The direct peroxisome proliferator-activated receptor target fasting-induced adipose factor (FIAF/PGAR/ANGPTL4) is present in blood plasma as a truncated protein that is increased by fenofibrate treatment.

The fasting-induced adipose factor (FIAF, ANGPTL4, PGAR, HFARP) was previously identified as a novel adipocytokine that was up-regulated by fasting, by peroxisome proliferator-activated receptor agonists, and by hypoxia. To further characterize FIAF, we studied regulation of FIAF mRNA and protein in liver and adipose cell lines as well as in human and mouse plasma. Expression of FIAF mRNA was up-regulated by peroxisome proliferator-activated receptor alpha (PPARalpha) and PPARbeta/delta agonists in rat and human hepatoma cell lines and by PPARgamma and PPARbeta/delta agonists in mouse and human adipocytes. Transactivation, chromatin immunoprecipitation, and gel shift experiments identified a functional PPAR response element within intron 3 of the FIAF gene. At the protein level, in human and mouse blood plasma, FIAF was found to be present both as the native protein and in a truncated form. Differentiation of mouse 3T3-L1 adipocytes was associated with the production of truncated FIAF, whereas in human white adipose tissue and SGBS adipocytes, only native FIAF could be detected. Interestingly, truncated FIAF was produced by human liver. Treatment with fenofibrate, a potent PPARalpha agonist, markedly increased plasma levels of truncated FIAF, but not native FIAF, in humans. Levels of both truncated and native FIAF showed marked interindividual variation but were not associated with body mass index and were not influenced by prolonged semistarvation. Together, these data suggest that FIAF, similar to other adipocytokines such as adiponectin, may partially exert its function via a truncated form.

carry a mutation in the gene for ANGPTL3, resulting in low plasma triglyceride and free fatty acid levels (20). The low plasma triglyceride levels are possibly due to elevated lipoprotein lipase activity, which was reported to be inhibited by ANGPTL3 in vitro, whereas the low plasma free fatty acid levels may be connected to the impaired stimulatory effect of ANGPTL3 on adipose tissue lipolysis (21,22). Similar to ANGPTL3, there is evidence that FIAF also may inhibit lipoprotein lipase activity and hereby influence plasma levels of triglycerides (18,23), thus connecting FIAF to lipid metabolism.
Expression of FIAF in liver and WAT was originally found to be up-regulated by the nuclear hormone receptors PPAR␣ and PPAR␥ (15,17,23). PPARs are ligand-activated transcription factors that mediate the effects of fibrates (PPAR␣) or thiazolidinediones (PPAR␥) on DNA transcription (24,25). PPAR␣ is mainly expressed in brown adipose tissue and liver and plays an important role in the hepatic fatty acid oxidation, whereas PPAR␥ is the master regulator of adipogenesis. However, it is still unclear whether FIAF is a direct PPAR target gene, with a functional PPAR response element in its promoter.
In order to close in on the potential function of FIAF, we studied the regulation of FIAF mRNA and protein expression in vitro and in human blood plasma. Our main conclusions are that FIAF is a classical PPAR target gene in both humans and rodents and that FIAF protein is mainly present in blood plasma in a truncated form, whose levels show a large interindividual variability. Plasma levels of the truncated form of FIAF are increased by treatment with fenofibrate.

MATERIALS AND METHODS
Chemicals-Wy14643 was obtained from ChemSyn Laboratories. Rosiglitazone was from Alexis. Recombinant human insulin (Actrapid) was from Novo Nordisk. SYBR Green was from Eurogentec. Dulbecco's modified Eagle's medium, fetal calf serum, calf serum, and penicillin/ streptomycin/fungizone were from BioWhittaker Europe (Cambrex Bioscience). Otherwise, chemicals were from Sigma.
