Pro-resolving lipid mediator lipoxin A4 attenuates neuro-inflammation by modulating T cell responses and modifies the spinal cord lipidome.

The chronic neuro-inflammatory character of multiple sclerosis (MS) suggests that the natural process to resolve inflammation is impaired. This protective process is orchestrated by specialized pro-resolving lipid mediators (SPMs), but to date, the role of SPMs in MS remains largely unknown. Here, we provide in vivo evidence that treatment with the SPM lipoxin A4 (LXA4) ameliorates clinical symptoms of experimental autoimmune encephalomyelitis (EAE) and inhibits CD4+ and CD8+ T cell infiltration into the central nervous system (CNS). Moreover, we show that LXA4 potently reduces encephalitogenic Th1 and Th17 effector functions, both in vivo and in isolated human T cells from healthy donors and patients with relapsing-remitting MS. Finally, we demonstrate that LXA4 affects the spinal cord lipidome by significantly reducing the levels of pro-inflammatory lipid mediators during EAE. Collectively, our findings provide mechanistic insight into LXA4-mediated amelioration of neuro-inflammation and highlight the potential clinical application of LXA4 for MS.

INTRODUCTION

Multiple sclerosis (MS) is the most common chronic inflammatory disorder of the central nervous system (CNS), characterized by blood-brain barrier disruption, immune cell infiltration into the CNS, demyelination, and neurodegeneration (Reich et al., 2018). Although it is still debated whether inflammation and/or autoimmunity are at the root of the disease (Stys et al., 2012), there is extensive evidence for a crucial role of the immune system in disease pathogenesis (Lassmann, 2005; Sawcer et al., 2011). Particularly, genome-wide association studies have identified over 230 genetic loci commonly associated with MS, the majority of which are linked to genes associated with the immune system, T cell activation, and proliferation, thereby implicating the peripherally activated CD4+ T cells in MS pathogenesis (Cotsapas and Mitrovic, 2018; Hellberg et al., 2016; Sawcer et al., 2011). Moreover, the abundant presence of inflammatory cells in CNS lesions of patients with MS or in its experimental model experimental autoimmune encephalomyelitis (EAE), has led to the hypothesis that MS is mediated by pathogenic T cell responses against myelin antigens, followed by neurodegeneration (Compston and Coles, 2008). As a result, the majority of the current therapies consist of immunosuppressive or anti-inflammatory agents, which are effective in early stages of the disease but often result in severe side effects and eventually fail to prevent disease progression (Chiurchiù, 2014). Therefore, there is an urgent need to develop alternative approaches allowing not only to interfere with chronic neuro-inflammation but also to promote local tissue recovery and subsequent homeostasis.

The chronic neuro-inflammatory character of MS suggests that the natural process to resolve inflammation (resolution of inflammation) is not properly functioning (Newson et al., 2014; Serhan, 2014). Resolution of inflammation is an active process mediated by specialized pro-resolving lipid mediators (SPMs), which are synthetized from ω−3 or ω−6 polyunsaturated fatty acids. Among these, E and D series resolvins, protectins, and maresins are derived from ω−3 polyunsaturated fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), respectively. SPMs of the lipoxin family include lipoxin A4 (LXA4) and LXB4 as well as their isomeric isoforms aspirin-triggered lipoxin (Serhan, 2005), which are biosynthesized from arachidonic acid (AA) (Calder, 2006; Serhan, 2014; Serhan and Petasis, 2011). In general, SPMs actively promote resolution of inflammation via specific mechanisms that include limiting neutrophil recruitment, enhancing efferocytosis, and pathogen removal, as well as reducing the secretion and production of pro-inflammatory mediators (Gilroy et al., 2004; Serhan, 2014) and reactive oxygen species (Leuti et al., 2019), thereby promoting the cardinal signs of resolution. Despite the beneficial effects in various chronic inflammatory disease models (Chiurchiù et al., 2018), the potential role of the resolution program and SPMs in MS and neuro-inflammation in general remains to be investigated.

Importantly, we recently showed that LXA4 levels were not increased in plasma of patients with relapsing-remitting MS compared to healthy controls (Kooij et al., 2020), providing first evidence for a potential resolution defect in MS. These findings are in line with earlier observations in the cerebrospinal fluid of patients with highly active MS compared to those with inactive disease (Prüss et al., 2013). Due to the robust neuro-inflammatory character of MS in combination with the potent anti-inflammatory and pro-resolving capacities of LXA4, we hypothesized that this SPM may represent an innovative strategy to limit neuro-inflammation in the context of MS. We here show that daily injections of LXA4 ameliorate EAE clinical symptoms, reduce CNS infiltration of CD4+ and CD8+ T cells, dampen Th1 and Th17 effector T cell functions, and potently modulate the inflammatory responses of encephalitogenic T lymphocytes from patients with relapsing-remitting MS. Moreover, we provide evidence that the spinal cord lipidome is severely altered during EAE and that LXA4 administration normalizes these effects by specifically dampening pro-inflammatory lipid mediators derived from AA. Overall, these findings highlight the therapeutic potential of LXA4 in MS and provide detailed insight into its mechanism of action in counteracting neuro-inflammation.

RESULTS

LXA4 treatment ameliorates EAE clinical signs

To study the effect of LXA4 treatment on neuro-inflammation in vivo, we induced EAE in female C57BL/6J mice and subjected them to a prophylactic treatment regimen. Mice received daily intraperitoneal injection of 100 ng LXA4/mouse or ethanol (1% in saline) as a vehicle control, starting prior to onset of the disease (day 8) until the end (Figure 1A). Interestingly, LXA4 treatment ameliorated EAE clinical scores and counteracted the reduction in the body weight compared to EAE vehicle-treated mice (Figures 1B and 1C). In particular, we observed a strong trend toward reduction in the severity of the disease between the LXA4-prophylactic-treated mice and the vehicle-treated mice (Figure 1D; Tables S1 and S2). To provide more insights into the mode of action of LXA4 and to understand in which time window it exerts its protective actions, we investigated the effect of LXA4 withdrawal on clinical signs of EAE. To achieve this, LXA4 was injected from day eight but withdrawn after day 21, leaving the animals untreated for the remainder of the experiment (until day 30; Figure 1E). Importantly, LXA4 treatment significantly reduced clinical signs of EAE (Figure 1F), the mean clinical score at the peak of the disease, and disease severity compared to the vehicle-treated EAE mice (Figure 1H; Tables S1 and S3), with a slight impact on weight changes of the animals (Figure 1G). Interestingly, LXA4 cessation at day 21 elicited a rebound effect, resulting in a worsening of EAE symptoms when compared to the last day of the treatment period. In contrast, the vehicle-EAE group displayed a reduction in EAE scores (Figure 1I; Table S3). Overall, these findings suggest that LXA4 ameliorates EAE and that its presence is required to continuously dampen neuro-inflammation and suppress clinical symptoms of EAE.

Figure 1.

LXA4 treatment ameliorates clinical signs of EAE

(A) Timeline illustrating the experimental setup of the treatment regimen of EAE mice.

