Metabolomics has shown considerable potential for identifying new molecules associated with the occurrence of several pathological disorders and even for classifying patients based on the degree to which a specific organ is affected. By leveraging the advantages of this technique, we explored the usefulness of metabolomics for identifying the metabolic profiles associated with kidney damage in urine samples from patients with SLE using LC–QTOF–MS and GC–QTOF–MS to identify small molecules that may contribute to the noninvasive diagnosis of LN.
In the context of a metabolomics study utilizing urine samples for SLE and LN the issue of normalization becomes particularly complex, especially in the presence of kidney damage associated with LN. Normalizing metabolite levels is a critical step to ensure accurate comparisons and interpretations, but when kidney function is compromised, as in LN, traditional normalization methods may face challenges. Correlating metabolite levels to creatinine, a commonly used approach, may not be sufficient in diseases like LN, where creatinine itself could be influenced by renal dysfunction, so an additional normalization strategy, useful MS signal was applied14. In this study we found that even after normalization, there were many metabolites that increased or decreased between the SLE and LN groups, which suggests that the observed metabolomic changes may not solely result from kidney leakage. This could imply the presence of broader systemic metabolic alterations associated with the disease, extending beyond renal dysfunction.
LC–QTOF–MS and GC–QTOF–MS allowed for the identification of 50 differentially expressed metabolites, revealing a statistically significant difference between the SLE and LN groups. The urine samples from patients with LN yielded higher concentrations of certain metabolite groups, such as amino acids, lipids, and organic acids. We also identified an relevant statistical increase in the levels of monopalmitin and glutamic acid, as well as a decrease of glycolic acid—that could help discriminate between patients with SLE with and without kidney involvement; these metabolites are proposed as potential biomarkers for LN diagnosis.
Monopalmitin is a monoacylglycerol and is the final product of the intestinal digestion of dietary fats. Intestinal cells transform monoacylglycerols into triacylglycerols (TAGs), which are subsequently transported to the liver. Ouyang, et al. reported that lipid metabolites were considerably increased in the serum of patients with SLE compared with healthy individual15. In addition, Shin, et al. reported marked elevations in the concentrations of certain lipids, such as palmitoleic acid, oleic acid, and eicosanoic acid, in the plasma of patients with SLE compared with healthy individuals16. Some studies have also proposed the lipid nephrotoxicity hypothesis, in which TAG and fatty acid deposits in renal tissue induce glomerular diseases, such as LN17. In these cases, the inflammatory stress secondary to renal disease changes lipid homeostasis, thus increasing cholesterol absorption, decreasing cholesterol efflux, and changing cholesterol synthesis. Consequently, cholesterol accumulates in renal tissue and leads to renal failure18. Other studies have also reported an increase in sterol regulatory element-binding protein (SREBP) expression in patients with kidney damage, thus contributing to TAG and cholesterol accumulation and resulting in glomerulosclerosis and proteinuria19. Our study showed a significant increase in lysophosphatidylethanolamine glycerophospholipid (18:4) levels in patients with LN compared with patients with SLE (fold change: 297), as did 3-oxo-4-pentenoic fatty acid levels (fold change: 15.02), which supports the idea that dyslipidemia contributes to kidney damage in patients with SLE.
A hyperlipidemic environment is associated with lipid peroxidation. Oxidized low-density lipoprotein (LDL) directly damages podocytes through chemokine ligand 16 (CXCL16) by inducing the production of reactive oxygen species (ROS). Thus, high CXCL16 and oxidized LDL levels have been reported in the renal tissue of patients with different glomerular diseases20,21. Several studies have shown a close relationship between oxidative stress and inflammation and between oxidative stress and autoimmune responses in patients with SLE. These studies have emphasized that oxidized phospholipids and metabolites resulting from increased oxidative stress may act as antigenic epitopes in patients with SLE, thus enhancing excessive antibody production and significantly accelerating LN progression22,23.
