Mice subjected to aP2-Cre mediated ablation of microsomal triglyceride transfer protein are resistant to high fat diet induced obesity

In this study, we asked whether MTP plays a role in adipocyte biology. Our studies showed that MTP is induced early during differentiation of 3T3-L1 cells before the induction of PPAR?, a critical transcription factor for the differentiation of adipocytes. Further, we show for the first time that A-Mttp
^?/?
mice are lean and resistant to diet-induced obesity when fed a high fat diet. The adipose tissue in these mice contained more number of smaller size adipocytes and had less macrophage infiltration. These studies suggest that MTP plays a role in determining size, number, and triglyceride content of adipocyte most likely by assisting in greater accretion of triglyceride and enhancing the size of adipocyte in high fat fed mice.

MTP expression changes during differentiation of 3T3-L1 preadipocytes into adipocytes. Undifferentiated 3T3-L1 preadipocytes had lipid transfer activity suggesting for the presence of pre-existing pools of MTP. Mttp gene was induced early during differentiation that preceded transcriptional regulation of key markers of differentiation such as PPAR? consistent with other studies [21]. Mechanistic studies by other investigators have suggested that the induction of MTP in early differentiation might be due to the binding of C/EBP? and C/EBP? transcription factors to a specific region in the exon 1 of the MTTP gene [21]. Thus, MTP is an early response gene involved in adipocyte differentiation.

Our study shows that early expression of MTP during differentiation is critical for optimal lipid storage. Molecular approaches using siRNA-mediated knockdown of MTP and overexpression of MTP demonstrated that changes in MTP affect intracellular lipid storage and droplet formation (Fig. 2). Overexpression of MTP increases the accretion of triglyceride in differentiated 3T3-L1 cells, whereas knockdown reduces their accumulation. These data are consistent with those of Rakhshandehroo et al. [21] who reported that MTP depletion (stable shRNA knockdown of MTP) does not influence the differentiation of 3T3-L1 preadipocytes, as assessed by FABP4 expression. These data suggest that MTP plays a role in promoting adipogenesis and determines the amounts of lipids accreted by differentiated adipocytes. In contrast to these studies, MTPi did not interfere with triglyceride accumulation. Thus MTP protein, not its lipid transfer activity, might play a role during adipocyte differentiation by assisting in the fusion of small droplets into larger droplets.

Surprisingly, treatment of cells with MTPi did not affect adipocyte differentiation and adipogenesis (Fig. 1). This is consistent with a recent finding demonstrating that the lipid transfer activity of MTP is neither critical for the mobilization of fatty acids from adipocytes nor for cell differentiation as assessed by the number of intracellular lipid droplets [44]. Our findings are also in agreement with another study dealing with the CD1d-mediated lipid self-antigen presentation in adipocytes showing that treatment of adipocytes with BMS-212122 at high dose (13 ?M) for 24 h did not influence the differentiation potential of the preadipocytes [21]. Taken together, our results and observations from others suggest that MTP protein, but not its lipid transfer activity, may play a pivotal role in adipocyte lipid droplet formation and maturation. It is possible that MTP protein might be involved in the fusion of smaller lipid droplets to from larger droplets during adipocyte differentiation.

Understanding the precise function of MTP during adipocyte differentiation is critical in delineating its role in larger lipid droplet formation. MTP might participate in lipid accumulation via other mechanisms. Analysis of the expression pattern of relevant genes controlling adipocyte differentiation and adipogenesis identified that the induction of PPAR? was reduced in A-Mttp
^?/?
mice and the maximum expression of MTP occurs prior to the maximum expression of PPAR? in 3T3-L1 cells. PPAR? is known to activate nearly all of the genes required for the formation of larger droplets, including FABP4 which is required for the transport of free fatty acids, and perilipin which is on the surface of mature lipid droplets. Thus, it is possible that reduced expression of PPAR? might be one mechanism contributing to reduced fat accumulation in MTP deficient adipocytes.

Another possibility is that MTP might interact with lipid droplet associated proteins, such as perilipin, and assist in droplet formation and/or stabilization of larger droplets. Immunohistochemical studies have shown that MTP surrounds small lipid droplets in 3T3-L1 adipocytes as well as in white and brown fat in mice [14, 15]. MTP has been found to physically interact with apoB in lipoprotein producing cells [10, 18]. Therefore, it is possible that MTP may interact with another protein(s) in adipocytes and act as a chaperone. For example, MTP could interact with one or more structural proteins on the surface of lipid droplets such as perilipin, adipophilin, TIP47, OXPAT/MLDP, and S3-12 [45–47], and assist in droplet fusion process by providing close access to triglycerides for smaller droplets, promoting lipid droplet growth and expansion. In fact, Love et al. demonstrated that MTP and perilipin-2 were associated with the same isolated lipid droplets from adipocytes, but they concluded that these proteins do not interact with each other [44]. More studies are needed to explore possible physical interactions between MTP and lipid droplet-coating proteins.