(B) Clinical EAE scores of vehicle-treated mice (1% ethanol; n = 10 mice), LXA4-treated mice (n = 8 mice; 100 ng/animal/day) and CFA/control mice (n = 9) until 30 days after EAE induction (treatment period highlighted in gray).

(C) Change in body weight, expressed in grams compared to day 0.

(D) Disease severity at day 30 p.i. of vehicle or LXA4-treated mice.

(B–D) Results are presented as mean ± SEM.

(E) Timeline illustrating the experimental setup of the ‘‘LXA4-withdrawal’’ EAE experiment.

(F) Clinical EAE scores of vehicle-treated mice (n = 9), LXA4-treated mice (treatment period: day 8 until day 21 p.i.; n = 10), and CFA/control mice (n = 10) until 30 days after EAE induction (treatment period highlighted in gray).

(G) Change in body weight, expressed in grams compared to day 0.

(H) Disease severity at day 21 p.i. of vehicle or LXA4-treated mice. Results are presented as mean ± SEM. *p < 0.05 determined by Mann Whitney test.

(I) Rebound effect: change in EAE clinical scores of vehicle-treated mice (n = 6 mice) and LXA4-treated mice (n = 7 mice) after LXA4 treatment cessation (day 21 versus day 30 after EAE induction), presented as mean ± SEM.

(J and K) Representative pictures of CD45+ (red), CD4+ (J), CD8+ (K) cells (green), collagen IV (cyan; vasculature), and DAPI (blue; nuclei) immunohistochemical stainings of cervical spinal cord sections from EAE-vehicle-treated and LXA4-treated mice (day 8–30 p.i.) sacrificed at the peak of the disease (day 21 p.i.; n = 7–9 biological replicates/group). Scale bars, overview: 1 mm and inset: 50 µm.

(L) Quantification of CD45+, CD4+, and CD8+ infiltrated cells in the spinal cord parenchyma of LXA4- and vehicle-treated EAE mice. Left graph displays average number of infiltrated cells per EAE (n = 3 EAE experiments, 3 mice/group/EAE; dotted line connects values of individual EAE experiments); right graph displays the average number of infiltrated cells of individual mice (n = 7–9 biological replicates/group); data points are color-coded according to the individual EAE experiment. Results are presented as mean ± SEM. *p < 0.05 determined by paired (left graph) or unpaired (right graph) Student’s t test.

LXA4 inhibits CD4+ and CD8+ T cell migration into the CNS during EAE

Because the main neuro-inflammatory hallmark in MS pathology is represented by an abundant infiltration of peripheral leukocytes into the CNS, and particularly in the spinal cord, we next sought to determine whether LXA4 was able to modulate leukocyte recruitment in EAE mice. Accordingly, we isolated the spinal cord of LXA4-treated versus vehicle-treated mice and performed immunohistochemical analysis to quantitatively assess the extent of leukocyte infiltration into the CNS parenchyma at the peak of the disease (Figures 1J and 1K). Interestingly, upon LXA4 treatment, we observed a significant reduction of infiltrating CD45+ leukocytes in the spinal cord parenchyma compared to vehicle-treated EAE mice at the peak of disease in three independent experiments (Figures 1J–1L). As the CD45+ population contains both pro- and anti-inflammatory cells, we next focused on the main pathogenic cells that mediate myelin damage during MS, which are autoreactive CD4+ and CD8+ T lymphocytes, especially interferon (IFN)-γ-producing Th1 and interleukin-17 (IL-17)-producing Th17 subsets, whose presence is abundant in cerebrospinal fluid and MS lesions (Compston and Coles, 2008; Han et al., 2014; Lassmann, 2014). Importantly, further characterization of the CD45+ spinal cord infiltrates revealed that both CD4+ and CD8+ T cell subsets were significantly decreased upon LXA4 treatment compared to vehicle-treated EAE animals (Figures 1J–1L). Overall, these in vivo findings suggest that LXA4 treatment ameliorates EAE clinical symptoms and attenuates neuro-inflammation by affecting CD4+ and CD8+ T cell infiltration into the CNS.

LXA4 modulates central and peripheral mouse Th1 and Th17 responses

To further address the underlying mechanism of action, we next assessed whether LXA4 was able to impact effector T cell responses in terms of cytokine production. For this, we examined lymphocytes from both the spinal cord and blood of EAE mice upon LXA4 treatment compared to vehicle-treated EAE mice at the peak of disease. We specifically evaluated encephalitogenic Th1 and Th17 responses in terms of tumor necrosis factor alpha (TNF-α), IFN-γ (for both CD4+ and CD8+ T cells), and IL-17 (for CD4+ T cells; see Figure S1 for gating strategies). Importantly, in the spinal cord, LXA4 significantly reduced the percentage of infiltrated CD4+ T cells producing TNF-α, IFN-γ, and IL-17, whereas a similar trend was observed for CD8+ T cells (Figure 2A). Moreover, upon anti-CD3 and anti-CD28 stimulation of peripheral lymphocytes isolated from LXA4-treated versus vehicle-treated EAE mice, we observed a significant reduction in both CD4+ and CD8+ T cell responses in terms of TNF-α, IFN-γ, and IL-17 cytokine production (Figure 2B). Overall, these data suggest that the LXA4-induced amelioration of EAE course and severity might be due to a reduction of Th1 and Th17 effector functions, both centrally and peripherally.

Figure 2.

LXA4 reduces central and peripheral cytokine responses

Immune cell populations were isolated from LXA4- or vehicle-treated EAE mice at peak of disease and analyzed by flow cytometry, as described in the STAR Methods.

(A) Representative flow cytometry plots and bar graphs from spinal cord infiltrates showing percentages of intracellular production of TNF-α, IFN-γ, and IL-17 from CD4+ T cells (left panel) or TNF-α and IFN-γ from CD8+ T cells (right panel). Results are presented as mean ± SEM of 6–8 individual mice/group. *p < 0.05 and **p < 0.01, as determined by Student’s t test.

(B) Representative flow cytometry plots and bar graphs of matching peripheral blood samples showing percentages of intracellular production of TNF-α, IFN-γ, and IL-17 from CD4+ T cells (left panel) or TNF-α and IFN-γ from CD8+ T cells (right panel) upon re-stimulation with anti-CD3/CD28. Results are presented as mean ± SEM of 6–8 individual mice/group. *p < 0.05 and **p < 0.01, as determined by Student’s t test.

