A New Synthetic Analogue of Protectin DX With Enhanced Therapeutic Potential Against Metabolic‐Associated Steatotic Liver Disease

ABSTRACT

Obesity, characterized by chronic low‐grade inflammation, promotes numerous complications such as type 2 diabetes (T2D) and metabolic dysfunction‐associated steatotic liver disease (MASLD). A class of lipid mediators known as specialized pro‐resolving mediators (SPMs) has garnered interest in this field due to their capacity to promote the resolution of inflammation. One such SPM is Protectin DX (PDX), the stereoisomer of Protectin D1 (PD1). We previously reported that PDX treatment protects against lipid‐induced and obesity‐linked insulin resistance and attenuates end‐stage renal failure in T2D animal models. Our group recently developed a cost‐efficient synthesis of PDX and structural analogues to accelerate research on PDX functions and to scale up the production of these molecules to facilitate their pharmaceutical development. After synthesizing and screening over 30 PDX analogues for their bioactivity in relevant cellular models, two analogues, AN‐44 and AN‐48, were selected for their ability to reduce macrophage inflammation and stimulate muscle glucose uptake in vitro. Since AN‐48 also lowers plasma TNF‐α in a hamster model of metabolic endotoxemia, it was selected for longer‐term in vivo studies. AN‐48 (50 ng/g) administered orally daily was found to fully prevent hepatic triglyceride accretion in diet‐induced obese hamsters. AN‐48 also prevented fasting hyperinsulinemia, insulin resistance, and reduced hepatic inflammation as compared to vehicle or PDX treatments. These results identify AN‐48 as a cost‐efficient and novel PDX analogue with high therapeutic potential against obesity‐linked T2D and MASLD.

The pro‐resolving lipid mediator PDX, and a novel synthetic analogue AN‐48, were orally administered (50 ng/g) daily to hamsters fed an obesity‐inducing diet. AN‐48 prevented hepatic triglyceride accretion and reduced the expression of fatty acid synthase (FASN), responsible for de novo lipogenesis. AN‐48 also prevented fasting hyperinsulinemia, insulin resistance, and reduced hepatic inflammation as compared to vehicle or PDX treatments. These results identify AN‐48 as a cost‐efficient and novel PDX analogue with high therapeutic potential against obesity‐linked T2D and MASLD.

1. Introduction

Obesity is an important risk factor for the development of chronic low‐grade inflammation in which inflammatory resolution mechanisms are defective [1, 2, 3]. This metabolic inflammation is an important factor in insulin resistance and promotes the development of metabolic‐associated steatotic liver disease (MASLD) and atherosclerosis [3]. Omega‐3 fatty acids supplementation in humans has been linked to a reduced risk of liver disease, including MASLD [4, 5]. There is, however, a lot of heterogeneity in the response across the literature. A meta‐analysis of recent clinical studies concluded that the impact of omega‐3 fatty acids on liver fat content remains inconclusive, but their effect on liver function and metabolic health (e.g., liver enzymes, serum lipids, insulin sensitivity, inflammatory markers) overall was beneficial [6]. The production of anti‐inflammatory metabolites derived from omega‐3 fatty acids, called specialized pro‐resolving mediators (SPM), is defective in obesity [1, 2, 7]. SPM production is essential for the return to homeostasis [1, 2, 3]. PDX, derived from the omega‐3 fatty acid docosahexaenoic (DHA), is particularly interesting for its potential glucoregulatory and hepatoprotective effects. In an acute mouse model of insulin resistance induced by lipid infusion, we previously showed that PDX acts on skeletal muscle to stimulate the synthesis of the myokine IL‐6 [8]. This activates a myokine‐liver axis resulting in a reduction of hepatic glucose production through STAT‐3‐mediated inhibition of gluconeogenesis and improvement of insulin sensitivity [8]. PDX also has IL‐6‐independent anti‐inflammatory and glucoregulatory functions, as demonstrated by the persistence of some of these effects in IL‐6 knockout mice [8]. Short‐term administration of PDX to obese db/db mice also reduced insulin resistance without affecting adipose tissue inflammation in this model of severe diabetes [8].

We also previously reported that PDX can mitigate end‐stage renal disease (ESRD) in high‐fat/high‐sucrose (HFHS)‐fed LDLr^−/−/ApoB^100/100 diabetic mice by alleviating kidney fibrosis and mesangial expansion [9]. Another group has also reported the efficacy of PDX in preventing lipid accumulation in HepG2 cells and limiting hepatic fat deposition in high fat‐fed C57BL/6J mice [10]. These results suggest that PDX can mitigate T2D and related liver (MASLD) and renal (ESRD) complications. However, PDX has some limitations as a potential therapeutic agent due to its complex and costly chemical synthesis and weak chemical/thermal stability as well as a short in vivo half‐life.

To circumvent these limitations, our group recently developed a new method for the total synthesis of high‐quality PDX [11] leading at the same time to the generation of a series of PDX synthetic analogues [12] that were screened for their bioactivity to blunt LPS‐mediated inflammation in macrophages and to promote glucose uptake in muscle cells. Two analogues were selected for further evaluation using in vivo models of metabolic endotoxemia and diet‐induced obesity and MASLD.

2. Methods

2.1. Synthesis of PDX and Analogues

PDX and its analogues AN‐44 and AN‐48 were synthesized following our previously published methods [11, 12, 13, 14]. The molecular structures of the analogues, AN‐48 and AN‐44, plus the AN‐48 metabolite AN‐43 and their natural counterparts (PDX and PD1) are shown in Table S1.

