We generated preadipocyte cell lines impaired in adrenomedullin production through integration of an adrenomedullin small interfering RNA expression vector. The reduction of adrenomedullin synthesis strongly accelerated adipose differentiation. These results were bolstered when overexpression of active adrenomedullin peptide led to delayed differentiation. Therefore, we propose that adrenomedullin is an antiadipogenic factor. Moreover, we checked whether insulin, a proadipogenic factor, regulates expression of adrenomedullin. We observed that insulin had an inhibitory effect on adrenomedullin expression in isolated human adipocyte cells. This response was dose dependent and was reversed by resistin, a new anti-insulin agent. We quantified circulating adrenomedullin in healthy obese patients and observed a threefold increase of adrenomedullin compared with lean patients. Furthermore, adrenomedullin plasma levels are negatively correlated to plasma insulin levels in these obese patients. The insulin inhibitory response was also observed in vivo in Sprague-Dawley rats but not in the insulin-resistant Zucker rat, suggesting that adrenomedullin expression is upregulated in insulin-resistant adipose cells. Using adrenomedullin promoter-luciferase reporter gene constructs, we have shown that the adrenomedullin response to insulin is mediated by insulin-responsive elements. These findings provide new insight into fat mass development and the relationship between obesity and elevated circulating adrenomedullin levels in diabetic patients.
Obesity prevalence is rapidly increasing in industrialized countries and is a significant risk factor for many serious illnesses, such as diabetes and other disorders, having insulin resistance as a common pathogenic denominator (1). Obesity is related to energy balance dysregulation, leading to increased fat mass due to hyperplasia and hypertrophy of white adipose cells. Therefore, particular attention is paid to adipokines regulating the developmental program of adipose tissue. We, and others, have recently shown that adrenomedullin belongs to the family of adipokines (2–5). This 52–amino acid peptide was first isolated from a phaeochromocytoma as a vasoactive and cardioprotective factor (6,7).
We have shown that adrenomedullin is expressed and secreted by animal and human adipose tissue in physiologically relevant amounts (3). Moreover, circulating adrenomedullin is elevated in type 2 diabetic patients (8,9). In addition, it has been shown that adrenomedullin inhibits insulin secretion by pancreatic β-cells (10) and stimulates synthesis and secretion of interleukin (IL)-6 (4,11), a cytokine involved in insulin resistance mechanisms (12). Animal models of obesity and diabetes show that adrenomedullin synthesis increases in adipose tissue and is elevated after a high-fat diet (2,5). Furthermore, we previously have shown that insulin induction of adipose cell differentiation leads to an inhibition of expression of adrenomedullin (3).
The molecular events linked to insulin adipogenic action are complex and not entirely understood. The study of established preadipocyte cell lines helped to elucidate the major steps of adipocyte differentiation in vitro, and it can now be followed by analysis of a set of gene markers: well-defined transcription factors and adipocyte enzymes (rev. in 13,14). Several experimental approaches have revealed a critical adipogenic role for the CCAAT/enhancer binding protein α (C/EBPα) and the peroxisome proliferator–activated receptor (PPAR)γ2 isoform (14). The earlier molecular events trigger the expression of later adipocyte-specific genes that generate the metabolic functions of the mature adipocyte. The adipocyte-specific genes commonly studied are genes encoding proteins associated with lipogenic function such as lipoprotein lipase, lipolysis-related enzymes such as hormone-sensitive lipase (HSL), and, especially, the adipocyte-selective fatty acid binding protein aP2/FABP4 (14).
All together, these observations led us to investigate the role of adrenomedullin in adipocyte differentiation and the mechanism implicated in the insulin-induced inhibition of adrenomedullin expression. Adrenomedullin regulation of adipose cell differentiation was assessed using adrenomedullin treatment of differentiating 3T3-F442A cell line and human adipose tissue stromal cells. This putative role of adrenomedullin was further investigated by generating 3T3-F442A clones overexpressing or knocked down for adrenomedullin. In addition, we investigated adrenomedullin gene expression regulation by insulin both in vivo and in vitro. Using a luciferase reporter gene assay in 3T3-F442A cells, we investigated 20 kb of the 5′-regulatory sequence of the adrenomedullin promoter for the putative transcriptional regulation through interaction with insulin responsive elements (IREs).
