GSK2256294 decreases soluble epoxide hydrolase activity in plasma, muscle and adipose and reduces F2-isoprostanes but does not alter insulin sensitivity in humans

Abstract

Epoxyeicosatrienoic acids (EETs) reduce blood pressure by acting in the vasculature and kidney, and interventions to increase circulating EETs improve insulin sensitivity and prevent diabetes in animal models. Inhibition of EET hydrolysis with a soluble epoxide hydrolase (sEH) inhibitor is an attractive approach for hypertension and diabetes.

We tested the hypothesis that sEH inhibition increases circulating EETs, reduces blood pressure, and improves insulin sensitivity, blood flow, and inflammation in a randomized, double-blind, placebo-controlled cross-over study. Sixteen participants with obesity and pre-diabetes were randomized to GSK2256294 10 mg once daily or placebo for seven days, insulin sensitivity was assessed by hyperinsulinemic-euglycemic clamp, and adipose and muscle tissues biopsies were performed to assess insulin-stimulated pAkt signaling. We assessed tissue and plasma EETs and their respective diol concentrations and sEH activity within plasma, muscle, and adipose tissues.

GSK2256294 reduced circulating and adipose tissue sEH activity but blood pressure, circulating EET and tissue EETs were unchanged. Plasma sEH activity correlated with muscle and adipose tissue sEH activity. Insulin sensitivity assessed during hyperinsulinemic clamps, as well as adipose and muscle pAkt/Akt expression were similar during GSK2256294 and placebo. sEH inhibition with GSK2256294 reduced plasma F2-isoprostanes (50.7±15.8 versus 37.2±17.3 pg/mL, P=0.03) but not IL-6. Resting blood pressure, forearm blood flow, and renal plasma flow were similar during GSK2256294 and placebo.

We demonstrate that GSK2256294 administration for seven days effectively inhibits sEH activity in plasma, muscle, and adipose tissue and reduces F2-isoprostanes, a marker of oxidative stress, but does not improve insulin sensitivity or blood pressure.

Introduction

Epoxyeicosatrienoic acids (EETs) are formed from arachidonic acid by the action of P450 epoxygenases (CYP2C and CYP2J).¹ EETs act as potent vasodilators and have been identified as endothelium-derived hyperpolarizing factors.² In the kidney, EETs promote sodium excretion by inhibiting the translocation of the Na+/H+ exchanger (NHE3) in the proximal tubule.³ EETs reduce inflammation by preventing NFκB activation.⁴ Increasing the actions of EETs in rodent models protects against hypertension, endothelial dysfunction, cardiovascular remodeling, and renal injury.^{1, 2, 5, 6} Soluble epoxide hydrolase (sEH) limits the effects of EETs by hydrolysis to the less active dihydroxyepoxyeicosatrienoic acids (DHETs), and therefore reducing sEH activity is one strategy to increase EETs.⁷ Pharmacological sEH inhibitors have been under development in clinical studies for the treatment of hypertension and lung disease.^8–10

Studies in rodents point to an important role for EETs and sEH in the regulation of glucose homeostasis and insulin sensitivity. Inhibiting sEH or disrupting Ephx2, the gene encoding sEH, increases insulin sensitivity in rodent models of T2DM and insulin resistance.^11–13 These effects have been attributed to enhanced insulin signaling in liver, muscle, and adipose tissue, as well as to beneficial effects on muscle capillary blood volume and microvascular blood flow. ^11–14 Administration of exogenous EETs replicates these effects.¹⁵ Other studies have identified a beneficial effect of EETs or sEH inhibition on islet cell function or apoptosis.^{11, 12, 16, 17} Studies in rodents also suggest that tissue sEH activity increases with obesity, potentially implicating this pathway in obesity complications.^{18, 19}

EETs and EPHX2 also affect glucose homeostasis in humans. Circulating EETs correlate with insulin sensitivity and are significantly decreased in individuals with metabolic syndrome and insulin resistance.^{20, 21} In addition, insulin sensitivity is increased in overweight carriers of a loss-of-function variant in EPHX2 (rs751141 or Arg287Gln), but not in obese carriers of the variant.²¹ The 287Gln allele is also associated with decreased vascular resistance.²² This study tests the hypothesis that treatment with a specific pharmacologic sEH inhibitor (GSK2256294) decreases sEH activity and increases insulin-stimulated vasodilation and insulin sensitivity via effects on tissue (muscle or adipose) insulin signaling and/or increased microvasculature.

Methods

Participants

Obese (BMI ≥ 30 kg/m²) and pre-diabetic volunteers, age 21 through 60 years, were enrolled. Individuals were defined as pre-diabetic if they had a fasting plasma glucose of 100–125 mg/dL, a two-hour plasma glucose of 140–199 mg/dL during a 75-gram oral glucose tolerance test (OGTT), or a hemoglobin (Hb) A1c of 5.7–6.4%. Participants with any history of cancer including skin cancer, cardiovascular disease (other than hypertension), renal, pulmonary, endocrine (other than insulin resistance or hyperlipidemia), or hematologic disease were excluded. Individuals with diabetes, resistant hypertension, or any history of smoking were excluded. Pregnancy was excluded by urine β-HCG testing.

