Changes of SPMs in human blood have been proposed as diagnostic markers of health and disease outcomes (Colas et al., 2014; Norris et al., 2018). Postmortem uveitic retinas are not readily available considering uveitis is not a fatal condition, therefore, we analyzed serum samples from 41 healthy subjects and 78 uveitis patients using ELISA to assess whether LXA₄ levels are affected in patients diagnosed with posterior uveitis (Figure 1A). The 78 uveitis patients had posterior uveitis that was either standalone or associated with various primary conditions such as Vogt-Koyanagi-Harada disease (VKH), birdshot chorioretinopathy, panuveitis, intermediate uveitis, choroidopathy and retinopathy. Average serum level of LXA₄ in patients were significantly lower than in healthy subjects (p=0.0135). Most uveitis diagnoses (n = 2 to 23 per diagnosis) had mean serum LXA₄ levels that were lower than healthy controls, with the exception of VKH. However, differences between individual diagnoses and healthy subjects did not reach significance, except for the group of intermediate uveitis (n = 5, p=0.019). The overall decrease in serum LXA₄ level in patients suggests that the LXA₄ pathway is physiologically relevant during pathogenesis of posterior segment uveitis.

To investigate the role of LXA₄ in posterior autoimmune uveitis, we induced EAU in C57BL/6J WT mice (Caspi, 2010; Caspi, 2003) and quantified LXA₄ and pathway-specific metabolite levels in the eye, submandibular lymph nodes, distal lymph nodes and inguinal lymph nodes that drain the immunization sites. Samples were collected from naive and immunized mice at disease onset (day 10) and peak disease (day 16) (Figure 1B and C). LXA₄ and its 5-LOX and 12/15-LOX pathway markers (5-HETE and 15-HETE) were significantly elevated in eyes at peak disease compared to naive unimmunized mice (Figure 1C). By contrast, LXA₄, 5-HETE and 15-HETE levels were significantly downregulated at peak disease in the inguinal lymph nodes (Figure 1B and C). LXA₄ levels did not change in the distal lymph nodes or eye-draining submandibular lymph nodes. Serum was analyzed at onset and peak of EAU (Figure 1—figure supplement 1A) to determine if the induced autoimmune response in mice would replicate changes in serum LXA₄ observed in uveitis patients (Figure 1A). While serum LXA₄ levels in EAU-challenged mice did not change compared to naïve mice, pathway markers 5-LOX and 15-LOX showed significant and progressive decreases during EAU (naïve vs. EAU day 16, p=0.0078 and p=0.0048 for 5-HETE and 15-HETE respectively). Analytes in lipidomic analysis also included DHA- and EPA-derived SPMs and leukotrienes. Pathway markers for DHA-derived SPMs (4-HDHA, 7-HDHA, 14-HDHA and 17-HDHA) were detected in all tissues, but DHA- or EPA- derived SPMs were not robustly detected or were below the signal-to-noise threshold (5:1) in our method. Leukotriene B₄ (LTB₄), a 5-LOX product, was detected in lymph nodes of healthy mice and during the time course of EAU (Figure 1—figure supplement 1B). However, unlike LXA₄, LTB₄ levels did not change significantly in inguinal lymph nodes during EAU pathogenesis. The finding is consistent with our previous lipidomic analysis that identified changes in LXA₄, but not LTB₄, in eye draining lymph nodes of an immune-driven dry eye disease model (Gao et al., 2015; Gao et al., 2018). Altogether, the current findings indicate selective and differential regulation of LXA₄ formation at inductive and effector sites of autoimmunity in EAU.

