Examination of the shared genetic architecture between multiple sclerosis and systemic lupus erythematosus facilitates discovery of novel lupus risk loci.

Systemic Lupus Erythematosus (SLE) is an autoimmune disease with heterogeneous manifestations, including neurological and psychiatric symptoms. Genetic association studies in SLE have been hampered by insufficient sample size and limited power compared to many other diseases. Multiple Sclerosis (MS) is a chronic relapsing autoimmune disease of the central nervous system (CNS) that also manifests neurological and immunological features. Here, we identify a method of leveraging large-scale genome wide association studies (GWAS) in MS to identify novel genetic risk loci in SLE. Statistical genetic comparison methods including linkage disequilibrium score regression (LDSC) and cross-phenotype association analysis (CPASSOC) to identify genetic overlap in disease pathophysiology, traditional 2-sample and novel PPI-based mendelian randomization to identify causal associations and Bayesian colocalization were applied to association studies conducted in MS to facilitate discovery in the smaller, more limited datasets available for SLE. Pathway analysis using SNP-to-gene mapping identified biological networks composed of molecular pathways with causal implications for CNS disease in SLE specifically, as well as pathways likely causal of both pathologies, providing key insights for therapeutic selection.

Single-cell RNA sequencing analysis reveals the heterogeneity of IL-10 producing regulatory B cells in lupus-prone mice.

Introduction: B cells can have both pathogenic and protective roles in autoimmune diseases, including systemic lupus erythematosus (SLE). Deficiencies in the number or immunosuppressive function of IL-10 producing regulatory B cells (Bregs) can cause exacerbated autoimmune inflammation. However, the exact role of Bregs in lupus pathogenesis has not been elucidated. Methods: We carried out gene expression analysis by scRNA-seq to characterize differences in splenic Breg subsets and molecular profiles through stages of disease progression in lupus-prone mice. Transcriptome-based changes in Bregs from mice with active disease were confirmed by phenotypic analysis. Results: We found that a loss of marginal zone (MZ) lineage Bregs, an increase in plasmablast/plasma cell (PB-PC) lineage Bregs, and overall increases in inflammatory gene signatures were characteristic of active disease as compared to Bregs from the pre-disease stage. However, the frequencies of both MZ Bregs and PB-PCs expressing IL-10 were significantly decreased in active-disease mice. Conclusion: Overall, we have identified changes to the repertoire and transcriptional landscape of Breg subsets associated with active disease that provide insights into the role of Bregs in lupus pathogenesis. These results could inform the design of Breg-targeted therapies and interventions to restore Breg suppressive function in autoimmunity.

An interpretable machine learning pipeline based on transcriptomics predicts phenotypes of lupus patients.

Machine learning (ML) has the potential to identify subsets of patients with distinct phenotypes from gene expression data. However, phenotype prediction using ML has often relied on identifying important genes without a systems biology context. To address this, we created an interpretable ML approach based on blood transcriptomics to predict phenotype in systemic lupus erythematosus (SLE), a heterogeneous autoimmune disease. We employed a sequential grouped feature importance algorithm to assess the performance of gene sets, including immune and metabolic pathways and cell types, known to be abnormal in SLE in predicting disease activity and organ involvement. Gene sets related to interferon, tumor necrosis factor, the mitoribosome, and T cell activation were the best predictors of phenotype with excellent performance. These results suggest potential relationships between the molecular pathways identified in each model and manifestations of SLE. This ML approach to phenotype prediction can be applied to other diseases and tissues.

Analysis of transcriptomic features reveals molecular endotypes of SLE with clinical implications.

