A novel role for lipoxin A4 in driving a lymph node–eye axis that controls autoimmunity to the neuroretina

The eicosanoid lipoxin A4 (LXA4) has emerging roles in lymphocyte-driven diseases. We identified reduced LXA4 levels in posterior segment uveitis patients and investigated the role of LXA4 in the pathogenesis of experimental autoimmune uveitis (EAU). Immunization for EAU with a retinal self-antigen caused selective downregulation of LXA4 in lymph nodes draining the site of immunization, while at the same time amplifying LXA4 in the inflamed target tissue. T cell effector function, migration and glycolytic responses were amplified in LXA4-deficient mice, which correlated with more severe pathology, whereas LXA4 treatment attenuated disease. In vivo deletion or supplementation of LXA4 identified modulation of CC-chemokine receptor 7 (CCR7) and sphingosine 1- phosphate receptor-1 (S1PR1) expression and glucose metabolism in CD4+ T cells as potential mechanisms for LXA4 regulation of T cell effector function and trafficking. Our results demonstrate the intrinsic lymph node LXA4 pathway as a significant checkpoint in the development and severity of adaptive immunity.


Introduction
Understanding the development of adaptive immune responses to self-antigens is imperative, as prevalence of autoimmunity has increased steadfastly in recent decades (Lerner and Matthias, 2015;Bach, 2002). Posterior autoimmune uveitis accounts for 10-15% of blindness in the Western world and primarily affects working age adults (Durrani et al., 2004;Gritz and Wong, 2004). Autoimmune uveitis is mediated by pathogenic Th1 and Th17 responses to retina-specific antigens, such as interphotoreceptor retinoid-binding protein (IRBP) and retinal arrestin. EAU in mice closely recapitulates human disease. Upon peripheral immunization with retinal antigens in complete Freund's adjuvant (CFA), T cells are primed in the inguinal lymph nodes that drain the immunization sites. Activated retina-specific T cells then traffic to the eye and elicit inflammation that starts 9-11 days and peaks 14-16 days after immunization (Caspi, 2010).
LXA 4 is an endogenous arachidonic acid-derived signaling molecule that exerts anti-inflammatory and pro-resolving activities via formyl peptide receptor 2 (ALX/FPR2) in mice and humans. To control the severity and duration of acute inflammation, cells expressing 5-lipoxygenase (5-LOX) and 12/15lipoxygenase (12/15-LOX) oxygenate arachidonic acid in tandem to synthesize LXA 4 locally. 5-LOX is the only enzyme in mammals that can generate the 5,6 epoxide intermediate leukotriene A 4 required for the formation of LXA 4 and other epoxide intermediates of several specialized proresolving mediators (SPMs) Serhan, 2014). Even though 5-LOX also initiates the generation of pro-inflammatory leukotrienes by leukocytes, 5-LOX in the retina has protective roles (Sapieha et al., 2011;Livne-Bar et al., 2017). Emerging evidence indicates that LXA 4 is an important tissue-resident pathway in eyes and draining lymph nodes that controls ocular surface inflammation and immune responses, and it is neuroprotective in the retina Livne-Bar et al., 2017;Kenchegowda and Bazan, 2010).
The defining functions of LXA 4 include inhibition of neutrophil infiltration to the site of inflammation, promotion of macrophage efferocytosis, and reduction of inflammatory cytokine and chemokine production (Serhan, 2014;Serhan et al., 2008). In vitro studies demonstrated that LXA 4 and SPMs have a role in T follicular and helper cell differentiation (Nagaya et al., 2017;Chiurchiù et al., 2016), and neutrophil-derived LXA 4 was previously shown to inhibit T effector cell responses in lymph nodes in immune-driven dry eye disease (Gao et al., 2015). ALX/FPR2 is expressed on T cells (Ariel et al., 2003), providing a modality for LXA 4 regulation of adaptive immunity. However, role of LXA 4 and its direct action on T lymphocytes in autoimmune responses has been minimally explored.
Specific aim of this study was to uncover whether and how LXA 4 regulates T effector cell function in the pathogenesis of autoimmune uveitis. Here, we report that expression of the homeostatic LXA 4 pathway and endogenous LXA 4 formation are dynamic and temporally defined in the target tissue, that is the eye, and the inciting lymph nodes that drain the site of immunization. T cells from LXA 4deficient (Alox5 -/-) mice exhibited enhanced effector function with augmented cytokine production and altered metabolism that resulted in exacerbated disease, while LXA 4 treatment in wild type (WT) mice attenuated development of autoimmune uveitis. The resident LXA 4 pathway in draining lymph nodes was identified as a significant regulator of CCR7 and S1PR1, which control egress of T effector cells from lymph nodes. Lastly, LXA 4 deficiency as well as LXA 4 treatment regulated T cell glycolysis, an essential metabolic pathway required for T effector cell function. Our findings identify novel mechanisms and a homeostatic role for LXA 4 in lymph nodes to control the development of adaptive autoimmune responses.