We next assessed gene expression of the LXA₄ pathway during EAU. Retinas and inguinal lymph nodes were harvested from naïve and immunized mice on day 3, day 7, and day 14 post-immunization. Expression of 5-LOX (Alox5), 12/15-LOX (Alox15), and LXA₄ receptor (Fpr2) was measured by RT-PCR. In the retina, Alox5 expression was upregulated by approximately five-fold on day 14 post-immunization in comparison to earlier time points and to naive mice (Figure 1D), which directly correlated with upregulation of LXA₄ formation in the eye. Alox15 expression remained comparable to naïve mice during the entire time course of EAU, whereas Fpr2 expression changed throughout the time course of EAU, likely reflecting changes in circulating leukocytes populations that express varying levels of Fpr2 in the lymph node, and in temporal regulation of the homeostatic LXA₄ pathway by resident cells in the retina.

In the inguinal lymph nodes, Alox5 expression was significantly upregulated on day 3 and day 14 post-immunization, whereas Fpr2 expression was highly upregulated on day 3, but not on day 14 post-immunization (Figure 1E). Increased RNA levels of Alox5 in inguinal lymph nodes did not correlate with 5-LOX activity and LXA₄ production as both 5-HETE and LXA₄ levels were decreased on day 10 and day 16 (Figure 1C). Since EAU pathogenesis is mediated by retina-specific CD4⁺ T cells, we evaluated Fpr2 expression on CD4⁺ T cells isolated from inguinal lymph nodes and found that the receptor expression was significantly upregulated at peak disease in comparison to CD4⁺ T cells isolated from inguinal lymph nodes of naïve mice (Figure 1F). RNA expression of lipoxygenases (5-LOX and 12/15-LOX) demonstrates selective and temporal modulation of tissue-resident LXA₄ pathway during EAU, while increased Fpr2 expression in CD4⁺ T cells in the inguinal lymph nodes potentiates their ability to interact with LXA₄ during the effector phase of the immune response.

In light of LXA₄ pathway modulation observed during EAU, we investigated whether endogenous LXA₄ deficiency would affect the course of autoimmune uveitis. Alox5^-/- mice are LXA₄ deficient (Aliberti et al., 2002), since 5-LOX is a necessary enzyme required for LXA₄ biosynthesis (Figure 2A). The Alox5^-/- mouse line is an established Jackson Laboratory strain that has been used to study the role of leukotrienes and LXA₄ in vivo. There are no reports of compensatory upregulation of other eicosanoid pathways (cyclooxygenases, 12-LOX or 15-LOX) in this mouse line. Lipidomic analysis did not detect significant differences in basal levels of 12-HETE, 15-HETE, prostaglandin D₂ (PGD₂), prostaglandin E₂ (PGE₂), or thromboxane B₂ (TXB₂) in inguinal lymph nodes of Alox5^-/- mice (Figure 2—figure supplement 1). Low amounts of 5-HETE and LXA₄ were detected in Alox5^-/- mice (Figure 2A). It is possible that one of the other mouse lipoxygenase enzymes, which do not exist in humans (e.g. 8-LOX), is able to catalyze LXA₄ formation in mice with low efficiency. Age-matched Alox5^-/- and WT mice were immunized with IRBP_651-670 peptide (Mattapallil et al., 2015) and EAU pathogenesis was assessed by fundoscopy. Disease scores of Alox5^-/- mice were significantly higher than WT mice starting on day 16 post-immunization, and retinal inflammation in Alox5^-/- persisted while inflammation gradually resolved in WT by day 20 post-immunization (Figure 2B). Representative Micron III fundus images of WT and Alox5^-/- mice revealed more inflammation in the Alox5^-/- eye, with retinal haze resulting from severe cellular infiltration in the retina and vitreous, as opposed to clear vitreous in the WT eye allowing visualization of the underlying retinal folds (Figure 2C). Electroretinography showed that visual function of immunized Alox5^-/- mice was significantly impaired compared to WT mice (Figure 2D), in line with the more severe disease phenotype in Alox5^-/- mice. Flow cytometric analyses of ocular infiltrating cells revealed that there were significantly more cell infiltrates including CD4⁺ T cells and CD4⁺ IFN-γ producing T cells in eyes of Alox5^-/- mice than in WT mice on day 18 post-immunization (Figure 2E,F). Lymphocytes from Alox5^-/- mice exhibited higher antigen-specific proliferative responses to IRBP_651-670 peptide than WT mice, as assessed by ³H thymidine-incorporation assay (Figure 2G). These results demonstrate more severe EAU development in the target tissue and at the site of T cell priming in the absence of LXA₄.