Primary Human Adipocyte Differentiation-Isolation of stromal vascular cells was done as follows. Subcutaneous and visceral adipose tissues were obtained during gastric restriction surgery. Adipose tissue was collected in phosphate-buffered saline and cut into 3 ϫ 3-mm pieces with scissors. The 3 ϫ 3-mm pieces were further processed with a scalpel. Next, the pieces of adipose tissue were digested in DMEM-high glucose containing 4% bovine serum albumin and 2 mg/ml collagenase. 1-2.5 g of adipose tissue was digested in 5 ml of this solution at 37°C on a shaking platform for 2 h. Next, the digest was transferred to a 5-ml syringe and gently pressed over a 500-m sterile pore size disposable nylon mesh. Stromal vascular cells were separated from adipose cells by centrifugation (1 min, 170 g). Adipose cells were removed, and the stromal vascular cells were precipitated by centrifugation (5 min, 350 g). Red blood cells were lysed by resuspending the cell pellet in 10 ml of red cell lysis solution (154 mM NH 4 Cl, 10 mM KHCO 3 , 0.1 mM EDTA). After 5 min, stromal vascular cells were spun down (5 min, 350 g) and resuspended in DMEM/F-12 containing 10% fetal calf serum (FCS), 2 mM glutamine, 100 IU/ml penicillin, and 100 g/ml streptomycin and further referred to as complete medium. Approximately 10 7 cells were plated in a 75-cm 2 flask. After 48 h, the medium was replaced by differentiation medium consisting of complete medium (instead of 10% FCS, 2% FCS was used) plus 15 mM NaHCO 3, 15 mM HEPES, 33 M biotin, 17 M panthothenate, 200 pM T3, 1 M dexamethasone, 500 nM insulin, 4 g/ml transferrin, and 10 M cPGI 2 . After an average period of 30 days of differentiation, extraction of total RNA was performed.
SGBS Cell Line Culture and Induction of Adipogenesis-The culture of the SGBS cells and their induction into mature human adipocytes were performed exactly as previously published (26).
3T3-L1 Adipogenesis Assay-3T3-L1 fibroblasts were amplified in DMEM plus 10% calf serum and plated for final differentiation in DMEM plus 10% FCS. Two days after reaching confluence (which was day 0), the medium was changed, and the following compounds were added: isobutyl methylxanthine (0.5 mM), dexamethasone (1 M), and insulin (5 g/ml). On day 3, the medium was changed to DMEM plus 10% FCS and insulin (5 g/ml). On day 6, the medium was changed to DMEM plus 10% FCS, which was changed every 3 days.
Isolation of Total RNA and RT-PCR-Total RNA was extracted from cells or tissue with Trizol reagent (Invitrogen) following the supplier's protocol. 3-5 g of total RNA was treated with DNase I amplification grade (Invitrogen) and then reverse-transcribed with oligo(dT) using Superscript II RT RNase H (Invitrogen).
PCR was carried out using Platinum Taq polymerase (Invitrogen) and SYBR Green on an iCycler PCR machine (Bio-Rad). Expression was related to actin, which did not change under any of the experimental conditions studied.
Transactivation Assay-A 350-nucleotide fragment surrounding the putative PPRE within intron 3 of the human or mouse FIAF gene was PCR-amplified from human or mouse genomic DNA (mouse strain C57/B6) and subcloned into the KpnI and BglII sites of pTAL-SEAP (Clontech). This reporter vector was transfected into human hepatoma HepG2 cells together with an expression vector for PPAR (mPPAR␣, mPPAR␤/␦, or mPPAR␥ in pSG5) in the presence or absence of Wy14643 (50 M), L165041 (5 M), or rosiglitazone (5 M), respectively. Transfections were carried out by calcium phosphate precipitation. A ␤-galactosidase reporter was co-transfected to normalize for differences in transfection efficiency. Secreted alkaline phosphatase activity was measured in the medium 48 h post-transfection via the chemiluminescent SEAP reporter assay (Roche Applied Science). ␤-Galactosidase activity was measured in the cell lysate by standard assay using 2-nitrophenyl-␤-D-galactopyranoside as a substrate.
