AMP-activated protein kinase activation ameliorates eicosanoid dysregulation in high-fat-induced kidney disease in mice.

High-fat diet (HFD) causes renal lipotoxicity that is ameliorated with AMP-activated protein kinase (AMPK) activation. Although bioactive eicosanoids increase with HFD and are essential in regulation of renal disease, their role in the inflammatory response to HFD-induced kidney disease and their modulation by AMPK activation remain un-explored. In a mouse model, we explored the effects of HFD on eicosanoid synthesis and the role of AMPK activation in ameliorating these changes. We used targeted lipidomic profiling with quantitative MS to determine PUFA and eicosanoid content in kidneys, urine, and renal arterial and venous circulation. HFD increased phospholipase expression as well as the total and free pro-inflammatory arachidonic acid (AA) and anti-inflammatory DHA in kidneys. Consistent with the parent levels, the AA- and DHA-derived lipoxygenase (LOX),

risk factor for both the initiation and progression of kidney disease independent of hypertension and diabetes (2)(3)(4). Excessive intake of calorie-dense lipids leads to organ dysfunction both by direct lipotoxicity and inflammation. In fact, the Western diet enriched in saturated animal fats has been shown to increase albuminuria and cause a faster decline in renal function (5). In our previous studies, we established that high-fat diet (HFD)-induced kidney disease is characterized by renal hypertrophy, increased albuminuria, and elevated markers of renal fibrosis and inflammation (6). These HFD-induced markers of inflammation, oxidative stress, and fibrosis are reversed by AMP-activated protein kinase (AMPK) activation (6,7).
HFD alters the activity of the crucial lipid metabolism enzymes, acetyl-CoA carboxylase (ACC) and HMG-CoA reductase (HMGCR), contributing to lipid accumulation in the kidney (6). Total cholesterol esters and phosphatidylcholine content in the kidney are elevated, while the FA and triglyceride content is unchanged. Phospholipid accumulation in the proximal tubules is associated with lysosomal dysfunction, stagnant autophagic flux, mitochondrial dysfunction, and inflammasome activation (8). High-fat feeding for long periods causes recruitment of macrophages, switch to macrophage pro-inflammatory phenotype, and increased inflammatory mediators like TNF, monocyte chemoattractant protein-1 (MCP-1), IL-6, cyclooxygenase (COX)-2, and IL-1 (9)(10)(11)(12). Diets rich in PUFAs are known to change plasma lipids, renal phospholipid content, and, subsequently, PUFA-derived eicosanoid inflammation in a rat model of nephrotic syndrome (13). However, the influence of renal eicosanoid synthesis and eicosanoid-derived inflammation in HFD-induced kidney disease is unknown.
Eicosanoids are oxylipins derived from arachidonic acid (AA) or related PUFAs and are inextricably related to inflammation in the kidney. The primary PUFAs for the n-6 series and n-3 series, linoleic acid (LA) and -linolenic acid (ALA), respectively, are both derived from the diet. These 18-carbon PUFAs are then metabolized by various desaturase and elongase enzymes in a stepwise fashion. However, both LA and ALA are acted on by the same enzymes, resulting in a competition between the n-3 and n-6 series (14). LA is metabolized through multiple steps to dihomo-linolenic acid (DGLA; 20:3n6) and, ultimately, to AA (20:4n6). On the other hand, ALA is metabolized to EPA (20:5 n-3) and subsequently to DHA (22:6 n-3) (15). These PUFAs are incorporated into membrane phospholipids and released by phospholipase A2 (PLA2) under the influence of various stimuli. In subsequent reactions, COXs, lipoxygenases (LOXs), and cytochrome P450 (P450) enzymes act on free PUFAs to form eicosanoids. Some eicosanoids can also be formed from PUFAs via nonenzymatic reactions [nonenzymatic degradation (NE)], e.g., isoprostanes.
