Sexual dimorphism of early GPR183-dependent B cell responses in systemic lupus erythematosus

Summary

Sensing oxidized cholesterol (oxysterol) ligands by GPR183-expressing B cells spatially regulates B cell activation within secondary lymphoid tissues, which, when dysregulated, could contribute to B cell-driven autoimmunity. Here, we show that in systemic lupus erythematosus, GPR183-expressing B cells are reduced in both humans and mice with established disease, irrespective of sex. However, we further show that GPR183-expressing splenic B cells are increased during the initiation phase of lupus-like disease in female but not male mice, leading to sex differences in the ability of B cells to migrate toward GPR183’s principal ligand 7α,25dihydroxycholesterol in vitro and B cell activation in vivo. Accordingly, disrupting early, but not late, GPR183-dependent responses reduces B cell activation and suppresses the severity of lupus-like disease in female mice but not male mice. These data demonstrate that the impact of GPR183-expressing B cells on disease pathology may change over the course of SLE and is sexually dimorphic.

Graphical abstract

Highlights

• •GPR183⁺ B cells are increased early post-lupus initiation in female, not male, mice
• •GPR183 blockade disrupts early B cell activation post-lupus initiation in female, not male, mice
• •GPR183 blockade at lupus initiation reduces disease severity in female, not male, mice

Biological sciences; immunology

Introduction

Systemic lupus erythematosus (SLE) is a complex multi-system disease presenting a wide variety of clinical symptoms, whose immunopathology is characterized by the production of autoantibodies, including anti-nuclear autoantibodies (ANA) by B cells.¹ Another notable feature of SLE is the strong female sex-bias observed in individuals with disease onset between post-pubertal and pre-menopausal years.² Sexual dimorphism in the risk of developing SLE is thought to be due to a complex interplay between sex hormones and sex chromosome complement. Recent evidence has firmly placed extrafollicularly activated B cells, those that do not go through the germinal center (GC), as an important pool of autoreactive B cells in SLE³ driven by both exposure to type 1 interferons (IFN1) and overactive toll-like receptor (TLR)7 signaling.⁴ However, it is unknown whether additional signals drive aberrant B cell activation in SLE.

GPR183 is a G protein-coupled receptor, which was first identified as the Epstein-Barr virus (EBV)-induced molecule EBI2.^5,6 As well as the historical link between EBV infection and increased SLE risk,⁷ GPR183 gene and protein expression are dysregulated in SLE. GPR183 was identified as a key dysregulated gene in an integrated analysis of RNA sequencing data that generated an SLE MetaSignature⁸ and GPR183 expression has been reported to be abnormally low on both T cells and B cells in the peripheral blood of people with adult-onset SLE.⁹ No equivalent studies exist in juvenile-onset SLE (JSLE), which is known to have a worse disease trajectory and outcomes¹⁰ than people who develop SLE in adulthood.

GPR183 has been shown to regulate the spatial location of B cells within secondary lymphoid tissues. Following exposure to antigen, B cells upregulate GPR183, guiding B cells along gradients of GPR183’s main ligand 7α,25dihydroxycholesterol into extrafollicular regions where they differentiate into plasmablasts.^11,12,13 This gradient is controlled by stromal cells and dendritic cells which metabolise cholesterol into 7α,25dihydroxycholesterol via expression of the enzymes cholesterol-25-hydroxylase (CH25H) and cytochrome P450 family 7 subfamily B member 1 (Cyp7B1).¹⁴ After approximately 6–7 days following antigen exposure, extrafollicular responses are dampened: GPR183 expression downregulates and CXCR5 upregulates, and B cells move into the follicles to take part in GC reactions.^3,4,11,13 However, the studies identifying these processes were carried out using model systems such as sheep’s red blood cells^11,15 and no equivalent studies exist using stimuli that induce an autoimmune response.

Here, we demonstrate that, similarly to previous data generated from people with adult-onset SLE, B cells expressing GPR183 are reduced in the peripheral blood of people with JSLE who have established disease compared to age-matched healthy controls. Using the parent into F1 model of chronic graft versus host disease, which recapitulates some features of lupus-like disease, including an autoantibody-driven glomerulonephritis and has a strong sex bias,^16,17 we also show that B cells expressing GPR183 are reduced in the spleen of mice with established nephritis. However, the temporal assessment of GPR183 expression over the course of disease showed that GPR183 is upregulated very early post-initiation of disease on splenic B cells but only in female, not male, mice. Splenic B cells isolated from female mice during this period showed enhanced chemoattractant sensitivity to 7α,25dihydroxycholesterol, GPR183's main ligand, compared to male mice. In vivo, this influenced sex differences in the magnitude of splenic B cell activation following disease induction. Accordingly, treatment with a small molecule antagonist for GPR183, NIBR189, during the initiation phase of disease suppressed plasmablast differentiation and prevented nephritis development in female but not male mice. Interestingly, targeting GPR183-dependent responses later in the disease course had no impact on disease severity. Collectively, our data show sexual dimorphism in early GPR183-dependent B cell activation and demonstrate that GPR183 expression and regulation change over the course of lupus-like disease.

Results

GPR183-expressing B cells are reduced in people with established juvenile-onset systemic lupus erythematosus

Considering historical data showing that GPR183 is reduced on B cells in people with adult-onset SLE, we first performed the in-depth immunophenotyping of GPR183-expressing B cells in the peripheral blood to determine if we could see similar signals in our cohort of people living with established JSLE. This demonstrated that the percentage of GPR183⁺ B cells, as well as the gpr183 transcript level were significantly lower in B cells from people with JSLE compared to those from age- and sex-matched healthy controls (Figures 1A and 1B). Interestingly, in our cohort, there was no significant difference in GPR183 expression at the protein or transcript level in any other major immune cell subsets, namely CD8⁺, CD4⁺, or CD14⁺ cells (Figures S1A and S1B). Overlaying GPR183⁺ B cells on canonical B cell gating based on the expression of CD24 and CD38 demonstrated that GPR183-expressing B cells mainly overlapped with CD38⁻CD24⁺ (memory) B cells (Figure 1C), with a clear population co-expressing CD27 and GPR183 (Figure 1D), matching previous reports from multiple sclerosis showing that GPR183 expression is enriched in memory B cells.¹⁸ Similarly to previous reports that both adult- and juvenile-onset SLE are characterized by a reduction in peripheral blood memory B cells,^19,20 there was a stark reduction in the frequency of GPR183⁺CD27⁺ memory B cells when comparing individuals with JSLE to controls (Figure 1D). In JSLE, GPR183⁺CD27⁺ memory B cells could be defined by hallmarks of SLE-driven B cell dysfunction such as reduced expression of the follicle-positioning molecule CXCR5, increased expression of CD11c, a strong IFN1 signature, and increased propensity to be class-switched to IgG in people with JSLE compared to controls (Figures 1E–1G and S1C). These phenotypic differences were not accompanied by statistically significant changes in the abundance of GPR183’s main ligands 7α,25-dihydroxycholesterol (7α25diOHC) and 7α,27-dihydroxycholesterol (7α27diOHC) in the serum (Figure S1D). Stratification of our phenotypic data by clinical parameters demonstrated that the reduction in GPR183⁺CD27⁺ memory B cells in people with JSLE was not influenced by current disease activity, history of nephritis, or sex (Figures 1H–1J). While the proportions of IgG⁺GPR183⁺CD27⁺ memory B cells were also not affected by current disease activity or history of nephritis (Figures 1K and 1L), healthy females had a significantly higher proportion of IgG⁺GPR183⁺CD27⁺ memory B cells compared to healthy males (Figure 1M). This sex bias was lost when comparing the proportion of IgG⁺GPR183⁺CD27⁺ memory B cells between males and females with JSLE (Figure 1L).

