Article Text
PDF
Abstract
Objective Circadian rhythm disruption (CRD) has been associated with inflammation and immune disorders, but its role in SLE progression is unclear. We aimed to investigate the impact of circadian rhythms on immune function and inflammation and their contribution to SLE progression to lupus nephritis (LN).
Methods This study retrospectively analysed the clinical characteristics and transcriptional profiles of 373 samples using bioinformatics and machine-learning methods. A flare risk score (FRS) was established to predict overall disease progression for patients with lupus. Mendelian randomisation was used to analyse the causal relationship between CRD and SLE progression.
Results Abnormalities in the circadian pathway were detected in patients with SLE, and lower enrichment levels suggested a disease state (normalised enrichment score=0.6714, p=0.0062). The disruption of circadian rhythms was found to be closely linked to lupus flares, with the FRS showing a strong ability to predict disease progression (area under the curve (AUC) of 5-year prediction: 0.76). The accuracy of disease prediction was improved by using a prognostic nomogram based on FRS (AUC=0.77). Additionally, Mendelian randomisation analysis revealed an inverse causal relationship between CRD and SLE (OR 0.6284 (95% CI 0.3630 to 1.0881), p=0.0485) and a positive causal relationship with glomerular disorders (OR 0.0337 (95% CI 1.634e-3 to 6.934e-1), p=0.0280).
Conclusion Our study reveals that genetic characteristics arising from CRD can serve as biomarkers for predicting the exacerbation of SLE. This highlights the crucial impact of CRD on the progression of lupus.
- systemic lupus erythematosus
- lupus nephritis
- inflammation
- risk factors
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
SLE is a complex autoimmune disease, and lupus nephritis (LN) is a severe complication.
The role of circadian rhythm disruption (CRD) in SLE and its progression to LN is not well-understood.
Previous studies have shown that CRD can exacerbate autoimmune diseases by altering immune cell function and inflammatory responses.
WHAT THIS STUDY ADDS
The study reveals a strong link between CRD and SLE progression.
It introduces the flare risk score (FRS) for predicting SLE severity and establishes a causal relationship between CRD and LN.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study suggests clinicians should incorporate circadian rhythm assessments into SLE management, potentially leading to tailored treatments that address CRDs.
The FRS introduced by the study offers a novel predictive tool for SLE, guiding more effective patient monitoring and intervention strategies.
Introduction
SLE is an autoimmune disease affecting millions of people worldwide. It is characterised by the production of autoantibodies, which can attack healthy tissues and organs such as the skin, kidneys and brain.1 Lupus nephritis (LN) is a severe complication of SLE, and it is estimated that up to 60% of patients with SLE will develop LN over time.2 LN is a type of kidney inflammation that can progress to chronic kidney disease, end-stage renal disease and death. The pathogenesis of SLE and LN involves a complex interplay between genetic, environmental and immune factors.3
Circadian rhythms are endogenous biological rhythms that regulate various physiological processes, including sleep-wake cycles, metabolism and immune function.4 These rhythms are driven by a central clock in the hypothalamus, which synchronises with environmental cues such as light and temperature. In recent years, there has been growing interest in the role of circadian rhythms in immune function and inflammation. Recently, studies have shown that circadian rhythm disruption (CRD) may exacerbate autoimmune diseases, such as rheumatoid arthritis and multiple sclerosis,5 6 by altering immune cell function and inflammatory responses. CRD has been demonstrated to decrease the number and activity of T regulatory cells, which suppress autoimmune responses and inflammation, leading to the exacerbation of autoimmune diseases.7 Moreover, CRD can alter the levels of cytokines, which are proteins that mediate immune and inflammatory responses, leading to an increase in proinflammatory cytokines and a decrease in anti-inflammatory cytokines.8–10 This imbalance in cytokine levels can exacerbate autoimmune diseases, leading to more severe symptoms and complications.
