Markers of macrophage infiltration and measures of lipolysis in human abdominal adipose tissues

Authors


  • Disclosure: The authors have no competing interests.

Abstract

Objective

We tested the hypothesis that high lipolytic responsiveness is related to increased expression of ATM genes in human adipose tissues.

Design and Methods

Omental (OM) and subcutaneous (SC) fat samples were obtained surgically in 46 women (age: 47.2 ± 4.7 years, BMI: 26.9 ± 5.2 kg/m2). Body composition and fat distribution were measured using dual energy X-ray absorptiometry and computed tomography. Lipolysis was measured by glycerol release in mature adipocytes isolated by collagenase digestion under basal-, isoproterenol (10−5M)-, and forskolin (10−5M)-stimulated conditions. Quantification of macrophage gene mRNA expression (CD11b, CD11c, and CD68) in whole adipose tissue was performed using real-time RT-PCR.

Results

SC CD68 mRNA abundance was positively associated with isoproterenol-stimulated lipolysis (r = 0.36, P < 0.05). This association remained significant after adjustment for total body fat mass (r = 0.34, P ≤ 0.05). In the OM depot, CD11b mRNA abundance was positively associated with isoproterenol-stimulated lipolysis (r = 0.42, P ≤ 0.005). This association remained significant after adjustment for total body fat mass (r = 0.41, P ≤ 0.01). In subgroup analyses, high lipolytic rates in SC adipocytes were related to increased whole tissue expression of CD68 and CD11b in this compartment, independent of adiposity and fat cell size (P ≤ 0.001 and P ≤ 0.05). High lipolytic rates in OM adipocytes were related to increased whole tissue OM expression of CD11b, independent of adiposity and fat cell size (P ≤ 0.05).

Conclusions

High adipocyte lipolytic responsiveness is related to increased expression of ATM markers in the corresponding compartment, independent of adiposity and fat cell size.

Introduction

Human and animal studies have shown that obesity is related to increased macrophage infiltration in adipose tissues [1]. Adipose tissue macrophage (ATM) content seems to contribute to both local and systemic inflammation [4], and has been implicated in the development of obesity-linked comorbidities such as insulin resistance and nonalcoholic fatty liver disease [5]. CD68, CD11b, and CD11c are surface markers involved in the binding of antigens, adhesion molecules, and macrophage-specific growth factors [6, 7]. CD68 has been used to identify total macrophage infiltration in human adipose tissues as CD68+ cells are viewed as resident cells at the junctions of two or more adipocytes [8]. CD11c+ cells, classically activated macrophages (M1) that express the inflammatory cytokines interleukin (IL)-6 and TNF-α, are increased in the adipose tissues of insulin resistant patients [8, 9].

There are many different molecules and processes that seem to be implicated in ATM recruitment and activation [10]. Cinti et al. [11] suggested that necrotic hypertrophied fat cells may stimulate ATM accumulation. Indeed, ATM of massively obese women typically aggregate in crown-like structures around adipocytes of the subcutaneous adipose tissue depot, suggesting that they may have phagocytic activity [11]. Others hypothesized that adipocyte hypertrophy creating local hypoxic conditions may also be involved in macrophage attraction [1, 12]. The involvement of chemotactic signals such as MCP-1 has also been suggested as a mechanism underlying macrophage recruitment in obese adipose tissue [3, 15]. Putative metabolic factors that could regulate the inflammatory response to obesity and ATM accumulation have also been investigated.

Obesity is related to increased adipocyte size and number, as well as enhanced nonesterified fatty acids (NEFA) and glycerol release resulting from adipose tissue lipolysis [16]. Kosteli et al. demonstrated that local lipid fluxes and especially adipose tissue lipolysis may drive ATM accumulation. For example, an increase in ATM number is observed upon induction of weight loss in obese mice and coincides with the peak in circulating concentrations of NEFA and adipose tissue lipolysis. Moreover, adipose tissue triglyceride lipase null mice fail to increase ATM recruitment following fasting, suggesting that excess adipose tissue lipolysis and the release of NEFA actually cause ATM accumulation [17]. Whether lipolytic rates are related to markers of ATM infiltration has never been examined in human adipose tissues.