Figure 1.

GPR183-expressing B cells are reduced in patients with juvenile-onset SLE compared to age- and sex-matched healthy controls

(A) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of GPR183⁺ out of total B cells in the peripheral blood (PB) of healthy controls (HC) (n = 74) and patients with juvenile-onset systemic lupus erythematosus (JSLE) (n = 58).

(B) Dot plot shows the relative expression of GPR183 against the housekeeping gene 18S from sorted B cells isolated from the PB of HC (n = 5) and patients with JSLE (n = 5) as measured by qPCR.

(C) Representative flow cytometry plot shows an overlay of GPR183⁺ B cells (red) on total B cells (black) from one representative healthy donor.

(D) Representative flow cytometry plots (left) and dot plot (right) show the percentage of GPR183⁺CD27⁺ B cells out of total B cells in HC (n = 74) and patients with JSLE (n = 58).

(E) Representative flow cytometry plots (left) and dot plot (right) show the percentage of CXCR5⁺ cells out of GPR183⁺CD27⁺ B cells in the peripheral blood (PB) of HC (n = 30) and patients with JSLE (n = 28).

(F) Representative flow cytometry plots (left) and dot plot (right) show the percentage of CD11c⁺ cells out of GPR183⁺CD27⁺ B cells in the PB of HC (n = 30) and patients with JSLE (n = 28).

(G) Representative flow cytometry plots (left) and dot plot (right) show the percentage of IgG⁺ cells out of GPR183⁺CD27⁺ B cells in the PB of HC (n = 33) and patients with JSLE (n = 30).

(H) Dot plot shows the percentage of GPR183⁺CD27⁺ B cells out of total B cells in the PB of inactive (n = 49) and active (n = 9) JSLE patients.

(I) Dot plot shows the percentage of GPR183⁺CD27⁺ B cells out of total B cells in the PB of JSLE patients without nephritis (n = 30) and JSLE patients with nephritis of class II-V (n = 20).

(J) Dot plot shows the percentage of GPR183⁺CD27⁺ B cells out of total B cells in the PB of female HC (n = 49), male HC (n = 26), female JSLE patients (n = 42), and male JSLE patients (n = 16).

(K) Dot plot shows the percentage of IgG⁺GPR183⁺CD27⁺ B cells out of total B cells in the PB of inactive (n = 25) and active (n = 5) JSLE patients.

(L) Dot plot shows the percentage of IgG⁺GPR183⁺CD27⁺ B cells out of total B cells in the PB of JSLE patients without nephritis (n = 18) and JSLE patients with nephritis of class II-V (n = 10).

(M) Dot plot showing the percentage of IgG⁺GPR183⁺CD27⁺ B cells out of total B cells in the PB of female HC (n = 19), male HC (n = 14), female JSLE patients (n = 17), and male JSLE patients (n = 11). Lines in dot plots represent mean ± SEM. Cumulative data of at least 3 independent experiments is shown. A–I, K, L Unpaired t test and J, M One-way ANOVA test were used to determine the significance of the difference between groups.

B-cell activation is heightened in females compared to males following the induction of experimental nephritis

To explore our findings further and to evaluate secondary lymphoid tissues, an important site for GPR183 functionality, we next utilised the parent into F1 model of chronic graft versus host disease (cGVHD). Notably, cGVHD recapitulates some features of human lupus pathology including an autoantibody-dependent nephritis, which usually develops 4–5 weeks post-induction, and has a strong sex bias (¹⁷; Figure S2A). In line with the evidence that estrogen has a strong influence on B cell phenotypic profiles²¹ and sex-biased risks in developing lupus,²² treatment with the oestrogen-receptor antagonist fulvestrant also significantly suppressed nephritis severity in female mice (Figure S2B). Similarly to our data from humans, there was a reduction in GPR183-expressing splenic B cells in mice that have already developed nephritis (9 weeks post-lupus induction), compared to naive controls, regardless of sex (Figures 2A and 2B). However, when assessing the temporal regulation of GPR183 on splenic B cells post-nephritis induction, we found that there was a significant, albeit small, upregulation in GPR183 very early post-disease induction (24–72 h) during the initiation phase of disease in female but not male mice, which was already downregulated by day 14 post-disease induction (Figures 2C–2E). B cells isolated from female mice during this early phase of disease also demonstrated enhanced ability to migrate toward 7α,25dihydroxycholesterol compared to equivalent B cells from male mice, suggesting that this small change in GPR183 expression was biologically relevant (Figures 2F and 2G).

Figure 2.

GPR183-expressing B cells from female mice increase in frequency early post-nephritis induction and demonstrate stronger chemotactic potential toward 7α25dihydroxycholesterol unlike their male counterparts

(A) Representative flow cytometry plots show the percentage of GPR183⁺ out of total B cells in female and male naive mice, and at 9 weeks post-nephritis induction (established disease).

(B) Summary dot plot shows the frequency of GPR183⁺B cells in naive and 9 weeks post-disease induction females and males (n = 3 per group).

(C) Representative flow cytometry plots show the percentage of GPR183⁺ out of total B cells in female (top) and male (bottom) naive mice, at 24-72H post-nephritis induction and at day 14 post-nephritis induction.

(D) Summary dot plot shows the frequency of GPR183⁺B cells in naive females (n = 23), females at 24-72H post-nephritis induction (n = 12), and females at day 14 post-nephritis induction (n = 14).

(E) Summary dot plot shows the frequency of GPR183⁺B cells in naive males (n = 19), males at 24-72H post-nephritis induction (n = 12), and males 14 days post-nephritis induction (n = 8).

(F) Schematic showing the transwell experiment for assessing the migration of B cells isolated from female/male mice at 24 post-nephritis induction toward the GPR183 ligand 7α,25dihydroxycholesterol (7α25-diHC).

(G) Summary dot plot shows the fold change of migrated B cells toward 7α25-diHC versus migrated B cells with no chemoattract signal (control migrated B cells) from samples isolated from female (n = 9) and male mice (n = 6) 24H post-nephritis induction.

(H) Representative immunofluorescence images (scale bar = 50 μm) showing the number and size of germinal centers (GL7⁺) within B cell follicles in the spleen of female and male naive mice, mice at 24-72H post-nephritis induction, and mice 14 days post-lupus nephritis induction (left) and enumeration of the size of the germinal centers (right) (n = 5 per group).

(I) Representative immunofluorescence images (scale bar = 50 μm) showing the number and size of extrafollicular foci (CD138⁺) in the spleen of female and male naive mice, mice at 24-72H post-nephritis induction, and mice 14 days post-lupus nephritis induction (left) and enumeration of the size of the extrafollicular foci (right) (n = 5 per group). Lines in dot plots represent mean ± SEM. C–I Cumulative data of at least 3 independent experiments is shown. B, D, E One-way ANOVA test and G–I Mann-Whitney test were used to determine the significance of the difference between groups.

