Obesity has been linked to cardiovascular disease, hypertension, diabetes and the metabolic syndrome, with elevated markers of systemic inflammation. Intercellular adhesion molecule-1 (ICAM-1) is a transmembrane adhesion molecule involved in leukocyte migration to sites of inflammation. In human obesity, elevated expression of the soluble form of ICAM-1 (sICAM-1) is positively correlated with abdominal fat deposition. Increases in adiposity have also been correlated with macrophage infiltration into adipose tissue. Here we investigate adipose tissue production and transcriptional regulation of ICAM-1 in a mouse model of dietary obesity. After feeding mice a high-fat diet, ICAM-1 expression in serum and adipose tissue was analyzed by ELISA, Northern blotting, real-time quantitative PCR, and flow cytometry. After 6 mo on the high-fat diet, sICAM-1 levels significantly correlated with body weight and abdominal fat mass. ICAM-1 mRNA was expressed in adipose tissue of mice, with significantly higher levels in males than females. After only 3 wk, there were adipose tissue-specific increases in mRNAs for ICAM-1, IL-6, and monocyte chemoattractant protein-1 (MCP-1) in male mice. Analysis of the stromal-vascular fraction of male adipose tissue revealed CD11b-negative cells with increased surface ICAM-1 and CD34. We also found two populations of F4/80+, CD11b+, ICAM-1+ cells, one of which also expressed CD14 and CD11c and was increased in response to a high-fat diet. These results indicate that within 3 wk on a high-fat diet, male mice exhibited significant increases in pro-inflammatory factors and immune cell infiltration in adipose tissue that may represent links between obesity and its associated inflammatory complications.
intercellular adhesion molecule-1/CD54 (ICAM-1) is a member of the immunoglobulin superfamily (37, 40) and is expressed on a wide variety of cells constitutively or under conditions of inflammation (13). One well-documented function is as a leukocyte adhesion receptor that, in response to inflammatory stimuli, is typically expressed on the surface of endothelial cells, where it mediates the migration of leukocytes into tissues (21). ICAM-1 is also expressed under inflammatory conditions on a variety of parenchymal cells [e.g., hepatocytes (16) and cardiac myocytes (39)] and supports cytotoxic adhesive interactions with leukocytes (14). In addition, ICAM-1 contributes to adaptive immunity, serving as an accessory adhesion molecule on antigen presenting cells (27). A soluble form of ICAM-1 (sICAM-1) is found to enter the systemic circulation, as a result of proteolytic cleavage at the cell surface, releasing the extracellular domain (44). sICAM-1 is considered one of the prototypic markers of inflammation, and its expression is elevated in a broad array of disease states, including bacterial sepsis (3), type II diabetes (19), pre-eclampsia (11), and atherosclerosis (5). sICAM-1 levels are also increased in response to a “Western-style” diet, i.e., one that is either high in atherogenic lipids (5, 26) or hypercholesterolemic (35). Several studies in humans have shown that sICAM-1 levels are also elevated in obesity, and are positively correlated with central adiposity (28, 43, 51) and insulin resistance (23, 41). The function of sICAM-1 is unclear, but there is evidence that it may play a modulating role in inflammation (18).
Since ICAM-1 is involved in leukocyte migration and can be cleaved to produce sICAM-1, we sought to investigate ICAM-1 production in adipose tissue in a murine model of obesity. We examined sICAM-1 production in response to a high-fat diet, ICAM-1 and chemokine expression in adipose tissue, and changes in adipose macrophage populations. We demonstrate sex-dependent increases in ICAM-1 expression, specific to adipose tissue, as well as changes in discrete populations of adipose tissue macrophages as early as 3 wk after feeding mice a high-fat diet. The increase in ICAM-1 expression in adipose tissue, as well as the accumulation of specific adipose tissue macrophage subsets early in the onset of obesity, may be linked to complications, such as type II diabetes and atherosclerosis, which are increasingly seen as having inflammatory characteristics.