Eicosanoids play an essential role in the regulation of renal physiology and disease by modulating renal blood flow, glomerular filtration rate, autoregulation, tubular glomerular feedback, excretion of renal water and sodium, and release of renin and erythropoietin. HFD feeding causes an increase in circulating eicosanoids. In the kidney, these eicosanoids are produced by all different cell types: mesangial cells, renal microvessels, and tubular cells. This makes it difficult to pinpoint the actual origin of these autacoids without actual profiling of the various compartments. Local production in the kidney will be reflected in the kidney tissue, renal venous compartment, and urine. Recent advances in eicosanoid analysis using highly sensitive MS have enabled us to profile over 150 different eicosanoid metabolites reliably in all tissues, enabling us to systematically profile the changes in the metabolic pathways with HFD and 5-aminoimidazole-4-carboxamide-1--D-furanosyl 5′-monophosphate (AICAR) therapy.
AMPK is a ubiquitous heterotrimeric kinase that acts as a cellular energy sensor that responds to changes in the intracellular AMP/ATP ratio (16). AICAR acts as a specific AMPK agonist (17). AMPK activation leads to inhibition of energy-requiring biochemical processes, like FA synthesis, and stimulation of energy-producing biochemical pathways, like -oxidation, to improve energy efficiency (18). Metabolic stress, such as diabetes or obesity, impairs the activity of AMPK, and AMPK activation reduces the initial and sustained inflammatory response in the kidney of the HFD-induced kidney disease model (6). Along with lipid accumulation, the markers of inflammation were modulated with AICAR use (7). AMPK signaling has been shown to influence the secretory PLA2 expression in vascular smooth muscle cells (19) and control triglyceride content in adipocytes (20). AMPK activation also decreases the formation of 15-LOX metabolites of AA in macrophages (21). While AMPK activation is beneficial in lipid and eicosanoid metabolism in other tissues, the effect of HFD and AMPK activation on eicosanoid pathways in the kidney is unknown. We hypothesized that the high-fat exposure triggers inflammation involving the eicosanoid pathway and that eicosanoid production is ameliorated with AMPK activation. We used a targeted lipidomic platform to systematically investigate the HFD-associated eicosanoid synthesis induced in mice consuming HFD with or without AMPK activation in order to better understand the pathophysiological processes involved in HFD-induced kidney disease.

Animals
All animal procedures were approved by the Institutional Animal Care and Use Committee of University of California, San Diego. Male 6-week-old C57BL/6 mice were purchased from Jackson Laboratory (Bar Harbor, ME) and fed either a standard diet (STD) [5% fat (PUFA, 2.1%; n-6, 1.9%; n-3, 0.2%), 24.5% protein, 40% carbohydrate] or a HFD [60% of total calories from fat (90% lard + 10% soybean oil; PUFA, 16.9%; n-6, 15.1%; n-3, 1.7%), 20% protein, 20% carbohydrate] (D12492; Research Diets, New Brunswick, NJ) for 14 weeks. The mice on the STD were treated with PBS and mice on HFD were treated with AICAR (0.5 mg/g body weight; Toronto Chemicals) or PBS via intraperitoneal injections for 5 days a week for a total of 14 weeks (6). Mice were placed in metabolic cages for 24 h urine collection before the start of the diets and after 14 weeks. Mice were euthanized after 14 weeks of diet, plasma was collected from the renal vein, and arterial blood was obtained from direct cardiac puncture. After perfusion with PBS, kidneys were snap-frozen in liquid nitrogen for further analysis.

Quantitative total and free FA analysis
Total and free FAs were analyzed by GC-MS as previously described (22,23). Briefly, kidney tissue (10 mg of tissue) was homogenized in 1 ml of PBS containing 10% methanol. For the analysis of total (esterified and free) FAs, kidney homogenates (50 l) were spiked with deuterium-labeled FAs, acidified with 200 l of 0.1 N hydrochloric acid, and extracted with 500 l of methanol. Phase separation was achieved by the addition of 500 l of dichloromethane (CH 2 Cl 2 ), and the organic phase was removed. The extraction with dichloromethane was repeated, and the combined extracts were dried under argon gas. For saponification, the dried lipids were resuspended in 250 l of methanol and 250 l of 4 N potassium hydroxide, vortexed for 30 s, and incubated at 37°C for 1 h. The lipid hydrolysates were then neutralized with 260 l of 4 N HCl and the free FAs were extracted twice with 2 ml of isooctane, and the combined extracts containing the hydrolyzed FAs were dried under argon before analysis with GC-MS.