BACKGROUND: Systemic lupus erythematosus (SLE) is known to be clinically heterogeneous. Previous efforts to characterize subsets of SLE patients based on gene expression analysis have not been reproduced because of small sample sizes or technical problems. The aim of this study was to develop a robust patient stratification system using gene expression profiling to characterize individual lupus patients. METHODS: We employed gene set variation analysis (GSVA) of informative gene modules to identify molecular endotypes of SLE patients, machine learning (ML) to classify individual patients into molecular subsets, and logistic regression to develop a composite metric estimating the scope of immunologic perturbations. SHapley Additive ExPlanations (SHAP) revealed the impact of specific features on patient sub-setting. RESULTS: Using five datasets comprising 2183 patients, eight SLE endotypes were identified. Expanded analysis of 3166 samples in 17 datasets revealed that each endotype had unique gene enrichment patterns, but not all endotypes were observed in all datasets. ML algorithms trained on 2183 patients and tested on 983 patients not used to develop the model demonstrated effective classification into one of eight endotypes. SHAP indicated a unique array of features influential in sorting individual samples into each of the endotypes. A composite molecular score was calculated for each patient and significantly correlated with standard laboratory measures. Significant differences in clinical characteristics were associated with different endotypes, with those with the least perturbed transcriptional profile manifesting lower disease severity. The more abnormal endotypes were significantly more likely to experience a severe flare over the subsequent 52 weeks while on standard-of-care medication and specific endotypes were more likely to be clinical responders to the investigational product tested in one clinical trial analyzed (tabalumab). CONCLUSIONS: Transcriptomic profiling and ML reproducibly separated lupus patients into molecular endotypes with significant differences in clinical features, outcomes, and responsiveness to therapy. Our classification approach using a composite scoring system based on underlying molecular abnormalities has both staging and prognostic relevance.

Differential regulation of the interferon response in systemic lupus erythematosus distinguishes patients of Asian ancestry.

OBJECTIVES: Type I interferon (IFN) plays a role in the pathogenesis of systemic lupus erythematosus (SLE), but insufficient attention has been directed to the differences in IFN responses between ancestral populations. Here, we explored the expression of the interferon gene signatures (IGSs) in SLE patients of European ancestry (EA) and Asian ancestry (AsA). METHODS: We used gene set variation analysis with multiple IGS encompassing the response to both type 1 and type 2 IFN in isolated CD14+ monocytes, CD19+B cells, CD4+T cells and Natural Killer (NK) cells from patients with SLE stratified by self-identified ancestry. The expression of genes upstream of the IGS and influenced by lupus-associated risk alleles was also examined. Lastly, we employed machine learning (ML) models to assess the most important features classifying patients by disease activity. RESULTS: AsA patients with SLE exhibited greater enrichment in the IFN core and IFNA2 IGS compared with EA patients in all cell types examined and, in the presence and absence of autoantibodies. Overall, AsA patients with SLE demonstrated higher expression of genes upstream of the IGS than EA counterparts. ML with feature importance analysis indicated that IGS expression in NK cells, anti-dsDNA, complement levels and AsA status contributed to disease activity. CONCLUSIONS: AsA patients with SLE exhibited higher IGS than EA patients in all cell types regardless of autoantibody status, with enhanced expression of genetically associated genes upstream of the IGS potentially contributing. AsA, along with the IGS in NK cells, anti-dsDNA and complement, independently influenced SLE disease activity.

Glucose oxidation-dependent survival of activated B cells provides a putative novel therapeutic target for lupus treatment.

Aberrant metabolic demand is observed in immune/inflammatory disorders, yet the role in pathogenesis remains unclear. Here, we discover that in lupus, activated B cells, including germinal center B (GCB) cells, have remarkably high glycolytic requirement for survival over T cell populations, as demonstrated by increased metabolic activity in lupus-activated B cells compared to immunization-induced cells. The augmented reliance on glucose oxidation makes GCB cells vulnerable to mitochondrial ROS-induced oxidative stress and apoptosis. Short-term glycolysis inhibition selectively reduces pathogenic activated B in lupus-prone mice, extending their lifespan, without affecting T follicular helper cells. Particularly, BCMA-expressing GCB cells rely heavily on glucose oxidation. Depleting BCMA-expressing activated B cells with APRIL-based CAR-T cells significantly prolongs the lifespan of mice with severe autoimmune disease. These results reveal that glycolysis-dependent activated B and GCB cells, especially those expressing BCMA, are potentially key lupus mediators, and could be targeted to improve disease outcomes.

Molecular pathways identified from single nucleotide polymorphisms demonstrate mechanistic differences in systemic lupus erythematosus patients of Asian and European ancestry.