In the context of SLE, several lines of evidence suggest that CRD may contribute to disease pathogenesis and progression. For instance, patients with SLE often report disrupted sleep-wake cycles, and circadian gene expression profiles are altered in patients with SLE compared with healthy controls.11 12 Moreover, animal studies have shown that circadian clock disruption accelerates the development of lupus-like autoimmunity and renal disease in lupus-prone mice.13 14 These findings suggest that CRD may play a role in the development and progression of LN in patients with SLE.
Given the potential link between CRD and autoimmune disease, it is important to understand the effects of CRD on SLE and LN. To investigate the potential exacerbating effects of CRD on the progression of SLE to LN, our study aims to explore the relationship between CRD and disease activity in patients with SLE and LN, as well as to identify a novel avenue for intervention that may be applicable in the long-term management of SLE.
Materials and methods
Datasets and samples
The acquisition of datasets for this study was procured through the use of Gene Expression Omnibus (GEO) (http://www.ncbi.nlm.nih.gov/geo). The inclusion criteria for the selection of datasets for analysis included a meticulous selection process that factored in the presence of SLE, human gene expression profiles and pertinent clinical data that would inform the research analysis. The specificities of the four GEO datasets (GSE72326,15 GSE81622,16 GSE99967,17 GSE7279818) used in this study are comprehensively detailed in online supplemental table S1, providing essential information for further analysis. The study comprised a total of 301 patients with SLE and 72 control subjects, who were included to ensure a diverse and representative sample. Online supplemental table S2 presents a concise overview of the clinical features of cases in the study group. In figure 1, we present a flow chart of the study design, which provides a graphical representation of the study methodology and serves as a vital tool for enhancing comprehension and visualising the research process. Data processing is detailed in the online supplemental methods.
Supplemental material
Supplemental material
Flow chart of the study design. We included information from the Gene Expression Omnibus (GEO) database for a total of 373 cases of patients with lupus and normal individuals and then conducted extensive downstream analyses. Furthermore, we explored causality using MR with the data from the UK Biobank database. GSEA, gene set enrichment analysis; LASSO, Least Absolute Shrinkage and Selection Operator; LN, lupus nephritis; MR, Mendelian randomisation; SNP, single-nucleotide polymorphism; UMAP, Uniform Manifold Approximation and Projection.
Construction of the flare related gene signatures
Before the construction of a flare signature, an unsupervised consensus clustering analysis was performed to elucidate the intricate and multifaceted relationship between the lupus flare and the progression of LN. Cluster analysis was performed using ConsensusClusterPlus19 using agglomerative k-means clustering with a 1−Pearson’s correlation distance and resampling 80% of the samples for 10 repetitions. The optimal number of clusters was determined through the empirical cumulative distribution function plot. Details on model building can be found in the online supplemental methods.
Mendelian randomisation analysis
The current study used genome-wide association studies (GWAS) data from three summary statistics files. There was no sample overlap between each other. Further information on the cohorts and measures used in all the original GWAS is available in online supplemental tabel S3. Single-nucleotide polymorphisms (SNPs) were defined as instrumental variables (IVs). For the exposure datasets, we selected significant SNPs with a p value <5×10−8 from a GWAS, and these variants exhibited a robust mean F-statistic of 42.62. To evaluate causal effects, the genetic variants must satisfy the following three instrumental variable assumptions: the genetic variants indexing the exposure must be1 associated with the exposure (relevance)2; independent of confounders of the exposure-outcome relationship (exchangeability) and3 only associated with the outcome through the exposure (exclusion restriction). Prior to the main analyses, we conducted quality control procedures on the GWAS summary statistics, including harmonisation and clumping using default parameters to ensure proper alignment of effect alleles. The same clumping and harmonisation parameters were used in all univariable and multivariable Mendelian randomisation (MR) analyses. We undertook sensitivity analyses to identify and mitigate potential biases and compared outcomes across various MR methodologies, including inverse variance weighted (IVW) and MR-Egger.
