macrophages, which may have a role for attracting other immune cells.

We show in this study that upon ER stress, adipocytes express various chemokines such as Ccl-2, -3, -5, -9 (Figure 7), for which CD11c-ATMs and DN-ATMs express respective receptors (Ccr1, -2, -3, -5; Figure 5). In accordance with chemokine induction in stressed adipocytes, CD11c-ATMs and DN-ATMs are enriched in crown-like structures (Figure 3). These structures consist of an accumulation of macrophages around stressed, dying adipocytes8 and have previously been shown to be enriched with CD11c-ATMs.12 Whether there are variations of crown-like structures as indicated by different enrichment with CD11c-ATMs and DN-ATMs, respectively (Figure 3c), remains to be elucidated. Interestingly, several chemokines that are downregulated in adipocytes upon ER stress (such as Ccl8, Ccl11 and Cx3cl1) are highly expressed in MR-ATMs (Figures 6 and 7; Supplementary Figures 1 and 2), indicating that MR-ATMs and adipocytes cooperate in attracting immune cells to the AT (see also Figure 8).

Elucidation of mechanisms for attraction of monocytes/macrophages to the AT is crucial for future pharmacological interference with obesity-induced inflammation and development of related disorders such as type 2 diabetes. One of the best-described chemokine/receptor pairs in AT inflammation is Mcp-1/Ccr2. Although several studies corroborate our result (Figure 5) that Mcp-1 (Ccl2) is not essentially involved in the attraction of ATMs,22, 23 prevention of obesity-induced AT inflammation and associated deterioration of metabolic parameters has been reported in Ccr2-knockout animals25 and by pharmacological blockade of Ccr2.26, 30 Although blockade of Ccr2 could be a promising approach for treatment of obesity-induced inflammation and insulin resistance according to these publications by reduction of CD11c-ATM, a simultaneous lack of DN-ATM may likely occur. Owing to their proposed function in tissue repair, a lack of DN-ATMs could counteract beneficial effects of Ccr2 blockade in the obese AT. Similar considerations would apply for blockade of other chemokine receptors common to CD11c-ATMs and DN-ATMs such as Ccr3, Ccr5 or Cx3cr1 (Figure 4), as well as for promising target chemokines, such as for example, the major Ccr5 ligand RANTES (Ccl5), the expression of which is also increased in human obesity.16, 18, 31 Hence, successful targeting of inflammatory ATM accumulation in the obese AT requires elucidation of specific mechanisms of attraction for the inflammatory CD11c-ATMs. However, it should be noted that the rate of apoptosis32 as well as local proliferation and (re-)differentiation may be factors determining the number of ATM and the proportions of their sub-population and, therefore, are worthwhile to be studied.

Selectivity for attracting the inflammatory CD11c-ATM could according to our data be achieved by the Ccl19/Ccr7 ligand/receptor pair. We found the gene for Ccl19 expressed in MR-ATMs, DN-ATMs and in tunicamycin-stressed adipocytes. As shown in Figure 4, Ccr7 is specifically expressed in CD11c-ATMs and thus might be a possible target for blocking attraction of this particularly proinflammatory population. Moreover, Ccr7 expression is more pronounced in the inflammation-prone and metabolically more relevant visceral compared with the subcutaneous AT.33 DN-ATMs, on the other hand, predominantly express Ccr9 (Figure 4), as well as the gene for its ligand Ccl25 (Figure 6). As DN-ATMs could promote accumulation of other DN-ATMs by Ccl25, but also of CD11c-ATMs through Ccl19 (Figure 6), DN-ATMs could be important for balancing the immigration of proinflammatory and anti-inflammatory ATMs in obesity.

Not only CD11c but also Ccr7 are markers of dendritic cells, the prototypical function of which is the stimulation of naive T cells. However, the absence of T-cell costimulatory receptors and CD1d in murine CD11c-ATMs (data not shown) as in human ATMs14 suggest designate designation of CD11c-ATMs as macrophages rather than dendritic cells, despite recent evidence for a role of T cells in the obese AT, which implicates an involvement of antigen-presenting cells.34, 35, 36 In addition, DN-ATMs selectively express a gene for a dendritic cell chemokine receptor, namely Ccr9 (Figure 5), which is a marker of murine tolerogenic, suppressive plasmacytoid dendritic cells.37 Thus, together with the expression of M2 marker genes, Ccr9 expression indicates that DN-ATMs are of an anti-inflammatory phenotype.

