Calcium-dependent release of adipocyte fatty acid binding protein from human adipocytes

Abstract

Background:

Fatty acid binding protein 4 (FABP4) is a predominantly cytosolic protein of the adipocytes, but also abundantly present in human plasma; its plasma concentrations were linked to obesity and metabolic syndrome. Recent studies have suggested a direct extracellular effect of FABP4 in the regulation of glucose metabolism and heart function independently of its effect as a carrier protein. Interestingly, FABP4 has no secretory signal sequence; hence, the mechanisms how FABP4 is released from adipocytes are unclear.

Methods and results:

In this study we investigated the mechanisms for FABP4 secretion from human adipocytes by using isolated primary pre-adipocytes (PAs) and the human adipocyte cell strain Simpson–Golabi–Behmel syndrome. In undifferentiated PAs, FABP4 expression was barely detectable and increased continuously during differentiation. The increase in FABP4 mRNA expression was accompanied by high levels of FABP4 secretion. In differentiated human adipocytes, FABP4 secretion was not abolished by blocking the Golgi-dependent secretory pathway in vitro, supporting a non-classical secretion mechanism for FABP4. However, raising intracellular Ca²⁺ levels enhanced FABP4 secretion in a concentration-dependent manner.

Conclusion:

This study shows that FABP4 is actively released from human adipocytes in vitro via a non-classical, calcium-dependent mechanism.

Introduction

Fatty acid binding proteins (FABPs) are cytosolic proteins with a molecular mass of 14–15 kDa and a high-affinity for long-chain fatty acids and other hydrophobic ligands.¹ Adipocyte FABP4 is predominantly expressed in adipocytes and accounts for ∼1% of total cytosolic protein in human adipose tissue,² but is also produced in macrophages³ and endothelial cells.⁴ The major function of FABP4 is mediating intracellular trafficking and targeting of fatty acids.¹ Recent data highlight an important role of FABP4 in the pathogenesis of obesity-related diseases, such as type 2 diabetes and atherosclerosis. Experimental animal models show that FABP4 knockout significantly improved glucose and lipid metabolism in obese mice.^{5, 6} Furthermore, apolipoprotein E-deficient mice, lacking FABP4, were protected against developing atherosclerosis.^{7, 8}

In humans, FABP4 is abundantly present in the plasma, presumably due to its release from adipocytes,^{9, 10} and plasma concentrations have recently been found to be directly linked with obesity and metabolic syndrome.^{11, 12} In addition, we previously reported a cardiodepressant effect of FABP4 in vitro.¹⁰ A previous study identified FABP4 in the supernatant of differentiated rodent adipocytes.¹¹ Similarly, we showed that mature human isolated adipocytes release FABP4 into extracellular media in vitro.¹⁰ Considering these data, the question arises whether FABP4 is an actively secreted protein, thus acting as an endocrine/paracrine hormone, linking adipose tissue to target organs. The primary sequence of FABP4 does not harbor a secretory signal sequence; therefore, the mechanism of its secretion remains to be elucidated.

In the present study we demonstrate that FABP4 is actively released from human adipocytes and THP-1 macrophages via a nonclassical, calcium-dependent mechanism.

Material and methods

Cell culture experiments

Culture of Simpson–Golabi–Behmel syndrome adipocytes

Human Simpson–Golabi–Behmel syndrome (SGBS) cells were derived from the subcutaneous white adipose tissue of a patient with SGBS; they possess a high capacity for adipogenic differentiation, showing typical characteristics of mature adipocytes in vitro.¹³ SGBS adipocytes were maintained in Dulbecco’s modified Eagle’s medium/F-12 (1:1) (Life Technologies, Karlsruhe, Germany) supplemented with 33 μM biotin, 17 μM pantothenate (Sigma, Munich, Germany), 10% fetal calf serum (Life Technologies), 100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin (Life Technologies; 0F medium). For experimental purposes, cells were plated on six-well plates and grown to near confluence. For induction of adipogenic differentiation, cells were washed repeatedly with phosphate-buffered saline (Sigma) and then cultured in serum-free 0F medium containing 10 μg ml⁻¹ apo-transferrin, 20 nM insulin, 100 nM hydrocortisone, 0.2 nM triiodothyronin, 25 nM dexamethasone, 250 μM isobutylmethylxanthine (all Sigma) and 2 μM rosiglitazone (Cayman Chemicals, Ann Arbor, MI, USA). After 4 days, medium was replaced by serum-free 0F medium containing 10 μg ml⁻¹ apo-transferrin, 20 nM insulin, 100 nM hydrocortisone and 0.2 nM triiodothyronin. Medium was replenished every 3 to 4 days.