Chromatin Immunoprecipitation (ChIP)-Pure-bred wild-type or PPAR␣ null mice on a sv129 background were used. Mice were fed by gavage with either Wy14643 (50 mg/kg/day) or vehicle (0.5% carboxymethylcellulose) for 5 days. Alternatively, mice were fasted or not for 24 h. After the indicated treatment, mice were sacrificed by cervical dislocation. The liver was rapidly perfused with prewarmed (37°C) phosphate-buffered saline for 5 min followed by 0.2% collagenase for 10 min. The liver was diced and forced through a stainless steel sieve, and the hepatocytes were collected directly into DMEM containing 1% formaldehyde. After incubation at 37°C for 15 min, the hepatocytes were pelleted, and ChIP was done using mouse PPAR␣-specific antibodies as previously described (27). Sequences of primers used for PCR were 5Ј-TCTGGGTCTGCCCCCACTCCTGG-3Ј (forward) and 5Ј-GTGTGT-GTGTGGGATACGGCTAT-3Ј (reverse). Control primers used were 5Ј-AGTAACTTTGACAGGAACCAGGGGTC-3Ј (forward) and 5Ј-TTTG-GACTGGGAACTTAGCTTAGTTG-3Ј (reverse).
3T3-L1 cells were differentiated as described above. After cell lysis and sonication, the supernatant was diluted 20-fold in re-chIP dilution buffer (1 mM EDTA, 20 mM Tris-HCl, pH 8.1, 50 mM NaCl, 1% Triton X) prior to incubation with mouse PPAR␥ antibody. The remainder of the assay was carried out as described previously (27).
Western Blot-The mouse polyclonal antibody used was directed against the epitope CQGPKGKDAPFKDSE located in the N-terminal part of the mouse FIAF protein. The human polyclonal antibody used was directed against the epitope CQGTEGSTDLPLAPE also located in the N-terminal part of the human FIAF protein. The peptide affinitypurified antibodies were generated in rabbit and ordered via Eurogentec's customized antibody production service. Western blotting was carried out using an ECL system (Amersham Biosciences) according to the manufacturer's instructions. The primary antibody was used at a dilution of 1:1000 (mouse) or between 1:2000 and 1:5000 (human), and the secondary antibody (anti-rabbit IgG, Sigma) was used at a dilution of 1:8000. All incubations were performed in 1ϫ Tris-buffered saline, pH 7.5, with 0.1% Tween 20 and 5% dry milk, except for the final washings, when milk was omitted.
Human Subjects-In experiment 1, blood was taken from 16 young adults after an overnight fast. In experiment 2, blood was taken after an overnight fast from 28 subjects before and after a 4-week treatment with 250 mg of micronized fenofibrate daily. In experiment 3, blood was taken after an overnight fast from 20 men (body mass index ranging 22.7 to 39.8). Samples were from a published study (28). In experiment 4, blood was taken after an overnight fast from 22 subjects before and after a 46-day semistarvation program (2.1 MJ/day). Samples were from a published study (29). All human experiments were approved by the medical ethics committee of Wageningen University, Maastricht University, or the University of Ulm.

FIAF Expression Is Regulated by All Three PPARs-Previous
studies have indicated that expression of FIAF/PGAR/AN-GPTL4 is up-regulated by PPAR␣ and PPAR␥ in mice. Several genes are known that are targets of PPAR␣ in mice but not in humans (30,31). To investigate whether expression of FIAF is under control of PPAR␣ in other species, rat hepatoma FAO cells were treated with the PPAR␣ agonist Wy14643 (Fig. 1A). According to real time quantitative PCR, FAO cells express relatively high levels of PPAR␣ as well as PPAR␤/␦, whereas PPAR␥ mRNA was below our detection limit (Fig. 1B). Basal expression of FIAF in FAO cells was extremely low but was dramatically increased by Wy14643, either alone or in combination with the RXR agonist Lg100268 (Fig. 1A). The synthetic PPAR␤/␦ agonist L165041 also strongly increased FIAF mRNA, suggesting that PPAR␤/␦ stimulates FIAF gene expression too. Finally, the PPAR␥ agonist ciglitazone had little effect on rat FIAF gene expression, which may be explained by the low expression of PPAR␥ mRNA in these cells.