Owing to their high abundance even in lean animals, their low expression of chemokine receptors, and their widely distributed localization in the AT as compared with the other ATM types (Figure 3), MR-ATMs may constitute resident macrophages. This notion is supported by results of a study with Ccr2-negative ATMs generated by bone marrow transplantation in genetically obese (ob/ob) mice.26 In these mice, the number of ATMs and their expression of mRNA for CD11c was reduced compared with wt ATMs, whereas MR expression was unchanged.26 Notably, the decrease in the percentage of MR-ATMs (Figure 1a) was not due to a decrease in absolute MR-ATM numbers, which were moderately increased upon obesity in our studies (1.6±0.4-fold, as assessed by the proportion within total stromal-vascular cells, not shown), as recently corroborated by gene expression data in another study.38 MR-ATMs also express MGL-1 on their surface (data not shown), a recently described marker for resident ATMs.12 However, MGL-1+ ATMs highly express Il10 and Arg1,12 the latter of which is not substantially expressed in MR-ATM (Figure 2 and Fujisaka et al.38) indicating that MGL-1 is not a selective marker for the MR+CD11c (MR-ATM) population. Hence, further research is warranted to characterize these putatively resident ATMs with respect to surface phenotype and function.

In conclusion, by investigating ATM populations defined by CD11c and MR surface expression, we show in this study that not only detrimental inflammatory but also a so far undescribed alternatively activated and putatively beneficial ATM population is increased in obesity. Chemokine and chemokine receptor expression analyses revealed a complex network of attraction mechanisms. As illustrated in a schematic overview in Figure 8, our conclusive hypothesis is, briefly, that upon obesity-induced ER stress dying adipocytes produce various chemokines that attract macrophages (DN-ATMs) competent for tissue repair. This attraction is supported by resident macrophages (MR-ATMs), which may be capable of removing glycoproteins and debris from their surroundings but are otherwise relatively quiescent. Possibly as a ‘side effect’ of DN-ATM attraction that may be important for tissue homeostasis, also CD11c-ATMs are attracted, which by release of inflammatory factors promote the vicious circle of obesity-mediated AT inflammation with its detrimental consequences on insulin sensitivity.

The identification of chemokine/chemokine receptor interactions selective for inflammatory ATMs could lead to the development of anti-inflammatory therapeutic strategies for preventing obesity-associated metabolic and cardiovascular complications.


  1. 1

    de Luca C, Olefsky JM . Inflammation and insulin resistance. FEBS Lett 2008; 582: 97–105.

    CAS  Article  Google Scholar 

  2. 2

    Fain JN . Release of interleukins and other inflammatory cytokines by human adipose tissue is enhanced in obesity and primarily due to the nonfat cells. Vitam Horm 2006; 74: 443–477.

    CAS  Article  Google Scholar 

  3. 3

    Zeyda M, Stulnig TM . Adipose tissue macrophages. Immunol Lett 2007; 112: 61–67.

    CAS  Article  Google Scholar 

  4. 4

    Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante Jr AW . Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest 2003; 112: 1796–1808.

    CAS  Article  Google Scholar 

  5. 5

    Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ et al. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest 2003; 112: 1821–1830.

    CAS  Article  Google Scholar 

  6. 6

    Curat CA, Wegner V, Sengenes C, Miranville A, Tonus C, Busse R et al. Macrophages in human visceral adipose tissue: increased accumulation in obesity and a source of resistin and visfatin. Diabetologia 2006; 49: 744–747.

    CAS  Article  Google Scholar 

  7. 7

    Ozcan U, Cao Q, Yilmaz E, Lee AH, Iwakoshi NN, Ozdelen E et al. Endoplasmic reticulum stress links obesity, insulin action, and type 2 diabetes. Science 2004; 306: 457–461.

    Article  Google Scholar 

  8. 8

    Cinti S, Mitchell G, Barbatelli G, Murano I, Ceresi E, Faloia E et al. Adipocyte death defines macrophage localization and function in adipose tissue of obese mice and humans. J Lipid Res 2005; 46: 2347–2355.

    CAS  Article  Google Scholar 

  9. 9

    Lumeng CN, Bodzin JL, Saltiel AR . Obesity induces a phenotypic switch in adipose tissue macrophage polarization. J Clin Invest 2007; 117: 175–184.

    CAS  Article  Google Scholar 

  10. 10

    Gordon S . Alternative activation of macrophages. Nat Rev Immunol 2003; 3: 23–35.

    CAS  Article  Google Scholar 

  11. 11

    Mantovani A, Sozzani S, Locati M, Allavena P, Sica A . Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol 2002; 23: 549–555.