2.2. In Vitro Studies

2.2.1. Glucose Uptake in Myocytes

C2C12 mouse myoblasts (RRID:CVCL_0188) were maintained in DMEM containing 10% FBS. Myotube differentiation was initiated by replacing the media with DMEM containing 2% horse serum. All experiments were conducted 7 days after the addition of the differentiation medium. Cells were serum deprived for 3 h and then treated for a further 2 h with PDX, PD1, or the analogues, with the addition of insulin (10 nM) or vehicle 45 min before the experiment's conclusion. Briefly, cells were incubated in transport solution [140 mM NaCl, 20 mM HEPES‐Na, 2.5 mM MgSO₄, 1 mM CaCl₂, 5 mM KCl, and 0.5 μCi/mL 2‐deoxy‐D‐[³H] glucose (pH 7.4)] for 8 min at room temperature. Cells were washed three times with ice‐cold saline solution and then lysed with 1 M KOH, and aliquots were transferred to scintillation vials for ³H radioactivity counting. Cellular glucose uptake activity was first calculated as pmol of glucose per minute per milligram protein and then converted to fold of vehicle.

2.3. Hepatic Glucose Production

FAO rat hepatocytes were grown and maintained in monolayer culture in Roswell Park Memorial Institute medium (RPMI) containing 10% FBS in an atmosphere of 5% CO₂ at 37°C. FAO hepatocytes were plated in 24‐well plates at 4 × 10⁶ cells/plate. After a 16 h serum deprivation (with or without insulin, 0.1 nmol), cells were washed 3 times with PBS. Cells were then incubated for 5 h (37°C, 5% CO₂) in the presence or absence of insulin (0.1 nmol) in a glucose production medium [glucose‐free DMEM containing 2 mmol sodium pyruvate, 20 mmol sodium L‐lactate, and sodium bicarbonate (3.7 g/L)] in which AN‐48 or AN‐43 was present. Glucose production was measured in the medium by using the Amplex Red Glucose/Glucose Oxidase Assay kit (Invitrogen). Hepatic glucose production results are presented as a percentage of suppression versus the untreated control.

2.4. iNOS Inhibition in Macrophages

The accumulation of nitrite in macrophage cell medium exposed to LPS was used as an index of NO production by inducible nitric oxide synthase (iNOS). J774A.1 mouse macrophages (RRID:CVCL_0358) were maintained in RPMI (10% bovine growth serum (BGS)) until 80% confluence and distributed in 96‐well plates. Macrophages were activated by the addition of 5 ng/mL LPS (Millipore Sigma, E. coli 055:B5, L2880) in the media with simultaneous addition of PDX, PD1, or the PDX analogues overnight (20 h). Nitrates were measured by Griess assay [15] and normalized on cell lysate protein concentration (PierceTM BCA Protein Assay kit, 23225).

2.5. PPARγ Response Element Luciferase Reporter Assay

HEK293T cells (RRID: CVCL_0063) were maintained in DMEM (10% FBS). HEK293T cells were co‐transfected with plasmids containing PPARγ response element‐luciferase (pGL3‐PPRE) and a second vector exogenously expressing mouse Pparγ2 (pSV‐SPORT1‐Pparg2, kindly provided by Dr. F. Picard CRIUCPQ‐UL, Québec, QC) [16] or an empty vector and a final expression vector containing β‐Gal. 32 h after transfection, vehicle (ethanol), rosiglitazone maleate, PDX, PD1, or PDX analogues was added to the cells for 16 h. Cell lysates were corrected for protein content and analyzed for luciferase activity and normalized on β‐Gal activity.

2.6. Animals

2.6.1. Pharmacokinetic Studies

Five to 6 week‐old female hamsters were purchased from Charles‐River Inc. (St‐Constant, Qc., Canada). Hamsters were individually housed (12‐h light–dark cycle) and fed a chow diet (2018SX, Envigo, USA), and water ad libitum. On day 1 of the study, hamsters were randomly assigned to one of three groups according to body weight.

2.6.2. Metabolic Endotoxemia and Metabolic Study

Eight‐week‐old male and female golden Syrian hamsters ( Mesocricetus auratus ) were purchased from Envigo and housed individually upon arrival with ad libitum access to food and water and kept in a 12‐h light–dark cycle. All in vivo evaluations and non‐terminal samples were performed/collected during daylight. All animal procedures were approved in accordance with the Université Laval Institutional Animal Care and User Committee (CPAUL).

2.7. Interventions and Therapeutic Treatments

2.7.1. Pharmacokinetic Studies

AN‐48 was dissolved in DMSO, followed by the addition of propylene glycol (vehicle 92:8 propylene glycol:DMSO). AN‐48 was intraperitoneally (IP) injected at assigned concentrations (5, 25, and 50 mg/kg of body weight). Blood samples collected by lateral saphenous vein were drawn for determination of AN‐48 and its metabolite AN‐43 at various time points (Figure S1). At the end of the experiment, the hamsters, under isoflurane anesthesia, were sacrificed by asphyxiation using CO₂. Blood samples were collected in potassium‐EDTA‐coated tubes (Sarstedt, Montreal, Canada) and centrifuged at 3200 rpm for 10 min at 4°C. The plasma was collected and stored at −80°C for further analysis.