Protocol

All participants gave written informed consent. Participants reported to the Vanderbilt Clinical Research Center (CRC) in the fasting state to give their medical history and undergo a physical examination, screening laboratory and electrocardiogram (ECG). They returned to the CRC in the fasting state for measurement of waist and hip circumference, an OGTT, and body composition measurement by Dual Energy X-ray Absorptiometry (DEXA). This study was registered prior to enrollment of the first participant (NCT03486223). The data that support the findings of this study are available from the corresponding author upon reasonable request.

Participants were then randomized using a permuted-block algorithm to treatment with the sEH inhibitor GSK2256294 (10 mg once daily) or matching placebo for one week in a double-blind, crossover design study. Phase I studies indicate that the maximal concentration and complete inhibition of sEH was achieved within hours of administration of the first dose of 10 mg. With daily dosing, steady-state drug concentrations were achieved within seven days. There was more than 95% sEH inhibition throughout the dosing interval, and inhibition persisted for >60 hours after the last dose.^{9, 10} Study drug was a generous gift of GlaxoSmithKline and was prepared by the Vanderbilt Investigational Drug Service. Participants were also provided a weight-maintenance diet containing 150 mmol/day sodium and 80 mmol/day potassium for one day prior to study. On the seventh day of drug treatment, participants reported to the CRC at 6–8 AM after an overnight fast to undergo a hyperinsulinemic-euglycemic clamp and adipose biopsies. The final dose of treatment drug was administered on the morning of the study and compliance was assessed by pill count. Participants collected urine overnight for eight hours prior to the study day, and again after clamp completion.

On each study day, an IV catheter was placed in both arms. Para-amino-hippurate (PAH) 8 mg/kg was given intravenously as a loading dose, followed by a 12 mg/min continuous infusion for measurement of renal plasma flow (RPF).^{23, 24} After the participant had been supine for 45 minutes, forearm blood flow (FBF), BP, and heart rate were measured in triplicate. FBF was measured using strain gauge plethysmography as previously described.²² Forearm vascular resistance (FVR) was calculated as FBF/mean arterial pressure (MAP). An adipose tissue biopsy was obtained by liposuction, and if the participant consented, a muscle biopsy was obtained from the vastus lateralis muscle. Blood samples were obtained for measurement of PAH, plasma EETs, DHETs, epoxy-12Z-octadecenoic acids (EpOMEs), dihydroxy-12Z-octadecenoic acids (DiHOMEs), interleukin 6 (IL-6), F₂ isoprostanes, vascular endothelial growth factor (VEGF), and potassium. Participants then underwent a hyperinsulinemic-euglycemic clamp (described below). Insulin sensitivity was assessed by the response to low (20mU/m²/min) and high (80mU/m²/min) insulin infusion rates. At the end of the clamp, we repeated FBF measurements, adipose tissue biopsy and optional muscle biopsy. At the end of each study day, we obtained safety laboratory studies including CBC with differential, complete metabolic profile, urinalysis, and ECG.

After completion of the first study day, participants underwent a seven-week washout period and then received the opposite drug for one week. On the sixth day of treatment they were provided the same diet. On the seventh day of treatment they reported to the CRC after an overnight fast and the study day protocol was repeated.

FBF measurements

We measured FBF before and at the end of hyperinsulinemia using venous occlusion mercury-in-sialastic strain-gauge plethysmography. The wrist was supported in a sling to raise the forearm to above the level of the atrium, and the strain gauge was placed at the widest part of the forearm. The strain gauge was connected to a plethysmograph (model EX-5; D. E. Hokanson, Inc., Bellevue, WA), calibrated to measure the percentage change in volume and connected to a chart recorder. For each measurement, a cuff placed around the upper arm was inflated to 40 mm Hg with a rapid cuff inflator (model E-10; D. E. Hokanson, Inc., Bellevue, WA) to occlude venous outflow from the extremity. The hand was excluded from the measurement of blood flow by inflation of a pediatric sphygmomanometer cuff around the wrist to 200 mm Hg before and during measurements of FBF. Flow measurements were recorded for ~7 of 15 s, and the slope was derived from the first three or four pulses; five to seven readings were obtained for each mean value.