LXA4 modulates human Th1 and Th17 responses

To translate these findings to a human setting, we next addressed whether LXA4 was able to also impact human T cell responses and whether it was able to affect their transendothelial migration capacity. For this, we investigated human T cells from patients with relapsing-remitting MS as well as healthy controls (Table S4). In particular, the immunomodulatory role of LXA4 was investigated upon polyclonal activation of the T cell receptor with anti-CD3 and anti-CD28 (Figure 3A). We found that activated T cells showed a significantly induced expression of CD25, which was not affected by LXA4 in both control and relapsing-remitting MS patient-derived T cells (Figure 3B). To further assess the effect of LXA4 on specific T cell subsets, we also investigated cytotoxic CD8+ T cells and CD4+ Th1 and Th17 cells. As expected and previously shown (Gentile et al., 2020), anti-CD3 and anti-CD28 stimulation resulted in high percentages of TNF-α-, IFN-γ-, and IL-17-producing T cells in both controls and patients with relapsing-remitting MS (Figures 3D–3G). Of note, cytokine production from all T cell subsets was higher in patients with relapsing-remitting MS, accounting for the hyper-activation of these pathogenic cells in MS. Importantly, LXA4 treatment significantly reduced the percentage of TNF-α-producing CD8+ T cells of healthy subjects (Figure 3C) and IFN-γ-producing CD8+ T cells of both healthy subjects and patients with relapsing-remitting MS (Figure 3D). Moreover, LXA4 significantly reduced the percentage of IL-17- and TNF-α-producing CD4+ T cells of healthy subjects, but not from patients with relapsing-remitting MS (Figures 3E and 3F), although the percentage of IFN-γ-producing cells was reduced by LXA4 both in healthy and relapsing-remitting-MS-patient-derived CD4+ T cells (Figure 3G). In view of these results, we next sought to investigate the effect of LXA4 on the transendothelial migration capacity of CD4+ and CD8+ T cells by using a human in vitro blood-brain barrier model (Kooij et al., 2010, 2014). As shown in Figures 3H and 3I, treatment with LXA4, significantly inhibited the migration of both human T cell subsets across inflamed brain endothelial cells. Overall, these results indicated that the production of specific cytokines that characterize the main pro-inflammatory T cell subsets from control or patients with MS was significantly reduced or showed a trend toward significance upon pre-treatment with LXA4, indicating that LXA4 regulates human T cell responses as well as their transmigration capacity.

Figure 3.

LXA4 reduces cytokine responses in activated human T cell populations from control and patients with relapsing-remitting (RR)-MS

(A) Schematic representation of the experimental procedure: peripheral blood mononuclear cells (1 × 106 cells per well) were left untreated (black bars) or treated with vehicle (red bars) or LXA4 (10 nM; blue bars) for 30 min. Cells were then stimulated with anti-CD3/CD28 for 8 h, stained at the cell surface and intracellularly, and analyzed by flow cytometry.

(B) Mean fluorescence intensity (MFI) of CD25 surface expression in total CD3+ T cells.

(C and D) Percentages of intracellular production of (C)TNF-α and (D) IFN-γ from CD8+ T cells.

(E–G) Percentages of intracellular production of (E) IL-17, (F) TNF-α, and (G) IFN-γ from CD4+ T cells.

(B–G) Results are presented as mean ± SEM of six biological replicates. **p < 0.01 and ***p < 0.001, as determined by one-way ANOVA followed by post hoc Bonferroni correction. Confluent brain endothelial cells (BECs) were stimulated for 24 h with TNF. Human CD4+ T or CD8+ T cells (1 × 105 cells/well) were treated with LXA4 (10 nM) or vehicle (ethanol) for 30 min prior to being plated on BECs. Cells were incubated in a co-culture setting for 4 h before harvesting the transmigrated cells.

(H and I) Percentage of (H) CD4+ or (I) CD8+ T cell transmigration evaluated by flow cytometry. Data are presented as mean ± SEM of 3 biological replicates done in triplicate. **p < 0.01 compared to control (vehicle-treated T cells) determined by Student’s t test.

Lipid mediator alterations in the spinal cord during EAE

Besides extensive infiltration of various leukocyte subsets into the CNS, EAE is also characterized by local microglia and astrocyte activation, which together results in clinical signs (Engelhardt and Ransohoff, 2005). To further characterize this pro-inflammatory CNS milieu during EAE and address the question whether CNS inflammation is reflected by changes in lipid mediator profiles, we performed targeted lipid mediator metabololipidomics using liquid chromatography-tandem mass spectrometry (LC-MS/MS) by analyzing 42 distinct lipid mediators based on published criteria (i.e., matching chromatographic retention times [RTs], fragmentation patterns, and six characteristic and diagnostic ions; Colas et al., 2014). For this, we analyzed spinal cords from control and EAE mice during the peak and chronic phase of the disease. Accordingly, we identified lipid mediators from each of the ω−6 AA, ω−3 DHA, and EPA bioactive metabolomes (Table 1; complete lipid mediator stereochemistry and annotated biological functions have been previously described; Dalli and Serhan, 2012). We identified several SPMs in the spinal cord of control and EAE mice, including LXA4, LXB4, resolvin D2 (RvD2), resolvin D3 (RvD3), maresin-1 (Mar1), resolvin E2 (RvE2), aspirin-triggered PD1 (AT-PD1), and resolvin E3 (RvE3) as well as various key pathway markers (Table 1; Figure S2). Importantly, we observed that neuro-inflammation predominantly affected the AA bioactive metabolome by revealing significantly increased expression levels of pro-inflammatory prostaglandins like prostaglandin E2 (PGE2), prostaglandin D2 (PGD2), prostaglandin F2a (PGF2a), leukotriene B4 (LTB4), and thromboxane (TXB2) during the chronic phase of disease as well as the peak of disease (PGE2) compared to (Complete Freund’s Adjuvant, CFA) control mice (Table 1; Figures 4A–4D and S2). We also observed significantly increased levels of various AA and EPA pathway markers, like 5-HETE, 15-HETE, 5-HEPE, and 18-HEPE, during different stages of EAE compared to CFA (Table 1). Together, these findings provide a detailed characterization of lipid mediator profiles in the spinal cord of control and EAE mice.

Table 1.

Mouse EAE spinal cord lipid mediator (LM)-SPM signature profile upon LXA4 treatment