2.8. Acute Metabolic Endotoxemia Study

Following a 1‐week acclimation period, animals were exposed to a high fat high fructose mixed protein diet (HFFMP) (D12079, Ssniff, Germany) for 3 weeks before treatment (Figure S2C). Four males and four females were randomly assigned to either control, AN‐44, or AN‐48 groups (Figure S2C). Hamsters were fasted 12 h before first receiving a single IP dose of analogue (50 μg/g) or vehicle (92:8 Propylene glycol: DMSO), followed 3 min later by a second IP injection containing 0.6 μg/g LPS ( E. coli 0111:B4, Invivogen, #cat: tlrl‐3pelps) in saline. Blood glucose was measured from the saphenous vein using a handheld glucometer (Accucheck) prior to the injections, as well as 30 min, 1, 2, and 3 h after. Euthanasia was carried out under isoflurane by cardiac puncture and decapitation. Tissues were weighed and snap frozen in liquid nitrogen.

2.9. Chronic Metabolic Study in HFFMP‐Fed Hamsters

Hamsters (84 in total) were randomly assigned to either the chow reference group (Chow) (2018 Teklad Global 18% Protein Rodent Diet, Envigo) (6 males, 6 females) or a HFFMP group (Figure 3A). An intention‐to‐treat protocol was used where hamsters were pre‐fed the HFFMP diet for 2 weeks before random assignment to a treatment group. For optimal group allocation, we randomly distributed matched sets of animals [17] using initial weight, weight gained during the pre‐feeding period, and initial fasting blood glucose. Treatment groups remained on the HFFMP diet and were orally gavaged daily with either vehicle (corn oil), PDX (50 ng/g), or AN‐48 (50 ng/g). Chow‐fed hamsters were kept on this diet and used as a healthy reference group throughout the study. Hamsters were fasted 6 h before every blood collection or oral glucose tolerance test (oGTT). On day 32 (End of treatment, EOT), after a 6‐h fasting period, animals received either a single dose of insulin (10 U/kg, IP) or saline (1/3 of the animals) 5 min before euthanasia. Animals received their daily oral dose of either vehicle, PDX, or AN‐48 2 h before the euthanasia. Euthanasia was carried out by cardiac puncture under isoflurane general anesthesia followed by decapitation.

FIGURE 3.

Systemic and hepatic effects of PDX and AN‐48 on HFFMP‐fed hamsters. (A) Study design. (B) Liver transaminases concentrations at EOT. (C) Food intake over the 32 days of treatment. (D) Body weight and (E) Liver weight at EOT (F) Liver triglycerides, (G) Liver cholesterol and (H) Liver glycogen at EOT. All data are mean ± SEM, ##p < 0.01, ###p < 0.001 vs. chow and **p < 0.01 vs. vehicle (N = 23–24).

2.10. Oral Glucose Tolerance Test and Calculations

Animals undergoing the metabolic study were subjected to an oGTT on Day 28. Hamsters were fasted 6 h the morning of the experiment. They received their daily oral dose of either vehicle, PDX, or AN‐48 2 h before the oGTT. Blood glucose from the saphenous vein was measured using a handheld glucometer (Accucheck) prior to the oral administration of a D‐glucose bolus (2 mg/g BW) as well as 15, 30, 60, and 120 min after. Glucose excursion was calculated as the area under the curve (AUC), and the HOMA‐IR was utilized as a measure of insulin resistance [18].

2.11. Lipid Analysis

Liver triglycerides and cholesterol were extracted with chloroform‐methanol [19]. Triglycerides and cholesterol were then quantified using the Infinity Triglycerides (cat# TR22421, Thermo Fisher Scientific) and RANDOX Total Cholesterol Assay (cat # CH200, Randox Laboratories, United Kingdom) kits, respectively.

2.12. Glycogen Determination

Hepatic glycogen concentration was determined according to Rasouli et al. [20]. Briefly, 30–40 mg of liver tissue was boiled in 200 μL 30% KOH for 10 min and then cooled on ice for another 10 min. 275 μL 100% EtOH was added to the sample and vortexed, followed by centrifugation for 10 min at 1700 g . After drying down, the sample was resuspended in distilled water and vortexed until homogenous. To measure glycogen, 50 μL of samples was added to phenol/H₂SO₄/water solution (10:50:5), incubated at room temperature for 30 min, and read at 490 nM.

2.13. Immunoblotting

Liver tissues from hamsters were pulverized in liquid nitrogen and then homogenized in complete RIPA buffer using a VWR bead mill homogenizer. Immunoblotting of lysates was then performed [1]. Briefly, 7–20 μg of protein were loaded on a 7.5% acrylamide gel, subjected to SDS‐PAGE, and then transferred onto nitrocellulose membranes. Membranes were exposed to primary antibodies overnight at 4°C. Primary antibodies for AKT (RRID:AB_329827), p‐AKT Ser473 (RRID:AB_329825), p‐AKT Thr308 (RRID:AB_329828), eEF2 (RRID:AB_10693546, 1:4000), FASN (RRID:AB_2100798), NF‐κB p65 (RRID:AB_330561), p‐NF‐κB p65(Ser536) (RRID:AB_331284), STAT3 (RRID:AB_331588), and p‐STAT3 Ser727 (RRID:AB_331589) were purchased from Cell Signaling Technology, CPT1A (RRID:AB_2864319) from Abcam, and Anti‐Actin clone 4 (RRID:AB_2223041, 1:40 000) from Millipore Sigma. Unless specified otherwise, primary antibodies were diluted at a concentration of 1:2000 in 5% BSA and 0.02% sodium azide. Goat‐Anti‐Rabbit IgG‐HRP (RRID:AB_2307391) and Goat anti‐Mouse IgG‐HRP (RRID:AB_2307392) antibodies were purchased from Jackson ImmunoResearch Labs. Secondary antibodies were diluted at 1:10 000 in either 2% BSA when used with antibodies targeting phosphorylated residues or 5% milk for total protein.