Detailed information on the SNPs is presented in online supplemental table S4 and S5. Individuals participating in the GWAS were of predominantly European ancestry. We followed the Strengthening the Reporting of Observational Studies in Epidemiology-MR guidelines20 (see online supplemental table S6). Detailed MR analysis methods are provided in the online supplemental methods.
Supplemental material
Supplemental material
Results
Enrichment of the circadian rhythm pathway in SLE
To minimise the effects of red blood cells and serum, we separately analysed datasets of whole blood and peripheral blood mononuclear cells (PBMCs) (online supplemental figure S1A,B). We examined 229 patients with SLE and 30 controls using transcriptional microarray analysis of whole blood from the GSE72326 and GSE72798 datasets. We identified significant changes in transcriptional levels in SLE whole blood (online supplemental figure S2A).
We performed gene set enrichment analysis (GSEA) to identify transcriptional differences between whole blood from patients with SLE and controls. We found that the circadian rhythm pathway was significantly enriched (normalised enrichment score (NES)=0.6714, normalised p=0.0062) (online supplemental figure S2B). We also compared transcriptional differences in PBMCs between 72 patients with SLE and 42 normal individuals. Similar results were obtained from PBMC analysis (online supplemental figure S3A,B), with the circadian rhythm pathway also being significantly enriched (NES=0.7305, normalised p=0.0020). Both whole blood and PBMCs from patients with SLE displayed similar upregulation of inflammation-related genes such as MX1, HERC5 and STAT1 (online supplemental figure S2C, S3C). Additionally, we found higher expression of some genes in the LN group than in the SLE group.
Overall, our findings suggest that the circadian rhythm pathway is commonly enriched in patients with lupus compared with healthy controls.
Identification of circadian rhythm subtypes
To explore the molecular changes in the progression of lupus to LN, we partitioned all SLE cases into subpopulations based on genotyping platforms. We consolidated the GSE81622 data and removed the batch effect. Before the correction, the sample distribution of each dataset had a batch effect, with GSE72798 showing significant differences from the other two groups. However, after correction, different batches were at the same level, with obvious differences before and after correction (online supplemental figure S4).
We achieved optimal clustering stability by analysing the expression level of circadian rhythm for k=2–10 and found that k=7 yielded the best results (figure 2A and online supplemental figure S5A,B). We divided the 255 patients with SLE into 7 subtypes: cluster 1 (n=24), cluster 2 (n=59), cluster 3 (n=34), cluster 4 (n=21), cluster 5 (n=73), cluster 6 (n=21) and cluster 7 (n=22) (figure 2A and B). The Uniform Manifold Approximation and Projection analysis revealed significant differences between clusters, such as cluster 1 vs cluster 7 (p=0.24) and cluster 5 vs cluster 6 (p=6.2e-3) (online supplemental figure S5C,D). By plotting cluster-stratified Kaplan-Meier curves based on lupus flare data, we found that patients with SLE with cluster 2 and cluster 5 had a lower frequency of lupus outbreaks (figure 2B), while those with cluster 6 had the most rapid disease progression. GSEA results showed more enrichment of the JAK/STAT signalling pathway and complement and coagulation cascades pathway in cluster 6, reflecting more severe lupus (figure 2C). Additionally, the circadian rhythm pathway was more enriched in cluster 5 than in cluster 6. The heatmap (figure 2D) also showed that the expression patterns of circadian rhythm inflammation-related genes were different in cluster 5 and cluster 6. Compared with cluster 5, genes such as CSNK1D, PER1, TYK2, IFNA21 and IFNGR1 were significantly upregulated in cluster 6, while several genes, including CLOCK, CRY1, PDE12 and CD44, were significantly downregulated in cluster 6 (figure 2D and E). We used the CIBERSORT algorithm to analyse immune cell differences in cluster 6. We found that there were increased neutrophils and endothelial cells in cluster 6 (figure 2F). In summary, our results revealed that subpopulations with lower circadian rhythm pathway enrichment demonstrated a hyperinflammatory state, which may aggravate the progression of LN.