Systemic lupus erythematosus (SLE) is a multi-organ autoimmune disorder with a prominent genetic component. Individuals of Asian-Ancestry (AsA) disproportionately experience more severe SLE compared to individuals of European-Ancestry (EA), including increased renal involvement and tissue damage. However, the mechanisms underlying elevated severity in the AsA population remain unclear. Here, we utilized available gene expression data and genotype data based on all non-HLA SNP associations in EA and AsA SLE patients detected using the Immunochip genotyping array. We identified 2778 ancestry-specific and 327 trans-ancestry SLE-risk polymorphisms. Genetic associations were examined using connectivity mapping and gene signatures based on predicted biological pathways and were used to interrogate gene expression datasets. SLE-associated pathways in AsA patients included elevated oxidative stress, altered metabolism and mitochondrial dysfunction, whereas SLE-associated pathways in EA patients included a robust interferon response (type I and II) related to enhanced cytosolic nucleic acid sensing and signaling. An independent dataset derived from summary genome-wide association data in an AsA cohort was interrogated and identified similar molecular pathways. Finally, gene expression data from AsA SLE patients corroborated the molecular pathways predicted by SNP associations. Identifying ancestry-related molecular pathways predicted by genetic SLE risk may help to disentangle the population differences in clinical severity that impact AsA and EA individuals with SLE.

Molecular mechanisms governing the progression of nephritis in lupus prone mice and human lupus patients.

Introduction: Pathologic inflammation is a major driver of kidney damage in lupus nephritis (LN), but the immune mechanisms of disease progression and risk factors for end organ damage are poorly understood. Methods: To characterize molecular profiles through the development of LN, we carried out gene expression analysis of microdissected kidneys from lupus-prone NZM2328 mice. We examined male mice and the congenic NZM2328.R27 strain as a means to define mechanisms associated with resistance to chronic nephritis. Gene expression profiles in lupus mice were compared with those in human LN. Results: NZM2328 mice exhibited progress from acute to transitional and then to chronic glomerulonephritis (GN). Each stage manifested a unique molecular profile. Neither male mice nor R27 mice progressed past the acute GN stage, with the former exhibiting minimal immune infiltration and the latter enrichment of immunoregulatory gene signatures in conjunction with robust kidney tubule cell profiles indicative of resistance to cellular damage. The gene expression profiles of human LN were similar to those noted in the NZM2328 mouse suggesting comparable stages of LN progression. Conclusions: Overall, this work provides a comprehensive examination of the immune processes involved in progression of murine LN and thus contributes to our understanding of the risk factors for end-stage renal disease. In addition, this work presents a foundation for improved classification of LN and illustrates the applicability of murine models to identify the stages of human disease.

Abnormalities in intron retention characterize patients with systemic lupus erythematosus.

Regulation of intron retention (IR), a form of alternative splicing, is a newly recognized checkpoint in gene expression. Since there are numerous abnormalities in gene expression in the prototypic autoimmune disease systemic lupus erythematosus (SLE), we sought to determine whether IR was intact in patients with this disease. We, therefore, studied global gene expression and IR patterns of lymphocytes in SLE patients. We analyzed RNA-seq data from peripheral blood T cell samples from 14 patients suffering from systemic lupus erythematosus (SLE) and 4 healthy controls and a second, independent data set of RNA-seq data from B cells from16 SLE patients and 4 healthy controls. We identified intron retention levels from 26,372 well annotated genes as well as differential gene expression and tested for differences between cases and controls using unbiased hierarchical clustering and principal component analysis. We followed with gene-disease enrichment analysis and gene-ontology enrichment analysis. Finally, we then tested for significant differences in intron retention between cases and controls both globally and with respect to specific genes. Overall decreased IR was found in T cells from one cohort and B cells from another cohort of patients with SLE and was associated with increased expression of numerous genes, including those encoding spliceosome components. Different introns within the same gene displayed both up- and down-regulated retention profiles indicating a complex regulatory mechanism. These results indicate that decreased IR in immune cells is characteristic of patients with active SLE and may contribute to the abnormal expression of specific genes in this autoimmune disease.

Classification of COVID-19 Patients into Clinically Relevant Subsets by a Novel Machine Learning Pipeline Using Transcriptomic Features.