2.14. Quantification Method of Analogues and Metabolite

25 μL of plasma was diluted in 2 mL of water containing 0.1% formic acid and 0.005% butylated hydroxytoluene (BHT; Sigma, St. Louis, MO), followed by the addition of 50 μL of internal standard (leukotriene B4‐d4; LTB4‐d4, 50 ng/mL; and Resolvin D2‐d5, RvD2‐d5, 20 ng/mL; Cayman chemicals, Michigan, USA). Solid phase extraction (SPE) Strata‐X 60 mg columns (Phenomenex, Torrance, CA, USA) were conditioned with 2 × 1 mL methanol and 2 × 1 mL water 0.1% formic acid. Samples were added to the columns and successively washed with 2 × 1 mL of H₂O and 2 × 1 mL of water: methanol (80:20) 0.1% formic acid. Analytes were eluted with 2 × 1 mL of methyl formate (Sigma, St. Louis, MO). The elutes were completely evaporated at 25°C under N2 and reconstituted in 100 μL of water: MeOH (50:50). The same procedure was also applied to the calibration standards which correspond to the dilution of 25 μL of the working solutions (0.5–5000 ng/mL in ethanol) in 25 μL of stripped plasma. The chromatographic separations were achieved with an ultra‐high pressure liquid chromatography (UHPLC) Nexera Separations Module (Shimadzu Scientific Instruments Inc., Columbia, MD, USA) using a 150 × 2.1 mm Poroshell 120 EC‐C18 column (2.7 μΜ particles) (Agilent, Santa Clara, CA) at 40°C. The flow rate was set at 0.3 mL/min and the following mobile phases were used: 3.5 mM ammonium formate in water 0.01% acetic acid (solvent A), and MeOH 0.01% acetic acid (solvent B; (WVR, Montreal, Quebec, Canada)). 10 μL was injected for the analysis of analogs AN‐43 (metabolite) and AN‐48. The following chromatographic program was used: initial conditions (57% B) were followed by a linear gradient to 59% B over the next 6 min and a second linear gradient to 80% B over 9 min; the column was then flushed with 100% B for 5 min and re‐equilibrated to the initial conditions for 7 min. 2 μL was injected for the analysis of analog AN‐44 and the following chromatographic program was used: initial conditions (45% B) were maintained for 2 min followed by a linear gradient to 70% B over the next 8 min; the column was then flushed with 100% B for 5 min and re‐equilibrated to the initial conditions for 7 min. All analytes were quantified by tandem mass spectrometry (MS/MS) using an API6500 instrument (Applied Biosystems, Concord, ON, Canada) in positive mode. The temperature was set at 300°C, the entrance potential (EP) at 10 V, and the declustering potential (DP) at 70 V. MS/MS parameters (ion transition (MRM), and collision energy (CE)) were as follows: AN‐43:404.2^369.1, 10 V; AN‐48: 418.3^365.2, 12 V; AN‐44: 266.3^213.0, 10 V; LTB4‐d4:358.2^305.2, 12 V; and RvD2‐d5: 399.2^364.1, 10 V.

2.15. ELISA

Hamster insulin (cat # 90336 Crystal Chem Elk Grove Village IL, USA) and hamster IL‐6 (MyBioSource San Diego CA, USA cat# MBS700950) ELISA kits. Briefly, blood was collected in EDTA tubes, centrifuged for 15 min at 1000 g at 4°C to obtain plasma. For insulin, 5 μL of plasma was utilized and the manufacturer's instructions were followed. For IL‐6 measurements, 100 μL of plasma was utilized and the manufacturer's instructions were followed.

2.16. Histology

At euthanasia, the liver was rapidly isolated and a section of the left lateral lobe was fixed in 4% paraformaldehyde for 24 h, then transferred to 70% ethanol. The tissue was paraffin embedded and 4 μm sections were stained with hematoxylin and eosin by the Department of Pathology, Quebec Heart and Lung Institute. Images of the H&E‐stained slides were captured using an Axio Observer Z1 microscope and ZEN lite software (Zeiss, Oberkochen, Germany).

2.17. Real‐Time Polymerase Chain Reaction

Livers were collected at euthanasia and were snap frozen. RNA was extracted from 20 mg of tissue via mechanical homogenization in TRizol reagent (Cat# 15596018, ThermoFisher Scientific) at 6 m/s for 15 s using a VWR beadmill. The RNA was then extracted with the Direct‐zol RNA Miniprep Plus (cat# R2072, Zymo Research). Subsequently, cDNA was prepared using 1 μg total RNA using a High‐capacity Reverse Transcriptase kit (ThermoFisher Applied Biosystems cat# 4368814) following the manufacturer's instructions. A Quant Studio 6 was used to perform the RT‐qPCR; the 2^(−ΔΔCT) method was used for data analysis.

