Fat-specific protein 27 modulates nuclear factor of activated T cells 5 and the cellular response to stress.

Fat-specific protein 27 (FSP27), a member of the cell death-inducing DNA fragmentation factor α-like effector (Cide) family, is highly expressed in adipose tissues and is a lipid droplet (LD)-associated protein that induces the accumulation of LDs. Using a yeast two-hybrid system to examine potential interactions of FSP27 with other proteins, a direct interaction with the N-terminal region of nuclear factor of activated T cells 5 (NFAT5) was identified. NFAT5 is a transcription factor that induces osmoprotective and inflammatory genes after its translocation to the nucleus. The interaction between FSP27 and NFAT5 was confirmed by bimolecular fluorescence complementation and coimmunoprecipitation. Using immunocytochemistry, NFAT5 is detected in the cytoplasm and in the nucleus under isotonic conditions; however, overexpression of FSP27 inhibited the hypertonic-induced nuclear translocation of NFAT5. Consistent with the suppression of NFAT5 nuclear translocation, in cells transfected with a reporter construct containing the NFAT5 response element from the monocyte chemoattractant protein 1 (MCP1) promoter, FSP27 overexpression repressed hypertonic-induced luciferase activity and the expression of NFAT5 target genes. Knockdown of FSP27 in differentiated 3T3-L1 adipocytes increased the NFAT5-mediated rise in MCP1. These results suggest that FSP27 not only modulates LD homeostasis but also modulates the response to osmotic stress via a physical interaction with NFAT5 at the LD surface.


RT-PCR analysis
Cells were homogenized in TRIzol® Reagent (Invitrogen), and total RNA was extracted and purifi ed following the manufacturer's protocol. Total RNA was reverse-transcribed by Super-Script® II Reverse Transcriptase (Invitrogen) with random primers. Real-time PCR was performed in triplicate with an Rt2Real-Time™ SYBER Green/Rox PCR master mix (Qiagen, Hilden, Germany) and specifi c primer pairs ( Table 1 ). The relative mass of specifi c RNA was calculated by the comparative cycle of threshold detection method according to the manufacturer's instructions. Acid ribosomal phosphoprotein P0 (36B4) was used as an internal control.

Immunoprecipitation
Cells were homogenized in TNE buffer (10 mM Tris-HCl [pH 7.8], 150 mM NaCl, 1 mM EDTA, and 1% NP40) with 10 µg/ml leupeptin. An aliquot (100 g) was incubated with 0.5 µg of anti-HA-Ab. After overnight incubation at 4°C, the immune complexes were incubated with 20 µg of Protein A/G beads (sc-2003; Santa Cruz) for 2 h at 4°C and washed fi ve times with TNE, and the pellet was resuspended in 63 mM Tris-HCl (pH 6.8) with 5% 2-mercaptoethanol, 2% SDS, 5% (w/v) sucrose, and 0.005% BPB and used for Western blotting. Immunoprecipitation from adipose tissue was performed as described previously ( 13 ). Briefl y, adipose cells were isolated from normal mouse epididymal fat depots, and the cells were homogenized in 50 mM Tris-HCl (pH 7.4), 8% sucrose, 1 mM EDTA, 0.1 mM Na 3 VO 4 , and 50 mM NaF with 10 mg/ml leupeptin. Cytosol and fat cake fractions were separated by centrifugation at 10,000 g for 15 min. The fat cake was further extracted with lysis buffer containing 5% SDS for 30 min. An aliquot (250 g) was precleared with Protein A beads and incubated with an immunomatrix consisting of rabbit polyclonal anti-NFAT5 IgG (sc-5501; Santa Cruz) and protein A or normal mouse IgG as a control. After overnight incubation at 4°C, the immune complex was centrifuged at 10,000 g for 15 min and washed twice in PBS with 0.05% BSA and then twice in PBS. The pellet was resuspended in SDS-PAGE loading buffer (0.063 M Tris-HCl [pH 6.8], with 1% 2-mercaptoethanol, 1% SDS, and 13% [v/v] glycerol), boiled for 5 min, electrophoresed on 12% SDS-PAGE, transferred to nitrocellulose paper, and immunoblotted with anti-CIDE3 IgG (1:1,000, GWB-269046; Genway Biotech, Inc., mice ( 5,6 ), which display lean phenotypes, accompanied by smaller LDs in WAT and larger LDs in BAT, resulting in enhancement of insulin sensitivity and resistance to dietinduced obesity. However, a truncated mutation of Cidec was found in an individual with partial lipodystrophy and insulin-resistant diabetes ( 12 ), perhaps suggesting a different response in mice and humans.
FSP27 can form a complex with its family member Cidea ( 9 ), but the interaction of FSP27 with other cellular proteins has not been reported. In the current study, we sought to identify proteins that interact with FSP27 using a yeast two-hybrid system. Our results revealed that FSP27 directly interacts with the N-terminal region of tonicity response element binding protein (TonEBP, also known as nuclear factor of activated T cells 5 [NFAT5]). The interaction between FSP27 and NFAT5 was confi rmed by bimolecular fl uorescence complementation (BiFC) and coimmunoprecipitation. Moreover, we showed that FSP27 inhibits the normal translocation of NFAT5 from the cytoplasm to the nucleus under hypertonic conditions and attenuates its transcriptional activity. Thus, FSP27 appears not only to have an important role in LD metabolism but also to modulate cellular signaling in response to stress.