The persistent impact of the COVID-19 pandemic and heterogeneity in disease manifestations point to a need for innovative approaches to identify drivers of immune pathology and predict whether infected patients will present with mild/moderate or severe disease. We have developed a novel iterative machine learning pipeline that utilizes gene enrichment profiles from blood transcriptome data to stratify COVID-19 patients based on disease severity and differentiate severe COVID cases from other patients with acute hypoxic respiratory failure. The pattern of gene module enrichment in COVID-19 patients overall reflected broad cellular expansion and metabolic dysfunction, whereas increased neutrophils, activated B cells, T-cell lymphopenia, and proinflammatory cytokine production were specific to severe COVID patients. Using this pipeline, we also identified small blood gene signatures indicative of COVID-19 diagnosis and severity that could be used as biomarker panels in the clinical setting.

Hypovitaminosis A Drives the Progression of Tubulointerstitial Lupus Nephritis through Potentiating Predisease Cellular Autoreactivity.

Vitamin A (VA) deficiency (VAD) is observed in both humans and mice with lupus nephritis. However, whether VAD is a driving factor for accelerated progression of lupus nephritis is unclear. In this study, we investigated the effect of VAD on the progression of lupus nephritis in a lupus-prone mouse model, MRL/lpr. We initiated VAD either during gestation or after weaning to reveal a potential time-dependent effect. We found exacerbated lupus nephritis at ∼15 wk of age with both types of VAD that provoked tubulointerstitial nephritis leading to renal failure. This was concomitant with significantly higher mortality in all VAD mice. Importantly, restoration of VA levels after weaning reversed VAD-induced mortality. These results suggest VAD-driven acceleration of tubulointerstitial lupus nephritis. Mechanistically, at the earlier time point of 7 wk of age and before the onset of clinical lupus nephritis, continued VAD (from gestation until postweaning) enhanced plasma cell activation and augmented their autoantibody production, while also increasing the expansion of T lymphocytes that could promote plasma cell autoreactivity. Moreover, continued VAD increased the renal infiltration of plasmacytoid dendritic cells. VAD initiated after weaning, in contrast, showed modest effects on autoantibodies and renal plasmacytoid dendritic cells that were not statistically significant. Remarkably, analysis of gene expression in human kidney revealed that the retinoic acid pathway was decreased in the tubulointerstitial region of lupus nephritis, supporting our findings in MRL/lpr mice. Future studies will elucidate the underlying mechanisms of how VAD modulates cellular functions to exacerbate tubulointerstitial lupus nephritis.

Mendelian randomization and pathway analysis demonstrate shared genetic associations between lupus and coronary artery disease.

Coronary artery disease (CAD) is a leading cause of death in patients with systemic lupus erythematosus (SLE). Despite clinical evidence supporting an association between SLE and CAD, pleiotropy-adjusted genetic association studies are limited and focus on only a few common risk loci. Here, we identify a net positive causal estimate of SLE-associated non-HLA SNPs on CAD by traditional Mendelian randomization (MR) approaches. Pathway analysis using SNP-to-gene mapping followed by unsupervised clustering based on protein-protein interactions (PPIs) identifies biological networks composed of positive and negative causal sets of genes. In addition, we confirm the casual effects of specific SNP-to-gene modules on CAD using only SNP mapping to each PPI-defined functional gene set as instrumental variables. This PPI-based MR approach elucidates various molecular pathways with causal implications between SLE and CAD and identifies biological pathways likely causative of both pathologies, revealing known and novel therapeutic interventions for managing CAD in SLE.

COVID-19 patients exhibit unique transcriptional signatures indicative of disease severity.

COVID-19 manifests a spectrum of respiratory symptoms, with the more severe often requiring hospitalization. To identify markers for disease progression, we analyzed longitudinal gene expression data from patients with confirmed SARS-CoV-2 infection admitted to the intensive care unit (ICU) for acute hypoxic respiratory failure (AHRF) as well as other ICU patients with or without AHRF and correlated results of gene set enrichment analysis with clinical features. The results were then compared with a second dataset of COVID-19 patients separated by disease stage and severity. Transcriptomic analysis revealed that enrichment of plasma cells (PCs) was characteristic of all COVID-19 patients whereas enrichment of interferon (IFN) and neutrophil gene signatures was specific to patients requiring hospitalization. Furthermore, gene expression results were used to divide AHRF COVID-19 patients into 2 groups with differences in immune profiles and clinical features indicative of severe disease. Thus, transcriptomic analysis reveals gene signatures unique to COVID-19 patients and provides opportunities for identification of the most at-risk individuals.