AA bioactive metabolome	Q1	Q3	CFA	EAE peak + vehicle	EAE peak + LXA4	EAE end + vehicle	EAE end + LXA4
AA	303	259	5,606.1 ± 353.2	5,532.0 ± 543.7	5,285.2 ± 457.0	5,225.0 ± 866.2	4,386.1 ± 161.6
LTB4	335	195	1.98 ± 0.30	1.11± 0.04	1.11 ± 0.07	5.58 ± 1.72[a]	1.02 ± 0.25[b]
20-OH-LTB4	351	195	‒	‒	‒	‒	‒
20-COOH-LTB4	365	195	‒	‒	‒	‒	‒
5S,12S-diHETE	335	195	0.36 ± 0.08	0.29 ± 0.02	0.24 ± 0.07	0.71 ± 0.34	0.23 ± 0.09
5S,15S-diHETE	335	115	‒	‒	‒	‒	‒
PGD2	351	189	542.0 ± 99.3	857.2 ± 74.6	694.1 ± 27.3	1,564.9 ± 162.5[c]	558.3 ± 139.4[b]
PGE2	351	189	337.8 ± 89.6	1,445.0 ± 156.6[c]	651.4 ± 122.5[d]	1,564.2 ± 148.8[c]	871.4 ± 99.3[e]
PGF2a	353	193	95.21 ± 8.68	136.47 ± 4.73	138.39 ± 23.81	392.64 ± 79.56[c]	179.54 ± 7.70[b]
TXB2	369	169	175.1 ± 85.4	309.3 ± 43.5	314.76 ± 93.5	926.8 ± 104.4[c]	246.4 ± 48.0[e]
15-HETE	319	219	231.1 ± 43.3	273.1 ± 39.3	298.3 ± 71.6	846.7 ± 107.9[c]	309.4 ± 34.7[e]
12-HETE	319	179	167.18 ± 54.14	477.53 ± 193.36	181.13 ± 28.68	74.44 ± 6.58	89.10 ± 15.57
5-HETE	319	115	30.80 ± 2.65	86.99 ± 24.07[a]	29.15 ± 3.28[d]	19.75 ± 2.26	22.47 ± 2.81
LXA4	351	115	0.33 ± 0.02	0.34 ± 0.03	0.56 ± 0.05[f]	0.25 ± 0.03	0.49 ± 0.05[g]
LXB4	351	115	1.30 ± 0.17	2.13 ± 0.43	1.80 ± 0.24	2.10 ± 0.64	0.72 ± 0.18[g]
AT-LXA4	351	115	0.18 ± 0.04	0.22 ± 0.07	0.24 ± 0.04	0.40 ± 0.09	0.14 ± 0.12
AT-LXB4	351	221	‒	‒	‒	‒	‒
DHA bioactive metabolome
DHA	327	283	13,765 ± 958	14,624 ± 884	14,365 ± 690	12,542 ± 1,699	10,931 ± 556
RvD1	375	215	‒	‒	‒	‒	‒
RvD2	375	215	0.25 ± 0.04	0.26 ± 0.10	0.27 ± 0.13	0.17 ± 0.05	0.29 ± 0.11
RvD3	375	181	0.11 ± 0.03	0.13 ± 0.03	0.12 ± 0.02	0.15 ± 0.04	0.19 ± 0.04
RvD4	375	255	‒	‒	‒	‒	‒
RvD5	359	199	‒	‒	‒	‒	‒
RvD6	359	101	‒	‒	‒	‒	‒
PD1	359	153	‒	‒	‒	‒	‒
AT-PD1	359	153	0.32 ± 0.13	0.31 ± 0.04	0.18 ± 0.06	0.20 ± 0.05	0.19 ± 0.06
PDX	359	153	0.75 ± 0.27	0.22 ± 0.05	0.44 ± 0.14	1.25 ± 0.38	0.56 ± 0.16
Maresin 1	359	221	0.19 ± 0.02	0.54 ± 0.18	0.53 ± 0.27	0.65 ± 0.32	0.47 ± 0.06
7S,14S-diHDHA	359	221	0.41 ± 0.09	0.76 ± 0.12	0.42 ± 0.12	0.51 ± 0.19	0.59 ± 0.18
17-HDHA	343	245	13.61 ± 4.05	6.11 ± 1.22	7.83 ± 1.53	23.18 ± 8.10	9.48 ± 1.01
14-HDHA	343	205	36.82 ± 19.48	57.72 ± 25.88	18.50 ± 1.62	7.99 ± 0.68	10.06 ± 2.67
7-HDHA	343	141	4.47 ± 0.86	8.11 ± 1.56	4.53 ± 0.26[f]	3.07 ± 0.27	3.09 ± 0.35
4-HDHA	343	101	10.77 ± 1.39	13.32 ± 0.83	11.00 ± 2.20	8.69 ± 0.49	8.42 ± 2.05
EPA bioactive metabolome
EPA	301	257	1,693.4 ± 220.4	1,771.6 ± 279.5	2,114.1 ± 793.8	1,723.0 ± 214.1	1,440.7 ± 112.2
RvE1	349	195	‒	‒	‒	‒	‒
RvE2	333	253	1.57 ± 0.33	1.57 ± 0.27	0.78 ± 0.19	0.74 ± 0.15	0.64 ± 0.13
RvE3	333	201	6.31 ± 0.38	8.66 ± 0.84	6.76 ± 1.51	6.04 ± 1.42	4.83 ± 0.33
18-HEPE	317	259	11.40 ± 1.38	10.67 ± 1.28	16.74 ± 5.55	27.45 ± 7.27[a]	15.27 ± 0.63
15-HEPE	317	219	29.90 ± 18.17	27.93 ± 12.33	12.67 ± 0.28	18.63 ± 2.83	14.63 ± 4.48
12-HEPE	317	179	20.77 ± 4.11	54.24 ± 20.14	25.46 ± 6.02	11.78 ± 0.99	14.36 ± 5.19
5-HEPE	317	115	4.22 ± 0.33	7.92 ± 1.26[h]	4.96 ± 0.72[f]	3.87 ± 0.32	4.13 ± 1.04

AA, DHA, and EPA bioactive metabolome signature profile (pg/100 mg tissue) determined from the spinal cord of vehicle-treated mice, LXA4-treated mice (day 8–30 p.i.), and CFA control mice at the peak of the disease (day 21 p.i.) and at the end of the disease (day 30 p.i.; CFA: n = 5 mice; all other groups: n = 3 mice). Data are presented as means (pg/100 mg spinal cord) ± SEM, and statistical significance is determined by one-way ANOVA followed by post-hoc Bonferroni correction

p < 0.05 compared to CFA

p < 0.01 compared to EAE end + vehicle

p < 0.001 compared to CFA

p < 0.01 compared to EAE peak + vehicle

p < 0.001 compared to EAE end + vehicle

p < 0.05 compared to EAE peak + vehicle

p < 0.05 compared to EAE end + vehicle

p < 0.01 compared to CFA

Figure 4.

Normalization of EAE-induced spinal cord lipidome upon LXA4 treatment

Lipid mediators were isolated from total spinal cords of control CFA (black bars) mice, EAE-vehicle (red bars)-treated mice, or LXA4-treated (day 8–30 p.i.; blue bars), sacrificed at the peak of the disease (day 21 p.i.) and at the end of the disease (day 30 p.i.; CFA: n = 5 mice; all other groups: n = 3 mice), and individually analyzed by LC-MS/MS.

(A–C) Representative MS/MS spectra of PGE2 (A), PGD2 (B), and LTB4 (C) and data are presented as mean (pg/100 mg spinal cord) ± SEM. **p < 0.01 and ***p < 0.001 determined by one-way ANOVA followed by post hoc Bonferroni correction.

(D) Heatmap of lipid mediator profiles under control or during different EAE phases upon LXA4/vehicle treatment. Columns represent different experimental groups and individual mice. The relative value for each lipid mediator is depicted by color intensity, with blue indicating decreased and red indicating increased lipid mediator levels.

(E) Unbiased principal-component analysis reveals temporal clustering of identified lipid mediator signatures in each experimental condition.