To examine whether the human FIAF gene is also up-regulated by PPARs, human hepatoma HepG2 cells were treated with PPAR agonists (Fig. 1C). HepG2 cells express all three PPARs, with PPAR␤/␦ being the most abundant (Fig. 1D). Similar to what was observed in FAO cells, although with much more modest -fold inductions, FIAF mRNA was increased by Wy14643 and Lg100268, either alone or used in combination (Fig. 1C). The PPAR␤/␦ agonist L165041 also induced FIAF mRNA, but no additional effect of Lg100268 was observed. In contrast to PPAR␣ and PPAR␤/␦ agonists, the PPAR␥ agonist ciglitazone reduced FIAF expression, which was maintained in the presence of Lg100268. Taken together, these results indicate that FIAF is up-regulated by PPAR␣ and PPAR␤/␦, but probably not by PPAR␥, in human and rat hepatoma cells.
To better examine regulation of human FIAF expression by PPAR␥, we turned to primary human preadipocytes. Upon stimulation with a mixture of hormones, these cells can be differentiated into adipocytes. Stromal vascular cells from both subcutaneous and visceral adipose tissue were isolated and induced to differentiate into adipocytes. Expression of FIAF was higher in adipocytes versus preadipocytes in all three subjects with cells from both subcutaneous and visceral origin ( Fig. 2A). A similar induction of expression was observed for PPAR␥, suggesting that FIAF is up-regulated by PPAR␥ during human adipocyte differentiation. In differentiated human SGBS adipocytes, both rosiglitazone and L165041 caused an induction of FIAF mRNA (Fig. 2, B and C, respectively), indicating that both PPAR␥ and PPAR␤/␦ regulate FIAF expression in human adipocytes. Together, these data suggest that FIAF is a PPAR␥ and possibly a PPAR␤/␦ target gene in human adipocytes.
FIAF Is a Direct Target Gene of PPAR-To unequivocally determine FIAF as a direct target gene of PPARs, direct binding of PPAR to the FIAF promoter needs to be demonstrated. Comparative analysis of the hFIAF and mFIAF gene sequence upstream of the transcription start site did not reveal any conserved stretches of DNA that might harbor a PPRE. Transactivation studies with several kilobases of the immediate upstream sequence from both the mouse and human FIAF gene did not yield any significant activation of a reporter gene, suggesting that the responsive element may be located elsewhere. While scanning the FIAF gene sequence for PPREs, a putative PPRE was identified in a conserved region of intron 3 of the human and mouse FIAF gene (AGG(G/A)GAAAGGTC(G/ A)) that differed little from the consensus PPRE (Fig. 3A). To determine whether this PPRE binds PPAR in vitro, gel shift experiments were carried out with in vitro translated/transcribed PPAR␣ and RXR␣. For both the human and mouse PPRE, a retarded complex was only observed in the presence of both PPAR␣ and RXR␣ (Fig. 3B), indicating that this complex represents a PPAR␣/RXR␣ heterodimer. The complex disappeared in the presence of an excess of cold specific oligonucleotide but not nonspecific oligonucleotide. Similar results were observed for PPAR␥ (data not shown). These data indicate that PPAR is able to bind to the human and mouse PPRE within intron 3 in vitro.
To assess whether the PPRE within intron 3 is able to mediate PPAR-dependent transactivation, a 350-nucleotide fragment surrounding the human or mouse PPRE was cloned in front of the thymidine kinase promoter followed by a SEAP reporter. In HepG2 cells, co-transfection of the reporter vector with a PPAR␣, PPAR␤/␦, or PPAR␥ expression vector increased SEAP activity, which was further enhanced by the addition of ligand (Fig. 3C). In this assay, PPAR␣ seemed to be the most potent activator, followed by PPAR␤/␦ and PPAR␥. These data suggest that the PPRE identified in intron 3 of the FIAF gene is able to mediate PPAR-dependent transactivation.