For glutamic acid, glutaminase has been reported to promote Th17 cell proliferation and activation. The expression of this enzyme is regulated by the transcription factor cAMP-response element modulator (CREM), which is overexpressed in the T cells of patients with LSE and MRL/lpr mice prone to SLE24. Additionally, inhibition of this enzyme improves SLE activity in LMR/lpr mice25,26. Glutamate oxaloacetate transaminase 1 (GOT1) also enhances Th17 cell differentiation, and its selective inhibition also significantly decreases Th17 differentiation in murine T cells27,28. In our study, we found higher glutamic acid levels in participants with LN than in participants with SLE with no kidney involvement (fold change = 4292.06). These findings are consistent with previous studies that related glutamine metabolism to kidney damage. However, in addition to its role in facilitating the inflammatory process, increased glutamate levels may be explained by the fact that patients with terminal renal disease have decreased bioactivity of insulin-like growth factor (IGF-1), whose activity significantly decreases urinary glutamate levels26.
The third metabolite identified as a possible LN biomarker is glycolic acid, a secondary bile acid. Several studies have reported that high bile acid plasma levels are associated with chronic kidney disease (CKD) progression. However, decreased bile acid levels have been reported in urine samples from patients with CKD, similar to our study20. The bile acid concentration in blood and urine depends on glomerular filtration performed through apical sodium-dependent bile acid transporters, multidrug resistance associated protein 2 (MRP2), and the organic solute transporters alpha and beta. The pathological changes observed in patients with CKD—such as mesangial and endothelial cell proliferation, glomerular sclerosis, renal interstitial fibrosis, and intrarenal vascular sclerosis—reduce glomerular filtration; thus, bile acid filtration tends to decrease in patients with renal diseases29,30.
In our study, MSEA was conducted using the KEGG pathway database. This approach also allowed for the identification of the metabolic pathways associated with kidney damage progression in patients with SLE, which include primary bile acid biosynthesis, BCAA synthesis and catabolism, and tryptophan metabolism. Primary bile acid biosynthesis has been associated with the lipid profile alterations typically observed in patients with SLE31,32. In this case, high cholesterol and glycosphingolipid levels in the T-cell membrane change the composition of signaling platforms, thus favoring proinflammatory signaling19,33. Moreover, as discussed above, lipid metabolism alterations are associated with lipid nephrotoxicity and its role in LN pathophysiology34. Certain bile acids, such as deoxycholic acid, glycolic acid, ursodeoxycholic acid, and arachidonic acid, are significantly correlated with patients’ systemic lupus erythematosus disease activity index (SLEDAI) score and have shown adequate power to predict disease activity31.
In addition to their role in lipid metabolism, bile acids are signaling molecules that act through the activation of bile acid receptors. A study reported decreased levels of farnesoid X receptors (FXRs) in patients with SLE and murine MRL/lpr models of lupus and hepatic failure. In this study, the use of chenodeoxycholic acid, an agonist of FXR, suppressed the expression of inflammatory cytokines such as TNF-α, interferon-gamma (IFN-γ), and interleukin 6 (IL-6) in mice34. Another study revealed the modulatory effect of bile acids on intestinal immunity and showed that metabolites derived from lithocholic acid (LCA), 3-oxoLCA, and isoallolLCA can inhibit Th17 differentiation by directly binding with the key transcription factor retinoid-related orphan receptor γt (RORγt). Moreover, these metabolites improve FOXP3 gene expression by producing mitochondrial ROS (mitoROS), which results in regulatory T-cell expansion35,36.
The role of BCAAs in immunity is mediated through the phosphoinositide 3-kinase-protein kinase B-mammalian target of rapamycin (PI3K/AKT/mTOR) signaling pathway37. Mammalian target of rapamycin (mTOR) activity is regulated by amino acid availability, energy levels, and growth factors. In mammalian cells, mTOR forms two different complexes: mTORC1 and mTORC2. mTORC1 detects various stress signals, including the accumulation of amino acids such as leucine, isoleucine, kynurenine, and glutamine. mTORC activity increases in Th17 cells and T cells that produce IL-4, leading to the proinflammatory profile observed in patients with SLE. mTOR is required for cell differentiation toward the Th17 subtype through the induction of hypoxia-inducible factor 1α (HIF1α), which enhances glycolysis in inflammatory cells during the pseudohypoxia that typically occurs in patients with SLE38,39.