GPR183 upregulation on B cells following exposure to antigen has been previously shown to control the early trafficking of B cells into the outer follicular regions of lymphoid tissue for the differentiation of plasmablasts.^11,12 Thus, we next wanted to assess whether the greater upregulation of GPR183 on female compared to male B cells post-disease initiation was associated with sex differences in the spatial regulation of B cell responses post-disease induction. We found that both GL7⁺ cells in B cell follicles (GCs) and outer follicular foci containing CD138⁺ plasmablasts increased in size and number over time in both male and female animals post-nephritis induction compared to naive controls (Figures 2H and 2I). However, by day 14 post-disease induction, both GCs and outer follicular foci containing CD138⁺ plasmablasts were significantly larger in size in female compared to male mice (Figures 2H and 2I). To assess whether these sex differences were dependent upon the very early upregulation of GPR183 we administered NIBR189, a small molecule antagonist of GPR183²³, which has been previously used to modulate GPR183-dependent B cell responses in vivo,^24,25 five times per week to both female and male mice. We next assessed the frequency of CD19⁺CD95⁺GL7⁺ GC B cells and CD138⁺Blimp-1⁺ plasmablasts at day 14 post-disease induction using flow cytometry. Similarly to our analysis using immunofluorescence, this analysis demonstrated that when compared to naive mice, there was a much greater increase in CD19⁺CD95⁺GL7⁺ GC B cells in female mice than in male mice, while CD138⁺Blimp-1⁺ plasmablasts were only expanded in female mice (Figures 3A–3D). It also demonstrated that while NIBR189 treatment did not impact the frequency of CD19⁺CD95⁺GL7⁺ GC B cells in either sex, it significantly suppressed female-biased plasmablast induction (Figures 3A–3D). We also assessed the frequency of CD11c⁺ B cells at day 14 post-disease induction. In agreement with previously published data,²⁶ CD11c⁺ B cell populations were only expanded in female mice at day 14 post-disease induction, an increase that was also suppressed by NIBR189 (Figures S2C and S2D).

Figure 3.

Early plasmablast differentiation post-nephritis induction is suppressed by GPR183 antagonism exclusively in female mice

(A) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of CD19⁺CD95⁺GL7⁺ GC B cells in naive (n = 21) as well as NIBR189-treated (n = 11) and control (n = 23) female mice at 14 days post-lupus nephritis induction.

(B) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of CD19⁺CD95⁺GL7⁺ GC B cells in naive (n = 19) as well as NIBR189-treated (n = 11) and control (n = 20) male mice at 14 days post-lupus nephritis induction.

(C) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of CD19⁺CD138⁺Blimp-1⁺ plasmablasts in naive (n = 21) as well as NIBR189-treated (n = 13) and control (n = 22) female mice at 14 days post-lupus nephritis induction.

(D) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of CD19⁺CD138⁺Blimp-1⁺ plasmablasts in naive (n = 19) as well as NIBR189-treated (n = 13) and control (n = 21) male mice at 14 days post-lupus nephritis induction. Lines in dot plots represent mean ± SEM. Cumulative data of at least 3 independent experiments is shown. A–D One-way ANOVA test was used to determine the significance of the difference between groups.

GPR183 antagonism reduces the severity of experimental lupus nephritis by suppressing the differentiation of antibody-secreting cells in female but not male mice

Having established that the upregulation of GPR183 during the initiation phase of disease regulates, at least in part, the early differentiation of plasmablasts in female but not male mice post-nephritis induction, we next wanted to interrogate how this influenced disease severity. To address this, we administered NIBR189 5 times weekly during the initiation phase of disease (0–2 weeks post disease initiation) and then as a maintenance dose (twice per week) throughout the rest of the disease course in female and male mice. We found that NIBR189-treated female mice developed significantly less severe nephritis as measured by proteinuria compared to untreated female mice (controls) (Figure 4A). Histological assessment of nephritis severity and kidney damage also demonstrated a reduction in disease severity in NIBR189-treated compared to control female mice (Figure 4B). Assessment of the splenic B cell compartment between NIBR189-treated and control female mice with nephritis demonstrated that NIBR189 treatment did not impact the frequency of CD19⁺CD95⁺GL7⁺ GC B cells but suppressed the frequency of CD19⁺CD138⁺Blimp-1⁺ plasmablasts, as well as total IgG and ANA levels in the serum (Figures 4C–4F). Conversely, in male animals, NIBR189 treatment did not suppress disease severity as measured by proteinuria or assessment of kidney histology or have any impact on CD19⁺CD95⁺GL7⁺ GC B cells, CD19⁺CD138⁺Blimp-1⁺ plasmablasts, total IgG levels, or ANA levels in the serum (Figures 5A–5F). Importantly, in line with our observations that GPR183 is downregulated later on in the disease course in both humans and mice, when NIBR189 was administered only later at both 5 weeks and 9 weeks during the disease course as nephritis progresses, it had no impact on disease severity regardless of sex (Figures S2E–S2H). Altogether, these data demonstrate that GPR183 seems to predominantly impact the processes that control sexual dimorphism during early B cell activation rather than pathways directly controlling disease progression.

Figure 4.

GPR183 antagonism suppresses nephritis severity, plasmablast differentiation, and autoantibody production in female mice

(A) Line graph shows urine proteinuria score in NIBR189-treated (n = 14) compared to control (n = 17) female mice with lupus nephritis.

(B) Representative images (left, scale bar = 100 μm) and summary dot plot (right) show histological scoring of kidney pathology in NIBR189-treated (n = 11) compared to controls (n = 13) female mice with lupus nephritis.

(C) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of splenic CD19⁺CD95⁺GL7⁺ Germinal Center (GC) B cells in NIBR189-treated (n = 14) and control (n = 18) female mice with lupus nephritis.

(D) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of CD19⁺CD138⁺Blimp-1⁺ plasmablasts in NIBR189-treated (n = 14) and control (n = 17) female mice with lupus nephritis.

(E) Summary dot plot shows the abundance of total IgG (μg/mL) in the sera of NIBR189-treated (n = 7) and control (n = 8) female mice with lupus nephritis.

(F) Summary dot plot shows the abundance of anti-nuclear antibodies (μg/mL) in the sera of NIBR189-treated (n = 11) and control (n = 14) female mice with lupus nephritis. Lines in line graphs and dot plots represent mean ± SEM. Cumulative data of at least 3 independent experiments is shown. A two-way ANOVA test and B–F Mann-Whitney test were used to determine the significance of the difference between groups.

Figure 5.

GPR183 antagonism does not suppress nephritis severity, plasmablast differentiation, and autoantibody production in male mice

(A) Line graph shows urine proteinuria score in NIBR189-treated (n = 14) compared to control (n = 23) male mice with lupus nephritis.

(B) Representative images (left, scale bar = 100 μm) and summary dot plot (right) show histological scoring of kidney pathology in NIBR189-treated (n = 9) compared to controls (n = 14) male mice with lupus nephritis.

(C) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of splenic CD19⁺CD95⁺GL7⁺ Germinal Center (GC) B cells in NIBR189-treated (n = 11) and control (n = 17) male mice with lupus nephritis.

(D) Representative flow cytometry plots (left) and summary dot plot (right) show the percentage of CD19⁺CD138⁺Blimp-1⁺ plasmablasts in NIBR189-treated (n = 13) and control (n = 20) male mice with lupus nephritis.

(E) Summary dot plot shows abundance of total IgG (μg/mL) in the sera of NIBR189-treated (n = 7) and control (n = 8) male mice with lupus nephritis.