Machine learning reveals distinct gene signature profiles in lesional and nonlesional regions of inflammatory skin diseases.

Analysis of gene expression from cutaneous lupus erythematosus, psoriasis, atopic dermatitis, and systemic sclerosis using gene set variation analysis (GSVA) revealed that lesional samples from each condition had unique features, but all four diseases displayed common enrichment in multiple inflammatory signatures. These findings were confirmed by both classification and regression tree analysis and machine learning (ML) models. Nonlesional samples from each disease also differed from normal samples and each other by ML. Notably, the features used in classification of nonlesional disease were more distinct than their lesional counterparts, and GSVA confirmed unique features of nonlesional disease. These data show that lesional and nonlesional skin samples from inflammatory skin diseases have unique profiles of gene expression abnormalities, especially in nonlesional skin, and suggest a model in which disease-specific abnormalities in "prelesional" skin may permit environmental stimuli to trigger inflammatory responses leading to both the unique and shared manifestations of each disease.

Deconvoluting the heterogeneity of SLE: The contribution of ancestry.

Systemic lupus erythematosus (SLE) is a multiorgan autoimmune disorder with a prominent genetic component. Evidence has shown that individuals of non-European ancestry experience the disease more severely, exhibiting an increased incidence of cardiovascular disease, renal involvement, and tissue damage compared with European ancestry populations. Furthermore, there seems to be variability in the response of individuals within different ancestral groups to standard medications, including cyclophosphamide, mycophenolate, rituximab, and belimumab. Although the widespread application of candidate gene, Immunochip, and genome-wide association studies has contributed to our understanding of the link between genetic variation (typically single nucleotide polymorphisms) and SLE, despite decades of research it is still unclear why ancestry remains a key determinant of poorer outcome in non-European-ancestry patients with SLE. Here, we will discuss the impact of ancestry on SLE disease burden in patients from diverse backgrounds and highlight how research efforts using novel bioinformatic and pathway-based approaches have begun to disentangle the complex genetic architecture linking ancestry to SLE susceptibility. Finally, we will illustrate how genomic and gene expression analyses can be combined to help identify novel molecular pathways and drug candidates that might uniquely impact SLE among different ancestral populations.

Nucleic Acid-Sensing and Interferon-Inducible Pathways Show Differential Methylation in MZ Twins Discordant for Lupus and Overexpression in Independent Lupus Samples: Implications for Pathogenic Mechanism and Drug Targeting.

Systemic lupus erythematosus (SLE) is a chronic, multisystem, autoimmune inflammatory disease with genomic and non-genomic contributions to risk. We hypothesize that epigenetic factors are a significant contributor to SLE risk and may be informative for identifying pathogenic mechanisms and therapeutic targets. To test this hypothesis while controlling for genetic background, we performed an epigenome-wide analysis of DNA methylation in genomic DNA from whole blood in three pairs of female monozygotic (MZ) twins of European ancestry, discordant for SLE. Results were replicated on the same array in four cell types from a set of four Danish female MZ twin pairs discordant for SLE. Genes implicated by the epigenetic analyses were then evaluated in 10 independent SLE gene expression datasets from the Gene Expression Omnibus (GEO). There were 59 differentially methylated loci between unaffected and affected MZ twins in whole blood, including 11 novel loci. All but two of these loci were hypomethylated in the SLE twins relative to the unaffected twins. The genes harboring these hypomethylated loci exhibited increased expression in multiple independent datasets of SLE patients. This pattern was largely consistent regardless of disease activity, cell type, or renal tissue type. The genes proximal to CpGs exhibiting differential methylation (DM) in the SLE-discordant MZ twins and exhibiting differential expression (DE) in independent SLE GEO cohorts (DM-DE genes) clustered into two pathways: the nucleic acid-sensing pathway and the type I interferon pathway. The DM-DE genes were also informatically queried for potential gene-drug interactions, yielding a list of 41 drugs including a known SLE therapy. The DM-DE genes delineate two important biologic pathways that are not only reflective of the heterogeneity of SLE but may also correlate with distinct IFN responses that depend on the source, type, and location of nucleic acid molecules and the activated receptors in individual patients. Cell- and tissue-specific analyses will be critical to the understanding of genetic factors dysregulating the nucleic acid-sensing and IFN pathways and whether these factors could be appropriate targets for therapeutic intervention.