We next assessed if amplifying the LXA₄ pathway could limit EAU progression. To this end, WT mice were immunized with IRBP_651-670 peptide and were treated with either 1 µg of LXA₄ or vehicle control daily, starting on the day of immunization. LXA₄-treated immunized mice developed significantly less disease compared to the vehicle-treated group (Figure 3A). Retinal histology assessed by H and E stained sections confirmed that LXA₄ treatment conferred protection in comparison to vehicle-treated group (Figure 3B and C). The retinas of vehicle-treated group exhibited more severe cellular infiltration that contributed to retinal folds. In contrast, the posterior segment of eyes from LXA₄-treated mice maintained largely normal retinal morphology with no inflammatory exudate resembling that of naïve unimmunized mice, despite presence of intravitreal cell infiltrates consistent with a mild form of uveitis (Figure 3C). Inducible cyclooxygenases-2 (COX-2), a hallmark of many inflammatory diseases, was used as an inflammation marker for the retina. Immunohistochemistry revealed COX-2 localization in the outer plexiform, choroid, retinal pigmented epithelium (RPE) and photoreceptor layers of the eye, while CD4⁺ T cell infiltration was observed in the choroid, RPE and into the photoreceptor layers (Figure 3D). In LXA₄-treated mice, the number of CD4⁺ T cells and the level of COX-2 expression were significantly reduced in comparison to vehicle-treated mice (Figure 3E).

Th1 and Th17 are T cell subsets that drive EAU pathogenesis (Damsker et al., 2010). Flow cytometry analyses of eyes harvested during peak inflammation confirmed that LXA₄ treatment significantly reduced ocular infiltration of pathogenic IFN-γ and IL-17A producing CD4⁺ T cells (Figure 3F and G). Immune cells from inguinal lymph nodes of vehicle-treated mice proliferated dose-dependently after IRBP_651-670 peptide challenge. In contrast, antigen-specific recall responses of lymphocytes from inguinal lymph nodes of LXA₄-treated mice were markedly decreased (Figure 3H).

Since LXA₄ treatment reduced antigen-specific recall responses, we investigated whether LXA₄ could regulate disease during the effector phase. We used in vitro-activated T cells from T cell receptor transgenic mice that express a receptor specific for IRBP_161-180 (R161H) on the B10. RIII background and develop spontaneous uveitis (Horai et al., 2015). Lymph node T cells from CD90.1 congenic R161H mice were cultured with IRBP_161-180 peptide for three days in vitro, which resulted in ~70% of IFN-γ producing T cells (data not shown). IRBP-primed cells were injected into B10.RIII WT CD90.2 recipient mice and received either LXA₄ or vehicle control treatment starting on the day of cell transfer. LXA₄-treated mice developed significantly lower disease in a dose-dependent manner than vehicle-treated mice (Figure 3I). These findings provide direct evidence that LXA₄ inhibits the actions of antigen-primed lymphocytes in vivo and also regulates the effector phase of the adaptive immune response.

To identify the genes that are regulated by LXA₄ treatment, we compared RNA expression in eyes and peripheral lymph nodes from immunized mice treated with LXA₄ or vehicle. Tissues were harvested on day 16 post-immunization and gene expression analyzed by nanoString. Eyes of vehicle-treated mice exhibited increased expression of inflammatory markers Icam1, Il1b, Mapk1, Mapk3, Nfkb1, Ptgs2 (Cox-2), Tgfb, Tnfaip3, and Vcam1 in eyes, whereas LXA₄ treatment downregulated the expression of these markers during EAU. By contrast, expression of these genes was comparable in the inguinal lymph nodes of vehicle vs. LXA₄-treated animals (Figure 4A). These data suggest that systemic LXA₄ treatment reduced expression of inflammatory genes in the effector site, that is eyes, but not in the inductive site, that is lymph nodes draining the site of immunization.