The aim of this study was to examine the relationship between isolated mature adipocyte lipolysis and whole tissue expression of ATM infiltration markers in lean to obese women. We tested the hypothesis that high lipolytic responsiveness is related to increased expression of ATM genes in whole abdominal subcutaneous and omental adipose tissues of women, independent of adiposity and cell size.

Methods

Subjects

The study sample included 46 healthy women aged 39.6-59.4 years. They were recruited through the elective surgery schedule of the Gynecology Unit at Laval University Medical Center. Women of the study elected for total (n = 43) or subtotal (n = 3) abdominal hysterectomies, some with salpingo-oophorectomy of one (n = 11) or two (n = 20) ovaries. Reason(s) for surgery included one or more of the following: menorrhagia/menometrorrhagia (n = 23), myoma/fibroids (n = 36), adenomyosis (n = 1), incapacitating dysmenorrhea (n = 9), pelvic pain (n = 1), benign cyst (n = 10), endometriosis (n = 8), pelvic adhesions (n = 4), benign cystadenoma (n = 1), ovarian thecoma (n = 1), endometrial hyperplasia (n = 4), or polyp (n = 1). All patients were asked to be nil-by-mouth from midnight prior to surgery. They were allowed to drink clear liquids 3 hours before their admission time. According to the time of the operation, patients fasted for at least 8-16 hours before surgery. This study was approved by the Research Ethics Committees of Laval University Medical Center. All subjects provided written informed consent before their inclusion in the study.

Body fatness and body fat distribution measurements

These measurements were performed on the morning of or within a few days before or after the surgery. Measures of total body fat mass, fat percentage, and fat-free mass were determined using dual energy X-ray absorptiometry (DEXA) using a Hologic QDR-2000 densitometer and the enhanced array whole-body software V5.73A (Hologic Inc., Bedford, MA, USA). Measurement of abdominal subcutaneous and visceral adipose tissue cross-sectional areas was performed by computed tomography using a GE Light Speed 1.1 CT scanner (General Electric Medical Systems, Milwaukee, WI, USA), as previously described [18].

Adipose tissue sampling

Subcutaneous and omental adipose tissue samples were collected during the surgical procedure at the site of incision (lower abdomen) and at the distal portion of the greater omentum, respectively. Adipocyte isolation was performed with a portion of the fresh sample and the remaining tissue was immediately stored at −80°C for subsequent analyses.

Adipocyte isolation and lipolysis

Fresh tissue samples were digested with collagenase type I in Krebs-Ringer-Henseleit buffer for 45 min at 37°C according to a modified version of the Rodbell method [19], as described [20]. Average adipocyte diameter was used in analyses.

Lipolysis experiments were performed by incubating isolated cell suspensions for 2 hours at 37°C in Krebs–Ringer Henseleit buffer (∼5,000 cells per condition), with or without ß-adrenergic receptor agonist isoproterenol in concentrations ranging from 10–10 to 10–5 mol/L or postreceptor-acting agents dibutyryl cyclic AMP (10–3 mol/L) and forskolin (10–5 mol/L). Glycerol release was assessed as described [20]. Lipid weight of the cell suspension was measured using Dole's extraction. Adipocyte weight and cell number in the suspensions were calculated using lipid weight, average cell volume, and the density of triolein. Lipolysis data were previously published in a different study on regional differences in lipolytic responsiveness [20]. On the basis of the demonstration that adipocyte size appears to be a critical determinant of regional differences in lipolysis, we expressed lipolytic responsiveness as fold over basal. However, in this study, analyses were also performed using lipolytic responsiveness expressed per cell number and generated similar results.