Adipose harvest

Adipose tissue was obtained from the periumbilical area using a Tulip^™ Medical closed syringe system for lipoaspiration. Under aseptic conditions and local lidocaine anesthesia, a small incision was made in the skin. The GEMS Johnnie Snap lock was placed into a 60-cc syringe, the Tulip liposuction cannula with 60-cc syringe attached was inserted at an angle through the incision to below Scarpa’s fascia, and suction was applied until the syringe activates the clicker lock. The needle was moved in and out at a rate of approximately 1 Hz without breaking suction with a twisting motion. The sampling continued until approximately 5 g of tissue was removed, and then the sample was immediately separated into pieces, frozen in liquid nitrogen and stored at −80°C for measurement of pAkt and Akt expression and of sEH activity, EETs, DHETs, EpOMEs, and DiHOMEs.

Muscle biopsies

Biopsies were obtained from the vastus lateralis under aseptic conditions and local lidocaine anesthesia by percutaneous needle biopsy using the modified Bergström technique. Briefly, after proper aseptic technique and local anesthesia, a 5 mm incision was made in the skin. The needle was inserted through the incision to the skeletal muscle. The inner trocar of the needle was retracted, and suction applied to pull muscle into the outer trocar. The inner trocar was then closed to cut the muscle. This needle biopsy was then rotated (90°) and the procedure was repeated twice. After removing the needle, tissue was immediately placed in ice-cold 0.9% NaCl, and the subcutaneous fat, if any, separated from muscle. Muscle biopsies were divided, and a fresh sample was taken on ice to the laboratory for measurement of sEH activity, EETs, DHETs, EpOMEs, and DiHOMES; other pieces were frozen in liquid nitrogen and stored at −80°C for Western blotting.

Hyperinsulinemic-euglycemic clamp

After a priming dose, insulin was infused at 20mU/m²/min for 120 minutes to suppress hepatic glucose production partially, and then at 80mU/m²/min to suppress hepatic glucose secretion fully and to stimulate peripheral glucose utilization maximally in insulin-resistant participants. Plasma glucose was measured every 2.5 to 5 min and 20% glucose was given to control the rate of fall of glucose and to maintain the plasma glucose at a target of 90–95 mg/dL.²⁵ Potassium chloride 40 mEq was given orally prior to hyperinsulinemia to maintain potassium levels. Heart rate and ECG were monitored continuously during glucose clamp studies. Participants were instructed to empty their bladder prior to insulin infusion, at the end of the low dose insulin infusion, and at the end of the high dose insulin infusion. To control for inter-individual variations in plasma insulin, the insulin sensitivity index was calculated by dividing M by average steady-state insulin concentration.

Laboratory Analyses

Plasma glucose was measured by the glucose oxidase method with a YSI glucose analyzer (YSI Life Sciences, Yellow Springs, OH). Plasma insulin concentrations were determined by radioimmunoassay (RIA; Millipore, St. Charles, MO). Urine sodium and potassium concentrations were determined by an Instrumentation Laboratory 943 flame photometer. Urine was not diluted prior to measurement. Urine creatinine was determined by 1D ¹H NMR spectra (600 MHz Bruker NMR spectrometer) with normalization to the internal standard dodium trimethylsilylpropanesulfonate.

Plasma EETs, DHETs, EpOMEs, and DiHOMEs were quantified using ultra-performance liquid chromatography/tandem mass spectrometry (UPLC/MS/MS) with select reaction monitoring (SRM) on a triple quadrupole mass spectrometer equipped with an electrospray source (ESI) operated in the negative ion-mode.

Plasma sEH activity was measured by the hydrolysis of pharmacologic concentrations of EETs or EpOMEs to their respective DHETs or DiHOMEs. Human adipose or muscle tissues were homogenized (4°C in 0.1M NaPi, PH 7.4) containing 0.1 mg/mL BSA to have 10% (wt/wt) homogenates in an T8-Ultra-Turrax (IKA, Staufen, Germany) homogenizer. Activity was performed in 50 μL of human plasma with 50 μL of 0.1M NaPi (PH 7.4) containing 0.1 mg/mL BSA or 100 μL of plasma, or 30 μL of human tissue homogenate with 70 μL of 0.1M NaPi (PH 7.4) containing 0.1 mg/mL BSA. After pre-warming the solution to 37°C, 50 μM (1 mM, 5 μL EtOH) of EET or EpOME was added. Samples were then incubated for 10 min at 37°C in a shaking water bath. Reactions were stopped by the addition of 5 mL of AcOEt containing 0.5% AcOH (v/v), followed by addition of 2 ng of D11-DHETs or D4-DiHOMEs (internal standards) and 1 mL of 0.15M KCl, and centrifugation at 3000 × g for 1 min. The organic phase was separated from the mixture and evaporated to dryness under a N2 stream in a 37°C water bath. Each sample was dissolved in a mixture of 15 mM-aqueous AcONH4 (PH 8.3): CH3CN (1:1, v/v), followed by centrifugation at 23,000 × g for 3 min. The supernatant (10 μL) was injected onto a Waters Acquity UPLC coupled with a TSQ Vantage Mass Spectrometer for product- DHETs or DiHOMEs -quantification as previously described.²⁶

