Lysophosphatidic Acid Inhibits Adipocyte Differentiation via Lysophosphatidic Acid 1 Receptor-dependent Down-regulation of Peroxisome Proliferator-activated Receptor γ2*

Lysophosphatidic acid (LPA) is a bioactive phospholipid acting via specific G protein-coupled receptors that is synthesized at the extracellular face of adipocytes by a secreted lysophospholipase D (autotaxin). Preadipocytes mainly express the LPA1 receptor subtype, and LPA increases their proliferation. In monocytes and CV1 cells LPA was recently reported to bind and activate peroxisome proliferator-activated receptor γ (PPARγ), a transcription factor also known to play a pivotal role in adipogenesis. Here we show that, unlike the PPARγ agonist rosiglitazone, LPA was unable to increase transcription of PPARγ-sensitive genes (PEPCK and ALBP) in the mouse preadipose cell line 3T3F442A. In contrast, treatment with LPA decreased PPARγ2 expression, impaired the response of PPARγ-sensitive genes to rosiglitazone, reduced triglyceride accumulation, and reduced the expression of adipocyte mRNA markers. The anti-adipogenic activity of LPA was also observed in the human SGBS (Simpson-Golabi-Behmel syndrome) preadipocyte cell line, as well as in primary preadipocytes isolated from wild type mice. Conversely, the anti-adipogenic activity of LPA was not observed in primary preadipocytes from LPA1 receptor knock-out mice, which, in parallel, exhibited a higher adiposity than wild type mice. In conclusion, LPA does not behave as a potent PPARγ agonist in adipocytes but, conversely, inhibits PPARγ expression and adipogenesis via LPA1 receptor activation. The local production of LPA may exert a tonic inhibitory effect on the development of adipose tissue.

Enlargement of adipose tissue is conditioned by the ability of adipocytes to store triglycerides as well as by the ability of preadipocytes to differentiate into adipocytes (adipogenesis). The genetic program set up for adipogenesis is tightly controlled by the coordinated interplay of several transcription factors, the most important being peroxisome proliferator-acti-vated receptor ␥ (PPAR␥) 1 (mainly the isoform PPAR␥2) (11). Identifying the factors that control and/or regulate PPAR␥ activity and adipogenesis is of major interest for understanding the normal growth and the pathologic growth of adipose tissue. Many circulating factors (insulin, insulin growth factor I, growth hormone, glucocorticoids, thyroid hormone, etc.) are known to promote proliferation and/or differentiation of preadipocytes (see Ref. 1 for review). In addition, the production of paracrine and autocrine factors within adipose tissue could also play an important role in its development. Adipocytes release several peptides (leptin, adipsin, adiponectin, angiotensinogen, etc.), proteins (lipoprotein lipase, autotaxin, etc.), and lipids (fatty acids, prostaglandins, lysophosphatidic acid, etc.) involved in preadipocyte growth and/or differentiation (see Ref. 2

for review).
Lysophosphatidic acid (LPA) is a potent bioactive phospholipid able to regulate several cellular responses via activation of specific G protein-coupled receptors. Four LPA receptor subtypes have been identified, namely LPA 1 , LPA 2 , LPA 3 , and LPA 4 (3). LPA 1 (edg-2 in the former nomenclature) was the first identified LPA receptor subtype. It is abundantly expressed in the central nervous system but is also present in numerous peripheral tissues. Invalidation of LPA 1 receptor in mouse is associated with impaired suckling behavior in neonate pups and reduced body size and weight of the adults (4).
Our group has demonstrated that LPA is produced in the extracellular medium of adipocytes (5) as the result of the secretion of lysophospholipase D (autotaxin) (6,7). Unlike adipocytes, preadipocytes do not produce LPA (6,7). Extracellular LPA activates the mitogen-activated protein kinases ERK1 and ERK2 and increases the proliferation of growing 3T3F442A preadipocytes, which mainly express the LPA 1 receptor subtype (5,8). Because preadipocytes are known to be present in adipose tissue in the close environment of adipocytes, extracellular LPA produced by adipocytes could be involved in paracrine control of the number of preadipocytes in adipose tissue. However, the possible influence of this regulation on in vivo enlargement of adipose tissue is conditioned by the ability of preadipocytes to differentiate into adipocytes.
