Characterization of the long pentraxin PTX3 as a TNF�induced, secreted protein of adipose cells

Exposure of preadipocytes to long-chain fatty acids induces the expression of several markers of adipocyte differentiation. In an attempt to identify novel genes and proteins that are regulated by fatty acids in preadipocytes, we performed a substractive hybridization screening and identified PTX3, a protein of the pentraxin family. PTX3 mRNA expression is transient during adipocyte differentiation of clonal cell lines and is absent in fully differentiated cells. Stable overexpression of PTX3 in preadipocytes has no effect on adipocyte differentiation. In line with this, PTX3 mRNA is expressed in the stromal-vascular fraction of adipose tissue, but not in the adipocyte fraction; however, in 3T3-F442A adipocytes, the PTX3 gene can be reinduced by tumor necrosis factor alpha (TNFalpha) in a dose-dependent manner. This effect is accompanied by PTX3 protein secretion from both 3T3-F442A adipocytes and explants of mouse adipose tissue. PTX3 mRNA levels are found to be higher in adipose tissue of genetically obese mice versus control mice, consistent with their increased TNFalpha levels. In conclusion, PTX3 appears as a TNFalpha-induced protein that provides a new link between chronic low-level inflammatory state and obesity.

An increasing body of evidence correlates the obese phenotype with chronically elevated systemic levels of acute-phase reactants and inflammatory cytokines (1)(2)(3).These elevated levels may in turn contribute directly or in-directly to the increased incidence of cardiovascular diseases (4).A significant correlation has been found between body mass index and circulating levels of C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor ␣ (TNF ␣ ) (5)(6)(7).Recent evidence shows that adipocytes secrete these cytokines as well as inflammatory proteins, i.e., serum amyloid A3 and haptoglobin (8)(9)(10).Although the role of these reactants has remained largely unknown, TNF ␣ and IL-6 have been shown to inhibit insulin signaling and to induce both hypertriglyceridemia and endothelial activation (11)(12)(13).Altogether, these observations suggest strongly that adipose tissue through adipocytes is an important determinant of a chronic lowlevel inflammatory state (14)(15)(16).Clearly, secretion of cytokines in adipose tissue is not confined to adipocytes, as we have previously reported that preadipocytes secrete leukemia inhibitory factor (LIF) transiently at a time when their cognate cell surface receptor (LIF-R) is also expressed (17).The LIF/LIF-R system then promotes adipocyte differentiation via the activation of extracellular signal-regulated kinases that mediate the expression of CCAAT/enhancer binding proteins ␤ and ␦ (18), and LIF production ceases with the ongoing differentiation process.During the course of our investigations in searching early-expressed, fatty acid-responsive genes, we have identified a novel mRNA that encodes a long form of pentraxin, i.e., PTX3.Up to now, PTX3, also known as TSG14, has been known as a member of the pentraxin gene family that is expressed at extrahepatic sites, predominantly in vascular endothelial cells of heart and skeletal muscle (19)(20)(21).We show herein that, despite the fact that PTX3 expression becomes low, if any, in adipocytes, its expression and secretion are induced dramatically during TNF ␣ exposure.We show also that expression of the PTX3 gene is elevated in mouse models of monogenic obesity, strongly suggesting that PTX3 provides a new link between the chronic low-level inflammatory state and obesity.

