Resistin is a key mediator of glucose-dependent insulinotropic polypeptide (GIP) stimulation of lipoprotein lipase (LPL) activity in adipocytes

Studies on the physiological roles of the incretin hormone, glucose-dependent insulinotropic polypeptide (GIP) have largely focused on its insulinotropic action and ability to regulate β -cell mass. In previous studies on the stimulatory effect of GIP on adipocyte lipoprotein lipase (LPL), a pathway was identified involving increased phosphorylation of protein kinase B (PKB) and reduced phosphorylation of LKB1 and AMP-activated protein kinase (AMPK). The slow time of onset of the responses suggested that GIP have induced release of an intermediary molecule, and the current studies focused on the possible contribution of the adipokine resistin. In differentiated 3T3-L1 adipocytes, GIP, in the presence of insulin, increased resistin secretion through a pathway involving p38 mitogen activated protein kinase (p38 MAPK) and the stress-activated protein kinase/Jun-amino-terminal kinase (SAPK/JNK). The other major incretin hormone, glucagon-like peptide-1 (GLP-1), exhibited no significant effects. Chronic elevation of circulating GIP levels in the Vancouver Diabetic Fatty (VDF) Zucker rat resulted in increases in circulating resistin levels and activation of p38 MAPK or SAPK/JNK in epididymal fat tissue, suggesting the existence of identical pathways in vivo as well as in vitro . Knockdown resistin RNA


Abstract
Studies on the physiological roles of the incretin hormone, glucose-dependent insulinotropic polypeptide (GIP) have largely focused on its insulinotropic action and ability to regulate β-cell mass. In previous studies on the stimulatory effect of GIP on adipocyte lipoprotein lipase (LPL), a pathway was identified involving increased phosphorylation of protein kinase B (PKB) and reduced phosphorylation of LKB1 and AMPactivated protein kinase (AMPK). The slow time of onset of the responses suggested that GIP have induced release of an intermediary molecule, and the current studies focused on the possible contribution of the adipokine resistin.
In differentiated 3T3-L1 adipocytes, GIP, in the presence of insulin, increased resistin secretion through a pathway involving p38 mitogen activated protein kinase (p38 MAPK) and the stressactivated protein kinase/Jun-aminoterminal kinase (SAPK/JNK). The other major incretin hormone, glucagon-like peptide-1 (GLP-1), exhibited no significant effects. Chronic elevation of circulating GIP levels in the Vancouver Diabetic Fatty (VDF) Zucker rat resulted in increases in circulating resistin levels and activation of p38 MAPK or SAPK/JNK in epididymal fat tissue, suggesting the existence of identical pathways in vivo as well as in vitro. Administration of resistin to 3T3-L1 adipocytes mimicked the effects of GIP on the PKB/LKB1/AMPK/LPL pathway: increasing phosphorylation of PKB, reducing levels of phosphorylated LKB1 and AMPK, and increasing LPL activity. Knockdown of resistin using RNA interference (RNAi) attenuated the effect of GIP on
There is strong evidence that GIP plays an important regulatory role in fat metabolism (7,8,15). During a meal, GIP is secreted in response to long chain fatty acids (FAs) released from triglycerides (TGs) (16,17). GIP promotes the clearance of chylomicronassociated TG from blood (18) and infusion of GIP lowers rat plasma TG responses to intraduodenal fat (19). A direct adipogenic role was suggested by the demonstrations of GIP enhancement of adipose tissue FA synthesis from acetate (20) and its incorporation into triglyceride (21). The physiological importance of these effects was emphasized by the finding that GIP receptor (GIPR) knockout mice demonstrated reduced adipose tissue accretion on feeding a high fat diet (22).
In an earlier study we demonstrated that GIP increased lipoprotein lipase (LPL) activity and triglyceride (TG) accumulation in differentiated 3T3-L1 cells and human subcutaneous adipocytes through a pathway involving increased phosphorylation of protein kinase B (PKB) and decreases in LKB1 and AMP-activated protein kinase (AMPK) phosphorylation (23). An anomaly of these studies was the slow time of onset of the responses, suggesting that GIP may have acted through release of an intermediary molecule. Recently, Hansotia and coworkers administered the long-acting GIPR agonist, [D-Ala 2 ]GIP, to mice and observed an increase in plasma resistin that was absent in GIPR null mice and was not mimicked by the long-acting GLP-1R agonist, exendin-4 (24). We have now examined the possibility that specific actions of GIP are mediated through the direct regulation of resistin release at the level of the adipocyte, resulting in activation of PKB and increased LPL activity.