Because LXA₄ treatment reduced T cell migration into the eye (Figure 3G), we examined gene expression of the T cell trafficking marker Ccr7 and its cognate ligands Ccl19 and Ccl21a, as well as the Th1-associated chemokine receptor Cxcr3, and Th17-associated chemokine receptor Ccr6. Compared to vehicle- treated mice, eyes of LXA₄ treated mice had downregulated expression of Ccr7, Cxcr3 and Ccl19 (Figure 4B), which paralleled the reduced numbers of total CD4⁺ T cells and CD4⁺ IFN-γ⁺ T cells in the ocular infiltrate (Figure 3G). In contrast, Ccr7 and Ccl21a were highly upregulated in the inguinal lymph nodes of LXA₄-treated mice compared to controls (Figure 4B). We next assessed whether LXA₄ treatment also impacts expression of S1pr1, which regulates lymphocyte egress from lymphoid tissues (Garris et al., 2014). LXA₄ treatment downregulated expression of S1pr1 in the inguinal lymph nodes that drain the site of immunization, but not in the eye, submandibular lymph nodes that drain the eye, or distal lymph nodes (Figure 4C). To examine whether downregulation of S1pr1 under conditions of LXA₄ deficiency can be detected in uveitis-relevant T cells, we assessed S1pr1 expression in CD4⁺ T cells from inguinal lymph nodes of EAU-challenged Alox5^-/- and WT mice by RT-PCR. S1pr1 expression showed a trend of higher expression in CD4⁺ T cells from Alox5^-/- mice (Figure 4D). The opposing expression of lymphocyte migration-associated receptors and ligands between the eyes and the inguinal lymph nodes identifies a potential mechanism for LXA₄ to control the magnitude of an adaptive immune response, specifically by regulating of Ccr7 and S1pr1 expression that mediate T cell egress from lymph nodes.

Next, we assessed whether LXA₄ deficiency impacts migration of CD4⁺ T cells from secondary lymphoid tissues in vitro. CD4⁺ T cells were isolated from inguinal lymph nodes and spleens of EAU-challenged WT and Alox5^-/- mice on day 13 post-immunization and stimulated in vitro with anti-CD3 and anti-CD28 for 18 hr prior to being subjected to a transwell migration assay. Significantly higher numbers of CD4⁺ T cells from LXA₄-deficient mice migrated toward CCL19 and CCL21 chemokine gradients than CD4⁺ T cells from WT mice (Figure 4E). These results suggest that the absence of tissue-resident LXA₄ increases migration of activated lymph node T cells and may explain exacerbated EAU in LXA₄-deficient mice (Figure 2). To confirm the direct regulation of chemokine receptors on T cells by LXA₄, S1PR1 and CCR7 expression was assessed by flow cytometry on in vitro stimulated T cells treated with LXA₄. CCR7 expression was modestly upregulated (Figure 4F) while S1PR1 expression was downregulated (Figure 4G) in a dose-dependent manner. Together, these data are compatible with the interpretation that LXA₄ acts directly on T cells and modulates trafficking through regulation of S1PR1 and CCR7 expression.

Next, we examined whether enhanced EAU in the absence of LXA₄ is a T cell-mediated effect. We immunized WT and Alox5^-/- donor mice to expand IRBP-specific T cells, and on day 11 purified CD3⁺ T cells from inguinal lymph nodes and spleens. The donor T cells were transferred into TCRb^-/- mice, which lack their own T cells, and three weeks after transfer the recipients were challenged for EAU with IRBP_651-670 peptide (Figure 5A). Consistent with the phenotype of exacerbated EAU in Alox5^-/- mice (Figure 2), recipients of Alox5^-/- T cells developed more severe EAU, with more granulomatous lesions and damage in the photoreceptor cell layer, than recipients of WT T cells (Figure 5B and C). Similar results, more severe EAU, were also observed in TCRb^-/- recipients of Fpr2^-/- T cells (Figure 5—figure supplement 1).