Clustering analysis based on circadian rhythm-related gene expression profiles identifies the prognosis of patients with SLE. (A) Consensus clustering matrix for k=7. (B) Kaplan-Meier survival analysis for patients with SLE in cluster (C)5 and C6. (C) Gene set enrichment analysis results of different expression levels of C5 and C6. (D) Heatmap depicting the expression levels of circadian rhythm-related genes and inflammation-related genes among C5 and C6. (E) Violin plot showing the differential circadian rhythm-related genes between C5 and C6. (F) Box plot depicting the relative immune cell infiltration differences in the C5 and C6 clusters. *P<0.05; **p<0.01; ***p<0.001 and ****p<0.0001. LN, lupus nephritis.
Construction of the flare risk score
To confirm the contribution of circadian rhythms to lupus progression, we developed an FRS by integrating lupus activity status, disease duration and enriched pathway genes. Using the Cox method, we evaluated the prognostic significance of characteristic genes in all samples, with an overall significant difference (logtest=0.0028, sctest=0.0008, Wald test=0.0220) and a C-index of 0.7772. The top related genes are illustrated in figure 3A. Using the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm, we filtered the most robust prognostic genes with an optimal lambda value of 0.08, resulting in eight genes, as shown in figure 3B and C.
Integrated analyses of lupus flare risk factors and outcomes. (A) Forest map of the hub target genes related to lupus flares, analysed by univariate Cox regression. (B) LASSO analysis of prognostic featured genes. (C) LASSO coefficient spectra of eight featured genes. (D) Patients were stratified by risk scores (upper); distribution of risk scores per patient (middle); heatmap of eight genes related to FRS. (E) Kaplan-Meier survival analysis for patients with high and low FRS. (F) ROC curves for the lupus flare risk prediction model. (G) Sankey plot of the four-cluster distribution in groups with different FRS. (H) Violin plot of the analysis of differences in SLEDAI in the low-FRS and high-FRS groups. (I) Diagram of correlation between SLEDAI and FRS. AUC, area under the curve; FRS, flare risk score; LASSO, Least Absolute Shrinkage and Selection Operator; ROC, receiver operating characteristic curve; SLEDAI, Systemic Lupus Erythematosu Disease Activity Index.
To further investigate the relationship between different risk scores and disease duration, disease activity and gene expression changes, we observed a significant decrease in patient healing with increasing risk scores. The protective genes PER2, PDE12, GBP4 and CD44 showed a downregulation trend with increasing risk scores (figure 3D). The Kaplan-Meier analysis for patients with SLE indicated that high FRS had a higher likelihood of lupus flares (figure 3E). The receiver operating characteristic (ROC) curve was generated to assess the predictive ability of the Cox regression hazard model, indicating an acceptable accuracy (area under the curve (AUC) of 1-year survival: 0.64, AUC of 3-year survival: 0.67 and AUC of 5-year survival: 0.76), as shown in figure 3F.
We calculated risk scores for each patient using the formula in the ‘Building flare predicting model’ section under ‘Supplementary Methods’ section in online supplemental file 1, and high-risk patients showed more rapid progression of lupus than low-risk patients (p=3.2e-22). The Sankey plots in figure 3G demonstrated that cluster 1, cluster 2 and cluster 3 patients were mostly concentrated in the low-FRS group, while cluster 5, cluster 6 and cluster 7 subtypes with poorer prognoses were largely distributed in the high-FRS group. We also assessed the clinical implication of the model by evaluating the association between FRS and related SLE Disease Activity Index (SLEDAI) parameters (figure 3H,I). The results showed that SLEDAI was significantly higher in the high-FRS group than in the low-FRS group, and a significant positive correlation existed between SLEDAI and FRS (p=2.1e-6, r=0.29).
