Abstract

   Systemic lupus erythematosus (SLE) is a complex autoimmune disease with
   abnormal activation of the immune system. Recent attention is
   increasing about how aberrant lipid and cholesterol metabolism is
   linked together with type I interferon (IFN‐I) signaling in the
   regulation of the pathogenesis of SLE. Here, a metabonomic analysis is
   performed and increased plasma concentrations of oxysterols, especially
   7α, 25‐dihydroxycholesterol (7α, 25‐OHC), are identified in SLE
   patients. The authors find that 7α, 25‐OHC binding to its receptor
   Epstein–Barr virus‐induced gene 2 (EBI2) in macrophages can suppress
   STAT activation and the production of IFN‐β, chemokines, and cytokines.
   Importantly, monocytes/macrophages from SLE patients and mice show
   significantly reduced EBI2 expression, which can be triggered by IFN‐γ
   produced in activated T cells. Previous findings suggest that
   EBI2 enhances immune cell migration. Opposite to this effect, the
   authors demonstrate that EBI2‐deficient macrophages produce higher
   levels of chemokines and cytokines, which recruits and activates
   myeloid cells,T and B lymphocytes to exacerbate
   tetramethylpentadecane‐induced SLE. Together, via sensing the oxysterol
   7α, 25‐OHC, EBI2 in macrophages can modulate innate and adaptive immune
   responses, which may be used as a potential diagnostic marker and
   therapeutic target for SLE.

   Keywords: Epstein–Barr virus‐induced gene 2, interferon, macrophages,
   oxysterols, systemic lupus erythematosus
     __________________________________________________________________

   This study has identified the increased plasma 7α,
   25‐dihydroxycholesterol in patients with systemic lupus erythematosus
   (SLE), which binds to Epstein–Barr virus‐induced gene 2 (EBI2) to
   inhibit the type I interferon (IFN‐I) response in macrophages. IFN‐γ
   stimulation reduces EBI2  expression to induce increased levels of
   IFN‐I and chemokines, resulting in massive infiltration and activation
   of T/B lymphocytes to exacerbate SLE.

   graphic file with name ADVS-10-2207108-g008.jpg

1. Introduction

   Systemic lupus erythematosus (SLE) is a complex autoimmune disease with
   different pathogenic mechanisms, leading to extremely varied clinical
   manifestations, including kidney damage, arthritis, skin disease, blood
   cell abnormalities, and neurological complications.^[ [64]^1 ^]
   Alongside genetic factors, environmental triggers, and hormones,^[
   [65]^2 ^] recent studies have highlighted the role of mitochondrial
   dysfunction, defective clearance of apoptotic cells, or inefficient
   degradation of DNA‐containing neutrophil extracellular traps (NETs) in
   the initiation of SLE.^[ [66]^3 ^] Notably, abnormal release of nuclear
   acids and cellular contents from damaged mitochondria, apoptotic cells,
   or NETs could be sensed by pattern recognition receptors (PRRs) to
   activate the innate immune system, leading to type I interferon (IFN‐I)
   production. A rise in serum IFN‐I levels has been noted in SLE
   patients,^[ [67]^4 ^] and IFN‐I is positively correlated with the SLE
   disease activity index (SLEDAI) scores and anti‐dsDNA (double‐stranded
   DNA) levels.^[ [68]^1b ^] Moreover, many SLE patients display an
   increased expression of IFN‐stimulated genes (ISGs) (referred as IFN
   signature) in multiple types of immune cells.^[ [69]^5 ^] Critically,
   anifrolumab, a monoclonal antibody antagonist of the IFN‐I receptor
   (IFNAR), was recently approved in the United States for the treatment
   of adult patients with moderate to severe SLE who are receiving
   standard therapy.^[ [70]^6 ^] Hence, IFN‐I and IFN‐I‐induced responses
   have drawn increasing attention to understand SLE pathogenesis.

   Macrophages are located in multiple organs or tissues, which can
   respond to diverse stimuli to polarize into different subsets, that is,
   producing pro‐inflammatory or anti‐inflammatory cytokines, and
   expressing different receptors to activate or inhibit T cells. This
   macrophage function plasticity could impact the development of SLE.^[
   [71]^7 ^] Upon recognizing exogenous nuclear acids or endogenous
   inappropriately aggregated self‐DNA, PRRs such as toll‐like receptors
   (TLRs), cytosolic cGAS (cyclic guanosine monophosphate‐adenosine
   monophosphate [cGAMP] synthase), or RIG‐I‐like receptors were activated
   to trigger a series of signaling cascades for the activation of the
   transcription factors IFN‐regulatory factor 3 (IRF3) and nuclear
   factor‐κB (NF‐κB) to produce IFN‐I and pro‐inflammatory cytokines.^[
   [72]^8 ^] IFN‐I then activates the IFNAR‐mediated signaling pathway,
   leading to the transcription of a wide range of ISGs, including
   multiple chemokines that further recruit numerous types of immune cells
   to amplify SLE. Accordingly, we aimed to investigate key molecules in
   macrophages that can precisely control IFN‐I and ISGs during the
   development of SLE, which might introduce potential targets for the
   diagnosis or treatment of SLE.

   Mounting evidences have shown that dyslipidemia is prevalent in SLE
   patients.^[ [73]^9 ^] About 30% SLE patients exhibit
   hypercholesterolemia, which rises to over 60% 3 years after
   diagnosis.^[ [74]^10 ^] Sex hormones, derivatives of cholesterol, have
   been demonstrated to closely related to the development and
   susceptibility of SLE.^[ [75]^11 ^] Excitingly, several advanced
   studies have indicated that cholesterol or its intermediates could
   regulate IFN‐I response. Opposite to that reducing cholesterol
   biosynthesis flux could enhance IFN‐I and ISGs expression in
   macrophages,^[ [76]^12 ^] our previous study revealed that the
   accumulation of the cholesterol precursor 7‐dehydrocholesterol (7‐DHC)
   could promote IFN‐I production in macrophages against viral
   infection.^[ [77]^13 ^] Consequently, it is crucial to identify which
   cholesterol metabolites are abnormally produced and are functionally
   linked with IFN and ISGs to control the pathogenesis of SLE.

   In this study, we conducted an assessment of the expression levels of
   enzymes and metabolites relevant to cholesterol synthesis, uptake,
   efflux, and esterification. The plasma concentration of 7α,
   25‐dihydroxycholesterol (7α, 25‐OHC) was markedly increased along with
   the enhanced expression of the enzyme 25‐hydroxycholesterol
   7alpha‐hydroxylase (CYP7B1) in peripheral blood mononuclear cells
   (PBMCs) of SLE patients. 7α, 25‐OHC was identified as the most potent
   endogenous ligand for the G protein‐coupled receptor (GPCR)
   Epstein–Barr virus‐induced gene 2 (EBI2, also known as GPR183).^[
   [78]^14 ^] EBI2 and 7α, 25‐OHC have been elucidated to promote
   migration of immune cells to regulate many physiological and
   pathological processes. EBI2 mediates migration of T cells to the outer
   T zone to augment follicular helper T (T[FH]) cell differentiation,^[
   [79]^15 ^] and promotes migration of encephalitogenic T[H]17 cells into
   the central nervous system to aggravate experimental autoimmune
   encephalomyelitis.^[ [80]^16 ^] Additionally, EBI2 directs B cells to
   the outer follicular and induces early plasmablast responses.^[ [81]^17
   ^] Since abnormal production of IFN and activation of T/B cells are
   important for the development of SLE, we asked whether 7α,
   25‐OHC/EBI2 affect the production of IFN and ISGs in innate immune
   cells, which then might promote the cross‐talk with adaptive immune
   cells to regulate the development of SLE.

   In this study, we have identified the increased plasma 7α, 25‐OHC in
   SLE patients, which binds to EBI2 in macrophages to reduce the
   production of IFN‐I and ISGs. EBI2‐deficient macrophages augment the
   production of IFN‐I and chemokines, resulting in massive infiltration
   and activation of myeloid cells and T and B lymphocytes to exacerbate
   the development of SLE. We therefore propose a new role for EBI2 in
   macrophages, whereby it modulates innate and adaptive immune responses
   to protect the host from SLE.

2. Results

2.1. Abnormal Cholesterol and Oxysterol Metabolism in Blood from SLE Patients

   Emerging evidence suggests that lipid and cholesterol metabolism has a
   complex impact on the pathogenesis and progression of SLE.^[ [82]^18 ^]
   To identify the dominant disrupted cholesterol metabolism in SLE
   patients, we used liquid chromatography‐mass spectrometry (LC‐MS) to
   detect cholesterol intermediates in the plasma of SLE patients. Plasma
   concentrations of cholesterol, 7‐DHC and dihydrocholesterol in SLE
   patients were higher than in healthy controls (HCs); furthermore, other
   cholesterol intermediates including lanosterol, lathosterol, and
   desmosterol were not altered (Figure [83] 1A). Conversely, the levels
   of steroid hormones, including glucocorticoids (hydrocortisone and
   cortisone) and androgens (dehydroepiandrosterone, testosterone, and
   androsterone) were reduced in the plasma of SLE patients, although
   estrogen such as 2‐methoxystradiol, estradiol benzoate, and
   pregnenolone did not differ (Figure [84]1B and Figure [85]S1A,
   Supporting Information). Decreased serum/plasma levels of cortisol,
   dehydroepiandrosterone and testosterone have been previously reported
   in SLE patients.^[ [86]^19 ^] Importantly, we discovered that several
   oxysterols including 24‐hydroxycholesterol (24‐HC), 25‐HC, 27‐HC, and
   7α, 25‐OHC were prominently increased in the plasma of SLE patients
   (Figure [87]1C, except for 7α‐HC and 22 (R)‐HC).

Figure 1.