Reduced pro-inflammatory lipid mediators upon LXA4 treatment during EAE

We next addressed the question whether the LXA4-mediated amelioration of clinical signs could be explained by local and beneficial changes in the spinal cord lipidome during the peak and chronic phase of EAE. Importantly, LXA4 treatment significantly increased LXA4 expression levels in the spinal cord during different stages of EAE (Table 1), thereby revealing LXA4-CNS incorporation upon treatment. As shown in Figures 4A–4D, we observed that LXA4 treatment predominantly suppressed the AA bioactive metabolome by revealing significantly reduced levels of pro-inflammatory prostaglandins, like PGE2, at both the peak and during the chronic phase of EAE, whereas PGD2, PGF2a, LTB4, and TXB2 were significantly reduced during the chronic phase of EAE upon LXA4 treatment (Figures 4A–4D; Table 1). In contrast, a key AA pathway marker (15-HETE) was significantly reduced upon LXA4 treatment during the chronic phase of EAE, whereas 5-HETE was significantly reduced upon LXA4 treatment at the peak of EAE, similar to DHA and EPA pathway markers, like 7-HDHA and 5-HEPE, respectively (Table 1). Of note, besides LXA4, none of the identified SPMs in the spinal cord were significantly affected by LXA4 treatment (Figure 4D; Table 1). Such targeted metabololipidomic fingerprinting was also confirmed by using unbiased principal-component analysis (PCA), showing that each experimental group was associated to a distinct lipid mediator profile and that LXA4 treatment resulted in a normalized lipid mediator signature that closely resembles the control group (Figure 4E). Taken together, these findings indicate that LXA4 treatment either directly and/or indirectly (via the regulation of T cell responses) affects the expression levels of pro-inflammatory lipid mediators during EAE, thereby suppressing ongoing neuro-inflammation in the spinal cord.

DISCUSSION

This study provides critical evidence that boosting a protective resolution response by LXA4 treatment suppresses clinical signs of EAE in vivo and reduces neuro-inflammation, therefore illustrating the therapeutic potential of LXA4. Particularly, we observed a significant reduction in disease severity upon LXA4 treatment as well as a reduction of CD4+ and CD8+ T cell infiltration into the CNS during EAE. Moreover, LXA4 treatment suppressed the pro-inflammatory T cell responses both centrally and peripherally during EAE but also in healthy and MS-derived T cells and blocked their migratory capacity, thereby substantiating the immunomodulatory properties of LXA4 in a human setting. Finally, we provided evidence for substantial lipid mediator alterations in the spinal cord during EAE and show that LXA4 administration normalizes these effects by specifically dampening pro-inflammatory lipid mediators, like prostanoids. Overall, our findings indicate that LXA4 may represent a promising approach to curtail neuro-inflammation and can be potentially considered as a therapeutic tool to limit MS disease progression.

Dysfunction of endogenous resolution programs may contribute to chronic inflammation, a pathological feature of several diseases (Serhan, 2014; Tabas and Glass, 2013). The evidence for both anti-inflammatory and pro-resolving properties of several SPMs, including LXs, indicates that SPMs are powerful pro-resolving lipid mediators that can profoundly affect several aspects associated with neurological disorders, like Alzheimer’s disease (Medeiros et al., 2013; Wang et al., 2015), Parkinson’s disease (Krashia et al., 2019), focal cerebral ischemia (Hawkins et al., 2014; Wu et al., 2010, 2013; Ye et al., 2010), spinal cord injury (Martini et al., 2016), traumatic brain injury (Bisicchia et al., 2018; Luo et al., 2013; Ren et al., 2017), and also EAE (Poisson et al., 2015). Particularly, AT-LXA4 treatment reduced nuclear factor κB (NF-κB) activation and levels of pro-inflammatory mediators, as well as increased levels of anti-inflammatory cytokines, ultimately promoting alternative-activated microglia and reducing AD-like pathology in mice (Medeiros et al., 2013). Moreover, it has been shown that the levels of LXA4 in the cerebrospinal fluid and in the hippocampus of patients with Alzheimer’s disease were lower compared to the level of control subjects and that this decrease was correlated with the degree of cognitive deficit and tissue accumulation of tau protein (Wang et al., 2015). Nevertheless, the potential impact of LXs and other SPMs in pathological aspects of specific CNS disorders, such as MS, is still largely unknown (Bogie et al., 2020). To date, only one study has shown an impaired resolution program in the cerebrospinal fluid of patients with MS (Prüss et al., 2013). However, the aforementioned study has several limitations, mainly due to the limited amount of patient material and the lack of appropriate control groups. Nevertheless, the authors showed that, in particular, LXA4 levels were not increased in patients with highly active MS. Importantly, these findings are in line with our recent observations in plasma of patients with relapsing-remitting MS compared to healthy controls (Kooij et al., 2020), which together prompted us to elucidate the role of LXA4 during neuro-inflammation in more detail in the present manuscript.

Here, we show direct beneficial effects of LXA4 treatment on neuro-inflammation during EAE via reducing clinical signs, limiting CD4+ and CD8+ T cell infiltration into the CNS and dampening central and peripheral Th1 and Th17 effector functions. LXA4 withdrawal during the chronic phase of EAE resulted in a rebound effect, which suggests that presence of LXA4 is required to continuously dampen neuro-inflammation and prevent resurgence of clinical signs of the disease. Our results confirm and extend previous observations where DHA (Kong et al., 2011) and RvD1 (Poisson et al., 2015) ameliorate neuro-inflammation and reduce clinical signs of EAE, as well as recent studies showing LXA4-mediated inhibition of microglial activation and reduction of neuro-inflammation in spinal cord injury (Martini et al., 2016) and ischemic stroke (Hawkins et al., 2017). In a mouse EAE study, RvD1 induced a profound reduction in the number of CD4+ T cells infiltrating the CNS (Poisson et al., 2015). In light of these findings and to address the abundance of T cells in the cerebrospinal fluid of patients with MS (Han et al., 2014) as well as in MS lesions (Lassmann, 2014), we investigated the potential immunomodulatory effect of LXA4 pre-treatment on different T cell subsets from control or patients with relapsing-remitting MS. The observed LXA4-induced reduction of the inflammatory profile in both activated cytotoxic human CD8+ T cells as well as CD4+ T cells is supported by recent findings showing that other SPMs, like RvD1, RvD2, and Mar1, modulate T cell responses (Chiurchiù et al., 2016), suggesting that these SPMs are also involved in the modulation of chronic inflammation (Serhan, 2014) via acting on cells of the adaptive arm of the immune system (Chiurchiù et al., 2016; Hsiao et al., 2015; Poisson et al., 2015; Schwanke et al., 2016). However, T cell subsets from patients with relapsing-remitting MS are less responsive to the immunomodulatory activity of LXA4 compared to T cells from healthy controls. A possible explanation might be that these cells express reduced levels of its receptor ALX/FPR2 (Hodges et al., 2017; Maddox et al., 1997) or show an impaired intracellular signaling. For instance, it was recently reported that T cells of patients with chronic heart failure display reduced levels of another pro-resolving receptor, i.e., GPR32, making them less responsive and dysfunctional to the action of another SPM, i.e., RvD1 (Chiurchiù et al., 2019). Moreover, LXA4 receptor (ALX/FPR2) is also expressed by other cell types involved in the neuro-inflammatory process, including macrophages, astrocytes, and microglia (Chiang and Serhan, 2017), and further research is warranted to define potential LXA4-mediated effects on such cell types.