Isolation and culture of primary human pre-adipocytes

Primary human pre-adipocytes (PAs) from subcutaneous white adipose tissue were isolated from human lipoaspirates, obtained during liposuction surgery of the outer thigh or the abdomen. Female patients undergoing surgery were healthy and non-obese (aged between 25 and 43 years, body mass index ranging between 20 and 25 kg m⁻², n=3 patients). Cells were isolated and maintained as follows. Briefly, lipoaspirates were immediately transported to the laboratory under sterile conditions. After washing with phosphate-buffered saline, a collagenase digestion was performed using 100 U ml⁻¹ collagenase (NB4 standard grade; Serva, Heidelberg, Germany) diluted in phosphate-buffered saline containing 2% fatty acid-free bovine serum albumin (PAA, Cölbe, Germany). Cells were then pelleted for 10 min at 1000 r.p.m.; the resuspended pellet was filtered through a 150-μM filter. After a repeated centrifugation step, the pellet was resuspended in cold Dulbecco’s modified Eagle’s medium/F-12 and subsequently filtered through a 70-μM filter (BD Biosciences, Heidelberg, Germany). Isolated PAs were kept at 37 °C in a humidified atmosphere of 5% CO₂ and cultured in Dulbecco’s modified Eagle’s medium/F-12 (1:1), supplemented with 10% fetal calf serum, 100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin.

Cells were plated on 24-well plates and grown to near confluence before adipogenic differentiation was induced by adding differentiation medium 1 (Dulbecco’s modified Eagle’s medium/F-12 containing 100 nM hydrocortisone, 66 nM insulin, 10 μg ml⁻¹ apo-transferrin, 1 nM triiodothyronin, 1 μg ml⁻¹ troglitazone and 0.5 mM isobutylmethylxanthine). After 3 days differentiation, medium 1 was replaced by differentiation medium without troglitazone and isobutylmethylxanthine.

At day 11 of differentiation, SGBS cells or primary human PAs were treated with either 5 μg ml⁻¹ brefeldin A, 3 μM monensin A or the corresponding vehicle for 6 h. Another set of cells was treated with either 0.5 μM, 1 μM or 3 μM ionomycin or vehicle for 6 h (all from Sigma). Subsequently, supernatants were collected and centrifuged for 10 min at 1000 r.p.m. at 4 °C to remove cellular debris. Supernatants were stored at −80 °C until further use.

Culture of THP-1cells

The human monocytic cell line (THP-1; ATCC, Manassas, VA, USA) was maintained in RPMI 1640-Glutamax (Life Technologies) supplemented with 100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin, and 10% fetal calf serum. For experimental purposes, monocytes were stimulated for 3 h with 0.5 μM phorbol 12-myristate 13-acetate (Life Technologies) and subsequently seeded at a density of 100 000 cells per well on 24-well plates. Differentiated THP-1 cells were then treated with 0.5, 1 or 3 μM ionomycin or the corresponding vehicle for 24 h. Subsequently, supernatants were collected and centrifuged for 10 min at 1000 r.p.m. at 4 °C to remove cellular debris. Supernatants were stored at −80 °C until further use.

Reverse-transcription PCR and quantitative real-time PCR

Total RNA from SGBS adipocytes and human PAs was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany). DNA was eliminated using gDNA eliminator columns during RNA preparation according to the manufacturer’s protocol. For reverse transcription, up to 1 μg of the total RNA was converted to first-strand cDNA using M-MLV reverse transcriptase, reaction buffer, RNase inhibitor, dNTP mix and oligo(dT)15/random hexamer primer according to the manufacturer’s instructions (Promega, Mannheim, Germany). Absolute quantitative real-time PCR was performed using SYBR green (Qiagen) and the Roche Light Cycler 1.5 (Roche, Mannheim, Germany). Primers were designed to span at least one intron to prevent unspecific amplification of DNA remnants. To normalize data, β-actin was used as an internal control gene. For absolute quantitative real-time PCR, a standard curve was generated for each target gene as well as the internal control gene, facilitating quantitation of unknown samples based on the defined standard curve. The standards were generated by cloning of the desired target sequence into a plasmid vector using the TOPO TA Cloning Kit (Life Technologies), according to the manufacturer’s protocol. After cloning and purification using the Invisorb Plasmid Maxi Kit (Stratec Molecular, Berlin, Germany), the exact amount of copies was determined and serial dilutions were used for the quantitation. Previous evaluation of different housekeeping genes in our laboratory (for example, β-actin, glyceraldehyde 3-phosphate dehydrogenase, eukaryotic elongation factor 2 and 14–3–3 protein zeta/delta) revealed β-actin to be most suited for our system, as the expression of β-actin showed the least variations (data not shown).