To elucidate effector T cell phenotype and expression of trafficking molecules on uveitogenic CD4⁺ donor T cells, transferred T cells were harvested from eyes and inguinal lymph nodes of TCRb^-/- recipient mice on day 20 post-immunization and analyzed by flow cytometry. Eyes of Alox5^-/- T cell recipients had significantly more total ocular infiltrates, specifically total CD4⁺ T cells and IFN-γ producing CD4⁺ T cells than WT donor T cell recipients (Figure 5D). However, there was no apparent difference in expression of the trafficking markers CXCR3, CCR6 and CCR7 by T cells that had infiltrated the eye (Figure 5E). In the inguinal lymph nodes, the frequency of IFN-γ⁺ donor CD4⁺ T cells from Alox5^-/- mice, but not total or IL-17A⁺ donor CD4⁺ T cells, was higher than that from WT (Figure 5F). Migration marker analysis showed higher frequency of CXCR3⁺ in Alox5^-/- donor CD4⁺ T cells in the inguinal lymph nodes, whereas frequency of activated CD4⁺ CCR7⁺ donor T cells from Alox5^-/- mice was significantly lower (Figure 5G). The downregulation of CCR7 expression and increased ocular infiltration of CD4⁺ T cells support the hypothesis that egress of effector T cells from inguinal draining lymph nodes is augmented in LXA₄-deficient mice.

T cell metabolic programming closely shapes T cell development and responses. Upon TCR stimulation, activated T cells engage in aerobic glycolysis to support effector T cell function (Almeida et al., 2016; Buck et al., 2015). We investigated whether LXA₄ deficiency impacts effector T cell metabolism since our findings indicate that the lymph node resident LXA₄ pathway regulates effector T cell function and migration (Figure 4E and 5G). CD4⁺ T cells were isolated from inguinal lymph nodes and spleens of EAU-challenged WT and Alox5^-/- mice and were stimulated for 18 hr with anti-CD3 and anti-CD28 antibodies. Cell metabolism was analyzed by the Seahorse Glycolytic Rate Assay. To accurately quantify glycolysis-derived extracellular acidification rate, mitochondrial CO₂-contributed acidification was measured and subtracted from total proton efflux rate, yielding the glycolytic proton efflux rate (glycoPER). CD4⁺ T cells from Alox5^-/- mice achieved significantly higher glycoPER at baseline as well as after inhibition of the mitochondrial electron transport chain with rotenone and antimycin A (Rot/AA) (Figure 6A–C). Seahorse Glycolytic Rate assay was also used to assess the direct effect of LXA₄ on T cell glycolysis. CD4⁺ T cells from WT mice were treated in vitro with LXA₄ or vehicle control. LXA₄ treated CD4⁺ T cells had significantly lower glycoPER at base line and after mitochondrial inhibition (Figure 6D), as well as lower basal (Figure 6E) and compensatory (Figure 6F) glycolysis. Thus, both ex vivo evidence from CD4⁺ T cells of EAU-challenged LXA₄-deficient mice and in vitro experiments of LXA₄-treated CD4⁺ WT T cells indicate that lymph node resident LXA₄ has a direct role in regulating metabolic functions of CD4⁺ T cells.

Male and female patients ages 30–76 with clinical diagnosis of non-infectious posterior segment uveitis were enrolled in the National Eye Institute protocol number 16-EI-0046. Healthy controls were NIH blood bank donors of both sexes with a similar age range. Serum samples were obtained from male and female patients ages 30–76 with clinical diagnosis of non-infectious posterior uveitis, healthy controls were NIH blood bank donors of both sexes with a similar age range whose samples were de-identified and sent to the lab. Patients were enrolled from May 2017 to July 2018 under a clinical research protocol (NCT02656381), approved by the institutional review board of the National Institutes of Health. Informed consent (including publishing language as required by NIH IRB) were obtained from all subjects. The study adhered to the tenets of the Declaration of Helsinki.