Interestingly, LPA was recently proposed to behave as a PPAR␥ agonist (9,10). Such a conclusion was based on the ability of LPA to compete the binding of the classical PPAR␥ agonist rosiglitazone on the purified PPAR␥ protein as well as on its ability to increase the transcription of a peroxisome proliferator response element reporter gene after transfection in RAW 264.7 monocytes and CV1 cells.
Knowing the pivotal role of PPAR␥ in adipogenesis, the initial objective of the present study was to determine whether LPA could regulate PPAR␥ activity in adipocytes and whether this could influence adipogenesis. We observed that LPA does not activate PPAR␥ in adipocytes but, conversely, down-regulates PPAR␥2 expression and impairs adipogenesis via LPA 1 receptor activation. Therefore, the local production of LPA by adipocytes may exert a tonic inhibitory effect on the recruitment of new adipocytes into adipose tissue.

MATERIALS AND METHODS
Animals-LPA 1 receptor null male mice (LPA 1 -KO) and their wild type (WT) litter mates (4) were handled in accordance with the principles and guidelines established by the National Institute of Medical Research (INSERM). They were housed conventionally in an animal room with constant temperature (20 -22°C) and humidity (50 -60%) and with a 12-h light/12-h dark cycle (lights on at 8:00 am). All mice had free access to food from UAR, Epinay, France (energy contents in percentage of kilocalories: 20% protein, 70% carbohydrate, and 15% fat) and water throughout the experiment. On the day of sacrifice the blood was collected on heparin, and glucose was immediately measured with a glucose meter. Plasma concentrations of insulin (Diagnostics Pasteur, Paris, France) and leptin (Linco) were determined with a radioimmunoassay kit. Plasma concentrations of triglycerides and free fatty acids were determined using a colorimetric kit (Wako).
Separation of Adipocytes and Stroma-vascular Cells from Adipose Tissue-Adipose tissue was dissected out and weighed before separating adipocytes from stroma-vascular cells as described previously (12). Adipose tissue was minced and incubated in 5 ml of Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 1 mg/ml collagenase and 1% bovine serum albumin for 30 -45 min at 37°C under shaking. The digestate was filtrated through a 100-m screen and centrifuged at 2100 rpm for 10 min to separate adipocytes as a floating fat cake and stroma-vascular cells in a pellet. Pelleted cells were induced to differentiate into adipocytes in primary culture as described in the next section.
Adipocyte Differentiation in Culture-Mouse 3T3F442A preadipocytes (13) were grown and differentiated into adipocytes as described previously (14). The preadipocytes were grown to confluence in DMEM supplemented with 10% donor calf serum (Invitrogen) and then cultured in an adipogenic medium consisting of DMEM supplemented with 10% fetal calf serum plus 50 nM insulin for 4 or 7 days (the medium was changed every 2 days). Triglyceride content was quantified using a colorimetric kit (Wako Chemicals) Proteins were quantified with a Bradford assay kit (Protassay, Bio-Rad). LPA present in adipogenic medium was quantified using a radioenzymatic assay as described previously (15). For LPA treatments, 1-oleoyl-LPA (18:1) solubilized in phosphate-buffered saline containing 1% fatty acid-free bovine serum albumin was used.