Of interest to the current study is the finding that macrophages accumulate in adipose tissue of animals on a high-fat diet in a sex-dependent fashion (32, 50). In those studies, macrophages in adipose tissue were primarily identified by expression of F4/80 (22). In the current study, we have used additional immune and stem cell markers to define characteristics of the stromal-vascular (S-V) cell fraction of adipose tissue. A marker of hematopoietic origin, CD11b, is of interest since it binds ICAM-1. CD14, another commonly used marker for monocytes and macrophages, binds bacterial LPS and is expressed on cells of monocytic lineage and is downregulated on cells that develop into mature myeloid dendritic cells (34, 52). CD11c is often considered a hallmark of dendritic cells (8), however, several reports have shown that some macrophage subsets can also express CD11c (20, 31). Using this panel of markers, we have identified for the first time increases in CD11c-positive macrophages in adipose tissue, in response to a high-fat diet.
Animal protocols were approved by the Baylor College of Medicine Institutional Animal Care and Use Committee, and mice were cared for according to the NIH Guide for the Care and Use of Laboratory Animals. All experiments utilized C57Bl/6 mice that had been weaned at 3 wk of age, housed at two to four per cage, and given food and water ad libitum. Mice were fed a high-fat diet (no. 112734; Dyets Inc, Bethlehem, PA) containing 21% fat (milk fat), beginning at 6 wk of age, or a standard chow (no. 5053; LabDiet, St. Louis, MO) containing 4.5% fat. Mice were maintained on the diet for 24 wk for long term studies, or 3 wk for short-term studies. Mice were then anesthetized, weighed, and blood samples, obtained by cardiac puncture, were collected into nonheparinized or EDTA-treated tubes for an automated, complete blood count (Center for Comparative Medicine, Baylor College of Medicine, Houston, TX). Nonheparinized blood samples were allowed to clot, were centrifuged, and serum was removed for ELISA analysis. Perigonadal fat pads (epididymal fat pad in males and periovarian plus periuterine fat pads in females) were removed, weighed, and either frozen in liquid nitrogen for the preparation of RNA, or digested with collagenase for isolation of adipocytes and S-V cells.
sICAM-1 was analyzed in serum using a mouse sICAM-1 ELISA kit (Endogen Pierce, Woburn, MA). Briefly, 50 μl of a 1:100 dilution of serum was incubated on the ELISA plate for 2 h, followed by washing and incubation with secondary reagents as supplied. Plates were read using a Spectramax Plus Microplate Spectrophotometer and Softmax Pro analysis software (Molecular Devices, Sunnyvale, CA).
Northern blot analysis.
Adipose tissue was homogenized in TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA) and RNA prepared according to the manufacturer's instructions, including the optional centrifugation for removal of excess lipid. Standard Northern blot analysis was performed using 20 μg of RNA per lane and full-length mouse ICAM-1 cDNA as the hybridization probe. Results were quantified using a Storm 860 (with ImageQuant 4.2A software) phosphor image analysis system (Molecular Dynamics, Sunnyvale, CA).
Real-time quantitative PCR.
Adipose tissue, spleen, and lung samples were homogenized in QIAzol lysis reagent, and total RNA was isolated with the RNeasy Mini kit (Qiagen, Valencia, CA). Quality of RNA was verified by agarose gel electrophoresis. Synthesis of cDNA was performed with the 1st Strand cDNA Synthesis Kit for RT-PCR (AMV) (Roche, Indianapolis, IN), using 500 ng of total RNA and random hexamer primers per manufacturer's recommended concentrations. For the amplification of ICAM-1, IL-6, TNF, CDH5 (VE-cadherin gene), and monocyte chemoattractant protein-1 (MCP-1), 1 μl of cDNA was added to corresponding 20 × TaqMan MGB probe-primer sets for each message, multiplexed with primers for ribosomal 18S and 2× TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA). PCR was performed in a 7500 Real Time PCR System (Applied Biosystems, Foster City, CA) using the manufacturer's suggested thermal settings: one cycle of 2 min at 50°C followed by 10 min at 95°C, and 40 cycles of 15 s at 95°C, 60 s at 60°C. Relative mRNA expression was calculated by comparative Ct-method. Ribosomal 18S RNA was used as the endogenous control. Mice fed the control diet were used as the calibrator and set to 100%.