Besides T cell infiltrates, an EAE/MS lesion is also characterized by glial cell activation, resulting in a pro-inflammatory environment that causes neuronal tissue damage. We here characterized this inflammatory milieu in more detail by performing LC-MS/MS-based lipid mediator metabololipidomics together with authentic biologically derived and synthetic standards to establish a spinal cord lipid mediator signature profile under control and neuro-inflammatory conditions. We identified lipid mediators from each of the ω−6 AA, ω−3 DHA, and EPA bioactive metabolomes and revealed that CNS inflammation predominantly affected the AA bioactive metabolome, confirmed by significantly increased expression levels of pro-inflammatory lipid mediators, like PGE2, PGD2, PGF2a, LTB4, and TXB2. These mediators are functionally characterized in inflammation and its resolution, and their biosynthetic pathways give rise to specific stereochemistry for each of these local mediators (Serhan and Petasis, 2011). Interestingly, we observed that LXA4 treatment exclusively altered the expression levels of such pro-inflammatory lipid mediators during EAE, while leaving other SPMs unaffected. Principal-component analysis further revealed that the lipid mediator signature of various stages of EAE was normalized to a homeostatic signature upon LXA4 treatment. Hence, bioactive lipid mediators are amenable to provide information regarding the potential inflammatory and/or resolution status of a given target tissue (for example, human lymph node tissues as described previously; Colas et al., 2014; Mangalam et al., 2013), which in turn can be used for both diagnosis in terms of biomarkers as well as personalized medicine (Norris and Serhan, 2018).

Our results, together with earlier findings (Chiurchiù et al., 2016; Geem et al., 2015; Luo et al., 2016; Ohse et al., 2004; Poisson et al., 2015; Schwab et al., 2007), suggest that LXA4-mediated neuro-protection during EAE might be the result of different mechanisms, pointing toward a reduction in T cell CNS numbers by affecting their migration capacity as well as downregulation of pro-inflammatory capacities of the T cells. Additionally, we provide evidence that LXA4 may also locally affect the lipidome of EAE spinal cord tissues, either directly or as a result of mechanisms mentioned above. Hence, further research is needed in order to better characterize molecular mechanisms associated with the protective effect of LXA4 treatment during EAE. Finally, optimization of the injection route, doses, therapeutic approach, and mode of action are vital steps to be undertaken toward human clinical translation. Overall, our findings in the experimental model of MS, supported also by the analogous data on patients with MS, suggest that LXA4 treatment can be used to resolve neuro-inflammation and may represent an innovative and safe strategy to manage MS pathogenesis and potentially its progression.

STAR⁎METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact Gijs Kooij (g.kooij@amsterdamumc.nl).

Materials availability

This study did not generate new unique reagents.

Data and code availability

The published article includes all datasets generated or analyzed during this study

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice and EAE induction

All applicable international, national, and institutional guidelines for the care and use of animals were followed. In detail, in vivo experiments were performed in accordance with the Swiss legislation on the protection of animals and were approved by the Veterinary Office of the Canton of Bern (licenses BE 42/14 and BE 28/16). Animal research was performed in accordance to the Swiss legislation on the protection of animals and in compliance with the ARRIVE guidelines for reporting animal research (https://arriveguidelines.org). To investigate the effect of LXA4 treatment on EAE, we used a chemically stable variant called LXA4-methyl ester (LXA4-ME) and disease was induced in eight weeks old C57BL/6J female mice. Animals were purchased from Janvier (France), randomly assigned to different experimental groups of 10 mice and housed in individually ventilated cages (Table S1). Besides water and breeding chow ad libitum, the mice received tissues and paper rolls as animation tools. C57BL/6J female mice were immunized subcutaneously into both flanks and tail root with 200 μg of myelin oligodendrocyte glycoprotein aa35–55 (MOG33–55) emulsified in 100 μL complete Freund’s adjuvant (CFA) supplemented with Mycobacterium tuberculosis. At day zero and day two after immunization, 300 ng of pertussis toxin was applied intraperitoneal (i.p.). A control group was included that received only CFA and pertussis toxin. To study the prophylactic effect of LXA4, EAE mice were injected daily from day eight of EAE with LXA4 i.p. (100 ng/animal in a volume of 0.1 mL of 0.9% saline) until the end of the experiment (day 30). Animal welfare was assessed twice-daily; weight and disease progression was documented once daily and scored as follows: 0.5 (limp tail), 1 (hind leg weakness), 2 (hind leg paraparesis), 3 (hind leg paraparesis and incontinence) (Tietz et al., 2016). Wet chow was provided in a Petri dish and renewed daily as soon as the first mouse reached a clinical score of 1. To investigate whether LXA4 withdrawal causes a rebound effect, weight and clinical scores were continuously assessed until the end of the experiment. At the peak and the end of the experiment, i.e., at day 21 and 30 p.i., mice were deeply anesthetized with isoflurane and perfused transcardial with 1% paraformaldehyde in PBS, pH 7.4. Afterward, brain and spinal cord were collected and processed for immunohistological analysis. For LC-MS/MS and flow cytometry analysis, anesthetized mice were perfused transcardial with PBS, pH 7.4 only.

Human cells

Peripheral blood mononuclear cells (PBMC) from healthy controls (n = 6, all women, mean 33.8 ± 3.7 years) and patients with RR-MS (n = 6, all women, mean age 33.7 ± 6.5 years, see Table S4 for demographics) were separated by density gradient using Ficoll, according to standard procedures. The diagnosis of MS was established at the end of the diagnostic protocol by clinical, laboratory, and magnetic resonance imaging parameters, and matched published criteria. The Expanded Disability Status Scale scores were always < 3 for patients with relapsing-remitting MS and all were treatment-naive at the time of sampling. Healthy control subjects had no history of autoimmune or degenerative diseases of the central or peripheral nervous system. All the subjects gave their written informed consent to the study. The ethics committees of Tor Vergata Hospital and of San Camillo Hospital approved the study.

METHOD DETAILS

Human T-lymphocyte treatments

PBMC from healthy subjects or patients with MS were left untreated or were pretreated with LXA4 (10 nM, the most efficient working dose on immune cells, as reported (Bonnans et al., 2002; Chiurchiù et al., 2016) and then stimulated with Dynabeads CD3/CD28 T Cell Expander (one bead per cell) for eight hours to allow cytokine synthesis and in presence of 10 μg/ml brefeldin-A to inhibit cytokine secretion. Cells were subsequently analyzed by flow cytometry analysis (see below).