RESEARCH DESIGN AND METHODS
Animal housing and treatment.
Male Sprague-Dawley rats (aged 17 weeks, weight 401 ± 22 g, and glycemia 96 ± 10 mg/dl) were obtained from Harlan Laboratory. Obese (fa/fa) male Zucker diabetic fatty (ZDF) rats (aged 17 weeks, weight 490 ± 35 g, and glycemia 148 ± 13 mg/dl) were obtained from Charles River Laboratories. Animals were housed in a conventional animal room. Rats were intraperitoneally injected with 2 IU insulin (Sigma) per rat or saline and killed 6 h after injection. Visceral intraperitoneal fat depot was recovered to isolate mature adipocytes. The experiments were conducted following ethical European Union guidelines.
Cell culture and isolation of mature adipocytes.
Human subcutaneous abdominal adipose tissue was obtained from healthy female patients (aged 42 ± 9 years, BMI 28.0 ± 5.4 kg/m2, n = 10) undergoing plastic surgery. Isolated mature adipocytes were obtained according to the method of Rodbell (15). Human adipocytes (1 million cells) were incubated in Dulbecco’s modified Eagle’s medium supplemented with 10% FCS. 3T3-F442A preadipocytes were seeded in 12-well plates (8.103 cells per well) and cultured as previously described (3). 3T3-F442A preadipocytes were grown in Dulbecco’s modified Eagle’s medium containing 10% donor calf serum (Life Technologies). Adipocyte differentiation was initiated at confluence by insulin (50 nmol/l) and 10% of FCS. The isolation of human adipose tissue–derived stromal cells and the culture of stromal preadipocytes differentiated into adipocytes were performed as previously described (16). Murine and human adrenomedullin were purchased from Phoenix Pharmaceuticals.
Real-time quantitative PCR.
Gene expression analysis was performed with an ABI Prism 7000 (Applied Biosystems). cDNA were synthesized from 600 ng of total RNA using Thermoscript reverse transcriptase (Invitrogen). Primer design, real-time PCR, and data analysis were performed as previously reported (3).
Adrenomedullin 5′-regulatory elements–luciferase reporter constructs.
The minimal promoter of the adrenomedullin gene was inserted into the HindIII site of the promotorless pGL3 basic luciferase expression vector (Promega). Mouse adrenomedullin gene 5′-regulatory elements were amplified by PCR using Isis (Qbiogene) and subcloned into the pGL3AM plasmid in the KpnI site located upstream of the adrenomedullin minimal promoter. The DNA sequence and orientation of the inserted fragments were checked by using the Big Dye Terminator v 3.1 sequencing kit (Applied Biosystems). Primers (Proligo) used to obtain the regulatory fragments were 5′-CCCAAGCTTAGGCAGTCAGGTCAACTCGC-3′ (sense) and 5′-CCCAAGCTTAAGAAGTCCTTCTGACAGTAGC-3′ (antisense) for the −761/+124 segment and 5′-CGGGGTACCTGACCTTCATTTGGAGTACCCG-3′ (sense) and 5′-CGGGGTACCGCAATTCTTCCTCTCTCTAAGC-3′ (antisense) for the −7,413/−5,379 fragment, with +1 referring to the transcription start site.
Plasmids transfections, luciferase, and β-galactosidase assays.
3T3-F442A cells (3.106) were mixed with 25 μg of pGL3 basic–derived plasmids and 9 μg of pGK-LacZ control vector. Cells were electroporated by a single pulse (200 V, 192 Ohms, 900 μF) and were allowed to grow in Dulbecco’s modified Eagle’s medium containing 10% donor calf serum alone or supplemented with insulin for 48 h before cell lysis for reporter assays. Luciferase activity was assayed with the luciferase assay system (Promega) according to the manufacturer’s protocol. Data were normalized to β-galactosidase activity determined as previously described (17).
Cloning, transfection, and selection of 3T3-F442A stable clones conditionally overexpressing adrenomedullin.