Resistin
ELISA (Enzyme-Linked ImmunoSorbent Assay)-For studies on the effect of GIP or GLP-1 on the secretion of resistin, INS-1 β-cells or 3T3-L1 adipocytes were treated with GIP or GLP-1 as indicated in the figure legend. The levels of resistin in the media were determined using a Resistin ELISA kit (Cayman Chemical, Ann Arbor, MI).

Quantitative
Real-time RT-PCR (Reverse Transcriptase-Polymerase Chain Reaction)-Total RNA was extracted and cDNAs generated by reverse transcription. 100 ng cDNA were used in the real-time polymerase chain reaction (PCR) to measure resistin expression, whereas 10 ng cDNA were used for the β-actin internal control . The  primer and probe sequences used for the  amplification of resistin cDNA are as  follows: forward primer, 5'-AAC CTT  TCA TTT CCC CTC CTT TTC-3';  reverse primer, 5'-GGA AGC GAC  CTG CAG CTT ACA GCA-3'; probe, 5'FAM-AGT CTC CTC CAG AGG GAA GTT GG-3'TAMRA. All reactions followed the typical sigmoidal reaction profile, and cycle threshold was used as a measurement of amplicon abundance.

Immunohistochemistry-
Paraffin sections containing epididymal fat tissue were prepared from GIP-treated and control rats and subjected to immunostaining for resistin. The sections were visualized with Alexa fluor ® 488 conjugated anti-mouse secondary antibody, and imaged using a Zeiss laser scanning confocal microscope (Axioskop2).
Statistical Analysis-Data are expressed as means ± Standard Errors of the Mean (SEM) with the number of individual experiments presented in the figure legend. Data were analyzed using the non-linear regression analysis program PRISM (GraphPad, San Diego, CA) and significance was tested using analysis of variance (ANOVA) with Newman-Keuls hoc test (P < 0.05) as indicated in figure legends.