Results
LXA 4 is generated during autoimmune uveitis in a time-and sitedependent manner 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 LXA 4 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 LXA 4 in patients were significantly lower than in healthy subjects (p=0.0135). Most uveitis diagnoses (n = 2 to 23 per diagnosis) had mean serum LXA 4 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 LXA 4 level in patients suggests that the LXA 4 pathway is physiologically relevant during pathogenesis of posterior segment uveitis.
To investigate the role of LXA 4 in posterior autoimmune uveitis, we induced EAU in C57BL/6J WT mice (Caspi, 2010;Caspi, 2003) and quantified LXA 4 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). LXA 4 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, LXA 4 , 5-HETE and 15-HETE levels were significantly downregulated at peak disease in the inguinal lymph nodes ( Figure 1B and C). LXA 4 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 LXA 4 observed in uveitis patients ( Figure 1A). While serum LXA 4 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 DHAderived 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 B 4 (LTB 4 ), 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 LXA 4 , LTB 4 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 LXA 4 , but not LTB 4 , 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 LXA 4 formation at inductive and effector sites of autoimmunity in EAU. We next assessed gene expression of the LXA 4 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 LXA 4 receptor (Fpr2) was measured by RT-PCR. In the retina, Alox5 expression was upregulated by approximately five-fold on day 14 postimmunization in comparison to earlier time points and to naive mice ( Figure 1D), which directly correlated with upregulation of LXA 4 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 LXA 4 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 LXA 4 production as both 5-HETE and LXA 4 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 LXA 4 pathway during EAU, while increased Fpr2 expression in CD4 + T cells in the inguinal lymph nodes potentiates their ability to interact with LXA 4 during the effector phase of the immune response.

Endogenous LXA 4 limits development and progression of EAU
In light of LXA 4 pathway modulation observed during EAU, we investigated whether endogenous LXA 4 deficiency would affect the course of autoimmune uveitis. Alox5 -/mice are LXA 4 deficient (Aliberti et al., 2002), since 5-LOX is a necessary enzyme required for LXA 4 biosynthesis ( Figure 2A). The Alox5 -/mouse line is an established Jackson Laboratory strain that has been used to study the role of leukotrienes and LXA 4 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 D 2 (PGD 2 ), prostaglandin E 2 (PGE 2 ), or thromboxane B 2 (TXB 2 ) in inguinal lymph nodes of Alox5 -/mice ( Figure 2-figure supplement 1). Low amounts of 5-HETE and LXA 4 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 LXA 4 formation in mice with low efficiency. Agematched 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-g 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 3 H 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 LXA 4 .