Free FAs were extracted from 50 l of the kidney tissue homogenates that were supplemented with a set of deuterated FAs that served as internal standards. The extraction was initiated by the addition of 25 l of 1 N hydrochloric acid and 500 l of methanol, and a biphasic solution was formed by the addition of 1 ml of isooctane. The isooctane phase containing the free FA fraction is removed, the extraction is repeated once more, and the combined extracts are dried under argon before analysis with the GC-MS.
In preparation for GC-MS analysis, the FAs (both total FA after saponification and free FA extracted from samples and the quantitative standards) were taken up in 25 l of 1% diisopropylethylamine in acetonitrile and derivatized with 25 l of 1% pentafluorobenzyl bromide. The FA esters were analyzed on an Agilent 6890N gas chromatograph equipped with an Agilent 5973 mass selective detector (Agilent, Santa Clara, CA) operated in the negative chemical ionization mode. Fatty acid quantitation was achieved by the stable isotope dilution method.

Quantitative eicosanoid analysis
Eicosanoids were analyzed by LC-MS, as described before (23,24). For tissue analysis, eicosanoids were isolated from 0.9 ml of kidney homogenate supplemented with a set of internal standards consisting of 26 deuterated eicosanoids (Cayman Chemical) by solid phase extraction (Strata-X; Phenomenex, Torrance, CA) and eluted into 1 ml of methanol. For eicosanoid analysis in urine, 600 l of sample were processed identically. The extracted samples were brought to dryness, reconstituted in 100 L of LC buffer consisting of 63% water, 37% acetonitrile, and 0.02% formic acid, and a 40 L aliquot was separated by reverse phase LC using a Synergy C18 column (2.1 × 250 mm, 4 m; Phenomenex). Eicosanoids were analyzed using a tandem quadrupole mass spectrometer (MDS SCIEX 4000 Q Trap; Applied Biosystems, Foster City, CA) via scheduled multiple reaction monitoring in the negative ionization mode. Eicosanoids were identified in samples by matching the multiple reaction monitoring signal and LC retention times with those of pure standards. Data analysis was performed using MultiQuant 2.1 software (Applied Biosystems). Quantitative eicosanoid determination was performed using the stable isotope dilution method.

Immunoblot
Kidney tissue homogenates were electrophoretically separated on NuPAGE bis-Tris gels (Life Technologies) and transferred onto nitrocellulose membrane (Life Technologies). Antibodies to ELOLV5, cytosolic phospho-PLA2 and secretory PLA2 (Abcam), FADS1/FADS2 (Biorbyt, UK), and actin (Sigma) were used on the blots. Detection was performed with ECL Plus detection reagents (GE Healthcare) after treatment with appropriate secondary antibodies.

Statistical analysis
Results are presented as mean values ± SD. The kidney values were normalized with tissue weight and plasma values were normalized to volume. Urinary metabolites are all normalized to urine creatinine values. All metabolites were log transformed and auto-scaled before undergoing statistical analysis using the MetaboAnalyst version 3.0 software (25). Figures were generated using GraphPad Prism software version 4.03 . The difference between data groups was evaluated for significance using two-way ANOVA and Fisher's least significant difference method post hoc tests for multiple comparisons. A P-value less than 0.05 was defined as statistically significant in the non-MS experiments. To account for multiple comparisons with a large number of lipid metabolites, a false discovery rate (FDR; q value) of <0.05 was considered statistically significant. The highly correlated metabolites derived from a single enzyme pathway from a specific parent FA in each tissue compartment and then were separately aggregated into one secondary variable representative of the corresponding pathway using principal component (PC) analysis. We have then used the PC explaining the highest variance in the variable for further analysis to detect trends using SPSS version 25 (IBM Corp.) (26).

HFD increased phospholipase expression and free levels of PUFAs, and this action is reversed by AMPK activation
Both cytosolic and secretory PLA2 act on membrane phospholipids to release PUFAs, the precursors of various eicosanoids. Levels of cytosolic phospho-PLA2 (group IV) and secretory PLA2 were both increased with HFD in the kidney, and AICAR treatment decreased their levels ( Fig. 1).
