Interesterified Fats Induce Deleterious Effects on Adipose Tissue and Liver in LDLr-KO Mice

Interesterified fats are being widely used by the food industry in an attempt to replace trans fatty acids. The effect of interesterified fats containing palmitic or stearic acids on lipid metabolism and inflammatory signaling pathways in adipose and hepatic tissues was evaluated. Male LDLr-KO mice were fed a high-fat diet containing polyunsaturated (PUFA), palmitic (PALM), palmitic interesterified (PALM INTER), stearic (STEAR), or stearic interesterified (STEAR INTER) fats for 16 weeks. The expression of genes and protein levels involved in lipid metabolism and inflammatory processes in liver and white adipose tissue was determined by quantitative RT-PCR and by Western blot, respectively. The infiltration of inflammatory cells in hepatic and adipose tissues was determined by eosin and hematoxylin, while liver collagen content was determined by Sirius Red staining. Both interesterified fats increased liver collagen content and JNK phosphorylation. Additionally, the STEAR INTER group developed nonalcoholic steatohepatitis (NASH) associated with higher neutrophil infiltration. PALM INTER induced adipose tissue expansion and enlargement of adipocytes. Furthermore, PALM INTER triggered increased IKK phosphorylation and TNFα protein content, conditions associated with the upstream activation of the NFkB signaling pathway. STEAR INTER induced NASH, while PALM INTER triggered hepatic fibrosis and adipocyte hypertrophy with inflammatory response in LDLr-KO mice.


Introduction
The adipose tissue (AT) is a central metabolic organ involved in the regulation of whole-body energy homeostasis [1]. The expansion of AT, notably the visceral fat depot, relates to obesity-associated

Mouse Model and Diet
Male LDLr-KO mice (C57BL/6J background) were purchased from Jackson Laboratory (Bar Harbor, USA) and housed in a conventional animal facility (12 h light/dark cycle) in a temperature-controlled environment and ad libitum food and water. For each group, 18 animals were enrolled which were recognized as appropriate models to evaluate the effects of different diets on metabolic syndrome, fatty liver disease, obesity, and diabetes [32][33][34], since they develop similar lipid profiles when compared to humans after a high-fat diet [35]. Furthermore, this animal model presents the absorption of fatty acids at the sn-2 position similar to that observed for humans [32,33]. All the animal protocols were approved by the Ethics Committee of the Faculty of Medical Sciences of the University of São Paulo (CEP 026/12) and were conducted in agreement with NIH guidelines for the care and use of laboratory animals. After weaning, mice were randomly divided into five groups and fed on high-fat diets (40% of energy as fat) enriched with polyunsaturated (PUFA), palmitic (PALM), palmitic interesterified (PALM INTER), stearic (STEAR), or stearic interesterified (STEAR INTER) fats for 16 weeks. The composition of the diets was compliant with the recommendations of the AIN [36] without cholesterol addition and had similar amounts of fiber. The fatty acid composition of the fats used in the diets was determined by gas liquid chromatography analyses (GLC) and the position of the fatty acids on the glycerol backbone was shown by nuclear magnetic resonance (NMR) employing 13 C [37,38] (Table 1). All fats were provided by Cargill ® (São Paulo, SP, Brazil). Mice were weighed weekly and food intake was estimated as the difference between the food offered and the residual food in the cages. After a 16-week period, mice were fasted overnight for a period of 9 to 12 h and euthanasia was performed using a CO 2 chamber with a gradual fill method (displacement rate from 10% to 30% of the chamber volume/min). Blood was drawn from the abdominal aorta into 0.1% EDTA tubes. After perfusion with phosphate-buffered saline (PBS), liver and epididymal and subcutaneous fat pads were weighed, frozen in liquid nitrogen, and stored at −80 • C.

Liver Lipid Concentrations
Measurement of liver cholesterol and TG was performed as previously described [39]. Briefly, ∼200 mg of tissue was homogenized and lipids were extracted in a chloroform:methanol (2:1) solution (6 mL). The organic phase was separated from aqueous phase by the addition of 0.05% H 2 SO 4 . Chloroform containing Triton X-100 (0.5% Triton X-100, v/v) was then added to the lower phase and dried under nitrogen. The tubes were rinsed with chloroform and dried again. Deionized water (1 mL) was added and the samples were incubated at 37 • C for 15 min under shaking. After this period, the samples were vortexed and subsequently used for the enzymatic determination of lipid content.