Yeast two-hybrid system
Yeast two-hybrid screening was performed as previously described ( 13 ). Full-length rat FSP27 cDNA was cloned into the Sal I/ Nco I site of the pAS1-CYH2 yeast two-hybrid vector ( 14 ) to yield an FSP27-GAL4 DNA-binding domain vector that was used as bait for the screening of a rat adipocyte library.

Construction of plasmids
The vectors pBiFC-VN173 and pBiFC-VC155 were kindly provided by Dr. Chang-Deng Hu from Purdue University ( 15 ). Fulllength mouse FSP27 cDNA was cloned into the EcoR I/ Xho I site of the pBiFC-VC155. Rat NFAT5 (residues 66-274) was cloned into the EcoR I/ Xba I site of the pBiFC-VN173. The reporter constructs containing the putative NFAT5 element enhancer reporter gene pGL3-Wt for the luciferase assay was kindly provided by Dr. Ryoji Kojima from Meijo University, Japan ( 16 ).

DNA transfection
HEK293 cells were plated onto 48-well plates at 2 × 10 4 cells/ well and transfected with expression plasmids. Transfections were performed with Lipofectamine™ 2000(Invitrogen) following the manufacturer's protocol.

Identifi cation of NFAT5 as a partner of FSP27
To identify protein(s) that interact with FSP27, we used the yeast two-hybrid system to screen a rat adipocyte expression library for proteins that would interact with fulllength rat FSP27 used as bait. Ten positive clones were identifi ed out of the 5 × 10 6 colonies screened, and sequencing results revealed that one of the clones contained an in-frame fragment of TonEBP/NFAT5 (amino acid residues 157-269). This domain includes the nuclear localization signaling motif, which is highly conserved among rat, mouse, and human (>99% at the amino acid level) ( Fig. 1 ).
NFAT5 is a widely expressed transcription factor that is a major inducer of osmoprotective gene products in mammalian cells ( 19,20 ). To identify tissues where FSP27 and NFAT5 are coexpressed and could potentially interact, we examined their relative mRNA expression in various tissues. FSP27 was detected in all WAT and BAT ( Fig. 2A ) and in heart and muscle in the mouse. NFAT5 mRNA was expressed in all tissues we tested, including WAT and BAT, with relatively similar levels of expression ( Fig. 2B) . The expression of FSP27 ( Fig. 2C ) and NFAT5 ( Fig. 2D ) increased during differentiation in 3T3-L1 adipocytes, but FSP27 was much more highly expressed after differentiation.
To confi rm the interaction of FSP27 with NFAT5, we used a BiFC assay to visualize the complexes in living cells. We constructed plasmids that express the genes of interest (NFAT5 66-274 and full-length FSP27) as fusion proteins San Diego, CA) and visualized using an Odyssey® Imaging System.