Gene	Forward (5′‐3′)	Reverse (5′‐3′)
Actb	TCCCAGCACCATGAAGATCAA	GTAAAACGCAGCTCAGTAACAG
Hprt1	TCCCAGCGTCGTGATTAGTG	GTGATGGCCTCCCATCTCTTT
Rpl13a	TCCGAAAGCGGATGAACACC	CCTCGCTTAGTCTTGTGGGG
PPARa	GTGGCTGCTATAATTTGCTGTG	AGCTTCGGGAAGAGAAAGGTAT

2.18. Statistical Analysis

Using previously published data [8, 9] and unpublished pilot study data, we performed power calculations (MetaboAnalyst [21]) showing that at least 12 animals are required for each condition and each sex to detect > 25% effects of the treatments on the measured variables; therefore, each treatment was allocated 12 males and 12 females. A two‐tailed unpaired Student's t‐test was utilized to compare differences between the Chow and Vehicle (effect of diet). A Mann–Whitney U test was used when standard deviations were significantly different or when data were not normally distributed. Comparison between all HFFMP fed groups was performed using one‐way ANOVA (Factor: treatment) followed by Dunnett's multiple comparison test. Glycemia and insulinemia during oGTT, as well as variables measured at Baseline and EOT, were statistically compared using two‐way ANOVA with Bonferroni post hoc test (Factors: Time, treatment, and interaction). Data are expressed as mean ± SEM. All results were considered statistically significant at p‐values < 0.05. Statistical analysis was performed using GraphPad Prism 10.5.0 (RRID:SCR_002798).

3. Results

3.1. Identifying PDX Analogues With Metabolic and Anti‐Inflammatory Effects

Over 30 synthetic analogues of PDX (10S,17S‐dihydroxy‐4Z,7Z,11E,13Z,15E,19Z‐ docosahexaenoic acid) in which the key E, Z, E trienic configuration system was conserved were synthesized [11, 12]. Analogues were screened to evaluate their bioactivity in relevant cellular models and compared to PDX and its stereoisomer PD1 (10R,17S‐dihydroxy‐4Z,7Z,11E,13E,15Z,19Z‐ docosahexaenoic acid) (Table S1). Two analogues (AN‐44 and AN‐48, Table S1) were selected based on their combined ability to increase muscle glucose uptake and to inhibit iNOS‐mediated NO production, a well‐known inflammatory mediator of insulin resistance [8] Both AN‐44 and AN‐48 increased glucose uptake in C2C12 myocytes, both in the presence and absence of insulin (Figure 1A). The effect of the analogues was similar to or greater than that of the parent PDX molecule, while PD1 had no effect on glucose uptake. The stereoselective effect of PDX vs PD1 is in agreement with our published work [8]. We also tested the effect of the molecules on L6 myocytes, which lack the ability to release IL‐6, and found that only AN‐44 and AN‐48 significantly increased glucose uptake in these cells (Figure S2A). This indicates that the PDX analogues exert their effects on muscle glucose uptake independently from IL‐6.

FIGURE 1.

In vitro effects of PDX analogues. (A) Glucose uptake in C2C12 myocytes, 2 h after compound or vehicle treatment and with (gray bar) or without (black bars) the addition of 100 nM insulin. (B) NO production in LPS stimulated J774 cells measured by nitrite accumulation in the medium after 14 h treatment. (C) PPARγ response element luciferase reporter assay in HEK293T cells 14 h post treatment. (D) Plasma cytokines in hamsters 3 h post 0.6 μg/g LPS IP injection preceded by administration of 50 μg/g of PDX analogues or vehicle. Data for cell culture experiments (A–C) mean ± SEM N = 3–7 experiments (each performed in triplicate). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. vehicle. Data are mean ± SEM N = 3–5 for animal studies (D).

Both AN‐44 and AN‐48 markedly inhibited iNOS activity in LPS‐activated J774 macrophages, as shown by reduced nitric oxide production (Figure 1B). Maximum inhibition was calculated to reach 82.83% and 100% (IC₅₀ 20.53 μM and 57.79 μM) with AN‐44 and AN‐48, respectively. Both analogues were > 10 times more potent on iNOS as compared to their natural PD1/PDX counterparts.

There are limited options regarding anti‐diabetic drugs directly targeting insulin sensitivity in skeletal muscle and liver. Thiazolidinediones were promising in this regard, but concerns were raised due to cardiovascular and renal systems side effects. If the analogues were able to combine gluco‐regulatory, insulin sensitizing, and pro‐resolving properties, assessing PPARγ activation became essential [22]. PPARγ transcriptional activity was next tested (Figure 1C). Both rosiglitazone (thiazolidinedione, TZD) and 15d‐PGJ₂, well known PPARγ ligands, increased PPARγ activity in this luciferase reporter assay. DHA, PDX, and PD1 (p‐value = 0.052) also moderately activated PPARγ at 10 μM. However, neither AN‐44 nor AN‐48 increased PPARγ activity at this concentration.

Neither AN‐44 nor AN‐48 showed any cellular toxicity based on MTT assays in HEK293T cells (Figure S2B).

AN‐44 and AN‐48 were also tested for their activity to blunt inflammation in vivo using golden Syrian hamsters injected with a low dose of LPS to simulate metabolic endotoxemia and low‐grade inflammation present in obesity. AN‐44 and AN‐48 significantly reduced plasma TNF‐α by 36.6% ± 4.1% and 52.0% ± 3.9%, respectively, in LPS‐challenged female but not male hamsters (Figure 1D). The analogues did not reduce plasma IL‐6 or IL‐1β (Figure S2D).