Histological Analysis
Hepatic tissue and epididymal adipose tissue fragments were fixed in 4% formaldehyde, dehydrated with ethanol, cleared with xylene, and embedded in paraffin. After inclusion, 4 µm sections were obtained and stained with hematoxylin and eosin or Picrosirius red. Photomicrographs were obtained by digital imaging system (Leica Qwin, Leica System, Cambridge, CB5 8PB, UK). The mean area of adipocytes (average surface area of 70-100 randomly sorted adipocytes per animal) was determined and quantified in Image J (Media Cybernetics, MD, USA).

Grading of the NAFLD Activity
Histological variables were blindly scored by one of the authors using a scoring system adapted from Kleiner et al. [40]. This system numerically rates macrosteatosis (0-3), lobular inflammatory changes (0-3), and hepatocyte ballooning (0-2) and fibrosis (0-2). Score of 5 or more was characterized as NASH; values between 3 and 4 were allocated in the categories of absence, presence, or probable NASH; and values below 3 indicated absence of NASH. The collagen volume (%) was expressed as the percentage of area of the centrilobular vein (CLV) and portal triad using a digital image system (Qwin image analysis software; Leica Imaging Systems, Cambridge, CB5 8PB, UK).

Neutrophil Infiltration
HE-stained slides were evaluated under a light microscope at 400× magnification. Neutrophils were identified by nuclear and cytoplasmic appearance and counted in 20 fields of the hepatic parenchyma (3.4 mm 2 /blade) [41]. Neutrophils of central or periportal regions as well as the present inside granulomas were not considered. The inflammatory infiltrate of neutrophils in the liver parenchyma is presented as number of cells/mm 2 .

Immunofluorescence Analyses of M1-like Macrophages
The localization of integrin-αM was detected using anti-integrin-αM antibody (CD11b + , rabbit polyclonal, Santa Cruz Biotechnology, code sc28664; 1:100) by indirect immunofluorescence staining, as previously described [42]. In short, sections were fixed in acetone and blocked with 2% rabbit serum for 30 min at room temperature, followed by incubation with primary antibody anti-integrin-αM (1:200) for 12 h, at 4 • C, in a moisture chamber. Sections incubated with the anti-F4/80 antibody were then incubated with an FITC-conjugated anti-rabbit IgG secondary antibody. FITC-conjugated mice anti-rabbit antibody was from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The sections were visualized under a fluorescence microscope (Leica FW 4500 B).

Statistical Analyses
Analyses were carried out on GraphPad Prism software (GraphPad Software Inc., San Diego, CA, USA). All data were checked for normality prior to statistical analysis. Comparisons were made utilizing one-way ANOVA followed by the post hoc test Newman-Keuls Multiple Comparison Test for parametric and Kruskal-Wallis for nonparametric comparisons. A value of p < 0.05 was considered statistically significant. The post hoc test was only applied after a difference between groups.

Results
To investigate the impact of interesterified fats on adipose tissue and hepatic lipid metabolism, weaning LDLr-KO mice were randomly distributed into five groups receiving high-fat diets (40% of Nutrients 2019, 11, 466 6 of 16 energy as fat). These animals develop obesity and comorbidities to a similar extent as humans submitted to a high-fat diet [32,33].
All animals presented the same body weight at baseline (p = 0.76) and dietary intake (p = 0.15) did not differ among the groups. After 16 weeks, PALM INTER presented increased body weight as compared to STEAR and STEAR INTER (p < 0.01), but not when compared to PUFA and PALM groups. Concurrently, PALM INTER had greater visceral (p < 0.001) and subcutaneous (p < 0.001) fat contents and epididymal adipocyte diameter (p < 0.001) (Figure 1(I)) when compared to the other groups (Table 2).
for parametric and Kruskal-Wallis for nonparametric comparisons. A value of p < 0.05 was considered statistically significant. The post hoc test was only applied after a difference between groups.