LXA 4 treatment attenuates development of uveitis
We next assessed if amplifying the LXA 4 pathway could limit EAU progression. To this end, WT mice were immunized with IRBP 651-670 peptide and were treated with either 1 mg of LXA 4 or vehicle control daily, starting on the day of immunization. LXA 4 -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 LXA 4 treatment conferred protection in comparison to vehicletreated 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 LXA 4treated 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 LXA 4 -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 LXA 4 treatment significantly reduced ocular infiltration of pathogenic IFN-g 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 LXA 4 -treated mice were markedly decreased ( Figure 3H).
Since LXA 4 treatment reduced antigen-specific recall responses, we investigated whether LXA 4 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-g producing T cells (data not shown). IRBP-primed cells were injected into B10.RIII WT CD90.2 recipient mice and received either LXA 4 or vehicle control treatment starting on the day of cell transfer. LXA 4 -treated mice developed significantly lower disease in a dose-dependent manner than vehicle-treated mice ( Figure 3I). These findings provide direct evidence that LXA 4 inhibits Figure 2 continued adaptation, n = 10 eyes per group. (E) Representative flow plots of CD4 + vs. CD8 + and IFN-g vs. IL-17A producing CD4 + cells in the eyes of WT and Alox5 -/mice. (F) Flow cytometry analyses of total numbers of ocular infiltrates in EAU-challenged WT and Alox5 -/gated on live cells, CD4 + , CD4 + CD44 + , CD4 + IFN-g + , CD4 + IL-17A + populations, n = 9-10 mice per group. (G) Representative bar graphs of antigen-specific cell proliferation performed on inguinal lymph nodes harvested on day 18 post-immunization from WT and Alox5 -/mice, pooled n = 6-9 mice per group. Showing one representative from three experiments with the same trend. *p<0.05, **p<0.01, Wilcoxon matched-pairs signed rank test and unpaired t test. The online version of this article includes the following figure supplement(s) for figure 2: the actions of antigen-primed lymphocytes in vivo and also regulates the effector phase of the adaptive immune response.

LXA 4 modulates T cell trafficking in inguinal lymph nodes that drain the immunization site
To identify the genes that are regulated by LXA 4 treatment, we compared RNA expression in eyes and peripheral lymph nodes from immunized mice treated with LXA 4 or vehicle. Tissues were harvested on day 16 post-immunization and gene expression analyzed by nanoString. Eyes of vehicletreated mice exhibited increased expression of inflammatory markers Icam1, Il1b, Mapk1, Mapk3, Nfkb1, Ptgs2 (Cox-2), Tgfb, Tnfaip3, and Vcam1 in eyes, whereas LXA 4 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. LXA 4 -treated animals ( Figure 4A). These data suggest that systemic LXA 4 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 LXA 4 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 LXA 4 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-g + T cells in the ocular infiltrate ( Figure 3G). In contrast, Ccr7 and Ccl21a were highly upregulated in the inguinal lymph nodes of LXA 4 -treated mice compared to controls ( Figure 4B). We next assessed whether LXA 4 treatment also impacts expression of S1pr1, which regulates lymphocyte egress from lymphoid tissues (Garris et al., 2014). LXA 4 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 LXA 4 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 LXA 4 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 LXA 4 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 EAUchallenged 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 LXA 4 -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 LXA 4 increases migration of activated lymph node T cells and may explain exacerbated EAU in LXA 4 -deficient mice (Figure 2). To confirm the direct regulation of chemokine receptors on T cells by LXA 4 , S1PR1 and CCR7 expression was assessed by flow cytometry on in vitro stimulated T cells treated with LXA 4 . 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 LXA 4 acts directly on T cells and modulates trafficking through regulation of S1PR1 and CCR7 expression.

Disease-limiting effect of LXA 4 is T cell-mediated
Next, we examined whether enhanced EAU in the absence of LXA 4 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-g 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-g + 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 LXA 4 -deficient mice.
LXA 4 pathway regulates the metabolism of CD4 + T cells 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 LXA 4 deficiency impacts effector T cell metabolism since our findings indicate that the lymph node resident LXA 4 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 CO 2 -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 LXA 4 on T cell glycolysis. CD4 + T cells from WT mice were treated in vitro with LXA 4 or vehicle control. LXA 4 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 LXA 4 -deficient mice and in vitro experiments of LXA 4 -treated CD4 + WT T cells indicate that lymph node resident LXA 4 has a direct role in regulating metabolic functions of CD4 + T cells.