Targeted lipidomic analysis using MS of kidney tissue revealed unique patterns in the total and free 3 and 6 PUFA series that are demonstrated in Fig. 2. The total FA (esterified and unesterified FA) and free FA (unesterified) profiles generated from the kidney of mice fed a low-fat standard chow (STD), HFD, or HFD with the AMPK activator AICAR (HFD+AICAR) are displayed in supplemental Table S1. The total EPA and free EPA are decreased in the kidneys with the HFD (Fig. 2A, B; P < 0.05) and further decreased with AICAR, while both total DHA and free DHA (an n-3 FA similar to EPA) are increased with HFD and reversed with the use of AICAR (Fig. 2C, D; P < 0.05). While DHA and EPA levels are modulated with HFD, the parent PUFA, ALA (total and unesterified), is unchanged.
The free and total n-6 FA, DGLA, is decreased in the kidney with HFD, but unchanged with AICAR ( Fig. 2E, F). Total AA and free AA in the kidney are increased by HFD, and the free AA decreased with AICAR ( Fig. 2G, H). Free and total levels of LA, the precursor for both DGLA and AA, are unchanged by the diet and AICAR in the kidney. Delta-6 desaturase and delta-5 desaturase catalyze desaturations at specific positions of FA substrates. Elongases extend the FA carbon chains by two carbons. The levels of both desaturases were not different between the three groups (supplemental Fig. S1).

HFD and AMPK activation alter the urinary eicosanoid profiles
HFD-induced changes in urine metabolites are illustrated in Table 2 and the metabolites that are altered with AICAR in Fig. 4. The overall trend of the kidney continues in the urine with a predominant increase of COX products of AA in the urine and a decrease of COX products of EPA. Adrenic acid and its COX metabolite, dihomo-prostaglandin (PG)F 2 (dihomo-PGF 2 ), are both increased with HFD in the urine, but only dihomo-PGF 2 is decreased with AICAR ( Fig. 4A). Dihomo-15-deoxy-PGD 2 (dihomo-15d-PGD 2 ), which is a NE byproduct of another adrenic acid COX metabolite, is decreased with HFD and increased with AICAR (Fig. 4B).

HFD and AMPK activation modulate circulating eicosanoid levels
In order to delineate the origin of the eicosanoids in the kidney, both the renal artery ( Table 3) and vein ( Table 4) were sampled after the mice were fed STD, HFD, and HFD with AICAR. The eicosanoid profile of the arterial and venous samples revealed that AA and its metabolite, 8,9-epoxyeicosatrienoic acid (EET) (a product of P450), were both increased with HFD. 5,6-EET and 11,12-EET are increased in the venous circulation, but not in the arterial circulation. Because these metabolites are also elevated in the kidney, this suggested the possible production of the metabolites in the kidney and release into the venous circulation.  EPA levels are decreased in both the arterial and venous circulation. EPA produces hydroxy-eicosapentaenoic acids (HEPEs), NE products; 18-HEPE, 11-HEPE, and 9-HEPE were decreased in the arterial and venous circulation. 5-HEPE, 12-HEPE, and 15-HEPE are produced from their respective LOXs and were decreased with HFD in both the arterial and venous circulation. 15-HEPE is decreased further with AICAR therapy in the arterial circulation. P450 metabolite, 14(15)-epoxyeicosatetraenoic acid [14(15)-EpETE], is decreased with HFD diet in both arterial and venous circulation and further with AICAR in the arterial circulation.

Specific eicosanoids are altered by AMPK activation
When taken as a whole, there were significant changes in eicosanoid pathways in arterial and venous circulation, urine, and kidney compartments (supplemental Table S2). Figure 6 depicts the changes influenced by AICAR therapy on HFD mice. To further detect trends in the changes, the highly correlated eicosanoids derived from a parent PUFA and acted on by specific enzymatic pathways were collapsed into PCs and analyzed for trends and statistical significance. Table 5 represents the trends in the kidney and urine of these highly correlated metabolites. In the kidney, HFD decreased EPA metabolites across all enzymatic pathways, while the COX metabolites alone were reversed by AICAR. Meanwhile, AA metabolites of urine were mostly elevated with HFD, whereas the COX EPA metabolites are decreased with HFD, and this is reversed with AICAR therapy. COX metabolites of adrenic acid are decreased with HFD in the urine and this was reversed with AICAR. Table 6 reveals that EPA, DHA, and ALA metabolites were all decreased with HFD in the arterial and venous blood.