GIP, but not GLP-1 increases the secretion of resistin by 3T3-L1
adipocytes. The effect of GIP and GLP-1 on the secretion of resistin in vitro was first examined in differentiated 3T3-L1 adipocytes. Incubation of 3T3-L1 adipocytes with GIP (100 nM) in the presence of insulin (1 nM) for 2 h resulted in ∼ 2.3-fold increases in resistin secretion, compared to control cells (Fig.  1A). Concentration dependent effects of GIP on resistin secretion were observed with EC 50 values of 9.6 ± 0.2 nM (Fig.  1B). In contrast, treatment of 3T3-L1 adipocytes with GLP-1 (100 nM) had no effect on resistin secretion. Additionally, although resistin was detectable in extracts from INS-1 (832/13) β-cells, neither GIP nor GLP-1 affected secretion ( Fig. 1A and B).  Figure 2) with GIP or GLP-1 (100 nM) was not found to alter resistin protein levels. One possible explanation for the lack of change in cell protein levels is that increases in expression only maintain sufficient newly synthesized resistin to support the increased secretion rate, without intracellular accumulation. To investigate this possibility, Retn mRNA levels were determined in 3T3-L1 adipocytes under basal and stimulated conditions. As shown in Figure 1C, GIP in the presence of insulin, increased Retn mRNA levels whereas GLP-1 had no effect. It is therefore likely that GIPpotentiated resistin secretion resulted from increased expression.
Signaling modules potentially involved in GIP-mediated resistin secretion were next studied. Because resistin secretion was evident within 1 h following GIP treatment, we first examined faster-acting MAPK signaling modules, including p42/44 MAPK, p38 MAPK and SAPK/ JNK. There was no significant effect of GIP or GLP-1 on the levels of phosphorylated p42/44 MAPK (Thr202/Tyr204) in 3T3-L1 adipocytes during the time-course of the study (Figure 2A and 2B). On the other hand, GIP, in the presence of insulin, was found to increase phosphorylation of p38MAPK (Thr180/Tyr182) transiently, peaking at 5-10 minutes ( Figures 2C and 2G), and SAPK/JNK (Thr183/Tyr185) in a more sustained manner ( Figures 2E and 2H Figure 3). These results indicate that synergistic action between insulin and GIP is required for phosphorylation.
p38 MAPK and SAPK/JNK are involved in GIP-mediated resistin secretion. In studies designed to examine the involvement of p38 MAPK and SAPK/JNK in GIP-mediated resistin secretion, 3T3-L1 adipocytes were incubated in the presence or absence of selective inhibitors of MAPK pathways. There was no significant effect on GIPmediated resistin secretion by treatment with the MAP Kinase Kinase (MEK) 1/2 inhibitors, PD098059 (75 µM) or U0126 (25 µM). However, resistin secretion in response to GIP, was greatly reduced by the p38 MAPK inhibitor, SB203580, and the SAPK/JNK inhibitor, SP600125 (50 µM) ( Figure 3A). There were no significant differences in cellular resistin levels following treatment with these inhibitors ( Figure 3B). These results strongly suggest that p38 MAPK and SAPK/JNK, but not ERK1/2, are involved in GIP-mediated resistin secretion in 3T3-L1 adipocytes and that changes in protein expression are not involved.

The effect of GIP on epididymal fat in vivo.
To determine whether GIP can regulate resistin secretion in vivo, peptide was systemically administrated to lean (Fa/?) or obese (fa/fa) VDF Zucker rats, using osmotic minipumps and conditions previously shown to increase adipose tissue LPL levels (23).
Resistin was shown to be expressed in epididymal fat tissue of lean and obese VDF Zucker rats ( Figure 4A), and 2 weeks continuous infusion of low concentrations of GIP (10 pmol/kg·min) resulted in increased circulating resistin levels ( Figure 4B), but no detectable change in resistin protein levels in epididymal fat tissue ( Figure 4C). However, GIP infusion did increase tissue levels of phosphorylated p38MAPK (Thr180/Tyr182) and SAPK/JNK (Thr183/Tyr185) in both lean and obese animals, when compared to controls ( Figures 4D and 4E). These results correlate well with the in vitro results in 3T3-L1 adipocytes, and support a role for GIP in the regulation of resistin secretion in vivo, acting through a p38MAPK and JNK/SAPK signaling module.