Discussion
In this study, we demonstrate that the resident LXA 4 pathway in lymph nodes is important in regulating T cell effector functions during development and progression of ocular autoimmunity. A key finding is the downregulation of basal LXA 4 levels in lymph nodes that accompany the adaptive immune response in a mouse model of autoimmune uveitis. Reduced levels of serum LXA 4 were also identified in serum of human patients with posterior uveitis diagnoses, suggesting that the LXA 4 pathway is dynamically regulated in human uveitis. In contrast, LXA 4 levels were elevated in the uveitic eyes of mice, which is consistent with the established role of LXA 4 as an induced counter-regulatory (D) Fold change in S1pr1 expression of CD4 + T cells isolated from WT and Alox5 -/on day 14 post-immunization, n = 11 from three experiments combined. (E) Transwell migration assay of CD4 + T cells isolated from immunized WT and Alox5 -/on day 13 post-immunization. CD4 + T cells were pre-stimulated with anti-CD3 and anti-CD28 antibodies for 18 hr and transferred to transwell culture plates and incubated for 4 hr in the absence or presence of CCL19 and CCL21, n = 12 per group. One representative experiment of 4 showing the same trend. **p<0.01, unpaired Welch's t test. (F-G) Representative flow plots of CCR7 and S1PR1 expression on in vitro anti-CD3 and anti-CD28 stimulated CD4 + T cells treated with vehicle control or 10 nM, 100 nM, 1 mM, and 2 mM of LXA 4 . Cells were gated on live single cells then CD4 + cells. Gray overlays indicate fluorescence minus one controls. Showing one representative experiment out of three. mediator that is generated at sites of inflammation. Despite the established role of LXA 4 in innate immune cell regulation, direct regulation of adaptive immune responses by endogenous LXA 4 is not well defined in vivo, as most research efforts have focused on LXA 4 's ability to dampen and resolve acute inflammation elicited by innate immune cells. Beneficial anti-inflammatory effects of LXA 4 have been previously reported in the LPS-induced anterior uveitis model, however, pathogenesis of LPSinduced uveitis is driven exclusively by innate immune responses (Karim et al., 2009). In a murine asthma model induced by allergen immunization, LXA 4 treatment reduced airway inflammation by increasing natural killer cell-mediated eosinophil apoptosis (Barnig et al., 2013), demonstrating LXA 4 regulation of innate responses in vivo. Our data indicate that downregulated endogenous LXA 4 , specifically in inguinal lymph nodes that drain the immunization site, is a key feature of  developing adaptive immunity, that is EAU. This is consistent with a previous report, which linked lower LXA 4 levels in draining lymph nodes to exacerbated dry eye disease in female mice (Gao et al., 2015). LXA 4 -deficient Alox5 -/mice and adoptive transfer in T cell-deficient mice also demonstrated that in vivo disruption of the LXA 4 pathway amplifies effector T cell responses in autoimmune uveitis. These experiments provide strong evidence for LXA 4 regulation of adaptive immune responses initiated in lymph nodes by demonstrating that (a) LXA 4 deficiency causes dysregulated T effector cell function that leads to exacerbated autoimmune pathogenesis and (b) supplementation with exogenous LXA 4 reverses this effect. Thus, we identified a new role for LXA 4 in T cell-driven adaptive autoimmunity and uncovered a lymph node -eye axis that regulates this process. It is not possible to capture changes in LXA 4 tissue levels after a low dose (1 mg) injection of LXA 4 at the time of tissue harvest since increases of LXA 4 in eye or lymph nodes are likely rapid and of short duration. LXA 4 , like all eicosanoids, is enzymatically inactivated in minutes when not bound to an endogenous carrier protein such as albumin. However, several research groups have established that tail vein, intraperitoneal, subcutaneous or subconjunctival injections of LXA 4 or other SPMs (10 ng-1 mg) in mouse disease models induce long-lasting protective actions through changes in lymphocyte and neutrophil function and macrophage phenotype at distant tissue sites (Livne-Bar et al., 2017;Serhan et al., 2008;Gao et al., 2015;Fosshaug et al., 2019;Wei and Gronert, 2019), which indicate direct actions at the target tissue.
Our findings provide evidence for LXA 4 regulation of T cell trafficking as a potential mechanism for controlling EAU progression. The S1PR1-sphingosine-1 phosphate axis is a primary mechanism for regulating lymph node T cell egress and lymphocyte recirculation. LXA 4 treatment selectively downregulated S1pr1 expression only in inguinal lymph nodes of WT EAU-challenged mice. Pharmacological downregulation of S1PR1 prevents trafficking of autoreactive T cells and is an FDAapproved treatment for multiple sclerosis (Garris et al., 2014). Hence, a role of the lymph node LXA 4 pathway in regulating S1pr1 expression in lymphocytes warrants further investigation into the crosstalk between these two classes of lipid mediators. The CCR7-CCL19/CCL21 axis facilitates naïve T cell homing to secondary lymphoid organs, activated T cell egress, and the interaction between T cells and dendritic cells (DC) to induce peripheral tolerance or to elicit an adaptive immune response (Comerford et al., 2013). LXA 4 treatment differentially regulated this axis at the effector and initiation sites of the adaptive immune response by reducing Ccr7 and Ccl19 expression in the eye and increasing Ccr7 and Ccl21 expression in inguinal lymph nodes.