Immunofl uorescence staining
HEK293 cells were transfected with pBiFC-FSP27-VC or pBiFC-VC155 as a control. Cells were incubated with 120 µM oleate and palmitate or 1% BSA as a control for 24 h, and osmotic stress was induced by 100 mM NaCl for 12 h. Cells were fi xed in 4% paraformaldehyde in PBS (pH 7.4) at 4°C for 1 h, washed with PBS with 0.2% Triton-X100 at 25°C for 30 min, and blocked with 3% BSA in PBS. Anti-NFAT5 Ab and anti-HA Ab (1:500, 2367S; Cell Signaling, Danvers, MA) were added to the cells at 4°C overnight and visualized by incubation with Alexa Fluor 488 or 568-conjugated secondary Ab (1:500, A11034 and A11031; Invitrogen).

Knockdown of FSP27
A recombinant AAV2 expressing shRNA targeting FSP27 was constructed by the Neuroscience Gene Vector and Virus Core, Stanford Institute of Neuro-innovation and Translational Neuroscience, Stanford University. The nucleotide sequences for the shRNA against FSP27 and a scrambled control were as follows: FSP27, AAAAGGAAGGTTCGCAAAGGCATCATTCGTGATGCC T TTGCGAACCTTCC; scrambled reverse, AAAAGCGCGCTTT-GTAGGATTCGTTCGCGAATCCTACAAAGCGCGC ( 3 ). For in vitro infection of AAV2-shRNA, fully differentiated 3T3-L1 adipocytes were incubated with 1.0 × 10 11 AAV2-shRNA for 24 h, and cells were harvested 48 h later.