Finally, to find out whether PPAR␣ and PPAR␥ are bound to this sequence in vivo, ChIP was performed using antibodies against PPAR␣ or PPAR␥. In human HepG2 cells, binding of PPAR␣ to the sequence spanning the putative PPRE within intron 3 was enhanced by Wy14643 (Fig. 4A). No immunoprecipitation was observed with preimmune serum, and no amplification was observed for a control sequence. In mice, treatment with Wy14643 enhanced binding of PPAR␣ to the PPRE sequence in liver, which was not observed in PPAR␣ null mice (Fig. 4B). Similarly, fasting enhanced binding of PPAR␣ to the PPRE sequence, which was not observed in the PPAR␣ null mice (Fig. 4C). With respect to PPAR␥, previous data had shown that FIAF is up-regulated during mouse 3T3-L1 adipogenesis (17), indicating that it may be a direct PPAR␥ target gene. Using ChIP, we observed binding of PPAR␥ to the PPRE sequence in differentiated 3T3-L1 adipocytes but not in preadipocytes (Fig. 4D). These data clearly demonstrate that PPAR␣ and PPAR␥ bind to the intronic sequence harboring the PPRE in vivo. Thus, FIAF can be formally classified as a direct PPAR target gene in human and mouse.
FIAF Protein Is Processed during Mouse Adipocyte Differentiation-The increased level of FIAF mRNA in primary differentiated adipocytes versus preadipocytes, regardless of the fat depot, indicates that FIAF is up-regulated during human adipocyte differentiation. Indeed, it was observed that FIAF mRNA increases during human SGBS adipocyte differentiation, displaying a dramatic up-regulation during early differentiation that diminished during prolonged differentiation (Fig. 5A). According to Western blot using an antibody that recognizes human FIAF (Fig. 5B), the mRNA expression profile of FIAF was mirrored at the protein level, with some delay (Fig.  5C). In the Western blot, a single band at the expected molecular mass (ϳ45 kDa) was observed.
In accordance with previous studies by Yoon et al. (17), an increase in FIAF mRNA during prolonged mouse 3T3-L1 adipogenesis was observed (Fig. 5D). However, we also observed that FIAF expression transiently peaks at day 3 of differentiation, reaching a level exceeding that of fully differentiated adipocytes. This effect could be attributed to IBMX, since incubation of confluent 3T3-L1 cells with only IBMX, which does not induce adipocyte differentiation, markedly increased FIAF mRNA (Fig. 5D, inset). IBMX is removed from the medium from day 3 onwards, explaining the precipitous drop in FIAF mRNA at day 4.
Whereas FIAF protein directly followed FIAF mRNA expression during human adipocyte differentiation (Fig. 5, A and C), a remarkable protein expression pattern was observed for mouse adipocyte differentiation (Fig. 5, D and E). In parallel with FIAF mRNA, with a delay of 1 day, native FIAF protein rose during early differentiation and peaked at day 4, 1 day after the maximal FIAF mRNA level. Thereafter, its level de-

FIG. 3. FIAF up-regulation by PPAR␣ is mediated by a PPRE present in intron 3.
A, alignment of the PPRE present in intron 3 of the human and mouse FIAF gene with known PPREs. Cons., consensus. B, binding of the PPAR/RXR heterodimer to putative PPRE as determined by gel shift. A double-stranded response element containing the human (left) or mouse (right) FIAF PPRE was incubated with in vitro transcribed/translated hPPAR␣ and hRXR␣, and binding complexes were resolved on a 6% nondenaturing polyacrylamide gel. -Fold excess of specific (malic enzyme PPRE) or nonspecific (ETS oligonucleotide) cold probe is indicated. C, HepG2 cells were transfected with a SEAP reporter vector containing a 350-nucleotide fragment of intron 3 of the human (left) or mouse (right) FIAF gene and a PPAR expression vector. SEAP activity was determined in the medium 24 -48 h post-transfection. creased (Fig. 5E). Interestingly, in the same immunoblot, an additional band with a molecular mass of about 32 kDa appeared at day 4 and further increased at days 6 and 10. Thus, the upper band, representing native FIAF, follows FIAF mRNA during early differentiation, whereas the lower band follows FIAF mRNA during prolonged adipocyte differentiation, suggesting it is derived from FIAF. We hypothesized that this band represents a truncated form of FIAF, which is observed in mouse but not human adipocytes.