Immunohistochemistry

For immunohistochemistry, collected spinal cord was embedded in Tissue-Tek, and frozen in a dry ice-cooled 2-methylbutane bath. Immunofluorescence stainings were performed on 6-μm cryostat sections (n = 3 mice/group) as follows: after fixation in ice-cold acetone, sections were permeabilized with 0.1% Triton X-100, incubated for 30 min with 10% donkey serum and incubated overnight at 4°C with primary antibody against CD45, CD4 and CD8. The primary antibodies were visualized by incubation with donkey anti-rabbit Alexa488 and goat anti-rat Alexa555 for 1h at room temperature (RT). Next, to visualize endothelial cells, sections were first incubated for 30 min with 10% rabbit serum followed by incubation with a biotin-conjugated Collagen IV primary antibody overnight at 4°C. Collagen IV was visualized by incubation with streptavidin Alexa647 for 1h at RT. Sections were treated with Sudan Black (0.3% Sudan Black in 70% ethanol). DAPI was used for nuclear staining and slides were mounted in Mowiol. Slides were scanned on the Vectra Polaris whole-slide scanner (PerkinElmer, USA) at 20x magnification. NIS elements general analysis 3 (Version 5.30.00, Nikon Instruments Inc., USA) was used to automatically quantify the number of all CD45+, CD45+CD4+ and CD45+CD8+ cells both in the whole tissue as well as in the vasculature (marked by Collagen IV). The final density of infiltrating cells was calculated by dividing the number of cells over the total area of the imaged section.

Spinal cord dissociation

At peak of EAE, eight animals per group were anesthetized and transcardially perfused with 10 mL of cold phosphate buffered saline (PBS) pH 7.4. Spinal cords were dissected, minced, and individually processed by enzymatic digestion in 1 mg/ml Collagenase and 1 μl/ml DNase in Hank’s Balanced Salt Solution (HBSS) at 37°C for 30 min. Using a glass potter, pellets were manually homogenized in dissection medium containing 50 mL HBSS, 750 μL HEPES buffer and 650 μL glucose (45%). Infiltrated cells were isolated by 37% Percoll density gradient centrifugation. Lastly, the cells were passed through a 70 μm cell strainer and kept on ice until further analysis.

Flow cytometry analysis

For the flow cytometry analysis of infiltrated T cells in the spinal cord, cells were stained for surface antigens with PE-Cy7-conjugated anti-CD45 (1:80), PerCP-eFluor710-conjugated anti-CD8 (1:100), APC/Fire750-conjugated anti-CD4 (1:100), FITC-conjugated anti-CD11b and anti-Ly6c (1:100), made permeable with BD Cytofix/Cytoperm (Fixation/Permeabilization Solution Kit), and then stained intracellular with BV650-conjugated anti-TNF-α (1:80), APC-conjugated anti-IFN-γ (1:80), PE-conjugated anti–IL-17 (1:40) in BD Perm/Wash buffer at 4°C for 30 min. Dead cells were stained with the fixable viability dye eFluor 506. Spinal cord cells were analyzed by Attune NxT flow cytometer (ThermoFisher) and analyzed using FlowJo software. For mouse T cell flow cytometry analysis, total PBMC were stained for surface antigens with PerCP5.5-conjugated anti-CD4 (1:100), FITC-conjugated anti-CD8 (1:100), made permeable with Cytofix/Cytoperm reagents, and then stained intracellular with PE-Cy7-conjugated anti-TNF-α (1:100), APC-conjugated anti-IFN-γ (1:100), PE-conjugated anti–IL-17 (1:30) in 0.5% saponin at room temperature for 30 min. For human T cell flow cytometry analysis, total PBMCs were stained for surface antigens with APC-e780-conjugated anti-CD14 (1:100), FITC-conjugated anti-CD3 (1:100), e780-conjugated anti-CD4 (1:100), Brilliant Violet-conjugated anti-CD8 (1:150) and ECD-conjugated anti-CD25 (1:40), made permeable with Cytofix/Cytoperm reagents (BD Biosciences), and then stained intracellular with PE-Cy7-conjugated anti-TNF-α (1:100, eBioscience), APC-conjugated anti-IFN-γ (1:100, eBioscience), PE-conjugated anti–IL-17 (1:30, PharMingen) in 0.5% saponin at room temperature for 30 min. The use of these markers allowed us to exclude CD14+ monocytes and gate on lymphoid cells based on CD3 expression and subsequently gating on both CD8+ and CD4+ T cells in order to evaluate the intracellular production of their respective cytokines. The expression of the activation marker CD25 was evaluated on total CD3+ T cells. Mouse and human peripheral blood cells were analyzed by Cytoflex flow cytometer (Beckman Coulter) and analyzed using FlowJo software.

LC-MS/MS based lipid mediator metabololipidomics

For LC-MS/MS analysis, 2ml ice-cold methanol was added to each spinal cord and the samples were profiled using metabololipidomics as described previously (Colas et al., 2014). In short, for lipid mediator profiling deuterated internal standards d8–5S-HETE, d4-LTB4, d5-LXA4, d4-PGE2 and d5-RvD2 (500 pg each) representing each chromatographic region of identified lipid mediators were added to samples to perform quantification and assessment of sample recovery. Proteins were allowed to precipitate at ‒20°C overnight prior to centrifugation (1200 g at 4°C for 10 min) and then supernatants were taken to solid-phase extraction on C18 columns. Products were eluted using methyl formate, brought to dryness under nitrogen and suspended in methanol-water (50:50) for lipid mediators (Colas et al., 2014). The system consisted of a Qtrap 6500 (AB Sciex, Framingham, MA, USA) equipped with a Shimadzu SIL-20AC autoinjector, LC-20AD binary pump (Shimadzu, Kyoto, Japan), and an Agilent Eclipse Plus C18 column (100 mm x 4.6 mm x 1,8 μm). To identify and quantify lipid mediators, multiple reaction monitoring (MRM) was used with signature ion pairs Q1 (parent ion) to Q3 (characteristic daughter ion) for each molecule (Table 1). Identification was conducted by using published criteria with six diagnostic ions and retention time matching of standards (Colas et al., 2014).

Principal component analysis

Principal component analysis was performed using SIMCA 13.0.3 software following mean centering and unit variance scaling of lipid mediator amounts. Principal component analysis serves as an unbiased, multivariate projection designed to identify the systematic variation in a data matrix (the overall bioactive lipid mediator profile of each experimental condition) with lower dimensional plane using score plots and loading plots. The score plot shows the systematic clusters among the observations (closer plots presenting higher similarity in the data matrix). Loading plots describe the magnitude and the manner (positive or negative correlation) in which the measured lipid mediators/SPM contribute to the cluster separation in the score plot.

Transwell migration of human CD4+ and CD8+ T cells

The in vitro CD4+ and CD8+ T cell transendothelial cell migration assay was performed using a Transwell system (Costar, the Netherlands) with polycarbonate filters pore size of 5 μm, which were coated with collagen type 1 (Sigma-Aldrich, the Netherlands). Human brain endothelial cells (hCMEC/D3 cells) (Weksler et al., 2005) were cultured until confluence and subsequently stimulated with TNF-α (5ng/ml) for 24 hours in serum free medium. Human CD4+ or CD8+ T cells were isolated from buffy coat as previously described (Kooij et al., 2010, 2014) and resuspended in serum free medium in the presence of LXA4 (10 nM) for 30 minutes. Ethanol-vehicle treated CD4+ or CD8+ T cells were used as control. Subsequently, CD4+ or CD8+ T cells were added to the transwell filters (1×105 cells/well) and incubated for 4 hours at 37°C and in 5% CO2. To determine the number of migrated cells, transmigrated cells were transferred to FACS tubes, and 20.000 beads (Beckman Coulter, USA) were added to each sample. Samples were analyzed using a FACS Calibur (Becton Dickinson, Belgium) and the number of migrated T cells was determined based on 5.000 gated beads. The absolute number of migrated T cells is presented compared to the total number of T cells added to the upper chamber as described (Kooij et al., 2011). All experiments were performed in triplicate with three different human donors.