Adipocyte differentiation of primary preadipocytes from adipose tissue was induced in serum-free medium as described previously (12). Briefly, stroma-vascular cells were prepared (see previous section) from 4-week-old mice. Pelleted cells were suspended in 1 ml of erythrocyte lysis buffer (16 mM Tris-HCl and 0.08% NH 4 Cl, pH 7.65) for 2 min and then diluted in 50 ml of DMEM before centrifugation. Pelleted cells were filtrated through a 100-m screen, seeded in 12-well plates (150,000 cells/well), and cultured in a serum-free adipogenic medium (DMEM/Ham's F12 (1:1) medium supplemented with 10 mg/ml transferrin, 33 mM biotin, 66 mM insulin, 1 nM triiodothyronine, and 17 mM pantothenate) for 7 days. After 7 days of culture under these conditions, the proportion of newly differentiated adipocytes was evaluated by counting the number of lipid-laden (light-refringent) cells under the microscope.
Gene Expression-Total RNAs were extracted using the RNeasy mini kit (Qiagen GmbH, Hilden, Germany). Gene expression was analyzed using real time PCR as described previously (7). Total RNA (1 g) was reverse-transcribed for 60 min at 37°C using Superscript II reverse transcriptase (Invitrogen) in the presence of a random hexamer. A minus reverse transcriptase reaction was performed in parallel to ensure the absence of genomic DNA contamination. Real time PCR was performed starting with 25 ng of cDNA and a 900 nM concentration of both sense and antisense primers in a final volume of 25 l using the SYBR Green TaqMan Universal PCR master mix (Applied Biosystems). Fluorescence was monitored and analyzed in a GeneAmp 5700 detection system instrument (Applied Biosystems). Analysis of 18 S ribosomal RNA was performed in parallel using the ribosomal RNA control TaqMan assay kit (Applied Biosystem) to normalize gene expression. Results are expressed as 2 (Ct18 S Ϫ Ctgene) ϫ 10,000, where Ct corresponds to the number of cycles needed to generate a fluorescent signal above a predefined threshold (Table I). Oligonucleotide primers were designed using the Primer Express software (PerkinElmer Life Sciences).

Expression of LPA Receptor Subtypes and PPAR␥2 in
3T3F442A Cells-Experiments were carried out in the mouse 3T3F442A cell line characterized previously for its ability to differentiate into adipocytes when cultured in an appropriate adipogenic medium (see "Materials and Methods"). Confluent cells expressed PPAR␥2, LPA 1 receptor, and LPA 4 receptor mRNAs. LPA 2 and LPA 3 were not detected ( Fig. 1 and Table I). The PPAR␥2 mRNA level increased 4-fold between confluence and 8-day post-confluence adipocytes (Fig. 1). In contrast, the LPA 1 receptor mRNA level decreased by 14-fold (Fig. 1). The LPA 4 receptor mRNA level did not change during the course of adipocyte differentiation (Fig. 1).
LPA Inhibits Agonist-mediated Activation of PPAR␥2 in 3T3F442A Cells-LPA was reported for its ability to bind and activate the nuclear receptor PPAR␥ in monocytes and CV1 cells (9,10). PPAR␥ is a transcription factor known for its pivotal role in adipogenesis (17). Our initial objective was to determine whether LPA could activate endogenously expressed PPAR␥ in 3T3F442A cells. One way to evaluate PPAR␥ activation was to measure the ability of a PPAR␥-agonist such as rosiglitazone to increase the transcription of endogenously expressed genes containing a peroxisome proliferator response element in their promoters. As demonstrated previously (18), PEPCK is one of the most sensitive PPAR␥-sensitive genes in 3T3F442A adipocytes.
