Cholesterol and Stearamide Mediate Lymphocyte Apoptosis and Cytokine Secretion in Systemic Lupus Erythematosus

Summary

Here, we present a study that integrates two-sample Mendelian randomization, liquid chromatography-mass spectrometry (LC-MS)-based metabolomics validation, and functional assays using Systemic Lupus Erythematosus (SLE) patient-derived lymphocytes to assess apoptosis and cytokine responses following metabolite stimulation, aiming to identify lipid-related biomarkers and investigate their immunoregulatory roles in SLE.

Abstract

Alterations in serum metabolite composition have been increasingly associated with systemic lupus erythematosus (SLE), yet the direct effects of these metabolites on immune cell function remain poorly defined. This study aimed to identify peripheral blood metabolites associated with SLE and evaluate their impact on apoptosis and cytokine secretion in lymphocytes derived from SLE patients.

We performed a two-sample Mendelian randomization (MR) analysis to investigate the relationship between 565 serum metabolites and SLE, using the inverse-variance weighted model as the primary analytical approach. Significant findings were cross-validated using previously published untargeted metabolomics data based on liquid chromatography-mass spectrometry (LC-MS). In functional experiments, lymphocytes isolated from SLE patients were stimulated with target metabolites, followed by assessment of apoptosis and cytokine secretion.

MR analysis identified 28 metabolites significantly associated with SLE. Of these, cholesterol (OR = 1.462, 95% CI: 1.100-1.940, P = 0.008) and stearamide (OR = 0.125, 95% CI: 0.020-0.660, P = 0.014) were validated through LC-MS and found to be elevated in SLE patients. Receiver operating characteristic analysis demonstrated strong diagnostic performance (AUC = 0.999 for cholesterol; AUC = 1.000 for stearamide). Functionally, both metabolites induced increased apoptosis and elevated secretion of TNF-α, IFN-γ, and TGF-β1 in SLE lymphocytes.

In summary, our integrated approach combining genetic association, metabolite profiling, and functional assays reveals that cholesterol may contribute to immune dysregulation in SLE. Although stearamide showed a negative association with SLE risk in MR analysis, its in vitro effects mirrored those of cholesterol, suggesting a complex and context-dependent role. These findings underscore the importance of lipid metabolism in SLE and support further mechanistic and clinical investigation of its immunological impact.

Introduction

Systemic Lupus Erythematosus (SLE) is a chronic autoimmune disease, distinguished by complex pathogenesis and diverse clinical manifestations, that affects millions of people worldwide¹. The heterogeneity of SLE-where symptoms differ widely among individuals and evolve over time-creates significant challenges in diagnosis, monitoring disease activity, and tailoring personalized treatment plans^2,3. Although the 2019 European League Against Rheumatism and American College of Rheumatology (EULAR/ACR) SLE classification criteria have improved diagnostic accuracy⁴, these criteria do not capture all disease manifestations and often depend on invasive procedures like renal biopsy, which can be impractical and burdensome for patients.

In recent years, research on metabolites as biomarkers in SLE has been increasing, demonstrating their tremendous potential in early disease diagnosis, typing, and prognostic assessment^5,6,7. However, the translational potential of metabolite-based research in SLE remains limited by several key challenges. Firstly, small sample sizes are a major constraint⁸. Most metabolomic studies to date have involved relatively few participants, which limits their ability to capture the full clinical heterogeneity of SLE. Given that SLE can range from mild cutaneous involvement to severe renal or neurological manifestations, studies with insufficient sample diversity may suffer from poor reproducibility and limited generalizability⁹. Secondly, mechanistic depth remains lacking. While previous studies have demonstrated associations between specific metabolites and disease activity, organ damage, or particular clinical features, most findings remain descriptive in nature¹⁰. These investigations have largely focused on quantitative differences in metabolite levels and their statistical correlations with clinical indicators, without exploring the causal roles of metabolites in the immunopathology of SLE. As immune dysregulation-including aberrant lymphocyte apoptosis and imbalanced cytokine secretion-is central to SLE pathogenesis, understanding how metabolites influence these processes is crucial¹¹. Currently, few studies have directly investigated how specific metabolites modulate immune cell apoptosis or cytokine production in the context of SLE.

