Metabolic determinants of autoimmune kidney diseases

Abstract

Autoantibody-driven autoimmune diseases, such as systemic lupus erythematosus, frequently affect organs such as the kidney. The differentiation and function of pathogenic immune cells that drive these diseases are in part controlled by their metabolic programming. For diseases that affect the kidney, the response of kidney cells to immune-mediated injury is also in part controlled by metabolic changes. Immune cells that promote the production of autoantibodies and/or infiltrate the kidney in lupus nephritis are sustained by enhanced glycolysis and mitochondrial oxidation. These metabolic processes are also enhanced in the mesangial and glomerular endothelial cells of patients with lupus nephritis and animal models of lupus nephritis, which may contribute to tissue injury. Similar alterations in metabolic processes might be involved in other autoimmune diseases that affect the kidney, including IgA nephropathy and ANCA-associated vasculitis. Insights into metabolic changes that occur in the context of autoimmune-mediated kidney diseases might have therapeutic implications. Despite the complexity of metabolic alterations presented by specific immune and renal cells in these autoimmune diseases, targeting of glycolysis, mitochondrial oxidation or iron metabolism could offer novel opportunities to enhance existing treatments for autoimmune-mediated kidney injury.

Introduction

Cellular metabolism is the process by which a cell obtains and uses its energy to sustain its functions. Most cell types rely on specific metabolic processes for their differentiation and effector functions, which can change in response to nutrient availability, energy requirements and cell activation status. The immune system is composed of a complex array of cells that adapt rapidly to respond to infectious or tumour challenges before returning to a homeostatic state to prevent tissue damage once the pathogen is cleared. This rapid response requires immune cells to change their effector functions, such as their ability to migrate to the infection site or produce cytokines. T cells differentiate into effector subsets that are specialized to the type of pathogen encountered. B cells undergo a complex differentiation process in germinal centres to produce high affinity class-switched antibodies that are secreted following their differentiation into plasma cells. Studies first performed in T cells¹ and macrophages² showed that their activation and differentiation requires profound shifts in metabolic programmes, largely towards an increase in glycolysis replacing the mitochondrial metabolism that sustains quiescent cells. Subsequently, multiple studies have shown that metabolic reprograming drives the fate and function of every type of immune cell (Box 1).

Box 1 |. Key metabolic pathways.

Glucose, glutamine and the β-oxidation of fatty acids are used to generate ATP and metabolites such as nucleotides to support proliferation and cellular effector functions. In addition, glucose metabolites, fatty acids and amino acids can act as autocrine or paracrine signalling molecules that alter the fate and functions of immune cells²²³. Mitochondria are the central hub of cellular metabolism, and many studies have identified mitochondrial dysfunction as a driver of autoimmune pathology, including systemic lupus erythematosus (SLE)⁵ . Finally, the mTOR complex integrates signals that relate to the metabolic status of the cell to promote translation and anabolic processes when activated and to block these processes and promote autophagy when its repressive counterpart AMP-activated protein kinase is activated. Overactivation of mTOR is a common finding in autoimmune pathologies, including SLE¹⁰⁵.

• Glycolysis: glucose is converted into pyruvate, producing a limited amount of ATP and precursors for biosynthesis. Glycolysis also feeds the tricarboxylic acid (TCA) cycle from pyruvate to generate acetyl-CoA. In addition, several glycolytic enzymes regulate immune functions through non-enzymatic functions.

• Oxidative phosphorylation: an efficient mitochondrial pathway that generates ATP from NADH and FADH₂, produced by the TCA cycle from substrates such as pyruvate (derived from glycolysis), glutamine and fatty acids.

• Pentose phosphate pathway: a pathway that branches off from the main glycolysis pathway and generates NADPH (which is important for redox balance) and pentoses (for nucleotide synthesis).

• Fatty acid metabolism: includes fatty acid oxidation for the production of energy and fatty acid synthesis, which produces membrane building blocks and signalling molecules.

• Amino acid metabolism: glutamine is a crucial fuel source and precursor, and other amino acids are vital for protein synthesis, regulatory processes and epigenetic modifications. Moreover, amino acid content is a sensor of metabolism homeostasis. Low amino acid content initiates the general control non-derepressible 2 response, which inhibits global protein synthesis while promoting the translation of key survival genes. High amino acid content, especially of leucine and other branched chain amino acids, activates the mTOR pathway to promote cell growth and anabolic processes.