DISCUSSION
This is the first study to systematically explore the effects of HFD on eicosanoid synthesis in the kidney and the role of AMPK activation in ameliorating these changes. We find that HFD increases PLA2 expression and activity (through phosphorylation) in the kidney that is largely corrected by AMPK activation. HFD increases both the pro-inflammatory AA and the anti-inflammatory DHA and their respective downstream products in the kidney and urine, and these changes are reversed with AICAR therapy. However, HFD decreases EPA, DGLA, LOX, and NE metabolites of EPA, while AICAR therapy decreases free EPA levels further in the kidney. The P450 and NE metabolites of LA were decreased in the arterial circulation with HFD. This decrease was prevented with AICAR. These changes highlight the changes in eicosanoid pathways in the kidney with HFD that are ameliorated by AICAR therapy. Obesity and metabolic syndrome patients have demonstrated altered FA and eicosanoid metabolism compared with lean subjects. Increased serum DGLA levels and low delta-5 desaturase activity were associated with hepatic steatosis over and above conventional risk factors in metabolic syndrome subjects (27). Like our HFD mouse model, COX, LOX, and P450 metabolites of AA and other PUFAs are elevated in the plasma of obese subjects compared with lean subjects (28). Increased TXA 2 can cause vasoconstriction and induce transcription of collagen in the glomerular matrix. In contrast, PGI 2 counteracts the deleterious TXA 2 effects associated with progressive glomerular damage by acting as a vasodilator. A decreased urinary PGI 2 /TXA 2 ratio has been demonstrated in diabetic humans compared with controls (29). In the Japanese population, urinary 8-iso-PGF 2 , which is an AA metabolite, predicts metabolic risks like obesity, hypertension, and glucose tolerance (30). Meanwhile, low serum lipoxin A4 levels (AA metabolite) are associated with metabolic syndrome risk in the Chinese population, similar to the elevated kidney levels of lipoxin A4 and its metabolite in our study (31).
Obesity-induced kidney disease is characterized by glomerular enlargement, podocyte loss, and proteinuria, eventually leading to renal failure (2,32,33). In animal models, HFD increased kidney weight and glomerular area with periodic acid-Schiff-positive matrix and increased Oil red-O-positive vacuolated cells. HFD also caused albuminuria, increased glomerular inflammation via the NFB pathway, and increased urinary MCP-1 and hydrogen peroxide (6,34). Increased fat intake increases kidney triglyceride content via increased SREBP-1 and -2 and lipogenesis along with reduced AMPK and increased ACC expression and activity in mice (7,10,35). Also, HFD decreased renal lipolysis along with carnitine palmitoyl acyl-CoA transferase-1 expression (36). Mice fed with HFD demonstrated vacuolization of the proximal tubule enriched with multi-lamellar lysosomal bodies with phospholipids the phenomenon was confirmed with human biopsies (37). Human studies have described the accumulation of ectopic lipid in obesity-induced kidney disease (4,38). Mice lacking the innate immune receptor, Nlrp3, did not develop the pathological markers of HFD-induced kidney disease, indicating that pathways regulated by Nlrp3 could control the development of HFD-induced renal pathology (39). Clearly, HFD causes dysregulation of mainstream lipid metabolism in the kidney, but the gateway for many inflammatory mediator's eicosanoid generation with HFD has not been clearly examined.  