Here, this study aims to identify SLE-associated metabolic biomarkers through the integration of Mendelian randomization (MR) and LC-MS untargeted metabolomics analyses, and elucidate the functional roles of these candidate metabolites in modulating lymphocyte apoptosis and inflammatory cytokine production using targeted cell-based assays, as shown in Figure 1.

This is the first study to integrate MR, metabolomic validation, and in vitro functional assays to investigate the causal and mechanistic roles of lipid metabolites-particularly cholesterol and stearamide-in the pathogenesis of SLE. By combining genetic, metabolic, and experimental evidence, this work provides a comprehensive and multilayered insight into how specific lipid alterations may influence immune dysregulation in SLE. This integrative approach offers a novel framework for identifying potential metabolic biomarkers and therapeutic targets in autoimmune diseases¹².

Protocol

​This study was conducted per the 2019 EULAR/ACR classification criteria. This study was approved by Shenzhen People's Hospital (Ethics Committee Approval Number LL-KY-2019514).

Mendelian randomization (MR) analysis

GWAS data sources
Genetic association data for SLE and 565 metabolites were obtained from the OpenGWAS database (https://gwas.mrcieu.ac.uk/). The disease GWAS datasets are summarized in Table 1, while the sources and sample sizes for the metabolite GWAS data are detailed in Table 2. These datasets provided the foundation for the MR analysis to explore causal relationships between metabolites and SLE^13,14,15,16.

Selection of instrumental variables
Genetic variants associated with blood metabolites and SLE were selected as instrumental variables (IVs) for the MR analysis. The IVs were chosen based on the three core assumptions of MR: relevance, independence, and exclusion restriction¹⁷. Specifically, single nucleotide polymorphisms (SNPs) associated with metabolites at a genome-wide significance threshold of P < 5 × 10^-8 were selected¹⁸. To ensure independence among the IVs, linkage disequilibrium (LD) clumping was performed with a threshold of r² < 0.001 and a physical distance of 10,000 kb using the European population as the reference panel¹⁹. This process minimized redundancy and ensured that each IV provided unique information for the MR analysis.

Statistical analyses
The primary statistical approach for the MR analysis was the inverse-variance weighted (IVW) method, which combines the effects of multiple IVs to estimate the causal effect of metabolites on SLE²⁰. To account for potential pleiotropy, sensitivity analyses were conducted using MR-Egger regression²¹. The strength of each IV was assessed using F-statistics, with only SNPs exhibiting F > 10 retained to ensure robust instrument strength²². SNPs with lower F-statistics were excluded to avoid weak instrument bias. The final MR analyses were performed only when no significant heterogeneity or pleiotropy was detected, ensuring the reliability of the causal estimates²³.

Sensitivity analyses
Several sensitivity analyses were employed to assess the robustness of the MR findings. Heterogeneity among the IVs was evaluated using the Cochran's Q test, while pleiotropy was assessed via the MR-Egger regression²⁴. Additionally, a leave-one-out analysis was conducted to determine whether the MR results were disproportionately influenced by any single SNP²⁵. Forest plots were generated to visualize the individual and combined effects of the IVs. These analyses collectively ensured that the MR results were not biased by pleiotropic effects or outlier SNPs.

Metabolomics data source
Metabolomics data used in this study were derived from a previously published study by our research group¹⁰. In that study, we performed metabolomics analyses on serum samples from 121 SLE patients and 106 healthy individuals. Untargeted metabolomics profiling was conducted using ultra-high-performance liquid chromatography coupled with mass spectrometry (UHPLC-MS). The data were processed using standard metabolomics workflows, including peak detection, alignment, and normalization, to identify metabolites significantly associated with SLE. These metabolites were then cross-referenced with the results from the MR analysis to identify overlapping metabolites. More detailed information can be found in the previous publication¹⁰. Metabolites significantly associated in the metabolomics analysis and those with significant causal associations in the MR analysis were mapped to their respective HMDB IDs using the MetaboAnalyst platform. The common metabolites were then identified by intersecting the two sets of HMDB IDs.

Cytokine measurements and apoptosis assay
A total of 3 SLE patients were enrolled (Table 3), in accordance with the 2019 EULAR/ACR classification criteria.