Western blot analysis was used to measure phosphorylated and total Akt protein expression in skeletal muscle and adipose tissue biopsies. For this purpose, 30 μg of muscle and adipose tissue were homogenized with RIPA buffer in the presence of proteinase and phosphatase inhibitors. The homogenate was centrifuged, and the lipid layer was removed from the adipose tissue. This process was repeated two more times to ensure complete removal of the lipid layer. Samples were mixed with Laemli buffer and resolved electrophoretically in 4–20% TGX Stain Free^™ precast acrylamide gels (Bio-Rad, Hercules, CA). Protein in the gels was transferred to polyvinylidene difluoride (PVDF) membranes (Immobilon-FL, Millipore, Billerica, MA). Membranes were then incubated with primary antibody against phosphorylated-Akt (Ser473, Cell Signaling, Danvers MA, USA, catalog number 4060) and anti-rabbit fluorescence secondary antibody. After detection of the band corresponding to phosphorylated-Akt, membranes were stripped using NewBlot^™ stripping buffer (LI-COR Biosciences, Lincoln, NE) according to manufacturer instructions. Membranes were then incubated with primary antibody against total Akt (Cell Signaling, catalog number 4691) and fluorescence secondary antibody. Odyssey® Infrared Imaging System (LI-COR Biosciences, Lincoln, NE) was used to detect the antibody fluorescence intensity. Band densities were analyzed using NIH Image J software.²⁷

Statistical analysis

Data are presented as mean ± standard deviation unless otherwise indicated. A Wilcoxon-signed rank test was used for the within-subject comparison of GSK2256294 versus placebo. Correlation between different outcomes was assessed by Spearman’s correlation coefficient. The T-test approach of Jones and Kenward²⁸ was used to test for a carry-over effect and no significantly differential carry-over effects were detected. For outcomes with multiple repeated measures during hyperinsulinemic clamp, generalized least squares models were fitted on drug and insulin, and compound symmetry correlation structure was used to account for within-subject correlation. To account for multiple comparisons, we set false discovery rate (FDR) at 0.05 using the FDR-controlling procedures by Benjamini and Hochberg within a group of endpoints.²⁹ A two-sided p-value <0.05 was considered to be the threshold for significance.

The study was powered to detect a difference in insulin sensitivity between treatment periods similar to the difference observed previously in carriers of the EPHX2 287Gln allele versus wild-type. Specifically, with 16 participants, the study was powered to have 86% power to detect a within-subject difference of 3.76±4.6 mg/kg/min per μU/mL*100 with α=0.05. All analyses were performed using R version ≥ 4.0.³⁰

Results

Participant characteristics

Sixteen obese, pre-diabetic individuals were randomized to treatment order (Figure 1). One participant dropped out after randomization but prior to treatment with study medication. Fifteen participants completed both study days. One participant completed all basal study day procedures but did not complete the second hyperinsulinemic clamp due to inadequate IV access during the high-dose insulin portion. Participant characteristics appear in Table 1.

Figure 1.

CONSORT Study enrollment flow-diagram.

Table 1:

Participant characteristics

Parameter	N or mean±SD
Gender (F:M)	14:2
Race/Ethnicity (W:B:>1 race)	11:3:2
Age (years)	46.5±8.6
Body mass index (kg/m2)	40.8±8.5
Fasting plasma glucose (mg/dL)	101.6±8.2
2-hr glucose (mg/dL)	164.9±32.3
HbA1c (%)	5.7±0.3
LDL Cholesterol (mg/dL)	112.6±20.7
HDL Cholesterol (mg/dL)	48.5±9.3
Triglycerides (mg/dL)	122.2±55.2
Body fat by DEXA (%)	49.2±6.3

Effect of GSK2256294 on sEH activity and EETs

Fasting measurements obtained prior to the hyperinsulinemic clamp during each treatment appear in Table 2. There was no difference in urinary sodium excretion or urinary sodium/potassium ratio. During treatment with GSK2256294, sEH activity was significantly decreased in plasma and adipose tissue, as measured by the conversion of 14,15-EET to 14,15-DHET (Table 3). In plasma, the conversion of 11,12-EET to 11,12-DHET also decreased during GSK2256294. In adipose tissue, the conversion of 12,13-EpOME to 12,13-DiHOME was significantly decreased during GSK2256294. Although only a subset of participants consented for muscle biopsy, muscle sEH activity for the conversion of 14,15-EET to 14,15-DHET was significantly decreased (P=0.03) during GSK2256294.