   Figure 1
   [88]Open in a new tab

   Abnormal cholesterol and oxysterol metabolism in blood from SLE
   patients. A–C) Plasma concentrations of cholesterol synthesis (A),
   oxidization (B), and metabolism (C) intermediates in healthy donors
   (n = 10) and SLE patients (n = 20) by LC‐MS. D–F) The relative mRNA
   expression of cholesterol synthesis, oxidization, and
   metabolism‐related enzyme genes in PBMCs of healthy donors (n = 23) and
   SLE patients (n = 24). G,H) Schematic diagrams of cholesterol (G) and
   oxysterols (H) biosynthesis pathways. The up‐regulated intermediates or
   enzyme genes were shown in red, while the down‐regulated intermediates
   or enzyme genes were shown in blue. Data are shown as mean ± SEM. ns,
   not significant (p > 0.05); *p < 0.05, **p < 0.01, ***p < 0.001, and
   ****p < 0.0001, using a two‐tailed, unpaired Student's t‐test (A–F).

   Considering that abnormal changes in cholesterol and its intermediates
   are mainly regulated by a series of metabolic enzymes, we examined the
   mRNA levels of enzymes involved in cholesterol synthesis, uptake,
   efflux, and esterification. In PBMCs of SLE patients, most of these
   metabolic enzymes either declined or remained unchanged
   (Figure [89]1D), which was inconsistent with the elevated levels of
   metabolites. To explain this inconsistency, we propose that when
   abnormal cholesterol metabolism occurs in SLE patients,^[ [90]^20 ^]
   the body may attempt to restore disrupted cholesterol balance toward
   homeostasis, possibly by altering the expression of certain metabolic
   enzymes. Additionally, the mRNA levels of some enzymes responsible for
   catalyzing cholesterol to steroid hormones were reduced
   (Figure [91]1E). In contrast, we observed significant up‐regulation of
   cholesterol 25‐hydroxylase (CH25H) and CYP7B1 (Figure [92]1F), which
   aligns with the enhanced levels of the metabolites named 25‐HC and 7α,
   25‐OHC (Figure [93]1C). Flow diagrams illustrating cholesterol
   biosynthesis and the generation of different derivatives were presented
   in Figure [94]1G,H and Figure [95]S1B, Supporting Information, with
   blue labels indicating reduced enzymes or metabolites and red labels
   representing enhanced enzymes or metabolites. Moreover, we observed
   similar changes in the levels of certain cholesterol metabolites and
   enzymes in peritoneal irrigation fluid (PIF) or peritoneal cells of
   tetramethylpentadecane (TMPD)‐induced SLE mice, particularly an
   increase in CH25H and 25‐HC for the generation of 7α, 25‐OHC (Figure
   [96]S1C,D, Supporting Information). We further confirmed the increased
   concentration of 7α, 25‐OHC in the serum of TMPD‐induced mice (Figure
   [97]S1E, Supporting Information). Together, our findings have
   demonstrated abnormal cholesterol metabolism in the blood of SLE
   patients, highlighting the up‐regulation of 7α, 25‐OHC as particularly
   intriguing.

2.2. The Down‐Regulated Oxysterol Receptor EBI2 in Monocytes of SLE Patients

   Aberrant cholesterol metabolism may lead to changes in the activation
   of cellular signaling pathways, thereby affecting cellular functions.
   We therefore analyzed RNA sequencing (RNA‐seq) data in the Gene
   Expression Omnibus (GEO) database^[ [98]^5b ^] to investigate
   cholesterol‐related differentially expressed genes (DEGs) in PBMCs of
   SLE patients. Undoubtedly, ISGs were significantly up‐regulated in
   classical monocytes (cMos) of SLE patients. Surprisingly, we identified
   the down‐regulated transcription of EBI2, the receptor for 7α, 25‐OHC,
   in cMos of SLE patients (p = 0.0052). Importantly, EBI2 expression was
   further decreased in cMos from IFN‐positive SLE patients (p < 0.0001),
   in contrast to normal expression in IFN‐negative SLE patients
   (p = 0.1663) (Figure [99] 2A). Again, based on multiplexed single‐cell
   RNA‐seq (scRNA‐seq) data from a different study,^[ [100]^21 ^] EBI2 was
   among the top 11 down‐regulated genes in cMos of SLE patients with
   Asian ancestry (Figure [101]S2A, Supporting Information). To verify
   this finding, we collected PBMCs and CD14^+ monocytes from SLE patients
   and healthy donors to determine the transcription of IFNA/B and EBI2.
   In contrast to enhanced transcription of IFNA/B (Figure [102]S2B,
   Supporting Information), mRNA and protein expression of EBI2 was
   significantly reduced in PBMCs of SLE patients (Figure [103]2B,C).
   Notably, the percentages of monocytes were substantially increased in
   PBMCs of SLE patients (Figure [104]S2C, Supporting Information), and
   the decreased transcription and surface expression of EBI2 were
   confirmed in monocytes of SLE patients (Figure [105]2D,E). In both
   PBMCs and CD14^+ monocytes of SLE patients, mRNA levels of the enzymes
   for 7α, 25‐OHC generation including CH25H and CYP7B1 were augmented
   (Figures [106]1F and [107]2F). We then analyzed the correlation between
   EBI2 and IFN using scRNA‐seq data from PBMCs of SLE patients published
   previously.^[ [108]^22 ^] Pediatric and adult SLE patients’ monocytes
   were clustered, followed by a 2D uniform manifold approximation and
   projection (UMAP), and then the EBI2‐high expression (EBI2^hi) subset
   was compared with the EBI2‐low expression (EBI2^lo) subset
   (Figure [109]2G and Figure [110]S2D, Supporting Information). The
   volcano plot of DEGs revealed that the expression of EBI2 was
   negatively related to that of ISGs, such as MX1, MX2, and IFI44L
   (Figure [111]2H and Figure [112]S2E, Supporting Information). Moreover,
   SLEDAI was inversely correlated with EBI2 expression in monocytes
   (Figure [113]2I). In addition, we found that EBI2 expression remained
   unchanged in the monocytes of patients with rheumatoid arthritis (RA),
   dermatomyositis (DM), and Sjögren's syndrome (SS) (Figure [114]2J).
   Together, we propose EBI2 as an IFN‐related gene, and its expression is
   reduced in monocyte from SLE patients.

Figure 2.

   Figure 2
   [115]Open in a new tab

   The down‐regulated oxysterol receptor EBI2 in monocyte of SLE patients.
   A) Heat map of the DEGs in cMos of SLE (IFN‐positive or IFN‐negative)
   patients versus HCs according to the GEO database ([116]GSE149050). B)
   EBI2 mRNA levels in PBMCs of healthy donors (n = 41) and SLE patients
   (n = 36). C) The protein expression of EBI2 in PBMCs of healthy donors
   (n = 4) and SLE patients (n = 5). D,F) The mRNA levels of EBI2 (D) and
   CH25H and CYP7B1 (F) in purified CD14^+ monocytes of healthy donors
   (n = 22) and SLE patients (n = 35). E) Surface EBI2 expression on
   CD11b^+CD14^+ monocytes from PBMCs of healthy donors (n = 24) and SLE
   patients (n = 24). G) scRNA‐seq data of monocytes from adult SLE
   patients or healthy donors were classified into molecularly distinct
   subclusters (SCs). UMAP plots representing SCs (n = 6), groups (healthy
   donors or SLE), and the expression values of EBI2. H) Volcano plot of
   DEGs from SC2 (EBI2^hi) and SC4 (EBI2^lo). I) Correlation between
   EBI2 expression in monocytes and the disease activity was measured by
   the SLEDAI (n = 20). J) Surface EBI2 expression on monocytes of healthy
   donors (n = 9) and patients with RA(n = 8), DM(n = 21), or SS (n = 8).
   Data are shown as mean ± SEM. *p < 0.05, **p < 0.01, and
   ****p < 0.0001, using a two‐tailed, unpaired Student's t‐test (B–E
   [right panels], F), one‐way ANOVA with Holm–Sidak's multiple
   comparisons test (J), or Pearson's test (I).

2.3. Macrophages from SLE Patients and Mice Reduce EBI2 Expression by IFN‐γ

   We next investigated EBI2 expression in TMPD‐induced SLE mice, since
   this model exhibits many clinical features of lupus patients.^[
   [117]^23 ^] The transcription of Ebi2 was decreased in peripheral blood
   cells (Figure [118] 3A), splenocytes (Figure [119]3B), and most
   significantly in peritoneal cells (Figure [120]3C) of TMPD‐induced SLE
   mice, but showed minor or no changes in liver, kidney, lung, and heart
   (Figure [121]S2F–I, Supporting Information). Consistently, monocytes
   sorted from spleens of TMPD‐induced SLE mice exhibited reduced
   Ebi2 transcription (Figure [122]3D). We also verified the reduced
   protein expression of EBI2 in spleen (Figure [123]3E) and peritoneal
   monocytes (Figure [124]3F) of TMPD‐induced mice.

Figure 3.

   Figure 3
   [125]Open in a new tab

   Macrophages from SLE patients and mice reduce EBI2 expression by IFN‐γ.
   A–C) Ebi2 mRNA expression in peripheral blood cells (A), splenocytes
   (B), or peritoneal cells (C) of mice treated with PBS (n = 5, and
   n = 4 only at week 10) or TMPD (n = 5, and n = 4 only at week 24). D)
   Ebi2 mRNA in splenic monocytes of PBS control (n = 4) or TMPD‐treated
   mice (n = 4). E) Immunofluorescence analysis of EBI2 expression in
   spleen of PBS control (n = 3) or TMPD‐treated mice (n = 5) (scale bar:
   25 µm). F) EBI2 protein expression in peritoneal monocytes of mice
   after PBS (n = 3) or TMPD (n = 3) treatment for 12 weeks. G) Ebi2 mRNA
   expression in PEMs after stimulated with various cytokines including
   IL‐2 (100 U mL^−1), IL‐6 (40 ng mL^−1), IL‐1β (10 ng mL^−1), IL‐10
   (10 ng mL^−1), IL‐12 (10 ng mL^−1), M‐CSF (20 ng mL^−1), GM‐CSF
   (20 ng mL^−1), TGF‐β (5 ng mL^−1), IFN‐β (500 ng mL^−1), or CD8^+ T C.M
   for 24 h (n = 3). H) Ebi2 mRNA expression in murine PEMs after
   stimulated with IFN‐γ (200 ng mL^−1) for different time (n = 4). I)
   CD8^+ T C.M was incubated with anti‐IFN‐γ antibody (10 µg mL^−1) at
   4 °C overnight, and then treated PEMs for 12 h to detect Ebi2 mRNA
   expression (n = 3). J) Immunofluorescence analysis of EBI2 expression
   in PEMs after stimulation with IFN‐γ (200 ng mL^−1) for 12 h (n = 3;
   scale bar: 20 µm). K) Human CD14^+ monocytes were stimulated with IFN‐γ
   (200 ng mL^−1) or CD3 C.M for 12 h to detect EBI2 transcription
   (n = 3). Data are shown as mean ± SEM or representative from at least
   three independent experiments (E,J [left panels], F [right panel]).
   *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, using one‐way
   ANOVA with Holm–Sidak's multiple comparisons test (G–I,K), two‐way
   ANOVA with Holm–Sidak's multiple comparisons test (A–C), or a
   two‐tailed, unpaired Student's t‐test (D–F,J).