The reversible tetracycline transactivator system (18) was used to produce inducible expression of adrenomedullin. The coding sequence of the mouse mRNA encoding adrenomedullin was amplified using the primers (Proligo) 5′-AAAACTGCAGGCCACCATGAAGCTGGTTTCCATCACCCTGATGTTA-3′ (sense) and 5′-ACGCGTCGACCTATATCCTAAAGAGTCTGGAGATGTGCGC-3′ (antisense) and inserted in the pBI-G Tet expression vector. In parallel, the pUHrT 62–1 DNA plasmid was cotransfected with a PGK-neo plasmid conferring resistance to G418 (Invitrogen) in native 3T3-F442A preadipocytes by the calcium phosphate protocol (19). Transformed cells were selected by 250 μg/ml G418. Several cell colonies, each obtained from a single cell, were harvested and subsequently expanded. To select cells with the highest induction rate of adrenomedullin expression, pBI-G vector was electroporated into the 11 clones obtained. The electroporated cells were cultured in the presence of 250 μg/ml G418 and 500 ng/ml doxycycline (Sigma) to induce LacZ reporter gene expression. Two clones (pUHrT-3 and pUHrT-8) showing the highest levels of β-galactosidase activity were electroporated with the pBI-G plasmid containing the adrenomedullin sequence and a PGK-hygro plasmid conferring resistance to hygromycin B (Invitrogen). Transformed cells were selected by 250 μg/ml G418 and 300 μg/ml hygromycin B. The selected clones were then cultured in presence of 500 ng/ml doxycycline to induce adrenomedullin expression, and adipogenesis was induced following the same procedure as for nontransfected cells.
Adrenomedullin expression knockdown: small interfering RNA preparation and small interfering RNA–expressing clone selection.
Small interfering RNA (siRNA) were designed according to the IMG-800 kit (IMGENEX) and cloned into the pSuppressorNeo vector DNA. The sequences of the oligonucleotides used were 5′-TCGAGAGATACTCCTTCGCAGTTCGAGTACTGGAACTGCGAAGGAGTATCTTTTT-3′ (sense) and 5′-CTAGAAAAAGATACTCCTTCGCAGTTCCAGTACTCGAACTGCGAAGGAGTATCTC-3′ (antisense). Plasmid DNA was transfected by the calcium phosphate protocol in 3T3-F442A preadipocytes at 60% confluence. Transformed cells were selected by 250 μg/ml G418. Colonies were harvested and subsequently expanded. The stability of the transgene was checked by PCR analysis on the genomic DNA of adrenomedullin siRNA clones. Twenty-four-hour-old conditioned media were harvested for radioimmunoassay (RIA) analysis. Clones were maintained and allowed to differentiate as described for wild-type cells.
RIA.
Conditioned media were analyzed using an RIA kit as recommended by the manufacturer (Phoenix Pharmaceuticals) to screen clonal cells for adrenomedullin levels secretion. The intra-assay and interassay coefficients of variance were 9.2% and 12.6%, respectively.
Immunofluorescence confocal laser microscopy.
Cells were seeded onto BD Falcon CultureSlides (BD Biosciences). Subconfluent cells were fixed in 4% formaldehyde for 15 min. Cells were washed 5× in PBS and incubated for 20 min at 37°C in protein block serum-free buffer (Dako) before a 90 min incubation at 37°C with a rabbit anti-adrenomedullin (rat) serum (1:200). Cells were then incubated for 60 min at 37°C with a Texas Red dye–conjugated goat anti-rabbit immunoglobulin secondary antibody (1:200) (Jackson ImmunoResearch laboratories). Cells were washed 5× with PBS and thereafter incubated for 30 s at room temperature with SYBR Green I (Molecular probes) diluted 1:1,000 in PBS. Cells were analyzed on a Zeiss LSM 510 confocal microscope.
Oil Red O staining.
Cells were fixed for 15 min in 10% formaldehyde and stained with Oil Red O (Sigma) for 30 min at room temperature. Dishes were washed twice with water and photographed. Oil Red O was prepared by diluting a stock solution (0.5 g of Oil Red O in 100 ml of isopropanol) with water (3:2), followed by filtration.
Data analysis.
Comparisons between matched pairs were performed using Student’s t test or Mann-Whitney test when equality of variance or normality tests were unsuccessful. Multiple comparisons were performed using one-way ANOVA with post hoc Bonferroni’s test. Correlation analysis were performed using Spearman correlation coefficients (r). P < 0.05 was considered statistically significant.