Table 2:

Hemodynamic and metabolic values during treatment with placebo and GSK2256294

Parameter	Placebo	GSK2256294	Δwithin-subject	P-value
Weight (kg)	109.4±25.6	108.8±25.8	−0.6±1.6	0.64
Systolic blood pressure (mmHg)	122.9±11.0	124.3±9.6	1.4±11.3	0.79
Diastolic blood pressure (mmHg)	77.4±6.8	78.9±8.7	1.5±8.9	0.79
Heart rate (bpm)	68.2±5.4	66.6±5.7	−1.5±5.9	0.64
QT duration (ms)	411.3±18.7	409.2±24.2	−2.1±19.8	0.73
Serum potassium (mEq/L)	3.9±0.4	3.9±0.4	−0.0±0.3	0.75
Serum sodium (mEq/L)	138.0±2.2	137.7±2.1	−0.3±1.8	0.73
Serum calcium (mmol/L)	9.3±0.2	9.2±0.5	−0.1±0.4	0.64
Serum Creatinine (mg/dL)	0.8±0.1	0.8±0.1	0.0±0.1	0.64
Fasting plasma glucose (mg/dL)	93.1±6.2	96.2±10.9	3.1±6.8	0.64
Fasting insulin (μU/mL)	24.2±14.3	26.6±20.4	2.6±10.6	0.79
Forearm blood flow (mL/min/100mL)	2.5±2.0	2.0±0.7	−0.5±1.8	0.73
Forearm vascular resistance (mL/min/mmHg per 100mL)	50.8±27.9	51.4±18.9	0.6±29.9	0.89
Urinary sodium excretion (mmol/24 hrs)	199.8±144.0	204.3±204.8	−8.1±100.9	0.94
Urinary potassium excretion (mmol/24 hrs)	64.8±62.8	73.6±90.0	6.2±52.6	0.83
Urinary sodium/potassium ratio	4.0±2.7	3.4±1.8	−0.9±2.6	0.71
Renal plasma flow (mL/min/1.73 m2) *	648.3±138.2	613.8±103.0	−34.5±107.5	0.88
Plasma IL-6 (pmol/mL)	1.48±0.61	1.48±0.69	4.7±40.5	0.79
Plasma VEGF (pg/mL)	31.9±18.6	29.0±24.4	−5.5±64.6	0.64

Results are mean±SD

^*N=8 for measurement of renal plasma flow due to delays in access to para-aminohippurate. P-value versus placebo using Wilcoxon signed rank and BH multiple comparison adjustment.

Table 3:

Soluble epoxide hydrolase activity during treatment with placebo and GSK2256294

Product	Placebo	GSK2256294	Within-subjectΔ%
Plasma (n=15)
14,15-DHET	4.08±4.53	1.93±2.62*	−44.3±47.4%
11,12-DHET	2.64±3.20	2.23±2.87*	−19.6±20.4%
8,9-DHET	4.04±9.54	3.95±8.06	33.9±121.8%
12,13-DiHOME	2.54±1.66	2.25±1.86	−7.7±62.2%
9,10-DiHOME	5.03±3.50	7.64±11.14	98.3±390.1%
Adipose (n=12)
14,15-DHET	2175.93±964.79	1279.85±675.15†	−42.8±21.4%
11,12-DHET	749.43±269.88	617.59±198.76	−14.9±38.7%
8,9-DHET	649.64±272.01	702.65±306.41	0.9±23.7%
12,13-DiHOME	1245.81±448.01	575.81±410.29†	−58.0±20.4%
9,10-DiHOME	837.50±301.00	824.33±447.00	12.9±70.4%
Muscle (n=7)
14,15-DHET	3.53±1.60	1.94±1.13	−41.3±28.0%
11,12-DHET	1.61±0.54	1.29±0.92	−24.5±31.6%
8,9-DHET	1.26±0.19	1.57±0.78	28.1±75.7%
12,13-DiHOME	2.11±1.26	1.17±0.57	−34.4±27.8%
9,10-DiHOME	1.64±0.75	1.48±0.97	−14.1±21.5%

Soluble epoxide hydrolase (sEH) activity was measured as the formation of the DHET or DiHOME from the added EET or EpOME substrate. sEH activity is expressed pmol/mL/hr for plasma and pmol/mg/hr for adipose and muscle tissue.

^*P≤0.05,

^†P≤0.01 versus placebo using Wilcoxon signed rank and BH multiple comparison adjustment.

Even though GSK2256294 significantly reduced plasma and adipose sEH activity, plasma and adipose EET concentrations were similar during GSK2256294 and placebo treatment periods (Table 4). DHET concentrations in plasma and adipose tissue were also unchanged during treatment (P=0.067 for each). Muscle DHET and EET concentrations were not measured due to limited tissue samples. There was no difference in the ratio of plasma or adipose 11,12-DiHOME:(DiHOME+EpOME) or 14,15-DHET:(DHET+EET) ratios (Figure S1 and Table S1). The plasma 11,12-DHET:(DHET+EET) ratio decreased during GSK2256294 (0.40±0.12 vs 0.36±0.18, P=0.049), but no difference was observed in adipose tissue (0.33±0.17 vs 0.32±0.11, P=0.65).