   To explore how EBI2 was down‐regulated in monocytes/macrophages of SLE
   patients and mice, we first treated murine peritoneal exudate
   macrophages (PEMs) with 7α, 25‐OHC, which did not affect the mRNA and
   protein expression of EBI2 (Figure [126]S2J,K, Supporting Information).
   As cytokines secreted by activated immune cells such as macrophages and
   T cells could play important roles in the development of SLE, we
   treated PEMs with various cytokines and supernatants from
   anti‐CD3/anti‐CD28‐activated CD8^+ T cells (CD8^+ T C.M) to screen
   which factors could reduce EBI2 expression. Intriguingly, we observed
   that activated CD8^+ T cell supernatants significantly decreased
   Ebi2 transcription in macrophages (Figure [127]3G). Furthermore,
   treatment with IFN‐γ, a crucial cytokine in CD8^+ T C.M, substantially
   inhibited Ebi2 transcription (Figure [128]3H), and neutralizing CD8^+ T
   C.M with anti‐IFN‐γ antibody could partially restore Ebi2 reduction
   (Figure [129]3I). A remarkable reduction in protein expression of
   EBI2 was also observed in IFN‐γ‐stimulated PEMs (Figure [130]3J). In
   agreement with this, EBI2 transcription was down‐regulated in human
   CD14^+ monocytes after treatment with supernatants from
   anti‐CD3/anti‐CD28‐activated human CD3^+ T cells (CD3^+ T C.M) or
   stimulation with IFN‐γ (Figure [131]3K). Taken together, these results
   suggest that cytokines produced by activated T cells including IFN‐γ
   could reduce EBI2 expression in monocyte/macrophages.

2.4. 7α, 25‐OHC Inhibits EBI2 and GNAI3‐Mediated STAT Activation to Reduce
IFN‐I Response Independent of TBK1

   Next, we asked whether EBI2 could regulate IFN‐I or ISG production
   during the progression of SLE. To this end, we generated
   Ebi2 conditional knockout (cKO) mice bearing two loxP sites flanking
   the second exon of the Ebi2 gene that were crossbred with the LysM‐Cre
   mice to specifically delete EBI2 in myeloid cells including macrophages
   (termed Ebi2‐cKO mice) (Figure [132]S3A,B, Supporting Information). The
   genetic ablation efficiency of Ebi2 in primary PEMs was confirmed by
   genome sequencing (Figure [133]S3C, Supporting Information),
   quantitative PCR, and immunoblotting (Figure [134]S3D, Supporting
   Information). Ebi2‐cKO mice have normal development of bone
   marrow‐derived macrophages (BMDMs), showing comparable percentages of
   macrophages (Figure [135]S3E,F, Supporting Information) and normal
   expression of Adgre1 and Mertk (Figure [136]S3G, Supporting
   Information). It was reported that mitochondrial DNA (mtDNA) in the
   cytosol or extracellular milieu can drive IFN‐I production to promote
   human lupus.^[ [137]^24 ^] Therefore, we treated WT and Ebi2‐cKO PEMs
   with the BCL‐X[L] and BCL‐2 inhibitor ABT‐737 plus the pan‐caspase
   inhibitor QVD‐OPh^[ [138]^25 ^] to release mtDNA into the cytoplasm,
   which increased the transcription of Ifnb, and ISGs including the
   chemokines Cxcl10 and Ccl2, the GTP‐binding protein Mx2, the
   IFN‐induced protein with tetratricopeptide repeats 2 (Ifit2), Irf7,
   Ifna4, and Ifna in Ebi2‐cKO PEMs (Figure [139] 4A and Figure [140]S4A,
   Supporting Information). The enhanced secretion of IFN‐β and CXCL10 in
   the culture supernatants of EBI2‐deficient macrophages was also
   confirmed (Figure [141]4B). Similar results were observed in ISD, a
   synthetic mimic of dsDNA, treated Ebi2‐cKO macrophages (Figure [142]4C
   and Figure [143]S4B, Supporting Information). Upon poly (dA:dT)
   stimulation, a synthetic analog of B‐DNA, EBI2‐deficient macrophages
   also up‐regulated the production of Ifnb and Cxcl10 (Figure [144]S4C,
   Supporting Information). Because virus infection such as Epstein–Barr
   virus has been reported in SLE patients,^[ [145]^26 ^] we next infected
   EBI2‐deficient macrophages with the DNA virus herpes simplex virus
   1(HSV‐1), which again resulted in a dramatic up‐regulation of Ifnb and
   Cxcl10 (Figure [146]S4D, Supporting Information).

Figure 4.

   Figure 4
   [147]Open in a new tab

   7α, 25‐OHC inhibits EBI2 and GNAI3‐mediated STAT activation to reduce
   IFN‐I response independent of TBK1. A,B) WT and Ebi2‐cKO PEMs were
   treated with QVD‐OPh (50 µm) plus ABT‐737 (5 µm) for 24 h to detect the
   mRNA or protein expression of IFN‐β, CXCL10 or Ifit2, Mx1, Mx2,
   Ccl2, and Irf7 (n = 3). C) WT and Ebi2‐cKO PEMs were treated with ISD
   (5 µg mL^−1) to detect the protein expression of IFN‐β and CXCL10
   (n = 3). D,E) PEMs were pre‐treated with 7α, 25‐OHC (10 µM) for 2 h
   followed by HSV‐1 (multiplicity of infection (MOI), 1) infection for
   6 h to detect IFN‐β production (D, n = 3), or infection for 12 h to
   detect the mRNAs of Cxcl10, Mx1, Mx2, Irf7, and Ccl2 (E, n = 3). F) WT
   and Ebi2‐cKO PEMs were treated with ISD for different time to detect
   phosphorylation levels of STAT1, STAT2, STAT3, ERK, and p38. G) PEMs
   were pre‐treated with DMSO or 7α, 25‐OHC for 2 h followed by ISD
   stimulation for different time to detect the phosphorylation levels of
   STAT1, STAT2, STAT3, ERK, and p38. H) The phosphorylation levels of
   STAT1 and STAT3 in PBMCs of SLE patients and HCs. I,J) PEMs were
   transfected with siRNAs (40 nm) targeting Jak1, Jak2, Stat1, and
   Stat2 for 48 h to detect the knockdown efficiency (I), or followed by
   ISD stimulation to measure the transcriptions of Ifnb (J, n = 4). K) WT
   and Gnai3 ^−/− BMDMs were treated with DMSO or 7α, 25‐OHC for 2 h
   followed by ISD stimulation to measure the transcription of Ifnb and
   Cxcl10 (n = 3). Data are shown as mean ± SEM or representative from
   three independent experiments (F–I). *p < 0.05, **p < 0.01,
   ***p < 0.001, and ****p < 0.0001, using two‐way ANOVA with Holm–Sidak's
   multiple comparisons test (D [left panel], J,K), or a two‐tailed,
   unpaired Student's t‐test (A–C,E).

   We also investigated the effect of EBI2 on IFN‐I production in
   macrophages upon activation of TLR7/9, RIG‐I, or in response to RNA
   virus infection. To this end, we treated WT or Ebi2‐cKO PEMs with the
   TLR7 and TLR8 agonist resiquimod (R848), the TLR7 agonist imiquimod
   (R837), the TLR9 agonist CpG, the TLR3 and RIG‐I agonist poly (I:C), or
   the RNA virus vesicular stomatitis virus (VSV). Ebi2‐cKO PEMs showed
   increased transcription of Ifnb after CpG stimulation, lipofectamine
   2000‐mediated transfection of poly (I:C), or VSV infection (Figure
   [148]S4E, Supporting Information), and the enhanced secretion of IFN‐β
   was confirmed in poly (I:C)‐transfected Ebi2‐cKO PEMs (Figure [149]S4F,
   Supporting Information). To further ascertain this phenotype, we
   constructed immortalized BMDMs (iBMDMs) that stably overexpressed human
   EBI2 (hEBI2) (Figure [150]S4G, Supporting Information).
   EBI2‐overexpressed macrophages showed a remarkable down‐regulation of
   the mRNA levels of Ifnb and ISGs including Cxcl10, Mx2, Irf7, Ifit2,
   Mx1, and Ccl2 (Figure [151]S4H, Supporting Information) as well as
   IFN‐β secretion (Figure [152]S4I, Supporting Information) in response
   to ISD stimulation.

   Given that the ligand 7α, 25‐OHC was found to be enhanced in SLE
   patients, we investigated how it affected IFN‐I and ISG production.
   Addition of 7α, 25‐OHC substantially decreased the production of IFN‐β
   and ISGs, including Cxcl10, Mx1, Mx2, Ccl2, and Irf7 after
   HSV‐1 infection (Figure [153]4D,E). Upon ISD and cGAMP stimulation, 7α,
   25‐OHC treatment also reduced Ifnb transcription (Figure [154]S4J,K,
   Supporting Information). These observations have elucidated that when
   macrophages sense stimuli such as cytosolic nucleic acids or viral
   infection, the binding of 7α, 25‐OHC to EBI2 could suppress the
   production of IFN‐I and ISGs.