Our finding of lean phenotype associated with reduced fat pad weight and lack of an increase of adipocyte size in HFD fed A-Mttp
^?/?
mice hints at a role of MTP in modulating adipocytes differentiation. Interestingly, reduction of fat mass in A-Mttp
^?/?
mice was accompanied with increased number of smaller adipocytes (Figs. 3 and 4). Thus, it is likely that basal cellular levels of triglycerides are unaffected by MTP. However, increased accumulation of fat during diet-induced obesity may require MTP. This is supported by the observation that A-Mttp
^?/?
mice on chow do not show any difference in body weight compared to controls.

Metabolic defects that alter adipose tissue fat accumulation are frequently associated with changes in glucose homeostasis. Unexpectedly, plasma glucose and insulin levels in the fasted HFD-fed animals did not differ between the A-Mttp
^?/?
and WT mice. Further, despite marked reduction in body weight and adiposity, A-Mttp
^?/?
mice did not exhibit any changes in whole-body glucose homeostasis and insulin sensitivity (Fig. 6). The reasons for no effect on glucose metabolism are not clear.

Using 1-¹⁴C acetate as substrate, we observed that the adipose tissue and liver accounted for about 93 % and 7 % of DNL in WT mice. This in agreement with previous studies suggesting that adipose tissue DNL is a major site of lipogenesis in rodents [48–50]. In A-Mttp
^?/?
mice, we found that DNL was significantly higher in the adipose tissue and was lower in the liver (Fig. 8). This is in good agreement with study showing significant elevation of adipose DNL in ob/ob mice in comparison with control mice [51]. Numerous studies indicated that enhanced adipose DNL, in contrast to the liver DNL, may be beneficial for whole-body metabolism. For example, liver-specific deletion of SCAP, a protein required for the cleavage of SREBP1c to its active form, has been shown to reduce hepatic lipogenesis and enhance adipose DNL by four-fold resulting in improved glucose homeostasis [52]. In another study, ablation of adipocyte/macrophage lipid chaperones aP2 (FABP4) and mal1 (FABP5) increased adipose DNL which protected mice from diet-induced obesity, fatty liver disease and insulin resistance [53]. The ob/ob mice that were rendered LPL deficient in adipose tissue also demonstrated increased DNL and diminished weight and fat mass [54]. Similar to these models, our finding suggest that increased DNL in adipocytes is associated with lean phenotype. The fact that DNL is upregulated while ?-oxidation is decreased indicates for a possible higher rate of intracellular lipolysis and secretion of fatty acid from adipose tissue of lean A-Mttp
^?/?
mice. More experiments are needed to illustrate potential impact of relevant lipases on the release of fatty acids and glycerol by adipose tissues.

Accumulation of adipose tissue macrophages has been well described in obese conditions in mice and humans [37, 55–57]. This process of macrophage infiltration is associated with adipocyte hypertrophy and metabolic dysfunction. In addition to the smaller size of adipocytes, we observed lesser macrophage infiltration in the adipose tissue of A-Mttp
^?/?
mice (Fig. 4). Thus, MTP deficiency in A-Mttp
^?/?
mice protects adipose tissue from HFD-induced macrophage infiltration.

Our studies also provide evidence that MTP deficiency in A-Mttp
^?/?
mice modulates the expression of genes involved in adipocyte differentiation and lipid formation and storage. The expression of PPAR? and its downstream targets FAS, lipins and FABP4 were significantly reduced (55 %-75 %) in the adipose tissue of A-Mttp
^?/?
mice. PPAR? is essential for adipocyte viability and regulates a number of genes involved in lipid uptake and storage [35, 58, 59]. Down regulation of PPAR? could provide protection from HFD induced inflammation and obesity [60–65]. It remains unclear how ablation of MTP in A-Mttp
^?/?
mice impacts transcription of PPAR? and protects mice from HFD induced inflammation and obesity.