Table 4:

Fasting plasma and adipose EET and DHET concentrations during treatment with placebo and GSK2256294

EET	Placebo	GSK2256294	Within-subjectΔ
Plasma
14,15-EET	6.69±2.07	7.31±2.91	0.21±4.69
11,12-EET	5.94±2.82	6.03±3.05	0.02±2.20
8,9-EET	8.78±5.49	9.06±3.87	0.66±1.79
Total EETs	21.41±8.43	22.39±9.31	0.89±7.40
14,15-DHET	1.90±2.23	1.65±1.38	−0.29±0.98
11,12-DHET	3.90±1.59	3.59±2.32	−0.28±1.43
8,9-DHET	4.01±1.53	3.98±1.35	0.03±0.80
Total DHETs	9.80±3.28	9.22±2.85	−0.53±2.34
Adipose
14,15-EET	50.33±51.00	25.32±13.67	−19.40±33.54
11,12-EET	35.28±52.57	14.36±8.49	−3.24±17.13
8,9-EET	109.40±208.76	37.41±22.45	2.20±36.60
Total EETs	175.73±290.64	66.08±43.99	−44.58±104.06
14,15-DHET	6.00±6.71	3.43±1.47	−0.87±2.33
11,12-DHET	13.11±22.00	6.39±2.73	−1.46±4.55
8,9-DHET	17.42±32.85	7.13±3.41	−1.94±3.04
Total DHETs	36.52±61.39	16.46±6.82	−4.67±8.74

Units are pmol/mL for plasma and pmol/mg tissue for adipose tissue. No comparisons were significant after BH multiple testing adjustment.

Relationship among sEH activity and EET concentrations measured in plasma, adipose and muscle

The activity of sEH for the conversion of 14,15-EET to 14,15-DHET in plasma correlated with sEH activity as measured by the conversion of 14,15-EET to 14,15-DHET and 12,13-EpOME to 12,13-DiHOME in both adipose and muscle tissue (Table S2). The highest correlation was observed for muscle sEH activity and plasma sEH activity conversion of 14,15-EET to 14,15-DHET (Table S2). Insulin-stimulated 14,15-EET, 11-12-EET, and total EET concentrations in plasma correlated significantly with basal 11,12-EET concentrations in adipose tissue (Table S3). Basal 8,9-EET and total EET-concentrations, as well as insulin-stimulated 14,15-EET, 11,12-EET, 8,9-EET, and total EET concentrations in plasma correlated significantly with basal 8,9-EET concentrations in adipose tissue. There was no effect of treatment order on plasma or adipose sEH activity as measured by 14,15-EET hydrolysis (Figure S2).

Effect of GSK2256294 on insulin sensitivity and basal and stimulated insulin signaling

As indicated in Figure 2, glucose concentrations were maintained at similar concentrations during GSK2256294 and placebo treatment throughout the euglycemic-hyperinsulinemic clamp. Achieved insulin concentrations were also similar. There was no difference in glucose infusion rate (GIR), a measure of insulin sensitivity, during either low-dose or high-dose insulin infusion between the GSK2256294 and placebo study days, suggesting no difference in global insulin sensitivity during supraphysiologic concentrations of insulin. The insulin sensitivity index measured during high-dose insulin correlated significantly with 11,12-EET during placebo but not during GSK.

Figure 2. Effect of sEH inhibition on insulin sensitivity.

A) Glucose concentration was maintained near a target of 95 mg/dL during Placebo (●) and GSK2256294 (▼). Grey shading indicates insulin infusion rates during low and high dose insulin (in mU/m2/min). B) The glucose infusion rate normalized to fat-free mass [FFM] (GIR) and C) insulin concentrations were similar during euglycemic-hyperinsulinemic clamp after one-week treatment with placebo or GSK2256294. Data are mean±SEM. *P<0.05, ***P≤0.001.

To assess insulin signaling in muscle and adipose tissue, we measured the ratio between pAkt and Akt expression before and after insulin clamp (representative Western blots shown in Figure S3). In muscle, the basal pAkt/Akt ratio was similar (Figure 2, 0.23±0.21 and 0.32±0.27 AU during placebo and GSK2256294 treatment, respectively, P=0.73) and increased similarly after insulin infusion during placebo and GSK2256294 (difference of 0.03, 95% CI −0.17 to 0.23, P=0.77 for effect of GSK; P<0.001 for effect of insulin). In adipose tissue, the basal pAkt/Akt ratio was not statistically different (Figure 2, 0.14±0.20 vs 0.30±0.28 AU during placebo vs GSK2256294, P=0.17). The insulin-stimulated pAkt/Akt ratio in adipose tissue was similar (0.21±0.14 vs 0.19±0.18 during placebo and GSK2256294, P=0.78; P=0.58 for effect of insulin).

