Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Sep 29.
Published in final edited form as: Ann Rheum Dis. 2026 Apr 23;85(8):1582–1595. doi: 10.1016/j.ard.2026.03.002

A pathogenic gut lipoglycan drives systemic thromboinflammation in lupus nephritis

Abhimanyu Amarnani 1, Cristobal F Rivera 2, Macintosh Cornwell 3, Tyler Weinstein 1, Zakia Azad 1, Susan RS Gottesman 2, Cynthia Loomis 4, Andy Lee 1, Nimat Ullah 1, Joshua Prasad 1, Mingyang Yi 1, Laura Cooney 5, Betsy J Barnes 6, Nicolas Gisch 7, Kelly V Ruggles 3, Bhama Ramkhelawon 2, Gregg J Silverman 1,*
PMCID: PMC13619169  NIHMSID: NIHMS2171972  PMID: 42031645

Abstract

Objectives:

The gut microbiome plays a crucial role in regulating systemic immunity and has been implicated in several chronic inflammatory diseases. Intestinal expansions of Ruminococcus gnavus (RG), a dominant gut commensal, correlate with disease flares in lupus nephritis (LN), but the underlying mechanism remains unknown.

Methods:

In a Pilot cohort of patients with biopsy-proven LN, subsetted by gut microbiota community, immune status was characterised using bulk-blood RNA sequencing libraries, serum levels of representative host proteins, and levels of immunoglobulin (Ig)G antibodies to the novel lipoglycan (LG) produced by pathogenic RG strains. A Validation LN cohort was evaluated for blood transcriptomic profiles and levels of anti-LG antibodies. In murine models, mechanistic hypotheses were tested after RG gut colonisation or after intraperitoneal injection with an LG preparation, with outcomes determined by transcriptomic analyses, platelet functional readouts, and tissue histology.

Results:

In a Pilot cohort of patients with LN, RG gut expansions were associated with high-level platelet, neutrophil, and monocyte activation. Serum levels of platelet factor 4 and release of neutrophil extracellular traps (NETs) were significantly higher in patients with high serum IgG antibody against the novel RG-specific LG, a marker of in vivo immune exposure. An LN Validation cohort confirmed these correlates and showed that anti-LG antibodies serve as a surrogate for thromboinflammatory profile in this LN-associated endotype. In mice, gut colonisation with LG-producing RG strains or a single LG injection caused megakaryocytosis and platelet activation; RG colonisation with LG-producing strains induced tubulointerstitial injury with NETosis. In vivo responses to LG toxin were Toll-like receptor 2-dependent.

Conclusions:

Gut expansions of the RG pathobiont may contribute to autoimmune pathogenesis through the LG toxin and cause LN flares through thromboinflammatory mechanisms in this previously unrecognised LN endotype.

INTRODUCTION

The gut microbiome is a key regulator of systemic health and is implicated as a significant driver in chronic autoimmune and inflammatory diseases [1,2]. We investigated the importance of this connection in systemic lupus erythematosus (SLE), an autoimmune disease characterised by considerable clinical heterogeneity in which recurrent flares often cause severe multiorgan damage. A fundamental quandary in SLE is how clinical and immunologic diversity arises and whether there are distinct patient subsets that each reflect unique underlying pathogenic mechanisms. This question may be fundamental to understanding the causes of lupus nephritis (LN), which affects half of SLE patients and is a major contributor to morbidity and mortality.

Despite recent therapeutic advancements, within 10 years of diagnosis, 1 in 5 patients progresses to end-stage renal disease [3]. In genetically predisposed individuals, environmental factors, including the gut microbiome, are suspected to play a role in the development of SLE and the subsequent triggering of disease flares; however, their contributions remain poorly understood. We hypothesise that a shift in the balance of intestinal commensals, leading to the release of microbial factors through a compromised gut barrier, is a key driver of disease flares [4,5]. This hypothesis underscores the significance of a comprehensive understanding of the interplay between disease predisposition, shifts in the gut microbiome, and the immunopathogenesis of SLE.

The gut microbiota plays a crucial role in immune development and homeostasis. Dysbiosis, or an imbalance in the gut microbiota, can increase gut permeability, contributing to the development of inflammatory and autoimmune diseases. In female patients with SLE, renal involvement is associated with expansions of specific strains of the pathobiont species that was recently renamed Mediterraneibacter gnavus [6]. Earlier work, to which we refer here, used the species name Ruminococcus gnavus (RG) [7–9]. RG, a member of the Lachnospiraceae family of obligate anaerobes, is a key species in the human gut microbiome [10,11]. Strains of RG generally first colonise humans during infancy [11]. In adults, several chronic inflammatory diseases have been associated with RG gut expansions [4,10]. In addition, substantial genomic diversity has been documented among RG strains [9,12], suggesting that not all strains may have pathogenic potential, and the pathways triggered by RG colonisation remain undefined.

In cross-sectional studies of cohorts of patients with SLE from diverse ethnic and racial backgrounds across 3 continents [7,13–15], RG gut expansions directly correlated with higher disease activity. Longitudinal studies have shown that, in individual patients, RG blooms are temporally associated with LN disease flares [9]. RG strains isolated from patients with active LN produce a novel form of lipoglycan (LG) [9], which has been characterised by mass spectrometry and nuclear magnetic resonance. This LG is a complex glycan with a unique structure featuring prominent mannoses and lipid anchors that attach it to the bacterial cell membrane. LGs produced by different RG strains share a conserved structure recognised by post-immunisation monoclonal antibodies [9] and also by spontaneously arising serum immunoglobulin (Ig)G2 anti-LG antibodies, which are found at high levels in approximately one-third or more of patients with active LN [7,9]. Evidence suggests a pathogenic role: murine colonisation RG strains, which produce the LG toxin and were isolated from lupus patients, translocate out of the gut and modulate the systemic immune response. LG-producing strains also accelerate the progression of lupus-like autoimmune disease, but the mechanisms underlying this effect remain undefined [16].

In this study, we investigated the mechanism(s) by which intestinal RG may directly influence autoimmune pathogenesis in LN patients. We specifically tested the hypothesis that gut colonisation with certain RG strains that produce the novel LG drives disease progression. Using patient cohorts with LN and relevant murine models, we aimed to elucidate the underlying pathogenic mechanisms. Our studies identified a significant subset of LN patients characterised by immune responses to LG. We delineate herein the role of this LG in activating platelets and neutrophils, which contribute to inflammatory renal injury.

RESULTS

LN patients in a Pilot cohort can be dichotomised based on the presence of RG intestinal blooms

In recent studies, we reported a significant correlation between high lupus disease activity and expansions of the gut commensal anaerobe, RG [7,9,14]. A longitudinal study of a cohort of female patients with SLE, as defined by the European League Against Rheumatism/American College of Rheumatology 2019 criteria [17], here termed the Pilot cohort, included 8 patients without evidence of renal disease and 8 who fulfilled the clinical criteria for LN (Fig 1A) [7,9] at the time of high disease activity. These LN patients all had elevated serum anti-double-stranded DNA (dsDNA) autoantibodies and hypocomplementemia [9], with renal biopsies documenting International Society of Nephrology/Renal Pathology Society histopathologic class III or IV proliferative glomerulonephritis [9] (Supplementary Table S1). All patients in our Pilot cohort had normal platelet counts and lacked antiphospholipid autoantibodies. Four LN patients had RG expansions, which we have termed ‘blooms’ when they represent >1% of the community [9]. This subset of LN patients with RG blooms also developed robust antibody responses against a structurally unique, membrane-associated LG produced by RG strains isolated from patients with active LN [7,9] (Fig 1B; Supplementary Data S1). Faecal RG abundance significantly correlated with the serum level of anti-LG IgG antibodies [7]. Within this Pilot cohort, LN patients with RG intestinal blooms displayed among the highest levels of serum IgG antibodies to this pathobiont species-associated LG, whereas the remaining LN patients showed neither expansions of RG nor identifiable antimicrobial antibody elevations compared to healthy subjects [7,9] (Fig 1B). LN patients with RG blooms also had evidence of increased systemic inflammation and immune activation, with significantly higher levels of the host acute phase reactant, lipopolysaccharide-binding protein (Supplementary Data S2), indicating increased intestinal permeability, microbial translocation, and a strong host innate immune response to circulating microbial components, including lipopolysaccharide [18].

Figure 1.

Figure 1.