If this is correct, it would be expected that the abundance of truncated FIAF would mirror the FIAF mRNA expression data in 3T3-L1 adipocytes treated with synthetic PPAR and RXR agonists. Indeed, induction of FIAF mRNA by the RXR agonist Lg100268 and by rosiglitazone in differentiated 3T3-L1 adipocytes was associated with an increased abundance of the lower molecular weight band (Fig. 6, A and B). Similarly, in livers of mice treated with Wy14643, which results in up-regulation of FIAF mRNA (Fig. 6C), the abundance of the lower molecular weight band was increased in parallel, providing compelling evidence that this band represents a truncated form of FIAF. Hereon, this form of FIAF is referred to as FIAF-S1 (for FIAF small form 1).
Because FIAF was initially found to be a protein secreted into the blood plasma, we set out to determine whether the same was true for FIAF-S1. Interestingly, besides native FIAF and FIAF-S1, another immunoreactive form of slightly higher molecular weight (about 2-3 kDa) was also detected, which we named FIAF-S2 and which was by far the most abundant (Fig.  6D). Preincubation of the mouse FIAF antibody with its peptide epitope completely abolished all three forms. Notice that in Fig.  6E native FIAF is barely visible because the blot was exposed for less time. It is not inconceivable that FIAF-S2 might rep- resent a phosphorylated or glycosylated form of FIAF-S1. Both FIAF-S1 and FIAF-S2 were well detected in mouse WAT, whereas only native FIAF and FIAF-S1 were detected in mouse liver (Fig. 6E). Together, these data suggest that FIAF is present in truncated forms in mouse blood plasma.
FIAF-S1 and FIAF-S2 Are Present in Human Blood Plasma-To establish that FIAF is also present in truncated forms in human plasma, Western blot was carried out on human blood plasma using an anti-hFIAF antibody. Almost copying the picture of mouse blood plasma, in human plasma both native human FIAF protein at 50 kDa but also two bands of lower molecular weight were observed, probably corresponding to FIAF-S1 and FIAF-S2 (Fig. 7A). Incubation with the peptide epitope caused the complete disappearance of all bands. The molecular weight of the putative FIAF-S1 and FIAF-S2 in human was lower than that of the same species in mice. Omission of dithiothreitol in the SDS-sample buffer led to the appearance of a very high molecular weight immunoreactive complex, suggesting that FIAF forms oligomers or possibly a high molecular weight complex involving other plasma proteins (Fig. 7B). Omission of dithiothreitol also slightly increased the mobility of native FIAF and FIAF-S2. Levels of putative FIAF-S2 after an overnight fast were very reproducible within subjects (not shown) but extremely variable between subjects (Fig. 7C). Levels of native FIAF also differed markedly between subjects but to a somewhat lesser extent. Together, these data indicate that FIAF is circulating in blood in several forms of different sizes at different concentrations.
Levels of FIAF-S2 in Human Blood Plasma Are Increased by Fenofibrate-Our data indicate that human FIAF mRNA is upregulated by PPAR␣ agonists in human hepatoma cells. If the lower molecular weight band in the immunoblot blot of human plasma indeed represents truncated FIAF protein, its level would be expected to increase after treatment with PPAR␣ agonists. To find out whether this is true, plasma levels of putative FIAF-S2 were assessed by Western blot in 28 subjects before and after treatment with fenofibrate, a potent PPAR␣ agonist (Fig. 7, D and E). In 24 of the 28 subjects, levels of FIAF-S2 rose after fenofibrate treatment, whereas four individuals showed a decrease or no change. The mean increase was 84.5% Ϯ 20.1 (S.E.) (paired Student's t test, p Ͻ 0.0001). Levels of native FIAF did not respond or only slightly responded to fenofibrate treatment. These data suggest that FIAF is mainly present in human blood plasma in a truncated form (FIAF-S2), whose level is increased by fenofibrate treatment.