The first and rate-limiting step of eicosanoid synthesis is the liberation of PUFAs from their esterified form in the sn-2 position of membrane phospholipids by cytosolic phospholipase. In our study, both the cytosolic (group IV) and secretory PLA2 were elevated with HFD and decreased with AICAR. Cytosolic PLA2 is essential for adipocyte differentiation in obesity models (40). Cytosolic PLA2 (group IV)-deficient mice are resistant to the hepatic effects of HFD due to the decrease in the production of downstream products, confirming the crucial role of this enzyme in high induced fat disease (41). Cytosolic PLA2 is also central to oxidative signaling in renal epithelial cells, releasing AA and its downstream products (42). Isoenzymes of PLA2 are differentially expressed in various tissues in response to inflammation; for example, expression of the group VII PLA2 gene is increased in the livers of Zucker obese rats, while expression of group IVA and VIA PLA2 was inhibited (43). PLA2 isoenzymes also demonstrate substrate selectivity to certain forms of phospholipids in membranes and specific PUFAs (44). Kidney desaturases and elongases that  are responsible for conversion between the various PUFAs are unaffected by HFD or AMPK activation in our study. However, livers of obese Zucker rats demonstrated increased expression of delta-6 desaturase and elongase-6 (43). Diets in the HFD and STD groups had similar n-6 to n-3 ratios, and the pattern in the circulation and the kidney do not reflect this ratio. AA is increased in the circulation, kidney, and urine with HFD in our study and this is consistent with increased circulating PUFAs demonstrated with HFD feeding in Sprague-Dawley rats (45). In excess nutrient states, phospholipase D suppresses AMPK activity through the mammalian/mechanistic target of rapamycin (mTOR); in return, AMPK activation decreased phospholipase D activity similar to the effect of AMPK activation on other phospholipases in vascular smooth muscles and endothelial cells (19,46,47). Thus, the effect of AMPK activation on phospholipases in our study could be responsible for the changes to the PUFA levels with AICAR therapy. Both COX-1 and COX-2 and their downstream products, PGD 2 , PGE 2, and PGI 2 , are involved in renal vasodilation. PGE 2 is the central PG synthesized in the kidney tubules and mesangial cells, and activation of its receptors causes contraction of vascular muscle cells and increased calcium in mesangial cells and inhibits sodium reabsorption at the thick ascending limb of the loop of Henle (48). In our study, PGE 2 is decreased in the kidney with HFD, but its metabolites, PGA 2 , PGB 2 , and PGEM, are increased in the urine with HFD. PGI 2 metabolites, 6,15 dk-,dh-PGF 1 and 6k-PGE 1 , are increased in the urine with HFD similar to the PGD 2 metabolites, PGJ 2 and dhkPGD 2 . Urinary AA-derived eicosanoids, primarily downstream of COX metabolite, PGH 2 , were increased with HFD and decreased with AICAR therapy. Increased flux in these pathways possibly results in decreased bioavailability of the renal vasodilators. In line with our findings, COX-2 expression was increased in the renal cortex, blood vessels, and urine in obese Zucker rats (49). Increased urinary TXB 2 and 6-keto-PGF 1 and decreased PGE 2 excretion rates were evident in obese Zucker rats. Similarly, a decreased urinary PGI 2 / TXA 2 ratio has been demonstrated in diabetic humans as well as in diabetic animal models. Rofecoxib, a COX-2 inhibitor, decreases vascular and glomerular damage when administered to obese Zucker rats (29). Decreased PGF 2 and 8-iso-PGF 2 levels contribute to the decreased glomerulosclerosis in obese Zucker rats treated with rofecoxib. COX-2 inhibition and decreases in 8-iso-PGF 2 levels have also been shown to ameliorate renal injury associated with hypertensive rats (50). This pattern is also reflected in the urinary levels of 8-iso-PGF 2 in our mouse model of HFDinduced kidney disease.