(F) Summary dot plot shows abundance of anti-nuclear antibodies (μg/mL) in the sera of NIBR189-treated (n = 11) and control (n = 18) male mice with lupus nephritis. Lines in line graphs and dot plots represent mean ± SEM. Cumulative data of at least 3 independent experiments is shown. A two-way ANOVA test and B–F Mann-Whitney test were used to determine the significance of the difference between groups.

Discussion

Studies comparing the immune system between males and females have long established that there are sex-based differences in B cell responses, with females displaying higher antibody titers post-vaccination than males and an increased risk of developing autoantibody-driven autoimmunity.²⁷ To date, sexual dimorphism in B cell responses is thought to be due to a complex interplay between sex hormone concentration and chromosome complement and we have recently demonstrated that increased levels of class-switched memory B cells in humans are uniquely dependent upon both estrogen and the presence of two X chromosomes.²¹ However, the exact downstream mechanisms controlling sex differences in the magnitude of B cell responses and class-switching profiles in different immunological situations remain ill-defined. Here, we demonstrate that sexual dimorphism in the expression of oxysterol receptor GPR183 following B cell activation and subsequent differences in the spatial location of B cells in secondary lymphoid tissues may be one such mechanism. We also show that the impact of GPR183 on immune responses seems to change over the course of lupus-like disease, highlighting the tight, time-dependent modulation of chemoattractant signaling across early and late phases of the immune response.

Oxysterols are formed as intermediates in the cholesterol biosynthesis pathway. Through their ability to interact with extracellular receptors (e.g., GPR183 and CXCR2) and binding to nuclear transcription factors (the liver X receptors; the RAR-related orphan receptors α,β,γ; estrogen receptors), oxysterols can influence a wide range of immune cell functions.²⁸ In mice, the interaction of GPR183 with its ligands controls B cell chemotaxis and the positioning of B cells within the outer follicular regions of B cell follicles in secondary lymphoid organs.²⁹ Our data complement these studies by also showing that in experimental lupus-like nephritis GPR183-oxysterol interactions control both the early induction of plasmablasts and CD11c⁺ B cells, which are thought to be derived extrafollicularly,³ and further demonstrate that these responses are sexually dimorphic. Oxysterol sensing by immune cells has been shown to play a critical role in numerous experimental models of inflammation. For example, oxysterol sensing by GPR183⁺ innate lymphoid cells (ILCs) is an essential signal for their localisation in isolated lymphoid follicles and the development of experimental colitis.³⁰ GPR183 also plays an important role in the migration of encephalitogenic CD4⁺ T cells into the central nervous system of mice with experimental autoimmune encephalitis (EAE) early in the disease course.³¹ Whether sexual dimorphism in GPR183-dependent responses also affects other immune cells and immune-mediated pathologies would be an interesting area for future study. As well as GPR183, GPR174 has been shown to regulate the sexually dimorphic spatial location of B cells in secondary lymphoid tissues in healthy immune responses. More specifically, testosterone conditions GC B cell responses in males by influencing the GPR174-CCL21 axis that regulates the movement of B cells into the central B cell follicle, leading to reduced GC responses.³² Considering we found that GPR183 antagonisms did not influence the frequency of GC B cells, together, our data suggest a possible scenario by which GPR174-CCL21 and GPR183-oxysterol interactions work together to modulate sexual dimorphisms in the spatial location of B cells within secondary lymphoid tissues. This would be an interesting area for future studies.

In line with reports that GPR183 regulates early pathogenic responses in EAE but is redundant in the progression of disease,³¹ we also found that blockade of GPR183-signalling did not act therapeutically to suppress ongoing proteinuria despite its potent anti-inflammatory effects when administered during the initiation phase of disease. Thus, GPR183-dependent B cell responses may be more likely to influence the increased risk of females developing SLE rather than driving ongoing pathology. In line with this, we found that there was a reduction of GPR183-expressing B cells in people with JSLE with established disease compared to controls, and we also observed that GPR183 was reduced in mice with established nephritis. Notably, a reduction in GPR183 protein and transcript expression has previously been identified in B cells from adult-onset SLE, suggesting a common mechanism across age. Notably, we did not find any difference in the abundance of GPR183 ligands in this JSLE cohort. A previous study utilizing serum samples from adult individuals with SLE has shown that there is an increase in 7α,25-dihydroxycholesterol as well as in 25-hydroxycholesterol, 27-hydroxycholesterol and other immune-activatory lipids.³³ Besides differences in age between these cohorts, the adult patients included in this previous study had a much higher disease activity than our cohort and were on different medications.³³ Considering the strong impact of medications such as steroids and inflammation on lipid metabolism pathways in SLE,^34,35 it is perhaps not surprising that oxysterol profiles will vary in studies with different clinical demographics.

In conclusion, this study sheds further light on the complexity of the physiological processes that control sex differences in the risk of developing immune-mediated disorders. We identify that sexual dimorphism in early GPR183-dependent B cell activation leads to altered B cell chemotaxis, potentially influencing the spatial location of B cell responses within secondary lymphoid tissues and plasmablast differentiation during the initiation phase of a female-biased autoimmune disease. As well as emphasizing the need for sex-disaggregated immunological studies, these data add to our understanding of the diverse impact that GPR183 and its oxysterol ligands have on immune cell activation over the different stages of an immune response.

Limitations of the study

Our study is not without its limitations. Our findings in human subjects comparing GPR183 B cell phenotype between healthy controls and people with JSLE utilized peripheral blood. This limits our ability to properly interrogate GPR183-oxysterol interactions as they occur in secondary lymphoid tissues. Due to the nature of our cohort and the rarity of treatment-naïve JSLE samples, we could only assess GPR183 in individuals with established disease, and usually many years post symptom onset or diagnosis. To address this limitation and to understand the temporal regulation of GPR183 following the induction of an autoimmune response, we used an inducible model of lupus nephritis, which is the parent into F1 model of chronic graft versus host disease. While these animals do develop an autoantibody-based nephritis dependent upon ANA production, it does not recapitulate all the features of human SLE. In addition, our data demonstrated stark differences in the percentages of B cells that express GPR183 in humans vs. mice (20–50% in human B cells in the peripheral blood vs. 3–5% B cells in the mouse spleen), potentially indicative of distinct B cell subsets across species and anatomical compartments. However, it is important to note that this may also reflect differences in the function of GPR183 across the two species investigated here. Our use of this mouse model also impeeded our ability to carry out genetic knockout studies of GPR183, as GPR183 KO mice are not available on the DBA/2 line. This means that we cannot fully exclude the possibility of off-target effects of the pharmacological inhibitor. Future genetic studies would be informative to fully test the influence of GPR183 on autoimmune disease progression in other models.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Elizabeth C. Rosser (e.rosser@ucl.ac.uk).

Materials availability

This study did not generate new unique reagents.

Data and code availability

The bulk RNA sequencing dataset generated during this study is available at NCBI Sequence Read Archive: PRJNA1248172. All other data reported in this article will be shared by the lead contact upon request. This article does not report original code; however, any additional information required to reanalyse the data reported in this article is available from the lead contact upon request.