• Mitochondria: mitochondria are the primary source of ATP. They also maintain redox balance by generating and detoxifying reactive oxygen species, which are important signalling molecules but can induce cell damage at high levels. Mitochondrial dynamics (fusion and fission) and quality control (mitophagy) are critical for regulating energy demand and preventing the release of inflammatory signals.

• mTOR pathway: a central hub that integrates signals from the environment, such as nutrients, growth factors and cytokines, to control cellular growth and metabolism, which are crucial for immune cell function. Activation of the mTOR pathway is linked to glycolysis, amino acid metabolism and mitochondrial homeostasis.

Another mechanism by which metabolism reprogrammes immune cell function is through the regulation of epigenetic modifications³. For example, acetyl-CoA and the tricarboxylic acid (TCA) metabolite α-ketoglutarate can support histone acetylation or methylation as well as DNA demethylation to influence chromatin accessibility and transcription. Moreover, cells sense and respond to available nutrients and inflammatory signals by adapting their function and/or differentiation. Multiple alterations have been reported in the immune cells of patients with systemic lupus erythematosus (SLE) and mouse models of SLE^4–7. For example, the abnormal phosphorylation of metabolic enzymes has been linked to SLE pathogenesis, including the serine/threonine protein phosphatase, PP2A, which regulates T cell metabolism and is activated in SLE^8,9. Correction of such alterations has demonstrated beneficial effects in mouse models of disease, and — for a small subset of interventions — in patients. This Review describes alterations in cellular metabolism associated with lupus nephritis — the most severe end organ manifestation of SLE. We next extend this discussion to two other autoimmune-induced renal diseases, IgA nephropathy (IgAN) and antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV). Finally, we discuss the potential of therapies that target immunometabolism in autoimmune-induced kidney diseases.

Pathogenesis of lupus nephritis

Progressive kidney disease that defines lupus nephritis occurs in up to 50% of patients with SLE and is associated with considerable morbidity and mortality¹⁰. Histologically, lupus nephritis is classified into six classes, with classes III and IV being the most severe¹¹. Tissue injury in SLE is induced by pathogenic autoantibodies, which are produced through a complex process. In addition to B cells and CD4⁺ T cells, several myeloid cell subsets contribute to autoantibody production¹². The exact mechanisms by which SLE progresses to lupus nephritis are not entirely clear. The glomerular deposition of immune complexes composed of antinuclear, anti-C1q, and crossreactive antiglomerular autoantibodies are thought to be predominant pathogenic drivers^13,14. However, differences in the function of kidney cells and their intrinsic susceptibility and response to injury influence disease pathogenesis independent of autoimmunity. Genetic predisposition also has an important role¹⁵; in particular, variants of several genes that regulate kidney functions have been linked to lupus nephritis¹⁶. Thus, lupus-associated autoimmunity contributes to lupus nephritis, but additional factors are required for the disease to progress to end organ damage.

Secondary to immune complex deposition in the glomeruli, locally produced chemokines such as MCP1, RANTES and MIP1 recruit immune cells to perpetuate kidney injury¹⁷. CD4⁺ T cells, CD8⁺ T cells and B cells undergo clonal expansion within the kidneys of patients and animal models, although these clones are also detected in the peripheral blood, suggesting a systemic origin¹⁸. Biopsy samples from patients with lupus nephritis contain B cells that recognize vimentin — an intracellular structural protein that is cleaved and extruded from apoptotic cells¹⁹. Antibodies to annexin 1 and α-enolase have also been detected in the serum and kidneys of patients with lupus nephritis²⁰, indicating either that immune complexes may form within the kidneys or that preformed circulating immune complexes may deposit in glomeruli. Patients with SLE also often produce antibodies against neutrophil granular proteins, such as lactoferrin, myeloperoxidase (MPO), proteinase 3 and elastase — known as ANCA²¹. The neutrophils of patients with SLE also release extracellular traps (NETs), which contain chromatin and intracellular proteins²² and act as a source of autoantigens²³. Macrophage infiltration is a common pathological finding, and is associated with poor outcomes in patients; moreover, therapies that target macrophages have been shown to alleviate kidney injury in animal models of lupus nephritis²⁴. Macrophages within the kidneys of patients with lupus nephritis can be classified as either kidney-resident macrophages (KRMs) or monocyte-derived macrophages (MDMs). These two subtypes have different origins and phenotypes^25,26, and contribute to lupus nephritis in different ways. KRMs promote leukocyte recruitment at sites of inflammation by expressing monocyte chemokines²⁷, whereas MDMs exacerbate autoimmune responses by presenting immune complex-associated antigens²⁸.