Open in a new tab

Whole-blood RNA sequence analysis of patients with lupus nephritis (LN) demonstrates increased platelet activation in patients with Ruminococcus gnavus (RG) intestinal blooms. (A) In the Pilot cohort, patients with LN (class III or IV based on renal biopsy) were dichotomised by serum levels of RG anti-lipoglycan (LG) immunoglobulin (Ig)G for RNA sequencing. (B) Antibody levels reactive with the RG LG in the sera of healthy controls (n = 20), nonrenal (systemic lupus erythematosus [SLE]) patients (n = 20), LN patients without gut blooms (n = 8), and LN patients with RG gut blooms (n = 4) in the Pilot New York University (NYU) cohort [7,9]. (C) Principal component analysis (PCA) of gene expression in whole-blood RNA sequencing libraries with top 25% variance between control, SLE without LN (SLE nonrenal), LN without RG bloom, and LN with RG gut bloom. In the prior study [9], a gut RG bloom was defined as >1% RG abundance (LN with RG bloom), which was ~10-fold higher than the mean abundance in healthy participants. PC1 40.9%, PC2 19.1%, and PC3 7.8%. (D) Scatter plot of Log2-fold change in gene expression comparing LN patients with RG blooms (LN with RG bloom) vs controls to LN patients without bloom vs healthy controls. Two hundred seventeen genes are significantly differentially expressed between the LN with RG bloom and the LN without bloom (adjusted P < .05), with 202 upregulated and 15 downregulated. Colour scheme adjusted P value of LN with RG bloom vs LN without bloom. (E) Gene set enrichment analysis results of ranked genes based on LN with RG bloom vs LN without bloom, showing significantly overrepresented pathways from genes increased in LN with RG bloom (right) or instead, increased in LN without bloom (left). The absolute value of the normalised enrichment score (NES) is shown. NES and Padj values are listed in Supplementary Table S2. (F) LN with RG bloom is devoid of B-cell activation and the classical complement pathway; instead, these pathways are associated with LN without bloom. Gene expression of LN with RG bloom vs LN without bloom, with genes related to platelet activation (right) and B-cell activation (left), noted. (G) Heatmap of gene expression involved in platelet activation in nonrenal SLE patients, LN patients without bloom, and LN patients with RG bloom. (H) Singscore modules assessing platelet activation pathway in LN patients with RG bloom vs those without RG bloom. (I) Expression of genes related to neutrophil degranulation (right) and interferon-stimulated genes (ISGs) (left), with (J) heatmap of gene expression in nonrenal SLE, LN without bloom, and LN with RG bloom patients, and (K) Singscore module of neutrophil degranulation genes in LN patients with RG bloom and without gut bloom; unpaired nonparametric t-test.

Differential blood transcript expression in patients with LN and RG blooms

To provide an unbiased global snapshot of a patient’s immune system, RNA sequencing (RNA-seq) libraries were prepared from whole blood and then analysed using unsupervised principal component analysis (PCA). As expected [19,20], in the Pilot cohort, we found a clear separation between the SLE patients and the control subjects (Fig 1C). As further anticipated based on earlier reports [21–23], there were features shared between patients with nonrenal and renal disease (Fig 1C; Supplementary Data S3). Surprisingly, LN patients with RG gut blooms formed a distinct cluster that was separate from those with LN and no blooms (Fig 1C). Although there was no significant difference in the steroid dose received (Supplementary Table S1), compared with the other LN patients, in the LN patients with RG blooms, 202 genes were upregulated and 15 were underexpressed (Fig 1D), which represented a tractable gene set amenable to meaningful interrogation.

In female LN patients with RG blooms in the Pilot cohort, gene set enrichment analysis (GSEA) [24] revealed that 5 of the top 20 most overexpressed pathways were associated with platelet activation and neutrophil degranulation compared with patients without blooms (Fig 1E; Supplementary Tables S2 and S3). Unlike the heterogeneous patterns found in SLE patients without prominent renal involvement, those with LN and gut RG blooms exhibited distinctly enhanced expression of platelet activation [25] (Fig 1F). These included transcripts for α-granule-related factors, such as platelet factor 4 (PF4), a chemokine that attracts immune cells and modulates platelet–immune cell interactions [26], and protein S (Pros1), which is crucial for clotting, apoptotic cell clearance, and immune modulation [27]. The LN patients with RG blooms also exhibited upregulation of transcripts for key platelet activation factors, including alpha-synuclein and platelet-derived growth factor A (Fig 1G,H). As expected [21,23], based on module analysis, some nonrenal patients showed elevated transcripts associated with platelet and neutrophil activation (Supplementary Data S3). However, anti-LG antibody levels were not increased in these patients (Supplementary Data S1A).

Patients with RG blooms also showed enhanced representation of pathways involved in neutrophil migration, chemotaxis, and degranulation, indicating broader immune activation. Furthermore, LN patients with RG blooms also showed transcript overexpression of specific genes, including LGALS3 and PRDX6 (Fig 1J, K), which are involved in neutrophil degranulation. In many cases, overexpression of genes involved in platelet activation correlated with expression of genes reflecting neutrophil degranulation (Supplementary Data S4A,B). However, these patients did not exhibit significantly elevated expression of interferon-stimulated genes (ISGs) (Supplementary Data S4A,B). Furthermore, LN patients without blooms instead had an overrepresentation of genes related to B-cell activation, humoral responses, and complement activation, which have been shown to play classical roles in LN disease development [19,28] (Fig 1E).

Although the number of longitudinal samples available for our study was limited, during intestinal RG blooms during disease flares, transcriptomic analyses showed that these LN patients had higher expression of the platelet activation module (Supplementary Data S5-S7). Furthermore, although it did not attain significance, the level of platelet gene activation was directly associated with that of neutrophil gene activation (rrm = 0.86, P = .062, Supplementary Data S6), which further supported a coordinated immune response involving platelets and neutrophils. Conversely, the expression level of platelet activation showed an inverse correlation trend with ISG modules (rrm = -0.88, P = .05, Supplementary Data S6), suggesting a reciprocal relationship during overall disease flares. Longitudinal evaluation of transcript modules showed a significant positive correlation between increased SLE disease activity (measured by the Systemic Lupus Erythematosus Disease Activity Index [SLEDAI]) and neutrophil module expression (rrm = 0.99, P = .001), while there was coordinated negative correlation of B-cell (rrm = -0.97, P = .006) and ISG modules (rrm = -0.96, P = .01) (Supplementary Data S7).

Serum anti-LG IgG antibody levels are markers of gut dysbiosis associated with RG blooms

In our earlier studies of SLE patients, higher levels of antibodies that bind to the LG produced by RG strains were correlated with higher lupus disease activity, as measured by the SLEDAI [29], and with serum IgG autoantibodies to native DNA [7]. Elevated levels of serum IgG antibodies against the novel surface LG from RG strains also significantly correlated with the faecal abundance of the RG species [4], a pattern here reproduced in a smaller cohort (Supplementary Data S1). To further investigate the immunobiologic implications of this specific host immune response, in patients of the Pilot cohort, we assessed levels of serum IgG to the RG-specific LG product [9] (Supplementary Data S1). We also studied a Validation cohort of patients with active biopsy-proven class III or IV LN from the standard-of-care control arm of the previously reported randomised multisite Abatacept and Cyclophosphamide Combination Therapy for Lupus Nephritis (ACCESS) trial (NCT00774852) [30] (Supplementary Table S4). Specifically, we sought to determine whether serum anti-LG IgG levels, which reflect the host response to RG colonisation and blooms [7,9], also correlated with systemic transcriptomic profiles at the time of LN flares (Supplementary Table S5).

Serum IgG antibody levels against the novel surface LG of the RG2 strain [4] were quantified with a previously reported immunoassay [9], and patient sera were stratified based on an anti-LG antibody cutoff value (59,000 U/mL). In the Pilot cohort, anti-LG IgG antibody levels were increased in the patients with LN with RG blooms compared with the patients with LN without RG bloom and the nonrenal patients (Supplementary Data S1A). In the Validation cohort, we found that 12 of these 20 active LN patients had anti-LG IgG antibody levels that exceeded the cutoff value (Supplementary Data S1).

These LN patients had been recruited from 11 of the 21 clinical sites in the ACCESS trial (Supplementary Data S8), geographically dispersed throughout the United States. Otherwise, these patients showed no associations between anti-LG antibody levels and patient demographics, antiphospholipid antibody positivity, anti-dsDNA autoantibody levels, complement levels, urinary protein-to-creatinine ratio, white and red blood cell counts, platelet counts, estimated glomerular filtration rate, or serum albumin level (Supplementary Table S3). Anti-LG antibody levels were unaffected by medications or corticosteroid exposure; however, the LN patients with elevated anti-LG antibodies were older (P = .02). In this cohort, each LN patient generally maintained a stable anti-LG antibody level over a 2-year follow-up period, with few exceptions [9] (Supplementary Data S1B).