Acknowledgments

This work was supported by a Medical Research Foundation Fellowship (MRF-057-0001-RG-ROSS-C0797) and a Senior Research Fellowship from the Kennedy Trust for Rheumatology Research (KENN 21 22 09). ECR is also a recipient of a Lister Institute for Preventive Medicine Research Prize. PJ is supported by a FOREUM research career grant (094) awarded to ECR. BJ was supported by a Fight 4 Sight Arthritis UK PhD studentship awarded to Professor Lucy Wedderburn and ECR (U/24VA22). This work was also supported by an Arthritis UK Centre for Excellence Grant awarded to Professor Lucy Wedderburn (21593) at the Centre for Adolescent Rheumatology at UCL, UCLH and GOSH. CC is supported by the National Institute for Health Research University College London Hospitals NHS Foundation Trust, Biomedical Research Centre (BRC4-III-CC). HP was supported by an Arthritis UK PhD Studentship awarded to CC (22203). The graphical abstract was created with Biorender.com.

Author contributions

D.E.M., N.G., and P.J. performed experiments and analyzed data. H.P. and A.R. processed and cataloged human samples and obtained clinical data. B.J. performed experiments. K.K. provided expert oversight in oxysterol measurements and analysis. C.C. critically reviewed the manuscript and obtained funding. E.C.R. conceptualized the study, obtained funding, performed experiments, analyzed data, and wrote the manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
CD45 PerCP/Cy5.5, clone HI30	BD Biosciences	Cat# 564106; RRID: AB_2744405
CD19 BV785, clone HIB19	Biolegend	Cat# 302240; RRID: AB_2563442
CD19 BV510, clone HIB19	Biolegend	Cat# 302242; RRID: AB_2561668
CD3 PE-Cy7, clone UCHT1	BD Biosciences	Cat# 563423; RRID: AB_2738196
CD3 BUV395, clone UCHT1	BD Biosciences	Cat# 563546; RRID: AB_2744387
CD3 BUV805, clone UCHT1	BD Biosciences	Cat# 612895; RRID: AB_2870183
CD4 APC-H7, clone SK3	BD Biosciences	Cat# 641398; RRID:AB_1645732
CD4 BUV395, clone SK3	BD Biosciences	Cat# 563550; RRID:AB_2738273
CD8 BV711, clone RPA-T8	BD Biosciences	Cat# 563676; RRID: AB_2744463
CD8 FITC, clone RPA-T8	BD Biosciences	Cat# 561948; RRID: AB_11154582
CD14 Alexa Fluor 700, clone 63D3	Biolegend	Cat# 367114; RRID: AB_2566715
CD14 APC-Cy7, clone 63D3	Biolegend	Cat# 367108; RRID: AB_2566710
CD24 APC-Cy7, clone ML5	Biolegend	Cat# 311132; RRID: AB_2566346
CD38 BV605, clone HIT2	Biolegend	Cat# 303532; RRID: AB_2561527
CD27 BV711, clone O323	Biolegend	Cat# 302834; RRID: AB_11219201
GPR183 Alexa Fluor 647, clone EBI2	Biolegend	Cat# 368904; RRID: AB_2566384
CD11c BV421, clone 3.9	Biolegend	Cat# 301628; RRID: AB_11203895
CD11c APC-Cy7, clone 3.9	Biolegend	Cat# 337218; RRID: AB_10662746
CXCR5 BV785, clone J252D4	Biolegend	Cat# 356936; RRID: AB_2629527
IgG BUV737, clone G18-145	BD Biosciences	Cat# 612819; RRID:AB_2738986
CD19 BV785, clone 6D5	Biolegend	Cat# 115543; RRID: AB_11218994
CD138 BV711, clone 281-2	Biolegend	Cat# 142519; RRID: AB_2562571
CD95 PE-Cy7, clone SA367H8	Biolegend	Cat# 152618; RRID: AB_2910313
GL7 BV421, clone GL7	BD Biosciences	Cat# 562967; RRID: AB_2737922
GPR183 polyclonal antibody, unconjugated	Thermo Scientific	Cat# PA5100950; RRID: AB_2850437
Blimp-1 Alexa Fluor 647, clone 5E7	Biolegend	Cat# 150004; RRID: AB_2565617
IgD Alexa Fluor 488, clone 11-26c.2a	Biolegend	Cat# 405718; RRID: AB_10730618
CD3 Alexa Fluor 647, clone 145-2C11	Biolegend	Cat# 100322; RRID: AB_492861
CD138 PE, clone 281-2	Biolegend	Cat# 142504; RRID: AB_10916119
GL7 PE, clone GL7	BD Biosciences	Cat# 561530; RRID: AB_10715834
CD11c FITC, clone N418	Biolegend	Cat# 117306RRID: AB_313775
Chemicals, peptides, and recombinant proteins
NIBR 189	BioTechne/Tocris	Cat# 5203
Fulvestrant	Selleckham	Cat# S1191
Fatty acid free bovine serum albumin (BSA)	Sigma-Aldrich	Cat# A8806
Methylated bovine serum albumin (mBSA)	Sigma-Aldrich	Cat# A1009
Incomplete Freund’s adjuvant (IFA)	Sigma-Aldrich	Cat# F5506
DAPI	Sigma-Aldrich	Cat# MBD0015
RNase-Free DNase set	QIAGEN	Cat# 79254
7a,25-dihidroxycholesterol	Sigma-Aldrich	Cat# SML0541
Recombinant mouse CXCL13	Biolegend	Cat# 583904
OxysterolSPLASH	Avanti Polar Lipids	Cat# 330700W
Cholesterol oxidase from Streptomyces sp.	Sigma-Aldrich	Cat# 228250
Critical commercial assays
Picopure™ RNA isolation kit	ThermoFisher Scientific	Cat# KIT0204
iScript™ cDNA synthesis kit	Biorad	Cat# 1708891
iQ™ SYBR® green supermix	Biorad	Cat# 1708882
NEBNext® Single Cell/Low Input RNA Library Prep Kit	Illumina	Cat# E6420S
Murine IgG ELISA kit	ThermoFisher Scientific	Cat# 88-50470-88
Murine anti-nuclear antibody (IgG) ELISA kit	Cusabio	Cat# CSB-E12912m
Deposited data
JSLE and HC GPR183+CD27+ B cell bulk RNAseq	This paper	NCBI Sequence Read Archive: PRJNA1248172
Experimental models: Organisms/strains
Mouse, C57BL/6	Envigo (UK) or Charles River (UK)	N/A
Mouse, DBA/2	Envigo (UK) or Charles River (France/Netherlands)	N/A
Oligonucleotides
Human GPR183 F	Thermo	5′-GAATCGGAGATGCCTTGTGT-3′
Human GPR183 R	–	5′- GCCTCCTGCTTTGACATAGG-3′
Human 18S F	–	5′-CTTATTAACGTTAGGGGCTA-3′
Human 18S R	–	5′-GCTAACCTACCAAATCACTC-3′
Software and algorithms
NDP.view2 Image viewing software	Hamamatsu	https://www.hamamatsu.com/eu/en/product/life-science-and-medical-systems/digital-slide-scanner/U12388-01.html
FlowJo 10 v9 or v10	FlowJo	https://www.flowjo.com/flowjo/download
BCL Convert Software v3.7.5	Illumina	https://emea.support.illumina.com/sequencing/sequencing_software/bcl-convert/downloads.html
R Studio version 4.4	R	https://cran.r-project.org/bin/macosx/
RNA-STAR	Encode	https://www.encodeproject.org/software/star/
Fiji (ImageJ)	ImageJ	https://imagej.net/software/fiji/downloads
ZEISS ZEN 3.9	Zeiss	https://www.zeiss.com/microscopy/en/products/software/zeiss-zen-lite.html
GraphPad Prism version 9 or 10	GraphPad	https://www.graphpad.com
Adobe Illustrator	Adobe	https://www.adobe.com/uk/products/illustrator
BioRender	BioRender	https://www.biorender.com
Other
RPMI-1640 media	Sigma-Aldrich	Cat# R8758
Foetal bovine serum	Cytiva	Cat# SV30160.03
Penicillin/Streptomycin	Sigma-Aldrich	Cat# P0781
Ethylenediaminetetraacetic acid	ThermoFisher Scientific	Cat# AM9260G
Red blood cell lysis buffer	Sigma-Aldrich	Cat# R7757
Hanks balanced salt solution	Gibco	Cat# 14175-053
Neutral buffered formalin	Scytek	Cat# FRN999
LIVE/DEAD Fixable Blue Dead Cell Stain	Invitrogen/Thermo	Cat# L34961
Lightning Link Conjugation Kit - Alexa Fluor 700	Abcam	Cat# Ab269824
Oligo dT beads	New England Biolabs	Cat# S1419S
Normal Goat Serum	Sigma-Aldrich	Cat# G9023
Brilliant stain buffer	BD Biosciences	Cat# 563794
M. tuberculosis H37 Ra, desiccated	BD Biosciences	Cat# 231141
O.C.T compound	Tissue-Tek	Cat# 23-730-571
Cytofix™ fixation buffer	BD Biosciences	Cat# 554714
Prolong Gold Antifade Mountant	ThermoFisher Scientific	Cat# P36961
Girard’s Reagent P	Santa Cruz Biotechnology	Cat# sc-295008