We used the CIBERSORT (Cell-type Identification By Estimating Relative Subsets Of RNA Transcripts) algorithm to analyse immune cell differences in patients with lupus with high FRS. We found that patients with higher scores had significantly increased neutrophils and endothelial cells (online supplemental figure S6A). Correlation analyses showed that neutrophils (r=0.77, p<0.001) and endothelial cells (r=0.6, p<0.001) were positively correlated with the FRS, while T cells (r=−0.77, p<0.001), CD8+ T cells (r=−0.46, p<0.001), cytotoxic lymphocytes (r=−0.46, p<0.001), natural killer cells (r=−0.24, p<0.001), monocytic lineage (r=−0.47, p<0.001) and myeloid dendritic cells (r=−0.13, p<0.001) were negatively correlated (online supplemental figure S6B-I).
Establishment of the prognostic nomogram combined with clinicopathological analysis
We created a predictive nomogram (online supplemental figure S7A) to measure the probability of lupus flares in patients with SLE at 1, 3 and 5 years, providing a quantitative measure of prognostic risk assessment. To determine whether the prognostic features of the signature were independent of other traditional clinical characteristics, including age, sex and SLEDAI, we performed univariate and multivariate Cox regression analyses (online supplemental figure S7B) on the entire cohort. Our results demonstrated that age (HR 0.9781, 95% CI 0.9614 to 0.9952), SLEDAI (HR 1.112, 95% CI 1.082 to 1.144) and risk score (HR 3.619, 95% CI 1.726 to 7.588) were independent risk factors for LN progression (online supplemental figure S7B). Additionally, we performed ROC analysis to verify the predictive value of the nomogram (online supplemental figure S7C) and found that the nomogram outperformed other single clinical parameters, making it the best option for lupus prognostic assessment. We further investigated the prognostic value in patients with SLE of different ages and SLEDAI scores, finding a significant negative correlation between age and risk scores and a positive correlation between SLEDAI scores and risk scores (online supplemental figure S7D,F). Finally, we observed rapid progression of lupus to LN in younger patients and those with high SLEDAI scores (online supplemental figure S7E,G).
Association of circadian rhythm with SLE from Mendelian randomisation study
We investigated the impact of CRD on the development of LN using the MR approach. We identified 71 genome-wide significant SNPs from exposure GWAS as IVs but excluded 3 SNPs due to linkage disequilibrium and lack of corresponding outcomes. The specific details of the GWAS data are shown in online supplemental table S3. Online supplemental table S2 presents the characteristics of the remaining SNPs in the European population. We conducted a two-sample MR analysis on SNPs with an F-statistic >10 and found that genetic changes in CRD were inversely associated with the risk of lupus progressing to LN. The MR-Egger (OR 0.0802 (95% CI 0.0096 to 0.6709), p=0.0229) and IVW (OR 0.6284 (95% CI 0.3630 to 1.0881), p=0.0485) methods consistently supported this finding (figure 4A, table 1). Our evaluation of horizontal pleiotropy between IVs and outcomes indicated no significant intercept, and sensitivity analysis using the leave-one-out approach confirmed the reliability of the MR analysis results (figure 4B). There was no significant heterogeneity across estimates of included SNPs (figure 4C).
Mendelian randomisation (MR) results in the effect of circadian rhythms on SLE. (A) Scatter plot of circadian rhythm single-nucleotide polymorphism (SNP) associations versus exposure-SNP associations. (B) Forest plot to visualise the causal effect of total SNPs. (C) The funnel plot of individual SNP effects showed a symmetrical distribution.
MR results of causal links between circadian rhythms and SLE risk
Furthermore, we investigated the causal relationship between CRD and glomerular nephritis by implementing MR analysis with circadian rhythms as the exposure and glomerular nephritis as the outcome (online supplemental table S4). Our results using the IVW method showed a positive causal association between CRD and glomerular nephritis (OR 0.0337 (95% CI 1.634e-3 to 6.934e-1), p=0.0280) (online supplemental figure S8A, table 2). Sensitivity analysis using the leave-one-out approach confirmed the credibility of causal relationships (online supplemental figure S8B), and no significant horizontal pleiotropy was observed. There was also no bias or heterogeneity in the included IVs (online supplemental figure S8C). Our findings suggest a correlation between increased CRD and a higher risk of SLE, with the negative coefficient suggesting a genetic link to increased SLE susceptibility.