The endogenous role of LXA 4 in regulating effector T cell function through Ccr7 was confirmed with Alox5 -/mice using T cell adoptive transfer and in vitro migration experiments. LXA 4 deficiency was coincident with dysregulated (i.e. lower) expression of CCR7 in T cells from inguinal lymph nodes, increased IFN-g production and an almost two-fold increase of CD4 + T cell migration towards CCL19/21 gradient. Ccr7 is expressed predominantly in DC and T cells, and its ligands CCL19 and CCL21 can be secreted by several cell types. In this study, we did not investigate the cellular source of CCL19 and CCL21 in tissues, the effect of LXA 4 on Ccr7 expression in DC or assess direct LXA 4 regulation of Ccr7 expression on T cells. LXA 4 has established actions with DC and other cell types, therefore, we cannot rule out that LXA 4 also controls T cell responses in inguinal lymph nodes through priming by DC and/or regulation of CCL19 and CCL21 production by other cell types.
T cell glycolysis is vital in regulating T cell effector function. Metabolic switch from oxidative phosphorylation to glycolysis is required to sustain T effector function and cytokine production (Coe et al., 2014;Salmond, 2018). Inhibition of glycolysis abrogates IFN-g production in activated CD4 + T cells, implicating glycolysis as regulator of sustained T effector responses (Chang et al., 2013). LXA 4 treatment decreased CD4 + T cells glycolytic responses and reduced IFN-g production from CD4 + T cells, providing evidence of direct regulation of T cell function by LXA 4 . Increased glycolytic responses of CD4 + T cells from Alox5 -/mice underscores the relevance of lymph node resident LXA 4 in controlling T cell effector function. T cell glycolysis and mitochondrial metabolism as potential therapeutic targets are supported by elevated glycolytic metabolism in T cells of systemic lupus erythematosus (SLE) patients and reduced IFN-g production and disease biomarkers upon inhibition of CD4 + T cell glycolysis in mouse models of SLE (Yin et al., 2015). The ability of LXA 4 to downregulate glycolysis in activated CD4 + T cell is of interest and suggests that the lymph node resident LXA 4 pathway is a potential checkpoint to limit inflammatory cytokine production and the magnitude of adaptive immune responses.
In mice and humans, LXA 4 cognate receptor ALX/FPR2 is expressed in T cells, B cells and innate lymphoid cells (Wei and Gronert, 2019), and has been shown to regulate CD4 + T cell cytokine production in vitro (Chiurchiù et al., 2016). During peak EAU, Fpr2 expression in CD4 + T cells was highly upregulated in the inguinal lymph nodes that drain the site of immunization, suggesting that increased capacity of LXA 4 signaling in T cells is a counter-regulatory mechanism in response to the development of adaptive immunity. Consistent with our hypotheses that dysregulated LXA 4 signaling allows for development of EAU and facilitates augmented effector T cell functional responses, adoptive transfer of either Alox5 -/or Fpr2 -/-T cells to TCRb -/recipients caused exacerbated EAU ( Figure 5-figure supplement 1A). Transferred T cells from Fpr2 -/and Alox5 -/mice induced different signatures of inflammatory ocular infiltrates ( Figure 5-figure supplement 1B and C), likely due to differential regulation of donor T cells and innate effector cells by the host LXA 4 pathway in TCRb -/recipient mice. However, unlike Alox5 -/mice, global Fpr2 -/mice (which generate LXA 4 ) did not develop more severe EAU than WT controls after immunization ( Figure 5-figure supplement  1D). A second LXA 4 receptor GPR32 has been identified in humans (Dalli and Serhan, 2019), which currently has no mouse homolog. Our data suggest that additional receptors for LXA 4 may exist in mice. Incomplete abrogation of LXA 4 signaling due to a putative second receptor may explain why immunized global Fpr2 -/mice did not develop more severe autoimmune disease.
EAU is an induced and well characterized model of autoimmune uveitis that recapitulates features of human posterior autoimmune uveitis. Although the immune-regulatory roles of LXA 4 in human draining lymph nodes remain to be defined, LXA 4 has been identified as an abundant SPM in human axillary lymph nodes (Colas et al., 2014). While it is important to note that there are significant differences in mouse and human immune responses (Seok et al., 2013), LXA 4 pathway and receptors are conserved in humans and mice and are targeted as novel therapies based on promising results from animal studies.
In summary, our data established that the basal tone of LXA 4 in lymph nodes is selectively downregulated in the draining lymph node during development of an adaptive immune response. We identified regulation of CCR7 and S1PR1 expression and glycolytic responses in CD4 + T cells as new mechanisms for LXA 4 to control effector T cell functions and egress from lymph nodes, which establishes LXA 4 as an important homeostatic regulator for fine tuning adaptive immune responses. Thus, amplification of the resident lymph node LXA 4 pathway may be a potential target to prevent the development of pathogenic adaptive immune responses and autoimmunity.