BCAA catabolism initially occurs through transamination by aminotransferases (BCAT) or decarboxylation by the branched-chain α-ketoacid dehydrogenase complex (BCKDC). Following these reactions, BCAA metabolites turn into acetyl-CoA and succinyl-CoA and participate in the tricarboxylic cycle (TCA cycle). In CD4+ T cells, BCAT negatively regulates mTOR and glycolysis. Activated T cells from mice with branched-chain amino acid aminotransferase (BCATc) deficiency show an increase in mTORC1 activation compared with the T cells from control mice39. In addition, another study reported that oral administration of ERG240, an analog of leucine, selectively inhibited BCAT1 activity, thus reducing the severity of collagen-induced arthritis and extracapillary proliferative glomerulonephritis in mice40.
To date, most of the amino acids analyzed in peripheral blood samples of patients with SLE, including gluconeogenic and ketogenic amino acids, have shown decreased levels. Ammonia is the catabolic product of amino groups; it can be converted to urea through the urea cycle and is subsequently excreted in the urine. A metabolomics study measured the metabolites associated with the urea cycle and revealed that both arginine, the immediate precursor of urea, and urea itself were increased in patients with SLE, thus suggesting increased activity in the urea cycle41,42. The kidneys play a significant role in amino acid homeostasis. In studies conducted with kidneys from LMR/lpr mice with SLE, BCAA concentrations were altered, which may suggest decreased protein synthesis, increased protein degradation, or both. The kidney is a dynamic organ with various enzymatic machinery components for amino acid catabolism and/or oxidization, particularly in the ascending limb of the loop of Henle, which provides the necessary energy for active ion transportation. The observed alterations in amino acid levels may be linked to the regulation of gene transcription, cell cycle progression, and immune and inflammatory responses. In the context of LN, these metabolic shifts could be a response to the underlying pathology, potentially reflecting an attempt by the body to regulate various cellular processes in the face of immune system dysregulation and inflammation43,44.
The LN samples analyzed in our study had higher amino acid levels than did the SLE samples. Patients with LN had higher concentrations of BCAAs and their metabolites, such as valine (SLE vs. LN: fold change = 5.81), hydroxyisoleucine (SLE vs. LN: fold change = 37.13), and aminoadipic acid (SLE vs. LN: fold change = 335). Therefore, the alteration of these metabolites and their metabolic pathways can be associated with renal disease progression in patients with SLE.
Finally, multiple mechanisms underlying the role of tryptophan in lupus progression have been proposed. Several studies have described biased metabolism of tryptophan toward the kynurenine pathway in patients with SLE, which is reflected by low tryptophan concentrations and high kynurenine levels in the serum of these patients45. Prior studies have shown that exogenous kynurenine enhances Th1 polarization of CD4+ T cells and reduces Treg cell polarization of cytotoxic T cells, thus suggesting that kynurenine promotes proinflammatory T-cell phenotypes. Moreover, kynurenine induces the activation of mTOR in human T cells46, contributing to the high level of mTOR activation typically observed in the CD4+ T cells of patients with SLE. In CD4+ T cells, active gene hypomethylation in the mTOR pathway increases the expression and activation of proinflammatory cytokines such as IFNγ and IL-1747, which are key for SLE pathophysiology. Studies conducted with mice have proven the efficacy of mTOR inhibition by rapamycin for treating LN in children. mTOR inhibition by rapamycin reduces STAT3 activation in effector T cells, as well as the migration of IL-17-producing T cells in inflamed kidneys, thus eliminating chronic inflammatory processes48.
Moreover, several tryptophan-derived metabolites, including indole-3-aldehyde, indol-3-acetic acid, 3-metilindole, tryptamine, and indoxyl sulfate, are ligands for aryl hydrocarbon receptors (AhRs). AhR signaling modulates many essential cell processes, such as cell cycle progression, apoptosis, and cell proliferation, by regulating P53, FasR, Bcl-2, and kinases of the cell cycle. AhR activation increases the regulation of genes encoding cytokines, such as IL-10, which regulate immune tolerance35,45. In patients with SLE, exposure to indoxyl sulfate, a metabolite of tryptophan degradation, increases AhR activity in the periglomerular region and in the proximal and distal renal tubules, causing renal fibrosis characterized by podocyte injury, progressive glomerular damage, and a proinflammatory phenotype associated with LN39.
The metabolic pathway of tryptophan and its particular relationship with LN have also been studied. In a metabolomics study conducted with urine samples to identify possible metabolites associated with membranous LNs, the picolinic acid–tryptophan ratio had considerable potential for LN diagnosis and classification. Therefore, the metabolites of this pathway are currently being considered potential alternative biomarkers for the noninvasive diagnosis of LN49. Our study showed a significant difference in the hydroxyanthranilic acid levels between samples from patients with LES and patients with LN (fold change = 0.66). Hydroxyanthranilic acid is produced by the metabolism of tryptophan through the kynurenine pathway, revealing the role of this pathway in LN occurrence. Several studies have reported high levels of hydroxyanthranilic acid in the plasma of patients with CKD. However, its renal excretion is limited under disease conditions; thus, its urinary concentration tends to decrease in patients with glomerular disease, as this study showed50.
Despite these findings, few studies have identified metabolic changes between SLE patients and LN patients. A study published by Guleria et al., which used 1H magnetic resonance spectroscopy, reported that patients with LN had higher serum lipid (LDL/very LDL) and creatinine levels and lower acetate levels than patients with SLE51. In general, studies aimed at determining the metabolomics profile of patients with SLE and LN based on urine samples are scarce. Although urine is an excellent option for identifying biomarkers of LN because it emerges directly from the affected renal tissue and is the most accurate biological fluid reflecting kidney dynamics4,52, a few studies have been conducted so far using this biofluid. In two of these studies, performed by Guleria et al.51, and Ganguly et al.53, a significant reduction in serum or urine levels of citrate was observed, when compared to healthy controls. In general, the behaviour of citrate in urine and serum is similar since it passes freely through the glomerulus, 60% of it being reabsorbed in the proximal tubule. Citrate is a tricarboxylic acid synthetized in the mitochondria that plays a key role in the TCA—therefore, it is reasonable to expect that when immune cells are activated and their energetic metabolism shifts, its levels in serum decrease, as the oxidative phosphorylation diminishes53. Acetate also showed significant changes in these two studies. Acetate is a product of the oxidation of fatty acids, and it was reported to be elevated in the serum samples of LN patients, compared to HC, which support the theory of disturbed lipid metabolism in LN patients43,51. Ganguly et al.53 observed higher acetate levels in LN, which exhibited a decreasing trend after treatment, possibly indicating tubular repair. Fatty acid oxidation primarily occurs in the mitochondria and peroxisomes of nephron tubules, especially the proximal tubules. Furthermore, toxins damaging the proximal tubules may lead to increased acetate excretion in urine, potentially explaining the differential behavior of acetate in serum and urine. Thanks to these study it has been demonstrated that factors such as the renal processes of filtration, reabsorption, and secretion of biomarkers, as well as the activity of active transporters in the kidneys can impact the levels of specific biomarkers in urine samples compared to serum or plasma. The behavior of these transporters may reflect changes in metabolic pathways or disruptions in kidney function, thereby establishing urine biomarkers as valuable indicators in metabolomics, especially for kidney related diseases.