MR results of causal links between circadian rhythms and glomerular nephritis
Neither horizontal pleiotropy nor heterogeneity (among IVs) was detected at statistically significant levels (p value for pleiotropy test >0.05, the p value for PRESSSO (Pleiotropy Residual Sum and Outlier) global test >0.05 and p for Cochran’s Q >0.05). The specific numerical results are shown in online supplemental table S7. This indicates a correlation between increased CRD and a higher risk of SLE, with the negative coefficient suggesting a genetic link to increased SLE susceptibility.
Discussion
Our study provides new insights into the role of CRD and lupus. We found that the circadian rhythm pathway was significantly enriched in patients with SLE, especially in patients with more severe diseases. Consistent with previous studies,21–23 circadian rhythms are involved in the initiation and progression of autoimmune diseases. We constructed an FRS and a prognostic nomogram that can be used to help predict the likelihood of lupus development based on clinical characteristics and CRD. Additionally, we used MR to establish a causal relationship between CRD and SLE.
Circadian rhythm is a natural, endogenous process that affects both innate and adaptive immunity and significantly enhances proinflammatory responses and susceptibility to autoimmune disease via strictly controlling the individual cellular components of the immune system that initiate and perpetuate the inflammation pathways.4 Several core circadian genes, particularly RORγt (Retinoic acid receptor-related Orphan Receptor gamma t), which can establish a feedback loop of the circadian clock to drive the expression of PER and CLOCK, have been reported to be involved in the pathogenesis of SLE.24 25 Furthermore, animal models have confirmed the existence of a dependence relationship between the expression of CLOCK and the activity of inducible nitric oxide synthase,26 which is found to be overproduced in the context of lupus activity.27 In addition, Dan et al found that PER2 gene SNPs were related to the genetic susceptibility of SLE, indicating a potential role of PER2 in the pathogenesis of SLE.28 This result is consistent with our findings. The expression of PER1 was higher in LN and flare-activated SLE, whereas CRY1 and CLOCK were higher in these subpopulations. Moreover, in CRY double-knockout B cells, the expression of C1q was significantly downregulated compared with non-mutant phenotype controls.21 The present prompt requires a description of the significance of the expression level of the CRD gene in immune cells, as opposed to the overall peripheral serum or plasma levels, in the inflammatory severity of lupus. In fact, we found that the severity of SLE, particularly the presence of LN, was linked to more severe disruption of circadian rhythms.
SLE flares are characterised by an increase in disease activity that typically requires alternative or intensified treatment.29 In our study, we used flares as the end point to evaluate the impact of circadian rhythms on lupus disease progression. We identified eight core genes from the differentially expressed genes between the severe and resting subtypes of the disease, which we then used to construct a flare risk signature using LASSO regression. The resulting FRS demonstrated stronger prognostic capacity than traditional clinicopathological parameters alone. A high FRS score indicates a higher likelihood of lupus flare occurrence and faster progression to LN. Intensive monitoring of patients with SLE who have achieved remission is essential for the early diagnosis and prompt initiation of appropriate immunosuppressive therapy to prevent lupus flares. The FRS we developed can serve as an evaluation tool to predict flares. Furthermore, our prognostic nomogram can be used to estimate the likelihood of LN development in individual patients based on their specific characteristics. Regular monitoring for the early detection and treatment of lupus flares could significantly improve patient outcomes. This personalised approach can aid in the early detection and treatment of renal flares, improving patient outcomes.