PPAR␥ activity was studied in 3T3F442A cells cultured in an adipogenic medium for 4 days (see "Materials and Methods"). At that stage the cells exhibited maximal expression of PPAR␥2 mRNA and minimal expression of LPA 1 receptor (Fig.  1). As a positive control of PPAR␥ activation, the influence of rosiglitazone on PEPCK mRNA was tested. Rosiglitazone led to a dose-dependent increase in PEPCK mRNA level with an EC 50 of 10 nM and a maximal effect of up to 17-fold when compared with control ( Fig. 2A). In contrast, LPA (10 M in 1% bovine serum albumin as vehicle) did not significantly modify the PEPCK mRNA level (Fig. 2, curve labeled ϩLPA, rosiglitazone at 0). When using other vehicles (ethanol, methanol, Me 2 SO, or translocase-3) (9), LPA alone still had no significant effect on PEPCK mRNA (data not shown). Interestingly, when LPA was used in co-treatment with rosiglitazone, the dose-response of rosiglitazone was significantly shifted to the right, leading to an EC 50 of up to 20 nM without modification of the maximal effect ( Fig. 2A). When the adipogenic medium was supplemented with 10 M LPA from confluence to day 4, the doseresponse of rosiglitazone on the PEPCK mRNA level was almost completely annealed (Fig. 2B). Similar results were obtained when analyzing the ALBP mRNA level as another PPAR␥-sensitive gene (not shown). These results suggested that LPA was not able to activate PPAR␥ in 3T3F442A cells but, conversely, inhibited agonist-dependent activation of PPAR␥.
LPA Inhibits PPAR␥2 Expression in 3T3F442A Cells-To test whether LPA-inhibition of agonist-mediated activation of PPAR␥ could result from a down-regulation of PPAR␥2 expression, the PPAR␥2 mRNA level was measured. In 4-day postconfluent 3T3F442A cells, 24 h of treatment with 10 M LPA led to a significant reduction (38%) in PPAR␥2 mRNA level when compared with control cells (Fig. 3A). When LPA was chronically present in the culture medium from confluence to day 4, PPAR␥2 mRNA level was further down-regulated (74%) when compared with control (Fig. 3A). In parallel, the PPAR␥ protein level was reduced after 4 days of treatment with LPA (Fig. 3B). No alteration of PPAR␥ protein level was observed after 24 h of LPA treatment. These results suggested that LPA-mediated inhibition of agonist-mediated activation of PPAR␥ resulted from down-regulation of PPAR␥2 expression.
LPA Inhibits Adipocyte Differentiation of 3T3F442A Cells-With PPAR␥2 playing a pivotal role in adipocyte differentiation, the influence of LPA on adipocyte differentiation of 3T3F442A cells was studied. When 3T3F442A cells were differentiated for 7 days in the presence of increasing concentra-  tion of LPA in the adipogenic medium, a dose-dependent reduction in triglyceride accumulation was observed (Fig. 4A). The effect of LPA was significantly detected at 4 M and was maximal at 8 M. LPA treatment also led to a significant reduction in the level of PPAR␥2, ALBP, and HSL mRNAs (Fig.  4B). These results revealed the anti-adipogenic activity of LPA.

LPA Normally Present in Serum Restrains
Adipocyte Differentiation of 3T3F442A Cells-As described under "Materials and Methods," the adipogenic medium used to allow differentiation of 3T3F442A cells contains 10% fetal calf serum, and this serum is known to contain LPA (15). We tested whether LPA brought by serum could influence adipogenesis. As shown in Fig. 5A, LPA concentration in fresh adipogenic medium measured using a radioenzymatic assay (15) was 0.45 M. As demonstrated previously, LPA can be hydrolyzed and inactivated by treatment with the lysophospholipase called phospholipase B (5, 7). As shown in Fig. 5A, 30 min of treatment with 0.5 units/ml phospholipase B led to a Ͼ90% reduction in LPA content of the adipogenic medium. When 3T3F442A cells were differentiated for 7 days in an adipogenic medium supplemented with 0.5 units/ml phospholipase B, the amount of triglycerides stored in the cells was significantly increased when compared with that for a control adipogenic medium (Fig. 5B). This was accompanied by a significant increase in the level of PPAR␥2, ALBP, and HSL mRNAs (Fig. 5C). These results showed that LPA brought by serum in the adipogenic medium restrains adipocyte differentiation of 3T3F442A preadipocytes. Most of the adipogenic medium used to differentiate preadipose cell lines contains serum. Our observations could help to optimize the composition of serum-containing adipogenic medium.