Whole tissue mRNA expression of ATM markers using quantitative real-time RT-PCR

Total RNA was isolated from whole subcutaneous and omental adipose tissue using the RNeasy lipid tissue extraction kit (Qiagen, Valencia CA, USA) and digested with Rnase-free Dnase (Qiagen, Missisauga, ON, Canada). RNA quantity and quality were assessed using an Agilent Technologies 2100 bioanalyser and RNA 6000 Nano LabChip kit (Agilent, Mountain View CA, USA). Complementary DNA (cDNA) was generated with Invitrogen Superscript II Rnase H-RT (Invitrogen Life Technologies, CA, USA) and purified with QIAquick PCR Purification Kit (Qiagen). Equal amounts of cDNA were run in triplicate and amplified in 2× Universal PCR Master Mix (Applied Biosystems, Foster City, Carlsbad CA, USA), with 10 nM Z-tailed forward primer, 100 nM reverse primer, 100 nM Amplifluor Uniprimer probe (Chemicon, Temecula, CA, USA). No template controls were used. The mixture was incubated at 50°C for 2 min, at 95°C for 4 min, and then cycled at 95°C for 15 s and at 55°C for 40 s 55 times using the Applied Biosystems Prism 7900HT Sequence Detector System. Target gene amplifications were normalized using expression of the 18 S subunit. Expression levels of this transcript were not associated with adiposity in our study sample. Primer sequences for CD68 (NM_001251.2; F 5′-Z-TGGATTCATGCAGGACCTCC-3′; R 5′-Z-CGCCATGTAGC TCAGGTAGACA-3′), CD11c (NM_000887; F 5′-GGCCATGC ACAGATACCAGGT-3′; R 5′-CTGGGGGTG CGATTTTCT CTG-3′), and CD11b (NM_000632.3; R 5′-Z-ACAGCCTTGACCTT ATGTCATGG-3′; R 5′-Z-CATTGCGTTTTCAGTGTCCAA-3′) were designed using Primer Express 2.0 (Applied Biosystems, Foster City, CA, USA). Forward primers containing the 5′Z sequence: AC TGAACCTGACCGTACA were used to detect amplicons. Expression data are presented in arbitrary quantity.

Statistical analyses

Pearson correlation coefficients were computed to quantify associations between characteristics of the sample or lipolysis measurements and CD68, CD11b, or CD11c mRNA expression. Statistical adjustments for total body fat mass were performed using covariance analysis. We performed linear regression analyses to predict lipolytic responsiveness in each fat compartment separately using measures of total body fat mass and adipocyte size. For each model, the study samples were stratified in two subgroups according to the median of residuals of the regression between fold-response glycerol release to isoproterenol 10−5 M, total body fat mass, and adipocyte size. Subjects with a residual below the median were considered as having lower-than-predicted lipolysis for a given adipocyte size and total body fat mass, and those with a residual above the median were considered as having higher-than-predicted lipolysis for a given adipocyte size and total body fat mass. Such stratification generated two subgroups of women with either low or high lipolytic responsiveness in each fat compartment that were matched for total body fat mass and adipocyte size. We found no significant difference between subgroups with low versus high lipolytic rates in each fat compartment for BMI and body fat distribution, even if these variables were not in the regression model. Student's t-tests were used to compare women with high versus low lipolytic rates in a given compartment. Linear regression analyses were performed to predict subcutaneous and omental CD68 or CD11b mRNA expression using all body composition and fat distribution variables available, as well as lipolysis measurements (expressed as fold over basal in response to 10−5 M isoproterenol). Pearson correlations between independent variables of the models varied between r = −0.07 and r = 0.87. Log10 and boxcox transformations were used for non-normally distributed variables. The following variables were log-transformed: BMI, total body fat mass, total adipose tissue area, subcutaneous adipose tissue area, visceral adipose tissue area, omental adipocyte diameter, CD68, CD11b, and CD11c mRNA expression levels in each compartment, omental isoproterenol-stimulated (10−5-10−9), forskolin-stimulated (10−5), and dibutyryl cyclic AMP-stimulated (10−3) lipolysis measurements and subcutaneous isoproterenol-stimulated (10−6, 10−8-10−10) and forskolin-stimulated (10−5) lipolysis measurements. The following variables were boxcox-transformed: omental isoproterenol-stimulated (10−10) lipolysis measurement, subcutaneous isoproterenol-stimulated (10−7) measurement, and dibutyryl cyclic AMP-stimulated (10−3) lipolysis measurement. Statistical analyses were performed with the JMP4 software (SAS Institute, Cary, NC, USA).

Results

Characteristics of the study sample are shown in Table 1. Women were approximately 47 years old and were slightly overweight according to their mean BMI value of 26.9 kg/m2. The range of adiposity covered the spectrum from lean to obese, with BMI values spanning from 19.1 to 41.3 kg/m2.