Cell culture
Cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 8% fetal bovine serum, 200 U/ml penicillin, and 50 g/ml streptomycin (standard medium) to confluence (hereafter referred to as Day 0).For differentiation, confluent 3T3-F442A cells were shifted to standard medium supplemented with 5 nM insulin and 2 nM triiodothyronine (differentiation medium).For experiments in serum-free medium, cells were first inoculated in standard medium at a density of 10 3 cells/cm 2 and washed 24 h later with DMEM-Ham's F12 (50:50, v/v).Cells were grown to confluence in 4-F medium (22) consisting of DMEM-Ham's F12 supplemented with insulin (5 g/ml), transferrin (10 g/ml), submaxillary gland extract (SMGE) (2 g/ml), and fibroblast growth factor (25 ng/ml).At confluence, cells were shifted to the same medium in the absence of fibroblast growth factor and SMGE, supplemented with fetuin (100 g/ml) growth hormone (2 nM), triiodothyronine (0.2 nM), BRL 49653 (rosiglitazone, 10 nM), ascorbate (100 M), and selenite (20 nM).Confluent 3T3-L1 cells were stimulated to differentiate by addition of hormonal cocktail [0.5 mM 1-methyl-3isobutylmethyl-xanthine, 0.5 M dexamethasone, and 170 nM insulin (MDI)] to the standard medium for 2 days and were maintained thereafter in standard medium with only 170 nM insulin.Typically, by Day 6, more than 95% of the cells fully differentiate into adipocytes.Media were changed every other day.

Substractive hybridization
Substractive hybridization of RNA from untreated and fatty acid-treated Ob1771 preadipocyte cells was performed by the use of a substraction kit (Stratagene, The Netherlands) according to the manufacturer's instructions.

Retrovirus production and transduction
BOSC23 cells were transfected with viral DNA at 50% confluence by the use of FuGENE 6 reagent (Roche Molecular Biochemicals, France) as described by the manufacturer.Forty-eight hours after transfection, viral supernatant was harvested, centrifuged, and filtered.Fifty percent confluent 3T3-F442A cells were transduced with viral supernatant diluted with one volume of fresh standard medium in the presence of 6 g/ml Polybrene.On the following day, cells were split and subjected to neomycin (Sigma) selection (200 g/ml).Stable cell populations were obtained after 7-10 days of selection.After confluence, cells were maintained in differentiation medium.

Culture of adipose tissue explants
Adipose tissue from epididymal fat pads were dissected under sterile conditions, freed as much as possible from blood capillaries, and bathed in DMEM medium containing antibiotics as above.Adipose tissue was then minced into ‫ف‬ 100 mm 3 pieces and incubated in differentiation medium for 2 h to 15 h.
Stromal-vascular and adipocyte fractions were obtained as already described in (23).

Secretion media
Adipose tissue explants or differentiated 3T3-F442A cells were washed twice in PBS (pH 7.4), once with ITT medium (DMEM-Ham's F12 supplemented with 5 nM insulin, 2 nM triiodothyronine, and 10 g/ml transferring), and then incubated in ITT medium for 2 h.Cytokines were added for 3 h and media were collected.Secretion media were centrifuged for 5 min at 2,000 rpm at 4 Њ C, then proteins were concentrated either by acetone precipitation or by centrifugation with Amicon filters as described by the manufacturer (Millipore, France).

Isolation and analysis of RNA
Total RNA from cells and mouse adipose tissue was extracted using Tri-Reagent™ kit (Euromedex, France) according to manufacturer's instructions and analyzed by Northern blot as described previously (24).Blots were subjected to digital imaging (FujixBAS 1000).

Plasmids
The retroviral construct containing full-length PTX3 cDNA was derived from pSG5-PTX3 and cloned into the Not I site of pAkvi-Bipe2 (gift of Dr. K. Kristiansen, University of Odense, Denmark).