Fenofibrate, which primarily acts on liver, influences plasma levels of FIAF-S2 but not native FIAF. At the same time, human SGBS adipocytes only produce native FIAF. This raises the possibility that human liver mainly produces FIAF-S2, whereas human WAT mainly synthesizes native FIAF. In agreement with this notion, we only detect FIAF-S2 in human liver, and native FIAF in human WAT (Fig. 8A).
Despite WAT possibly being the main contributor to native FIAF in plasma, plasma levels of native FIAF did not respond significantly in a group of 22 subjects undergoing a 46-day semistarvation program (2.1 MJ/day), although they lost an average of 12 kg (Fig. 8B, representative results of four subjects). The same was true for FIAF-S2. Also, no association was observed between body mass index and plasma levels of native FIAF or FIAF-S2 in a group of individuals with varying body mass index (22.7-39.8) (Fig. 8, C and D).

DISCUSSION
In the past decade, it has become clear that adipose tissue not merely serves to store energy but also has an important endocrine function, secreting an array of proteins that include leptin, resistin, adiponectin/ACRP30/adipoQ, interleukin-6, and tumor necrosis factor-␣. These so called adipokines or adipocytokines are involved in numerous processes and have been particularly studied as potential mediators of the link between obesity and obesity-related metabolic abnormalities, with special emphasis on insulin resistance (2).
An adipocytokine that received a lot of publicity lately is adiponectin. A special property of adiponectin is that it is cleaved to generate a smaller product called globular adiponectin, which is probably the physiologically active form (32). According to our data, FIAF may also become proteolytically processed to generate a protein of 20 -35 kDa, the exact size of which depends on the species and probably on glycosylation. Alternatively, FIAF-S could be generated through alternative splicing of FIAF mRNA, by differential initiation start sites, or by some unknown mechanisms. However, neither RT-PCR experiments using different primers, Northern blots, nor RNase protection provided any evidence of the generation of an additional mRNA. This suggests that FIAF-S1 and FIAF-S2 are generated by proteolytic processing. Considering that native FIAF is glycosylated, FIAF-S1 and FIAF-S2 may represent different glycosylated forms (15,16). Our data also indicate that human liver mainly synthesizes FIAF-S2, whereas human WAT seems to produce native FIAF exclusively. This suggests that FIAF-S2 and native FIAF in plasma may originate from different tissues. This is supported by the observation that fenofibrate, which mainly acts on liver, increases plasma levels of FIAF-S2 but not native FIAF. In mice, the contribution of various tissues to plasma FIAF is less transparent.
Proteolytic processing of prohormone precursor proteins is a common theme in endocrinology. Numerous protein and peptide hormones, including insulin, glucagon, and adipocytokines such as tumor necrosis factor-␣ and adiponectin, are proteolytically cleaved to generate the smaller functional form of the protein. The most common processing recognition site in prohormones consists of a doublet of basic amino acids (33), which is recognized by subtilisin-like proprotein convertases, although other types of motifs are also possible. Carboxypeptidase E is responsible for the removal of carboxyl-terminal basic residues exposed by the endoproteases (34). Interestingly, in the primary structure of FIAF, two conserved adjacent arginines could be identified, which might represent proteolytic recognition sites. Digestion around this site (Arg 229 and Arg 230 in hFIAF) would be compatible with the size of fragments FIAF-S1 and FIAF-S2. Recently, it was found that expression of the proprotein convertases PACE4, PC7, and furin increases during 3T3-L1 adipocyte differentiation, when processing of native FIAF to FIAF-S1 becomes apparent (35). Consequently, it is conceivable that these enzymes participate in the processing of FIAF in 3T3-L1 adipocytes.