An introduction to machine learning and analysis of its use in rheumatic diseases.

Machine learning (ML) is a computerized analytical technique that is being increasingly employed in biomedicine. ML often provides an advantage over explicitly programmed strategies in the analysis of multidimensional information by recognizing relationships in the data that were not previously appreciated. As such, the use of ML in rheumatology is increasing, and numerous studies have employed ML to classify patients with rheumatic autoimmune inflammatory diseases (RAIDs) from medical records and imaging, biometric or gene expression data. However, these studies are limited by sample size, the accuracy of sample labelling, and absence of datasets for external validation. In addition, there is potential for ML models to overfit or underfit the data and, thereby, these models might produce results that cannot be replicated in an unrelated dataset. In this Review, we introduce the basic principles of ML and discuss its current strengths and weaknesses in the classification of patients with RAIDs. Moreover, we highlight the successful analysis of the same type of input data (for example, medical records) with different algorithms, illustrating the potential plasticity of this analytical approach. Altogether, a better understanding of ML and the future application of advanced analytical techniques based on this approach, coupled with the increasing availability of biomedical data, may facilitate the development of meaningful precision medicine for patients with RAIDs.

Transcriptomics data: pointing the way to subclassification and personalized medicine in systemic lupus erythematosus.

PURPOSE OF REVIEW: To summarize recent studies stratifying SLE patients into subgroups based on gene expression profiling and suggest future improvements for employing transcriptomic data to foster precision medicine. RECENT FINDINGS: Bioinformatic & machine learning pipelines have been employed to dissect the transcriptomic heterogeneity of lupus patients and identify more homogenous subgroups. Some examples include the use of unsupervised random forest and k-means clustering to separate adult SLE patients into seven clusters and hierarchical clustering of single-cell RNA-sequencing (scRNA-seq) of immune cells yielding four clusters in a cohort of adult SLE and pediatric SLE participants. Random forest classification of bulk RNA-seq data from sorted blood cells enabled prediction of high or low disease activity in European and Asian SLE patients. Inferred transcription factor activity stratified adult and pediatric SLE into two subgroups. SUMMARY: Several different endotypes of SLE patients with differing molecular profiles have been reported but a global consensus of clinically actionable groups has not been reached. Moreover, heterogeneity between datasets, reproducibility of predictions as well as the most effective classification approach have not been resolved. Nevertheless, gene expression-based precision medicine remains an attractive option to subset lupus patients.

Altered expression of genes controlling metabolism characterizes the tissue response to immune injury in lupus.

To compare lupus pathogenesis in disparate tissues, we analyzed gene expression profiles of human discoid lupus erythematosus (DLE) and lupus nephritis (LN). We found common increases in myeloid cell-defining gene sets and decreases in genes controlling glucose and lipid metabolism in lupus-affected skin and kidney. Regression models in DLE indicated increased glycolysis was correlated with keratinocyte, endothelial, and inflammatory cell transcripts, and decreased tricarboxylic (TCA) cycle genes were correlated with the keratinocyte signature. In LN, regression models demonstrated decreased glycolysis and TCA cycle genes were correlated with increased endothelial or decreased kidney cell transcripts, respectively. Less severe glomerular LN exhibited similar alterations in metabolism and tissue cell transcripts before monocyte/myeloid cell infiltration in some patients. Additionally, changes to mitochondrial and peroxisomal transcripts were associated with specific cells rather than global signal changes. Examination of murine LN gene expression demonstrated metabolic changes were not driven by acute exposure to type I interferon and could be restored after immunosuppression. Finally, expression of HAVCR1, a tubule damage marker, was negatively correlated with the TCA cycle signature in LN models. These results indicate that altered metabolic dysfunction is a common, reversible change in lupus-affected tissues and appears to reflect damage downstream of immunologic processes.