The Anti-adipogenic Activity of LPA Is Mediated by the LPA 1 Receptor Subtype-To determine whether the LPA 1 receptor was involved in the anti-adipogenic activity of LPA, adipocyte differentiation was analyzed in primary preadipocytes from LPA 1 -KO mice and their WT litter mates (4). Preadipocytes are present in the stroma-vascular fraction of the adipose tissue and were isolated from adipocytes after collagenase dissociation (see "Materials and Methods"). When prepared from WT mice, the stroma-vascular fraction expressed LPA 1 , LPA 2 , and LPA 4 receptor mRNAs, with the LPA 1 receptor subtype being predominantly expressed (Table I). LPA 3 receptor mRNAs were undetectable. When the stroma-vascular fraction was prepared from LPA 1 -KO mice, LPA 1 mRNAs were undetectable, and LPA 2 and LPA 4 mRNA levels were not different from those in WT mice (Table I).
The stroma-vascular fraction of adipose tissue is not only composed of preadipocytes but also of endothelial cells and macrophages. To evaluate the proportion of preadipocytes present, the stroma-vascular fraction was cultured in an adipogenic medium (see "Materials and Methods") for 7 days. Under these conditions a certain proportion of cells became light-refringent as the result of triglyceride droplet accumulation (Fig. 6A) and the expression of adipocyte-specific genes (ALBP, PPAR␥2, and HSL) (not shown). Interestingly, the proportion of lipid-laden cells obtained with WT mice (24%) was significantly lower than that obtained with LPA 1 -KO mice (Fig. 6B). This finding was associated with a lower expression of adipocyte-specific genes (not shown). This observation suggested that the proportion of preadipocytes present in the stroma-vascular fraction from WT mice was lower than that from LPA 1 -KO mice.
When starting with WT mice, supplementation of the adipogenic medium with increasing concentrations of LPA during the 7 days of culture led to a dose-dependent decrease (maximal inhibition of 50% at 0.1 M LPA) in the proportion of lipid-laden cells when compared with control cells (Fig. 7). When starting with LPA 1 -KO mice, no changes in the proportion of lipid-laden cells (Fig. 7) were observed after LPA treatment. These data showed that the absence of an LPA 1 receptor in preadipocytes led to the suppression of the anti-adipogenic activity of LPA.
Adipose Tissue Phenotype of LPA 1 -KO Mice-To determine the possible consequences of LPA 1 invalidation on adipose tissue development, LPA 1 -KO mice were analyzed and compared with WT mice. Data presented were obtained with males, but similar results were obtained with females (not shown). LPA 1 -KO mice exhibited lower body weight than WT mice, whatever the age of the animals (Fig. 8A). This was accompanied by no differences in the mean daily food intake measured over 10 weeks, i.e. 1.0 Ϯ 0.2 and 0.8 Ϯ 0.1 gram of food per gram of body weight per day for WT (n ϭ 6) and LPA 1 -KO (n ϭ 3) mice, respectively. At 15 weeks of age and despite the lower body weight, perigonadic adipose tissue weight was significantly higher in LPA 1 -KO mice than in WT mice (Fig. 8B). Inguinal adipose tissue weight also tended to be higher in LPA 1 -KO mice than in WT mice, but this was not significant (Fig. 8B). At the plasma level, LPA 1 -KO mice exhibited a significantly higher (2-fold) concentration of leptin when compared with WT mice (Table II). In contrast, no differences in plasma concentration of insulin, glucose, triglycerides, and free-fatty acids were observed (Table II). Taken together, these observations showed that, despite their lower body weight, LPA 1 -KO mice exhibited higher adiposity than WT mice.