Although studies have largely focused on glomerular pathology in lupus nephritis, biopsy samples indicate that the extent of tubulointerstitial lesions is a better predictor of kidney pathology²⁹. The enhanced glomerular permeability secondary to injury leads to an over-absorption of proteins by proximal tubular epithelial cells (PTECs), triggering tubulointerstitial inflammation, scarring and deterioration of kidney function³⁰. The deposition of immune complexes is also observed in the tubular basement membrane of patients with lupus nephritis, especially those with class III or IV disease, and correlates with the severity of tubulointerstitial inflammation³⁰. Tubulointerstitial inflammation and tubular injury correlate strongly with loss of kidney function, independent of the severity of glomerular injury, and are indicative of poor outcomes^31,32. These findings position tubular injury as a consequence of glomerular pathology, and as a mediator of progressive kidney dysfunction, underscoring its utility in monitoring lupus nephritis severity. Thus, the collective responses of kidney parenchymal cells, infiltrating innate and adaptive immune cells synergize and worsen overall outcomes (Fig. 1).

Fig. 1 |. Mechanisms of lupus nephritis.

Capillaries within the glomerulus are lined by a fenestrated endothelium. This endothelium is contiguous with the mesangium, such that the mesangial cells are constantly exposed to blood components. The fenestrated endothelium is lined by the glomerular basement membrane, which itself is lined by podocytes and their foot processes. Lupus nephritis involves the deposition of immune complexes (ICs) in the mesangial region (classes I and II), subendothelial region (classes III and IV) or subepithelial region (class V) regions, which induces cellular pathology and inflammation. Some common pathological features of lupus nephritis include mesangial proliferation, basement membrane thickening, endocapillary proliferation, sclerosis and the formation of crescents with loss of podocyte foot processes. Immune cells, including T cells, dendritic cells (DCs), monocytes and neutrophils, also infiltrate the glomerulus where they can interact with the glomerular cells, secrete cytokines and amplify the inflammatory response. Glomerular injury and impairment of the glomerular filtration barrier exposes tubular epithelial cells (TECs) to filtered proteins and cytokines. Tubular IC deposits are also a common feature in lupus nephritis, and the activation of TECs by anti-double-stranded DNA antibodies and ICs further promotes the production of pro-inflammatory cytokines. IC deposition in the tubular basement membrane can also activate complement components, and induce the recruitment and activation of immune cells, including T cells, B cells, DCs, monocytes (which differentiate into macrophages or DCs) and neutrophils via the production of chemotactic factors. The collective contribution of glomerular and tubular injury, chemokine production and immune cell recruitment initiates and perpetuates kidney pathology.

Metabolic regulation of autoantibody production

The production of autoantibodies in SLE occurs in systemic lymphoid organs. Their production is associated with several metabolic alterations, as described in the following sections (Fig. 2).

Fig. 2 |. Metabolism of immune cells in lupus nephritis.

Immune cells contribute to lupus nephritis through two main processes. a, Production of autoantibodies that deposit in the glomeruli as immune complexes. These cells are mostly located in the secondary lymphoid organs. The release of neutrophil extracellular traps (NETs) by neutrophils through NETosis produces oxidized DNA that fuels the production of type I interferon (IFN), which activates the differentiation of B and CD4⁺ T cells. Autoreactive B cells differentiate into germinal centre (GC) B cells and extrafollicular B cells (double-negative (DN2) cells) with the help of T follicular helper (T_FH) cells and T peripheral helper (T_PH) cells, which then differentiate into long-lived plasma cells (LLPCs) and short-lived plasma cells (SLPCs), respectively. Both LLPCs and SLPCs produce autoantibodies. b, Immune cells infiltrate the inflamed lupus kidney to amplify inflammation and cause cell injury. Conventional B cells and innate-like, tissue-resident B1a cells can develop into SLPCs that produce autoantibodies in situ with the help of T cells. Infiltrating T cells also produce inflammatory and cytotoxic cytokines. NETosis also contributes to inflammation. Inflamed kidney-resident macrophages (KRMs) recruit monocytes which differentiate into monocyte-derived macrophages (MDMs). MDMs get in turn activated by internalization of immune complexes. The major metabolic pathways that sustain the cells in both processes are indicated by coloured boxes. In addition, renal T cells exhibit an increased HIF1α-driven programme, which presumably enables them to adapt to the hypoxic environment, and renal B cells express ion transporter Na⁺–K⁺ ATPase to counter the high sodium levels present in the tissue. Glycolysis is overactive in most of the immune cells responsible for producing autoantibodies that are deposited in the kidneys as well as the immune cells in the lupus kidney. MHC, major histocompatibility complex; TCR, T cell receptor.