Resistin
modulates PKB/LKB1/AMPK phosphorylation and LPL activity. Next, we examined the functional implications of GIPmediated resistin secretion. In the earlier study, we demonstrated that GIP increased LPL activity and TG accumulation through a pathway involving increased phosphorylation of PKB and reductions in phosphorylated LKB1 and AMPK (23). Responses to GIP under these conditions were relatively slow and the possible involvement of an intermediary could not be excluded. We therefore tested whether resistin could act as a mediator of GIP's action on adipocytes. Treatment of 3T3-L1 adipocytes with resistin (10 nM) mimicked the effects of GIP, resulting in increased levels of phosphorylated PKB (Ser473), and decreased levels of LKB1 (Ser428) and AMPK (Thr172) (Figures 5A, 5B and 5C). In control experiments, there were no significant changes in levels of phosphorylated PKB (Ser473), LKB1 (Ser428) or AMPK (Thr172) during the time course of the study, confirming that these effects were GIP-dependent (Supplementary Figure 4). Treatment of 3T3-L1 adipocytes with resistin (10 nM) for 12 h resulted in ∼3.4-fold increase in LPL activity, compared to control ( Figure  5D) and concentrationdependent effects of resistin on LPL activity were observed with EC 50 values of 15.8 ± 0.2 nM (Fig. 5E). Taken together, these results suggest that resistin could be an important mediator of GIP stimulation of LPL activity.  Figures  6A and 6B, 3T3-L1 adipocytes treated with 100 nM of a pool of 3 resistin siRNAs resulted in greatly reduced resistin protein levels and an approximate 65% and 75% reduction in basal-and GIP-stimulated resistin secretion, respectively. The reduction in GIP-mediated responses was associated with attenuation of the effects of GIP on phosphorylation of PKB, LKB1 and AMPK ( Figure 6C, 6D and 6E). These results complement the finding that resistin treatment resulted in decreased levels of phosphorylated LKB1 and AMPK ( Figure 5B and 5C). Furthermore, GIP-mediated LPL activation was also greatly reduced in resistin RNAi-treated cells ( Figure 6F), supporting a role for resistin as a mediator of GIP-mediated increases in LPL activity.

Resistin knockdown greatly reduces
Basal phospho-LKB1 and AMPK levels appeared to be slightly increased in RNAi-treated cells in the absence of GIP ( Figure 6D and E). Basal phospho-PKB levels were undetectable. However resistin may normally maintain a pool of PKB in the phosphorylated state, thus suppressing levels of phospho-LKB1 and AMPK, with loss of resistin removing this constraint. Since GIPtreated cells showed both a reduction in phospho-LKB1/AMPK and a small increase in LPL, there is likely an alternative pathway by which it can act.