Table 1. Characteristics of the 46 women included in the study
VariablesMean ± SDRange (min-max)
  1. Values are mean ± SD.
  2. an = 44.
  3. bn = 45.
Age (yrs)47.2 ± 4.739.6–59.4
Weight (kg)70.2 ± 16.049.5–110.5
BMI (kg/m2)26.9 ± 5.219.1–41.3
Total body fat mass (kg)25.4 ± 9.711.2–50.8
Lean body mass (kg)42.8 ± 7.132.5–63.0
Adipose tissue areas (cm2)
Totala420 ± 190137–991
Subcutaneousa325 ± 147103–759
Viscerala96 ± 4934–233
Adipocyte diameter (μm)
Subcutaneous99 ± 1367–123
Omentalb81 ± 1752–119

Table 2 shows Pearson correlation coefficients between whole tissue CD68, CD11b, or CD11c mRNA expression in each compartment and adiposity measurements. Expression of subcutaneous and visceral ATM markers increased with the degree of adiposity. Specifically, a positive correlation was observed between BMI and CD68 or CD11b mRNA expression in both adipose tissue compartments (0.05 ≤ P ≤ 0.10). Visceral and total adipose tissue areas as well as adipocyte diameters were positively related to CD68 mRNA expression in the subcutaneous fat depot (P ≤ 0.05). Adipose tissue areas as well as adipocyte diameters were also significantly correlated with CD11b mRNA expression in this compartment (P ≤ 0.05). Trends were observed between total or visceral adipose tissue areas and CD11b mRNA abundance in the omental fat compartment (P < 0.10). CD11c mRNA expression in subcutaneous adipose tissue tended to be correlated with BMI and subcutaneous adipocyte size (P < 0.10). No significant association was observed between omental CD11c mRNA abundance and measures of body composition, fat distribution, or adipocyte size.

Table 2. Correlation between whole tissue CD68, CD11b, and CD11c mRNA abundance and characteristics of the sample
VariablesCD68CD11bCD11c
SCOMSCOMSCOM
  1. Pearson correlation coefficients;
  2. CD68, CD11b or CD11c mRNA abundance in whole tissue from each site;
  3. Subcutaneous (SC) or omental (OM) adipose tissue;
  4. (-) No significant association,
  5. aP ≤ 0.005,
  6. bP ≤ 0.05,
  7. cP ≤ 0.10.
  8. dn = 46.
  9. en = 44.
  10. fn = 45.
  11. gn = 43.
Anthropometricsd
BMI0.29b0.25c0.37b0.34b0.28c
Total body fat mass0.26c0.42a
Adipose tissue arease
Total0.30b0.40b0.28c
Subcutaneous0.26c0.37b
Visceral0.34b0.41b0.26c
Adipocyte diameter
Subcutaneousf0.36b0.26c0.48a0.26c0.28c
Omentalg0.35b0.35b

Table 3 shows Pearson correlation coefficients between lipolysis measurements in isolated fat cells and CD68 or CD11b mRNA abundance in whole tissue from each site. CD68 mRNA abundance in whole omental adipose tissue was significantly and positively associated with isoproterenol-stimulated lipolysis (10−6, 10−7, 10−9, and 10−10 M) in isolated fat cells from the same fat compartment (P ≤ 0.05). CD11b mRNA expression in omental adipose tissue was positively correlated with omental adipocyte isoproterenol-stimulated (10−5-10−10 M) and forskolin-stimulated (10−5 M) lipolysis response (P ≤ 0.05). When adjusting for total body fat mass, lipolysis measurements in omental adipose tissue were still positively associated with CD68 and CD11b mRNA expression in the same depot. Some trends and significant associations were also observed between omental adipocyte lipolysis response and CD68 mRNA expression in subcutaneous adipose tissue. However, after adjustment for total body fat mass, these associations were no longer significant. No significant association was observed between CD11c mRNA expression in omental adipose tissue and lipolysis in the omental fat depot (data not shown). Regarding basal lipolytic measurements, we found no significant correlation between basal lipolytic rates and macrophage gene expression in either fat compartment, suggesting that only adipocyte lipolytic responsiveness to adrenergic receptor- and postreceptor-acting agents is associated with markers of ATM infiltration.