   To investigate EBI2‐related downstream signaling, we assessed the
   phosphorylation of TANK (TRAF family member associated NF‐κB
   activator)‐binding kinase 1 (TBK1) and IRF3 in ISD‐stimulated Ebi2‐cKO
   macrophages. Surprisingly, we found that WT and Ebi2‐cKO PEMs showed
   comparable levels of phosphorylation in TBK1 and IRF3 (Figure [155]S5A,
   Supporting Information). Since EBI2 inhibited the expression of ISGs,
   we next investigated its effect on the IFNAR signaling pathway. Indeed,
   ISD‐stimulated Ebi2‐cKO PEMs showed enhanced phosphorylation levels of
   STAT1, STAT2, STAT3, ERK, and p38 (Figure [156]4F and Figure [157]S5B,
   Supporting Information). In agreement with this, 7α, 25‐OHC‐treated WT
   PEMs showed decreased phosphorylation of STAT1, STAT2, STAT3, ERK, and
   p38, but not TBK1 and IRF3 (Figure [158]4G and Figure [159]S5C,D,
   Supporting Information).

   Since we observed the elevated phosphorylation of STAT3 and STAT1 in
   PBMCs of SLE patients (Figure [160]4H and Figure [161]S5E, Supporting
   Information), we asked whether blocking the JAK‐STAT signaling could
   reverse the enhanced IFN‐I response in EBI2‐deficient macrophages.
   Treatment with the pan‐JAK inhibitor tofacitinib or the JAK1 and
   TYK2 dual inhibitor PF‐06700841, as well as the STAT inhibitors
   SH‐4‐54 and static, could all abrogate the elevated transcription of
   Ifnb and Cxcl10 in Ebi2‐cKO PEMs (Figure [162]S6A, Supporting
   Information). We further confirmed that silencing Jak1, Jak2, Stat1,
   and Stat2 could effectively restore the heightened expression of Ifnb
   and Cxcl10 in ISD‐stimulated Ebi2‐cKO PEMs (Figure [163]4I,J and
   Figures [164]S5F and S6B, Supporting Information). To verify the
   inhibitory role of EBI2 on IFN‐I production, we overexpressed hEBI2 in
   primary Ebi2‐cKO BMDMs using retrovirus infection. The infection
   efficiency of BMDMs was determined by detecting the percentages of
   GFP^+ cells (Figure [165]S6C, Supporting Information) or measuring the
   mRNA levels of EBI2 by quantitative PCR (Figure [166]S6D, Supporting
   Information). Ebi2‐cKO BMDMs enhanced the levels of Ifnb, Cxcl10, and
   IFN‐β secretion; but once EBI2 was overexpressed in Ebi2‐cKO BMDMs, the
   levels of Ifnb, Cxcl10, and IFN‐β were rescued and reduced to the
   levels comparable to those in WT BMDMs (Figure [167]S6E,F, Supporting
   Information). In addition, the enhanced phosphorylation of STAT1 in
   Ebi2‐cKO BMDMs was abolished once EBI2 was overexpressed. In contrast,
   the phosphorylation of TBK1 and IRF3 was not affected by
   EBI2 deficiency or overexpression (Figure [168]S6G,H, Supporting
   Information), which was consistent with our previous findings. These
   results suggest that 7α, 25‐OHC‐EBI2 signaling inhibits IFN‐I response
   through suppressing STAT activation.

   Previous study has reported that EBI2 signals through the G protein
   alpha inhibitory family (Gα [i]),^[ [169]^27 ^] which includes three
   subunits: Gα [i1], Gα [i2], and Gα [i3].^[ [170]^28 ^] Using small
   interfering RNAs (siRNAs) to silence Gnai1, Gnai2, Gnai3, and Gnas,
   respectively (Figure [171]S7A, Supporting Information), we found that
   knockdown of Gnai1, Gnai2, or Gnas had no effect on the reduced
   expression of Ifnb by 7α, 25‐OHC treatment. However, knockdown of
   Gnai3 abolished the 7α, 25‐OHC‐induced reduction of Ifnb transcription
   (Figure [172]S7B, Supporting Information). Consistent with the effect
   by 7α, 25‐OHC, Gnai3 did not affect phosphorylation of TBK1 and IRF3
   (Figure [173]S7D,E, Supporting Information); Instead, the attenuated
   phosphorylation of STAT1 and STAT3 (Figure [174]S7F,G, Supporting
   Information), as well as the decreased expression of Ifnb and Cxcl10
   (Figure [175]4K), by 7α, 25‐OHC treatment were partially restored by
   Gnai3 KO (Figure [176]S7C, Supporting Information). This incomplete
   recovery of Gnai3 ^−/− macrophages might be due to other receptors,
   such as retinoic acid‐related orphan receptor alpha, which can also
   recognize 7α, 25‐OHC.^[ [177]^29 ^] Importantly, we found that
   GNAI3 interacted with the tyrosine kinases JAK1, JAK2, and TYK2 and
   weakly bound to JAK3 (Figure [178]S7H, Supporting Information).
   Together, we have elucidated that the 7α, 25‐OHC‐EBI2 axis inhibits
   STAT activation and IFN‐I response via the downstream GNAI3.

2.5. EBI2 Inhibits STAT Activation Depending on IFN‐I and IFN‐II Receptor

   Because IFN receptor‐mediated signaling could activate JAK/STAT, we
   next asked whether EBI2 was dependent on IFN receptors to inhibit STAT
   activation. PEMs from AG6 (IFN‐I and IFN‐II receptor deficient mice)
   and WT mice were transfected with Ebi2 siRNA or treated with 7α,
   25‐OHC, followed by ISD stimulation. We found that Ebi2 knockdown or
   7α, 25‐OHC treatment could enhance or reduce the phosphorylation levels
   of STAT1 and STAT3, respectively, in WT PEMs, whereas these effects
   were not observed in AG6 PEMs (Figure [179] 5A–D and Figure S[180]8A,B,
   Supporting Information). Furthermore, in ISD‐stimulated
   AG6 macrophages, the 7α, 25‐OHC‐induced reduction of Ifnb and
   Cxcl10 transcription was abrogated (Figure [181]5E). Consistently, in
   response to IFN‐β or IFN‐γ stimulation, the phosphorylation levels of
   STAT1 and STAT3 (Figure [182]5F and Figure [183]S8C, Supporting
   Information) or STAT1 alone (Figure [184]5H and Figure [185]S8D,
   Supporting Information), as well as the production of CXCL10
   (Figure [186]5G,I) were enhanced in Ebi2‐cKO PEMs. These data suggest
   that EBI2 inhibits the effects of IFN‐I (i.e., STAT activation and ISG
   production), which is dependent on the IFN‐I/IFN‐II Receptor.

Figure 5.

   Figure 5
   [187]Open in a new tab

   EBI2 inhibits STAT activation depending on IFN‐I and IFN‐II receptor.
   A,B) WT and AG6 PEMs were transfected with Ebi2 siRNA for 48 h followed
   by ISD stimulation for 3 or 6 h to detect the levels of p‐STAT1, STAT1,
   p‐STAT3, STAT3, p‐TBK1, TBK1, p‐IRF3, and IRF3 (A) and quantify the
   p‐STAT1 and p‐STAT3 (B, n = 3). C–E) WT and AG6 PEMs were treated with
   DMSO or 7α, 25‐OHC for 2 h followed by ISD stimulation for 3 or 6 h to
   detect the levels of p‐STAT1, STAT1, p‐STAT3, STAT3, p‐TBK1, TBK1,
   p‐IRF3, and IRF3 (C) and quantify the p‐STAT1, p‐STAT3 (D, n = 3), or
   to detect the transcriptions of Ifnb and Cxcl10 (E, n = 3). F–I) WT and
   Ebi2‐cKO PEMs were stimulated with IFN‐β (500 ng mL^−1) or IFN‐γ
   (200 ng mL^−1) for 15 or 30 min to detect the levels of p‐STAT1, STAT1,
   p‐STAT3, and STAT3 (F,H), or the expression of CXCL10 (G,I, n = 3).
   Data are shown as mean ± SEM or representative from three independent
   experiments (A,C,F,H). *p < 0.05, **p < 0.01, ***p < 0.001, and
   ****p < 0.0001, using two‐way ANOVA with Holm–Sidak's multiple
   comparisons test (B,D,E), or a two‐tailed, unpaired Student's t‐test
   (G,I).

2.6. Ebi2‐cKO Lupus Mice Unexpectedly Increase Infiltration of Myeloid and T
Cells

   To investigate the in vivo role of EBI2 in lupus, we constructed
   TMPD‐induced lupus model in WT and Ebi2‐cKO mice. After 2 weeks of
   induction, peritoneal monocytes were sorted for RNA‐seq (Figure
   [188]S9A,B, Supporting Information), and the global gene expression
   profile identified 267 DEGs (P < 0.05) including 110 up‐regulated and
   157 down‐regulated genes in Ebi2‐cKO monocytes. The substantially
   reduced Gpr183 expression confirmed the KO efficiency (Figure [189]
   6A). Ebi2‐cKO monocytes strikingly increased expression of chemokines
   and cytokines such as Cxcl10, Cxcl11, Ccl2, Ccl12, and Il6, whereas
   Tgfb3 expression was down‐regulated (Figure [190]6A). Intriguingly,
   serum concentrations of these up‐regulated genes in Ebi2‐cKO monocytes
   were previously reported to be increased in SLE patients,^[ [191]^30 ^]
   and TGF‐β3 was reported to negatively regulate SLE.^[ [192]^31 ^] Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis
   also revealed that cytokine and cytokine receptor interaction were the
   top two enriched signaling pathways (Figure [193]6B). Next, we
   substantiated the increased production of Cxcl10, Cxcl11, Ccl2, and
   Ccl12 in TMPD‐treated Ebi2‐cKO monocytes (Figure [194]6C), and
   CXCL10 and CCL2 in PIF of Ebi2‐cKO mice (Figure [195]6D). Also, several
   genes related to SLE, including Ifit2, Ifi208, and Tlr9, were increased
   in Ebi2‐cKO monocytes (Figure [196]S9C, Supporting Information).