Results
To investigate the impact of interesterified fats on adipose tissue and hepatic lipid metabolism, weaning LDLr-KO mice were randomly distributed into five groups receiving high-fat diets (40% of energy as fat). These animals develop obesity and comorbidities to a similar extent as humans submitted to a high-fat diet [32,33].
All animals presented the same body weight at baseline (p = 0.76) and dietary intake (p = 0.15) did not differ among the groups. After 16 weeks, PALM INTER presented increased body weight as compared to STEAR and STEAR INTER (p < 0.01), but not when compared to PUFA and PALM groups. Concurrently, PALM INTER had greater visceral (p < 0.001) and subcutaneous (p < 0.001) fat contents and epididymal adipocyte diameter (p < 0.001) ( Figure 1I) when compared to the other groups (Table 2). . Western blot analysis from Both stearic groups (STEAR and STEAR INTER groups) was not performed (A-D) because of the small amount of adipose tissue obtained in these groups. Data were checked for normality prior to statistical analysis. One-way ANOVA was performed, followed by the post hoc Newman-Keuls test. Statistically significant difference is presented by different letters (p < 0.05) with a > b > c values. Data are shown as mean ± SEM (n = 6).  . Western blot analysis from Both stearic groups (STEAR and STEAR INTER groups) was not performed (A-D) because of the small amount of adipose tissue obtained in these groups. Data were checked for normality prior to statistical analysis. One-way ANOVA was performed, followed by the post hoc Newman-Keuls test. Statistically significant difference is presented by different letters (p < 0.05) with a > b > c values. Data are shown as mean ± SEM (n = 6). Glucose and insulin concentrations were not different between PALM and PALM INTER groups nor between STEAR and STEAR INTER groups, meaning that the interesterification process did not alter these parameters (Table 3). However, glucose and insulin concentrations were significantly higher for PALM and PALM INTER (p < 0.05) as compared to the other treatments. PALM INTER induces inflammation. The adipocyte hypertrophy observed in the PALM INTER group led us to investigate the expression of genes related to lipolysis and inflammatory response. PALM INTER-fed mice had reduced expression of ATGL (Pnpla2) as compared to PALM, but no differences were observed regarding other lipid droplet-associated protein, such as perilipin (Plin, p = 0.34) and hormone-sensitive lipase (Lipe, p = 0.61) ( Table 4). A higher Scd1 expression was observed in STEAR-and STEAR INTER-fed mice as compared to PUFA (p < 0.05), since stearic acid is the main substrate for this enzyme. There were also no differences in the expression of genes involved in the inflammatory response, such as adiponectin (Acrp30, p = 0.07) and TNFα (Tnf, p = 0.65). Although we did not find any difference in cytokines gene expression among the groups, PALM INTER-fed mice presented increased TNFα (p < 0.001) and pIKK (p < 0.05) protein levels ( Figure 1(IIA,C), respectively). However, IL1β and pJNK levels were not different among groups, which suggests that PALM INTER induced TNFα protein concentration in adipose tissue through the activation of the upstream NFkB inflammatory signaling pathway.  (Table 3). STEAR  INTER-fed mice developed NASH-like lesions characterized by intermediate steatosis (grade 2), inflammatory infiltrate (grade 1), and fibrosis in both centrilobular and portal triad veins (grade 2), whereas PALM INTER-and PALM-fed mice had mild intermediate steatosis (grade 2) with minimal inflammatory process (grade 1) and fibrosis (grade 1) (Figure 2(I)). Hepatocellular ballooning was not observed in all samples and was therefore not scored. STEAR INTER induced more neutrophil infiltration (p < 0.001) compared to the others and higher collagen content in the centrilobular vein compared to STEAR and PUFA (p < 0.05). On the other hand, the PALM INTER group presented higher collagen content in the centrilobular (p < 0.05) and portal triad (p < 0.05) veins as compared to PALM and PUFA groups (Table 3; Figure 2(I)).  Interesterified fats induce JNK activity and PALM INTER triggers apoptosis. In order to investigate the molecular mechanisms underlying NASH development and the presence of fibrosis, the expression of genes related to hepatic lipid metabolism and inflammatory response was determined. The expression of Scd1 was threefold and twofold greater in PALM INTER-fed mice as compared to PUFA-and PALM-fed mice (p < 0.05), respectively (Table 5). Although STEAR INTER induced a twofold expression of Scd1 as compared to STEAR and PUFA, the difference was not significant (p = 0.11). There was also no difference in the expression of genes related to fatty acid oxidation (Ppar-a and Cpt-1), inflammation (Il-1b, Il-6, and Tnf ), and lipogenesis (Srebp-1c, Pparg and Mttp). Both interesterified fats increased JNK phosphorylation (p < 0.001; Figure 2(IIE)) and presented lower TNF-α (Figure 2(IIA)) and IL-1β (p < 0.001; Figure 2(IIB)) level compared to other fats. In addition, PALM INTER-and STEAR INTER-fed mice had less infiltration of M1-like macrophages in the liver, as shown by CD11b staining, which corroborates the immunoblotting results (Figure 2(III)). Regarding the apoptotic signaling, both STEAR and STEAR INTER showed higher caspase 3 protein levels (p < 0.001) than other experimental diets. Interestingly, PALM INTER presented a higher caspase-3 expression as compared to PALM (p < 0.01; Figure 2(IIG)). Corroborating this data, the antiapoptotic protein Bcl2 was reduced in PALM INTER (p < 0.01) as compared to PALM and PUFA (Figure 2(IIH)).