Apoptosis
For evaluation of apoptosis, HEK293 cells in 48-well plates were transfected with pBiFC-VC155 or pBiFC-FSP27-VC. Cells were incubated with 120 µM oleate and palmitate or 1% BSA as a control for 24 h to promote lipid accumulation, and osmotic stress was induced by 100 mM NaCl or 200 mM D-glucose for 12 h. After that, apoptotic cells were stained using an annexin-V-FLUOS staining kit (#11 858 777 001; Roche, Mannheim, Germany) following the manufacturer's protocol. and BiFC. As an additional means of documenting the interaction between FSP27 and NFAT5, Flag-NFAT5 and HA-FSP27 were expressed in HEK293 cells, immunoprecipitated with anti-HA antibody, and immunoblotted with anti-Flag antibody. Fig. 4A documents the expression of the Flag-tagged constructs. NFAT5 coimmunoprecipitated with FSP27 ( Fig. 4B ), but another cytosolic protein, Flag-HSL, expressed as a control, did not coimmunoprecipitate with FSP27; however, Flag-HSL was coimmunoprecipitated when HA-Plin1 was coexpressed, consistent with our previous observations ( 22 ). To document the inter action of endogenously expressed FSP27 and NFAT5, proteins extracted from the fat cake obtained from isolated normal mouse adipose cells were immunoprecipitated with anti-NFAT5 antibody and then immunoblotted with anti-FSP27 antibody ( Fig. 4C ). FSP27 was highly expressed in the fat cake (lane 1) and coimmunopre cipi tated with NFAT5 (lanes 3 and 5) but not when non immune IgG (lanes 2 and 4) was used for the immunoprecipitation, thus substantiating the physical with the N-terminal (VN173) or C-terminal (VC155) of Venus for the BiFC assay. If FSP27 and NFAT5 interact with each other, the interaction should bring VN173 and VC155 into close proximity and reconstitute an intact Venus molecule, which can be detected by direct visualization with fl uorescence imaging ( 21 ). To confi rm the interaction between the fusion proteins, we transiently expressed the proteins in HEK293 cells that had been loaded with fatty acids (FAs) to promote LD formation. Because HSL and Plin1 have previously been reported to interact ( 22 ), we used HSL-VN and Plin-VC as a positive control for the BiFC assay. Consistent with a physical interaction between HSL and Plin1, the BiFC assay showed a fl uorescent signal that appeared localized on the LDs ( Fig. 3A,  B ). In contrast, HSL-VN and FSP27-VC did not demonstrate any fl uorescent signal ( Fig. 2C, D ). Coexpression of NFAT5-VN and FSP27-VC yielded a fl uorescent signal that appeared to surround the LDs that had been stained with bodipy ( Fig. 3E-H ). Thus, FSP27 and NFAT5 appear to physically and specifi cally interact by genetic screening in HEK293 cells in the presence or absence of overexpressed FSP27. HEK293 cells were transiently transfected with FSP27-VC or VC155 as a control for transfection, incubated with fatty acids (FAs) to induce LDs, and then cultured in hypertonic medium with an additional 100 mM NaCl. Under basal conditions, VC155 is ubiquitously expressed throughout the cell ( Fig. 5B ), and endogenous NFAT5 is observed throughout the cytoplasm and nucleus ( Fig. 5D ), with the merged image of DAPI ( Fig. 5C ) and NFAT5 shown in Fig. 5E . After hypertonic stress, VC155 continues to be found throughout the cytoplasm and nucleus ( Fig. 5G ), whereas NFAT5 is now localized almost exclusively to the nucleus ( Fig. 5I ), with the merged image shown in Fig. 5J . Under basal conditions, FSP27-VC is found throughout the cytoplasm ( Fig. 5L ), and endogenous NFAT5 is again observed throughout the cytoplasm and nucleus ( Fig. 5N, O). However, after hypertonic stress, in cells that express FSP27-VC ( Fig. 5Q ), NFAT5 is not exclusively nuclear but remains localized to the cytoplasm and nucleus ( Fig. 5S, T ). As another means to evaluate the effects on the nuclear traffi cking of NFAT5, HEK293 cells were transiently transfected with FSP27-VC or VC155 as a control for transfection, incubated with FAs to induce LDs, after cultured in isotonic (Control) or hypertonic medium (100 mM NaCl added), and then the cells were lysed, nuclei were isolated, and nuclear proteins were immunoblotted for NFAT5 ( Fig. 6 ). The expression of total cellular NFAT5 was induced by osmotic stress, and this stimulation was not affected by overexpression of FSP27 ( Fig. 6A) . Cell fractionation experiments confi rmed the immunocyto chemistry fi ndings, demonstrating that the amount of NFAT5 found in the nuclear fraction was increased markedly after osmotic stress, but the amount of nuclear NFAT5 was decreased 60% by overexpression of FSP27 under both basal and osmotic stress conditions ( Fig. 6B ). Thus, heterologous expression of FSP27 decreases the nuclear traffi cking of NFAT5 normally observed after hypertonic stress.

FSP27 modulates the transcriptional activity of NFAT5
Because the expression of FSP27 attenuates the translocation of NFAT5 to the nucleus, we tested whether the transcriptional activity of NFAT5 was similarly inhibited in the presence of FSP27. Kojima et al. (16) identifi ed the cis -acting regulatory enhancer elements within the monocyte chemoattractant protein 1 (MCP1) gene that contains the putative NFAT5 binding sequence and that is responsive to hypertonic stress. We used a reporter gene construct containing the putative NFAT5 element from the MCP1 promoter with the luciferase reporter gene and the basic plasmid as a control for promoter analysis. HEK293 cells were transiently transfected with VC155 or FSP27-VC along with reporter plasmids, incubated with or without FAs, and cultured in hypertonic medium with an additional 100 mM NaCl. Promoter activity was increased 2.4fold by NaCl, 1.6-fold by FA loading, and 2.5-fold by FA loading along with NaCl ( Fig. 7A ) . Thus, although FA loading increased reporter activity, NaCl-induced hypertonic stress was most potent. We also incubated cells with D-glucose to induce osmotic stress, which resulted in a interaction between FSP27 and NFAT5 under conditions without forced overexpression.