Figure 6.

   Figure 6
   [197]Open in a new tab

   Ebi2‐cKO lupus mice unexpectedly increase infiltration of myeloid and T
   cells. A–C) RNA‐sequencing analysis of the DEGs (A), or KEGG pathway
   enrichment analysis of the DEGs (B), or mRNA expression of Cxcl10,
   Cxcl11, Ccl2, and Ccl12 (C, n = 5) in the sorted monocytes from
   peritoneal cells of WT and Ebi2‐cKO mice with TMPD treatment for
   2 weeks. D,E) WT and Ebi2‐cKO mice were intraperitoneally injected with
   TMPD for 2 or 12 weeks to detect the secretion of CXCL10 and CCL2 in
   the PIF (D, n = 4), or the cell numbers in the peritoneal cavity (E,
   n = 3). F–J) WT and Ebi2‐cKO mice were intraperitoneally injected with
   TMPD for 12 weeks to detect the percentages and numbers of
   Ly6C^intLy6G^+ (neutrophils), Ly6C^hiLy6G^− (monocytes), Ly6C^hiLy6G^+
   cells (F,G, n = 6), T and B cells in peritoneal cavity (H, n = 6), or
   the percentages and numbers of T cells in spleens (I,J, n = 6). K,L)
   Transwell assay to check migration ability of murine PEMs (K) or
   anti‐CD3/CD28‐stimulated CD8^+ T cells (L, n = 3) toward PIF from
   TMPD‐treated WT and Ebi2‐cKO mice or toward CXCL10 (100 ng mL^−1). Data
   are shown as mean ± SEM or representative from three independent
   experiments (E,F,I,K). *p < 0.05 and **p < 0.01, using a two‐tailed,
   unpaired Student's t‐test (C–E [right panel], G,H,J), or one‐way ANOVA
   with Holm–Sidak's multiple comparisons test (L).

   EBI2 has been reported to enhance immune cell migration to accelerate
   the development of multiple sclerosis.^[ [198]^16 ^] However, we
   surprisingly observed that the numbers of immune cells were robustly
   increased in the peritoneal cavity of Ebi2‐cKO mice upon TMPD
   administration (Figure [199]6E). Not only the elevated percentages of
   monocytes (Ly6C^hiLy6G^−), but also the increased numbers of neutrophil
   (Ly6C^hiLy6G^+), monocytes, CD8^+ T cells, and B cells were observed in
   the peritoneal cavity of Ebi2‐cKO mice (Figure [200]6F–H). Moreover,
   the percentages and numbers of CD4^+ and CD8^+ T cells were decreased
   in spleen of TMPD‐treated Ebi2‐cKO mice (Figure [201]6I,J). To further
   investigate whether the enhanced immune cell infiltration was
   associated with the elevated chemokines in PIF of TMPD‐treated Ebi2‐cKO
   mice, we performed a Transwell migration assay and found that PIF from
   Ebi2‐cKO mice attracted more PEMs or CD8^+ T cell toward the bottom
   than PIF from WT mice (Figure [202]6K,L and Figure [203]S9D, Supporting
   Information).

   It is possible that the augmented expression of chemokine receptors or
   enhanced cell proliferation may also contribute to cell accumulation in
   the peritoneal cavity of Ebi2‐cKO mice. However, we found that
   transcription levels of Cxcr3 (receptor for CXCL10 and CXCL11), Cxcr7
   (receptor for CXCL11), and Ccr2 (receptor for CCL2) were comparable in
   WT and Ebi2‐cKO peritoneal cells at 2 or 12 weeks after TMPD induction
   (Figure [204]S10A, Supporting Information). Moreover, mRNA levels of
   Ccr2 in sorted peritoneal monocytes or Cxcr3 and Cxcr7 in sorted
   peritoneal T cells, and surface CXCR3 expression in splenic or
   peritoneal T cells remained unchanged in TMPD‐treated Ebi2‐cKO mice
   (Figure [205]S10B–E, Supporting Information). Next, we detected the in
   vivo proliferation of myeloid cells, T cells, and B cells with EdU
   staining, which was all unaffected in Ebi2‐cKO mice (Figure
   [206]S10F–J, Supporting Information). Collectively, we suggest that
   Ebi2‐cKO macrophages/myeloid cells enhance chemokine production to
   shape the microenvironment, which remarkably promotes migration of
   myeloid and T cells to overcome the effect of EBI2 deficiency on cell
   migration.

2.7. Ebi2‐cKO Lupus Mice Show the Enhanced Lymphocyte Activation

   Abnormal activation of T and B cells is a hallmark of SLE, and
   macrophages can secret cytokines and present autoantigens to regulate
   lymphocyte activation.^[ [207]^32 ^] We next investigated whether
   EBI2‐deficient macrophages affect the adaptive immune response. Flow
   cytometric analysis showed increased percentages of effector memory
   (CD44^hi CD62L^lo) CD4^+ and CD8^+ T cells as well as decreased
   percentages of naïve (CD44^lo CD62L^hi) T cells in lymph nodes (LNs)
   and spleens of Ebi2‐cKO mice (Figure [208] 7A,B and Figure [209]S11A,B,
   Supporting Information). To explore how EBI2‐deficient macrophages
   promote T cell activation, we analyzed the RNA‐seq data and found
   several up‐regulated DEGs such as Il15, Il6, and Cd226 and
   down‐regulated DEG pdcd1lg2 in Ebi2‐cKO monocytes (Figure [210]S11C,
   Supporting Information). We confirmed these changes in TMPD‐treated
   Ebi2‐cKO monocytes or ISD‐stimulated Ebi2‐cKO PEMs (Figure [211]7C and
   Figure [212]S11D, Supporting Information). In addition to the decreased
   expression of PD‐L2 may weaken the inhibition of T cell activation,^[
   [213]^33 ^] the increased IL‐15 could also promote the proliferation of
   CD4^+ effector memory T cells and support both central and effector
   memory CD8^+ T cells.^[ [214]^34 ^] Besides, IL‐6 concentrations in
   serum and PIF (Figure [215]7D), or Il1b transcription levels in
   peritoneal cells (Figure [216]S11E, Supporting Information) were
   increased in Ebi2‐cKO mice. IL‐6 is involved in T[FH]
   differentiation,^[ [217]^35 ^] and IL‐6 or IL‐1β is an important driver
   for T[H]17 differentiation.^[ [218]^36 ^] Indeed, the frequencies of
   splenic T[H]17 (CD4^+IL‐17A^+) cells (Figure [219]7E) and T[FH]
   (CD4^+CXCR5^+PD‐1^+) cells (Figure [220]7F) were increased in Ebi2‐cKO
   mice. Furthermore, the production of IFN‐γ and TNF‐α was markedly
   increased in splenic CD8^+ T cells of Ebi2‐cKO mice (Figure [221]7G,H).
   These results suggest that EBI2‐deficient macrophages promote the
   activation of T[H]17, T[FH], and CD8^+ T cells via modulating the
   production of various cytokines during the development of TMPD‐induced
   SLE.

Figure 7.

   Figure 7
   [222]Open in a new tab

   Ebi2‐cKO lupus mice show the enhanced lymphocyte activation. A,B,D–J)
   WT and Ebi2‐cKO mice were treated with TMPD for 12 weeks.
   Representative dot plots of naïve (CD44^loCD62L^hi), central memory
   (CD44^hiCD62L^hi), and effector memory (CD44^hiCD62L^lo) T cell subsets
   in LN and spleen (A), statistical analysis of the percentages of
   effector memory T cells from LN or spleen (B, n = 6),
   IL‐6 concentrations in serum and PIF (D, n = 4), flow cytometric
   analysis of CD4^+IL‐17A^+ T[H]17 cells (E, n = 6) or CD4^+PD‐1^+
   CXCR5^+ T[FH] cells (F, n = 6) in spleen, the production of IFN‐γ and
   TNF‐α in splenic CD8^+ T cells (G,H, n = 6), MHCII expression on CD19^+
   B cells (I, n = 6) and the percentage of plasma cells (J, n = 6) in
   spleen. C) WT (n = 5) and Ebi2‐cKO (n = 5) mice were treated with TMPD
   for 2 weeks to detect the mRNA expression of Il15, Cd226, Pdcd1lg2, and
   Pdcd1lg1 in peritoneal monocytes. Data are shown as mean ± SEM or
   representative from three independent experiments (A,E–J [left
   panels]). *p < 0.05, **p < 0.01, and ***p < 0.001, using a two‐tailed,
   unpaired Student's t‐test (B–D, and E–J [right panels]).

   TGF‐β inhibits B cell proliferation, plasma cell differentiation, and
   antibody production, while IFN‐I, IL‐6, and T[FH] cells promote these
   processes.^[ [223]^31 , [224]^37 ^] In agreement with the reduced
   levels of TGF‐β and increased levels of IL‐6, IFN‐I, and percentages of
   T[FH] cells in EBI2‐deficient cells or mice (Figure [225]7D,F and
   Figure [226]S11F, Supporting Information), we found that Ebi2‐cKO lupus
   mice enhanced surface MHCII levels on B cells (Figure [227]7I) and
   increased the percentages of plasma cells (Figure [228]7J). These
   findings suggest that EBI2‐deficient macrophages promote the production
   of cytokines and chemokines, which might activate B cells and plasma
   cells to accelerate SLE development.