Primers used are listed below:

β-actin (NM_001101.3): forward (5′-GCCGTCTTCCCCTCCATCGTG-3′), reverse (5′-GGAGCCACACGCAGCTCATTGTAGA-3′). FABP4 (NM _001442.1): forward (5′-TGGGGATGTGATCACCATTAAATCT-3′), reverse (5′-CATTTCTGCACATGTACCAGGACAC-3′). Peroxisome proliferator-activated receptor-γ (PPARγ; NM_138711.3): forward (5′-CCCATTGAAGACATTCAAGACAACC-3′), reverse (5′-CCCTCAGAATAGTGCAACTGGAAGA-3′).

Immunoassays

FABP4 concentrations in cell culture supernatants were measured using the AFABP EIA kit according to the manufacturer’s protocol (Bertin Pharma, IBL, Hamburg, Germany). Sensitivity of the test was 0.1 ng ml⁻¹.

Measurement of adiponectin concentrations in cell culture supernatants were performed using the Adiponectin human enzyme-linked immunosorbent assay kit following the manufacturer’s instructions (Abcam, Cambridge, UK).

Statistical analyses

Data presented are expressed as mean±s.e.m. and represent at least three independent experiments (n=3). Statistical analysis was performed by unpaired Student’s t-test using GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego CA, USA). Values of P<0.05 were considered statistically significant.

Results

Expression and secretion of FABP4 during differentiation

FABP4 and PPARγ mRNA expression in SGBS cells were analyzed at several time points of differentiation (D0–D11). FABP4 expression was almost not detectable in undifferentiated SGBS cells (D0) but increased continuously during differentiation (Figure 1a). Expression of PPARγ, a molecular marker of adipocyte differentiation, was detected at low levels in undifferentiated SGBS cells (D0) and increased during differentiation with a peak on day 4 of differentiation (Figure 1a).

Figure 1

FABP4 mRNA expression and secretion and PPARγ expression during differentiation of SGBS adipocytes (a and b) and primary human preadipocytes (PAs; c and d). Expression of FABP4 and PPARγ was quantified by real-time PCR, using the absolute quantification method; in addition, data was normalized to β-actin mRNA levels. FABP4 secretion into the cell culture medium of SGBS adipocytes (b) and differentiating human PAs (d) was measured by immunoassay; n=3 independent cell culture experiments; **P<0.01, ***P<0.0001.

In contrast to high FABP4 expression, almost no secreted FABP4 was detected in the supernatant of SGBS cells at day 4 of differentiation. Starting from day 6 of differentiation, FABP4 secretion into the supernatant increased steadily, reaching high levels on day 11 of differentiation (Figure 1b).

FABP4 and PPARγ expression, as well as FABP4 secretion into the cell culture medium, were additionally analyzed in primary human PAs at indicated time points (Figures 1c and d). In undifferentiated PAs (D0), no FABP4 mRNA expression was detectable. After induction of differentiation, FABP4 mRNA levels increased strongly, showing a maximum at day 6 of differentiation. Similarly, PPARγ expression increased after induction of differentiation, with peak levels at day 3 of differentiation (Figure 1c).

Levels of secreted FABP4 were detectable starting from day 4 of differentiation and increased up to day 11 of differentiation (Figure 1d).

Effect of brefeldin A and monensin A on FABP4 secretion from differentiated adipocytes in vitro

To investigate the involvement of the endoplasmic reticulum–Golgi pathway in FABP4 release from adipocytes, differentiated SGBS adipocytes (D11) were incubated with brefeldin A or monensin A. Both agents are inhibitors of the Golgi-dependent secretory pathway.^{14, 15} Although adiponectin release into the cell culture medium was reduced significantly by both brefeldin A and monensin A, FABP4 secretion was not considerably affected (Figure 2).

Figure 2

Effect of brefeldin A and monensin A on FABP4 and adiponectin secretion from SGBS adipocytes in vitro. SGBS adipocytes (D11) were incubated with brefeldin A (5 μg ml⁻¹), monensin A (3 μM) or vehicle; FABP4 (a) and adiponectin (b) secretion were measured using immunoassay; n=3 independent cell culture experiments; *P<0.05; n.s., not significant.