Oil red O staining
Culture dishes were washed in PBS (pH 7.4) and cells were fixed in 3.7% formaldehyde for 1 h, followed by staining with Oil Red O for 1 h.Oil Red O was prepared by diluting a stock solution (0.5 g of Oil Red O) (Sigma) in 100 ml of isopropanol with water (60:40, v/v) followed by filtration.After staining, plates were washed twice in water and photographed.
Western blotting.Fifty to one hundred micrograms of protein were loaded in each lane.After SDS-polyacrylamide gel electrophoresis, proteins were blotted onto polyvinylidene difluoride membranes (Amersham Biosciences, France) using a Bio-Rad semidry blotter.Equal loading/transfer was confirmed by Ponceau S staining of membranes.Membranes were blocked overnight in TBS [10 mM Tris-Cl (pH 7.5), 150 mM NaCl] containing 0.1% Tween 20 (TBS-T) and 5% nonfat dry milk.Incubation with primary and secondary antibodies was performed in TBS-T containing 5% nonfat dry milk for 2 h at room temperature.After incubation with antibodies, membranes were washed in TBS-T.Rat anti-human PTX3 was already described (25), and cross-reacted well with the mouse PTX3.Anti-TBP (Santa-Cruz) was used to monitor the equal loading.Secondary antibodies were horseradish peroxidase-conjugated anti-rat (Jackson Laboratories) or anti-rabbit antibodies (Promega).Enhanced chemiluminescence (Amersham Biosciences) was used for detection.Stripping of membranes was done by boiling for 5 min in water.

Materials
Culture media, fetal calf serum, cytokines, and geneticin were from Life Technologies, Inc. (France).Other chemicals were purchased from Sigma and Aldrich (France).Radioactive materials, random priming kit, and nylon membranes were from Amersham Biosciences.BRL 49653 was a kind gift (S.Michel and U. Reicher, Galderma RandD, Sophia-Antipolis, France).Mice (C57BL/6J, ob/ob, db/db) were purchased from Janvier (Le Genest Saint Isle, France) and used at 6-8 weeks of age.Animals were killed by cervical dislocation.All experimental protocols were performed in accordance with the recommendations of the French Accreditation of Laboratory Animal Care.

Identification of PTX3 cDNA and deduced amino acid sequence
To search for new genes that are expressed early in preadipocytes and are responsive to fatty acids, we employed a substractive hybridization technique.Two cDNA libraries were obtained from RNA prepared either from Day 2 postconfluent Ob1771 cells (26) or from RNA of Day 1 postconfluent Ob1771 cells exposed to 300 M ␣ -linolenate for 24 h.After substractive hybridization screening among various clones, we identified a clone that contained a cDNA fragment of about 1.7 kb.Using this DNA fragment as a probe for Northern blot analysis, a single transcript around 2 kb was observed at higher levels in untreated and at very low levels in fatty acid-treated preadipocytes ( Fig. 1A ).As shown previously (24), aP2 mRNA expression was, in contrast, strongly induced by fatty acid treatment (Fig. 1A).A nearly full-length cDNA clone was subsequently sequenced, and Blastn searching of nucleic acid databases at the National Center for Biotechnology Information to look for sequence homologies (27) identified the cDNA sequence as identical to mouse PTX3 (GenBank™ accession number X83601).The PTX3 gene encodes a 42 kDa glycoprotein with a carboxy-terminal half that shares high homology with the entire sequence of CRP and serum amyloid protein, which are acute-phase proteins of the pentraxin family (28)(29)(30), whereas the NH2-terminal part does not show any significant homology with other known proteins.

