Beyond detoxification: a role for mouse mEH in the hepatic metabolism of endogenous lipids

Abstract

Microsomal and soluble epoxide hydrolase (mEH and sEH) fulfill apparently distinct roles: Whereas mEH detoxifies xenobiotics, sEH hydrolyzes fatty acid (FA) signaling molecules and is thus implicated in a variety of physiological functions. These epoxy FAs comprise epoxyeicosatrienoic acids (EETs) and epoxy-octadecenoic acids (EpOMEs), which are formed by CYP epoxygenases from arachidonic acid (AA) and linoleic acid, respectively, and then are hydrolyzed to their respective diols, the so-called DHETs and DiHOMEs. Although EETs and EpOMEs are also substrates for mEH, its role in lipid signaling is considered minor due to lower abundance and activity relative to sEH. Surprisingly, we found that in plasma from mEH KO mice, hydrolysis rates for 8,9-EET and 9,10-EpOME were reduced by 50% compared to WT plasma. This strongly suggests that mEH contributes substantially to the turnover of these FA epoxides-despite kinetic parameters being in favor of sEH. Given the crucial role of liver in controlling plasma diol levels, we next studied the capacity of sEH and mEH KO liver microsomes to synthesize DHETs with varying concentrations of AA (1-30 μM) and NADPH. mEH-generated DHET levels were similar to the ones generated by sEH, when AA concentrations were low (1 μM) or epoxygenase activity was curbed by modulating NADPH. With increasing AA concentrations sEH became more dominant and with 30 μM AA produced twice the level of DHETs compared to mEH. Immunohistochemistry of C57BL/6 liver slices further revealed that mEH expression was more widespread than sEH expression. mEH immunoreactivity was detected in hepatocytes, Kupffer cells, endothelial cells, and bile duct epithelial cells, while sEH immunoreactivity was confined to hepatocytes and bile duct epithelial cells. Finally, transcriptome analysis of WT, mEH KO, and sEH KO liver was carried out to discern transcriptional changes associated with the loss of EH genes along the CYP-epoxygenase-EH axis. We found several prominent dysregulations occurring in a parallel manner in both KO livers: (a) gene expression of Ephx1 (encoding for mEH protein) was increased 1.35-fold in sEH KO, while expression of Ephx2 (encoding for sEH protein) was increased 1.4-fold in mEH KO liver; (b) Cyp2c genes, encoding for the predominant epoxygenases in mouse liver, were mostly dysregulated in the same manner in both sEH and mEH KO mice, showing that loss of either EH has a similar impact. Taken together, mEH appears to play a leading role in the hydrolysis of 8,9-EET and 9,10-EpOME and also contributes to the hydrolysis of other FA epoxides. It probably profits from its high affinity for FA epoxides under non-saturating conditions and its close physical proximity to CYP epoxygenases, and compensates its lower abundance by a more widespread expression, being the only EH present in several sEH-lacking cell types.

Keywords: EET; EpOME; Lipid signaling; Liver; Microsomal epoxide hydrolase.

Fig. 1

Plasma diol and 20-HETE levels and diol:epoxide ratios in WT control, mEH KO, and sEH KO mice. a Levels for DHETs, the secondary metabolites from AA generated by EH conversion; 8,9-DHET levels were significantly lower in mEH KO relative to WT mice. b Levels for DiHOMEs, the secondary metabolites from linoleic acid generated by EH conversion; 9,10-DiHOME levels were significantly lower in mEH KO relative to WT mice. c The DHET:EET ratio, reflecting the hydrolysis rate, was lower for 8,9-, but higher for 14,15-EET in mEH KO compared to WT plasma. d The DiHOME:EpOME ratio was reduced for 9,10-EpOME, but increased for 12,13-EpOME in mEH KO relative to WT plasma. e 20-HETE levels were similar across genotypes. 1-way ANOVA followed by Dunnett’s post tests; statistically significant differences for WT vs KO are indicated. * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 2

sEH activity and protein expression are increased in mEH KO compared to WT liver. a Turnover assay with WT and mEH KO liver cytosol using 14,15-EET as substrate in presence and absence of the selective sEH inhibitor tAUCB. The conversion rate to 14,15-DHET was 1.4-fold higher in mEH KO compared to WT cytosol (n = 5; 1-way ANOVA followed by Bonferroni post tests). b Representative immunoblot comparing sEH expression in WT and mEH KO liver. c Quantification of immunoreactive bands revealed a 1.3-fold higher sEH expression in mEH KO compared to WT liver (n = 3; unpaired Student’s t test). * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 3