Although A-Mttp
^?/?
mice consumed similar amounts of food, they did not accumulate fat in the adipose tissue and the liver compared with WT mice. Further, we did not find differences in rectal temperatures of these mice (Fig. 3). Thus, the physiologic reason for the leaner phenotype in these mice is not clear. It is possible that the A-Mttp
^?/?
mice oxidize fat more efficiently and have higher metabolic rates than the WT mice. Future studies may identify the physiologic reasons for the leaner phenotype in A-Mttp
^?/?
mice when challenged with a high fat diet.

A-Mttp
^?/?
mice have higher plasma triglyceride (Fig. 7). These mice showed modest increase (~28 %) in hepatic MTP activity but mRNA levels did not change (Fig. 3). Thus, changes in hepatic MTP do not explain increases in plasma triglyceride in A-Mttp
^?/?
mice. Further, intestinal lipoprotein assembly and secretion was reduced, not increased. Therefore, increased lipoprotein production is unlikely to be responsible for higher plasma triglyceride in these mice. In addition, postprandial secretion of triglycerides in response to acute fat challenge was significantly decreased in A-Mttp
^?/?
mice after a 4 h oral fat tolerance test. These mice absorbed lesser amounts of fat than control. Further, we observed reduced lipid absorption in lipase inhibited A-Mttp
^?/?
mice (Fig. 9). It is plausible that triglyceride elevations might be secondary to reduced amounts of lipoprotein lipase. This is based on several studies that show that adipocyte lipoprotein lipase is reduced in high fat fed animals. In fact, in recent study [66], mice lacking prostaglandin E receptor subtype 4 (EP4) manifested disrupted lipid metabolism associated with a 69 % reduction in weight gain and fat mass following HFD feeding. Plasma triglycerides in these mice were elevated by 245 %. Authors found that these mice had reduced lipoprotein lipase activity, the key enzyme responsible for trafficking of plasma triglycerides into peripheral tissues.

The data obtained with 3T3-L1 cells and high fat fed A-Mttp
^?/?
mice appear to provide different results. If we assume that 3T3 cell culture studies are comparable to chow fed mice, then results in cells and mice are the same. Both cells and mice do not show any significant difference in fat accumulation during MTP deficiency. We only saw the effect of MTP deficiency when mice were fed a high fat diet. We did not do similar fat challenge studies in cells. Thus, adipose MTP might be more critical when cells are exposed to high lipids.

We are aware that adipocytes express two MTP isoforms, A and B. Dr. Swift’s laboratory was the first to establish the presence of MTP in adipose tissue as well as in the pre-adipocyte cell line, 3T3-L1 [14]. The group has previously generated anti-MTP antibodies from a 19-amino acid peptide representing residues 843 through 861 in the MTP protein. In a recent report [44], they were able to monitor mRNA and protein levels of both the MTP isoforms in adipocytes. They showed that adipocytes predominantly (95 %) express MTP B isoform and demonstrated that MTP protein expression increased linearly over the 8-day period of differentiation, but mRNA levels for MTP-A and MTP-B did not change significantly. In another study, MTP-B mRNA levels increased significantly within the first two days of differentiation and decreased slightly afterwards [21]. This later report is in good agreement with our observation for the decline of MTP expression at late stages of differentiation. Therefore, it is likely that MTP-B is the major isoform in adipocyte. Further, it should be pointed that no functional differences between these two isoforms have been reported. Therefore, expressions of these two different isoforms as determined by mRNA or protein levels do not appear to have any functional consequences.

In this study, we used aP2-Cre to ablate the Mttp gene. In these mice, MTP expression was greatly diminished in isolated adipocytes (80 %) and decreased by?~?50 % in adipose tissue which also includes non-adipocyte cell types. We are aware that aP2-Cre is expressed in adipocytes and macrophage. Therefore, the observed effects might be due to the deletion of MTP in both the adipose tissue and macrophages. In our study, we found significant reductions in the adipose tissue MTP but not in the spleen, a tissue rich in macrophages. Thus, it is likely that the observed effects are due to MTP deficiency in the adipose tissues. It is unlikely that reduction in T-cell mediated immune response due to possible deletion of MTP in macrophages could account for the lean phenotype and metabolic changes observed in A-Mttp
^?/?
mice for the following reasons: (1) iNKT cell population and CD1d expression is reduced in the adipose tissue of obese mice and humans compared to those of lean subjects [67]; (2) iNKT cell-specific J?18 knockout mice are obese and exhibit increased adipose tissue inflammation at the early stage of obesity [68]; (3) CD1d-deficient mice fed a high-fat diet show insulin resistance and hepatic steatosis [69–71]. Future studies may dissect the role of MTP in the adipose tissue and macrophages using cell-specific deletion of the gene.