Table 3. Correlation coefficients between lipolysis measurements in isolated fat cells and whole tissue CD68 or CD11b mRNA abundance
VariablesCD68CD11b
SCOMSCOM
  1. Pearson correlation coefficients.
  2. Isolated adipocyte glycerol release in each fat depot expressed as fold over basal in response to isoproterenol (10−5-10−10 M); forskolin (10−5 M) or dibutyryl cyclic AMP (10−3 M).
  3. CD68 or CD11b mRNA abundance in whole tissue from each site.
  4. SC, subcutaneous; OM, omental, AMP, adenosine monophosphate; Iso, isoproterenol; Fsk, forskolin; Dbc, dibutyryl cyclic AMP.
  5. aP ≤ 0.01,
  6. bP ≤ 0.05,
  7. cP ≤ 0.08.
  8. dAssociation that remained significant after adjustment for total body fat mass (P ≤ 0.05).
Omental adipocyte lipolysis
Iso 10−5 M0.27c0.42a, d
Iso 10−6 M0.30b0.38b, d0.56a, d
Iso 10−7 M0.28c0.35b0.50a, d
Iso 10−8 M0.38b0.41a, d
Iso 10−9 M0.34b0.40a
Iso 10−10 M0.48a, d0.46a, d
Fsk 10−5 M0.27c0.30c0.37b, d
Dbc 10−3 M0.29c
Subcutaneous adipocyte lipolysis
Iso 10−5 M0.36b, d0.27c0.40a, d
Iso 10−6 M0.33b, d0.30b, d0.33b, d
Iso 10−7 M0.35b, d0.32b, d
Iso 10−8 M0.33b, d0.27c
Iso 10−9 M
Iso 10−10 M
Fsk 10−5 M0.30b0.27c
Dbc 10−3 M

CD68 mRNA expression in whole subcutaneous adipose tissue was positively and significantly correlated with isoproterenol-stimulated (10−5-10−8 M) and forskolin-stimulated (10−5 M) lipolysis in adipocytes isolated from this depot (Table 3). Only a trend was observed between CD11b mRNA expression in subcutaneous adipose tissue and isoproterenol-stimulated (10−8 M) lipolysis (P ≤ 0.08). CD11b mRNA abundance in omental adipose tissue was positively related to subcutaneous adipocyte isoproterenol-stimulated (10−5-10−7 M) lipolysis measurements (P ≤ 0.05). The vast majority of these associations remained significant after adjustment for total body fat mass. No significant association was observed between CD11c mRNA abundance in subcutaneous adipose tissue and lipolysis in subcutaneous adipose cells (data not shown).

Figure 1 shows subgroups of women with either low or high lipolytic responses in subcutaneous adipocytes, but with similar adipocyte sizes and total body fat mass. No significant difference was found between women with low versus high lipolytic response for BMI (25.7 ± 4.5 vs. 28.1 ± 5.6 kg/m2, respectively, P = 0.15), total body fat mass (24.2 ± 8.9 vs. 26.5 ± 10.8 kg, respectively, P = 0.45), visceral adipose tissue area (87 ± 50 vs. 102 ± 46 cm2, respectively, P = 0.25), subcutaneous adipose tissue area (306 ± 132 vs. 346 ± 164 cm2, respectively, P = 0.50), omental adipocyte diameter (77 ± 19 vs. 81 ± 12 μm, respectively, P = 0.27), and subcutaneous adipocyte diameter (98 ± 14 vs. 100 ± 13 μm, respectively, P = 0.41). As expected, isoproterenol- (10−5-10−9 M), forskolin-, and dibutyryl cAMP-stimulated lipolysis in subcutaneous adipocytes was significantly higher in the group with high lipolytic responsiveness compared with the group with low lipolytic responsiveness (P ≤ 0.07 or P ≤ 0.005) (Figure 1A). High lipolytic responsiveness in subcutaneous adipocytes was related to increased expression of CD68 and CD11b in this compartment, independent of adiposity and fat cell size (P ≤ 0.0001 and P ≤ 0.05, respectively) (Figure 1B). As expected, significant differences in isoproterenol- (10−5-10−8 M) stimulated lipolysis were also observed in omental adipocytes of these subgroups, which were stratified based on subcutaneous lipolytic rates (P ≤ 0.05) (Figure 1C). High lipolytic responsiveness in subcutaneous adipocytes was associated with significantly increased expression of CD68 and CD11b in omental adipose tissue, independent of adiposity and fat cell size (P ≤ 0.05) (Figure 1D).

Figure 1.