In understanding results from metabolomics studies of diseases like SLE and LN, it’s crucial to remember that factors beyond the disease itself can significantly impact the levels of metabolites. Medications commonly used for these conditions, such as glucocorticoids and immunosuppressants, are a major group of these influencing factors. Glucocorticoids, which are almost always used for disease control, can unfortunately bring unintended consequences, and it has been demonstrated how they affect metabolism, including imbalances in glucose, lipids, and proteins41.
In an experiment performed by Malkawi et al54, where rats were treated with Dexamethasone at a dose of 2.5 mg/kg twice a week for 14 weeks, it was demonstrated that the serum metabolome was characterized by a decrease in phenylalanine, lysine, and arginine, while levels of tyrosine, hydroxyproline, and acylcarnitines were increased. These changes suggest that Dexamethaose may affect processes like the production of glucose from non-carbohydrates, the breakdown of proteins, and the breakdown of fat tissue. A similar study was performed in healthy volunteers, where a single 4-mg dose of dexamethasone caused major changes in over 150 plasma metabolites, characterized by an increase of blood sugar, lactate, mannose, and some amino acids, and a decrease in of cholesterol and fatty acids55. Interestingly, in another study performed with patients with Cushing’s syndrome or adrenocortical adenomas with or without hypercortisolism that were compared with hormonally normal controls, those with high cortisol had lower levels of certain fats and amino acids, but higher levels of polyamines. This suggests that some of the metabolic changes in SLE might be caused by the body’s own natural steroids, while others might not be as affected by these hormones56.
Additionally, a study found no clear link between taking glucocorticoids and changes in the blood related to oxidative stress, production of glutathione and specific inflammatory pathways. This includes substances like MDA, glutathione itself, leukotriene B4, and gamma-glutamyltransferase 1 (GGT1)41,42.
Unlike other immunosuppressant medications used for SLE, hydroxychloroquine (HCQ) seems to have a beneficial effect on cholesterol and blood sugar41. Studies have shown that HCQ induces the decrease of low density lipoprotein, triglycerides, and very low-density lipoprotein57,58. On the other hand, medications like cyclosporine and tacrolimus have been reported to worsen cholesterol and blood sugar levels59. These negative effects are dose-dependent, meaning the higher the dose, the worse the impact, and can eventually lead to hyperglycemia and hyperlipidemia60. Thankfully, azathioprine and cyclophosphamide, other SLE medications, seem to have no influence in the metabolome of SLE patients41.
Since medications can affect the results of metabolic studies in SLE and LN patients, it’s crucial to consider their influence before drawing conclusions. To get a clearer picture of how lupus itself affects metabolism, future studies should ideally involve patients who haven’t received any medications yet (drug-naive, new-onset SLE). Additionally, studying metabolism in mice with lupus might provide valuable insights in this area. In our study, due to the limitations in the data regarding medications used by SLE and LN patients, including the specific medications, dosages, and treatment durations, it is not possible to definitively determine or precisely estimate the impact of these treatments on the observed metabolic changes. However, this important consideration will be factored into the design of future research to ensure a more accurate understanding of the metabolic effects of SLE and LN independent of medication influence. Because of the complex pathophysiology of SLE and the difficulty in reaching an adequate diagnosis, metabolomics may be a good alternative approach for identifying new noninvasive biomarkers. However, multiple confounding factors may arise in metabolomics studies, which results in study limitations. We believe that factors such as participants’ concomitant medications, eating habits, alcohol use, smoking, and consumption of other substances cause the heterogeneity of the data collected during analysis. Therefore, in our upcoming studies, we aimed to control for these confounding factors and expand the sample size, including newly diagnosed patients, to compare subgroups based on clinical data, including the SLEDAI score, sex, and treatment. Moreover, validating the metabolites with the highest predictive power through targeted metabolomics utilizing standardized references, will be highly valuable for confirming our findings and proposing these metabolites as potential biomarkers for LN diagnosis.
- The Renal Warrior Project. Join Now
- Source: https://www.nature.com/articles/s41598-024-60217-0