Anti-LG antibody levels are correlated with platelet activation and neutrophil degranulation gene signatures

In the Pilot cohort, the LN subset with RG blooms had significantly higher expression of transcripts for platelet activation and neutrophil degranulation (P = .03) (Fig 1). However, there was no correlation with differences in the levels of specific cell types (eg, platelets or neutrophils) in the blood (Supplementary Data S9). In the Validation LN cohort, PCA of bulk-blood RNA-seq libraries, based on anti-LG antibody positivity, did not distinguish patients with global gene representation (Supplementary Data S10A). In contrast, anti-LG antibody positive and negative subsets significantly distinguished those with platelet module-specific gene expression, which was consistent with the findings that elevated anti-LG antibody levels correlated with significantly higher expression of platelet activation genes (P = .04) (Fig 2A-D [31]; Supplementary Tables S2 and S3; Supplementary Data S10B,C). Hence the multiscale approach validated both the biological specificity and the reproducibility of the platelet activation signature as a pathway-level biomarker in LN. Furthermore, in the Validation cohort, the levels of anti-LG antibodies also significantly directly correlated with platelet activation (r = 0.47, P < .05), with neutrophil extracellular trap (NET) fragmentation (r = 0.45, P < .05), and with a positive numerical trend with the neutrophil degranulation transcript module (r = 0.43, P = .06) (Supplementary Data S11 and S12; Supplementary Table S5).

Figure 2.

Figure 2.

Open in a new tab

Patients with anti-lipoglycan (LG) positive antibody levels in the Validation cohort displayed transcript profile evidence of platelet activation. (A) Using a quantitative Luminex-based assay, lupus nephritis (LN) patients were dichotomised based on a cutoff for immunoglobulin (Ig)G anti-LG reactivity (horizontal dashed line), distinguishing those with negative (n = 8) vs positive (≥5.9 × 104 U/mL; n = 12) antibody levels. Validation cohort LN samples were from the ITN034AI trial (NCT00774852) (see methods in the Supplementary Materials). (B) Pathways and (C) genes related to platelet and neutrophil activation were increased in anti-LG-positive patients. Pathway normalised enrichment score (NES) values and statistics are listed in Supplementary Table S2. *Indicates an unusual patient with a negative anti-LG antibody level who had a significantly elevated urine protein/creatine ratio of >20 g/g (see text and Supplementary Data S4). (D) Platelet gene expression module and (E) Neutrophil degranulation gene module scores in the anti-LG negative patients (n = 8) vs anti-LG-positive patients (n = 12) in the Validation cohort. (F) Platelet factor 4 (PF4) protein levels in the serum of patients with class III or class IV LN in the Validation cohort (n = 17). (G) Serum neutrophil extracellular trap (NET) fragment detection by enzyme-linked immunosorbent assay for citrullinated histone 3 (CitH3), myeloperoxidase (MPO), and DNA, which are serum markers for NETosis in the patients in the Validation cohort (n = 20). NET fragment detection utilised a published methodology for simultaneous detection of MPO-CitH3 and DNA fragments [31]. Statistics represent an unpaired, nonparametric t-test. AU, arbitrary units.

Hierarchical clustering of these overexpressed pathways in the anti-LG positive patients in the Validation cohort mirrored the patterns found in patients with RG blooms in the Pilot cohort (Supplementary Data S13). In contrast, there was no direct or indirect significant association with ISG expression (Supplementary Table S5). However, we also identified an exception—an LN patient with elevated platelet activation gene expression despite negative anti-LG antibody levels. This patient had a significantly elevated pathologic-range protein-to-creatinine ratio (>20 g/g) (Fig 2B; Supplementary Data S13), suggesting that serum antibody loss may have resulted from pathologic urinary excretion.

To assess these profiles in independent SLE cohorts, we analysed transcriptomic data from published reports on 390 patients with active LN [19,20,32,33]. We found inverse relationships between platelet activation/neutrophil degranulation genes with both ISG and B-cell activation/differentiation genes (Supplementary Data S4C,D), and the same relationships seen for modules for ISG and B-cell activation. However, in this composite of datasets from several reports, there was no identifiable relationship between platelets and neutrophil modules, and this was expected as there was no means to dichotomise this heterogeneous lupus dataset based on gut RG representation or circulating anti-LG antibody levels. Hence, RNA-seq analyses across cohorts provided circumstantial evidence for a distinct pathway of platelet activation and neutrophil degranulation that was first identified in LN patients with RG blooms, distinguishing this patient flare endotype, from flares involving previously implicated pathways, which are characterised by overexpression of B-cell activation and interferon-related genes [19].

Peripheral transcript profiles of platelet and neutrophil activation are associated with elevated serum PF4 and NETosis proteins

To further investigate the mechanisms underlying the observed platelet and neutrophil activation, we assessed for correlations between gene expression and serum levels of relevant self-proteins. Validation cohort patients with positive anti-LG antibody levels showed increased platelet activation gene scores and higher serum PF4 protein levels, confirming in vivo platelet activation (Fig 2C,D,F). These patients also exhibited gene signatures of increased neutrophil degranulation, as well as higher serum levels of NET fragments, including myeloperoxidase, citrulline-modified histone 3, and dsDNA, compared with anti-LG antibody-negative patients (Fig 2C,E,G). Significant correlations were found between levels of serum PF4 protein and platelet-related transcripts, including PROS1 (r = 0.52, P = .02), PPBP (r = 0.48, P = .03), and TREML1 (r = 0.48, P = .03), indicating coordinated platelet degranulation (Supplementary Data S14). Similarly, correlations between serum NET fragments, a hallmark of NETosis, and transcripts related to neutrophil degranulation and epithelial damage (KRT1: r = 0.58, P = .02; EGF: r = 0.53, P = .04) were observed (Supplementary Data S15). These direct transcript-protein relationships provide robust evidence for coordinated platelet, neutrophil activation, and NETosis as contributors to disease activity.

Intestinal colonisation of mice with LG-producing RG strains induces platelet activation

We next sought to survey for potential causal relationships between the influences of RG strains from LN patients and in vivo platelet activation [8,16] (Fig 3A). To assess the gut presence of RG akin to the RG-colonized LN subset, RG-specific quantitative polymerase chain reaction assays of faecal samples documented high-level colonisation in mice given RG strains isolated from LN patients, at 2 weeks post-gavage, but not in the controls (Fig 3B). Notably, after challenge, the rapid appearance of high serum fluorescein isothiocyanate-dextran levels documented impaired intestinal barrier functional integrity in RG-colonised mice (Fig 3C), akin to findings in LN patients with RG intestinal blooms (Supplementary Data S1). Overall, in RG gut colonisation studies, there was no evidence of sepsis, infection, changes in body temperature, or weight loss in treated mice. Colonisation with LG-producing RG induced significant global changes in the transcriptome, affecting over 700 enriched pathways (Supplementary Table S2). As shown by PCA, whole-blood RNA-seq libraries from peak RG colonisation revealed distinct gene expression profiles (Fig 3D). In RG-gavaged mice, there was significant overrepresentation of pathways for platelet activation and degranulation (Fig 3E; Supplementary Tables S2 and S3), mirroring findings in LN patients with RG blooms (Figs 1 and 2). Key genes involved in platelet activation and aggregation, such as Pros1, Pdgfa, Pf4, and Sparc, were notably overrepresented (Fig 3F, G). Although RG colonisation did not affect total white blood cell, lymphocyte, or platelet counts, mean platelet volume, or haemoglobin levels (Supplementary Data S16), RG-colonised mice did have significantly lower circulating neutrophil levels (P = .02), suggesting that RG-induced platelet activation may accelerate peripheral neutrophil consumption.

Figure 3.

Figure 3.