Comparison of (A) isolated subcutaneous (SC) adipocyte lipolysis; (B) SC CD11b, CD68, and CD11c mRNA expression; (C) isolated omental (OM) adipocyte lipolysis; and (D) OM CD11b, CD68, and CD11c mRNA expression in women with either low or high SC lipolytic rates, but similar adipocyte sizes and total body fat mass. **P ≤ 0.005, *P ≤ 0.05, #P ≤ 0.10.

Figure 2 shows subgroups of women with either low or high lipolytic responsiveness in omental adipocytes, but with similar adipocyte sizes and total body fat mass. No significant difference was found between women with low versus high lipolytic rates for BMI (25.6 ± 3.7 vs. 27.6 ± 5.6 kg/m2, respectively, P = 0.30), total body fat mass (22.9 ± 7.5 vs. 26.8 ± 11.0 kg, respectively, P = 0.25), visceral adipose tissue area (87 ± 45 vs. 101 ± 52 cm2, respectively, P = 0.44), subcutaneous adipose tissue area (288 ± 111 vs. 352 ± 171 cm2, respectively, P = 0.25), omental adipocyte diameter (79 ± 19 vs. 82 ± 15 μm, respectively, P = 0.51), and subcutaneous adipocyte diameter (97 ± 15 vs. 100 ± 12 μm, respectively, P = 0.44). As expected, isoproterenol- (10−5-10−9 M), forskolin-, and dibutyryl cAMP-stimulated lipolysis in omental adipocytes were significantly higher in women with high lipolytic responsiveness compared with women with low lipolytic responsiveness in omental adipocytes (P ≤ 0.07 or P ≤ 0.005) (Figure 2A). High lipolytic responsiveness in omental adipocytes was related to increased omental expression of CD11b, independent of adiposity and adipocyte size (P ≤ 0.05). No significant difference was observed in omental expression of CD68 and CD11c (Figure 2B). Significant differences were observed in isoproterenol- (10−5-10−9 M) and dibutyryl cAMP-stimulated lipolysis in subcutaneous adipocytes of these subgroups stratified according to omental adipocyte lipolysis (P ≤ 0.07 or P ≤ 0.05)) (Figure 2C). High lipolytic responsiveness in omental adipocytes was related to increased expression of CD11b and CD68 in subcutaneous adipose tissue, independent of adiposity or adipocyte size (P ≤ 0.05) (Figure 2D).

Figure 2.

Comparison of (A) isolated OM adipocyte lipolysis; (B) OM CD11b, CD68, and CD11c mRNA expression; (C) isolated SC adipocyte lipolysis; and (D) SC CD11b, CD68, and CD11c mRNA expression in women with either low or high OM lipolytic rates, but similar adipocyte sizes and total body fat mass. **P ≤ 0.005, *P ≤ 0.05, #P ≤ 0.10.

To identify the strongest predictors of ATM marker expression in each depot, we generated multiple linear regression models including adiposity measurements, cell size, and lipolysis measurements expressed as fold over basal in response to 10−5 M isoproterenol. Isoproterenol-stimulated lipolysis in cells from the omentum was the best predictor of CD68 and CD11b mRNA expression in whole omental adipose tissue, explaining 11% (P = 0.05) and 19% (P = 0.007) of the variance, respectively. Furthermore, isoproterenol-stimulated lipolysis in cells from subcutaneous fat was the best predictor of CD68 mRNA abundance in whole subcutaneous adipose tissue explaining 15% (P = 0.02) of the variance. Isoproterenol-stimulated lipolysis in cells from subcutaneous fat tended to be a predictor of CD11b mRNA abundance in subcutaneous adipose tissue, explaining 6% (P = 0.08) of the variance. We found similar results with responses to other lipolytic stimuli (data not shown).

Discussion

On the basis of the demonstration that lipolysis induces macrophage infiltration in mice [17], the objective of the study was to examine the relationship between isolated mature adipocyte lipolysis and markers of ATM infiltration in lean to obese women. We tested the hypothesis that high lipolytic responsiveness is related to increased whole tissue expression of ATM genes in abdominal subcutaneous and omental adipose tissues of women, independent of adiposity and fat cell size. Several positive correlations were observed between measures of lipolysis and whole tissue ATM marker expression. High lipolytic responsiveness in subcutaneous adipocytes was related to increased expression of CD68 and CD11b in both adipose tissue compartments, independent of adiposity and fat cell size, and high lipolytic responsiveness in omental adipocytes was related to increased expression of CD11b in both adipose tissue compartments, independent of adiposity and fat cell size. This is the first study to report that high adipocyte lipolytic responsiveness is associated with markers of ATM infiltration, independent of adiposity and fat cell size in humans.