Isolation of adipose tissue S-V cell fractions.
Perigonadal adipose tissue was excised, minced and placed in 3 ml of Krebs-Ringer bicarbonate buffer (KRB) per gram of adipose tissue. The KRB contained 10 mM glucose and 4% BSA plus 840 U/g collagenase Type I (Worthington Chemicals, Lakewood, NJ). The solution was gassed with 5% CO2 and was incubated at 37°C, with gentle agitation for 40 min. The tissue slurry was then filtered through chiffon mesh into a 50-ml tube and rinsed with KRB. The lipid-laden adipocytes floated to the surface, and the underlying infranatant fraction was removed. The infranatant fraction was centrifuged at 500g to pellet the S-V cells, which were then resuspended and washed 2× with 30 ml of PBS with 10 mM glucose. Cells were then resuspended in 500 μl of PBS plus glucose for further analysis.
The S-V cell fractions of five mice were pooled, and either incubated with antibody at 4°C in the dark for 30 min as a whole population, or separated by CD11b expression by magnetic cell sorting (MACS) (Miltenyi Biotec, Auburn, CA) per manufacturer's instructions. In short, cells were incubated for 15 min at 4°C with a buffer containing PBS, with 2 mM EDTA and 0.5% BSA, and 10 μl of magnetically labeled anti-CD11b MicroBeads. Cells were then washed with 1 ml of buffer, centrifuged at 300 g for 10 min, resuspended in 50 μl, and separated by MACS MS column (Miltenyi Biotec, Auburn, CA). Cells were then washed and stained in PBS with glucose, with the following antibodies for 30 min at 4°C: ICAM-1 clone, YN1/1.7.4; CD34 clone, RAM34; CD14 clone, Sa2–8 (eBioscience, San Diego, CA); F4/80 clone, CI:A3–1 (Serotec, Raleigh, NC); CD11b clone, M1/70; and CD11c clone, HL3 (BD Biosciences, San Jose, CA). Cells were incubated with nonspecific IgG to assess background fluorescence (BD Bioscience, San Jose, CA). Flow cytometry was performed using a BD-FACScan (Becton-Dickinson, San Jose, CA).
Results are expressed as means ± SE. Regression analysis techniques were used to assess the relationship between weight and abdominal fat (independent variables) with sICAM-1 (dependent variable). The effect of sex on this relationship was assessed by including an interaction term. Two-way analysis of variance was used to assess the effects of diet, sex, and their interaction on abdominal fat, weight, and sICAM-1. Male to female, or control to high-fat comparisons of real-time PCR, as well as mean fluorescence of ICAM-1, were compared using Student's t-test.
sICAM-1 levels were analyzed in sera of mice that were obese due to maintenance on a long-term (6 mo), high-fat diet. The average weights were 48.7 ± 1.0 g for males and 44.1 ± 1.8 g for females. Overall mean sICAM-1 levels were higher in the male mice (50.5 ± 2.8 μg/ml) than in females (42.2 ± 2.6 μg/ml) (P = 0.05). When related to body weight, sICAM-1 levels increased an average of 10 μg/ml for each 10 g increase in body weight, with correlation r = 0.50 (P < 0.001). This relationship did not differ statistically between males and females (Fig. 1A). When related to abdominal fat pad weight, which is an index of adiposity, sICAM-1 levels were also positively correlated with increasing amounts of abdominal adipose tissue (r = 0.40) (Fig. 1B), with an average increase in sICAM-1 of 4.6 μg/ml per gram increase in abdominal fat pad weight (P = 0.03). Slopes did not differ significantly between males and females (P = 0.60), but the mean sICAM-1 level for males was significantly increased compared with females (P = 0.003). Others have reported (25) and we have confirmed that C57BL/6 mice do not spontaneously increase sICAM-1 levels between 3 and 10 mo of age, a time frame inclusive of the ages of those mice with dietary obesity. These results demonstrate that in an animal model of obesity, as in humans, there is an association between sICAM-1 levels and abdominal adiposity. We sought to determine if adipose tissue is a potential source of sICAM-1. Northern blot analysis showed that ICAM-1 mRNA was expressed in the adipose tissue of both male and female mice after 6 mo on a high-fat diet (Fig. 1C).