Discussion
The strength of our investigation was that the interesterification process and the enrichment of palmitic acid on the sn-2 position of triglycerides elicits harmful effects on hepatic and adipose tissues in LDLr-KO mice, triggering the onset of obesity.
Since glucose and insulin plasma concentrations are also influenced by dietary fats [44], we hypothesized that the change of the position of fatty acids in the glycerol backbone could also interfere in these parameters. We observed an increase in glucose and insulin concentrations in PALM and PALM INTER as compared to the other groups. Accordingly, long-term exposure to palmitate culminates in beta cell death [45] and impaired glucose-induced generation of the second messenger cAMP, an important determinant of insulin secretion [46]. Furthermore, chronic exposure to palmitate promotes ceramide synthesis [45], which contributes to Akt inhibition and impairment in insulin sensitivity [47]. Although in our investigation SAFA impaired glucose and insulin concentration, no differences were observed between PALM and PALM INTER or between STEAR and STEAR INTER. Therefore, the set of these initial results shows that the interesterification did not alter biochemical parameters. Additionally, our previous study [48] showed that the interesterification process did not influence total plasma cholesterol, although it promoted cholesterol accumulation in LDL particles.
To the best of our knowledge, there are no investigations regarding the influence of interesterified fats on adipose tissue metabolism, as well as in the adipocytes inflammatory signaling pathways. Obesity is characterized by adipose tissue expansion with large adipocytes [49], proinflammatory cell infiltration, and subsequent secretion of inflammatory biomarkers [50]. Altogether, these conditions contribute to an impairment in insulin sensitivity [51] and higher lipolysis [52] in adipose tissue. An important finding of our investigation was that PALM INTER induced greater adipose tissue expansion and adipocytes hypertrophy. Therefore, our next goal was to elucidate the underlying molecular mechanisms involved in adipocyte hypertrophy. It has already been demonstrated in 3T3-L1 cells that palmitic acid increases the expression and secretion of inflammatory cytokines, such as IL-6 and TNF-α, as well as NF-κB activity in adipocytes [53]. When comparing the effects of different fatty acids on the expression of inflammatory cytokines, no differences were observed among the groups. In the case of the PUFA group, the inflammation was observed presumably because of the high concentration of omega-6 (18:2n-6), as also shown by others [54]. Arachidonic acid, the main product of 18:2n-6, is a substrate for eicosanoid synthesis which is enrolled in the control of intensity and duration of inflammatory responses [55]. However, PALM INTER presented higher TNF-α protein levels in comparison to PALM and PUFA, and one possible explanation is the higher incorporation of palmitic acid in adipocyte phospholipids, as observed in a previous study conducted in rats [29]. The adipocyte hypertrophy and TNF-α secretion are the most relevant contributors to the development of insulin resistance in adipose tissue [47]. TNFα stimulates lipolysis through attenuation of the insulin antilipolytic activity [6], increasing the release of free fatty acids, which are the major determinants of hepatic insulin resistance (IR) and lipogenesis, which culminates in NAFLD [49,50,52]. The hypertrophy of adipose tissue promotes the release of inflammatory cytokines leading to the overexpression of MCP-1, which supports the recruitment of macrophages to the adipose tissue [49]. Moreover, macrophages within expanded adipose tissue can secrete TNFα through the IKKβ-NFκB or JNK-AP1 (Jun N-terminal kinase-mitogen-activator protein-1) signaling pathway [6]. Both IKKβ and JNK are serine kinases that disrupt insulin signaling pathways, eliciting adipose tissue inflammation [51]. In our investigation, PALM INTER presented the highest phosphorylation of IKKα/β. It is well documented that the IKKβ-NF-κB signaling pathway not only induces the transcription of proinflammatory mediators, such as IL-1β, IL-6, TNF-α, but also contributes to IR and NAFLD development [52].