RESULTS
Adrenomedullin treatment impairs adipogenesis in 3T3-F442A cells and in human preadipocytes.
Daily 100-nmol/l adrenomedullin treatment of differentiating 3T3-F442A and human stromal preadipocytes decreased mRNA levels for all the adipocyte-specific markers tested, such as transcription factors sterol regulatory element–binding protein-1c, C/EBPα, and PPARγ2 and genes encoding proteins implicated in adipocyte-specific functions like HSL and aP2 (Fig. 1,A and B). This lowered differentiation state of human stromal preadipocytes was manifested as a reduction in the clusters of differentiation and the accumulated lipid droplets (Fig. 1C).
Knockdown of adrenomedullin mRNA expression accelerates 3T3-F442A adipocyte differentiation.
Cells stably expressing adrenomedullin siRNA were screened for adrenomedullin secretion levels using RIA. Two clones presented low adrenomedullin secretion levels compared with untransfected cells: adrenomedullin-1 and adrenomedullin-5 were secreting 9 and 13%, respectively, of native 3T3-F442A adrenomedullin levels (Fig. 2A). Clones selected after stable transfection of the empty plasmid (without the siRNA coding sequence) secreted similar levels of adrenomedullin peptide as wild-type cells (Fig. 2A, control), suggesting that reduced adrenomedullin secretion by adrenomedullin-1 or adrenomedullin-5 clones was not a consequence of transfection or selection of clones. The decrease in adrenomedullin peptide synthesis was confirmed for clone adrenomedullin-1 by immunofluorescence. Subcellular localization of adrenomedullin content revealed much less peptide in adrenomedullin-1 cytoplasm compared with native 3T3-F442A cells (Fig. 2B).
Cultured adrenomedullin-1 and adrenomedullin-5 clones revealed acceleration in their kinetics of differentiation. At day 3 of differentiation, siRNA-expressing cells displayed a similar morphology to native 3T3-F442A and control cells. However, by day 5, adrenomedullin-1 and adrenomedullin-5 clones had nonetheless accumulated lipid droplets that were rare in wild-type or control (Fig. 2C). This accelerated differentiation state was manifested as clusters of differentiation rather than as increases in lipid droplets in differentiated cells. This earlier adipocyte differentiation at day 5 was more apparent in clone adrenomedullin-1 than in adrenomedullin-5, and this was well visible with Oil Red O staining of the cells (Fig. 2C). At day 10, siRNA-expressing lines adrenomedullin-1 and adrenomedullin-5 harbored more differentiated clusters than wild-type and control transfected cells (data not shown).
Adrenomedullin mRNA knockdown leads to earlier expression of adipocyte differentiation marker genes.
mRNA levels of adipocyte-specific genes, such as transcription factors as well as the lipolytic pathway proteins, were significantly increased at day 3 in the adrenomedullin-1 clone, whereas the levels of expression seen are found much later in wild-type cells (Fig. 3). The increase in adrenomedullin-1 was more pronounced at day 5, and it also became appreciable in adrenomedullin-5 at this time point (Fig. 3).
Stable overexpression of adrenomedullin in 3T3-F442A clones lowers adipocyte differentiation marker gene expression.
Using stable clone selection, we obtained 3T3-F442A cell clones with different adrenomedullin mRNA levels. Cells were grown in the presence of doxycycline to induce adrenomedullin expression, and real-time PCR analysis of adipocyte-specific genes expression was performed at day 12 of differentiation. PPARγ2 and HSL mRNA levels were inversely correlated with adrenomedullin mRNA levels (Fig. 4).
Insulin downregulates adrenomedullin expression in 3T3-F442A and isolated human adipocytes.
Induction of confluent 3T3-F442A preadipocyte differentiation by 50 nmol/l of insulin resulted in a strong decrease of adrenomedullin expression and secretion (Fig. 5A). Low concentrations of insulin ranging from 10 nmol/l to 100 pmol/l also induced a decrease in adrenomedullin mRNA expression (data not shown). Moreover, a physiological dose of 1 nmol/l IGF-I mimicked insulin action (data not shown). These observations lead us to consider that adrenomedullin expression could be specifically regulated by insulin. Experiments were performed in order to answer to this question.