Renal P450 epoxidation generates EETs that are metabolized by soluble epoxide hydrolase (sEH) to less active dihydroxyeicosatrienoic acids. Decreased renal expression of CYP2C epoxygenase enzymes has been observed in diabetes, HFD-fed insulin resistance rats, and the kidney and mesenteric blood vessels of obese Zucker rats (51). Additionally, sEH expression is increased in blood vessels of obese rats and could further contribute to the decrease in EET bioavailability, implicating impaired endothelial dilator responses in obesity and diabetes. The cytokines, IL-6 Statistical analyses were performed by one-way ANOVA followed by Fisher post hoc tests. All P values corrected for Bonferroni FDR correction of *q  0.05. LA metabolites of P450 in the artery, 9,10-diHOME (A), 12,13-EpHOME (B), and 12,13-diHOME (C). Venous LOX metabolites of ALA, 9-HOTrE (D), and EPA, 5-HEPE (E). and TNF-, decrease expression of P450 epoxygenase enzymes. TNF- and EETs act via the NFB pathway causing elevated MCP-1 levels that have been observed in HFDinduced kidney disease in mice (6). Obese Zucker rats have increased renal vascular CYP4A expression and increased generation of 20-HETE (49). However, renal tubular CYP4A expression was decreased in Sprague-Dawley rats on HFD (52). The AA metabolites, 20-HETE, EET, and dihydroxyeicosatrienoic acid, are decreased with HFD in the kidney and associated with the diet-induced change in blood pressure and renal function of Sprague-Dawley rats (52,53). In our study, the kidney P450 products of AA, 9-HETE and EETs, are increased with HFD, just as many of the EETs are also increased in the circulation and urine. Pharmacological inhibition or transgenic deletion of sEH, the dominant enzyme in degrading EETs, has been shown to protect mice from various adverse effects induced by obesity (54). Thus, protective eicosanoid metabolites of AA-EET are increased with HFD in the kidney. In HFD-fed mice, plasma profiling revealed increases in circulating P450 metabolites of LA-EpOMEs (55), which is a different trend to the one observed in the arterial circulation of our HFD mice maintained on HFD for a longer period.
Central adipose tissue in obese Zucker rats shows increased LOX expression along with inflammatory markers (56). 5-LOX products induce NF-B expression and secretion of inflammatory cytokines like MCP-1, TNF-, macrophage inflammatory protein-1, and IL-6 by the adipose tissue. The 5-LOX pathway plays a major role in obesityinduced fatty liver disease (57). In our study, LOX metabolites of AA and AA metabolites, 6(R),15(R)-LXA 4 and 6(S)-LXA 4 , are increased in the kidney with HFD in our study, and LXA 4 has been shown by our group to attenuate inflammation in HFD-related kidney disease (10). Similarly, LOX metabolites of DHA-HDoHE were elevated in the kidney, whereas LOX metabolites of EPA-HEPEs were all decreased in the kidney and the circulation.
Omega 3 FA (DHA and EPA) supplementation in ob/ob mice induced AMPK phosphorylation and alleviated hepatic steatosis. Omega 3 supplementation is anti-inflammatory, promotes insulin sensitivity, and raises n-3 FA-derived resolvins and protectins (58). Obesity significantly decreased DHA-derived 17-hydroxydocosahexaenoic acid (resolvin Fig. 6. Schematic diagram of the eicosanoid metabolites regulated by AMPK activation to reverse the effects of HFD in the kidney, urine, and arterial and venous circulation. Adrenic acid metabolites, dihomo-PGF 2 and dihomo-15d-PGD 2 ; AA metabolites, PGEM and PGFM, and PGF 2 ; EPA metabolites, PGE 3 , TXB 3 , and 5-HEPE; ALA metabolite, 9-HOTrE; DHA metabolites, 19,20-DiHDPA, 16-HDoHE, 15(t)-PD 1 , 7-HDoHE, 11-HDoHE, 8-HDoHE, and 13-HDoHE; and finally LA metabolites, 9,10-diHOME; 12,13-EpOME, and 12,13-diHOME. Red font indicates that HFD increases these metabolites and AICAR reverses them, while blue font indicates the metabolite is decreased with HFD and AICAR reverses this trend. D1 precursor) and PD 1 levels in murine adipose tissue. Notably, 17-hydroxydocosahexaenoic acid treatment reduced adipose tissue expression of inflammatory cytokines, increased adiponectin expression, and improved glucose tolerance parallel to insulin sensitivity in obese mice (59). A concentrated formulation of n-3 PUFAs attenuates albuminuria, renal function, SREBP-1 expression, and triglyceride levels in the kidneys of type 2 diabetic obese db/db mice (60). DHA and DHA metabolites in the kidney, the various HDoHEs, EpDPE and DiHDPA produced by P450, LOX, and NE in the kidney are all increased with HFD in our study. While the kidney and urine DHA metabolites