Regarding the effects of the interesterified fats on obesity development, a study conducted on mice showed that the maternal intake of palmitic or interesterified fat enriched with palmitic acid during pregnancy and lactation predisposed the offspring to the development of obesity in adult life, which was attributed to interesterified fats adversely influencing epigenetics [53]. According to another study, a higher concentration of palmitic acid was incorporated into adipocyte phospholipids of rats submitted to a diet enriched with palmitic acid in the sn-2 position of the glycerol backbone [29].
We were also looking at identifying the impact of interesterified fats on hepatic lipid metabolism, since a high-fat diet [56], particularly high in saturated fat, is related to hepatic IR and TG accumulation [6], promoting inflammation and fibrosis [57]. We did not find differences in triglyceride content among groups, since all animals developed NAFLD because of the high-fat diet. Similar results were found by Ebbert & Jensen [58] in the same animal model submitted to a high-SAFA or high-PUFA diet.
Although PALM INTER presented a less inflammatory cytokine level (TNF-α and IL-1β), this group showed higher fibrosis and apoptosis as compared to PALM and PUFA, as well as higher JNK activation. Fibrosis and apoptosis processes are late events in NAFLD [59]. Furthermore, although inflammation classically precedes fibrosis, the amount of fibrosis is not automatically associated with the severity of inflammation, indicating that the mechanisms that regulate fibrogenesis may be different from those that regulate inflammation [60]. Thus, fibrosis is not permanently characterized by persistent inflammation that usually occurs after higher inflammation degree and may explain lower levels of TNF-α and IL-1β. JNK is a serine kinase involved in the inflammatory signaling pathway that also activates apoptotic proteins, such as Bax. It was demonstrated in vitro that hepatocytes treated with monounsaturated and saturated fatty acids induced the same degree of steatosis, higher apoptosis, related to JNK activation [61].
Moreover, STEAR INTER-fed mice presented higher pJNK without expressing apoptosis markers, but this was the only group presenting NASH-like lesions with higher neutrophils infiltration as compared to all groups, probably because of the interesterification process, since the amount of SAFAs in the sn-2 position remained the same. However, STEAR INTER-fed mice did not present higher adipose tissue expansion as compared to the others. In fact, it was already demonstrated that among SAFAs, stearic acid has a higher melting point, which hinders micelle formation, contributing to less absorption and adipose tissue formation in animals fed with these fats [62].
Altogether, our results showed that interesterified fats enriched with palmitic or stearic acids induced hepatic fibrosis, whereas PALM INTER triggered adipocytes hypertrophy in LDLr-KO mice. As interesterified fats are largely utilized by the food industry, the adverse effects shown here warrant further investigation in humans.

Limitations of the Study
The limitations of this study refer to the NMR analysis lacking the ability to determine the position of fatty acids in the glycerol backbone from triglycerides extracted from liver and adipose tissue. We were also not able to determine the composition of chylomicrons due to the complexity of the method when realized in mice. PALM