Serum LXA₄ levels were measured by ELISA (Neogen, KY) per manufacturer instruction and read on SpectraMax iD3 microplate reader (Molecular Devices, CA).

C57BL/6J mice between 8 and 12 weeks of age were purchased from the Jackson Laboratory and were used as the WT strain. All experimental procedures were approved by the Animal Care and Use Committee at University of California, Berkeley (protocol AUP-2016-04-8691-1), and the National Eye Institute at the National Institutes of Health (animal protocol NEI-581) and are in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All animals were maintained on the NIH-31 Open Formula diet: 4.7% Fat, 1.2% monounsaturated fatty acids, 2.1% polyunsaturated fatty acids, with 1.9% C18:2 ω6 linoleic acid and 0.2% C18:3 ω3 linolenic acid (Envigo, WI), starting at 3 weeks of age through the duration of the experiments. For induction of EAU, each mouse was immunized with 150 or 300 µg IRBP_651-670 peptide (Genscript, NJ or BioBasic, NY) as indicated, emulsified in CFA supplemented with 2.5 mg/ml Mycobacterium tuberculosis (Sigma, MO), and injected with 1 µg pertussis toxin (Sigma, MO) as reported earlier (Mattapallil et al., 2015). Alox5^-/- mice (B6.129S2-Alox5^tm1Fun/J, stock number 004155) and TCRb^-/- mice (B6.129P2-Tcrb^tm1Mom/J, stock number 002118) were purchased from the Jackson Laboratory. Fpr2^-/- mice (Dufton et al., 2010) were gifted by Dr. Asma Nusrat from the University of Michigan. For assessing therapeutic effects of LXA₄, either 1 µg/100 µl of LXA₄ (Cayman Chemical, MI), which was resolved in 100% EtOH and diluted to 10% in PBS, or 100 µl of 10% EtOH in PBS (vehicle control) was injected subcutaneously and intraperitoneally starting at day 0 of disease induction. IRBP-specific TCR transgenic R161H mice (Horai et al., 2013) generated on the congenic CD90.1 B10.RIII background and WT CD90.2 B10.RIII mice (Jackson Laboratory, ME) were maintained in house at the NIH animal facility.

Funduscopic scoring system was used to evaluate activity of inflammation in the retina. Mice were anesthetized with ketamine and xylazine (7:3) and the retinas were examined under a binocular microscope after pupil dilation with tropicamide and phenylephrine. Scores ranging from 0 to 4 were assigned to each eye according to the extent and severity of inflammatory lesions (Agarwal et al., 2012). For histological evaluation, eyes were enucleated, pre-fixed in 4% glutaraldehyde for 1 hr, post-fixed in 10% buffered formalin, embedded in paraffin and processed for H and E staining. Scores were determined in a blinded manner on a scale of 0–4 in half-point increments according to published EAU scoring criteria that are based on the type, size, severity and number of lesions (Caspi, 2003).

Mice were dark-adapted overnight, and experiments were performed under dim red illumination. Electroretinograms were recorded using Espion E2 System (Diagnosys LLC, MA) that generated and controlled light stimuli at 10 cds/m² for a-wave and b-wave (Chen et al., 2013). A reference electrode was inserted above the tongue, corneal electrodes were placed at the center of corneas with sterile lubricant eye gel applied to the corneal electrodes. The amplitudes of a-wave and b-wave were analyzed and measured using Espion software. Images was captured using Micron III retinal imaging microscope (Phoenix Research Labs, CA) for small rodents.