Open in a new tab

Ruminococcus gnavus (RG) gut gavage induces platelet-targeted transcriptomic changes and activates peripheral platelets. (A) The experimental model used antibiotic pretreatment, then subsequent murine gavage of RG or sham, and sample collection. (B) Polymerase chain reaction assessment of stool samples throughout the experiment identified peak RG colonisation of C57BL/6 mice at 2 weeks after the last gavage. Mice were pretreated for 2 weeks with an antibiotic cocktail of enrofloxacin, ampicillin, and metronidazole. Mice were then gavaged with either sham or RG every other day for a total of 5 administrations over 2 weeks. Timepoint 0 represents the end of this gavage period (n = 3 per group). (C) FITC-dextran levels in serum pre-and post-oral gavage of sham vs RG gavage-treated (RG-colonised) mice. n > 7 mice per group. (D) Principal component analysis of the top 25% varied genes in C57BL/6 mice at 2 weeks, and 4 weeks post-gavage, with either sham or RG. PC1 48.7%, PC2 28.3%. n = 3 per group. (E) Gene set enrichment analysis detected genes overexpressed in RG compared with sham at 2 weeks post-gavage. Pathway normalised enrichment scores are listed in Supplementary Table S2. (F) Expression at 2 weeks post-gavage of genes related to platelet activation. (G) Heatmap of whole-blood RNA sequencing (RNA-seq) genes related to platelet activation, in sham gavage (left) and RG-colonised (right) mice. *Indicates transcripts overrepresented in pathway analyses for platelet activation and/or degranulation in both murine RG colonisation model and human whole-blood RNA-seq studies (Figs 1, 2). (H, I) P-selectin+ platelets (relative to CD41+ FSClow) in C57BL/6 mice (H) and GP1b-Cre+ platelets in tdTomato-GP1bCre mice (I) with or without thrombin stimulation detected by flow cytometry. n > 3 mice per group. FITC, fluorescein isothiocyanate, FSC, forward scatter; GP1b, glycoprotein Ib.

To understand RG-driven platelet activation in thromboinflammatory conditions, we used a standard set of experimental agonists to assess ex vivo platelet responsiveness in RG-colonised mice. During thrombogenesis, P-selectin is packaged into plate-let cytoplasmic α-granule organelles. Using flow cytometry, we quantified P-selectin levels expressed on platelet membranes. We found that, upon ex vivo activation with thrombin, P-selectin rapidly translocated to the platelet surface in RG-gavaged mice, but not in sham-treated mice (Fig 3H; Supplementary Data S17), where it is accessible for cell-cell interactions, including those with neutrophils. Using RG-colonised tdTomato-GP1b transgenic mice (Supplementary Data S18), we found no differences in basal P-selectin expression. However, after in vitro challenge with thrombin, we documented enhanced P-selectin responses (Fig 3I). In contrast, no functional differences were observed with either type I collagen or adenosine diphosphate (ADP), which act via receptors distinct from those activated by thrombin (Supplementary Data S19).

RG colonisation leads to a robust local intestinal immune response

After gavage with an LG-producing RG strain, but not in shamgavaged control mice, RG colonisation led to large lymphoid aggregates in the ileal lamina propria. These contained IgA+CD45+ lymphocytes (Supplementary Data S20A-D). Additionally, RG-colonised mice showed reduced epithelial expression of claudin-1, a protein essential for maintaining intestinal barrier integrity (Supplementary Data S20E,F). These findings suggest that certain RG strains can compromise gut barrier function, potentially allowing systemic exposure to numerous microbial factors.

Intestinal colonisation increases the number of megakaryocytes in the bone marrow and spleen

Platelets, formed from the fragmentation of megakaryocyte cytoplasm in the bone marrow, lack nuclei and mitochondria. In GP1b-tdTomato platelet reporter mice, we examined the effect of RG gut colonisation on megakaryocytes. After colonisation with LG-producing RG, megakaryocyte numbers significantly increased in the bone marrow (Supplementary Data S21A). We also assessed the impact on the spleen, a key site for platelet storage and extramedullary production during periods of increased demand, such as in sepsis [34]. RG gavage resulted in significantly higher megakaryocyte numbers in the spleen compared with sham-treated mice (Supplementary Data S21B) (Supplementary Data S33).

Pathogenic LG-producing RG strains induce renal injury in lupus-prone mice

To assess the influence of LG production, we used bacterial strains with or without the capacity to produce LG in vitro [7]. These included newly isolated RG subclones obtained by subculture and colony isolation from the previously described S107–48 RG strain recovered from an LN patient [9]. For these studies, we used the nb90 subclone, which lacks LG production, and the nb92 subclone, which consistently produces LG. In wild-type mice, colonisation with LG-producing strains did not lead to renal pathology (Supplementary Data S22).

To consider potential effects on autoimmune pathogenesis and end-organ effects, we studied the effect of RG colonisation in lupus-prone FcγRIIb-/- mice, which are at risk of autoimmune renal injury [35,36]. FcγRIIb-/- mice were gut-colonised with the control RG1 (also called the ATCC 29149) strain (an LG-non-producing RG strain [7] from a healthy donor) or with the LG-producing nb92, or RG2 (also called CC55_001C) strains (Supplementary Data S32). The LG-producing strains significantly increased spleen megakaryocyte numbers, whereas the LG-non-producing strain did not (Fig 4A,B). In lupus-prone mice, colonisation with LG-producing RG strains (nb92 or RG2) in FcγRIIb-/- mice induced significant renal cellular infiltration, marked by tubulointerstitial lymphocytic infiltrates (Fig 4C,D,E) with prominent T-cell and macrophage populations identified by immunofluorescence (Fig 4F) (Supplementary Tables S6 and S7). Furthermore, in mice colonised with LG-producing RG strains, there was deposition of myeloperoxidase and citrullinated histone 3 markers consistent with neutrophil degranulation and death by NETosis (Fig 4F-I; Supplementary Data S23). This renal pathology was not present in shamgavaged mice or in mice colonised with the control LG-nonproducing strain. Hence, in genetically lupus-prone female mice, LG-producing RG strains drive both NETosis and renal injury.

Figure 4.

Figure 4.

Open in a new tab

Colonisation of lupus-prone mice causes splenic megakaryocytosis and increased renal cellular infiltration by macrophage/monocytes and lymphocytes with increased local NETosis. (A, B) Splenic immunohistochemistry shows an increased number of megakaryocytes, representing megakaryocytosis, after gut colonisation with the Ruminococcus gnavus RG2 strain and the nb92 subclone that produces lipoglycan (LG), but not after sham or colonisation with the nb90 subclone that does not produce LG. See Supplementary Data S32. Immunofluorescence of splenic megakaryocytes in 12-month-old, female, FcγRIIb-/- mice (as defined by CD41+ cells) after gavage with LG-positive RG subclone or strain (nb92, RG2) versus a subclone that does not produce LG (nb90). The cell counts represent the number of cells per low-power field (n > 3 mice per group). (C) Representative kidney sections (hematoxylin and eosin stained) after colonisation of FcγRIIb-/- mice with LG-producing RG strains (nb92, RG2), and LG-non-producing strain (nb90), 10 days after the last gavage. Bar = 50 μm. (D) Quantification of the renal activity index evaluated a blinded reviewer for the presence (score 0–3) of mononuclear tubulointerstitial cells, glomerular leukocyte infiltration, hyalinisation of glomeruli, closed capillary loops, or early crescent formation/fibrotic ring (see Supplementary Table S6). Statistics represent an unpaired t-test. (E) Example of glomeruli of RG2 (LG-producing RG strain)-colonised FcγRIIb-/- lupus-prone female mice, showing mononuclear interstitial cellular infiltrations (#), early crescent formation (*) and hyalinisation (arrow) (original magnification, 40×). Bar = 20 μm. (F) Representative images of merged immunofluorescence of macrophage/monocytes (F4/80), neutrophils (MPO), with evidence of NETosis (CitH3), and T cells (CD3). Bar = 50 μm. Regions of interest with glomeruli and surrounding macrophages are encircled with dashed lines. Arrow highlights area of NETosis adjacent to a glomerulus, based on MPO, CitH3, and F4/80 detection. (G) Cell percentage of F4/80+ cells (macrophages), (H) MPO+ cells (neutrophils), and (I) CD3+ cells (T cells) relative to all cells (DAPI) in low-power fields, with >3 fields evaluated per mouse (n = 3 mice per group). Statistics represent an unpaired t-test. CitH3, citrullinated histone 3; DAPI, 4′,6-diamidino-2-phenylindole; FcγRIIb, Fc gamma receptor IIb; MPO, myeloperoxidase; NET, neutrophil extracellular trap.