Effect of ionomycin on FABP4 secretion from adipocytes

To elucidate the role of free intracellular calcium (Ca²⁺) in the regulation of FABP4 secretion from differentiated adipocytes, SGBS adipocytes (D11) were treated with different concentrations of ionomycin, a substance that raises intracellular Ca²⁺ levels.^{16, 17} FABP4 concentrations in the supernatant significantly increased after treatment with 0.5, 1 and 3 μM ionomycin (Figure 3a). Adiponectin secretion, in contrast, was not significantly affected by treatment with ionomycin (Figure 3b). Incubation of differentiated primary human PAs (D11) with ionomycin also resulted in an increased FABP4 secretion into the culture medium, whereas adiponectin secretion was not significantly altered compared with vehicle (Figures 3c and d).

Figure 3

Effect of ionomycin on FABP4 and adiponectin secretion from SGBS adipocytes (a and b) and differentiated primary human preadipocytes (PAs; c and d) in vitro. Ionomycin increased significantly FABP4 release from SGBS adipocytes (D11) (a) as well as from human PAs (c). Adiponectin secretion from SGBS adipocytes (b) and from PAs (d) was not significantly affected by ionomycin treatment; n=3 independent cell culture experiments; *P<0.05; **P<0.01; n.s., not significant.

Specific effect of ionomycin on FABP4 secretion from differentiated adipocytes

To further clarify whether the increased secretion of FABP4 on treatment with ionomycin was due to stimulation of FABP4 gene expression rather than stimulation of FABP4 release, SGBS adipocytes (D11) were treated with ionomycin for 30 min, 1, 4 and 6 h. Both mRNA levels of FABP4 and FABP4 release into the supernatant were analyzed at the indicated time points. Already 30 min of ionomycin treatment significantly increased FABP4 release into the extracellular medium, whereas no changes were detected in FABP4 expression up to 6 h of treatment (Figure 4).

Figure 4

Effect of ionomycin on FABP4 expression and secretion. SGBS adipocytes (D11) were treated with 3 μM ionomycin, and FABP4 expression (a) and release (b) into the medium were analyzed at indicated time points; n=3 independent cell culture experiments; *P<0.05, **P<0.01.

Effect of ionomycin on FABP4 secretion from THP-1 macrophages

Differentiation of THP-1 monocytes into adherent macrophages, induced by phorbol 12-myristate 13-acetate treatment, resulted in an increased FABP4 release into the cell culture medium (Figure 5a). To analyze whether FABP4 secretion follows a similar mechanism in macrophages as in adipocytes, differentiated THP-1 macrophages were treated with 0.5, 1 and 3 μM ionomycin. Treatment with ionomycin significantly increased FABP4 secretion into the supernatant (Figure 5b).

Figure 5

Differentiation of THP-1 monocytes was induced by phorbol 12-myristate 13-acetate (PMA); differentiated THP-1 macrophages show increased FABP4 secretion into the medium (a). On treatment with ionomycin (0.5, 1 and 3 μM), FABP4 secretion was significantly elevated (b); n=3 independent cell culture experiments; **P<0.01.

Discussion

In the present study we show that FABP4 is actively released from differentiated human adipocytes in vitro via a nonclassical, calcium-dependent mechanism.

FABP4 has traditionally been implied to act as an intracellular lipid chaperone, mediating the intracellular trafficking and targeting of fatty acids.¹ However, recent studies emphasize an important role of circulating FABP4 in the pathogenesis of obesity-related complications such as diabetes mellitus and atherosclerosis.^{11, 18} Supporting an extracellular role of FABP4, we have previously shown a direct effect of FABP4 in suppressing cardiomyocyte contraction in vitro, independently of its function as a carrier protein.¹⁰ An extracellular role of FABP4 in regulating glucose production and gluconeogenic activity has also recently been suggested in hepatocytes in vitro.¹⁹ Consistent with an extracellular function of FABP4, we and others found high concentrations of FABP4 in the supernatants of differentiated rodent PAs and mature human adipocytes.^{10, 11}

Despite cumulating evidence of a significant extracellular function of FABP4, its secretory mechanisms were unclear so far. FABP4 lacks an N-terminal secretory signal sequence, necessary for the classical secretory pathway (endoplasmic reticulum/Golgi-dependent pathway). Alongside with the classical secretory pathway, proteins can also be actively secreted from eukaryotic cells via the endoplasmic reticulum- and Golgi-independent pathways.²⁰ It was recently reported that FABP4 can be actively imported into the nucleus, even though its primary sequence lacks a nuclear localization signal. Interestingly, the three-dimensional structure of FABP4 exhibits such a signal, enabling nuclear import of FABP4 on ligand binding.^{21, 22} To assess whether FABP4 secretion via the Golgi-dependent pathway might be facilitated by a similar mechanism, we evaluated FABP4 secretion after blocking this secretory pathway.