2.8. EBI2 Deficiency in Macrophages Aggravates the Lupus Nephritis

   Next, we investigated how EBI2‐deficient macrophages affect the
   development of SLE in vivo using TMPD‐induced lupus model, which
   features the development of lipogranulomas adhering to peritoneal
   mesothelial cell surface and diffuse pulmonary hemorrhage (DPH).^[
   [229]^23 , [230]^38 ^] Compared with WT mice, more Ebi2‐cKO mice
   developed complete DPH 2 weeks after TMPD injection and displayed more
   severe hemorrhagic spots in lung and more lipogranulomas at the end of
   12 weeks (Figure [231] 8A). Hematoxylin and eosin (H&E) staining
   demonstrated worsened histologic damage in Ebi2‐cKO mice, as evidenced
   by enlarged glomerulus and pulmonary hemorrhage (Figure [232]8B,C),
   more diffused immune cells in the red pulp of spleen, and more severe
   vacuolization in spleen and liver (Figure [233]8D,E). In addition,
   immunostaining analysis revealed enhanced deposition of immunoglobulin
   G (IgG) and complement 3 (C3) in the glomeruli of Ebi2‐cKO mice
   (Figure [234]8F). Consistently, serum levels of anti‐dsDNA antibody and
   anti‐nuclear antibody (ANA) were significantly increased in Ebi2‐cKO
   mice (Figure [235]8G). Quantitative analysis of urine samples exhibited
   elevated levels of proteinuria, urea nitrogen, and creatinine in
   Ebi2‐cKO mice (Figure [236]8H). These results indicate that
   EBI2 deficiency in macrophages promotes the development of lupus in
   mice.

Figure 8.

   Figure 8
   [237]Open in a new tab

   EBI2 deficiency in macrophages aggravates the lupus nephritis. A–H) WT
   and Ebi2‐cKO mice were treated with TMPD for 2 or 12 weeks.
   Representative images of lungs and lipogranulomas in peritoneal cavity
   (A) or H&E staining of kidney (B), lung (C), spleen (D), and liver (E,
   scale bar: 100 µm), or immunofluorescence of glomerular deposition of
   C3 and IgG (F, scale bar: 30 µm), or serum anti‐dsDNA antibody and
   anti‐nuclear antibody (G, n = 10), or levels of proteinuria, urea
   nitrogen, and creatinine in urine samples (H, n = 10). Scale bars:
   200 µm (B–D). Data are shown as mean ± SEM or representative from three
   independent experiments (A–F). *p < 0.01, **p < 0.01, and ***p < 0.001,
   using a two‐tailed, unpaired Student's t‐test (G,H).

   In summary, we propose that the enhanced levels of the ligand 7α,
   25‐OHC could bind and activate EBI2 to reduce the expression of IFN‐I
   and ISGs, thus protecting SLE patients. However, during the development
   of SLE, IFN‐γ secreted by activated T cell reduces EBI2 expression in
   macrophages, resulting in increased production of chemokines and
   cytokines to not only attract myeloid and T cells, but also enhance
   lymphocyte activation to worsen SLE pathogenesis. Together, we propose
   that the GPCR EBI2 acts as a key effector that links the crosstalk
   between innate and adaptive immunity, and it holds promise as a
   potential diagnostic marker or therapeutic target for SLE.

3. Discussion

   EBI2 is a well‐known chemotactic receptor that mediates the migration
   of various immune cells, including B cells, T cells, dendritic cells,
   astrocytes, and group 3 innate lymphoid cells, to sites containing the
   ligand 7α, 25‐OHC.^[ [238]^16 , [239]^39 ^] Therefore, we expected that
   Ebi2^fl ^/ ^fl ‐LysM‐Cre mice might exhibit reduced infiltration of
   myeloid cells during the development of SLE. Surprisingly, after TMPD
   induction, we found increased infiltration of myeloid cells in the
   peritoneal cavity of Ebi2‐cKO mice, along with enhanced production of
   IFNs and ISGs such as CXCL10 and CCL2, which suggests that this
   extrinsic microenvironment could overcome the intrinsic
   EBI2 deficiency‐mediated reduction of myeloid cell migration. In
   addition, we observed reduced mRNA and protein expression of EBI2 in
   PBMCs of SLE patients, as well as in splenocytes, peritoneal cells, and
   peripheral blood cells of SLE mice. These observations indicate that
   other immune cells besides monocyte/macrophages may also down‐regulate
   EBI2 expression during the progression of SLE. Consequently, further
   research is needed to investigate whether the 7α, 25‐OHC/EBI2 signaling
   in other immune cells with EBI2 down‐regulation is involved in the
   pathogenesis of SLE.

   An integrated genome‐wide association study suggests EBI2 as a negative
   regulator of the IRF7‐driven inflammatory network, which was associated
   with susceptibility to type 1 diabetes.^[ [240]^40 ^] In line with
   this, we have elucidated that EBI2 suppresses JAK‐STAT activation via
   GNAI3, resulting in the reduced ISGs production to limit the
   progression of SLE. Furthermore, we found that the significantly
   reduced expression of EBI2 in monocyte/macrophages of SLE patients and
   mice could be triggered by IFN‐γ in a paracrine manner. Since
   autoantigens‐activated CD4^+ T cells or CD8^+ T cells produce IFN‐γ
   during the development of SLE, this leads to the reduction of
   EBI2 expression in monocyte/macrophages, which alleviates its
   inhibitory effect on STAT activation to produce higher levels of
   chemokines and cytokines that not only further facilitates lymphocyte
   activation and IFN‐γ secretion but also promotes differentiation of
   lymphocytes into T[H]17, T[FH], and plasma cells, reinforcing the
   development of SLE. Together, we propose a feedback loop linking
   monocyte/macrophages and T cells via EBI2 during the progression of
   SLE. Additionally, we provided a new angle to understand EBI2 function,
   that is, reducing EBI2 expression in macrophages could enhance IFNs,
   cytokines and chemokines in the microenvironment, which reshape the
   innate and adaptive immune responses to promote SLE development.

   Although SLE patients exhibit profoundly reduced surface expression of
   EBI2 on monocytes, patients with RA, DM, or SS display normal
   EBI2 levels. A previous study has shown that 7α, 25‐OHC stimulation
   could induce β‐arrestin‐mediated internalization of EBI2.^[ [241]^14b
   ^] We observed that the transcription of enzymes for 7α, 25‐OHC
   biosynthesis, including CH25H and CYP7B1, was enhanced in PBMCs and
   monocytes of SLE patients, which is consistent with the increased
   plasma concentrations of 7α, 25‐OHC in SLE patients. Therefore, the
   down‐regulated surface expression of EBI2 in SLE patients might be
   partially due to its internalization and desensitization after sensing
   the elevated ligand 7α, 25‐OHC. Together, EBI2 serves as a key receptor
   for macrophages to sense cholesterol metabolites in the
   microenvironment, which could modulate cytokine production to reshape
   the microenvironment or the function of innate and adaptive immune
   cells. These characters make EBI2 a precisely controlled GPCR at its
   expression and signaling levels by IFN‐related regulators (i.e., IFN‐γ,
   oxysterol, STAT, IFN‐I, and ISGs) during the development of SLE.

   We have found that both the expression levels of plasma 7α, 25‐OHC and
   EBI2 on the surface of monocyte/macrophage may serve as diagnostic
   markers to predict the outcome of SLE. Moreover, as GPCRs are favorable
   drug targets, exploring potential agonists to maintain EBI2 surface
   expression might provide new therapeutic strategies for SLE. Given that
   the expression and signaling of EBI2 are controlled by multiple
   IFN‐related regulators (i.e., IFN‐γ, oxysterol, STAT, IFN‐I, and ISGs),
   it would be exciting to investigate whether detecting the surface
   expression of EBI2 could predict the therapeutic efficacy of SLE
   patients who receive Anifrolumab treatment.

4. Experimental Section

Mice

   The second exon of Ebi2 gene was flanked by two loxP sites to generate
   Ebi2 genetically ablated mice (Beijing Biocytogen Co., Ltd.). Ebi2^fl
   ^/− mice were crossbred with Lyz2‐Cre ^+/+ mice to generate Ebi2^fl ^/−
   Lyz2‐Cre ^+/− mice, which were then backcrossed with Ebi2^fl ^/− mice.
   Ebi2^fl ^/ ^flLyz2‐Cre ^+/− or Ebi2^fl ^/ ^flLyz2‐Cre ^+/+ (termed
   Ebi2‐cKO) mice specifically knockout Ebi2 in myeloid cells, and the
   littermate Ebi2^fl ^/ ^flLyz2‐Cre ^−/− (termed WT) were used as the
   controls. Genotyping was performed using the primers: forward,
   5′‐ACTTTCTGGGTTGACACAGATCAGATT‐3′; reverse,
   5′‐TACAGTTGGTGGCTTTGGCAGAACT‐3′. Gnai3 ^−/− mice were kindly provided
   by Dr. Beicheng Sun (The Affiliated Drum Tower Hospital of Nanjing
   University Medical School). All mice were C57BL/6 background, and were
   randomly allocated to experimental groups with the same age (8–12
   weeks) and sex. Mice were bred under specific pathogen‐free conditions
   and animal studies were approved by the Animal Care Facility of
   Shanghai Institute of Biochemistry and Cell Biology (Protocol No.
   IBCB0057), Chinese Academy of Science.