Intraperitoneal injection of LG alone drives platelet activation and in vivo megakaryocytosis

We next sought to determine whether the LG glycan moiety alone is sufficient to drive these features in vivo, and we therefore administered a purified LG preparation via intraperitoneal injection. As with our findings after RG colonisation (Fig 3), in LG-injected mice RNA library analysis revealed overexpression of genes associated with platelet activation compared with vehicle-injected mice (Fig 5A-C; Supplementary Tables S2 and S3; Supplementary Data S24). In both wild-type and lupus-prone mice, LG treatment significantly increased functional platelet responses to ex vivo thrombin stimulation (Fig 5D,E), without associated alterations in total white blood cell count, lymphocytes, neutrophils, haemoglobin, platelet count, or mean platelet volume compared with vehicle controls (Supplementary Data S25, S26). LG injection induced enhanced ex vivo platelet responses at 6 days post-injection, but not at earlier time points (8 hours or 2 days; Supplementary Data S27). Given the ~4-day platelet lifespan in mice [37], this suggests that the influence of LG is mediated by sensitisation of newly formed platelets to thrombin stimulation, which then mature, rather than by direct effects of the lipoglycan on mature platelets.

Figure 5.

Figure 5.

Open in a new tab

Intraperitoneal (IP) administration of lipoglycan (LG) preparation in C57BL/6, FcγRIIb-/- and TLR2-/- mice increases splenic megakaryocyte numbers, peripheral blood platelet activation, and P-selectin externalisation. (A) Platelet activation genes were increased in whole-blood RNA sequencing (RNA-seq) studies 6 days after a single IP injection of an LG preparation (1.5 μg/g), isolated from the RG2 strain, into FcγRIIb-/- mice. n greater than or equal to 3 mice per group. (B) Heatmap visualisation of genes related to platelet activation and degranulation and the associated immune activation pathways induced by RG bacterial products. (C) Gene set enrichment analysis (GSEA) detects pathways overrepresented in mice after LG injection. Pathway normalised enrichment scores (NES) are listed in Supplementary Table S2. (D, E) Functional platelet activation 6 days after IP LG injection was evaluated by flow cytometric monitoring of the percentage of surface P-selectin expressing cells after ex vivo thrombin stimulation in (D) C57BL/6 and (E) FcγRIIb-/- mice. n > 3 mice per group. (F) Immunofluorescence megakaryocyte quantification in the spleen of 5- to 8-month-old lupus-prone, female, FcγRIIb-/- mice 6 days after one-time LG IP injection. Bar = 50 μm. Cell counts are the number of cells per low-power field (n > 3 mice per group). (G-I) TLR2-dependence of RG LG preparation-mediated activities. (G) Increased megakaryocyte numbers in the spleen were observed in wild-type LG-injected mice, but not LG-injected TLR2-/- mice. Cell counts are the number of cells per low-power field (n = 3–4 fields per mouse, n > 3 mice per group). (H) Percentage of surface P-selectin-expressing cells, functional platelet activation 6 days after IP LG injection, in response to ex vivo thrombin stimulation in TLR2-/- mice (n > 3 mice per group). (I) Platelet activation genes were increased in RNA-seq studies of whole blood collected 6 days after LG injection in both C57BL/6 (B6) and TLR2-/- mice (n > 3 mice per group). (J) GSEA showed increased expression of genes related to platelet activation and degranulation pathways in B6 mice, but not in TLR2-/- mice, after IP injection of LG preparation. Pathway NES listed in Supplementary Table S2. In panels A and I, vertical dashed lines indicate log2(Fold Change) ±1, representing a 2-fold change in gene expression. Unpaired, nonparametric t-test. DAPI, 4′,6-diamidino-2-phenylindole; FcγRIIb, Fc gamma receptor IIb; TLR2, Toll-like receptor 2.

LG-injected mice also exhibited splenic megakaryocytosis (Fig 5F), indicating that LG affects platelet progenitors and subsequent platelet activation functions. However, there was no increase in functional aggregation rates after thrombin or ADP stimulation compared with vehicle-injected mice (Supplementary Data S28). Hence, LG specifically drives platelet activation but not the platelet aggregation that leads to clot formation.

RG drives platelet activation and immune modulation in both lupus patients and murine models

As described above, in both human and murine studies, we found consistent evidence of RG-driven platelet activation and innate immune modulation. Indeed, in RG-colonised mice, the pathways for platelet activation, degranulation, and NETosis closely matched those in RG-exposed LN patients. Under the influence of LG-producing strains, the shared overexpression pattern was also seen at the level of individual genes crucial for regulating platelet activation, degranulation, and NET formation in a pathway-dependent manner (Supplementary Data S29).

Megakaryocytosis, altered platelet function, and transcriptomic activation mediated by LG preparations in vivo are Toll-like receptor 2-dependent

Prior studies using a reporter cell line implicated Toll-like receptor (TLR)2 in the immune activation induced by an LG preparation isolated from RG by chromatography (see Methods [7]). As TLR2 is expressed on megakaryocytes and platelets [38], we therefore tested the in vivo responses to such LG preparations. In wild-type mice, a single LG injection induced megakaryocytosis expansion of these platelet precursors (Fig 5G), as well as platelet activation and TLR-mediated cascade pathways, whereas neither was observed in TLR2-/- mice (Fig 5H,I; Supplementary Data S30). In LG-injected wild-type mice, GSEA confirmed overrepresentation of pathways associated with platelet activation, degranulation, macrophage migration, and neutrophil degranulation, which was not observed in LG-injected TLR2-/- mice (Fig 5J; Supplementary Tables S2 and S3; Supplementary Data S31). These findings highlight the crucial in vivo role of the TLR2 cascade in mediating the thromboinflammatory response triggered by LG preparations (Fig 6 [4,39]).

Figure 6.

Figure 6.

Open in a new tab

A mechanistic hypothesis of the influence of RG on systemic lupus erythematosus pathogenesis. Impaired gut barrier function allows microbial products to translocate, exposing the immune system to RG surface lipoglycan (LG). RG can activate platelets via Toll-like receptor 2 (TLR2), and, potentially, via immune complexes acting through FcγR and complement/mannose-binding lectin, promoting platelet–immune cell interactions. This leads to lupus nephritis injury via P-selectin-mediated neutrophil priming, resulting in the release of neutrophil extracellular traps (NETs) that contain inflammatory factors and active neutrophil proteases. Activated platelets also release soluble factors, such as platelet factor 4 (PF-4), which drive granulocyte maturation and further sustain inflammation. NET, neutrophil extracellular trap; RG, Ruminococcus gnavus; SLEDAI, Systemic Lupus Erythematosus Disease Activity Index. Adapted from [4,39].

DISCUSSION

In this study, we identified a distinct pathophysiologic endotype in patients with LN characterised by gut expansion of the pathobiont RG and the release of a proinflammatory LG toxin. In 2 independent cohorts, unbiased whole-blood RNA-seq (Figs 1 and 2) provided a comprehensive view of the immune landscape. LN patients with RG blooms exhibited a gene expression profile indicative of platelet activation and neutrophil degranulation. In contrast, LN patients without gut bloom displayed distinct transcriptional patterns associated with B-cell and complement activation [19,20]. In support of this, analysis of publicly available whole-blood gene expression datasets underscored the clear separation of these molecular pathologic endotypes among patients with active LN, especially the inverse correlation between platelet activation and B-cell activation modules.

To further investigate the causative roles in lupus-like disease, we conducted animal studies. Our murine gut colonisation experiments demonstrated that only RG strains that produced the unique immunogenic surface LG toxin, or injection of an LG preparation from RG alone, induced increased tissue levels of megakaryocytes, platelet activation, and renal injury (Figs 3 and 4). Indeed, megakaryocytes and platelets express innate immune receptors, including TLRs, which can stimulate proliferation and differentiation, particularly in the context of inflammation, infection, or sepsis [38]. Transcript profiles were induced by LG injection that reiterated findings after colonisation with LG-producing RG strains (Fig 5A-C), as well as induced functional activation in platelets (Fig 5D, E). However, responses induced by an LG preparation from RG were not observed in TLR2-deficient mice (Fig 5). Notably, to induce platelet activation via specific signalling pathways, TLR2 functions as a heterodimer [40]. Our RNA-seq data demonstrated implicated activation via the TLR2 and the TLR1 signalling subunit. These findings align with a report that synthetic TLR2 agonists can activate platelets [41] and contribute to thromboinflammatory responses and NETosis [42]. However, it remains possible that other host innate immune receptors, such as C-type lectins, also contribute to RG-driven immune activation.