were increased, the circulatory DHA metabolites were decreased, suggesting local production in the kidney. With HFD exposure, the DHA metabolites, 15(t)-PD 1 and PDX, are both increased in the kidney and urine, respectively. In our study, HEPEs derived from the n-3 EPA were all decreased in the kidney, urine, and circulation. The EPA metabolite, PGH 3 , and its hydroxyl metabolites were all decreased in the circulation with HFD. Downregulation of HEPEs in our study represents suppression of anti-inflammatory signals. AMPK is a nutrient sensor and is inhibited by nutrient excess states. In our prior studies, the effects of HFD-induced kidney disease were reversed with AMPK activation with AICAR therapy. AMPK activation reversed the inflammatory profile, lipid accumulation, and low adiponectin levels that were caused by HFD (9,10,61). Diet reversal alone was unable to improve the renal inflammation and apoptosis in HFD-induced kidney damage, emphasizing the persistence of renal injury even after 8 weeks of dietary control (62). However, AMPK activation with metformin treatment was able to reduce the levels of inflammatory and apoptotic markers that remain residual despite diet reversal. Also, metformin ameliorated HFD-induced glomerular injury, renal fatty acid oxidation, and serum adipokine levels (62)(63)(64). Metformin reduces fat content by decreasing SREBP-1, FAS, and ACC expression in the kidney (65). In our study, AICAR reversed the upregulation of phospholipase and free PUFAs. The influence of AICAR is evident in reversing the increased hydroxyl DHA metabolites in the kidney and COX metabolites of AA in the urine. In the venous circulation, AICAR reverses the downregulation of the P450 metabolite of LA, 9-HOTrE, and the LOX metabolite of LA, 5-HEPE. HFD has been previously shown to increase leukotoxins, such as EpOMEs; while in our study, EpOME metabolites are decreased (55). AICAR also reverses the downregulation of 9,10-diHOME, 12,13-diHOME, and 12,13-EpOME in the arterial circulation in our study. Thus, AMPK activation is able to reverse key pathways in eicosanoid metabolism and helps to ameliorate changes resulting from HFD.
In this current study, we demonstrate modulation of PUFAs and their metabolites after HFD as well as AMPK activation in the kidney, but we cannot delineate the origin of the various eicosanoids in the circulation and urine, which could be from the diet, liver, muscle, pancreas, and/or gut microbiota. Again, the changes in the eicosanoids might be influenced by the upstream action of phospholipases on specific PUFAs whose pattern might be reflective of the specific diet; although both of the diets had a similar n-6:n-3 ratio, the specific composition of HFD might have influenced the eicosanoid changes. Also, as the influence of AICAR is not restricted to one specific oxygenase or PUFA, its influence on phospholipases might influence its action to follow the pattern of the predominant PUFA availability in various compartments. Though we have independently demonstrated the effects of HFD on the kidney pathology and the lipid and inflammatory profile, we cannot conclusively provide evidence that the eicosanoid changes are the sole cause of HFD-induced kidney disease or the reason for the AICAR-induced amelioration. Also, the eicosanoids might have differential changes based on temporal exposure to HFD, and this was not conclusively studied in this work. It is also difficult to tease out the source of inflammation in such models, as both HFD and AICAR therapy, in addition to altering inflammation, also influence body weight. Another factor to account for is that pharmacologic AMPK inhibition might not be complete, and genetic manipulation of AMPK might provide a clearer picture. Genetic manipulation of the pathways clearly influenced by AMPK will help us to study downstream and upstream changes and their interaction with the eicosanoid pathways and link them to renal pathology.
In summary, in this study using a targeted lipidomic approach, we demonstrate the dysregulation of eicosanoid synthesis and metabolism in the kidney with HFD and its amelioration with AMPK activation. Our work sheds light on the mechanism behind HFD-related toxicity and the eicosanoid pathways under the control of AMPK activation in the kidney. We also expose the various pro-and anti-inflammatory mechanisms in HFD-mediated kidney disease that are altered with AMPK activation. Future studies that manipulate these dysregulated pathways might open therapeutic avenues in the management of high-fat-induced kidney disease.