3.2. Effect of AN‐48 Treatment on Liver and Metabolic Endpoints in Western‐Diet Fed Obese Hamsters

We next compared the therapeutic effect of AN‐48 to PDX in hamsters pre‐fed the HFFMP diet to induce metabolic alterations. After 2 weeks of HFFMP feeding, animals received daily oral administration of either 50 ng/g PDX, AN‐48, or an equivalent volume of vehicle for the following 32 days. PDX was detected in plasma and the liver after 32 days of treatment (Figure S1D,G). On the other hand, AN‐48 was not detectable in either the plasma or livers of AN‐48 treated animals (Figure S1E,H). Pharmacokinetic data conducted in hamsters (Figure S1A–C) revealed that greater than 99% of AN‐48 (methyl ester) is transformed into AN‐43 (free acid) 1 h post IP administration; therefore, we measured AN‐43 and found that it was detected in the liver of many AN‐48 treated animals (Figure S2I). AN‐48 is likely a prodrug of AN‐43 since we observed that the latter inhibits both basal and insulin‐suppressed glucose production with ~100‐fold more potency than the former in FAO hepatocytes (Figure 2).

FIGURE 2.

Suppression of hepatic glucose production by PDX analogues. Glucose production by FAO hepatocytes was measured after 16 h serum deprivation and with (right panel) or without (left panel) 0.1 nmol insulin. The percentage of inhibition of hepatic glucose production is presented versus untreated control. *p < 0.05, **p < 0.01 vs. control: #p < 0.05, ##p < 0.01, ###p < 0.001 vs. AN‐48.

Treatments with both PDX and AN‐48 were well tolerated, with no adverse effects observed during the 32‐day treatment period. Transaminase (ALT, AST) levels remained unchanged (Figure 3B), and no abnormalities were observed during necropsies.

When compared to the healthy reference, chow‐fed animals, HFFMP‐Vehicle feeding enhanced caloric intake with no effect on body weight gain but significantly augmented liver weights (Figure 3E). The higher liver weight was related to higher accumulation of both hepatic TG and cholesterol levels (Figure 3F,G). Treatment with PDX and AN‐48 did not change food intake, body weight gain, or liver weight (Figure 3C–E) compared to the HFFMP‐Vehicle group. However, AN‐48 significantly reduced liver TG levels in comparison to Vehicle (Figure 3F). PDX treatment showed a trend of decrease liver TG; however, it is not statistically significant from the vehicle group (p = 0.06). The level of steatosis found in liver sections is representative of the hepatic TG concentrations measured biochemically (Figure S3C). Neither PDX nor AN‐48 changed hepatic cholesterol levels (Figure 3G) or hepatic glycogen content (Figure 3H) between groups.

Concomitant with the decreased hepatic TG, liver FASN expression was decreased by both PDX and AN‐48 treatments (Figure 4A). Moreover, CPT1A expression was increased by PDX (Figure 4B). Given that CPT1A expression is under the control of PPARα, a known regulator of both de novo lipogenesis and beta oxidation, we measured its expression and found no difference between treatment groups (Figure S3D). Importantly, phosphorylation‐dependent activation of both NF‐kB and c‐Jun N‐terminal kinase (JNK) in the liver was decreased by AN‐48, but not by PDX, indicating that AN‐48 reduced hepatic inflammation in HFFMP‐fed hamsters as compared to vehicle‐treated animals (Figure 4C,D). We also observed that AN‐48 reduced inhibitory phosphorylation of IRS1 on Ser307 levels (Figure 4E). Acute insulin administration significantly enhanced Akt Ser473 phosphorylation in the liver of vehicle‐treated HFMMP‐fed hamsters (p < 0.05), but insulin‐induced Akt signaling was more robustly activated in PDX (p < 0.01) and AN‐48 (p < 0.001) treated HFFMP‐fed animals (Figure 4F), suggesting that hepatic insulin action was improved by PDX and AN‐48 treatment, in line with the blunted IRS‐1 inhibitory Ser307 phosphorylation in the liver of these animals. Hepatic STAT3 phosphorylation was also increased by PDX, but this effect was only significant in the liver of female but not male hamsters (Figure S3F). In contrast, AN‐48 treatment did not increase STAT3 phosphorylation in either female or male hamsters (Figure S3F).

FIGURE 4.

Effects of PDX and AN‐48 on key hepatic lipid metabolism and inflammatory markers, and insulin signaling in HFFMP‐fed hamsters. Liver relative protein levels for (A) FASN, (B) CPT1A, (C) phospho‐NF‐κB Ser536, (D) phospho‐JNK Thr183/Tyr185, (E) pIRS1 Ser307, and (F) pAKT Ser473. All data are mean ± SEM, *p < 0.05, **p < 0.0, ***p < 0.001 vs. vehicle (N = 23–24).