Patients and healthy controls
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 ELISA assay
Serum LXA 4 levels were measured by ELISA (Neogen, KY) per manufacturer instruction and read on SpectraMax iD3 microplate reader (Molecular Devices, CA).

Mice, EAU induction and LXA 4 treatment
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 w6 linoleic acid and 0.2% C18:3 w3 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 mg 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 mg 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 LXA 4 , either 1 mg/100 ml of LXA 4 (Cayman Chemical, MI), which was resolved in 100% EtOH and diluted to 10% in PBS, or 100 ml of 10% EtOH in PBS (vehicle control) was injected subcutaneously and intraperitoneally starting at day 0 of disease induction. IRBPspecific TCR transgenic R161H mice ) 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.

Evaluation of EAU by fundoscopy and histology
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).

Electroretinography and fundus imaging
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/m 2 for a-wave and b-wave . 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.

Tissue harvest and processing
Eyes, lymph nodes and spleens were harvested from vehicle and LXA 4 -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 mm 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.

Real-time quantitative PCR
RNA was isolated from mouse tissues using RNeasy isolation kit (Qiagen, Germany), and reverse transcribed with High Capacity cDNA kit (Applied Biosystems, CA). Mouse b-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.

Detection of endogenous formation of eicosanoids
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 À80˚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 (PGE 2 -d4, LTB 4 -d4, 15-HETE-d8, LXA 4 -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).

Immunohistochemistry
Tissues were embedded in OCT and cryosectioned into 7 mm 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 Alex-aFluor 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.

Quantitation of immunohistochemistry images
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.
T cell proliferation assays 5 Â 10 5 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 mCi/well of [ 3 H] thymidine was added and incubated for 16 hr. Samples were harvested and counted by liquid scintillation (Perkin Elmer, MA).

Adoptive transfer of retina-specific T cells and LXA 4 treatment
To elicit EAU by adoptive T cell transfer, 5 Â 10 6 IRBP 161-180 in vitro-activated (2 mg/ml for 3 days) lymph node cells from R161H CD90.1 mice were injected intraperitoneally into B10.RIII WT CD90.2 mice. LXA 4 was administered at the doses of 1 mg 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 Â 10 7 each) were transferred intraperitoneally into TCRb -/recipients. Three weeks after the cell transfer, TCRb -/recipient mice were immunized with 150 mg IRBP 651-670 in CFA and pertussis toxin as described above.
NanoString assays for transcriptomic analysis 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.
Transwell migration assay and in vitro T cell assay 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 mg of anti-CD3 and 1 mg of anti-CD28 antibodies for 18 hr. Stimulated T cells were seeded at 2 Â 10 5 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 mM of LXA 4 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 metabolic assays
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 mM rotenone/antimycin A. For assays with LXA 4 treatment, CD4 + T cells were preincubated with 1 mM of LXA 4 or vehicle control for 1 hr before anti-CD3 and anti-CD28 stimulation.

Statistical analysis
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.