The mechanisms promoting ATM infiltration are not fully defined. Some studies have suggested that increased synthesis of chemotactic signals such as MCP-1 in obese adipose tissue may lead to adipose tissue monocyte/macrophage recruitment [3, 15]. Adipocyte hypertrophy creating local hypoxic conditions may also be implicated in macrophage recruitment by stimulating inflammatory pathways through JNK1-regulated chemokine release [1, 12]. Cinti et al. [11] hypothesized that death of hypertrophied fat cells may be a stimulus regulating ATM accumulation. Recently, Kosteli et al. [17] showed that in mice, NEFA released from adipocytes during early weight loss or fasting may also be central regulators of ATM accumulation. Consistent with this hypothesis [17], we report that high lipolytic responsiveness in human adipose tissue is associated with increased adipose tissue expression of macrophage genes (CD68 and CD11b), independent of adiposity and fat cell sizes. Other studies demonstrated that NEFA can contribute to the stimulation of local inflammation in adipose tissue. Using mouse models, Shi et al. [21] and Davis et al. [22] demonstrated that fatty acids have the capacity to bind and activate toll-like-receptor-4 (TLR4), which induce inflammatory signals in adipose tissue. On the basis of in vitro coculture systems composed of adipocytes and macrophages, Suganami et al. [23] demonstrated the existence of a paracrine loop involving saturated fatty acids and TNF-α derived from adipocytes and macrophages, respectively. This phenomenon may reflect a vicious cycle between adipocytes and macrophages, which possibly aggravates adipose tissue inflammation.

CD11b is a marker of macrophages as well as a marker of neutrophil infiltration in adipose tissue [24]. In fact, infiltrating immune cells in adipose tissue include macrophages but also lymphocytes, neutrophils, and other granulocytes [25]. Neutrophils infiltrating adipose tissues may produce chemokines, proinflammatory cytokines, and metabolites that could facilitate monocyte recruitment in adipose tissue, as seen in several disease models [28]. In various tissues, neutrophils are known to initiate the inflammatory response and to increase the chronic inflammatory state by recruiting macrophages. In this context, Elgazar-Carmon et al. [24] demonstrated in mice that neutrophil infiltration into adipose tissue precedes macrophage infiltration in response to high fat feeding. At this time, it is difficult to determine whether neutrophilic inflammation may actually explain our results better than a macrophage-driven hypothesis. In the context of lipolysis, fatty acids released by adipocytes may induce phagocytic macrophage infiltration to reduce lipid excess, which may possibly explain our results.

Recent studies reported that obesity is associated with changes in macrophage activation state, from the alternate (M2) to the classical (M1) macrophage phenotype. M1 macrophages, which produce large amounts of proinflammatory cytokines [9], are increased in adipose tissues of insulin-resistant patients [8]. Recent studies in humans have demonstrated that macrophages in crown-like patterns stained for CD86, CD40, and CD11c are compatible with the M1-like macrophage phenotype [8, 28, 29]. Wentworth et al. demonstrated using flow cytometry that obesity was associated with significant increases in densities of CD11c+CD206− (monocyte), CD11c+CD206+ (crown-like structure), and CD11-(resident) in human subcutaneous and omental adipose tissue [8]. Interestingly, we found no significant associations between lipolysis response and CD11c mRNA abundance in either depot. The characteristics of the sample we examined could potentially explain this discrepancy. Our study included a group of lean to obese women, whereas the other sample included formerly obese, obese, and severely obese women (BMI: 39-56 kg/m2). Our sample was very likely characterized by a lower inflammatory state compared with severely obese subjects. In a similar population, we previously observed that a very small number of women had adipose tissue-infiltrating macrophages typically dispersed as crown-like structures, which may represent more advanced stages of the inflammatory response [30]. In this previous study, the percentage of crown-like structures in subcutaneous adipose tissue was positively and significantly associated with CD68+ cell percentage as well as mRNA expression of CD11c, CD68, and CD11b in this depot, whereas the percentage of crown-like structures in omental adipose tissue was significantly associated with CD68+ cell percentage and CD11c mRNA abundance level in this depot [30]. We cannot exclude that investigating a larger sample or more obese subjects could lead to significant differences in CD11c mRNA expression between obese and lean women. The method used to quantify CD11c (RT-PCR quantitative vs. immunohistology or flow cytometry) may also explain our finding. Finally, it may be more difficult to clearly identify recently recruited macrophages and to unequivocally identify the M1 or M2 phenotype in humans compared with mice [31].