    CAS  Article  Google Scholar 

  12. 12

    Lumeng CN, Delproposto JB, Westcott DJ, Saltiel AR . Phenotypic switching of adipose tissue macrophages with obesity is generated by spatiotemporal differences in macrophage subtypes. Diabetes 2008; 57: 3239–3246.

    CAS  Article  Google Scholar 

  13. 13

    Nguyen MT, Favelyukis S, Nguyen AK, Reichart DD, Scott PA, Jenn A et al. A subpopulation of macrophages infiltrates hypertrophic adipose tissue and is activated by free fatty acids via Toll-like receptors 2 and 4 and JNK-dependent pathways. J Biol Chem 2007; 282: 35279–35292.

    CAS  Article  Google Scholar 

  14. 14

    Zeyda M, Farmer D, Todoric J, Aszmann O, Speiser M, Gyori G et al. Human adipose tissue macrophages are of an anti-inflammatory phenotype but capable of excessive pro-inflammatory mediator production. Int J Obes (Lond) 2007; 31: 1420–1428.

    CAS  Article  Google Scholar 

  15. 15

    Bourlier V, Zakaroff-Girard A, Miranville A, De Barros S, Maumus M, Sengenes C et al. Remodeling phenotype of human subcutaneous adipose tissue macrophages. Circulation 2008; 117: 806–815.

    CAS  Article  Google Scholar 

  16. 16

    Huber J, Kiefer FW, Zeyda M, Ludvik B, Silberhumer GR, Prager G et al. CC chemokine and CC chemokine receptor profiles in visceral and subcutaneous adipose tissue are altered in human obesity. J Clin Endocrinol Metab 2008; 93: 3215–3221.

    CAS  Article  Google Scholar 

  17. 17

    Moraes RC, Blondet A, Birkenkamp-Demtroeder K, Tirard J, Orntoft TF, Gertler A et al. Study of the alteration of gene expression in adipose tissue of diet-induced obese mice by microarray and reverse transcription-polymerase chain reaction analyses. Endocrinology 2003; 144: 4773–4782.

    CAS  Article  Google Scholar 

  18. 18

    Wu H, Ghosh S, Perrard XD, Feng L, Garcia GE, Perrard JL et al. T-cell accumulation and regulated on activation, normal T cell expressed and secreted upregulation in adipose tissue in obesity. Circulation 2007; 115: 1029–1038.

    CAS  Article  Google Scholar 

  19. 19

    Kamei N, Tobe K, Suzuki R, Ohsugi M, Watanabe T, Kubota N et al. Overexpression of monocyte chemoattractant protein-1 in adipose tissues causes macrophage recruitment and insulin resistance. J Biol Chem 2006; 281: 26602–26614.

    CAS  Article  Google Scholar 

  20. 20

    Kanda H, Tateya S, Tamori Y, Kotani K, Hiasa KI, Kitazawa R et al. MCP-1 contributes to macrophage infiltration into adipose tissue, insulin resistance, and hepatic steatosis in obesity. J Clin Invest 2006; 116: 1494–1505.

    CAS  Article  Google Scholar 

  21. 21

    Chow FY, Nikolic-Paterson DJ, Ma FY, Ozols E, Rollins BJ, Tesch GH . Monocyte chemoattractant protein-1-induced tissue inflammation is critical for the development of renal injury but not type 2 diabetes in obese db/db mice. Diabetologia 2007; 50: 471–480.

    CAS  Article  Google Scholar 

  22. 22

    Inouye KE, Shi H, Howard JK, Daly CH, Lord GM, Rollins BJ et al. Absence of CC chemokine ligand 2 does not limit obesity-associated infiltration of macrophages into adipose tissue. Diabetes 2007; 56: 2242–2250.

    CAS  Article  Google Scholar 

  23. 23

    Kirk EA, Sagawa ZK, McDonald TO, O’Brien KD, Heinecke JW . Macrophage chemoattractant protein-1 deficiency fails to restrain macrophage infiltration into adipose tissue. Diabetes 2008; 57: 1254–1261.

    CAS  Article  Google Scholar 

  24. 24

    Lumeng CN, Deyoung SM, Bodzin JL, Saltiel AR . Increased inflammatory properties of adipose tissue macrophages recruited during diet-induced obesity. Diabetes 2007; 56: 16–23.

    CAS  Article  Google Scholar 

  25. 25

    Weisberg SP, Hunter D, Huber R, Lemieux J, Slaymaker S, Vaddi K et al. CCR2 modulates inflammatory and metabolic effects of high-fat feeding. J Clin Invest 2006; 116: 115–124.

