Epoxy-oxylipins direct monocyte fate in inflammatory resolution in humans

Abstract

The role of cytochrome P450-derived epoxy-oxylipins and their metabolites in human inflammation and resolution is unknown. We report that epoxy-oxylipins are present in blood of healthy, male volunteers at baseline and following intradermal injection of UV-killed Escherichia coli, an experimental model of acute resolving inflammation. At the site of inflammation, cytochrome P450s and epoxide hydrolase (EH) isoforms, which catabolise oxylipins to corresponding diols, are differentially upregulated throughout the inflammatory response, as is the biosynthesis of epoxy-oxylipins. GSK2256294, a selective sEH inhibitor specifically elevates 12,13-EpOME and 14,15-EET. While inhibition of sEH hastens pain resolution, it has no effect on tissue heat, redness and swelling. GSK2256294, however, significantly reduces numbers of circulating intermediate monocytes that expand during inflammation. We find that 12,13-EpOME blocks the transition of classical to intermediate monocytes in a p38 MAPK-dependent manner, results that are recapitulated when blocking p38 MAPK in vitro and when administering the p38 MAPK inhibitor losmapimod in vivo to healthy volunteers. Furthermore, fewer intermediate monocytes are observed at the site of inflammation, accompanied by reduced tissue CD4 T cells. Hence, we have mapped the expression, activity and function of epoxy-oxylipins in human inflammation revealing new mechanisms of monocyte differentiation and resolution biology.

Fig. 1. Expression of CYP2J2 and sEH and lipid profiles during inflammation in human skin tissue.

Paraffin-embedded skin punch biopsies were obtained at baseline, 4, 8 and 24 h post UVKEc injection. Sections were incubated with A CYP2J2 and B sEH primary antibodies and visualised by avidin-biotin complex-based detection methods. The sections were subsequently counterstained with haematoxylin. C Skin biopsies at baseline, 4, 8 and 24 h were stained and the % area of staining quantified for (i) CYP2J2, (ii) Pan CYP2C, (iii) sEH, (iv) mEH (n = 1–3; biologically independent samples). D Healthy volunteers received an intradermal injection into the forearm of UV-killed E. coli (UV-KEc), resulting in a local inflammatory response. Local inflammatory exudate was collected at baseline, 4, 8, 14, 24, 48 and 72 h post UVKEc injection and analysed by mass spectrometry for lipidomic profiles. (i) Levels of polyunsaturated fatty acids (PUFAs) were quantified in pg/ml (n = 2–5; biologically independent samples). Cytochrome P450 products and Epoxide Hydrolase products (ii) 14,15-DHET, (iii) 12,13-DiHOME, (iv) 19,20-DiHDPA and (v) 17,18-DiHETE were quantified during inflammation in pg/ml (n = 4-5; biologically independent samples). (vi) 12,13-EpOME was quantified in pg/ml (n = 4-5; biologically independent samples). The dotted line represents baseline levels of the lipids. (vii) Total CYP450 profiles (viii) total COX products were normalised relative to baseline levels to show the changing profile during acute and resolving inflammation (each dot is representative of 4–5 biologically independent samples). Data were assessed for normalisation using the D’Agostino & Pearson test, the Shapiro–Wilk test and visualised using a QQ plot. Data are presented as mean ± SD. Source data are provided as a Source Data file.

Fig. 2. Study design with GSK2256294 for prophylactic and therapeutic dosing regimens and clinical score data.

A Prophylactic study design. Participants received GSK2256294 2 h before inflammation was induced by intradermal UVKEc injection. Blood samples were taken before dosing, 2 h after dosing, and at 4, 24 and 48 h post-inflammation. Blisters were raised at 4 and 24 h. B Therapeutic study design. Participants received GSK2256294 4 h after inflammation was induced by intradermal UVKEc injection. Blood was collected at baseline and at 4, 24 and 48 h post-inflammation. Blisters were raised at 24 and 48 h. C Clinical scores in the prophylactic arm of the study (i) average pain score scored from 0 to 10 in each arm; (ii) average temperature at the inflammatory site in degrees Celsius (°C); (iii) blister volume in ml (untreated: n = 12; GSK2256294: n = 12; biologically independent samples). D Clinical scores in the therapeutic arm of the study (i) average pain score scored from 0 to 10 in each arm; (ii) average temperature at the inflammatory site in °C; (iii) blister volume in ml (untreated: n = 12; GSK2256294: n = 6–11; biologically independent samples). Data for C, D are presented as mean ± SD. E Inflammatory sites were imaged at baseline, 4, 24 and 48 h after UVKEc injection using laser doppler imaging. Blood flow was quantified as perfusion units (flux × valid pixels) with a threshold of >300 flux units. (i) Representative doppler images of blood flow. (ii) Perfusion units with prophylactic dosing of GSK2256294 (untreated: n = 10; GSK2256294: n = 12; biologically independent samples). (iii) Perfusion units with therapeutic dosing of GSK2256294 (untreated: n = 10-11; GSK2256294: n = 12; biologically independent samples). Data were assessed for normalisation using the D’Agostino & Pearson test, the Shapiro–Wilk test and visualised using a QQ plot. For C(i–ii), D(i–ii), and E(ii–iii), data were analysed using two-way ANOVA mixed-effects models with Tukey’s multiple-comparison test. C(iii) and D(iii) were analysed using two-way ANOVA mixed-effects models with Šídák’s correction. Source data are provided as a Source Data file. Created in BioRender. Bracken, O. (2026) https://BioRender.com/xvt2mut.