Expression of PTX3 mRNA during adipocyte differentiation
White and brown adipose tissues were shown to express high and low levels of PTX3 mRNA in the stromal-vascular fraction and the adipocyte fraction, respectively.In contrast to PTX3, aP2 and peroxisome proliferator-activated receptor ␥ (PPAR ␥ ) were predominantly expressed in the adipocyte fraction (Fig. 1B, and not shown).These observations led us to examine PTX3 expression as a function of differentiation of 3T3-F442A (31), 3T3-L1 (32), and Ob1771 cells.As shown in Fig. 2A , PTX3 mRNA was already expressed in growing cells but became expressed at higher levels in confluent 3T3-F442A cells, whereas its expression dramatically decreased thereafter.Its kinetics of expression appeared opposite to those of aP2 and PPAR ␥ , which characterize the terminal differentiation process (33,34).Identical results were obtained with 3T3-L1 (Fig. 2A) and Ob1771 cells (data not shown).PTX3 mRNA expression appeared as a differentiation-related event as the message was maintained in the absence of induction of differentiation of 3T3-L1 cells (Fig. 2A), and a weak and continuous expression was observed in 3T3-C2 cells (data not shown), a fibroblast nonadipocyte clonal line (35).Consistent with this interpretation, in Ob1771 cells exposed to growth hormone or to fatty acids, in order to accelerate the differentiation process, the disappearance of PTX3 mRNA is faster than that observed in untreated cells (data not shown).The same pattern of expression was obtained in 3T3-F442A cells differentiated in a serumfree medium (Fig. 2B).This result excluded nonspecific effects of serum components, including cytokines, and favored the existence of a differentiation-specific event.Thus, the finding that PTX3, a member of the long pentraxin family, was transiently expressed in preadipocytes prompted us to examine its role, if any, in adipocyte differentiation.

Differentiation of stable 3T3-F442A cells overexpressing PTX3
Overexpression of PTX3 in 3T3-F442A cells was accomplished by retroviral infection as shown in Fig. 3 control cells, i.e., cells infected with the empty vector expressed the endogenous PTX3 mRNA at the expected size of 2 kb.Cells infected with the retroviral vector containing the PTX3 coding sequence expressed higher levels of PTX3 protein, and both the 2 kb signal and a 4 kb signal corresponding to the viral transcript were observed (Fig. 3A,   B).The expression of the 4 kb mRNA remained unchanged throughout the postconfluent period, consistent with similar levels of PTX3 protein during differentiation (Fig. 3A).As shown in Fig. 3B and C, adipocyte differentiation of PTX3-overexpressing cells was unimpaired compared with that of control cells, as assessed by microscope examination and expression of differentiation-specific markers, excluding a direct role of PTX3 in the differentiation process.However, owing to the involvement of PTX3 in the inflammatory process and its postulated role in the defense against bacterial infections (36)(37)(38), studies were performed with 3T3-F442A adipocytes in response to inflammatory cytokines, i.e., TNF ␣ , as PTX3 expression, which is also termed TSG-14, is highly induced by TNF ␣ , and as both human and mouse PTX3 promoters are known to be responsive to TNF ␣ (39,40).

Modulation by TNF ␣ of PTX3 mRNA and protein expression
Upon exposure of differentiated 3T3-F442A cells to TNF ␣ present in the nanomolar range of concentrations, the induction of the PTX3 gene occurred within 1 h.Peak levels were reached within 3 h and decreased somewhat thereafter ( Fig. 4 ).Interestingly, IL-6 and LIF were inactive at concentrations up to 1 nM, in contrast to lipopolysaccharide (LPS) which is active at 100 ng/ml.3T3-F442A cells were able to respond to TNF ␣ at concentrations as low as 0.01 nM (Fig. 4).Likewise, PTX3 protein expression was enhanced following TNF ␣ treatment.PTX3 protein was clearly detectable by Western blot analysis of incubation media from differentiated 3T3-F442A adipocytes as well as from adipose tissue explants ( Fig. 5 ).Collectively, these data emphasize that PTX3 is a TNF ␣induced protein secreted from adipocytes.As TNF ␣ levels were reported to be elevated in adipose tissue of genetically obese mice, PTX3 levels were examined in these animals.