Discussion
Since its identification as an adipocyte-secreted protein (25,26), there has been controversy over the physiological roles played by resistin (Adipocyte Secreted Factor; ADSF). Since resistin administration to C57BL/6J mice resulted in glucose intolerance and reduced insulin sensitivity, while anti-resistin serum improved insulin resistance, it was suggested that resistin was the long sought after link between obesity and type 2 diabetes (25). Resistin levels were found to be markedly elevated in serum from insulin resistant mice placed on a high-fat diet and in obese rats and mice (25). Subsequently a number of groups reported a reduction in resistin expression in white adipose tissue from a number of obese rodent models including ob/ob, db/db, tub/tub and KKA (y) mice (27)(28)(29). Despite this, elevated serum resistin levels have found in high fat fed or monogenic obese mice (28), similar to the elevated levels observed in the obese VDF Zucker rats.
A number of studies have shown that resistin can act in an autocrine/paracrine manner in white adipocytes, although its actual role is obscure. Resistin mRNA is undetectable in 3T3-L1 cell preadipocytes, but its expression increases during differentiation to the adipocyte phenotype (30). Addition of recombinant resistin, or increasing resistin expression via gene transfection promoted differentiation of preadipocytes to adipocytes. Conflicting results have been reported for resistin effects on mature adipocytes. Fat specific overexpression of resistin in a spontaneously hypertensive rat model was found to impair reesterification of FFAs to triglycerides (31), whereas treatment of mouse adipose explants with recombinant FIZZ3, the human resistin equivalent, resulted in increases in both lipolysis and reesterification (32). Resistin can also inhibit insulin-induced glucose uptake in 3T3-L1 adipocytes (25). Resistin is additionally expressed in both human and mouse pancreatic islets (33), and exogenous resistin impaired glucose-stimulated insulin secretion in isolated islets (34), although its physiological islet function is unknown.
We showed previously that GIP, in the presence of insulin, increased LPL enzyme activity and triglyceride accumulation, through a pathway involving increased phosphorylation of PKB and reductions in the phosphorylated forms of LKB1 and AMPK (23). Unlike GIP action on PKB in INS-1 beta cells (5), responses in 3T3-L1 adipocytes demonstrated a slow time of onset, suggesting responses involved release of an intermediary molecule that acted upstream of PI3-K/PKB. The present results suggest that resistin is a major contributor to the lipogenic effects of GIP in 3T3-L1 adipocytes since it induced identical changes in PKB, LKB1 and AMPK ( Figure 5). Knockdown of resistin using RNA interference resulted in attenuation of GIP's effects ( Figure 6). The incomplete abolition of GIP's effects is likely due to the involvement of other growth factors and/or hormones secreted by adipocytes.
The current studies therefore provide a plausible explanation for the delayed responses to GIP: increased secretion and actions of resistin. There are several ways by which GIP could have produced an increase in resistin secretion. It is unlikely to be through a major effect on resistin biosynthesis, since neither resistin levels in 3T3-L1 adipocytes nor those in INS-1 (832/13) β-cells, were influenced by GIP treatment (  (39,42). A fraction of adipocyte leptin appears to be localized in small vesicles (39,42,43). It is therefore possible that resistin also resides in intracellular vesicles and that GIP stimulates their production and secretion, as proposed for leptin (44). In the present study, although GIP was shown to stimulate resistin secretion, there was no detectable effect on the levels of resistin protein in either acute experiments on 3T3-L1 adipocytes or in vivo in VDF rats. A potential explanation for this observation is that GIP increases resistin expression so as to maintain sufficient newly synthesized protein to support the increased secretion rate, without intracellular accumulation. This possibility was supported by the finding that GIP treatment resulted in increased Retn mRNA expression ( Figure 1C). In studies designed to identify signaling pathways responsible for GIP-induced resistin secretion, we identified p38 MAPK and SAPK/JNK as potential targets (Figures 2G and 2H). Interestingly, increased phosphorylation of p38 MAPK was rapid, but transient, whereas increases in SAPK/Jnk did not reach peak levels until 1 h. This may explain why GIP-induced resistin secretion is maximal at 1-2 h following initiation of stimulation. As discussed previously (23), low concentrations of insulin were included in all the studies since it is essential for the stimulatory effect of GIP on LPL activity and triglyceride synthesis. Under these conditions the ability of GIP to stimulate adenylyl cyclase and activate protein kinase A is blocked. It is therefore unlikely that cyclic AMP-mediated pathways contributed to the stimulation of secretion. A pharmacological approach was taken to confirm the kinase pathways involved. Application of MEK inhibitors did not influence resistin secretion, whereas inhibition of either SAPK/JNK or p38 MAP kinase greatly reduced secretion. Exactly where in the secretory pathway these kinases act is currently unknown. An additional uncertainty is the exact mechanism by which resistin acts in the adipocyte. As previously discussed (23), application of the pharmacological inhibitors of PI3-K, LY294002 and wortmannin, resulted in reduced phosphorylated LKB1/AMPK strongly suggesting that PKB at least partially mediated these effects of GIP (23). The responses to resistin shown in the current studies did not reveal a definitive temporal relationship between PKB phosphorylation and reduced LKB1/AMPK, possibly due to close intracellular interaction between these enzymes. It has been shown that resistin increases phosphorylation of ERK 1/2 in cultured human endothelial cells (45) and aortic smooth muscle (46), whereas resistin's ability to impair insulinreceptor phosphorylation has been attributed to increasing gene expression of Suppressor of Cytokine Signaling 3 (SOCS3) (47). Resistin also induced a transient increase in phospho-PKB in endothelial cells (48), although studies on resistin null ob/ob mice indicated that resistin normally suppressed PKB phosphorylation in skeletal muscle and liver. Importantly, as found in 3T3-L1 adipocytes, resistin treatment reduced AMPK phosphorylation in both the liver (49,50) and skeletal muscle (51)         L e a n L e a n + G IP F a tty F a tty + G IP L e a n L e a n + G IP F a tty F a tty + G IP L e a n L e a n + G IP F a tty F a tty + G IP L e a n L e a n + G IP F a tty F a tty + G IP