Fig. 3. Cytochrome P450-derived lipids are elevated with prophylactic and therapeutic sEH inhibition in plasma.

The forearms of participants were intradermally injected with UV-killed E. coli (UVKEc), resulting in a local and peripheral inflammatory response. Two hours prior to (prophylactic) or 4 h after (therapeutic) UVKEc injection, participants were dosed with 15 mg of GSK2256294. Plasma was collected and subject to lipidomic analysis at all timepoints. A Quantification of (i) 14,15-EET, (ii) 14,15-DHET, (iii) the ratio of 14,15-EET:14,15-DHET, (iv) 12,13-EpOME, (v) 12,13-DiHOME and (vi) 12,13-EpOME:12,13-DiHOME prior to inflammation in participants dosed with GSK2256294 for 2 h (untreated: n = 12; GSK2256294: n = 12; biologically independent samples). B Quantification of (i) 14,15-EET, (ii) 14,15-DHET, (iii) the ratio of 14,15-EET:14,15-DHET, (iv) 12,13-EpOME, (v) 12,13-DiHOME and (vi) 12,13-EpOME:12,13-DiHOME during inflammation in the prophylactic arm of the study (untreated: n = 11–12; GSK2256294: n = 12; biologically independent samples). C Quantification of (i) 14,15-EET, (ii) 14,15-DHET, (iii) the ratio of 14,15-EET:14,15-DHET, (iv) 12,13-EpOME, (v) 12,13-DiHOME and (vi) 12,13-EpOME:12,13-DiHOME in the therapeutic arm of the study (untreated: n = 11–12; GSK2256294: n = 11–12; biologically independent samples). All data are expressed in pg/ml. Box-and-whisker plots show the median (centre line), the interquartile range (25th–75th percentiles; box), and the full data range (whiskers, minimum to maximum). Normality was assessed using the D’Agostino & Pearson test, the Shapiro–Wilk test and visualised using a QQ plot. Data in (A) were analysed using a two-tailed, paired, parametric t-test. Data in (B, C) were analysed using two-way ANOVA mixed effect analysis with Šídák’s multiple comparison test. Source data are provided as a Source Data file.

Fig. 4. sEH inhibition both prophylactically and therapeutically inhibits the expansion of intermediate monocytes.

The forearms of participants were intradermally injected with UV-killed E. coli (UVKEc), resulting in a local and peripheral inflammatory response. Two hours prior to (prophylactic) or 4 h after (therapeutic) UVKEc injection, participants were dosed with 15 mg of GSK2256294. Peripheral blood was collected, and leucocytes analysed at baseline and 4, 24 and 48 h post UVKEc injection using multiparameter flow cytometry. A Classical (CD14+CD16−), intermediate (CD14+CD16+) and non-classical monocytes (CD14–CD16+) were quantified using manual gating in FlowJo. (i) Classical, (ii) intermediate and (iii) non-classical monocyte numbers during inflammation with prophylactic sEH inhibition (untreated: n = 10-12; GSK2256294: n = 8; biologically independent samples). (iv) Classical, (v) intermediate and (vi) non-classical monocytes during inflammation with therapeutic sEH inhibition (untreated: n = 11–12; GSK2256294: n = 10–11; biologically independent samples). B UMAP of monocyte populations identified in peripheral blood in the therapeutic arm of the study. Total monocytes were extracted from FCS files in FlowJo and clustered using CATALYST; (i) six monocyte populations were identified; (ii) heatmap of marker expression in each cluster; (iii) the percentage of intermediate monocytes using supervised clustering was quantified as a % of total monocytes at 4 and 24 h in untreated and drug-treated participants. (untreated: n = 6; GSK2256294: n = 6; biologically independent samples). C Intermediate monocytes at 24 h were analysed using CATALYST for median expression of (i) ICAM-1, (ii) PSGL-1, (iii) VCAM-1, (iv) CCR2 and (v) CX3CR1 in untreated, prophylactic and therapeutic participants (untreated: n = 11; GSK2256294 prophylactic: n = 5; GSK2256294 therapeutic: n = 6; biologically independent samples). Data were assessed for normalisation using the D’Agostino & Pearson test, the Shapiro–Wilk test and visualised using a QQ plot. Box-and-whisker plots show the median (centre line), the interquartile range (25th–75th percentiles; box), and the full data range (whiskers, minimum to maximum). Data in (A) were analysed using two-way ANOVA mixed effect analysis with Tukey’s multiple comparisons test. Data in (B) were analysed using two-way ANOVA mixed effect analysis with Uncorrected Fisher’s LSD. Data in (C) were analysed using a One-Way ANOVA. Source data are provided as a Source Data file.