Antibodies and Reagents

   Antibodies recognizing p‐STAT1 (Tyr701 clone 58D6) (#9167), STAT1
   (D1K9Y) (#14994), p‐STAT2 (Tyr690 clone D3P2P) (#88410), STAT2 (D9J7L)
   (#72604), p‐STAT3 (Tyr705 clone D3A7) (#9145), STAT3 (124H6) (#9139),
   p‐IRF3 (Ser 396 clone 4D4G) (#4974), IRF3 (D83B9) (#4302), p‐TBK1
   (Ser172 clone D52C2) (#5483), TBK1 (clone D1B4) (#3504), p‐p38 MAPK
   (Thr180/Tyr182 clone D3F9) (#4511), p38 MAPK (D13E1) (#8690),
   p‐p44/42 MAPK (Erk1/2) (Thr202/Tyr204 clone D13.14.4E) (#4370),
   p44/42 MAPK (Erk1/2) (137F5) (#4695), and Myc (71D10) (#2278) were
   obtained from Cell Signaling Technology. Anti‐EBI2 polyclonal antibody
   (sc‐6640) was purchased from Santa Cruz. Horseradish peroxidase
   (HRP)‐conjugated GAPDH monoclonal antibody (HRP‐60004) and
   anti‐C3/C3b/C3c polyclonal antibody (21337‐1‐AP) were purchased from
   Proteintech. Anti‐EBI2 antibody (ab168744) and HRP‐β‐actin (ab49900)
   antibodies were obtained from Abcam. Monoclonal anti‐HA (H3663),
   poly(deoxyadenylic‐thymidylic) acid sodium salt (86828‐69‐5), 7α,
   25‐OHC (64907‐22‐8), and pristane (1921‐70‐6) were purchased from
   Sigma. Secondary HRP‐labeled antibodies against rabbit (A0208) or mouse
   (A0216) and BCA protein assay kit (P0012) were obtained from Beyotime.
   PE anti‐human GPR183 (368912) and PE anti‐mouse IgG2a, κ Isotype Ctrl
   (400212), APC‐cyanine 7 anti‐mouse F4/80 (123118), and Brilliant Violet
   421 anti‐CD45 (103133) antibodies were purchased from BioLegend.
   PerCP‐cyanine 5.5 anti‐human CD19 (45‐0199), FITC anti‐human CD14
   (11‐0149), APC anti‐human CD11b (17‐0118), FITC anti‐human CD4
   (11‐0049), APC anti‐human CD8a (17‐0087), FITC anti‐mouse CD11b
   (11‐0112), FITC anti‐mouse B220 (11‐0452), FITC anti‐mouse CD3e
   (11‐0031), APC anti‐mouse CD8a (17‐0081), PE anti‐mouse CD4 (12‐0041),
   FITC anti‐mouse Ly‐6G (11‐9668), APC anti‐mouse Ly‐6C (17‐5932),
   PE‐cyanine 7 anti‐mouse CD11c (25‐0114), PE anti‐mouse CD138 (281‐2),
   PerCP‐cyanine 5.5 anti‐mouse CD19 (45‐0193), APC anti‐mouse IFN‐γ
   (17‐7311), PE anti‐mouse IL‐17a (12‐7177), PE anti‐mouse TNF‐α
   (12‐7321), APC anti‐mouse CD183 (CXCR3) (17‐1831), Fixable Viability
   Dye eFluor 506 (65‐0866‐18) antibodies, and Rat anti‐mouse IFN gamma
   monoclonal antibody (R4‐6A2) (MM701) were obtained from eBioscience. PE
   anti‐mouse I‐A[b] (553552) antibody was obtained from BD Pharmingen.
   Lymphocyte separation medium (LSM, 0850494) was purchased from MP
   Biomedicals. CpG‐B (ODN1668) (tlrl‐1668‐1), R848 (resiquimod)
   (tlrl‐r848), imiquimod (R837) (tlrl‐imqs), and poly (I:C) (HMW)
   (tlrl‐pic) were purchased from InvivoGen. Polybrene transfection regent
   (TR‐1003‐G) was purchased from Merck Millipore. ANA Hep Screen ELISA
   (DE7020) was purchased from Demeditec. Anti‐dsDNA enzyme immunoassay
   test kit (0104000143) was obtained from Shanghai Kexin Biotech Co.,
   Ltd. Urea nitrogen (BUN) detection Kit (JN‐025698) was purchased from
   Shanghai Jining Industrial Co., Ltd. Creatinine assay kit
   (Colorimetric) (ab204537) was purchased from Abcam. Mouse
   IP‐10/CXCL10 ELISA Kit ([242]F10933) and mouse MCP‐1/CCL2 ELISA Kit
   ([243]F11130) were obtained from Shanghai Xitang Biotech Co., Ltd.
   Mouse IFN‐β ELISA Kit (42400) was purchased from PBL Assay Science.
   Mouse IL‐6 ELISA Kit (85‐88‐7964‐89) was purchased from eBioscience.
   Q‐VD‐OPh (HY‐12305) and ABT‐737 (HY‐50907) were purchased from MCE.

Immunoprecipitation and Immunoblotting Assay

   HEK293T cells or iBMDMs were washed with ice‐cold PBS and lysed with
   RIPA lysis buffer (50 mm Tris‐HCl, 150 mM NaCl, 1% Nonidet P40, 0.1%
   SDS, 1 mm EDTA, and 0.25% sodium‐deoxycholate) supplemented with
   protease inhibitors NaF, Na[3]VO[4], PMSF, and cocktails. Cell lysates
   were incubated with indicated antibodies, followed by protein G
   Sepharose beads incubation at 4 °C. The Sepharose beads were then
   washed three to five times with lysis buffer and resuspended in an
   appropriate amount of SDS‐PAGE loading buffer. The samples were
   separated by 8–10% SDS‐PAGE gel, followed by immunoblotting analysis
   with the indicated antibodies.

Plasmids and Transfection

   Human EBI2 was amplified from pcDNA3.1‐EBI2 plasmid (a gift from Dr.
   Hongwen Chen) and was cloned into pUltra vector or MIGR‐IRES‐GFP vector
   with N‐terminal FLAG tag, respectively. Mouse Gnai3 was amplified from
   reverse‐transcribed cDNA of PEMs and was cloned into the
   pcDNA3.1 vector with a C‐terminal HA tag. Human JAK1, JAK2, JAK3, and
   TYK2 cDNA were obtained from the 293T cell line and inserted into the
   MIGR‐IRES‐GFP vector with C‐terminal HA tag. pUltra‐EBI2 plasmids were
   transfected with the lentiviral packaging plasmids PASPX and pMD2 to
   HEK293T cells. Lentiviral supernatants were collected to infect iBMDMs,
   followed by GFP^+ cell sorting to generate stably transfected cells.
   For replenishing hEBI2 in primary Ebi2‐cKO BMDMs, BMs from WT and
   Ebi2‐cKO mice were collected and cultured with complete DMEM
   supplemented with recombinant mouse M‐CSF (25 ng µL^−1) (Novoprotein,
   CK02) for 2 days, then infected with retrovirus supernatants
   supplemented with M‐CSF (25 ng mL^−1) and polybrene (10 µg mL^−1) from
   MIGR‐EBI2 or MEGR‐GFP transfected platinum‐E (Plat‐E) cells every 12 h
   for 3 days. After infection, cells were again cultured with complete
   DMEM for another 12 h before ISD stimulation.

Cell Culture

   PEMs were harvested from mice with intraperitoneal (i.p.) injection of
   3% brewer thioglycollate medium (3 mL) for 4 days and then cultured in
   complete DMEM. BMDMs were generated after culturing BM cells for a week
   with complete DMEM containing M‐CSF (20 ng µL^−1). PEMs, BMDMs, iBMDMs,
   HEK293T, and plat‐E cells were cultured in complete DMEM supplemented
   with 10% v/v FBS, 1% v/v glutamine, and 1% penicillin/streptomycin
   (100 U mL^−1) at 37 °C with 5% CO[2].

LC‐MS Analysis of Sterols in Plasma and Peritoneal Irrigation Fluid

   Quantitative analysis of sterols including cholesterol, oxysterols, and
   steroid hormones was done using a reported LC‐MS/MS method^[ [244]^41
   ^] with some modifications.^[ [245]^42 ^] In brief, plasma (50 µL) was
   mixed with pre‐cooled methanol‐water (1 mL, 10:1, −20 °C) and methanol
   solution (10 µL) of internal standards including cholesterol‐d7,
   24‐hydroxycholesterol‐d7, campesterol‐d3, estrone‐d2, estradiol‐d2,
   dehydroepiandrosterone‐d2, and β‐sitosterol‐d7 (Steraloids, USA). After
   10 min centrifugation (12 000 rpm, 4 °C), supernatant was obtained and
   dried with nitrogen gas to obtain extract residues. The extract
   residues from each sample were, respectively, added with pyridine
   solution (100 µL) of derivatization reagents consisting of 2‐picolinic
   acid (8 mg), 2‐methyl‐6‐nitrobenzonic anhydride (22 mg), and
   dimethylaminopyridine (4 mg) from Aladdin (Shanghai, China). After
   standing at 80 °C for 60 min, the mixture was added with water (200 µL)
   and extracted with MTBE (1.5 mL). The upper layer was dried under
   nitrogen gas and then re‐dissolved in acetonitrile (100 µL) for
   LC‐MS/MS analysis.

   LC‐MS/MS analysis was conducted on an UHPLC‐MS system with a 30AD UHPLC
   (Shimadzu, Japan) and a QTRAP 6500plus mass spectrometer (SCIEX, USA).
   All data were acquired using Analyst (V1.7, Sciex) and processed with
   OS (V2.0, Sciex). The chromatographic separation was performed on a
   Zorbax Eclipse Plus C18 column (100 × 2.1 mm, 1.8 µm, Agilent, USA)
   with injection volume of 5 µL; gradient elution was done with water (A)
   and acetonitrile (B) containing 0.1% formic acid in both. The optimized
   mass spectrometer parameters for electrospray ionization source were
   employed including ion‐spray voltage (5500 V), curtain gas (40 psi),
   and source temperature (550 °C). Quantification of all detectable
   sterols was achieved using these internal standards based on method
   validation.^[ [246]^42 ^]

The Selection of Human Subjects

   The current study was approved by the Ethics Committee of Renji
   Hospital, Shanghai Jiao Tong University (Approval No. KY2022‐023‐B) and
   Shanghai Tenth People's Hospital, Tongji University (Approval No.
   SHYS‐IEC‐5.0/22K234/P01). All SLE patients fulfilled the 1997 American
   College of Rheumatology (ACR) classification criteria for SLE. Patients
   with DM, RA, and SS were also included as disease controls. DM patients
   fulfilled the Bohan and Peter's criteria. RA patients fulfilled the
   2010 ACR/EULAR RA classification criteria. SS patients fulfilled the
   2017 ACR/EULAR classification criteria for primary SS. All HCs had no
   history of autoimmune diseases. All study subjects were enrolled in the
   study after giving informed and written consent. Information of SLE
   patients, disease controls, and healthy donors were listed in Tables
   [247]S1 and [248]S2, Supporting Information.