Here we demonstrate that in differentiated human adipocytes, FABP4 secretion is not abolished by blocking the Golgi-dependent secretory pathway in vitro, supporting a nonclassical secretion mechanism for FABP4. This finding corresponds with recently published data by Cao et al.,¹⁹ showing that FABP4 secretion from a rodent PA cell line was not inhibited by blocking the classic secretory pathway. Adiponectin secretion however was reduced significantly by brefeldin A and monensin A, supporting secretion via the Golgi-dependent pathway for this adipokine. This finding is consistent with findings from Xie et al.^{23, 24}

Furthermore, we demonstrate that FABP4 secretion significantly increased on treatment of human adipocytes with ionomycin, an ionophore that raises intracellular Ca²⁺ levels.^{16, 17} Importantly, adiponectin release into the culture medium was not influenced by ionomycin, indicating that the increased amount of FABP4 in the culture medium was due to an active process connected to the raised levels of intracellular Ca²⁺ rather than an unspecific lysis of cells. In addition, we show that the stimulation of FABP4 secretion on ionomycin treatment is not the consequence of increased FABP4 expression but rather an acute and specific response to the raised levels of intracellular Ca²⁺. Herewith, we show that FABP4 is released actively from human adipocytes in a calcium-dependent manner. Interestingly, our data indicates that FABP4 is released from human macrophages by a similar mechanism.

In their study, Kralisch et al.³⁵ report similar findings in rodent PAs. In addition to a similar, nonclassical release of FABP4 into the extracellular space, they observed release of small amounts of FABP4 in the form of microvesicles. Consistent with this data, we have previously reported release of a small fraction of FABP4 via microvesicles by isolated mature human adipocytes.¹⁰ In accordance with their low FABP4 content, microvesicles isolated from mature adipocytes only had a minor cardiodepressant activity.¹⁰ Otherwise, microvesicle-free supernatant showed a significant cardiodepressant effect, along with high amounts of FABP4.¹⁰

An active release of FABP4 from adipocytes is of particular interest, as circulating FABP4 was not only positively correlated to key features of the metabolic syndrome in humans¹¹ but also seemed to have a critical role in the pathogenesis of insulin resistance and atherosclerosis in several animal models. FABP4 knockout mice have been shown to be protected against developing insulin resistance in the course of genetic- or diet-induced obesity.^{5, 6} FABP4 is also produced in macrophages where it seems to have a critical role in the development of atherosclerosis^{8, 25} and inflammation.²⁶ In fact, apolipoprotein E-deficient mice that are lacking FABP4 are markedly protected from developing atherosclerosis.^{25, 27} Furthermore, FABP4 directly inhibits rat cardiomyocyte contraction in vitro, thus mainly explaining the suppressive effect of adipocyte factors on heart contractile function.^{28, 29} Supporting these in vitro data, circulating FABP4 levels are associated with left ventricular diastolic dysfunction in obese subjects (Baessler et al., submitted) as well as with cardiac remodeling in obese women.³⁰ Therefore, inhibiting FABP4 or its secretion from adipocytes might be a promising possibility for the treatment of obesity-related metabolic and cardiovascular complications. The aromatic biphenyl azole compound BMS309403, a member of a group of small molecules that interact with the fatty acid-binding pocket of FABP4, thereby inhibiting the binding of fatty acids,^{31, 32} is considered a promising new therapeutic agent to treat complications of the metabolic syndrome. In a recent study, BMS309403 effectively prevented foam cell transformation and the cellular expression of pro-inflammatory cytokines. Furthermore, treatment of apolipoprotein E-deficient mice with this compound reduced atherosclerotic lesions and improved the glucose metabolism as well as insulin sensitivity in mice in vivo.³³ Nevertheless, we provided evidence that BMS309403 has alarming negative ionotropic effects on the myocardium in vitro, indicating limitations for the potential therapeutic use of this compound.³⁴ Thus, new strategies are needed to prevent the deleterious cardiometabolic effects of FABP4.

In summary, we demonstrate that FABP4 is actively secreted from human adipocytes in vitro via a nonclassical, calcium-dependent mechanism. An active secretory process of FABP4 supports the increasing evidence of an important role of extracellular FABP4 in the pathogenesis of metabolic and cardiovascular complications of obesity. The comprehension of the secretory mechanism of FABP4 from adipocytes will allow the development of new therapeutical strategies for the cardiometabolic complications of obesity.