We checked whether the insulin-induced decrease of adrenomedullin expression observed in Fig. 5A was limited to the 3T3-F442A cell line. Thus, we treated isolated human adipocytes for 18 h with increased insulin concentrations. We observed a dose-dependent downregulation of adrenomedullin mRNA levels (Fig. 5B). Moreover, adrenomedullin expression was not affected by insulin in human adipocytes preincubated for 2 h with recombinant resistin prepared as previously described (20) (Fig. 5C).
Insulin downregulates the expression of the adrenomedullin gene in Sprague-Dawley but not in insulin-resistant rat adipocytes.
The physiological relevance of the above results was investigated in vivo in male 17-week-old Sprague-Dawley and ZDF rats by intraperitoneal injection of 2 IU insulin. We observed a twofold decrease of adrenomedullin mRNA levels in visceral intraperitoneal mature adipocytes isolated 6 h after injection of insulin in Sprague-Dawley rats, whereas insulin failed to downregulate adrenomedullin mRNA expression in visceral adipocytes isolated from obese (fa/fa) ZDF rats (Fig. 5D).
Adrenomedullin and IL-6 mRNA levels are positively correlated during adipocyte differentiation.
The analysis of IL-6, a well-known adipocytokine involved in insulin resistance (12), and adrenomedullin synthesis in 3T3-F442A preadipocytes during differentiation by qPCR revealed a positive correlation between IL-6 and adrenomedullin mRNA synthesis upon differentiation (r = 0.68, P < 0.0001) (Fig. 5E).
The inhibitory effect of insulin on adrenomedullin gene expression is mediated through IREs in the adrenomedullin promoter 5′ flanking regulatory sequence.
To investigate possible regulation of the adrenomedullin gene by insulin at the transcriptional level, we searched within 20 kb upstream of the transcription start point of the adrenomedullin gene for homologies on both strands with known consensus DNA sequences for positive- and negative-acting IREs. We found strong consensus IREs in the minimal promoter region of the adrenomedullin gene and in a DNA segment upstream of the adrenomedullin gene (Fig. 6A). Luciferase reporter gene constructs containing the minimal promoter region of adrenomedullin alone or with the upstream DNA segment in both orientations were then transiently transfected into 3T3-F442A preadipocytes (Fig. 6B). The activity of the minimal promoter region spanning −761 and +124 was significantly induced by ∼40% in the presence of 5 nmol/l insulin (Fig. 6C). This positive effect may be mediated by the putative serum response element (SRE) (Fig. 6A), a positive consensus IRE located between −732 and −723 (the adrenomedullin transcription start point is taken as +1). The addition of the −7,413/–5,379 5′-regulatory DNA segment resulted in a significant inhibition of luciferase activity of the minimal promoter under insulin treatment. This inhibitory effect was obtained when the DNA segment was inserted in both orientations, although it was slightly stronger when in the reverse orientation (Fig. 6C). This downregulation could be mediated by the presence of two typical PEPCK-like negative IRE motifs (Fig. 6B).
Adrenomedullin plasma levels are increased in obese compared with nonobese subjects.
Ten obese and 10 nonobese subjects were included in the following study after they gave informed consent. All subjects underwent careful clinical and biological examinations performed by clinicians of the Rangueil Hospital, Toulouse, France. The recorded clinical parameters guaranteed that obese and lean patients were not suffering from any disease or common obesity-associated disorders like arterial hypertension, heart failure, type 2 diabetes, or dyslipidemia (Table 1). Plasma C-reactive protein levels were higher in the obese group, accordingly to the increased fat mass. Insulinemia, glycemia concentration, and homeostasis model assessment values were in the normal range and similar in the two groups of patients. Mean plasma adrenomedullin level was 3.6-fold higher in the obese group than in the control group (Table 1).
Adrenomedullin and insulin plasma levels are negatively correlated in obese subjects.
Quantification of adrenomedullin and insulin plasmatic levels from the group of uncomplicated obese patients (described in Table 1) displayed a high negative correlation (r = −0.85; P < 0.002).