Enhanced expression of PTX3 gene in adipose tissue of genetically obese mice
Real-time RT-PCR analysis of RNA from epididymal fat pads of obese (ob/ob), and obese diabetic (db/db) mice was performed and compared with that of the corresponding control mice.The pattern of expression shown in Fig. 6 shows a higher level of PTX3 expression in adipose tissue of genetically obese mice versus control mice, although it is difficult to estimate the relative contribution of stromalvascular cells and adipocytes in each tissue sample.Quantitative analysis allowed an average increase of 5-and 15-fold in ob/ob and db/db mice, respectively, compared with wild-type mice.An increase in the PTX3 level was also observed in Agouti mice compared with control mice (data not shown).However, when one considers i) the lack of PTX3 gene expression in the heart, which contains, like adipose tissue, both endothelial cells and various other cell types reported to express PTX3 (fibroblasts, monocytes/ macrophages), and ii) the enhanced expression of locally produced TNF␣ in adipose tissue of obese rodents or humans (7,41), it is likely that adipocytes are responsible for the observed up-regulation of PTX3 gene expression.By microarray techniques, Friedman and coworkers (42) have also reported an increase in PTX3 mRNA levels in adipose tissue of ob/ob mice compared with their lean littermates.

DISCUSSION
As early events of adipogenesis are critical to gaining insights into subsequent steps, our aim was to identify novel genes and proteins that are differentially regulated by nutrients in preadipocytes.For that purpose, substractive hybridization was used in early confluent preadipocytes treated or not with long-chain fatty acids (24).Unexpectedly, our strategy led to the identification of PTX3, a member of the long form of pentraxin proteins involved in inflammatory processes.The transient expression of PTX3 in the early 3T3-F442A preadipocytes period remains unclear, as its constitutive expression in infected cells shows no impairment of adipocyte differentiation.In this respect, it is of interest to note that myogenic differentiation gives rise to the expression of neuronal pentraxin 1 (NP1), another long form of the pentraxin family whose expression is induced by MyoD (43), and it is tempting to postulate that PTX3 and NP1 share common but still uncharacterized features in the commitment of mesoderm precursor cells to a specific lineage, adipoblast or myoblast, respectively.The identification of the signaling pathways involved in preadipocytes for the transient expression of PTX3 remains unknown.Cytokines would appear as putative effectors, but LIF can clearly be excluded as, despite the fact the LIF/LIF-R system is active in preadipocytes and that LIF-R remains expressed in adipocytes, LIF shows no effect, in contrast to TNF␣, on the expression of PTX3 in adipocytes.Of note is the fact that LIF and IL-6, which share a common gp130-related signaling pathway, do not induce activation of the PTX3 gene (44).The expression of PTX3 in preadipocytes can be linked to the   macrophage-like functions of preadipocytes, consistent with the involvement of adipose tissue in immune response during inflammation (45).
PTX3 is a long-form member of the pentraxin family, and we have identified the PTX3 gene as a TNF␣-responsive gene leading to the protein secretion.TNF␣ and IL-1, two major mediators of inflammation, have been reported to induce the transcriptional activation of the PTX3 gene (19,21,46).LPS, which raises PTX3 plasma levels after injection in mice, shows a direct effect by activating the PTX3 gene.It is assumed that, as in 3T3-L1 adipocytes, LPS binding to constitutively-expressed Toll-like receptor-4 (TLR4) results in the fast induction of TLR-2 and the synthesis of a set of secretory proteins that would include PTX3 (47).Although cultured adipocytes and adipose tissue explants respond to very low concentrations of TNF␣ (0.01-1 nM), it remains unclear whether Type 1 or Type 2 TNF␣ receptors are involved in this response.Further studies using adipocytes lacking each or both of these receptors should clarify this issue (48).Preliminary evidence shows that significant levels of PTX3 are present in mouse plasma, but a possible increase in plasma of genetically obese mice remains to be shown.If it is so, considering the fact that PTX3 binds to C1q (25), which is known to be homologous to adiponectin (49,50), it is tempting to postulate that higher circulating levels of PTX3 in a chronic low-level inflammatory state may lower the concentration of unbound adiponectin and may participate in the aggravation of the metabolic syndrome observed in obese animals and individuals.Studies in transgenic mice and gene-targeted mice suggest an important role for PTX3 in the regulation of inflammatory reactions and innate immunity (51)(52)(53)(54).The study of the development of adipose tissue in these knock-out animals as well as the use of obese TNF␣-deficient mice should shed some light on the physiological role of PTX3.