Experimental model and study participant details

Human samples

Peripheral blood mononuclear cells and serum were obtained from patients with juvenile-onset systemic lupus erythematosus (JSLE) attending clinics at University College London Hospital. Healthy controls (HCs) were recruited at UCL. Demographics for all samples are in Table S1. All samples were recruited under North Harrow ethics committee approval REC 11/0101. All patients were recruited in accordance with the Declaration of Helsinki. Patients with monogenic forms of SLE were excluded. JSLE patients were stratified based on the validated SLE Disease Activity Index (SLEDAI-2000) score, as active (SLEDAI >4) vs. inactive (SLEDAI ≤4, with no activity in any of the clinical domains)³⁶ or nephritis history based on histological diagnostic criteria from biopsies.³⁷

Mice

C57BL/6 were purchased from Envigo and Charles River (UK) and DBA/2 mice were purchased Charles River (France or Netherlands) or Envigo (UK). DBF1 mice were bred in house by crossing C57BL/6 and DBA/2 mice. Mice were assigned to experimental groups at random and, where possible, mixed amongst cages. Mice were housed in individually ventilated cages at 20-24°C, 45%–64% humidity, and at a 12-h light/dark cycle in specific opportunist- and pathogen-free (SOPF) conditions (health screening (Full-FELASA profile) was performed annually). Experimental mice were fed Harlan Teklad pellets 2018 (18% protein) and breeding mice were fed Harlan Teklad pellets 2010 (19% protein), with food and water available ad libitum and were used at 6-10 weeks of age. Male and female mice were used as specified in the figure legends. All animals were housed in a UCL Biological Services Unit and experiments were approved by the Animal Welfare and Ethical Review Body of University College London and authorised by the United Kingdom Home Office.

Method details

Parent into F1 mouse model of lupus nephritis

The parent into F1 mouse model of lupus nephritis was induced according to published protocols.³⁸ Spleens from naïve DBA/2 mice were collected post-mortem in complete RPMI-1640 (cRPMI) media with L-glutamine and NaHCO3 (Sigma-Aldrich), with 1:10 foetal bovine serum (FBS) (Biosera) and 1:100 Penicillin (10.000 units/ml)/Streptomycin (10 mg/ml) (Sigma-Aldrich) and single cell suspensions were obtained by dissociating tissues through 70 μm cell strainers (BD Biosciences). Erythrocytes were lysed using Red Blood Cell (RBC) Lysis buffer (Sigma-Aldrich) and disease was induced by injecting 70 million sex-matched splenocytes per DBF1 mouse in 500 μl of Hanks saline solution (Sigma-Aldrich) at 8-10 weeks of age. For NIBR189 treatment, NIBR189 (Avanti Polar Lipids) was reconstituted in DMSO, diluted in PBS (Gibco) and injected intraperitoneally into mice at a working concentration of 0.1 mg/kg (e.g. a 20 g mouse would receive 100 μl diluted in 10% DMSO).²⁴ Unless otherwise stated (i.e. beginning treatment 5- or 9-weeks post-nephritis induction) for the first two weeks post-disease initiation, mice were injected for 5 days, followed by a 2-day rest period, and then kept on a maintenance dose of twice per week for the rest of the experiment. Control mice received vehicle alone at the same DMSO concentration. For fulvestrant treatment (Selleckham), fulvestrant was reconstituted in DMSO, diluted in oil and injected subcutaneously into mice at a working concentration of 60 mg/kg (e.g. a 20 g mouse would receive 200 μl diluted in 10% DMSO).³⁹ Mice were injected the day before lupus induction, and then twice per week for the duration of the experiment. Control mice received vehicle alone at the same DMSO concentration. For scoring of urine proteinuria, from 2 weeks post-disease induction mice urine proteinuria was analysed using urinalysis dipsticks (Multistix 10SG from Siemens) and scored according to colour change or as follows: negative/trace of proteinuria (score 0), 30 mg/ml of proteinuria (+ or 1), 100 mg/ml of proteinuria (++ or 2), 300 mg/ml of proteinuria (+++ or 3), >2000 mg/ml of proteinuria (++++ or 4). For assessment of nephritis severity by histology, mouse kidneys were dissected to remove excess fat tissue and fixed in 10% neutral buffered formalin solution (FRN999, ScyTek) for 24 hours. Kidneys were moved to PBS, paraffin embedded and sectioned. Slides were stained with haematoxylin and eosin (H&E) staining, captured on a Hamamatsu Nanozoomer for image digitalisation, analysed using NDP.view2. Kidney sections were scored as follows: 1 - mildly irregular glomeruli, normal aspect of tubules and perivascular area; 2 - moderately irregular and somewhat enlarged glomeruli, normal aspect of tubules and perivascular area; 3 - irregular and enlarged glomeruli, some protein deposits (casts) in tubules, some enlarged tubules, perivascular lymphocyte infiltrates; 4 - very irregular and enlarged glomeruli, many protein casts, enlarged tubules, perivascular and -glomerular lymphocyte infiltrates.