LPA Inhibits Adipocyte Differentiation of Human Preadipocytes-The possible anti-adipogenic activity of LPA was tested in human preadipocytes. This was tested in the human preadipocyte cell strain from SGBS (16). These cells have previously been described for their ability to differentiate into adipocytes in a serum-free medium (see "Materials and Methods"). Confluent SGBS preadipocytes expressed both LPA 1 and LPA 2 receptor mRNAs (Table I), with the LPA 1 receptor RNAs being 150-fold more expressed than LPA 2 receptor RNAs (Table I). In contrast, LPA 3 receptor and LPA 4 receptor RNAs remained undetectable.
After 10 days of culture in an appropriate serum-free adipogenic medium, SGBS cells accumulated triglyceride droplets (Fig. 9A) and expressed adipocyte-specific genes such as ALBP and HSL (Fig. 9B). Supplementation of the adipogenic medium with LPA led to a striking reduction in triglyceride droplet accumulation (Fig. 9A) accompanied by a dose-dependent reduction in ALBP and HSL gene expression (Fig. 9B). These results showed that LPA was anti-adipogenic in human preadipocytes. DISCUSSION The recruitment of new fat cells in adipose tissue requires the differentiation of preadipocytes into adipocytes (adipogenesis), FIG. 5. Influence of phospholipase B on adipocyte differentiation of 3T3F442A cells. A, the adipogenic medium was treated or not treated (Cont) with 0.5 units/ml phospholipase B (PLB) for 30 min at 37°C before quantification of LPA concentration using a radioenzymatic assay (see "Materials and Methods"). B and C, confluent 3T3F442A cells were cultured in an adipogenic medium (see "Materials and Methods") in the presence or absence of 0.5 units/ml phospholipase B (PLB). After 10 days the amount of triglycerides accumulated in adipocytes (B) and the expression of adipocyte-specific genes (C) were measured as described under "Materials and Methods." Values are means Ϯ S.E. of three separate experiments. Comparison with control was performed using Student's paired t test; *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. a process tightly controlled by the transcription factor PPAR␥2. Factors locally produced in adipose tissue by adipocytes could contribute to the regulation of adipogenesis by exerting paracrine actions on preadipocytes. Among those paracrine factors LPA could play an important role, because it is produced at the extracellular face of adipocytes by autotaxin (7) and because preadipocytes express LPA receptors (mainly the LPA 1 subtype) (8). We demonstrated previously that, in 3T3F442A preadipocytes, LPA was able to increase phosphorylation of the mitogen-activated protein kinases ERK1 and ERK2 and to increases proliferation (8,19), but the influence of LPA on adipogenesis was not studied until the present work. Such study was further motivated by recent reports showing that LPA could, in parallel to its ability to activate G protein-coupled membrane receptors, behave as an agonist to the nuclear transcription factor PPAR␥ in monocytes and CV1 cells (9,10). Because PPAR␥ is known to play a pivotal role in the control of adipogenesis (17), determining whether LPA could activate PPAR␥ in adipocytes was of main interest in the context of adipose tissue.
The first part of the present study shows that, unlike the PPAR␥ agonist rosiglitazone, LPA was unable to activate PPAR␥ activity in adipocytes evaluated by measuring the induction of PPAR␥-sensitive genes (PEPCK and ALBP). These results lead us to conclude that LPA does not behave as a potent activator of PPAR␥ in adipocytes. This conclusion is not in agreement with that drawn previously from experiments performed in monocytes and CV1 cells (9,10). Although LPA was demonstrated to bind to PPAR␥ in an in vitro assay (9), the ability of LPA to activate PPAR␥ obviously appears to be dependent on the cell type. Possible activation of a nuclear receptor such as PPAR␥ by exogenous LPA would require that a high enough amount of LPA penetrated into the cell and reached the nucleus. As demonstrated previously by our group, when preadipocytes are exposed to radiolabeled LPA virtually no radiolabeled LPA can be detected in the cells because of the presence of a high ecto-lipid phosphate phosphohydrolase activity, which dephosphorylates and inactivates LPA (20). This could explain why we observed no activation of PPAR␥ by LPA in adipocytes. This finding also suggests that ecto-lipid phosphate phosphohydrolase activity could be weaker in monocytes and CV1 cells  than in adipocytes, allowing a higher amount of LPA to enter into the cells and activate PPAR␥. The most important result of the present study was the finding that chronic exposure of preadipocytes to LPA inhibits their differentiation into adipocytes, as attested by a reduction in triglyceride accumulation and a reduction in the expression of adipocyte specific genes. Therefore, LPA clearly behaves as an anti-adipogenic compound. In addition, the anti-adipogenic effect of LPA was not found in primary preadipocytes from LPA 1 knock-out mice, indicating that the LPA 1 receptor is fully responsible for the anti-adipogenic activity of LPA. These observations clearly support the concept that LPA is an antiadipogenic mediator acting via a specific receptor.