Beyond platelet effects, which included increased serum P-selectin expression, RG blooms in LN patients enhanced the expression of genes associated with NET formation (eg, myeloperoxidase, citrullinated histone H3, dsDNA) [31,43], features of the thromboinflammatory innate pathway also observed in sepsis [44]. In our patients with RG intestinal expansions, we observed no episodes of bacteraemia, epithelial barrier invasion, or associated tissue injury. Furthermore, unlike in sepsis, neither thrombosis nor vascular clotting was prominent, and these patients showed no signs of disseminated intravascular coagulation or thrombocytopenia. To better define the central features shared by patients with this LN endotype, we intentionally excluded patients with clinical antiphospholipid syndrome, thereby clarifying that these autoimmune features are not integral to this LG toxin-driven thromboinflammatory endotype.

Supporting the translational relevance of our murine models, comparative analysis of patient and murine RNA-seq libraries revealed conserved pathway overexpression (Supplementary Data S29). Hence, despite species-specific differences between mice and humans, and significant differences in LG exposure timelines (short-term in murine models vs years in LN patients), the transcriptomic profiles induced by murine RG gut colonisation remarkably mirrored those observed in LN patients with RG intestinal blooms.

In the kidneys of lupus-prone mice, within weeks of RG colonisation, a profound cellular infiltration by neutrophils, macrophages, and T cells was observed. This provided a novel model of early-stage immune cell infiltration driven by gut dysbiosis. In contrast, glomerulonephritis in patients with SLE typically develops over a longer period. Murine studies confirmed that these LG preparations alone can induce the central features of an early step in autoimmune renal pathogenesis.

In mice, we observed neutrophil infiltration and degranulation, as well as macrophage and T-cell entry into the kidneys of RG-colonised lupus-prone mice. Just weeks after murine RG gut colonisation, we observed immune cell infiltration, representing an early-stage model of dysbiosis-driven inflammation. However, refinements in experimental design can address the consequences of prolonged intestinal RG persistence and its impact on long-term renal pathology that is more typical in human LN pathogenesis. Overall, our findings from human and murine studies support the hypothesis that gut expansion of pathogenic RG strains in LN patients drives platelet and immune activation, as well as renal disease (Fig 6).

Among the limitations that warrant consideration, the Pilot cohort size was modest and from a single site. Because women are most affected by LN, the Pilot cohort comprised female patients, and the 20 LN patients in the Validation cohort included only one male patient. Although gut microbiota data were unavailable for the Validation cohort, the number of significant cross-data correlations in this cohort demonstrated that anti-LG IgG antibody levels served as a powerful surrogate for identifying patients with the thromboinflammatory endotype. The LG from RG strains is specifically recognised by a well-characterised recombinant antibody, although future studies will be needed to extend testing to additional strains and other glycan molecules. In addition, we appreciate that a copurified minor component may influence the host response to LG [7]. It remains possible that, at the molecular level, TLR2 dependence does not rely on a direct interaction with the LG itself but rather with a copurified factor [45–48].

The initial renal injury from RG colonisation and LG exposure mirrors aspects of early cellular infiltration seen in sepsis-driven acute kidney injury [49,50]. However, this RG-associated early inflammatory renal injury occurs without overt infection or widespread thromboses. We hypothesise that such repeated or prolonged insults are necessary for the development of all the features of classical LN in susceptible individuals.

Taken together, our findings define an LN endotype associated with blooms of LG-producing RG strains and the underlying mechanisms. In patients with LN, platelet activation has previously been linked to leukocyte recruitment and end-organ injury [22,51–53], roles central to this endotype. Unlike other LN endotypes, which include B-cell activation or interferon pathways, our findings suggest that a pathogenic subset of LN patients can be identified using the anti-LG antibody biomarker, reflecting the in vivo influence of a gut pathobiont that acts via specific disease mechanisms. Overall, this previously unknown LN endotype is triggered by systemic platelet activation—beyond their traditional roles in thrombosis—as immune sentinels that drive inflammation. As a result, our studies are reshaping our understanding of LN pathogenesis in lupus patients with RG blooms.

In-depth studies are now needed to investigate the relationship between elevated anti-LG antibody levels and the risk of renal failure, as well as the potential for impairment of clinical responses to therapy due to overwhelming microbiome-induced innate immune activation in the lupus host. Similarly, our understanding of RG ecobiology, particularly the relationship with time-dependent shifts in pathogenic strains and overall disease activity, is currently limited. We documented remarkably high antibody levels against the LG toxin in many patients with LN. The direct pathogenic properties of such antibodies, and/or their IgG immune complexes with the lipoglycan, need to be characterised in patients with this endotype. The effects on platelet biology and functionality, including transcript levels in isolated platelets from affected individuals, should also be investigated. The induced B-cell/antibody clones should also be evaluated for their potential contribution to pathogenesis through cross-reactivity with self- or renal antigens and by forming complexes in vivo. We found a significant difference in the mean age of the patients in the Validation cohort in anti-LG antibody high versus anti-LG antibody low subsets (Supplementary Table S3; P = .02), although for the categorical age distributions (18–24, 25–39, >40 years) there were no significant differences (P = .25). It currently appears unlikely that RG intestinal expansions occur only late in the course of LN pathogenesis, as we have detected high anti-LG antibody levels in individual patients on the day of LN biopsy and diagnosis [13]. However, to address this question, longitudinal studies of patients with LN enrolment at disease onset may be necessary.

In summary, our study highlights the unique pathophysiology underlying RG blooms in a subset of LN patients. Our findings reveal many parallels with poststreptococcal autoimmune syndromes, such as rheumatic fever and poststreptococcal glomerulonephritis, as recently discussed [4]. However, in the absence of overt infections and associated tissue barrier destruction, RG blooms provoke host immune responses that disrupt gut barrier function [4,5]. The detection of RG blooms and related systemic anti-LG antibody responses should serve as a biomarker to identify patients at higher risk for this endotype of disease flare and tissue damage, particularly in genetically susceptible individuals. Future work in SLE inception cohorts will therefore also be needed to assess the dynamics of RG strain expansion, competition, and persistence.

Early clinical testing of anti-LG antibody as a biomarker may help stratify patients by risk of microbiome-induced disease flares, potentially as early as their first clinical presentation. Detection of this pronounced platelet-driven immune response in an LN patient may inform therapy decisions more effectively, facilitating a personalised medicine approach [38]. These findings highlight the potential clinical opportunities to target RG-induced pathways, including through TLR2 blockade or selective antibiotic treatment, which may spare some patients the broad immunosuppressive effects and associated adverse events of current standard-of-care therapies for lupus.

METHODS

This study was conducted in accordance with the principles outlined in the Declaration of Helsinki. Written informed consent, approved by the New York University (NYU) Institutional Review Board, was obtained from all subjects before their enrolment in the study for research use and publication of their data. Detailed methods are provided in the Supplementary Material.

Supplementary Material

Suppl file1
supplementary table 3
Suppl file 3
supplementary table 2
Suppl fil 4

Supplementary materials

Supplementary material associated with this article can be found in the online version at doi:10.1016/j.ard.2026.03.002.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • The gut microbiome has been recognized as a regulator of systemic immunity.

  • Lupus nephritis (LN) is a serious illness, yet despite improvements in therapeutic interventions, many patients progress to renal failure.

  • Intestinal blooms of Ruminococcus gnavus (RG)(renamed Mediterraneibacter gnavus) are associated with higher lupus disease activity.

  • Strains isolated from lupus patients have been shown to produce a cell wall-associated novel lipoglycan (LG); however, the roles of this lipoglycan in the pathogenesis of lupus and disease progression remain unclear.

WHAT THIS STUDY ADDS

  • We identify a distinct LN endotype in which systemic innate immune thromboinflammation is linked to RG gut blooms and the immune response to a novel RG-associated LG.

  • In the Pilot and Validation cohorts, the serum level of anti-LG antibody was shown to be a reliable surrogate biomarker for this inflammatory profile (endotype), based on transcriptomic profiles of platelet activation and neutrophil degranulation, as well as immunoassay measures of Platelet Factor 4 levels (PF4) and neutrophil extracellular trap (NET) release.

  • Murine gut colonization demonstrated that only RG strains that produce the novel immunogenic surface LG toxin, or injection of an LG preparation from RG alone, induced increased megakaryocytosis, platelet activation, and renal injury, remarkably mirroring findings in LN patients with RG intestinal blooms.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • This study identified a previously unsuspected, yet common, LN endotype in which renal flares are driven by gut bacterial blooms.

  • Anti-LG antibody testing may help identify LN patients with this unique endotype, and therefore those most at risk for microbiome-driven thromboinflammation and renal involvement, with implications for disease presentation, therapeutic decision-making, and subsequent personalized treatment.