B cell metabolism

Autoreactive B cells differentiate into antibody-secreting cells in germinal centres where, with help from CD4⁺ T follicular helper (T_FH) cells, they undergo class switching and somatic hypermutation to generate high-affinity, long-lived plasma cells (LLPCs). Autoantibodies can also be produced outside germinal centres through an extrafollicular pathway by atypical memory B cells or double-negative (that is, IgD⁻CD27⁻) B cells in humans, and by age-associated B cells (ABCs), in mice³³. These extrafollicular B cells require TLR7 signalling as well as CD4⁺ T cell help — most probably from cells called T peripheral helper cells in humans and extrafollicular T helper cells in mice — and differentiate into short-lived plasma cells (SLPCs). Extrafollicular B cells are increasingly recognized as a major pathogenic contributor to both SLE and lupus nephritis, and substantial efforts have been devoted to understanding the mechanisms responsible for their expansion with the ultimate goal of selectively eliminating them³⁴.

The activation and early differentiation of non-autoreactive B cells into antibody-secreting cells relies on mitochondrial respiration³⁵ and glycolysis³⁶. However, within germinal centres (where B cells undergo affinity maturation and class switching in response to foreign antigens), B cells rely primarily on fatty acid oxidation (FAO) and exhibit relatively low glycolytic activity³⁷. A 2014 study showed that chronic exposure of B cells to their growth factor BAFF, as occurs in SLE, increases the dependence of differentiating antibody-secreting cells on glycolysis³⁸. This early finding suggested that autoreactivity might reprogramme B cell metabolism, and has since been confirmed by other studies.

B cells from lupus-prone mice are more glycolytic and have higher levels of mitochondrial respiration compared with B cells from healthy control mice³⁹. The methyltransferase, METTL1, produces a m⁷G tRNA modification in germinal centre B cells that affects mitochondrial electron transport chain activity through B cell receptor (BCR) signalling⁴⁰. Increased METTL1 expression in the B cells of mouse models of lupus and patients with SLE correlates with disease activity and autoantibody production, including the production of anti-double-stranded DNA (anti-dsDNA) IgG, which is associated with lupus nephritis. Furthermore, Mettl1 deficiency reduced the elevated respiration rate in lupus B cells as well as their differentiation into antibody-secreting cells. These findings link m⁷G tRNA modification to increased BCR signalling and mitochondrial metabolism, which is likely to promote the survival and expansion of autoreactive B cells.

Treatment with 2-deoxyglucose (2DG) — a glucose analogue that blocks the first step of glycolysis — normalized the high glycolysis and mitochondrial respiration rates of lupus-prone mice³⁹. 2DG also prevented the expansion of germinal centre B cells and eliminated the production of autoantibodies in mouse models of lupus⁴¹. Autoreactive B cells from the well-characterized AM14 mouse model of SLE are transgenic for a BCR that encodes a rheumatoid factor (that is, an antibody-binding antibody) that requires endosomal TLR signalling for activation⁴². A 2023 study found that autoreactive B cells from AM14 mice became more glycolytic following activation by their cognate antigen and that administration of 2DG inhibited their differentiation into antibody-secreting cells⁴³. Moreover, bystander non-antigen-specific B cells were not affected by 2DG treatment. These results support the notion that autoreactivity increases the glucose requirements of B cells. In the MRL/lpr mouse model of SLE, genetic depletion of ABCs reduced autoantibody production and the severity of lupus nephritis, demonstrating the contribution of these cells to antibody-mediated renal pathology⁴⁴. ABCs are highly glycolytic, which is at least in part driven by their response to IFNγ, which induces the expression of glycolytic genes through the transcription factor and ABC lineage marker, TBET⁴⁵. The same study showed that the inhibition of glycolysis prevented the differentiation of human B cells into TBET⁺CD11c⁺ ABC-like B cells, suggesting that pharmacological inhibition of glycolysis may also exert beneficial effects in patients with SLE through targeting of these cells.

Of note, plasma cells require glucose for the production and secretion of antibodies. