Metabolite Alterations in Autoimmune Diseases: A Systematic Review of Metabolomics Studies

Autoimmune diseases, characterized by the immune system’s loss of self-tolerance, lack definitive diagnostic tests, necessitating the search for reliable biomarkers. This systematic review aims to identify common metabolite changes across multiple autoimmune diseases. Following PRISMA guidelines, we conducted a systematic literature review by searching MEDLINE, ScienceDirect, Google Scholar, PubMed, and Scopus (Elsevier) using keywords “Metabolomics”, “Autoimmune diseases”, and “Metabolic changes”. Articles published in English up to March 2023 were included without a specific start date filter. Among 257 studies searched, 88 full-text articles met the inclusion criteria. The included articles were categorized based on analyzed biological fluids: 33 on serum, 21 on plasma, 15 on feces, 7 on urine, and 12 on other biological fluids. Each study presented different metabolites with indications of up-regulation or down-regulation when available. The current study’s findings suggest that amino acid metabolism may serve as a diagnostic biomarker for autoimmune diseases, particularly in systemic lupus erythematosus (SLE), multiple sclerosis (MS), and Crohn’s disease (CD). While other metabolic alterations were reported, it implies that autoimmune disorders trigger multi-metabolite changes rather than singular alterations. These shifts could be consequential outcomes of autoimmune disorders, representing a more complex interplay. Further studies are needed to validate the metabolomics findings associated with autoimmune diseases.


Introduction
Autoimmune diseases include a wide range of clinical disorders, including rheumatoid arthritis (RA), multiple sclerosis (MS), inflammatory bowel diseases (IBDs), autoimmune liver diseases, and systemic lupus erythematosus (SLE), characterized by loss of selftolerance by the immune system.Autoimmune diseases may be systemic or organ-specific, resulting in various complications and disabilities.Incidence of autoimmune disease varies owing to the diversity of diseases, affecting 5-10% of the population around the globe [1,2].Several factors (genetic, environmental, and epigenetic factors) are involved in the development of autoimmune diseases [3].Autoimmune diseases are poorly diagnosed owing to obscure symptoms and overlapping symptoms of various diseases.The majority of autoimmune diseases are multi-genic, with multiple susceptibility genes interacting to create the abnormal phenotype [4].Several gene variants have been discovered for autoimmune diseases; however, their relationship with disease susceptibility remains elusive.Hence, a novel approach is required for comprehensive understanding of autoimmune disease biology, especially underlying molecular mechanisms and treatment strategies.Metabolomics is an emerging technology that has drawn the attention of the scientific community in order to identify disease biomarkers due to its cost-effectiveness, short time period for repeated measurements, and very close observation of the metabolic state of patients [5,6].Metabolomics is employed to assess metabolites, which are the end products were retrieved for a comprehensive evaluation and final selection.Two reviewers (WFF and AAO) critically evaluated the quality and validity of the included studies.The first reviewer extracted the data, which were then verified by the second and third reviewers (KSA, AAO), and finally reviewed by the fourth reviewer (WFF) for accuracy and completeness.Consensus discussions were held to address any discrepancies and ensure study eligibility.The initial screening of titles was conducted by the first reviewer (AM).Subsequently, two reviewers (AM and KSA) independently assessed the title and abstracts using an eligibility checklist to exclude irrelevant studies.Full texts of potentially eligible studies were retrieved for a comprehensive evaluation and final selection.Two reviewers (WFF and AAO) critically evaluated the quality and validity of the included studies.The first reviewer extracted the data, which were then verified by the second and third reviewers (KSA, AAO), and finally reviewed by the fourth reviewer (WFF) for accuracy and completeness.Consensus discussions were held to address any discrepancies and ensure study eligibility.

Data Synthesis
The outcomes of the included studies were summarized in tables mentioning the author, analytical technique, biological fluids, models, and the number of patients and controls.Metabolic changes with respect to different biological fluids were also summarized in tables.

Risk of Bias Assessment
For both the fluid samples and the studies, the risk of bias was assessed by using the AMSTAR 2 tool.We assessed the patient recruitment process and examined the information available/lack of information about the patients.Contrasting targeted and non-targeted metabolic analysis tactics were also evaluated and, finally, fluid sample collection techniques were also taken into consideration.

Results
The systematic search of different databases of published articles produced 88 studies.General characteristics of the included studies are shown in Table 1, which includes the author, biological fluids, analytical techniques, models, and sample sizes of different groups which allowed for further categorization of metabolomics changes in the biological fluids that were analyzed in the included studies: plasma, serum, feces, urine, and other biological fluids (synovial fluids, CSF, tears, peripheral blood monocytes, in vivo white matter, peripheral blood, and lymphocytes).In 12 out of 88 studies, other fluid samples were used in contrast to plasma, urine, feces, and serum [29][30][31][32][33][34][35][36][37][38][39][40].

Serum
A total of 33 studies assessed the metabolite changes in the serum of patients.In all studies, the analysis was performed on a human model.While changes were observed in various metabolites, not a single metabolite was found to be statistically significant across all the studies.In 11 studies, aromatic amino acids (tyrosine, tryptophan, and phenylalanine) were altered [44,45,49,51,53,57,79,80,105,110,111], four of which were found to be associated with SLE and PBC.Ten studies reported an alteration in branched amino acids (leucine, isoleucine and valine) [44,45,49,52,53,59,79,98,105,110], of which four studies found an association with SLE [49,52,53,59].Alterations in fatty acids were observed in eight studies [44,50,53,57,61,104,106,108].Among the eight studies, three were related to SLE [50,53,57].The remaining metabolites that were shown to be significantly changed in the serum samples were linked to numerous metabolic pathways, including those related to lipid metabolism, ATP storage, nucleotide metabolism, oxidative stress, amino acid metabolism, and the TCA cycle (Table 2).

Plasma
A total of 21 studies assessed the metabolite changes in the plasma of patients.The analysis was conducted on human models in all studies.In all 21 plasma-based metabolite studies, not a single metabolite exhibited a consistent statistical significance across all experiments.Eleven out of twenty-one studies showed alterations in amino acid metabolism [41,46,62,[68][69][70]72,91,96,100,101], of which five were shown to be associated with MS [62,[68][69][70]72] and four with T1D [91,96,100,101].Seven out of twenty-one studies showed an alteration in aromatic amino acids [41,46,62,68,70,96,101].The alteration of metabolite levels in lipid metabolism has been reported in seven studies [41,56,65,66,93,99,101].Of these, three and two studies were associated with MS [65,66] and T1D [93,99,101].Fang et al. reported the alteration in membrane phosphoproteins and dihydroceramides [42].Åkesson et al. reported the metabolic alteration in kynurenine pathways [55].The remaining metabolites that were shown to be significantly changed in the plasma samples were linked to numerous metabolic pathways.These pathways included nucleotide metabolism, oxidative stress, amino acid metabolism, glycolytic metabolism, and the TCA cycle (Table 3).

Feces
A total of 15 studies assessed the metabolite changes in the feces of patients.The analysis was conducted on human models in all studies.Seven out of fifteen studies showed metabolic alterations in amino acid metabolism [58,74,75,77,81,85,88].Of these, five studies were linked to CD [74,77,81,85,88].Five out of fifteen studies showed an alteration in aromatic amino acids [58,74,75,78,81].Of these, three studies were linked to CD [74,78,81].Eight out of fifteen studies showed an alteration in bile acids [78,[83][84][85][86][87][88][89].Of these, five studies were linked to CD [78,[84][85][86]88].The remaining metabolites that were shown to be significantly changed in the fecal samples were linked to numerous metabolic pathways.These pathways included nucleotide metabolism, lipid metabolism, amino acid metabolism, and the TCA cycle (Table 4).Nonetheless, it is important to highlight that none of the identified metabolites exhibited consistent and significant alterations throughout all analyses.

Urine
A total of seven studies assessed the metabolite changes in the urine of patients.The analysis was conducted on human models in all studies.However, none of the identified metabolites showed consistent significant alterations across all studies.Four out of seven studies showed metabolic alterations in amino acid metabolism [54,71,92,94].Of these, two studies were linked to T1D [92,94].Three out of seven studies showed metabolic alterations in aromatic amino acids, especially tryptophan [54,71,94].One of the studies reported a decrease in trigonelline and hippurate [111].Deja et al. observed an increase in urea [92].Another study reported an increase in bile acids [109].The remaining metabolites that were shown to be significantly changed in the urine samples were linked to numerous metabolic

Other Biological Fluids
A total of 12 studies assessed the metabolite changes in the other biological fluids (synovial fluids, CSF, tears, peripheral blood monocytes, in vivo white matter, and peripheral blood and lymphocytes) of patients.The analysis was conducted on human models in all studies.In all 12 other biological fluid-based metabolite studies, not a single metabolite showed significant changes that were consistent across all experiments.Six out of twelve studies performed metabolomics analysis on CSF [32,[34][35][36]38,39].Two studies performed metabolomics analysis on synovial fluid [29,30].Two studies performed metabolomics analysis on peripheral blood lymphocytes and monocytes [31,40].One out of twelve studies performed metabolomics analysis on tears [37].One study carried out metabolomics analysis on in vivo white matter [33].Among the twelve studies, six reported metabolic changes in amino acid metabolism [30,31,33,[37][38][39].Of these, three studies were linked to MS [37][38][39].A study conducted on blood samples from patients with MS observed a decrease in glucose and lactate levels [40].There was another study conducted on CSF specimens from patients with MS, which observed a decrease in glycine, dimethylarginine, and glycerophospholipid PC-O (34:0), as well as hexoses [39].Podlecka-Piętowska et al. analyzed the metabolic alteration in CSF from MS patients and observed a decrease in acetone, choline, urea, 1,3-dimethylurate, creatinine, isoleucine, myo-inositol, leucine, 3-OH butyrate, and acetyl-CoA [38].A study conducted by Cicalini et al. reported an increase in amino acids and acylcarnitines in the tears of MS patients [37].A study performed by Herman et al. reported a decrease in 3-methoxytyramine and caffeine in the CSF of MS patients [36].Pieragostino et al. reported a decrease in phosphatydic acid and an increase in phosphatidylcholine and phosphatidylinositol in patients with MS [35].Vingara et al. analyzed the metabolic alteration in in vivo white matter and reported a decrease in lipid metabolism in patients with MS [33].Gonzalo et al. analyzed the metabolic alteration in CSF and reported a decrease in PPARγ and an increase in 8-iso-prostaglandin F2α in patients with MS [32].The remaining metabolites that were shown to be significantly changed in the other biological fluid samples were linked to numerous metabolic pathways.

Discussion
The metabolomics approach is a continuously evolving approach in the field of "omics" technology that offers a molecular view of disease pathophysiology and identifies disease biomarkers.Metabolomics also provides early diagnosis of diseases, better intervention, and monitoring of the progression of disease and the potency of treatment.The term autoimmune disease refers to a group of chronic disorders that are associated with a variety of metabolic changes that vary with the disease type.Given the absence of definitive cures for autoimmune diseases, patients are confronted with enduring illness and ongoing treatment throughout their lives.Hence, early diagnosis and recognition of various autoimmune diseases are essential to lessen disease progression and prevent painful conditions as well as co-morbidity and mortality caused by autoimmune diseases.The studies included in this systematic review analyzed the metabolic changes in various autoimmune diseases (rheumatoid arthritis, multiple sclerosis, systemic lupus erythematosus, Crohn's disease, primary sclerosing cholangitis, primary biliary cholangitis, inflammatory bowel disease, ulcerative colitis, and type 1 diabetes) in serum, plasma, feces, urine, and other biological fluids including synovial fluids, CSF, tears, in vivo white matter, and peripheral blood monocytes and lymphocytes.All studies were carried out on patients.Mass spectroscopy and nuclear magnetic resonance were used in these studies.In most of the studies, mass spectroscopy was utilized in combination with various separation techniques.Metabolites that are identified through metabolomics analysis of various biological fluids are either reported as increased or decreased in contrast to controls.Various metabolites were found to increase or decrease, belonging to various metabolic pathways including TCA, glycolytic, amino acid metabolism, ATP metabolism, nucleotide metabolism, oxidative stress, lipid metabolism, and carbohydrate metabolism.A relatively consistent change in the proportion of metabolites was observed.However, there were instances of variation between individual cases.For instance, one of the studies reported an increase in the level of phosphatidylcholine in CSF specimens [35], while some studies observed a decrease in the level of phosphatidylcholine in tears and plasma [37,90].We observed similar findings for other metabolites.It is possible that this may be due to interspecies differences in the metabolic process of patients, suggesting that further studies are required about pathophysiology and metabolomics.Further, for metabolomics findings to be applicable across species, it is imperative to identify both similarities and differences between animals and humans.Additionally, human clinical populations must be evaluated in order to confirm the utility of identified biomarker candidates in animal models.Different studies have reported metabolic changes associated with various autoimmune diseases.The metabolism of acylcarnitine and carnitine, changes in fatty acid metabolism, as well as TCA cycle metabolites have been linked to mitochondrial dysfunction [26,57,61,91,109,111,113]. Reactive oxygen species, antioxidant metabolites, glucogenic amino acid metabolites [58,114,115], and the accumulation of signaling metabolites were also reported [116].Developing metabolites associated with mitochondrial dysfunction may be a focus for future research.The metabolism of various amino acids and lipids has been found to be similar in a number of studies.However, an alteration in phosphorylcholine has only been reported in a limited number of studies.In many studies, altered amino acid metabolism and the ratio of aromatic to branched amino acids have been found to be diagnostic indicators of autoimmune diseases, particularly SLE, MS, and CD [58,117,118].However, the metabolic changes in the level of amino acids across the studies were different.This suggests that further studies are required to validate the ratio of aromatic and branched amino acids as a diagnostic indicator of autoimmune diseases.Maintaining body homeostasis requires the synthesis and degradation of proteins.Amino acid metabolism plays a crucial role in this biochemical process, including regulation of the innate and adaptive immune systems [119,120].The utilization of amino acid metabolism changes as diagnostic markers offers several compelling advantages [121].Amino acids, being stable and easily measurable in biological fluids, present a feasible and practical option for clinical assessment.Their involvement in a wide array of metabolic pathways makes them valuable indicators of physiological changes.Several studies have shown that alteration in amino acid metabolism is linked with various disease conditions, including cardiovascular disease [122], cancer [123,124], and autoimmune diseases [100][101][102][119][120][121].A case report has shown that serum levels of aspartic and glutamic acids are linked with the development of myasthenia gravis [125].Reports conducted on dietary protein restriction have demonstrated that branched amino acids contribute to promoting metabolic health [126,127].In the current study, there were changes in serum levels of aromatic amino acids in 11 studies and branched amino acids in 10.A significant alteration in amino acid metabolism was observed in 11 plasma reports.Seven studies reported significant alterations in amino acid metabolism in feces whereas four studies reported them in urine.The above findings indicated that branched amino acid metabolism may act as a diagnostic biomarker for autoimmune diseases, specifically SLE, CD, and MS.Altered amino acids in other biological fluids may be related to different stages or severity of autoimmune diseases.However, it is necessary to validate the method with a larger study sample before it can be applied to diagnostic practice, due to the multifactorial, heterogeneous, and complex nature of these diseases.Only 88 articles met the inclusion criteria for the current study.There are, however, several articles on metabolomics and autoimmune diseases that did not meet our inclusion criteria or did not appear in databases due to keywords or database limitations.Thus, these studies were not chosen for this systematic review.A study should identify and control for confounding factors (dietary habits, patient demographics, and concurrent medical conditions) since biological fluids, especially plasma, urine, and serum, all reflect systemic metabolism.These confounding factors may be involved in the metabolic alterations, indicating that statistical modeling is required for development of diagnostic biomarkers of autoimmune diseases.

Conclusions
The findings of the current study suggest that alterations in amino acid metabolism, particularly aromatic and branched amino acids, may serve as potential diagnostic biomarkers for autoimmune diseases such as SLE, MS, and CD.We also observed altered amino acid metabolism in various biological fluids including plasma, feces, urine, synovial fluids, CSF, tears, peripheral blood monocytes, in vivo white matter, peripheral blood, and lymphocytes.The study also emphasizes the complexity and heterogeneity of autoimmune disorders, since several other metabolic alterations have been reported.These alterations within various metabolic pathways were linked to energy metabolism, oxidative stress, lipid metabolism, and nucleotide metabolism, suggesting that these shifts are likely consequences of autoimmune disorders.However, biomarkers are changed owing to slight alterations in the experimental environment.Hence, metabolomics analyses must be carefully performed in the laboratory.While amino acid metabolism emerges as a promising diagnostic biomarker, the study emphasizes that further studies are required to validate the method with a larger study sample before it can be applied to diagnostic practice, due to the multifactorial, heterogeneous, and complex nature of these diseases.Researchers need to explore the correlation between the severity or stages of autoimmune disease and amino acid metabolism in different biological fluids.Furthermore, studies are required to evaluate the relationship between alterations in amino acid metabolism in various biological fluids and different autoimmune diseases.They are also required to investigate the potential therapeutic targets and conduct longitudinal studies to evaluate the efficacy of the identified biomarkers over time.

Figure 1 .
Figure 1.PRISMA flow chart illustrating the selection process of the studies.

Figure 1 .
Figure 1.PRISMA flow chart illustrating the selection process of the studies.

Table 1 .
General characteristics of the included studies.

Table 2 .
Metabolite changes found in serum.
* Based on a partial least squares discriminant analysis (PLS-DA).