The effects of PDX and AN‐48 on whole‐body glucose homeostasis and insulin sensitivity of HFFMP‐fed hamsters were next determined. Hamsters fed the HFFMP diet showed increased fasting plasma glucose at the end of treatment (EOT) vs baseline and as compared to chow‐fed animals (Figure 5A). Fasting insulin levels EOT also increased as compared to baseline in HFFMP‐fed hamsters treated with vehicle (Figure 5B). Fasting hyperinsulinemia at EOT was prevented in animals treated with either PDX or AN‐48 as shown by no significant difference between baseline and EOT values in contrast to significantly increased values observed in the vehicle‐treated animals (Figure 5A,B). In addition, the HOMA‐IR insulin resistance index significantly increased in both the vehicle and PDX groups, but not in the AN‐48 treated animals (Figure 5C), suggesting that the PDX analogue prevented the development of hepatic insulin resistance. The effect of both PDX and AN‐48 on fasting hyperinsulinemia and HOMA‐IR tended to be greater in male hamsters (Table S2). Glucose intolerance was also observed in HFFMP‐fed animals (Figure 5D) with minimal effect on glucose‐stimulated insulin secretion (GSIS) (Figure 5E). Neither PDX nor AN‐48 treatments were found to change glucose and insulin responses during the oGTT.

FIGURE 5.

Effects of PDX and AN‐48 on whole‐body glucose homeostasis and insulin resistance in HFFMP‐fed hamsters. (A) FBG at Baseline and EOT. (B) FPI at Baseline and EOT. (C) HOMA‐IR index at Baseline and EOT. (D) Glycemia during OGTT with AUC in inset, and (E) insulin levels during OGTT with AUC in inset. All data are mean ± SEM, #p < 0.05, ##p < 0.01, vs chow and *p < 0.05. **p < 0.01 vs. vehicle (N = 23–24).

3.3. Sex‐Specific Effects of PDX and AN‐48 on Liver and Metabolic Endpoints in HFFMP‐Diet Fed Obese Hamsters

To determine potential sex‐specific effects of the treatments, we further analyzed the data separately for male and female hamsters (Table S2 and Figure S3A–G). Females had a significantly greater increase in body weight in response to the HFFMP diet (Table S2), but there were no differences between the treatment groups. This increased body weight was accompanied by an increase in food intake, but again treatment had no influence. Not surprisingly, males had significantly larger visceral adipose deposits (Table S2). We found that the effect of PDX and AN‐48 treatments on liver TG and TC was greater in female than in male hamsters (Table S2 and Figure S3A,B), which was also apparent by histological observations (Figure S3C). In contrast, male hamsters had significantly lower FBG at baseline that persisted until the EOT (Table S2).

4. Discussion

There is growing evidence that the beneficial effect of long chain omega‐3 fatty acids on inflammation and metabolic health is mediated at least in part through the production of protectins and other members of the SPM family [7, 10, 23, 24, 25, 26, 27]. We previously reported that PDX, a DHA‐derived SPM, alleviates insulin resistance by activating a myokine‐liver glucoregulatory axis. In that study, acute intravenous (IV) PDX administration prior to the start of a lipid infusion paired to a hyperinsulinemic‐euglycemic clamp in C57BL/6J mice improved insulin sensitivity and glycemic control [8]. Part of these effects resulted from an increase in muscle IL‐6 secretion to promote STAT3‐mediated transcriptional suppression of the hepatic gluconeogenic program. Some, but not all, of the metabolic effects of PDX were abrogated in IL‐6 null mice [8]. In the present study, insulin sensitivity was preserved in AN‐48 treated HFFMP‐fed hamsters despite that STAT3 phosphorylation levels were not increased by AN‐48. STAT3 phosphorylation was only augmented in PDX treated hamsters and only significantly in female hamsters. The modest effect of PDX on STAT3 phosphorylation as compared to our previous study might be explained by differences in animal models used (mice vs. hamsters), the severity of the models (lipid infusion, genetic obesity in mice vs. diet‐induced model of insulin resistance in the hamster). It may also be related to the duration of the treatment (acute vs. chronic PDX treatment) or the mode of administration (IV vs. gavage) of PDX in these studies. The observation that AN‐48 did not increase STAT3 phosphorylation in the liver of HFFMP‐fed hamsters suggests that the metabolic effect of the PDX analogue is not mediated by the IL‐6‐STAT3 axis. This is consistent with the finding that AN‐48 increases glucose uptake not only in C2C12 myocytes but also in L6 myocytes that do not secrete IL‐6.

In addition to its pro‐resolution and glucoregulatory effect [1, 8, 22], the current work shows that AN‐48 (and to a lesser extent PDX) reduces hepatic steatosis in the HFFMP hamster model in addition to preventing the deterioration of insulin sensitivity induced by the dietary treatment. Both PDX [10] and PD1 [28, 29] have been reported to improve liver parameters in rodent models of MASLD and hepatitis. More importantly, the PDX analogue AN‐48, but not PDX itself, was found to further blunt hepatic inflammation based on reduced activation of NF‐kB and JNK inflammatory pathways, which are well known to be involved in hepatic insulin resistance and MASLD development [30, 31, 32]. Taken together with in vitro data, these animal studies suggest that the PDX analogue AN‐48 has improved therapeutic potential against inflammation, T2D, and MASLD as compared to its natural counterpart.

It has been shown that PDX limits fatty acid accumulation in HepG2 cells and high‐fat fed C57BL/6J mice in association with decreased FASN expression [10]. In agreement with these studies, we found that both PDX as well as AN‐48 reduced FASN expression in HFFMP‐fed hamsters. Other SPMs such as PD1 and RvD1 have also been reported to ameliorate liver steatosis in mice and HepG2 cells, an effect that is thought to be mediated through blunting the IRAK–JNK pathway in hepatocytes [23, 24, 28]. In our hamster model of hepatic steatosis we also found that the PDX analogue AN‐48 inhibited JNK activation, suggesting this is a mechanism that may be common to many SPMs and more potently activated by the PDX analogue. This reduction of liver inflammation by AN‐48 likely explains the concomitant decrease in inhibitory Ser307 phosphorylation of IRS‐1 in the liver and the improved insulin sensitivity of AN‐48 treated animals. The effect of JNK activation on IRS‐1 serine phosphorylation and in mitigating insulin signaling in the liver is well documented [33].