To evaluate early events in adipose tissue of animals on a high-fat diet, male and female mice were fed a short-term (3 wk), high-fat diet and compared with mice maintained on a standard chow diet. Mice fed the high-fat diet showed significant increases in body weight (Fig. 2A) and abdominal fat pad weight (Fig. 2B), with males accumulating slightly more fat. However, the ratio between abdominal fat mass and body weight remained the same between male and female mice (Fig. 2C). Comparing mice fed the control diet to mice fed a high-fat, short-term diet, automated WBC analysis revealed an increase in whole blood monocytes from 0.061 × 103 ± 0.007 cells/μl to 0.089 × 103 ± 0.012 cells/μl, respectively (P < 0.05, n = 10). The high-fat diet also resulted in a decrease in the overall percentage of lymphocytes from 81.79% ± 0.53 to 77.87% ± 1.67 (P < 0.02, n = 10). Real-time quantitative PCR of total RNA taken from adipose tissue showed that ICAM-1 mRNA expression is increased over 3.5 fold in male mice fed the high-fat diet (P < 0.001), compared with mice fed standard chow (Fig. 2D). Female mice fed the high-fat diet did not show a significant increase in ICAM-1 message levels. These results show that male mice not only have higher ICAM-1 expression after long-term, high-fat feeding, but males appear to show earlier responses than females. Furthermore, increases in ICAM-1 expression in the males appears at this early time to be limited to adipose tissue, since elevated expression was not found in either the spleen or liver (Fig. 2E).
Increases in ICAM-1 message seen in the adipose tissue of male mice fed a short-term, high-fat diet were then compared with other pro-inflammatory cytokines. As noted in Fig. 3A, male mice, but not female mice, exhibited marked upregulation of MCP-1 in adipose tissue. Studies were repeated in male adipose tissue and revealed that ICAM-1, MCP-1, and IL-6 message levels significantly increased (P < 0.01), while no significant increase was seen in TNF message levels (Fig. 3B). No significant increase in message was seen in the expression of VE-cadherin (mouse CDH5 gene), a cell surface protein expressed on vascular endothelial cells.
We then isolated the S-V compartment of male adipose tissue in an attempt to localize and quantify cell surface expression of ICAM-1. Three regions of clustered cells were revealed by flow cytometric analysis of forward vs. side scatter patterns, arbitrarily labeled regions 1–3 (Fig. 4A). Region 1 contained lymphocytes consistent with the previously published results (7, 32). ICAM-1 expression in this region was low, compared with cells in regions 2 and 3 (Fig. 4B). Only region 3 revealed a significant increase in cell-associated ICAM-1 after a short-term, high-fat diet (P < 0.001). Region 2 contained 10% of the total cells within the 3 regions, and most of these cells expressed CD11b and ICAM-1 (Fig. 4C). Region 2 showed a small decreasing trend in the number of CD11b positive and ICAM-1 positive cells while region 3 showed significant increases. Region 3 contained ∼45–50% of the total S-V fraction in both the control adipose tissue and adipose tissue from animals on high-fat diet. This region contained a heterogeneous population of CD11b-positive and -negative cells (Fig. 4D). Feeding animals a high-fat diet increased the proportion of ICAM-1/CD11b-positive cells in region 3 from 31.5% ± 2.8 to 58.6% ± 3.1. The mean anti-ICAM-1 fluorescence of CD11b-positive cells was not increased (Fig. 4E), indicating that increases in ICAM-1 message, as detected by real-time PCR, may be accounted for in part by an increased proportion of ICAM-1/CD11b-positive cells. In contrast, the CD11b-negative population of cells in region 3 revealed increased binding of anti-ICAM-1, indicating possible upregulation of ICAM-1 expression in this nonhematopoietic cell population.