Eyes, lymph nodes and spleens were harvested from vehicle and LXA₄-treated immunized mice 16 days post-immunization after extensive perfusion through the heart. Enucleated eyes were trimmed of lenses and connective tissues, minced and treated with 1 mg/ml collagenase D for 40 min at 37°C. Lymph nodes and spleens were mashed, cells filtered through 40 µm strainers and resuspended to make single cell suspensions. Red blood cells in spleens were lysed with ACK lysis buffer. Whole blood was allowed to clot for 30 min at room temperature. Samples were centrifuged and serum supernatant collected for analysis.

RNA was isolated from mouse tissues using RNeasy isolation kit (Qiagen, Germany), and reverse transcribed with High Capacity cDNA kit (Applied Biosystems, CA). Mouse β-actin was used as the endogenous reference gene. PCR was performed using Taqman gene expression assays (Thermo Fisher Scientific, MA) with following primer/probe sets. Alox5: Mm01182747_m1, Alox15: Mm00507789_m1, Fpr2: Mm00484464_s1, S1pr1: Mm00514644_m1.

Whole lymph nodes and eye globes were harvested at disease onset and peak inflammation from mice induced for EAU by immunization with IRBP_651-670 peptide. Tissues from naïve mice served as healthy controls. Tissues were homogenized in 66% MeOH containing deuterated internal standards with Bead Ruptor, and lipid mediators were extracted using solid phase C-18 columns. Mouse serum was combined with 2 volumes of ice cold MeOH containing deuterated internal standards. Diluted serum samples were vortexed and incubated at −80°C for 1 hr to precipitate proteins. Supernatants were collected and extracted using solid phase C-18 columns (Gao et al., 2018). Lipid mediators were then quantified by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Extraction recovery was calculated based on deuterated internal standards (PGE₂-d4, LTB₄-d4, 15-HETE-d8, LXA₄-d5, DHA-d5, AA-d8). The LC-MS/MS system was composed of Agilent 1200 series HPLC, Kinetex C18 minibore column (Phenomenex, CA), and AB Sciex QTRAP 4500 mass spectrometer (SCIEX, MA). Analyses were performed in negative ion mode with scheduled multiple reaction monitoring using 4–5 transition ions per lipid mediator (Livne-Bar et al., 2017; von Moltke et al., 2012).

Single cell suspensions were prepared from isolated tissues. Samples of 1 × 10⁶ cells were stained with anti-mouse antibodies conjugated with following fluorochromes: BV421, BV510/V500, FITC, PE, PerCP-Cy5.5/PerCP-eF710, PE-Cy7, APC, APC Cy7/APC-eF780, BV650/SB645, BV785. Surface staining was performed using antibodies against F4/80, Ly6G, CD3, CD4, CD11b, CD44, CCR6, CXCR3, CCR7, and S1PR1. Ghost Dye UV 450 or Red 780 (Tonbo, CA) were used to exclude dead cells. Intracellular staining was performed after stimulating cells with 50 ng/ml PMA and 1 µg/ml Ionomycin (MilliporeSigma, MA) in the presence of Brefeldin A (Golgi Plug, BD Biosciences, CA) for 4–6 hr at 37°C in cell culture media. Following the surface and viability staining, cells were fixed with 4% paraformaldehyde (PFA) and permeabilized with Triton buffer (0.5% Triton X-100% and 0.1% BSA in PBS). Fluorochrome-conjugated (see above) anti-mouse Foxp3, IL-17A, and IFN-γ were used for intracellular staining. Cells were acquired on BD LSR Fortessa (BD Biosciences, CA), MACSQuant Analyzer (Miltenyi Biotec, Germany), or Cytoflex (Beckham Coulter, CA). Fcs files were analyzed using FlowJo (FlowJo LLC, OR).

Tissues were embedded in OCT and cryosectioned into 7 µm sections. Tissues sections were fixed in 4% PFA and blocked with serum prior to primary antibody staining. Primary antibodies against mouse CD4 (Biolegend, CA) and COX-2 (Abcam, UK) were used. Anti-goat AlexaFluor 488 and AlexaFluor 568 (ThermoFisher Scientific, MA) secondary antibodies were used for immunofluorescent labeling. Cells were counterstained with DAPI at 1:3000 dilution (Sigma, MO). Fluorescent images were acquired with a Zeiss AxioImager microscope at 10X and 20X.