Flow cytometry and cell sorting

For human studies: multi-colour flow cytometry was performed on isolated peripheral blood mononuclear cells from JSLE patients and healthy controls. Cells were stained ex vivo first with LIVE/DEAD Fixable Blue Dead Cell Stain (Invitrogen) for 15 mins at room temperature and then with surface anti-human antibodies for flow cytometry for 30 min at 4°C as follows: CD45 PerCP/Cy5.5 (HI30, BD), CD19 BV785 or BV510 (HIB19, Biolegend), CD3 PE-Cy7 (UCHT1, Biolegend) or BUV395 (UCHT1, BD), CD4 APC-H7 (SK3, BD), CD8 BV711 or FITC (RPA-T8, BD), CD14 AF700 (63D3, Biolegend), CD24 APC-Cy7 (ML5, Biolegend), CD38 BV605 (HIT2, Biolegend), CD27 BV711 (O323, Biolegend), GPR183 AF647 (EBI2, Biolegend), CD11c BV421 or APC-CY7 (3.9, Biolegend), CXCR5 BV785 (J252D4, Biolegend), and IgG BUV737 (G18-145, BD). For cell sorting, peripheral blood mononuclear cells from JSLE patients and healthy controls were stained with zombie NIR viability die and then with surface anti-human antibodies for CD3 BUV805 (UCLHT1, BD), CD4 BUV395 (SK3, BD), CD8 BV711 (RPA-T8, DB), CD19 BV785 (HIB19, Biolegend) and CD14 APC-CY7 (63D3, Biolegend). Isolated B cells, monocytes, CD4 T cells and CD8 T cells were sorted using BD FACSDiscover S8 into RNAse free PBS and prepared for RNA extraction immediately as detailed below.

For mouse studies: single cell suspensions of splenocytes were stained with LIVE/DEAD Fixable Blue Dead Cell Stain (Invitrogen) for 15 mins at room temperature and then with surface anti-mouse antibodies for flow cytometry: CD19 BV785 (6D5, Biolegend), CD138 BV711 (281-2, Biolegend), CD95 PE-Cy7 (Jo2, Biolegend), GL7 BV421 (GL7, BD), CD11c FITC (N418, Biolegend). For GPR183 surface staining, polyclonal anti-mouse GPR183 antibody (PA5100950, Thermo Scientific) was conjugated in house to Alexa Fluor 700 Lightning Link conjugation kit (ab269824, abcam). For intracellular staining, cells were fixed with FoxP3 Fix/Perm Buffer Set (Biolegend) following surface stains for 20 mins at 4°C and then stained with anti-mouse Blimp-1 AF647 (5E7, Biolegend) for 40 mins at 4°C. Data were acquired on LSRII flow cytometer (BD Pharmigen) or LSRFortessa X-20 and flow cytometric data were analysed using FlowJo 10.

RNA extraction, cDNA transcription and RT-PCR

RNA from isolated human B cells, CD4/CD8 T cells and monocytes was extracted using Arcturus Picopure RNA isolation kit (ThermoFisher Scientific) and RNA was reverse transcribed using an iScript cDNA synthesis kit (Bio-Rad), according to manufacturer’s instructions. qPCR was carried out on the cDNA samples using iQ SYBR® Green Supermix (Bio-Rad), according to manufacturer’s instructions. Primers were used at a final concentration of 0.5 μM. Human GPR183 and housekeeping 18S primers were purchased from Thermo: GPR183 Forward (5′-GAATCGGAGATGCCTTGTGT-3′); GPR183 Reverse (5′- GCCTCCTGCTTTGACATAGG-3′); 18S Forward (5′-CTTATTAACGTTAGGGGCTA-3′); 18S Reverse (5′-GCTAACCTACCAAATCACTC-3′). qPCR data were calculated as the ratio of gene to 18S expression by the relative quantification method.

Bulk RNA sequencing

Live CD19⁺GPR183⁺CD27⁺ B cells were isolated from PBMCs on a FACSAria sorter (BD Biosciences) and RNA was extracted using PicoPure™ RNA Isolation Kit (Thermo Fisher) according to manufacturer’s instructions. RNA quality control and sequencing were performed by UCL Genomics (UCLG). RNA integrity was confirmed using the Agilent 4200 Tapestation (Standard Total RNA assay). Samples were processed using the NEBNext® Single Cell/Low Input RNA Library Prep Kit for Illumina (JSLE and HC samples). The cDNA library was prepared from mRNA that was isolated from total RNA by use of Oligo dT beads. Full length xGen adaptors (IDT), containing two unique 8bp sample-specific indexes, a unique molecular identifier and a T overhang, were ligated to the A-tailed cDNA and enriched with limited cycle PCR (16 PCR cycles).

Samples were sequenced on a NovaSeq (Illumina) at 300pM using a 101bp paired read run with corresponding 8bp dual sample index and 8bp unique molecular index reads. Run data were demultiplexed and converted to fastq files using BCL Convert Software v3.7.5 (Illumina). Transcript abundance was estimated using a custom Galaxy pipeline developed by UCLG to produce a raw count table. Data were adapted and quality trimmed using fastp before being aligned to the human genome (UCSC hg38) with RNA-STAR. Aligned reads were then UMI deduplicated using JE-Suite. FeatureCounts was used to estimate reads per transcript by counting uniquely mapped reads that fall within coding or UTR regions of the genome. All sequence and annotation data were obtained from the Illumina iGenomes repository. Normalised counts were used to create the IFN score bar plot in R using R Studio version 2022.12.0+353.

Measurement of oxysterol in peripheral blood serum - Extraction of non-esterified sterols

The extraction of non-esterified sterols was adapted from Griffiths and colleagues.⁴⁰ Briefly, 100 μL of serum was added to 1.050 mL of ethanol containing 100 μL of OxysterolSPLASH (Avanti Polar Lipids, USA) in a microcentrifuge tube under sonication in an ultrasonic bath. The solution was diluted with 350 μL of water to give a 70% alcohol solution. This was sonicated for a further 5 min, then centrifuged at 17,000 x g at 4 °C for 30 min. Oxysterols were separated from cholesterol and sterols of similar lipophilicity by solid phase extraction (SPE), using a “Sep-Pak tC₁₈” column (200 mg, Waters Inc, Elstree, Herts, UK). The column (SPE-1) was washed with absolute ethanol (4 mL), then conditioned with 70% ethanol (6 mL). The sterol extract from above in 70% alcohol (1.5 mL) was applied to the cartridge and allowed to flow at a rate of 0.20 mL/min. The column flow-through was collected and combined with subsequent a 5.5-mL column wash of 70% ethanol (total 7 mL). The column was washed further with 70% ethanol (4 mL), followed by two portions of 2 mL. All eluents were one SPE-1 fraction and dried under reduced pressure using a rotor evaporator.

Measurement of oxysterol in peripheral blood serum - Enzyme-assisted derivatisation of sterol analysis (EADSA)