To further characterize the S-V compartment, isolated cells were separated into CD11b-negative and -positive fractions (Fig. 5A). Compared with the total cell count, we detected 2.5% ± 0.45 cells in region 2 and 27.0% ± 2.5 cells in region 3 remaining in the CD11b-negative fraction (Fig. 5B). CD11b-positive cells in region 2 and region 3 showed marked differences in cell surface markers (Fig. 5C). Region 2 cells were CD11c negative, CD14 low and F4/80 positive, while region 3 cells were CD11c positive, CD14 positive and F4/80 positive. These relative expression patterns of cell surface markers did not change significantly between control mice and mice fed a high-fat diet (Fig. 5D). Taken together these data indicate that there are two distinct myeloid populations in adipose tissue. Only the CD11c/CD14-expressing population of region 3 increases in response to a high-fat diet.
Consistent with observations in Fig. 4E, the CD11b-negative fraction of region 3 increases the cell surface expression of ICAM-1 in response to the high-fat diet (Fig. 6A). There were too few cells in region 2 of the CD11b-negative fraction to quantify. We were unable to detect any endothelial cells in region 3 by anti-ICAM-2, or anti-VE-cadherin antibodies (data not shown). Interestingly, we found a cell surface protein expressed on many stem cell-like progenitor cells, CD34, expressed on 43.2% ± 2.6 of CD11b-negative cells (Fig. 6B), a portion of which coexpressed ICAM-1(Fig. 6C). The proportion of CD34-expressing cells increased to 50.1% ± 2.3, in response to a high-fat diet (Fig. 6, B and C).
Our results show that plasma sICAM-1 expression is increased in mice following 6 mo of a high-fat diet. These concentrations correlate with abdominal fat pad weight in both males and females. This is consistent with reports in humans in which increased levels of soluble ICAM-1 have been correlated with several pathological conditions, including obesity and obesity-related disorders (28, 43, 51). We show that adipose tissue, after 6 mo of a high-fat diet, expresses ICAM-1 mRNA, indicating that adipose tissue ICAM-1, via proteolytic cleavage (9, 33), may be a possible source of the circulating sICAM-1 that is seen in obesity. Comparing these data to a 3-wk, high-fat diet, we find that male mice express significantly more ICAM-1 and MCP-1 message than female mice and that this increase in ICAM-1 message was evident in adipose tissue, but not in spleen or liver at this early time point. Similarly, male mice show an increase in mRNA expression of IL-6, but not TNF. It is unclear why female mice, while still exhibiting a similar ratio of fat deposition after 3 wk on a high-fat diet, demonstrate a lesser cytokine response, but such intrinsic mechanisms may potentially underlie the sex-specific differences in fat deposition and susceptibility to obesity-related complications seen in humans.
The male-specific increase in ICAM-1 message seen, is reflected, at least in part, as an increase in cell surface protein expression on CD11b-negative cells of the S-V fraction. A population of these cells express CD34, and upregulate their CD34 expression in response to a short-term, high-fat diet. It is also recently thought that CD34, in addition to being expressed on hematopoietic progenitor cells (2), might be expressed on cells with adipocyte differentiation potential (36). We show for the first time in an in vivo model, a population of CD34+ cells in adipose tissue that increases their CD34 levels in response to a short-term, high-fat diet. Peripheral blood hematopoietic cells expressing CD34 are thought to potentially have an active role in immune function by producing cytokines such as MCP-1, RANTES, IL-1β, and IL-8 (45). These cells could potentially be a source of the MCP-1 seen at 3 wk of a high-fat diet, assisting in the local activation, differentiation, or recruitment of macrophages into adipose tissue.