Peripheral blood sample collection and lymphocytes isolation
Peripheral blood samples (3-6 mL) were collected from participants using vacuum tubes containing heparin sodium as an anticoagulant. Peripheral blood mononuclear cells (PBMCs) were isolated via Ficoll-Paque density gradient centrifugation (400-800 × g for 20-30 min). The isolated cells were washed twice with phosphate-buffered saline (PBS, HyClone) via low-speed centrifugation (250 × g for 10 min) and resuspended in complete RPMI-1640 medium supplemented with 10% fetal bovine serum. The cell density was adjusted to 1 × 10⁶ cells/mL, as verified by an automatic cell counter. The cell suspension was seeded into culture plates and pre-incubated for 2 hours at 37°C in a 5% CO₂ atmosphere to promote adherence. After pre-incubation, non-adherent cells in the culture supernatant were carefully collected and subjected to lymphocyte isolation. The collected suspension was gently layered over Ficoll-Paque and centrifuged again at 400 × g for 20 min without brake. The mononuclear cell layer at the plasma-Ficoll interface was carefully aspirated, washed twice with phosphate-buffered saline (PBS), and resuspended in RPMI-1640 medium containing 10% fetal bovine serum. Cell viability and concentration were assessed using the automatic cell counter, and the lymphocytes were subsequently used for downstream applications.

Cell treatment and culture
The cells were divided into five groups for treatment: Control group: Treated with an equal volume of DMSO. Cholesterol treatment group: Treated with cholesterol at a concentration of 1.5 mmol/L. The 1.5 mmol/L concentration of free cholesterol used in this study was selected to simulate a lipid-rich microenvironment beyond physiological levels. Circulating free cholesterol in healthy individuals typically ranges from 0.9-1.3 mmol/L. The slightly elevated concentration used in this experiments was intended to mimic pathological lipid accumulation that may occur in inflamed or metabolically disturbed tissues in SLE. Stearamide treatment groups: Treated with stearamide at concentrations of 5 µmol/L and 10 µmol/L. The concentrations of stearamide (5 and 10 µmol/L) were selected based on a prior report²⁶, which suggested that stearamide levels in the range of 5-10 µmol/L may mimic elevated lipid accumulation under pathological conditions. The intent here was to explore the pro-apoptotic and immunomodulatory effects of high-concentration lipid exposure on SLE lymphocytes, simulating inflammatory or metabolically dysregulated states. Cholesterol and stearamide were first dissolved in DMSO to prepare high-concentration stock solutions, which were then diluted in culture medium to achieve the target concentrations. The final DMSO concentration in all groups did not exceed 0.1% (v/v). After adding the respective treatments, the plates were gently shaken to ensure even distribution and incubated for 48 h at 37 °C in a 5% CO₂ environment.

Apoptosis assay
For the apoptosis experiments, 1 × 10⁶ to 3 × 10⁶ cells were collected and washed twice with pre-chilled PBS. A positive control group was prepared by resuspending cells in 500 µL of pre-chilled apoptosis-positive control solution and incubating on ice for 30 minutes, followed by another PBS wash. Experimental groups were treated as described in section 1.7, with control, cholesterol (1.5 mmol/L), and stearamide (10 µmol/L) treatments. After treatment, 1-10 × 10⁵ cells (including those from the culture supernatant) were collected and resuspended in 500 µL of 1× Binding Buffer (prepared by diluting 5× stock solution with double-distilled water). Each sample was mixed with an equal number of untreated live cells, and the volume was adjusted to 1.5 mL with pre-chilled 1× Binding Buffer. The suspension was then divided into three tubes: one for blank control and two for single-stain compensation (Annexin V-FITC and PI). To each experimental tube, 5 µL of Annexin V-FITC and 10 µL of PI staining solution were added, followed by vortexing and incubation at room temperature in the dark for 5 minutes. Flow cytometry was performed using a 488 nm excitation laser to detect Annexin V-FITC (emission at 530 nm) and PI (emission at 615 nm). The blank control tube was used to set forward scatter (FSC), side scatter (SSC), and fluorescence channel voltages, while the single-stain tubes were used for fluorescence compensation. Apoptosis was analyzed using dual-parameter scatter plots to assess the proportion of apoptotic cells^27,28.