Images were quantified in Fiji software. Particle mean intensity was measured in each fluorescent channel of retina and choroid tissues. Statistical significance was calculated on 3–7 mice per experimental group.

5 × 10⁵ lymph node cells resuspended in HL-1 media (Lonza, GA) were plated in 96-well round bottom plates and stimulated with serial dilutions of IRBP_651-670 peptide. Cells were cultured for 2 days and 1 µCi/well of [³H] thymidine was added and incubated for 16 hr. Samples were harvested and counted by liquid scintillation (Perkin Elmer, MA).

To elicit EAU by adoptive T cell transfer, 5 × 10⁶ IRBP_161-180 in vitro-activated (2 µg/ml for 3 days) lymph node cells from R161H CD90.1 mice were injected intraperitoneally into B10.RIII WT CD90.2 mice. LXA₄ was administered at the doses of 1 µg or 250 ng every day throughout the disease course. Fundoscopy was assessed starting on day four after cell transfer to track disease development. Alternatively, donor C57BL/6 WT and Alox5^-/- mice were immunized with IRBP_651-670 peptide, and CD3⁺ T cells purified with T cell enrichment columns (R and D Systems, MN) from lymph nodes that drain immunization sites and spleens on day 11 post-immunization. WT or Alox5^-/- T cells (1 × 10⁷ each) were transferred intraperitoneally into TCRb^-/- recipients. Three weeks after the cell transfer, TCRb^-/- recipient mice were immunized with 150 µg IRBP_651-670 in CFA and pertussis toxin as described above.

RNA was extracted from snap frozen tissues using RNeasy (Qiagen, Germany) and analyzed by nCounter Analysis Technology using the mouse cancer immunology panel (nanoString, WA). Data were normalized to housekeeping genes and analyzed using nSolver software. Gene expression threshold was set at 30 counts to eliminate background signals.

CD4⁺ T cells were isolated using CD4 T cell isolation kit (STEMCELL Technologies, Canada) from inguinal lymph nodes and spleens of mice immunized with IRBP_651-670 and stimulated in vitro with 2 µg of anti-CD3 and 1 µg of anti-CD28 antibodies for 18 hr. Stimulated T cells were seeded at 2 × 10⁵ per transwell insert (Corning, MA) and allowed to migrate toward 100 ng of CCL19 and 100 ng of CCL21 chemokine ligands (R and D systems, MN) in 2% FCS for 4 hr. 10 nM – 2 μM of LXA₄ was added to stimulated T cells cultured in HL-1 media supplemented with 0.5% normal mouse serum and stained for chemokine receptors. Cells were analyzed on Cytoflex.

Seahorse XFe96 Extracellular Flux Analyzer (Agilent Technologies, CA) was used to measure glycolysis. CD4⁺ T cells were harvested from WT or Alox5^-/- mice immunized with IRBP_651-670 and stimulated with anti-CD3 and anti-CD28 antibodies for 18 hr. The glycolysis proton efflux rate (glycoPER) for each well was calculated from cells subjected to XF Glycolytic Rate assays with the following concentrations of injected compounds: 50 mM 2-DG and 0.5 µM rotenone/antimycin A. For assays with LXA₄ treatment, CD4⁺ T cells were preincubated with 1 µM of LXA₄ or vehicle control for 1 hr before anti-CD3 and anti-CD28 stimulation.

Statistics were performed using GraphPad Prism 8 (GraphPad, CA). Statistical differences were determined by ANOVA, Welch’s t-tests, and Wilcoxon matched-pairs signed rank test for parametric variables and Mann-Whitney test for nonparametric variables. All error bars are standard error mean. p value < 0.05 was considered statistically significant and is denoted with *p≤0.05, **p≤0.01, ***p≤0.001 in the figures.