FSP27 inhibits the translocation of NFAT5 to the nucleus
Indirect immunofl uorescence analysis of NFAT5 has demonstrated that hypertonicity in cells increases the nuclear staining of NFAT5 ( 23 ), suggesting that nuclear traffi cking regulates NFAT5 transcriptional activity ( 19 ). Based on the results of the BiFC assay and the fact that FSP27 is localized to LD within the cytoplasm, we hypothesized that the interaction between FSP27 and NFAT5 should inhibit the nuclear translocation of NFAT5 after hypertonic stress. To test this hypothesis, the translocation of endogenous NFAT5 after hypertonic stress was examined conditions had no effect on apoptosis when VC155 was overexpressed as a control, but overexpression of FSP27 tended to increase apoptosis, and this was slightly augmented under hypertonic conditions such that differences compared with overexpression of VC155 were now statistically signifi cant in the presence of hypertonic NaCl or glucose.

FSP27 controls MCP1 expression in 3T3-L1 adipocytes
Because the prior studies were conducted in a cell where FSP27 is not normally expressed, we next tested the functional interaction of FSP27 and NFAT5 in differentiated 3T3-L1 adipocytes under hypertonic conditions ( Fig. 9A ). FSP27 mRNA was decreased 50% by NaCl (100 mM added) and 60% by D-glucose (200 mM) incubation. NFAT5 mRNA was increased 3.1-fold by NaCl and 3.5-fold by 1.8-fold increase in reporter activity. Importantly, coexpression of FSP27 signifi cantly decreased luciferase activity under each condition ( p < 0.05) ( Fig. 7A ). Furthermore, RT-PCR showed that FSP27 overexpression inhibited endogenous MCP1, PAI-1, TNF-␣ , and IL-6 expression in HEK293 cells under hypertonic conditions ( Fig. 7B ). Thus, heterologous expression of FSP27 not only inhibits the nuclear traffi cking of NFAT5 normally observed after hypertonic stress but also inhibits the transcriptional activity of NFAT5. Because FSP27 and its other cell death-inducing DNA fragmentation factor ␣ -like effector (Cide) family members are known to induce apoptosis ( 24 ), particularly in cells lacking LDs ( 25 ), we assessed the effects of overexpression of FSP27 on apoptosis in HEK293 cells under isotonic and hypertonic conditions ( Fig. 8 ). Hypertonic  by the NFAT5-mediated increase in MCP1 mRNA expression ( Fig. 9C ).