Our study confirms previous research that links SLE to increased monocytes and greater disease severity. Furthermore, research has shown that inflammation and immune dysregulation can lead to disruptions in circadian rhythm, resulting in sleep disturbances and other disruptions. Proinflammatory cytokines have been found to suppress the activity of circadian genes, adding to the complexity of the relationship between inflammation and circadian rhythms.30 However, currently, it is unclear whether CRD triggers the onset of lupus flare-ups (the ‘egg’) or is instead a modifiable factor that exacerbates disease progression once lupus is present (the ‘chicken’). This is a common puzzle in the study of circadian rhythm disorders and immune system diseases. We further conducted an MR analysis and found that circadian rhythm, as measured by one of its indicators, sleep duration, is inversely related to SLE. However, it is positively correlated with the onset of glomerulonephritis. Although the underlying mechanisms of this relationship are not yet clear, the imbalance of CR is an important factor in the exacerbation of lupus itself. Understanding and monitoring the patterns of CRD may show potential role in the clinical strategies and patient care, particularly in different subtypes of LN involvement.
Several limitations to our study should be taken into consideration. First, we lacked data on lupus lesions, which could have provided valuable insight into disease progression. Instead, we rely solely on peripheral blood data, which may not provide a complete picture of the disease. Furthermore, detailed LN subtype information would provide insights into the pathological and immunological variations across different forms of nephritis, improving the model accuracy. Second, lupus is a complex disease that is managed with various treatment options. The effectiveness of these treatments can affect disease management and the accuracy of disease progression modelling. Therefore, our results may be influenced by the treatment methods used by the patients in the study. Third, when screening GWAS data, we did not subdivide the different traits of circadian rhythm to explore the association of SLE. Instead, we used sleep duration as one of the measures of circadian rhythm, which may weaken the persuasiveness of the conclusions. Additionally, the limited sample size and homogeneous European population may also prevent us from providing a sufficiently precise estimate for clinical practice. Although our study provides some insight into the role of circadian rhythms in lupus disease progression, having access to more comprehensive case data would allow us to make more accurate conclusions.
While our study has limitations, it has provided valuable insight into the potential impact of time changes and rhythmic genes on the progression of lupus disease. The circadian rhythm serves as a basis for monitoring SLE progression and managing other chronic illnesses, which can aid in the development of targeted treatments and management strategies. As such, our findings represent a preliminary exploration, and further research is needed to fully understand the relationship between time changes, rhythmic genes and the progression of lupus disease.
Conclusion
In conclusion, this study provides evidence that CRD can be considered as a potential biomarker for predicting the progression of SLE.
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
The present research used publicly available summary data and did not involve contact with participants; thus, no extra ethical approval was needed. The research procedures were designed to adhere to the principles and guidelines set forth in the Declaration of Helsinki.
Acknowledgments
We sincerely thank the researchers for providing their GEO database information online, and it is our pleasure to acknowledge their contributions. We acknowledge the participants and investigators of the UK Biobank and FinnGen studies.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
LS, MH and XL are joint first authors.
GW and FZ contributed equally.
Contributors FZ and GW conceived the strategy for this study. LS and MH designed all the experiments. XL, QZ and HX collected data. LS and XL performed the data analyses. JW, NW and QL assisted with the statistical analysis. LS, MH and XL contributed equally to this work. FZ and LS wrote the manuscript with comments from all authors. FZ is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis, as well as the decision to submit for publication.
Funding This research was supported by the China National Nature Science Foundation (82073928); the Leading technology foundation research project of Jiangsu province (BK20192005, BK20232035); Haihe Laboratory of Cell Ecosystem Innovation Fund (22HHXBSS00005); Nanjing Scientific and Technological Special Project for Life and Health (No. 202110006); the Project of State Key Laboratory of Natural Medicines, China Pharmaceutical University (No. SKLNMZZ202302); 'Double First-Class' University Project (CPU2018GF01, China) and Jiangsu Province '333' Project, China.
Competing interests None declared.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.