Effect of GSK2256294 on forearm and renal blood flow

Baseline FBF correlated significantly with 11,12-EETs (ρ=0.41, P=0.03) and total EETs (ρ=0.40, P=0.03) in plasma. There was no effect of GSK2256294 on basal FBF, FVR or RPF (Table 2). Insulin infusion significantly and similarly increased FBF and RPF (Figure S4) during placebo and GSK2256294.

Effect of GSK2256294 on oxidative stress, inflammation, and VEGF concentrations

Basal concentrations of circulating F2-isoprostanes were significantly lower during GSK2256294 treatment (50.7 ± 15.8 vs. 37.2 ± 17.3 pg/mL, P=0.03, Figure S5). During placebo, F2-isoprostanes decreased after insulin infusion (P<0.001) while no further reduction was observed during GSK2256294 (P=0.34), so that F2-isoprostane concentrations after insulin were similar during GSK2256294 and placebo treatment (31.6 ± 9.7 vs. 34.6 ± 10.2 pg/mL, P=0.24). IL-6 concentrations were similar during GSK2256294 and placebo (Table 2). Circulating VEGF concentrations were also similar during GSK2256294 and placebo.

Safety

GSK2256294 was well-tolerated; adverse events appear in Table S4. There were no laboratory adverse events. There was no effect of drug on electrolytes, kidney function, or electrocardiographic QT interval.

Discussion

This study investigated the effect of pharmacologic sEH inhibition on blood pressure, oxidative stress, inflammation, and insulin sensitivity in patients with obesity and pre-diabetes. The results indicate that one-week treatment with GSK2256294 significantly decreases sEH activity in plasma, muscle, and adipose tissue but does not alter circulating EET concentrations. Plasma EET concentrations correlate with tissue EET concentrations and with FBF. Treatment with GSK2256294 decreases oxidative stress as measured by circulating F₂-isoprostanes. In contrast, one-week treatment with GSK2256294 does not alter blood pressure or insulin sensitivity as assessed using a euglycemic-hyperinsulinemic clamp.

This study is the first to report circulating EET and DHET concentrations during pharmacologic sEH inhibition in humans. Chen et al reported that urinary excretion of DHETs and DiHOMES was decreased following one-week treatment with the sEH inhibitor AR9821, but the authors did not detect urine EETs and did not report plasma EET or DHET concentrations.⁸ Lazaar et al reported that single doses of GSK2256294 inhibited sEH activity (specifically the hydrolysis of 14,15-EET-d11 to 14,15-DHET-d11) from an average of 41.9% (2 mg) to an average of 99.8% (20 mg) for up to 24 hours.⁹ Fourteen-day dosing with either 6 mg or 18 mg resulted in near maximal inhibition of sEH.⁹ The authors did not report plasma EET concentrations, however. We also found that GSK2256294 significantly decreased plasma sEH activity, based on the hydrolysis of 14,15-EET; nevertheless, EET concentrations were not increased during sEH inhibition by GSK2256294. EET concentrations reflect the balance of formation by CYP2C9 and CYP2J2 and hydration and inactivation by sEH. The finding that EET concentrations did not increase when sEH was inhibited suggest that there may be compensatory changes in EET production.

There was no effect of sEH inhibition on basal insulin signaling in muscle or on insulin sensitivity as measured by euglycemic-hyperinsulinemic clamp. While there was a trend toward increased basal insulin signaling (pAkt/Akt ratio in the setting of similar fasting insulin concentrations) in adipose tissue during treatment with GSK2256294, this was not significant. In in vitro studies, EETs increase insulin-stimulated AKT phosphorylation, and in cultured mesenchymal stem cells this effect is enhanced during sEH inhibition.^31–33 Studies in rodents have demonstrated that sEH inhibition increases the vasodilatory effect of insulin and microvascular blood flow, which has been proposed as a mechanism by which EETs increase insulin sensitivity.¹⁴ We observed no effect of GSK2256294 on stimulated insulin sensitivity or muscle blood flow determined by plethysmography, however. This in turn may reflect the lack of increase of circulating EET concentrations during GSK2256294 treatment. Alternatively, the duration of treatment may have been insufficient to observe an effect of sEH inhibition. In this regard, the previously reported association between a loss-of-function genetic variant in EPHX2 and insulin sensitivity may reflect the impact of prolonged decreases in sEH activity.²¹ Interventions which alter insulin sensitivity can act as quickly after one day of metformin³⁴ or after one week of dietary modification,^{35, 36} although a longer duration of treatment may be required for less potent agents. Although most antihypertensive agents reduce blood pressure over the course of hours to days, the blood pressure in our population was only mildly elevated at enrollment which could restrict the effects of sEH inhibition on this measure.