DISCUSSION
It is widely accepted that obesity and type 2 diabetes are characterized by evidence of a chronic, low-grade infl ammatory process ( 26,27 ), as manifested by elevated circulating levels of acute phase proteins (e.g., C-reactive protein) and cytokines (e.g., TNF-␣ , IL-1, and IL-6). In addition, adipose tissue in these settings has increased expression of chemokines, such as CC-chemokine ligand 2 (also known as MCP1), that can recruit macrophages. Infl ammation in obesity is associated with an increased infi ltration of macrophages, mast cells, and T cells into adipose tissue, with an associated increase in infl ammatory mediators. Several mechanisms have been proposed to explain the inciting insult leading to infl ammation, such as tissue hypoxia, cell death due to an enlargement of an adipose cell that exceeds the boundaries of its cell membrane, and cell stress due to lipotoxicity, ER stress, oxidative stress, or glucotoxicity. Each of these mechanisms for cell stress has been associated with the activation of several signaling pathways, such as inhibitor of B kinase, which activates nuclear factor B (NF-B) and controls the transcription of a wide variety of infl ammatory genes, and JNK, which controls the activation of several transcription factors. In addition, several other signaling pathways are activated by cell stress in diabetes and obesity ( 28 ), including PKR, ERK, PKC, and p38 MAPK.
NFAT5, also known as TonEBP and as osmotic response element (ORE) binding protein, is a member of the Rel family of transcription factors, which also includes NF-B D-glucose, and MCP1 mRNA was increased 4.5-fold and 6.1-fold by NaCl and D-glucose, respectively. Thus, NFAT5 and MCP1 mRNA expression was inversely correlated with FSP27. shRNA against FSP27 delivered via an AAV2 viral vector was able to knockdown FSP27 80% under control conditions and 90% after hypertonic stress in differentiated 3T3-L1 cells as compared with scrambled shRNA, whereas hypertonic stress alone reduced FSP27 expression 40% ( Fig. 9B ). The 80% reduction of FSP27 in differentiated 3T3-L1 cells dramatically increased MCP1 mRNA expression in the absence of hypertonic stress without altering the cellular response to hypertonic stress as assessed Fig. 6. FSP27 overexpression decreases nuclear NFAT5. HEK293 cells were transfected with HA-VC155 or with HA-FSP27-VC155, loaded with lipid, and incubated without (Control) or with an additional 100 mM NaCl. A: Total cellular proteins were used for immunoblotting with anti-NFAT5 and anti-␤ -actin Ab. The gel was scanned, and the densitometric ratio of NFAT5 to ␤ -actin (used as a loading control) is shown in the bar graph. B: Nuclei were isolated and nuclear proteins were used for immunoblotting with anti-NFAT5 and anti-HDAC4 Ab. The gel was scanned, and the densitometric ratio of NFAT5 to HDAC4 (used as a loading control for the nuclear fraction) is shown in the bar graph. Results are representative of two independent experiments. is localized to the cytoplasm under hypotonic conditions and translocates to the nucleus after osmotic (hypertonic) stress. This nucleocytoplasmic traffi cking is mediated via a nuclear localization signal regulated by a number of different kinases, including p38, Fyn, ATM, PKA, and CDK5 ( 19,31 ), and via a nuclear export sequence ( 32 ).
LDs are intracellular organelles that store neutral lipids within cells, where the LD serves as a reservoir for energy stores and to protect the cell from lipotoxic effects of FAs via their incorporation into TAG within the LD ( 1 ). Formation of LDs in response to overfeeding (obesity) is associated with ER stress ( 33 ), linking LDs with infl ammation. Proteomic analyses of LDs from a variety of cells have revealed the existence of many LD-associated proteins. The most abundant of these proteins belong to the PAT family, with Plin1 being the predominant member in adipocytes. In addition to PAT proteins, the Cide (cell death-inducing DNA fragmentation factor ␣ -like effector) family of proteins associates with and regulates LD physiology in addition to their role in apoptosis ( 25 ). The Cide family consists of three proteins, Cidea, Cideb, and Cidec (also called FSP27), and shares regions of homology with Plin that are outside the conserved PAT region. Cide proteins are predominantly expressed in adipose tissue and liver in mice, with Cidea primarily in BAT and FSP27 in WAT and Cideb in liver. Ectopic expression of Cide proteins promotes LD formation and reduces TAG hydrolysis. Moreover, knockdown of FSP27 decreases LD size and increases LD number in adipocytes ( 5,6,11 ), suggesting that FSP27 is involved in promoting the formation and maintenance of a unilocular LD within adipocytes, apparently by mediating the clustering and fusion of LDs and promoting lipid exchange ( 34,35 ).
In the current work we identifi ed a direct interaction of FSP27 with NFAT5 using a yeast two-hybrid screen. The interaction between FSP27 and NFAT5 was confi rmed by BiFC and by coimmunoprecipitation and appeared to occur on LDs where FSP27 is normally localized. We could not detect NFAT5 localization to LDs in the absence of FSP27 expression (data not shown). The interaction of FSP27 with NFAT5 seems to result in the sequestration of NFAT5 at the LD surface, thus attenuating the nuclear traffi cking of NFAT5 normally observed after activation of signaling pathways induced by osmotic stress.