Purification and Fractionation of Human PBMCs

   The peripheral blood of healthy donors and SLE patients were collected
   in anticoagulation tubes, diluted with PBS, and gently layered over an
   equal volume of LSM in a conical tube and centrifuged at 2200 rpm for
   25 min in a swinging bucket rotor without brake. After density gradient
   centrifugation, four layers formed, and the second layer containing
   mononuclear cells was carefully transferred into a new conical tube,
   filled with PBS and centrifuged at 1200 rpm for 10 min. After
   aspirating the supernatant, isolated PBMCs were immunostained for flow
   cytometric assay or were resuspended in Trizol reagent (Takara)for RNA
   extraction. The remaining PBMCs were further fractionated into
   T‐lymphocytes and monocytes by a T cell negative selection kit (#17951,
   STEMCELL Technologies) and a CD14 positive selection kit (#17818,
   STEMCELL Technologies). Briefly, PBMCs were mixed with the selection
   cocktails (5 µL per 10^7 cells) and incubation at room temperature (RT)
   for 5 min. After adding recommend medium to top up the sample to the
   indicated volume, the mixtures were subsequently placed into the magnet
   to acquire the supernatant or attached beads. The purified cells were
   resuspended in Trizol reagent for RNA extraction. For human CD3^+ T
   cells induction, 48‐well plate was pre‐coated with anti‐CD3
   (5 µg mL^−1) and anti‐CD28 (5 µg mL^−1) antibodies at 4 °C overnight,
   then discarded the coating buffer and washed with PBS for three times.
   CD3^+ T cells were seeded at 1 × 10^6 cells/well in 48‐well plate for
   24 h, and the supernatant (CD3 conditional medium, CD3 C.M) was
   collected to stimulated human CD14^+ monocytes.

RNA Interference, RNA Extraction, and Real‐Time Fluorescence Quantitative PCR

   RNA interference was performed with single siRNA or a mixture of two
   different siRNAs for one gene. siRNA oligonucleotides of Jak1, Jak2,
   Stat1, Stat2, Ebi2, Gani1, Gani2, Gani3, and Gnas were all designed by
   siRNA selection program and synthesized from GenePharma. Murine PEMs
   were transfected with siRNA (40 nm) using Lipofectamine RNAiMAX reagent
   (Invitrogen). The siRNA sequences are shown in Table [249]S3,
   Supporting Information. Total RNA was extracted with Trizol reagent,
   and cDNA was obtained with reverse transcriptase M‐MLV (Takara) for
   quantitative real‐time PCR (qRT‐PCR) using SYBR Green Dye (2043, DBI
   Bioscience) on a LightCycler480 (Roche). The primers for qRT‐PCR are
   shown in Table [250]S4, Supporting Information.

TMPD‐Induced IFN Signature and Nephritis

   To induce IFN signature, 8‐week‐old mice (WT and Ebi2‐cKO) received a
   single i.p. injection of TMPD (0.5 mL) as previously described. Control
   mice received PBS (0.5 mL) at the same time. Female mice were selected
   for the study because they were more susceptible to the disease.
   2 weeks after injection, mice were euthanized and peritoneal cavities
   were washed with PBS (5 mL) containing EDTA (2 mm) to harvest total
   peritoneal cells. For TMPD‐induced nephritis, mice were given TMPD by
   i.p. injection (0.5 mL per mouse) for 12 weeks. The urine of
   experimental animals was then collected for proteinuria, creatinine,
   and urea nitrogen analysis using BCA protein assay kit, creatinine
   assay kit, and urea nitrogen detection kit according to the
   manufacturer's instructions, respectively. The peripheral blood cells,
   peritoneal cells, LN cells, and splenocytes were collected for flow
   cytometric assay. The kidney, spleen, lung, and liver tissues of mice
   were collected after TMPD injection for 2 and 12 weeks to perform H&E
   staining or immunofluorescence assay. All animal procedures were
   conducted in strict accordance with the institutional guidelines and
   were approved by the Institutional Animal Care and Use Committee of
   Shanghai Institute of Biochemistry and Cell Biology (IBCB0057).

ELISA

   Concentrations of IL‐6, CXCL10, and CCL2 in serum or PIF of
   TMPD‐treated mice, or IFN‐β in cell culture supernatants were measured
   by commercial ELISA kits according to the manufacturer's instructions.

Histopathological and Immunofluorescence Analysis

   For histopathological analysis, the kidney, spleen, lung, and liver
   tissues of TMPD‐treated mice were fixed in 4% paraformaldehyde (PFA)
   and embedded in paraffin. After serially cutting, the tissue sections
   were stained with H&E. The pathological changes were evaluated by
   BX51 light microscope (Olympus). For immunofluorescence analysis,
   kidney tissues were fixed in 4% PFA and dehydrated with 30% sucrose,
   and embedded in OCT compound. Kidney slices (10 µm in thickness) were
   cut using microtome (Leica), and blocked with PBS containing 5% donkey
   serum for 30 min at RT, and then incubated with anti‐C3 antibody at
   appropriate final dilutions in a humidified chamber at RT for 2 h.
   After washing with PBS for three times, the tissue slices were
   incubated with Cy3 donkey anti‐rabbit and F488 donkey anti‐mouse
   antibodies for 1 h, followed by staining the nucleus with DAPI for
   10 min at RT. Tissue slices were mounted and then examined using a
   Leica TCS SP8 confocal microscope. Spleen slices (4 µm in thickness)
   from paraffin‐embedded tissues was stained with anti‐EBI2 antibody at
   RT for 2 h, followed by F488 donkey anti‐mouse antibody and DAPI
   staining for 1 h to detect EBI2 expression.

CD8^+ T Cells Migration Assay

   Naïve CD8^+ T cells were isolated from spleens of WT mice with BD IMag
   anti‐mouse CD8a particles and stimulated with anti‐CD3 (2 µg mL^−1) and
   anti‐CD28 (2 µg mL^−1) for 12 h. CD8^+ T cells were collected and
   washed by migration media (HBSS with 0.1% protease‐free BSA), and
   adjusted volume to 5 × 10^5 cells/100 µL. Transwell inserts were
   removed from the 24 well plates (Corning costar, 3421), and migration
   media (600 µL) was added with recombinant human CXCL10 (100 ng mL^−1)
   (Novoprotein, [251]P02778) or conditional medium (the ratio of complete
   RPMI1640 medium and PIF of TMPD‐treated WT or Ebi2‐cKO mice was 2:1).
   The Transwell inserts were placed onto the receiving plates and CD8^+ T
   cells (100 µL) were added into the inserts to observe the migration of
   T cells every 1 h. The inserts were removed at proper time, and Hoechst
   33342 was added to the receiving plates. The cell number in the
   receiving plates was recorded by Operetta (PerkinElmer).

PEMs Migration Assay

   For PEMs migration, PEMs (2 × 10^5) were seeded in the top of the
   Transwell inserts (Corning costar, 3422) in a complete DMEM, and
   conditional medium (the ratio of complete DMEM and PIF of TMPD‐treated
   WT or Ebi2‐cKO mice was 2:1) was added in the well below. After
   incubation for 6 h, the inserts were washed using 1× PBS and the cells
   were fixed with 100% methanol for 15 min, followed by cell staining
   using 0.5% crystal violet (filtered) for 30 min. The upper membrane of
   the inserts was next cleaned to remove excess stain and non‐migrated
   cells. Five microscopic fields of each insert (one field in the center
   and four fields in the periphery of the membrane) were captured by
   BX51 light microscope (Olympus).

RNA Sequencing Analysis

   Monocytes (CD11b^+Ly6c^+ cells) were sorted from peritoneal cells of WT
   or Ebi2‐cKO mice with TMPD treatment for 2 weeks. Total RNA was
   extracted with Trizol reagent (Invitrogen) and subjected to RNA‐seq
   analysis using an Illumina HiSeq‐X‐ten sequencer in Shanghai Majorbio
   Bio‐Pharm Technology Co., Ltd. Hisat2 was used for genome mapping, and
   the mapping ratio was over 90% across all the samples in the dataset.
   Fragments per kilobase of exon model per million mapped reads (FPKM)
   were used to normalize gene expression. StringTie was run with FPKM
   values for known gene models. edgeR was used to identify DEGs. The
   p‐value significance threshold in multiple tests was set by the false
   discovery rate (FDR). The fold‐changes were estimated according to the
   FPKM in each sample. The DEGs were selected using the following filter
   criteria: FDR ≤ 0.05 and fold‐change ≥ 2.

Statistical Analysis

   All sample sizes were large enough to ensure proper statistical
   analysis. Data were represented as the mean ± SEM from at least three
   different experiments. Statistical analysis was performed using
   GraphPad Prism8 software, version 8 (GraphPad Software, Inc.).
   Statistical significance was calculated using Student's two‐tailed
   unpaired t‐test or ANOVA with Holm–Sidak multiple comparisons test.
   Correlations were done using Pearson's test. ns, no significance (p >
   0.05). p < 0.05 was considered statistically significant (*p < 0.05,
   **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   F.Z., B.Z., and H.D. contributed equally to this work. B.W., H.W., and
   F.Z. designed research. F.Z. and B.Z. performed the majority
   experiments and statistical analysis. H.D. and N.S. provided the blood
   samples and information of patients with SLE, RA, SS, and DM and
   offered professional advice. Xiang.L., X.W., Xia.Z., Q.L., Q.F., L.Q.,
   D.Y., Xiao.L., Xin.Z., and Q.Z. participated in part of the
   experiments. M.H. contributed the scRNA‐seq analysis data. L.C. and
   H.T. detected cholesterol and its derivatives in plasma of SLE patients
   by LC‐MS. J.Q. and Z.Z. detected cholesterol and its derivatives in
   peritoneal lavage fluid of lupus mice by LC‐MS. F.Z. drafted the
   manuscript. B.W.,H.W.,F.Z., B.Z., and Xiang.L.reviewed and edited the
   manuscript. All authors read and approved the final manuscript.

Supporting information

   Supporting Information
   [252]Click here for additional data file.^ (3.3MB, pdf)

Acknowledgements