PPAR␥2 clearly plays a pivotal positive role in adipogenesis (17). Our data show that treatment with LPA leads to downregulation of PPAR␥2 expression and activity, as expected for an anti-adipogenic factor. It is therefore very likely that the anti-adipogenic activity of LPA is due to its negative impact on PPAR␥2. This conclusion is supported by previous reports showing that LPA is able to inhibit PPAR␥ activity in THP-1 monocytes and IMR-32 neuroblastomas (21,22). In this last cell line, LPA inhibits the capacity of PGJ2 to activate the transcription of a PPAR␥ responsive element-dependent reporter gene, and by using specific pharmacological inhibitors it was shown that the effect of LPA involves activation of the mitogen-activated protein kinase pathway as well as activation of the Rho kinase (22). In addition, it has previously been shown that increasing Rho-GTPase activity decreases adipocyte differentiation and PPAR␥ expression in mouse embryoderived fibroblasts (23). In preadipocytes, LPA activates the phosphorylation of ERK1 and ERK2 mitogen-activated protein kinases and activates the small G-protein Rho (19). It can therefore be proposed that LPA-mediated inhibition of PPAR␥2 activity and expression may result from the LPA 1 receptor-dependent activation of the mitogen-activated protein kinases and/or the Rho kinase.
What could be the physiological relevance of the anti-adipogenic activity of LPA? Because the LPA 1 receptor was responsible for the anti-adipogenic activity of LPA in preadipocytes, phenotypic analysis of LPA 1 -KO mice presented an excellent opportunity for determining the possible involvement of LPA on adipose tissue development. As demonstrated previously, LPA 1 -KO mice exhibit reduced the size and weight of their bodies (4). In the present study, we observed that despite their lower body weight, LPA 1 -KO mice exhibited higher adiposity than WT mice. This was associated with a higher plasma concentration of leptin, a cytokine known to be tightly associated with adipose tissue mass (24). Because LPA 1 -KO mice exhibited the same food intake as WT mice, the higher adiposity of LPA 1 -KO mice cannot be explained by an alteration of their feeding behavior. Enlargement of adipose tissue not only results from increased accumulation of triglyceride accumulation in adipocytes but can also be influenced by the recruitment of new adipocytes resulting from adipogenesis. Interestingly, the proportion of preadipocytes present in adipose tissue from LPA 1 -KO mice was higher than that of WT mice (Fig. 6). It is therefore possible that increased adiposity of LPA 1 -KO mice may result from the suppression of the anti-adipogenic activity of LPA normally existing in WT mice.
In conclusion, LPA is produced in the extracellular medium of adipocytes as the result of the lysophospholipase D activity of autotaxin (6,7). Adipocytes and preadipocytes are both present in adipose tissue in a close environment. Because preadipocytes express LPA 1 receptor and are sensitive to LPA, it is likely that local production of LPA by adipocytes could exert a paracrine anti-adipogenic activity on surrounding preadipocytes and therefore exert a negative paracrine effect on adipose tissue development.