Besides being proteolytically processed, adiponectin also forms higher order oligomers, which may have a different functional activity than monomeric adiponectin. Resistin has been shown to self-associate as well (36,37), which again may influence functional activity. According to our data, FIAF may also be present in human blood plasma as higher order oligomers, although the exact composition of the observed higher molecular weight complex(es) remains to be determined. Similar to adiponectin and resistin, oligomerization of FIAF may influence functional activity.
Very recently, Ono et al. reported that ANGPTL3 is cleaved in vivo, and, similar to our observations for FIAF (or AN-GPTL4), is present in mouse blood plasma in several forms of around 30 kDa (38). Interestingly, it was found that the resulting N-terminal fragment is probably responsible for the plasma triglyceride-raising effect of ANGPTL3. Furthermore, while our manuscript was in preparation, data were published showing that recombinant FIAF protein is truncated and forms oligomers in HEK293 cells and in vivo (39). No data were provided on endogenous FIAF, in contrast to the present paper. Although details about the site of truncation seem to be different between the two papers, together they suggest that proteolytic processing, and perhaps oligomerization, may be important for FIAF function. Thus, proteolytic processing may be common among members of this protein family and may serve to regulate functional activity.
Previous studies have indicated that, at least in mouse, FIAF mRNA is most highly expressed in white adipose tissue (15). According to our data, human WAT mainly produces native FIAF. The lack of a significant association between body mass index and plasma levels of native FIAF and the absence of an effect of prolonged weight loss on native FIAF suggest that either the size of WAT has little impact on the total amount of native FIAF released from WAT into blood plasma or that adipose tissue may not be the primary source for native FIAF in human plasma.
Experiments in mice have shown that both hepatic and adipose expression of FIAF are elevated by fasting (15). With respect to FIAF in plasma, levels of native FIAF were found to be elevated after fasting (15), whereas levels of FIAF-S2 or FIAF-S1 did not seem to be affected. 2 Preliminary data indicate that the fasting-induced up-regulation of FIAF mRNA in adipose tissue may not be observed in mice of the FVB strain. With regard to humans, it is unclear whether fasting causes upregulation of FIAF mRNA in liver and WAT. Levels of FIAF in plasma do not appear to be influenced by short term fasting (data not shown) or long term semistarvation. Thus, the term fasting-induced adipose factor may not aptly describe the behavior of FIAF in several species.
Previously, we and others have demonstrated that in mice FIAF is up-regulated by both PPAR␣ and PPAR␥ (15,17). Here it is shown that this regulation also occurs in humans, in contrast to many other PPAR␣ target genes. Furthermore, besides PPAR␣ and PPAR␥, PPAR␤ is similarly able to induce FIAF expression in hepatocytes and adipocytes. It is also shown that up-regulation of FIAF expression by PPARs is, at least partly, mediated by a PPRE present in intron 3. Via chromatin immunoprecipitation on livers of fasted and fed or Wy14643-treated mice, direct in vivo binding of PPAR␣ to intron 3 was demonstrated, which was enhanced by fasting and by Wy14643. Furthermore, binding of PPAR␣ to the same sequence was enhanced by Wy14643 in human HepG2 cells. Finally, binding of PPAR␥ to the sequence could be demonstrated in differentiated 3T3-L1 adipocytes but not preadipocytes. Thus, FIAF can be added to the list of direct PPAR target genes. Although the presence of a functional PPRE within an intron is remarkable, it is not completely uncommon. Indeed, recently the presence of a functional PPRE within intron 3 of the rat peroxisomal thiolase B gene was demonstrated (40).
Plasma levels of FIAF-S2 are increased by fenofibrate treatment. Inasmuch as there is evidence that FIAF is involved in lipid metabolism, it can be speculated that the effects of synthetic PPAR␣ agonists on plasma lipid levels may be partially mediated via changes in FIAF expression. Further studies are necessary to ascertain the potential of FIAF as a target for treatment of various forms of dyslipidemia.