Fig. 5. The effect of 12,13-EpOME and p38 inhibition on monocyte differentiation.

A PBMCs ($2\times {10}^6$/well) were seeded and treated with 1 μM epoxy-oxylipin or vehicle for 24 h. Cells were phenotyped by flow cytometry. (i) % classical monocytes (CM), (ii) % intermediate monocytes (IM), (iii) % non-classical monocytes (NCM) with 1 µM 12,13-EpOME vs. methyl acetate (MA) (MA: n = 6; 12,13-EpOME: n = 7; biologically independent samples). (iv–vi) % CM, IM, and NCM with 1 µM 14,15-EET vs ethanol (EtOH) (EtOH: n = 3; 14,15-EET: n = 4; biologically independent samples). B (i) Representative image of a western blot for P-p38 in monocyte subsets. (ii) Quantification of P-p38 expression in monocyte subsets relative to GAPDH (n = 4; biologically independent samples). C PBMCs ($2\times {10}^6$/well) were treated with 1 μM 12,13-EpOME or vehicle (MA) for 24 h, then re-stimulated with 1 μM 12,13-EpOME or MA plus 100 ng/ml LPS for 30 min. (i) Phosphorylation of p38 in PBMCs treated with 1 μM 12,13-EpOME or MA (n = 10; biologically independent samples). (ii) Representative histogram of P-p38 MFI in PBMCs treated with LPS + MA, LPS + 12,13-EpOME, LPS + losmapimod (Los), or unstimulated. D PBMCs ($2\times {10}^6$/well) were treated with 3 μM losmapimod or DMSO (vehicle) for 24 h, then phenotyped by flow cytometry. Shown are % (i) classical (CM), (ii) intermediate (IM), and (iii) non-classical monocytes (NCM) (n = 8; biologically independent samples). E Older people (>65 years old) were dosed with losmapimod daily for four days. Peripheral blood was taken both before and after dosing with losmapimod and analysed for monocyte subsets using flow cytometry. % of (i) CM, (ii) IM, (iii) NCM in peripheral blood at steady state before and after dosing with losmapimod (n = 8; biologically independent samples). Data were assessed for normalisation using the D’Agostino & Pearson test, the Shapiro–Wilk test and visualised using a QQ plot. Parametric data are presented as mean ± SD. Data in (A, C, D, E) were analysed using a two-tailed parametric, paired t-test. Data in (B) were analysed using a one-way ANOVA. Source data are provided as a Source Data file.

Fig. 6. sEH inhibition significantly increases the ratio of 12,13-EpOME:12,13-DiHOME and reduces numbers of intermediate monocytes at the inflammatory site.