In addition to the CD34+ cells, we identified two distinct populations of macrophages within adipose tissue that are F4/80+, CD11b+, and ICAM-1+. This is the first demonstration of relatively rapid changes in a specific population of macrophages in adipose tissue, corroborating longer term studies that have correlated increases in adiposity with increased F4/80+ macrophage infiltration into adipose tissue (32, 50, 53). Only the F4/80+ populations expressing CD11c and CD14 increase in response to a high-fat diet. This is the first time to the authors' knowledge that CD11c-expressing macrophages have been reported in adipose tissue. These CD11c-expressing cells may represent a unique inflammatory signaling subset of macrophages in adipose tissue.
Obesity and its complications exhibit many characteristics in common with chronic inflammation (47). Furthermore, evidence is rapidly accumulating that the links between adiposity and heightened immune cell activity are responsible for at least some manifestations of the metabolic syndrome, particularly atherosclerosis (1, 29). Soluble ICAM is associated with obesity in humans, insulin resistance and type 2 diabetes (41, 51). We show that similar to humans, mice on a long-term diet increase sICAM-1 levels. Transgenic mice which over express sICAM also show increased susceptibility to weight gain on a high-fat diet and these mice also demonstrate a decreased capacity to recruit neutrophils and macrophages into sites of inflammation (48). These results are similar to reports from ICAM-1 knockout mice (12, 38), however, these knockout mice have since been shown to express alternatively spliced forms of ICAM-1 that can be found in a soluble form (24, 46). These isoforms have been shown to have the capacity to act as accessory molecules in CD4+ T-cell signaling (33). Heightened levels of soluble forms of ICAM may play an immunoregulatory role in an obese state, leading to obesity-related disorders.
How excess body fat triggers inflammatory-like responses is not clear, but it is thought to involve elevated levels of cytokines, particularly TNF and IL-6 (54). Several cytokines such as TNF, IL-1, and IL-6 have been shown to upregulate expression of ICAM-1 (4, 30) and MCP-1 (15, 17). We demonstrate that at 3 wk, an increase in IL-6 and MCP-1 occurs before TNF upregulation, in response to a high-fat diet in adipose tissue. We also show an increase in the numbers of whole blood monocytes, consistent with reports that CD11b+ monocytes are increased in mice fed a 6-mo, high-fat diet, or after treatment with MCP-1 (42). ICAM-1 and MCP-1 both act to facilitate transmigration and chemotaxis of monocytes and macrophages into sites of inflammation. Knockout mouse models of CCR2, the receptor of MCP-1, have shown that CCR2 deficiency results in decreased macrophage infiltration into adipose tissue (49). Although both the adipocytes and the S-V cell fractions have the potential to secrete MCP-1 and IL-6 (10, 17), there is evidence that the primary source of these cytokines, in response to a high-fat diet, is the S-V fraction, possibly from macrophages and preadipocytes (6, 18).
Macrophages reside in all tissues and take up tissue-specific form and function. Adipose tissue contains macrophages and macrophage precursors, but their functions have not yet been defined. We have shown sex-specific inflammatory responses to a high-fat diet within adipose tissue. Here we have demonstrated an increase in the expression of cytokines and ICAM-1 in adipose tissue that may assist in the recruitment and signaling of macrophages in an obese state. In addition, our identification of a previously unrecognized macrophage subset within adipose tissue, suggests that these cells have important roles within an expanding fat mass. Elucidating these mechanisms is key to understanding adipose tissue biology, as well as links between inflammation and obesity-related complications.
This work was supported by grants from the United States Department of Agriculture 6250–51000-046–01A (to H. J. Mersmann and C. W. Smith), an NIH Training Grant in Pediatric Gastroenterology DK07664 (to R. L. Robker) and an Individual National Research Service Award DK60381 (to R. L. Robker) Partially supported through grant numbers DGE-0086397 and DGE-0440525 of the National Science Foundation (to D. K. Brake).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society