The dried sample was reconstituted in propan-2-ol (100 μL) and thoroughly vortexed and 1 mL of 50 mM phosphate buffer (KH₂PO₄, pH 7, 1.000 mL) containing cholesterol oxidase from Streptomyces sp. (3 μL, 2 μg/μL in water, 44 mU/μg protein, Sigma-Aldrich, UK) was added and the mixture incubated at 37 °C for 1 h, after which the reaction was quenched with methanol (2 mL). Glacial acetic acid (150 μL) was then added and the solution thoroughly vortexed. GP reagent (190 mg, Santa Cruz Biotechnology) was added to this solution which was thoroughly vortexed and incubated at room temperature overnight in the dark. Following overnight incubation, the oxidised/GP-derivatised sterols was then separated from excess GP reagent utilising a second SPE step a “recycling” method. The C₁₈ SPE column was conditioned with 10 mL of methanol, and 10 mL of 10% methanol in water, and finally with 5 mL of 70% methanol in water prior to sample application. The O/GP reaction mixture (∼3 mL in 70% methanol) was directly applied on the C₁₈ SPE column, followed by 1 mL of 70% methanol (this is a wash of the reaction vessel), and 1 mL of 35% methanol. The effluent was collected into a beaker (a sample tube). The combined effluent (now 5 mL) was diluted with 4 mL of water. The resulting mixture (now 9 mL in 35% methanol) was again applied to the C₁₈ SPE column, followed by a wash with 1 mL of 17% methanol. The effluent was collected into the sample tube. To the combined effluent, 9 mL of water was added to give 19 mL of about 17.5% methanol. This was again applied to the C₁₈ SPE column followed by a wash with 6 mL of 10% methanol in water to remove excess of GP hydrazine. At this point, all the oxidised/GP-derivatised sterols had been extracted on the C₁₈ column. The O/GP-derivatised oxysterols were then eluted with three portions of 1 mL of methanol (SPE2-Fr1, -Fr2 and -Fr3), followed by 1 mL of ethanol (SPE-Fr4). Reference standards 7α,25-dihydroxycholesterol (cholest- 5-ene-3β,7α,25-triol) and 7α,27-dihydroxycholesterol (cholest-5-ene-3β,7α,27-triol) (Avanti Polar Lipids) were processed in the same way. Prior to LC-MS analyses, an aliquot of 70 μL of each SPE-Fr2, 3, 4 fraction was combined and dried nearly to dryness in a SpeedVac, sterols were re-dissolved to 100 μL with methanol (SPE-Fr_combined). An aliquot 70 μL of SPE-Fr1 and 70 μL of SPE-Fr_combined were combined, and the resulting sample was vortexed for 2 min and analysed three times by cap-LC-MSⁿ. All fractions were stored at -20⁰C.

Measurement of oxysterol in peripheral blood serum - LC-MS analyses

The Vanquisher LC system connected to the Orbitrap Q Exactive mass spectrometer with a HESI probe (Thermo Fisher Scientific, UK) was utilised for LC-MS analysis. Twenty μL of a sample was injected on a Hypersil Gold C18 (150 x 2.1 mm, 1.9 μm column (Thermo Fisher Scientific, UK). Mobile phase A consisted of 33.3% methanol, 16.7% acetonitrile, containing 0.1% formic acid. Mobile phase B consisted of 63.3% methanol, 31.7% acetonitrile containing 0.1% formic acid. The gradient was as follows: 20% B was for 1 min and after a linear increase to 80% B over 8 min, and left at 80% B for 4 min, followed by the linear increase to 99% B over 1 min which followed by change to 20% B in 0.1 min and left at 20% B for another 2.9 min. The total LC-MS analysis time was 21 min, and the flow rate was 0.2 mL/min. The HESI ion source parameters were: spray voltage 3500, capillary temperature 320°C, sheath gas 25, auxiliary gas 10 and probe heater 310°C, in source fragmentation set at 20 eV, with positive ionisation mode. On the Q Exactive instrument six scan events we performed: one high resolution (70,000 at m/z 400) with the mass range m/z 100 to 800 Da and five MS² scan events at resolution 17,500 at m/z 400, precursor ion isolation width 0.4 m/z and MS² loop count 3, collision energy was 50%. The instrument was externally calibrated using caffeine, MRFA (MET-ARG-PHE-ALA) and Ultramark 1621. Quantification was performed by stable isotope dilution, and the identification was based on MS² spectra and its retention on the reverse phase column as compared to authentic standards.

Serum anti-nuclear antibodies (ANA) and total IgG ELISA’s

Serum was collected from naïve mice and mice with lupus nephritis post-mortem and frozen at -80°C. Anti-nuclear antibody (IgG) levels in the serum (1:200 dilution) were measured by ELISA according to manufacturer’s protocol (CSB-E12912m, Cusabio). Total IgG in the serum was measured by ELISA according to manufacturer’s protocol (88-50470-88, ThermoFisher Scientific).

Cell migration assay

To perform a lymphocyte in vitro transmigration assay according to published protocols,⁴¹ splenocytes isolated from lupus female and male mice 24 hours post-nephritis induction were resuspended at 5x10⁶/ml in migration media (RPMI+50 U penicillin/streptomycin+10 mM HEPES+0.5% fatty acid-free BSA) and left to rest for 30 min at 37°C. From each mouse 0.5x10⁶ cells/condition were then plated on transwells (5 μm pores, Corning) and placed in 96-well plates with or without 7α,25-diHC (1 μM, Sigma) or with 1 μg of CXCL13 (Biolegend) as positive control. Cells were allowed to migrate for 3hr at 37°C. Cells that have migrated to the bottom of the transwell were washed and resuspended in PBS for flow cytometry analysis as described above.

Immunofluorescent imagining of mouse spleens

Spleens from naïve mice and mice with lupus nephritis were collected post-mortem. Spleens were fixed in 20% Cytofix™ fixation buffer (BD Biosciences) in PBS for 4-5 hours and then soaked in 30% sucrose in PBS (w/v) overnight, for cryopreservation. Spleens were then embedded in O.C.T compound (Tissue-Tek), and snap-frozen in dry ice-cold isopentane (Sigma). Frozen spleens were cut into 6μm sections (HM525 NX cryostat, Thermo Scientific) and adhered to Superfrost Plus slides (VWR). The cryo-sections were rehydrated in PBS, then blocked with 5% normal goat serum (G9023-5ML, Sigma) with 0.3% Triton X-100 (Sigma) for 20 min at room temperature (RT). The tissues were incubated with primary antibodies overnight at 4°C: anti-mouse IgD AF488 (11-26c.2a, Biolegend); anti-mouse CD3 AF647 (145-2C11, Biolegend); anti-mouse CD138 PE (281-2, Biolegend), anti-mouse GL7 PE (GL7, BD). DAPI (MBD0015, Sigma) was added the following day after 2X PBS washes and incubated for 10 min at RT. The slides were finally mounted in Prolong Gold Antifade Mountant (Thermo Fisher), imaged on a Zeiss Axioscan fluorescence microscope using Zen lite (Blue edition) software and analysed using Fiji (ImageJ) and ZEISS ZEN 3.9. The areas of GL7⁺ GC B cells and CD138⁺ plasmablasts were measured with the ZEISS ZEN 3.9 microscopy software using the spline curve tool, and statistical analysis was carried out using measurements from a minimum of 24 and 39 IgD⁺ B cell follicles for GL7⁺ GC B cells and CD138⁺ plasmablasts, respectively.

Quantification and statistical analysis

For animal experiments, scoring of proteinuria and kidney histology was performed blinded and treatment groups were mixed randomly amongst cages of age- and sex-matched mice. All data are expressed as mean ± SEM and were analysed using GraphPad Prism version 9 or 10 (La Jolla, CA, USA). Plots show datapoints pooled across multiple experiments. When two groups were compared, paired or unpaired t test (parametric) or Mann-Whitney U test (non-parametric) were used to determine significance of difference between groups. When more than two groups were compared, one-way ANOVA (parametric) or Kruskal-Wallis test (non-parametric) were used to determine significance of difference between groups. For clinical proteinuria, two-way ANOVA was used to compare differences between groups and the column factor is shown as a p value on graphs. All analyses met assumption of statistical tests. For all figures, p values less than 0.1 are listed on figures as numeric values. Figures were made in Adobe Illustrator and graphical abstract made using Biorender.

Published: February 18, 2026