QUANTIFICATION AND STATISTICAL ANALYSIS

Results are shown as mean values with standard error of the mean (SEM). Statistical analysis was performed using GraphPad Prism software (v7 GraphPad Software). Comparisons between 2 experimental groups were made by unpaired Student t test and between > 2 groups by one-way ANOVA followed by post hoc Bonferroni correction. Clinical EAE scores and clinical parameters were analyzed by non-parametric (Mann-Whitney). Differences were considered significant at p < 0.05. All statistical tests are described in the figure legends.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
rabbit-anti-CD4	Abcam	Cat# ab183685; RRID: AB_2686917
rat-anti-CD45	BDBioscience	Cat# 553076; RRID: AB_394606
rabbit-anti-CD8	Abcam	Cat# ab217344; RRID: B_2890649
rabbit-anti-collagen IV	Abcam	RRID: AB_305584
donkey-anti-rabbit Alexa488	Molecular Probes	Cat# A21206; RRID: AB_2535792
goat-anti-rat Alexa555	Molecular Probes	Cat# A21434; RRID: AB_141733
Streptavidin Alexa647	Molecular Probes	Cat# S32357; RRID: AB_2336066
anti-mouse CD4 PerCP5.5	Biolegend	Cat# RM4–5; RRID: AB_893325
anti-mouse CD8 FITC	Biolegend	Cat# 5H10–1; RRID: AB_312764
anti-mouse TNF-α PE-Cy7	Biolegend	Cat# MP6-XT22; RRID: AB_2204356
anti-mouse IFN-γ APC	Biolegend	Cat# XMG1.2; RRID: AB_315404
anti-mouse IL-17 PE	Biolegend	Cat# TC11–18H10.1; RRID: AB_315464
anti-mouse CD4 APC/Fire-750	Biolegend	Cat# RM4–5; RRID: AB_2629699
anti-mouse CD8 PerCP-eFluor 710	BD Biosciences	Cat# 53–6.7; RRID: AB_1603266
anti-mouse CD45 PE-Cy7	LSBio	Cat# LS‑C810960
anti-mouse CD11b FITC	BD Biosciences	Cat# M1/70; RRID: AB_394774
anti-mouse Ly6c FITC	Biolegend	Cat# HK1.4; RRID: AB_1186134
anti-mouse TNF-α BV650	BD Biosciences	Cat# MP6-XT22; RRID: AB_2738498
anti-mouse IFN-γ APC	Biolegend	Cat# XMG1.2; RRID: AB_315404
anti-mouse IL-17 PE	Biolegend	Cat# TC11–18H10.1; RRID: AB_315464
anti-human CD4 e780	BD Biosciences	Cat# RPA-T4; RRID: AB_1272044
anti-human CD8 Brilliant Violet 421	Biolegend	Cat# SK1; RRID: AB_2629583
anti-human CD25 ECD	Beckman Coulter	Cat# 6607112
anti-human TNF-α PE-Cy7	BD Biosciences	Cat# Mab11; RRID: AB_1727578
anti-human IFN-γ APC	BD Biosciences	Cat# 4S.B3; RRID: AB_398505
anti-human IL-17 PE	BD Biosciences	Cat# 033–782; RRID: AB_1645512
Biological samples
Human PBMC	Santa Lucia Foundation Biobank (Rome, Italy)	N/A
Chemicals, peptides, and recombinant proteins
Lipoxin A4 (LXA4)	Cayman Chemical	Cat# 90410
Lipoxin A4 methyl ester (LXA4-ME)	Cayman Chemical	Cat# 10033
Myelin oligodendrocyte glycoprotein aa35–55	GenScript	Cat# RP10245
Incomplete Freund’s adjuvant	Santa Cruz	Cat# sc-24648; RRID: AB_10186533
Mycobacterium tuberculosis	BD Biosciences	Cat# 231141
Pertussis toxin	List Biologica Laboratories	Cat# 180
Saline	Braun	Cat# 2132363
Isoflurane	Piramal Healthcare	Cat# NDC 66794–017-25
Paraformaldehyde	Merck	Cat# 1.04005.1000
DPBS	GIBCO	Cat# 14190–094
Ficoll-Paque PLUS	GE Healthcare	Cat# 17–1440-02
Brefeldin-A	Sigma Aldrich	Cat# B7651
Tissue-Tek OCT compound	Sakura Europe	Cat# 25608–930
2-methylbutane	Sigma-Aldrich	Cat# 78–78-4
Triton X-100	Sigma-Aldrich	Cat# 9002–93-1
Normal goat serum	MP biomedicals	Cat# 8642921; RRID: AB_2335031
DAPI	Invitrogen	Cat# D1306; RRID: AB_2629482
Mowiol	Sigma-Aldrich	Cat# 81381
Sudan black	Sigma-Aldrich	Cat# 4197–25-5
collagenase type VIII	Sigma-Aldrich	Cat# C2139
DNase I	Sigma-Aldrich	Cat# 11284932001
Percoll	GE Healthcare	Cat# 17–0891-01
Zombie Near-IR Fixable viability dye	Biolegend	Cat# 423106
eFluor 506 Fixable viability dye	eBioscience	Cat# 65–0866-14
Cytofix/cytoperm	BD Biosciences	Cat# 554715; RRID: AB_2869008
Saponin	Sigma-Aldrich	Cat# 8047–15-2
Methanol	Sigma Aldrich	Cat# 32213
d8–5S-HETE	Cayman Chemical	Cat# 334230
d4-LTB4	Cayman Chemical	Cat# 320110
d5-LXA4	Cayman Chemical	Cat# 10007737
d4-PGE2	Cayman Chemical	Cat# 314010
d5-RvD2	Cayman Chemical	Cat# 11184
C18 columns	Thermo Fisher	Cat# SI928975–906
Methyl formate	Sigma Aldrich	Cat# 291056
Collagen type 1	Sigma-Aldrich	Cat# C9791
Serum free medium	Invitrogen	Cat# A3705001
Beads / flow count fluorospheres	Beckman Coulter	Cat# 7547053
Critical commercial assays
Dynabeads Human T-activator CD3/CD28 for T cell expansion and activation	ThermoFisher	Cat# 11161D
Experimental models: organisms/strains
C57BL/6J female mice	Janvier (France)	N/A
Software and algorithms
Graphpad Prism	https://www.graphpad.com	Version 7
FlowJo software	https://www.flowjo.com	v.10 (TreeStar)
SIMCA software	MKS Data Analytics Solution Umea	13.0.3
Imaris x64 software	Bitplane AG	version 9.1.2