Fig. 1 .
Fig. 1.Effect of fatty acid on PTX3 mRNA expression in preadipocytes and its distribution in adipose tissue.A: One day post-confluent Ob1771 cells were maintained for 24 h in standard medium in the absence (Ϫ) or presence (ϩ) of 300 M ␣-linolenate.Twenty micrograms of RNA were analyzed as described in Materials and Methods.The results are representative of four independent experiments.B: RNA from stromal-vascular and adipocyte fractions of white adipose tissue (WAT) and brown adipose tissue from six mice was blotted and hybridized with PTX3 probe.The results are representative of two independent experiments.

Fig. 2 .
Fig. 2. PTX3 mRNA expression in 3T3-F442A and 3T3-L1 cells during adipocyte differentiation in serumcontaining and in serum-free medium.A: RNA was extracted from 3T3-F442A or 3T3-L1 cells that had been maintained in serum-supplemented medium at different intervals relative to confluence.3T3-L1 cells have been induced to differentiate (ϩMDI) or not (ϪMDI) in the presence of an hormonal cocktail (0.5 mM 1-methyl-3-isobutylmethyl-xanthine, 0.5 M dexamethasone, and 170 nM insulin) for 2 days.B: Total RNA was extracted from 3T3-F442A cells that had been maintained in serum-free medium at different inter vals relative to confluence.Twenty micrograms of RNA of each sample were loaded and analyzed by Northern blot as described in Materials and Methods.The results are representative of two independent experiments.

Fig. 3 .
Fig. 3. Ectopic expression of PTX3 and differentiation of 3T3-F442A preadipocytes.Stable 3T3-F442A-vector cells and 3T3-F442A-PTX3 transfectants were maintained after confluence in differentiation medium.A: Total protein was extracted at Days 0 and 13, and 50 g was analyzed by Western blot.TATA-binding protein (TBP) was used as the internal standard.B: RNA was extracted at Days 2 and 10 after confluence, and 20 g were analyzed by Northern blot.C: Cells were fixed and stained with Oil red O at Day 10 after confluence and photographed.The results are representative of two independent experiments.

Fig. 4 .
Fig. 4. Effect of cytokines on PTX3 mRNA expression in 3T3-F442A adipocytes.3T3-F442A cells were maintained after confluence in differentiation medium.At Day 7 after confluence, differentiated cells were (A) treated for different time periods with 1 nM tumor necrosis factor ␣ (TNF␣), or (B) treated for 3 h with increasing concentrations of TNF␣ or interleukin-6 (IL-6) as indicated, or with 10 ng/ml leukemia inhibitory factor or 100 ng/ml lipopolysaccharide (LPS).Twenty micrograms of RNA were analyzed by Northern blot.The results are representative of three independent experiments.

Fig. 5 .
Fig. 5. PTX3 protein secretion from 3T3-F442A adipocytes and adipose tissue explants.Differentiated 3T3-F442A cells (upper panel) or adipose tissue explants (lower panel) were maintained in serum-free medium and treated with increasing concentrations of TNF␣ as indicated, 100 ng/ml LPS, or 1 nM IL-6 for 3 h.Secretion media were collected and analyzed by Western blot as described in Materials and Methods.The results are representative of two independent experiments.

Fig. 6 .
Fig. 6.PTX3 mRNA expression in WAT of genetically-obese mice.Total RNA was extracted from epididymal adipose tissue of wild-type (wt), obese (ob/ob), or obese diabetic (db/db) mice.One microgram of RNA was analyzed by real-time RT-PCR as described in Materials and Methods.The calculations of PTX3 amounts (relative units) were obtained by using the comparative-Ct method.Each point represents measurements from three mice determined in triplicate.TBP served as reference.Asterisk indicates statistically significant differences, P Ͻ 0.05.