  • This work opens the door to new therapeutic directions, including newly characterized RG-induced pathogenic pathways, TLR2 blockade, and the therapeutic targeting of intestinal blooms of pathobiont strains.

  • Taken together, our findings may provide a means to treat a common driver of LN flares without the immunosuppression inherent to current treatments.

Acknowledgements

We are grateful to Parthiban Rajasekaran and Sanofi for advice, and to Sharon Chung and the Immune Tolerance Network for biosamples used in the Validation cohort. Histological analyses were aided by Mark Alu and Gyles Ward, as well as the NYU GSOM Experimental Pathology Core. RNA sequencing studies were conducted at the NYU Genome Technology Center, with contributions from Gael Westby, Emily Guzman, Shyanon Rai, and Paul Zappile, under the supervision of Adriana Heguy. Flow cytometry studies were completed through the NYU Cytometry and Cell Sorting Laboratory, with guidance from Catherine Rapelje, Florenal Joseph, and Jiangwei Li. These NYU facilities are funded in part by the Cancer Center Support Grant NIH/NCI 5 P30CA16087 at the Laura and Isaac Perlmutter Cancer Center. We are grateful to Kate Trujillo, Uzair Chaudhary, Steven Medvedovsky, Sophia DelGiudice, and Heiko Käßner for their technical assistance. We also appreciate the advice and assistance of Bharati Matta and Tatiana Borja with neutrophil assays and platelet activation flow cytometry assays. We are grateful to Jason Knight, Marc Scherlinger, Patrick Blanco, Eric Morand, Diance, Kamen, Karen Costenbader, Laurence Morel and Brad Rovin for key discussions. We thank Peter Izmirly and Jill Buyon for clinical support. We thank the clinical coordinators of this study and all our patients for their participation.

Funding

This study was in part funded by NIH P50-AR070591, R01 AI143313, R21AI180737, the Lupus Foundation of America Gilkeson Career Development Award (AA), NIH 5T32AR069515–07 (AA), and the Department of Defense CDMRP LRP W81XWH-18–1-0674 (BJ), and by Sanofi, Lyon, France. Prior studies were funded in part by the Lupus Research Alliance (GJS) and the Colton Center (GJS).

Footnotes

Competing interests

GJS reports financial support was provided by Sanofi and National Institutes of Health. GJS has patent pending to NYU. Grant support from Novartis and Genentech. No direct relationship to this project. All the other authors declare they have no competing interests.

Patient consent for publication

Written informed consent was obtained.

Ethics approval

Supervised by the NYU IRB.

Provenance and peer review

Not commissioned; externally peer reviewed.

Data availability statement

Clinical data for patients in our Pilot cohort are available in our prior publications [7,9], and for the 35 Validation cohort included in the Supplementary data, from the Immune Tolerance Network [26] (itntrialshare.org) of the NIAID. Datasets for RNA sequence data will be made available to online repositories. Additional primary data are available to qualified parties on request.