We previously reported that both PDX and PD1 activate PPARγ, thus possibly contributing to their anti‐inflammatory and glucose sensitizing effects [22]. We thus determined whether PPARγ could be activated by the PDX analogues. Neither AN‐44 nor AN‐48 was found to activate PPARγ at the same concentrations required by PD1 and PDX to exert such effects. This suggests that PDX analogues are unlikely to induce their beneficial effects through PPARγ and that their metabolic and hepatic effects are mediated via other mechanisms.

More work will be necessary to further understand the mechanism by which AN‐48 exerts its action. Unlike other SPMs, the receptor involved in PDX's beneficial action remains unknown. Recent studies have revealed that the G protein‐coupled receptor GPR37 is essential for the pro‐resolution effects of PD1 [34, 35]. Whether GPR37 is also implicated in the biological effects of PDX, a stereoisomer of PD1, and of AN‐48, remains to be established. Future studies will also be needed to better understand how AN‐48 is metabolized as compared to PDX, which may also explain its increased potency against macrophage and liver inflammation. Initial PK studies revealed enrichment of plasma and liver PDX in HFFMP‐fed hamsters treated with PDX, but we were not able to detect AN‐48 in animals treated with the PDX analogue. As we suspected that AN‐48 may be metabolized by liver esterases, we conducted MS/MS studies to also detect the presence of a theoretical degradation product (named AN‐43), which was indeed detected in the livers of many AN‐48 treated animals and even in the plasma of a few animals. Since gavage of PDX and AN‐48 was given about 2 h before tissue collection, these data indicate that AN‐48 can reach the liver following oral administration but may be quickly metabolized to AN‐43, likely a more stable metabolite, which we have shown to be ~100‐fold more potent to suppress glucose production in hepatic cells. While treatment with AN‐48 was well tolerated, with no adverse effects observed even after 1 month of chronic exposure, further dose–response and toxicological studies in additional animal models are required to ascertain formally its safety before future clinical studies.

In conclusion, our data indicate that AN‐48 is a novel PDX analogue with therapeutic efficacy against hepatic steatosis and inflammation. Combined with the finding that AN‐48 also prevented diet‐induced insulin resistance, we suggest that the PDX analogue may be a potential new drug candidate for the treatment of MASLD.

Author Contributions

F.D., B.M., G.G., J.T. and P.L.M. performed experiments; R.M. and J.‐Y.S. produced the protectins and analogues used in the experiments; F.D., B.M., G.G., P.L.M., A.M. analyzed results; F.D., R.M., P.L.M., O.B., D.P. and A.M. provided input into the study design and interpretation of results; F.D., P.L.M. and A.M. wrote the paper with input from all co‐authors, who read, edited and approved the final manuscript.

Conflicts of Interest

André Marette, René Maltais, Jean‐Yves Sancéau, and Donald Poirier hold a patent for the total synthesis of specialized pro‐resolving mediators, structural isomers, and analogs (US20230303474A1). All other authors declare no conflicts of interest.

Supporting information

Figure S1: (A) Structure of AN‐48 (pro‐drug) and it's metabolite (AN‐43). Plasma concentration of AN43 between 0 and 2 h (B) or 0–24 h (C) following IP administration of AN‐48 at 5, 25, or 50 mg/kg. Plasma (D–F) and liver (G–I) concentrations or PDX, AN‐48 and AN‐43 concentrations following 32 days of treatment of diet‐induced obese hamsters. All data are mean ± SEM (N = 23–24).

Figure S2: (A) Glucose uptake in L6 myocytes after 2 h compound treatment with (gray bar) or without (black bars) the addition of 100 nM insulin. (B) MTT cell viability assay in HEK293T cells following 2‐h incubation with compound indicated. (C) Study design for LPS induced metabolic endotoxemia. (D) Plasma cytokines measured in the LPS treated hamsters.

All data are mean ± SEM (N = 3–7), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. vehicle (N = 2–5).

Figure S3: Metabolic parameters, (A) liver TG, (B) liver TC, (C) Representative hepatic histology (D) PPARα relative expression in hamster liver (E) FBG, (F) change in FBG, (G) FPI at Baseline and (H) pSTAT3 by sex, measured before (BSL) and after (EOT) 32 days of treatment. All data are mean ± SEM (male n = 23 and female n = 24). *p < 0.05, **p < 0.01 significantly different for main treatment effect; ####p < 0.0001, ###p < 0.001, ##p < 0.01, #p < 0.05 indicates significant difference for main sex effect.

Table S1: Molecular structure of PDX, PD1 and synthetic analogues.

Table S2: Metabolic parameters, shown by sex, measured before and after 32 days of treatment.

Desmarais F., Maltais R., Sancéau J.‐Y., et al., “A New Synthetic Analogue of Protectin DX With Enhanced Therapeutic Potential Against Metabolic‐Associated Steatotic Liver Disease,” The FASEB Journal 39, no. 20 (2025): e71172, 10.1096/fj.202502152RRRR.

Data Availability Statement

The data that support the findings of this study are available in the Materials and Methods, Results and/or Supporting Information of this article.