Immune factor detection
Following the 48-h incubation, the cell suspensions were centrifuged (250 × g for 5 min) to collect the culture supernatants, which were stored at -80 °C until analysis. Cytokine quantification was performed using a commercial Human Essential Immune Response Panel, following the manufacturer's instructions with optimizations for the instrument. Briefly, 25 µL of assay buffer and 25 µL of standards (ranging from 2.4 to 10,000 pg/mL) or samples were added to each well of a microplate. The bead mixture was vortexed for 30 s and added to the wells, bringing the total reaction volume to 75 µL. The plate was incubated with shaking (800 rpm) at room temperature for 2 h in the dark. After two automated washes (1× wash buffer, 250 × g for 5 min), 25 µL of biotinylated detection antibody was added to each well and incubated for 1 h. Subsequently, 25 µL of streptavidin-phycoerythrin (SA-PE) was added for 30 min to amplify the signal.

Data acquisition was performed using a flow cytometer equipped with a 488 nm excitation laser and a 575/26 nm filter for PE detection. Prior to each experiment, calibration beads were used to optimize photomultiplier tube voltages and spectral compensation. Data were analyzed using LEGENDplex software v8.0, with cytokine concentrations determined via five-parameter logistic (5PL) curve fitting. The detection limits for each cytokine ranged from 0.68 pg/mL to 1.97 pg/mL, and experiments were conducted in duplicate to ensure precision (inter-well coefficient of variation <10%). Results were expressed as mean ±± standard deviation, and statistical differences between groups were assessed using one-way ANOVA followed by Tukey's post-hoc test in an appropriate statistical analysis software. The 48-h incubation period was selected based on the dynamic changes in cytokine secretion to capture the immune response comprehensively. Daily instrument calibration and control of DMSO concentration (≤0.1%) ensured the reliability and reproducibility of the results^29,30,31.

Results

Association between key metabolites and clinical manifestations in SLE

In this study, MR analysis of 565 metabolites with SLE identified significant causal associations. As shown in Figure 2, a total of 28 metabolites significantly associated with SLE were identified, encompassing categories such as peptides, amino acids, energy, lipids, carbohydrates, and fatty acids. Peptides such as gamma-glutamylphenylalanine and pro-hydroxy-pro, amino acids i...

Discussion

Our study demonstrates that 1.5 mmol/L cholesterol, 5 µmol/L stearamide, and 10 µmol/L stearamide induce similar patterns of apoptosis and cytokine secretion in lymphocytes from SLE patients, although variations in magnitude were observed. Specifically, both 1.5 mmol/L cholesterol and 5 µmol/L stearamide significantly increased apoptosis in SLE lymphocytes. Additionally, treatment with 1.5 mmol/L cholesterol in combination with either 5 µmol/L or 10 µmol/L stearamide further enhanced the secretio...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Annexin V-FITC/PI Apoptosis Kit	Multi Sciences	AP101	For apoptosis detection via flow cytometry
Cholesterol	Shanghai Jinpan Biotech Co., Ltd.	JP1210051	Used for stimulation experiments in PBMCs (10 g)
Counterwell Insert	Corning	CBG10001	Used to physically separate cells in coculture experiments
Flow Cytometer	Beckman Coulter	CytoFLEX	Flow cytometric analysis
GraphPad Prism Software	GraphPad		Statistical analysis and graphing
LEGENDplex Human Essential Immune Response Panel	BioLegend	740390	Multiplex detection of immune-related cytokines
Micropipette Set	Eppendorf	Various	Liquid handling and volume transfer (0.1–10 μL, 10–100 μL, 100–1000 μL)
Peripheral Blood Lymphocyte Separation Medium	Domestic Manufacturer	B1.1420A	Used for PBMC isolation (200 mL)
RPMI-1640 Medium	Gibco	C22400500BT	Culture medium for PBMCs
Stearamide	Shanghai Jinpan Biotech Co., Ltd.	1267285	Used for lipid stimulation of PBMCs (100 g)
Sterile 24-well Plate	Domestic Manufacturer	BLJ182A	For cell culture
CountStar	Shanghai Ruiyu Biotechnology	CS200	Automatic cell counter
PBS	Hyclone	10010023	Cell washing
Ficoll-Paque	Cytiva	17-5442-03	Synovial cell isolation
LEGENDplex software	BioLegend	Version 8	Flow cytometry data analysis
GraphPad Prism 9.0	GraphPad Software	Version 9	Statistical analysis