DISCUSSION
Adrenomedullin peptide synthesis and secretion during 3T3-F442A differentiation have a particular kinetic pattern (Fig. 5A): Confluent preadipocytes express and secrete high levels of this peptide just before insulin induction of adipocyte differentiation. Differentiation is reflected in a strong and prolonged decrease of adrenomedullin gene expression. These observations lead us to speculate that adrenomedullin expression could be regulated by insulin and that adrenomedullin could be an inhibitor of adipocyte differentiation. We tested first this last hypothesis by chronically treating differentiating human stromal and 3T3-F442A preadipocytes with adrenomedullin, which showed a decrease of adipose cell differentiation markers and clusters of differentiation (Fig. 1), suggesting that adrenomedullin inhibits adipogenesis. To confirm this role of adrenomedullin, we investigated the consequences of adrenomedullin gene knockdown and overexpression. We first used the reverse tetracycline transactivator system to induce the overexpression of adrenomedullin in 3T3-F442A adipocytes. Several clones presented variable adrenomedullin mRNA levels under doxycycline induction. Real-time PCR analysis of the differentiated clonal cells revealed an inverse correlation between adrenomedullin and adipocyte marker mRNA levels. The clones presenting the highest adrenomedullin mRNA levels seemed to be resistant to differentiation with fewer clusters of differentiation (data not shown). Stable siRNA knockdown of adrenomedullin expression in 3T3-F442A cells generated several clones with various adrenomedullin levels. Among the clones selected, we obtained two potentially interesting clones secreting very low adrenomedullin levels. The two clones presented an impressive accelerated kinetic of differentiation. The effect was clearly visible under the microscope or by Oil Red O analysis, and real-time PCR measurements confirmed a significant increase of mRNA levels for all adipocyte-specific markers tested. The adrenomedullin-1 clone had the highest adipogenic properties and showed a strong increase in mRNA expression for the major adipocyte-specific transcription factors PPARγ2 and C/EBPα at days 3 and 5. The knockdown of the gene encoding adrenomedullin greatly facilitates the early phase of terminal differentiation. Overall, these results confirmed that adrenomedullin has the capacity to slow adipocyte differentiation. However, the molecular mechanisms of this effect need to be elucidated. Inflammatory cytokines, including tumor necrosis factor-α and IL-6, have been shown to inhibit insulin-induced adipocyte differentiation in 3T3-L1 cells by blocking induction of PPARγ and C/EBPα (21). Adrenomedullin recently has been shown to induce IL-6 expression in mesenchymal stem cell–derived human adipocytes (4). Furthermore, we observed a significant positive correlation between adrenomedullin and IL-6 expression during adipocyte differentiation (Fig. 5E). Therefore, it could be speculated that adrenomedullin inhibits adipocyte differentiation by promoting IL-6 production.
A 24-h treatment of confluent preadipocytes with IGF-I also resulted in a significant decrease in adrenomedullin mRNA expression (data not shown). However the ability of physiological concentrations of insulin (100 pmol/l to 10 nmol/l) to reduce adrenomedullin mRNA expression suggested that this effect may be mediated through insulin receptor activation.
Insulin-mediated adrenomedullin downregulation was also observed by Fukai et al. (2) in the 3T3-L1 model, another murine preadipocyte cell line, in which insulin also is commonly used to induce adipocyte differentiation. Interestingly, downregulation of adrenomedullin mRNA expression by insulin also was detected in a dose-dependent manner in human mature adipocytes. Moreover, this effect was abolished by pretreating adipocytes with resistin, an adipocyte hormone that was shown to attenuate multiple effects of insulin in 3T3-L1 adipocytes (22). The treatment of male Sprague-Dawley rats by insulin also resulted in a significant decrease of adrenomedullin mRNA expression in mature adipocytes isolated from visceral adipose tissue, making the observations performed on isolated human adipocytes physiologically relevant. We also showed that plasmatic adrenomedullin levels are negatively correlated to plasmatic insulin levels in patients with uncomplicated obesity but not in lean patients. Thus, as insulin levels are elevated, inhibition of adrenomedullin expression in adipose cells is stronger and adrenomedullin plasmatic levels are