To the best of our knowledge this is the first study to report EETs, DHETs, and sEH activity in human adipose tissues. De Taeye et al reported sEH expression in adipose tissue in mice. High fat intake did not alter sEH activity normalized per tissue weight but dramatically increased total sEH activity attributable to fat.¹⁹ In obese humans, we found that GSK2256294 significantly reduced sEH activity in adipose tissue, and that sEH activity in these tissues correlated with sEH activity in plasma. Although the reduction of muscle sEH activity did not remain significant after multiple testing adjustment, we observed the strongest correlation between muscle and plasma sEH activity, despite a smaller sample size than in adipose. These results support the use of plasma measurements of sEH as a correlate of tissue sEH activity during pharmacologic inhibition. Future studies should determine whether genetic variation in EPHX2 is associated with sEH activity in adipose and muscle tissues. In addition, EET concentrations measured in adipose tissue correlated with EET concentrations measured following the hyperinsulinemic clamp, suggesting that adipose tissue may provide a source of circulating EETs, a hypothesis that merits further investigation.

This is the first study to demonstrate that sEH inhibition reduces circulating F2-isoprostanes, which are free radical-induced peroxidation products of arachidonic acid. F2-isoprostanes are increased in patients with obesity, diabetes, and smoking, and are reduced after treatment or reversal of these conditions.^37–39 Increasing EETs by exogenous administration or by sEH inhibition decreases oxidative stress in vitro and in vivo in animal models, as assessed by expression of inflammatory cytokines (e.g., IL-1β, IL-6, and VCAM-1) and signaling pathway activation (e.g., IκB).^{4, 5, 40} We found that sEH inhibition did not decrease plasma IL-6 but significantly reduced circulating F2-isoprostanes independent of changes in weight, blood pressure, or glycemia. This effect of sEH on F2-isoprostanes is previously unexplored and suggests this class of medications could reduce oxidative stress in humans. The present study is also the first to demonstrate that insulin administration during euglycemic-hyperinsulinemic clamp decrease F2-isoprostanes. Reduction of F2-isoprostanes in patients during treatment of metabolic dysfunction could provide a biomarker for cardiovascular risk reduction, although this remains to be proven.

In summary, we report for the first time the effect of pharmacological sEH inhibition on tissue sEH activity, as well as plasma sEH activity, and on tissue and plasma EET concentrations. We confirm the inhibition of circulating and adipose sEH after one-week therapy with GSK2256294. F2-isoprostanes, markers of oxidative stress, are also decreased by sEH inhibition. Pharmacological sEH inhibition does not alter blood flow or sensitivity to insulin. Pharmacologic sEH inhibition with GSK2256294 does not alter circulating or tissue EET concentrations, which could explain these findings. Additional studies are needed to understand the regulation of EET concentrations when sEH is inhibited.

Perspectives.

EETs are endothelium-derived relaxing factors which have been previously demonstrated to improve insulin sensitivity in animal studies. Inhibition of sEH prevents hydrolysis of EETs to their less active diols (DHETs), and this new drug class has been proposed as a treatment for hypertension, insulin sensitivity, and inflammation. The current study demonstrates that GSK2256294 administration for one week effectively inhibits sEH activity in plasma, muscle, and adipose tissue and reduces the F2-isoprostane oxidative stress marker but does not improve insulin sensitivity or reduce blood pressure. Plasma, adipose and muscle tissue measures of EETs and sEH activity correlated significantly, demonstrating that circulating measures may reflect tissue concentrations in vivo.

EETs are important mediators of vascular reactivity, salt sensitivity and hypertension and targeting this pathway may improve blood pressure and insulin sensitivity in the future. Although blood pressure and insulin sensitivity were not affected in this study with one week treatment, longer term effects cannot be excluded. The positive effects of sEH inhibition on oxidative stress could provide support for a longer treatment in the future.

Novelty and Significance.

What is new

• In this randomized cross-over study, we examined the effects of sEH inhibition with GSK2256294 for seven days on plasma, adipose, and muscle measures of sEH activity, insulin sensitivity, and inflammation in humans with pre-diabetes.

• GSK2256294 inhibits sEH activity in plasma, muscle, and adipose tissue and reduces plasma F2-isoprostanes, a marker of oxidative stress, but does not improve insulin sensitivity or reduce blood pressure.

• Plasma, adipose and muscle tissue measures of EETs and sEH activity correlated significantly, demonstrating that circulating measures may reflect tissue concentrations in vivo.

What is relevant

• EETs are important mediators of vascular reactivity, salt sensitivity and hypertension in animal models and targeting this pathway may improve blood pressure and insulin sensitivity. These effects have not been previously studied using an sEH inhibitor in humans.

Summary

Pharmacological sEH inhibition for one week decreased sEH activity without increasing EETs or improving blood pressure or insulin sensitivity. sEH inhibition decreased oxidative stress, however, suggesting that a longer treatment study may be warranted.