Consistent with the attenuation of nuclear traffi cking of NFAT5, overexpression of FSP27 suppressed the transcriptional activity of a luciferase reporter construct containing the putative ORE from the MCP1 promoter and attenuated the increase in MCP1 mRNA expression as well as the expression of mRNA of other infl ammatory genes, such as TNF-␣ , PAI-1, and IL-6, observed after osmotic stress induced by NaCl or glucose in HEK293 cells. FA loading of HEK293 cells also increased the activity of the ORE promoter reporter construct and MCP1 expression under isotonic conditions. Compatible with the FA induction of MCP1 being mediated via NFAT5, this, too, was attenuated by overexpression of FSP27. This suggests that one of the mechanisms through which elevated FA levels and higher rates of lipolysis induce infl ammation and the and NFAT1-4 ( 19,20 ). In contrast to NFAT1-4, NFAT5 does not contain a calcineurin-sensitive domain and is not responsive to calcineurin inhibitors; however, it shares sequence homology with all other Rel family members in its DNA binding domain. All Rel family members, including NFAT5, function as homo-and heterodimers. NFAT5 is the largest Rel protein, containing almost 1,500 amino acids, and is the major transcription factor activated in response to osmotic stress, where it regulates an osmoprotective gene expression program via binding to OREs, which increases enzymes and transporters that elevate intracellular organic osmolytes and heat shock proteins. NFAT5 is ubiquitously expressed and functions not only in the renal medulla but also in other tissues, where it regulates a variety of other genes not directly involved in osmoregulation, including infl ammatory genes such as MCP1 ( 16 ) and TNF-␣ , among others ( 29 ). Moreover, NFAT5 has recently been shown to interact with NF-B, thereby enhancing the transcriptional activity of NF-B ( 30 ). NFAT5  recruitment of macrophages into adipose tissue ( 36,37 ) might be mediated via NFAT5. Nonetheless, osmotic stress resulted in a maximal response of MCP1 transcription in the current studies that was not infl uenced by FA loading. Not surprisingly, the overexpression of FSP27 was associated with an increase in apoptosis, though this is minimized in the presence of LD induction by fatty acid loading ( 25 ); however, it seems unlikely that the small changes observed in apoptosis substantially contributed to the ability of FSP27 to inhibit the nuclear traffi cking and repress the transcriptional activity of NFAT5 because overexpression of FSP27 had no effect on the induction of NFAT5 by osmotic stress. Moreover, our observations were not confi ned to HEK293 cells because exposure of differentiated 3T3-L1 adipocytes to osmotic stress (either NaCl or glucose) resulted in an increased expression of NFAT5 and MCP1 mRNA. In these adipocytes where FSP27 is endogenously expressed, osmotic stress resulted in the simultaneous suppression of FSP27, whereas NFAT5 and MCP1 were increased. This suggests that the decrease of FSP27 that occurs as a part of the response to osmotic stress removes the restraint on the movement of NFAT5 to the nucleus and magnifi es the transcriptional function of NFAT5. The results of the knockdown of FSP27 in differentiated 3T3-L1 adipocytes are consistent with this interpretation because the removal of FSP27 resulted in an increase in MCP1 even in the absence of osmotic stress. In addition to removing the restraint of the movement of NFAT5 to the nucleus, the osmotic stress-induced suppression of FSP27 would be expected to increase lipolysis by alleviating the barrier function of FSP27 toward intracellular lipases.
Therefore, these experiments have documented the modulation of the infl ammatory response to stress by an LD-associated protein (FSP27) and that this occurs through the interruption of the normal nuclear traffi cking and, thus, activity of a transcription factor (NFAT5), which is important in the cellular response to stress. Interestingly, similar to the current observations with FSP27 and NFAT5, lipin, a protein important in TAG synthesis and LD formation and which is known to interact with several nuclear proteins ( 38 ), has been reported to interact with NFATc4 and to repress infl ammatory genes ( 39 ).
Hyperglycemia results in signifi cant osmotic stress, and postprandial hyperglycemia occurs in many patients even with apparently normal fasting glucose values ( 40 ). In addition, the DNA binding activity of NFAT5 is increased in diabetics with microvascular complications ( 41 ). Moreover, aldose reductase, a key enzyme in the polyol pathway that has been mechanistically linked with hyperglycemiainduced complications ( 42 ), is a target gene of NFAT5 ( 43 ). In addition to TNF-␣ and MCP1, which are direct NFAT5 target genes ( 16,44 ), IL-1, IL-6, and IL-18 contain putative NFAT5 consensus sites in their promoter regions and can be regulated by NFAT5 under hypertonic conditions ( 20 ). Thus, the results of the current experiments provide a potential mechanistic link in dissecting the regulation of infl ammation in obesity and diabetes, where reduced expression of LD-associated proteins Cidea, Cidec/ FSP27, and Plin1 is associated with insulin resistance ( 8 ).