EH-mediated generation of FA diols in WT, mEH KO, and sEH KO liver homogenates under saturating conditions. a The formation rates for the four DHET regioisomers were generally higher in mEH KO, but lower in sEH KO compared to WT (statistical significances are indicated for WT vs KO; n = 5 for each genotype; 2-way ANOVA, followed by Dunnett’s post tests). b The formation rates for total DHETs, comprising 8,9-, 11,12-, and 14,15-DHET. In sEH KO liver, mEH could not entirely compensate for sEH loss, but still generated 60% of DHETs formed in WT liver by both EHs. c, d Formation rates for 9,10- and 12,13-DiHOMEs, which are higher in mEH KO compared to WT. 12,13-DiHOME generation was significantly reduced in sEH KO, pointing towards a pivotal role for sEH in its formation. e EETs and f EpOMEs were exclusively detectable in sEH KO, but not in WT and mEH KO liver homogenates after 30 min incubation with AA. g Turnover of 8,9-EET with WT liver without inhibitors and in presence of the sEH inhibitor tAUCB alone and in combination with the mEH inhibitors TCPO/NSPA (TC/NS). DHET formation was reduced by 50% in presence of tAUCB and by 95% in presence of all blockers, supporting the notion that mEH and sEH are the only EET-conversing EHs in mouse liver. h 20-HETE formation was similar across genotypes. n = 5 for each genotype and metabolite or treatment group; for b–d, g, h 1-way ANOVA followed by Dunnett’s post test. * p < 0.05, ** p < 0.01, *** p < 0.001

Fig. 4

mEH contribution to DHET formation depends on the availability of AA and regenerating system. a In mEH KO liver microsomes formation rate of [EETs and DHETs], assumed to reflect epoxygenase activity, increased steadily with rising AA concentrations and incubation time, indicating that hepatic epoxygenases are not saturated under these conditions. b The same as a for sEH KO microsomes (n = 3 for a, b). c 11,12-DHET formation rate in mEH KO liver microsomes in absence and presence of the sEH inhibitor tAUCB. tAUCB blocked 98% of 11,12-DHET formation. d The same as c with sEH KO microsomes and the mEH inhibitors TCPO/NSPA, which blocked 97% of the DHET formation. e–g Liver microsomes were incubated with AA concentrations varying from 1 to 30 µM in the presence of a regenerating system and 2 mM NADPH. With lower AA concentrations, DHET formation rates were similar between genotypes for the mEH-preferred EET regioisomers (8,9 and 11,12), but were significantly distinct with high AA concentration. h DHET formation rates were similar between the two genotypes even with 30 μM AA, if the regenerating system including NADPH was omitted (n = 4, 2-way ANOVA with Bonferroni post tests). * p < 0.05, *** p < 0.001

Fig. 5

mEH and sEH expression in C57BL/6 mouse liver a mEH-positive cells include endothelial cells (EC), lining branches of the portal vein (PV) and hepatic artery (HA) (arrows), bile duct (BD) epithelial cells, and hepatocytes. b Centrilobular (CL) mEH expression in hepatocytes is slightly stronger than periportal (PP) expression. c No mEH expression is detected in the liver of mEH KO mice. d sEH expression is confined to hepatocytes and bile duct epithelium. e Similar to mEH, sEH shows a gradient of expression ranging from intense in the centrilobular areas to moderate in the periportal hepatocytes. f No sEH expression is detected in the liver of sEH KO mice. g Liver section showing a branch of the central vein stained for the endothelial cell marker CD31 and h mEH; note the mEH-positive Kupffer cells (arrows), which show stronger mEH immunoreactivity compared to the surrounding hepatocytes. i Detail from H (box). Endothelial cells in the crossline are positive for CD31 (red) and mEH (green, middle panel); colocalization shown in yellow (right panel). Scale bar in a, c, d, and f equals 50 μm. Scale bar in b and e equals 250 μm, in g and h 30 μm and in i 10 μm

Fig. 6

Changes in genes of the CYP-epoxygenase–EH pathway in mEH KO and sEH KO liver transcriptome. a Ephx1, encoding for the mEH protein, is 1.35-fold upregulated in sEH KO relative to WT liver. b Ephx2, encoding for the sEH protein, is 1.4-fold upregulated in mEH KO relative to WT liver. Note, that in WT liver Ephx2 is 2.5 times more abundant than Ephx1 (n = 3 for each genotype, 1-way ANOVA followed by Bonferroni post tests; only significant differences between WT and KO are indicated). c Expression changes in Cyp2c and Cyp2j genes, encoding for epoxygenases, are shown as Log2 ratio of the respective KO over WT; the Cyp gene with the strongest change in expression (Cyp2a4) is displayed for comparison. A 1.5-fold increase relative to WT is represented by 0.5 on the Log2 ratio scale, e.g., see the Cyp2c44 bar for mEH KO. Only Cyp2c/2j genes with statistically significant changes (cut-off p < 0.05) relative to WT are depicted as bar graphs. Note that for several of these genes, expression changes occur in parallel for both KO livers. d The same as c for ω-hydroxylase genes. Expression changes in the 20-HETE synthases only occur with low-expression ω-hydroxylase genes