    CAS  Article  Google Scholar 

  26. 26

    Ito A, Suganami T, Yamauchi A, Degawa-Yamauchi M, Tanaka M, Kouyama R et al. Role of CC chemokine receptor 2 in bone marrow cells in the recruitment of macrophages into obese adipose tissue. J Biol Chem 2008; 283: 35715–35723.

    CAS  Article  Google Scholar 

  27. 27

    Kiefer FW, Zeyda M, Todoric J, Huber J, Geyeregger R, Weichhart T et al. Osteopontin expression in human and murine obesity: extensive local up-regulation in adipose tissue but minimal systemic alterations. Endocrinology 2008; 149: 1350–1357.

    CAS  Article  Google Scholar 

  28. 28

    Nomiyama T, Perez-Tilve D, Ogawa D, Gizard F, Zhao Y, Heywood EB et al. Osteopontin mediates obesity-induced adipose tissue macrophage infiltration and insulin resistance in mice. J Clin Invest 2007; 117: 2877–2888.

    CAS  Article  Google Scholar 

  29. 29

    Patsouris D, Li PP, Thapar D, Chapman J, Olefsky JM, Neels JG . Ablation of CD11c-positive cells normalizes insulin sensitivity in obese insulin resistant animals. Cell Metab 2008; 8: 301–309.

    CAS  Article  Google Scholar 

  30. 30

    Yang SJ, Iglayreger HB, Kadouh HC, Bodary PF . Inhibition of the chemokine (C-C motif) ligand 2/chemokine (C-C motif) receptor 2 pathway attenuates hyperglycaemia and inflammation in a mouse model of hepatic steatosis and lipoatrophy. Diabetologia 2009; 52: 972–981.

    CAS  Article  Google Scholar 

  31. 31

    Maury E, Ehala-Aleksejev K, Guiot Y, Detry R, Vandenhooft A, Brichard SM . Adipokines oversecreted by omental adipose tissue in human obesity. Am J Physiol 2007; 293: E656–E665.

    CAS  Google Scholar 

  32. 32

    Kiefer FW, Zeyda M, Gollinger K, Pfau B, Neuhofer A, Weichhart T et al. Neutralization of osteopontin inhibits obesity-induced inflammation and insulin resistance. Diabetes 2010; 59: 935–946.

    CAS  Article  Google Scholar 

  33. 33

    Lee HS, Park JH, Kang JH, Kawada T, Yu R, Han IS . Chemokine and chemokine receptor gene expression in the mesenteric adipose tissue of KKAy mice. Cytokine 2009; 46: 160–165.

    CAS  Article  Google Scholar 

  34. 34

    Winer S, Chan Y, Paltser G, Truong D, Tsui H, Bahrami J et al. Normalization of obesity-associated insulin resistance through immunotherapy. Nat Med 2009; 15: 921–929.

    CAS  Article  Google Scholar 

  35. 35

    Feuerer M, Herrero L, Cipolletta D, Naaz A, Wong J, Nayer A et al. Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nat Med 2009; 15: 930–939.

    CAS  Article  Google Scholar 

  36. 36

    Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M et al. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nat Med 2009; 15: 914–920.

    CAS  Article  Google Scholar 

  37. 37

    Hadeiba H, Sato T, Habtezion A, Oderup C, Pan J, Butcher EC . CCR9 expression defines tolerogenic plasmacytoid dendritic cells able to suppress acute graft-versus-host disease. Nat Immunol 2008; 9: 1253–1260.

    CAS  Article  Google Scholar 

  38. 38

    Fujisaka S, Usui I, Bukhari A, Ikutani M, Oya T, Kanatani Y et al. Regulatory mechanisms for adipose tissue M1 and M2 macrophages in diet-induced obese mice. Diabetes 2009; 58: 2574–2582.

    CAS  Article  Google Scholar 

Download references


This work was supported by the Austrian Science Fund (project no. P18776-B11 and as part of the CCHD doctoral program W1205-B09), and the European Community's Seventh Framework Programme (FP7/2007–2013) under grant agreement no. 201608 (TOBI—Targeting OBesity-driven Inflammation; all to TMS).

Author information



Corresponding author

Correspondence to T M Stulnig.

Ethics declarations

Competing interests

The authors declare no conflict of interests

Additional information

Supplementary Information accompanies the paper on International Journal of Obesity website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Zeyda, M., Gollinger, K., Kriehuber, E. et al. Newly identified adipose tissue macrophage populations in obesity with distinct chemokine and chemokine receptor expression. Int J Obes 34, 1684–1694 (2010).

Download citation


  • adipose tissue inflammation
  • macrophages
  • chemokines
  • chemokine receptors

Further reading