Multi-omics profiling reveals distinct molecular identities across four autoimmune diseases

Evotec查看公司主页 has completed a multi-omics study that maps the distinct molecular identities of four autoimmune diseases, revealing both disease-specific and shared biological processes while uncovering a range of candidate biomarkers. Leveraging a combination of metabolomics, lipidomics, and proteomics, the study profiled plasma samples from 76 individuals—comprising healthy volunteers and patients with systemic lupus erythematosus (搜索)搜索该疾病 (SLE), multiple sclerosis (搜索)搜索该疾病 (MS), Crohn's disease (搜索)搜索该疾病 (CD), and ulcerative colitis (搜索)搜索该疾病 (UC)—to better understand the molecular mechanisms driving these related yet distinct disorders.

The work addresses a central challenge in autoimmune disease: patients diagnosed with the same condition often present with very different symptoms, treatment responses, and disease trajectories. This heterogeneity implies the presence of numerous underlying disease endotypes that are not fully captured by current clinical classifications, creating an urgent demand for biomarkers able to improve patient stratification, support earlier diagnosis, predict disease progression, and facilitate more precise monitoring of therapeutic response.

The molecular landscape of autoimmune disease

Untargeted metabolomics and lipidomics were used alongside proteomics profiles, generating multiple immunoassay platforms to provide insight into disease-related biological processes. The study initially sought to determine whether molecular profiling could distinguish autoimmune diseases from one another.

Unsupervised analyses showed that SLE exhibited the most distinguishing molecular signature, with SLE patients forming a distinct cluster that highlighted extensive metabolic dysregulation and the disease's systemic nature. CD and UC showed substantial overlap, a result anticipated from a biological perspective given that both conditions share common inflammatory mechanisms on the inflammatory bowel disease (IBD) spectrum. MS patients clustered much closer to healthy volunteers, reflecting the fact that individuals in the cohort were receiving treatment and in a relapsing-remitting phase of disease.

These results confirmed that metabolomics and lipidomics are sensitive enough to capture both disease-specific biology and common inflammatory processes across autoimmune disorders.

Lupus: A state of systemic metabolic stress

SLE exhibited the strongest perturbation among all diseases studied, with differential analysis identifying several considerably dysregulated metabolites and lipids—a finding suggesting profound metabolic reprogramming. Mapping these alterations onto biological pathways revealed links to mitochondrial dysfunction, oxidative stress, protein catabolism, immunometabolism rewiring, and metabolic demand associated with chronic immune activation. Signals linked to neuroimmune pathways further emphasized the disease's multisystem nature.

Taken together, these findings describe a condition characterized by redox imbalance, systemic immune activation, and widespread metabolic stress, offering potentially valuable leads for future functional and mechanistic studies.

Multiple sclerosis: Subtle signals, strong biological relevance

Overall metabolic disruption in MS was more limited than in SLE, but the identified alterations were highly consistent with known disease biology, suggesting the presence of active pathological processes despite treatment. Key pathways included myelin lipid turnover, sphingolipid metabolism, neuroinflammation, histidine-related immune regulation, and mitochondrial dysfunction.

These signatures suggest that biologically meaningful molecular alterations remain detectable, although relapsing-remitting patients could potentially appear metabolically similar to healthy individuals at a global level. These features may prove especially beneficial as biomarkers of disease progression, activity, or relapse risk.

The gut-microbiome axis in inflammatory bowel disease

The investigated inflammatory bowel diseases revealed a specific molecular landscape centered on host-microbiome interactions and intestinal biology. CD showed altered metabolites linked to host-microbiome co-metabolism, bile acid dysregulation, protein catabolism, and inflammatory lipid signaling associated with intestinal barrier dysfunction—signatures consistent with microbial dysbiosis and chronic intestinal inflammation, both key drivers of disease pathogenesis.

UC shared a number of these features while exhibiting further alterations involving epithelial energy stress, microbiome-derived tryptophan metabolism, and mucosal inflammation. The resulting metabolic signatures reflected altered interactions between the intestinal microbiota and the host immune system, as well as epithelial barrier dysfunction.

These findings highlight one of metabolomics' distinct strengths: its capacity to capture biological information originating from both the microbiome and the host, providing a highly beneficial window into the sophisticated molecular dialog taking place in IBD.

Disease-specific remodeling versus a common autoimmune signature

Cross-disease comparisons showed minimal overlap among significantly altered metabolites. A small common core was believed to reflect shared immunometabolic and inflammatory processes, but most molecular changes were found to be disease-specific. Hierarchical clustering verified this observation, with disease groups displaying clearly differentiated abundance patterns: SLE exhibited the most distinct profile, the two IBDs were distinguishable despite being clustered closer together, and MS occupied an intermediate position closer to that of healthy volunteers.

These results support the assertion that autoimmune diseases are driven by distinct underlying biological mechanisms despite sharing common immunological features.

Predictive biomarker signatures identified via machine learning

Regularized regression approaches were applied to determine whether these molecular signatures could support disease classification. Elastic net modeling enabled the identification of compact metabolite and lipid panels that distinguished disease groups from healthy volunteers, with classification models achieving F1 scores from approximately 0.84 to nearly 1, depending on the disease in question.

Selected features were consistent with biological pathways identified via differential analysis. For example, sphingolipids were prominent markers in MS; SLE signatures were driven by amino acid and redox metabolism; and IBDs were characterized by microbiome-associated metabolites.

Revealing novel biomarker candidates via multi-omics integration

Metabolomics, lipidomics, and proteomics data were integrated using the DIABLO multi-omics framework to further expand biological insight. This approach revealed coordinated molecular networks rather than isolated markers, highlighting metabolite, lipid, and protein networks that collectively characterized each disease. The resulting multi-omics signature featured 230 metabolites and lipids, along with 30 proteins.

A large number of integrated signatures reinforced known disease mechanisms identified in the study, while also uncovering historically unreported combinations of proteins and metabolites. Of the top 60 identified candidate biomarkers, 27 had been previously reported, while 33 represented potentially novel candidate biomarkers.

These findings showcase the benefits of integrating multiple molecular layers when investigating complex diseases. The study also validated a ready-to-use workflow suitable for multi-omics data generation, analysis, and integration. Though autoimmune diseases served as the use case, the same framework can be readily applied across therapeutic areas to characterize disease endotypes, identify biomarkers, and generate novel biological insights.

The study was produced from materials originally authored by Michaël Méret, Senior Research Scientist of Metabolomics at Evotec查看公司主页.