The forearms of participants were intradermally injected with UV-killed E. coli (UVKEc) resulting in a local inflammatory response. Two hours prior to (prophylactic) or 4 h after (therapeutic) UVKEc injection, participants were dosed with 15 mg of GSK2256294. Local inflammatory exudate was collected and subject to lipidomic and flow cytometric analysis at 4 and 24 h (prophylactic) and 24 and 48 h (therapeutic). A Quantification of (i) 12,13-EpOME, (ii) 12,13-DiHOME and (iii) 12,13-EpOME:12,13-DiHOME in blister fluid at 4 and 24 h (prophylactic) (untreated: n = 12; GSK2256294: n = 12; biologically independent samples). Quantification of (iv) 12,13-EpOME, (v) 12,13-DiHOME and (vi) 12,13-EpOME:12,13-DiHOME in blister fluid at 24 and 48 h (therapeutic) (untreated: n = 6; GSK2256294: n = 12; biologically independent samples). Data are expressed in pg/ml. B UMAP of monocyte populations identified in local inflammatory exudate (therapeutic arm). Total monocytes were extracted from FCS files in FlowJo and analysed using CATALYST. Eight monocyte populations were identified and labelled. C Heatmap of marker expression in each cluster. D UMAP of monocyte populations faceted by time and treatment (B–D representative of untreated: n = 6; GSK2256294: n = 6–7; biologically independent samples). E Classical monocytes (CD14+CD16−), intermediate monocytes (CD14+CD16+) and non-classical monocytes (CD14−CD16+) were quantified using manual gating in FlowJo. (i) Intermediate monocytes/blister at 24 h (prophylactic) (untreated: n = 12; GSK2256294: n = 10; biologically independent samples). (ii) Intermediate monocytes/blister at 24 and 48 h (therapeutic) (untreated: n = 6–12; GSK2256294: n = 6–11; biologically independent samples). (iii) CM-IM and IM populations from the supervised clustering analysis were combined and analysed as a % of total monocytes at 24 and 48 h (therapeutic) (untreated: n = 6; GSK2256294: n = 6–7; biologically independent samples). Normality was assessed using the D’Agostino & Pearson test, the Shapiro–Wilk test and visualised using a QQ plot. Box-and-whisker plots show the median (centre line), the interquartile range (25th–75th percentiles; box), and the full data range (whiskers, minimum to maximum). Data in A, E(ii–iii) were analysed using two-way ANOVA mixed effect analysis with Uncorrected Fisher’s LSD. Data in E(i) were analysed using a two-tailed, Mann–Whitney test. Source data are provided as a Source Data file.

Fig. 7. CD4 T cells are reduced at the inflammatory site with therapeutic sEH inhibition.

The forearms of participants were intradermally injected with UV-killed E. coli (UV-KEc). Two hours prior to (prophylactic) or 4 h after (therapeutic) UV-KEc injection, participants were dosed with 15 mg of GSK2256294. Local inflammatory exudate was subject to flow cytometric analysis at 4 and 24 h (prophylactic) and 24 and 48 h (therapeutic). A (i) CD4, (ii) CD4 T-regulatory, (iii) CD8 T cell numbers per blister at 48 h (therapeutic) (untreated: n = 5; GSK2256294: n = 5; biologically independent samples). (iv) % dead T cells at 24 and 48 h (therapeutic) (untreated: n = 6-7; GSK2256294: n = 6; biologically independent samples). B Concentration of IL-1α in pg/ml in blister fluid at 24 and 48 h (therapeutic) (untreated: n = 10–11; GSK2256294: n = 6–12; biologically independent samples). C Monocyte subset numbers/blister at 4, 24 and 48 h in untreated participants (n = 11–18; biologically independent samples). D Time course of CD4 and CD8 subset numbers/blister in untreated participants (n = 6; biologically independent samples). E Intermediate monocytes were co-cultured with CD4 T cells at a 5:1 (T cell:monocyte) ratio for 48 h and analysed by spectral flow cytometry. The % expression on total CD4 cells was visualised for (i) CD25, (ii) CD39, (iii) CD69, (iv) CD103, (v) CLA, (vi) CTLA-4, (vii) HLA-DR, (viii) Ki67, (ix) CD45RO and (x) FOXP3+CD25+CD127− (n = 3; technical repeat). F Classical, intermediate and non-classical monocytes and CD8 T cells were co-cultured for 4 days at a 5:1 (T cell:monocyte) ratio, after which a cytotoxicity assay against K562 cells was performed (n = 2; biologically independent samples). Normality was assessed using the D’Agostino & Pearson test, the Shapiro–Wilk test and visualised using a QQ plot. Box-and-whisker plots show the median (centre line), the interquartile range (25th–75th percentiles; box), and the full data range (whiskers, minimum to maximum). Parametric data are presented as mean ± SD. Non-parametric data are presented as median ± 95% CI. Data in (A(i-iii)) were analysed using a two-tailed, Mann–Whitney t-test. Data in (A(iv), B) were analysed using two-way ANOVA mixed effect analysis with Uncorrected Fisher’s LSD. Data in (C) were analysed using two-way ANOVA mixed effect analysis with Šídák multiple comparison test. Data in (E) were analysed using a two-tailed, parametric, paired t-test. Source data are provided as a Source Data file.