lowered. The insulin-resistant state that inexorably develops during obesity could be the source of the observed plasmatic adrenomedullin upregulation in diabetic patients (Fig. 7) (8,9,23). This is in accordance with the fact that insulin failed to inhibit adrenomedullin expression in obese and insulin-resistant ZDF rat adipocytes (Fig. 5D). This could also be supported by the published data (24) demonstrating adrenomedullin resistance in these obese rats but not in their lean counterparts and by the rise in adrenomedullin expression levels in fat from animal obesity models like the genetically obese ob/ob mice (2,5). This state could explain the decrease of catecholamine-induced lipolytic rates in subcutaneous adipose tissue of obese subjects, because we previously have shown that adrenomedullin inhibits lipolysis through β-adrenergic agonist oxidation (3). We previously did not notice any lipolytic effect of adrenomedullin (3), whereas Linscheid et al. (4) observed a glycerol release from human mesenchymal stem cell–derived adipocytes subjected to an extended adrenomedullin treatment (24 h). But as mentioned above, this group also reported the induction of IL-6 expression when the cells were subjected to adrenomedullin. Since IL-6 is thought to stimulate lipolysis in humans (25,26), the reported effect may be indirectly caused by adrenomedullin-induced IL-6 expression. The consequences could be extended throughout the body since adipose tissue is now proposed to be one of the major sites of adrenomedullin production and is age and BMI dependent (27,28). However, these data displayed some discrepancy on whether the main secreting adipose tissue was omental or subcutaneous. An elevation of adrenomedullin circulating levels caused by insulin resistance (Fig. 7) could be involved in the development of some pathophysiological states encountered in severe obesity. Whether adipocyte-derived adrenomedullin promotes or reduces the development of obesity-associated diseases remains to be investigated.
Insulin is known to affect mRNA transcription of several genes by the recruitment of trans-acting factors to specific promoter sequences termed insulin response elements (29). We therefore speculated that the adrenomedullin gene 5′-regulatory DNA sequence may contain IREs that regulate adrenomedullin mRNA expression in adipocytes. Sequence analysis of 20 kb of 5′ flanking sequence of the adrenomedullin gene revealed the presence of putative positive and negative IRE consensus motifs. We used a luciferase reporter gene to test a set of constructs with the DNA segments containing the IREs. The SRE, through its interaction with many growth factors in serum, has been shown to induce a positive transcriptional effect by insulin on genes like c-fos (30) and β-actin (31). The SRE motif detected at position −732/−723 could be related to the significant increase in luciferase activity seen with the −761/+124 DNA segment. However, the addition of −7,413/−5,379 to this construct resulted in a significant inhibition of luciferase activity after insulin treatment. The fact that this inhibitory effect was obtained when the −7,413/−5,379 promoter fragment was inserted in both orientations further strengthens the involvement of cis-acting elements. This region contains two PEPCK-like motifs T(G/A)TTTT(G/T). This consensus heptanucleotide sequence has been linked to the inhibition of transcription of several genes by insulin, including the genes encoding PEPCK (32), IGFBP-1 (32), glucose-6-phosphatase (33), fatty acid transport protein (34), and also the adipose-specific glycerol channel aquaporin adipose (35). Moreover, it is interesting to note that insulin is hypothesized to mediate its negative effect on PEPCK and IGFBP-1 by inhibiting glucocorticoid-stimulated gene transcription (29). This mechanism could influence adrenomedullin since glucocorticoids have been shown to upregulate adrenomedullin gene expression both in vitro and in vivo (36,37).
In conclusion, this work reinforces the potential role of adrenomedullin as a crucial agent in the regulation of adipose tissue metabolism and development. It also underlines the possible existence of a tight regulation loop between adrenomedullin produced in adipose tissue and insulin. The effects of adipocyte-derived adrenomedullin on pancreatic functions during obesity are potentially interesting fields of investigation. Efforts should be carried out in the development of pharmacological tools involving adrenomedullin mimetic drugs to better understand its biological functions and its impact on body homeostasis.