Strong correlations were found between macrophage markers in omental adipose tissue and lipolysis measurements in this fat depot. Most measurements of subcutaneous adipocyte lipolysis were also positively associated with CD68 mRNA expression in subcutaneous adipose tissue. The vast majority of these associations remained significant after adjustment for total body fat mass. These findings are consistent with abundant literature showing that obesity is positively associated with both adipocyte size and ATM infiltration [1]. Large adipocytes have higher rates of triglyceride uptake and release more NEFAs during lipolysis than smaller adipocytes [32]. To exclude the fact that high lipolytic rates are found in women with large adipocytes, and consequently exclude cell size as the explanatory factor in the relationship between lipolysis measurements and markers of ATM infiltration, we subdivided women into subgroups with either low or high lipolytic rates in omental or subcutaneous adipocytes, but with similar adipocyte sizes and total body fat mass. To further control for adipocyte size, we used lipolysis measurements expressed as fold over basal in our analyses. We observed no association between these variables and adipocyte size. On the basis of these results, adipocyte size variation cannot be considered as a plausible confounder of the association between high lipolytic responsiveness and CD68 or CD11b mRNA expression reported here.

Previous studies have already shown that subcutaneous and/or visceral ATM accumulation increase with the degree of adiposity [1, 11, 33], which is highly consistent with our results. We also recently demonstrated in a similar sample of women that CD68+ cell percentage and CD68 mRNA expression in both adipose tissues were positively and significantly correlated with most measures of adiposity including adipose tissue areas and adipocyte diameters [30]. All these results are concordant with those presented in this study.

Limitations of this study need to be acknowledged. Our results are cross sectional, which prevents us for concluding on cause-and-effect relationships. Thus, it is not possible to conclude that high lipolytic rates are the cause of ATM infiltration and ATM marker expression. The duration of fasting prior to surgery could have influenced our findings. According to the timing of the surgery, patients fasted for at least 8-12 h before surgery if they were scheduled in the morning or for at least 12-16 h before surgery if they were scheduled in the afternoon. When women were subdivided into two subgroups according to the time of the surgery (women who had surgery during the morning vs. women who had surgery during the afternoon), we found no significant difference between the two subgroups for lipolysis and ATM markers in either fat compartment, after adjustment for total body fat mass, body fat distribution, and adipocyte size. Whether characteristics of our patients influence fasting-induced macrophage infiltration is uncertain. However, our association was seen in models adjusted for total body fat mass and other characteristics. Furthermore, we measured lipolytic responsiveness in vitro by incubating isolated cell suspensions with adrenergic receptor- and postreceptor-acting agents. Consequently, we measured the capacity of mature adipocytes to respond to lipolytic stimuli. For these reasons, we suggest that lipolytic responsiveness is unlikely to be affected by the duration of fasting prior to surgery. Our study also did not include males. We examined only women because of the difficulty of setting-up similar studies in lean to moderately obese men. Posttranscriptional and translational modifications may also affect the relationships currently observed with macrophage gene expression. Further studies are required to better understand the molecular mechanisms underlying increased macrophage infiltration in highly lipolytic adipose tissues.

In summary, high lipolytic rates in human adipocytes are related to increased whole tissue expression of ATM markers in the corresponding compartment, independent of adiposity and fat cell size. Our findings are concordant with the proposal that local release of fatty acids induces ATM recruitment, which may also represent an important mechanism of adipose tissue homeostasis regulation in lean to obese women.

Acknowledgments

This study was supported by operating funds from the Canadian Institutes of Health Research-Institute of Gender and Health to André Tchernof (MOP-64182). We acknowledge the contribution of the study coordinator, gynecologists, nurses and radiology technicians as well as the study participants. Andréanne Michaud was funded by Diabète Québec and Fondation Jean-Paul Houle.

Ancillary