REFERENCES

  • [1].Meadows V, Antonio JM, Ferraris RP, Gao N. Ruminococcus gnavus in the gut: driver, contributor, or innocent bystander in steatotic liver disease? FEBS J 2025;292(6):1252–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Miyauchi E, Shimokawa C, Steimle A, Desai MS, Ohno H. The impact of the gut microbiome on extra-intestinal autoimmune diseases. Nat Rev Immunol 2023;23(1):9–23. [DOI] [PubMed] [Google Scholar]
  • [3].Anders HJ, Saxena R, Zhao MH, Parodis I, Salmon JE, Mohan C. Lupus nephritis. Nat Rev Dis Primers 2020;6(1):7. [DOI] [PubMed] [Google Scholar]
  • [4].Silverman GJ, Azzouz DF, Gisch N, Amarnani A. The gut microbiome in systemic lupus erythematosus: lessons from rheumatic fever. Nat Rev Rheumatol 2024;20:143–57. [DOI] [PubMed] [Google Scholar]
  • [5].Amarnani A, Silverman GJ. Understanding the roles of the microbiome in autoimmune rheumatic diseases. Rheumatol Immunol Res 2023;4(4):177–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Togo AH, Diop A, Bittar F, Maraninchi M, Valero R, Armstrong N, et al. Description of Mediterraneibacter massiliensis, gen. nov., sp. nov., a new genus isolated from the gut microbiota of an obese patient and reclassification of Ruminococcus faecis, Ruminococcus lactaris, Ruminococcus torques, Ruminococcus gnavus and Clostridium glycyrrhizinilyticum as Mediterraneibacter faecis comb. nov., Mediterraneibacter lactaris comb. nov., Mediterraneibacter torques comb. nov., Mediterraneibacter gnavus comb. nov. and Mediterraneibacter glycyrrhizinilyticus comb. nov. Antonie Van Leeuwenhoek. 2018;111:2107–28. [DOI] [PubMed] [Google Scholar]
  • [7].Azzouz D, Omarbekova A, Heguy A, Schwudke D, Gisch N, Rovin BH, et al. Lupus nephritis is linked to disease-activity associated expansions and immunity to a gut commensal. Ann Rheum Dis 2019;78(7):947–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Silverman GJ, Deng J, Azzouz DF. Sex-dependent Lupus Blautia (Ruminococcus) gnavus strain induction of zonulin-mediated intestinal permeability and autoimmunity. Front Immunol 2022;13:897971. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Azzouz DF, Chen Z, Izmirly PM, Chen LA, Li Z, Zhang C, et al. Longitudinal gut microbiome analyses and blooms of pathogenic strains during lupus disease flares. Ann Rheum Dis 2023;82(10):1315–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Juge N Microbe profile: Ruminococcus gnavus: the yin and yang of human gut symbionts. Microbiology (Reading) 2023;169(8):001383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Sagheddu V, Patrone V, Miragoli F, Puglisi E, Morelli L. Infant early gut colonization by Lachnospiraceae: high frequency of Ruminococcus gnavus. Front Pediatr 2016;4:57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Nooij S, Plomp N, Sanders IMJG, Schout L, van der Meulen AE, Terveer EM, et al. Metagenomic global survey and in-depth genomic analyses of Ruminococcus gnavus reveal differences across host lifestyle and health status. Nat Commun 2025;16(1):1182. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Silverman G, Azzouz D, Grönwal C, Gunnarsson I, Svenungsson E. Validation of a serologic antibody biomarker against a candidate gut pathobiont for the diagnosis of lupus nephritis [Abstract]. Arthritis Rheumatol 2019;71(Suppl 10):1915. [Google Scholar]
  • [14].Chen BD, Jia XM, Xu JY, Zhao LD, Ji JY, Wu BX, et al. An autoimmunogenic and proinflammatory profile defined by the gut microbiota of patients with untreated systemic lupus erythematosus. Arthritis Rheumatol 2021;73(2):232–43. [DOI] [PubMed] [Google Scholar]
  • [15].Toumi E, Goutorbe B, Plauzolles A, Bonnet M, Mezouar S, Militello M, et al. Gut microbiota in systemic lupus erythematosus patients and lupus mouse model: a cross species comparative analysis for biomarker discovery. Front Immunol 2022;13:943241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Ma L, Ge Y, Six N, Choi SC, Brown J, Castellanos Garcia A, et al. Gut expansion of a human lupus pathobiont is associated with autoantibody production and T cell dysregulation. ACR Open Rheumatol 2025;7(5):e70033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Aringer M, Costenbader K, Daikh D, Brinks R, Mosca M, Ramsey-Goldman R, et al. 2019 European League Against Rheumatism/American College of Rheumatology classification criteria for systemic lupus erythematosus. Arthritis Rheumatol 2019;71(9):1400–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Fasano A All disease begins in the (leaky) gut: role of zonulin-mediated gut permeability in the pathogenesis of some chronic inflammatory diseases. F1000Res. 2020;9:F1000 Faculty Rev-69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Banchereau R, Hong S, Cantarel B, Baldwin N, Baisch J, Edens M, et al. Personalized immunomonitoring uncovers molecular networks that stratify lupus patients. Cell 2016;165(3):551–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Figgett WA, Monaghan K, Ng M, Alhamdoosh M, Maraskovsky E, Wilson NJ, et al. Machine learning applied to whole-blood RNA-sequencing data uncovers distinct subsets of patients with systemic lupus erythematosus. Clin Transl Immunol 2019;8(12):e01093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Muller MA, Luttrell-Williams E, Bash H, Cornwell MG, Belmont HM, Izmirly P, et al. Platelet gene expression in systemic lupus erythematosus and cardiovascular health. JACC Basic Transl Sci 2025;10(12):101395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Berger JS, Cornwell MG, Xia Y, Muller MA, Smilowitz NR, Newman JD, et al. A Platelet Reactivity Expression Score derived from patients with peripheral artery disease predicts cardiovascular risk. Nat Commun 2024;15(1):6902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Lood C, Amisten S, Gullstrand B, Jönsen A, Allhorn M, Truedsson L, et al. Platelet transcriptional profile and protein expression in patients with systemic lupus erythematosus: up-regulation of the type I interferon system is strongly associated with vascular disease. Blood 2010;116(11):1951–7. [DOI] [PubMed] [Google Scholar]
  • [24].Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005;102(43):15545–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Foroutan M, Bhuva DD, Lyu R, Horan K, Cursons J, Davis MJ. Single sample scoring of molecular phenotypes. BMC Bioinformatics 2018;19(1):404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Liu Q, Zhu W, Wen X, Da Y. The role of platelet–neutrophil interactions in driving autoimmune diseases. Immunology 2025;175(1):1–15. [DOI] [PubMed] [Google Scholar]
  • [27].Lemke G How macrophages deal with death. Nat Rev Immunol 2019;19(9):539–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Dai X, Fan Y, Zhao X. Systemic lupus erythematosus: updated insights on the pathogenesis, diagnosis, prevention and therapeutics. Signal Transduct Target Ther 2025;10(1):102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Gladman DD, Iban¸ ez D, Urowitz MB. Systemic lupus erythematosus disease activity index 2000. J Rheumatol 2002;29(2):288–91. [PubMed] [Google Scholar]
  • [30].ACCESS Trial Group. Treatment of lupus nephritis with abatacept: the abatacept and cyclophosphamide combination efficacy and safety study. Arthritis Rheumatol 2014;66(11):3096–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].Matta B, Battaglia J, Barnes BJ. Detection of neutrophil extracellular traps in patient plasma: method development and validation in systemic lupus erythematosus and healthy donors that carry IRF5 genetic risk. Front Immunol 2022;13:951254. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [32].Wither JE, Prokopec SD, Noamani B, Chang N-H, Bonilla D, Touma Z, et al. Identification of a neutrophil-related gene expression signature that is enriched in adult systemic lupus erythematosus patients with active nephritis: clinical/pathologic associations and etiologic mechanisms. PLoS One 2018;13(5):e0196117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [33].Chiche L, Jourde-Chiche N, Whalen E, Presnell S, Gersuk V, Dang K, et al. Modular transcriptional repertoire analyses of adults with systemic lupus erythematosus reveal distinct type I and type II interferon signatures. Arthritis Rheumatol 2014;66(6):1583–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Puhm F, Laroche A, Boilard E. Diversity of megakaryocytes. Arterioscler Thromb Vasc Biol 2023;43(11):2088–98. [DOI] [PubMed] [Google Scholar]
  • [35].Horvei KD, Pedersen HL, Fismen S, Thiyagarajan D, Schneider A, Rekvig OP, et al. Lupus nephritis progression in FcγRIIB-/-yaa mice is associated with early development of glomerular electron dense deposits and loss of renal DNase I in severe disease. PLoS One 2017;12(11):e0188863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Thim-Uam A, Surawut S, Issara-Amphorn J, Jaroonwitchawan T, Hiengrach P, Chatthanathon P, et al. Leaky-gut enhanced lupus progression in the Fc gamma receptor-IIb deficient and pristane-induced mouse models of lupus. Sci Rep 2020;10(1):777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [37].Armstrong PC, Allan HE, Kirkby NS, Gutmann C, Joshi A, Crescente M, et al. Temporal in vivo platelet labeling in mice reveals age-dependent receptor expression and conservation of specific mRNAs. Blood Adv 2022;6(23):6028–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Beaulieu LM, Lin E, Morin KM, Tanriverdi K, Freedman JE. Regulatory effects of TLR2 on megakaryocytic cell function. Blood 2011;117(22):5963–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].Scherlinger M, Richez C, Tsokos GC, Boilard E, Blanco P. The role of platelets in immune-mediated inflammatory diseases. Nat Rev Immunol 2023;23(8):495–510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [40].Blair P, Rex S, Vitseva O, Beaulieu L, Tanriverdi K, Chakrabarti S, et al. Stimulation of Toll-like receptor 2 in human platelets induces a thromboinflammatory response through activation of phosphoinositide 3-kinase. Circ Res 2009;104(3):346–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [41].Cognasse F, Nguyen KA, Damien P, McNicol A, Pozzetto B, Hamzeh-Cognasse H, et al. The inflammatory role of platelets via their TLRs and Siglec receptors. Front Immunol 2015;6:83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [42].Clark SR, Ma AC, Tavener SA, McDonald B, Goodarzi Z, Kelly MM, et al. Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in septic blood. Nat Med 2007;13(4):463–9. [DOI] [PubMed] [Google Scholar]
  • [43].Etulain J, Martinod K, Wong SL, Cifuni SM, Schattner M, Wagner DD. P-selectin promotes neutrophil extracellular trap formation in mice. Blood 2015;126(2):242–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [44].Sun R, Huang J, Yang Y, Liu L, Shao Y, Li L, et al. Dysfunction of low-density neutrophils in peripheral circulation in patients with sepsis. Sci Rep 2022;12(1):685. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [45].Gisch N, Kohler T, Ulmer AJ, Müthing J, Pribyl T, Fischer K, et al. Structural reevaluation of Streptococcus pneumoniae lipoteichoic acid and new insights into its immunostimulatory potency. J Biol Chem 2013;288(22):15654–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [46].Zähringer U, Lindner B, Inamura S, Heine H, Alexander C. TLR2 - promiscuous or specific? A critical re-evaluation of a receptor expressing apparent broad specificity. Immunobiology 2008;213(3–4):205–24. [DOI] [PubMed] [Google Scholar]
  • [47].Kang JY, Nan X, Jin MS, Youn S-J, Ryu YH, Mah S, et al. Recognition of lipo-peptide patterns by Toll-like receptor 2-Toll-like receptor 6 heterodimer. Immunity 2009;31(6):873–84. [DOI] [PubMed] [Google Scholar]
  • [48].Lavagna A, Auger J-P, Dumesnil A, Roy D, Girardin SE, Gisch N, et al. Interleukin-1 signaling induced by Streptococcus suis serotype 2 is strain-dependent and contributes to bacterial clearance and inflammation during systemic disease in a mouse model of infection. Vet Res 2019;50(1):52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [49].Deng B, Wang S, Zhou P, Ding F. New insights into immune cell diversity in acute kidney injury. Cell Mol Immunol 2023;20(6):680–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [50].Dellepiane S, Leventhal JS, Cravedi P. T cells and acute kidney injury: a two-way relationship. Front Immunol 2020;11:1546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [51].El Bannoudi H, Cornwell M, Luttrell-Williams E, Engel A, Rolling C, Barrett TJ, et al. Platelet LGALS3BP as a mediator of myeloid inflammation in systemic lupus erythematosus. Arthritis Rheumatol 2023;75(5):711–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [52].Morrell CN, Pariser DN, Hilt ZT, Vega Ocasio D. The platelet Napoleon complex-small cells, but big immune regulatory functions. Ann Rev Immunol 2019;37(1):125–44. [DOI] [PubMed] [Google Scholar]
  • [53].Liu Z, Li L, Zhang H, Pang X, Qiu Z, Xiang Q, et al. Platelet factor 4(PF4) and its multiple roles in diseases. Blood Rev 2024;64:101155. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Suppl file1
NIHMS2171972-supplement-Suppl_file1.pdf (4.7MB, pdf)
supplementary table 3
NIHMS2171972-supplement-supplementary_table_3.xlsx (2.3MB, xlsx)
Suppl file 3
NIHMS2171972-supplement-Suppl_file_3.pdf (263.7KB, pdf)
supplementary table 2
NIHMS2171972-supplement-supplementary_table_2.xlsx (511.3KB, xlsx)
Suppl fil 4
NIHMS2171972-supplement-Suppl_fil_4.pdf (165.7KB, pdf)

Data Availability Statement

Clinical data for patients in our Pilot cohort are available in our prior publications [7,9], and for the 35 Validation cohort included in the Supplementary data, from the Immune Tolerance Network [26] (itntrialshare.org) of the NIAID. Datasets for RNA sequence data will be made available to online repositories. Additional primary data are available to qualified parties on request.

RESOURCES