Clinical parameters . | Normal range . | Lean group . | Obese group . |
---|---|---|---|
BMI (kg/m2) | 20<X<25 | 21.6 ± 1.7 | 38.1 ± 5.9* |
n (male/female) | 6/7 | 7/6 | |
Age (years) | 34 ± 6 | 33 ± 5 | |
Blood pressure (mmHg) | X<140/90 | 118 ± 2/74 ± 3 | 121 ± 4/72 ± 2 |
Heart rate (bpm) | 64 ± 3 | 66 ± 6 | |
C-reactive protein (mg/l) | X<10 | 9.4 ± 2.1 | 13.7 ± 5.4† |
A1C (%) | 4<X<6 | 5.0 ± 0.3 | 5.3 ± 0.5 |
C-peptide (ng/ml) | 0.5<X<3 | ND | 3.2 ± 0.9 |
Insulin (μU/ml) | 2<X<17 | 8.00 ± 1.00 | 8.97 ± 0.98 |
HOMA of insulin sensitivity | 1.87 ± 0.34 | 2.00 ± 0.24 | |
Glycemia (mmol/l) | 3.9<X<5.8 | 4.87 ± 0.55 | 5.00 ± 0.49 |
Triglycerides (mmol/l) | 0.6<X<1.7 | 1.08 ± 0.39 | 1.34 ± 0.40 |
Total cholesterol (mmol/l) | 3.8<X<6.2 | 4.8 ± 0.8 | 5.6 ± 0.6 |
HDL (mmol/l) | 1.25<X<1.90 | 1.45 ± 0.36 | 1.25 ± 0.34 |
LDL (mmol/l) | 2.5<X<4 | 2.6 ± 0.5 | 3.4 ± 0.8 |
Creatinine (μmol/l) | 44<X<106 | 80 ± 9 | 80 ± 14 |
Adrenomedullin (pmol/l) | 17.1 ± 2.4 | 61.6 ± 12.5* |
Clinical parameters . | Normal range . | Lean group . | Obese group . |
---|---|---|---|
BMI (kg/m2) | 20<X<25 | 21.6 ± 1.7 | 38.1 ± 5.9* |
n (male/female) | 6/7 | 7/6 | |
Age (years) | 34 ± 6 | 33 ± 5 | |
Blood pressure (mmHg) | X<140/90 | 118 ± 2/74 ± 3 | 121 ± 4/72 ± 2 |
Heart rate (bpm) | 64 ± 3 | 66 ± 6 | |
C-reactive protein (mg/l) | X<10 | 9.4 ± 2.1 | 13.7 ± 5.4† |
A1C (%) | 4<X<6 | 5.0 ± 0.3 | 5.3 ± 0.5 |
C-peptide (ng/ml) | 0.5<X<3 | ND | 3.2 ± 0.9 |
Insulin (μU/ml) | 2<X<17 | 8.00 ± 1.00 | 8.97 ± 0.98 |
HOMA of insulin sensitivity | 1.87 ± 0.34 | 2.00 ± 0.24 | |
Glycemia (mmol/l) | 3.9<X<5.8 | 4.87 ± 0.55 | 5.00 ± 0.49 |
Triglycerides (mmol/l) | 0.6<X<1.7 | 1.08 ± 0.39 | 1.34 ± 0.40 |
Total cholesterol (mmol/l) | 3.8<X<6.2 | 4.8 ± 0.8 | 5.6 ± 0.6 |
HDL (mmol/l) | 1.25<X<1.90 | 1.45 ± 0.36 | 1.25 ± 0.34 |
LDL (mmol/l) | 2.5<X<4 | 2.6 ± 0.5 | 3.4 ± 0.8 |
Creatinine (μmol/l) | 44<X<106 | 80 ± 9 | 80 ± 14 |
Adrenomedullin (pmol/l) | 17.1 ± 2.4 | 61.6 ± 12.5* |
Data are means ± SE, unless otherwise indicated. Ten obese and 10 healthy nonobese subjects were included in the following study after they gave informed consent. All subjects underwent careful clinical and biological examinations performed by clinicians of the Rangueil Hospital, Toulouse, France. Obese patients were not suffering from any common obesity-associated disorders like arterial hypertension, heart failure, type 2 diabetes, or dyslipidemia. Homeostasis model assessment (HOMA) was used to evaluate insulin sensitivity, according to the following formula: HOMA of insulin sensitivity = fasting insulin (μU/ml) × fasting glucose (mmol/l)/22.5.
P < 0.01 and
P < 0.05 vs. lean group.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Article Information
We are indebted to Dr. Peter Romanienko (Memorial Sloan-Kettering Cancer Center, New York) for critical reading of the manuscript. The authors thank Marie-Adeline Marques for isolating human mature adipocytes.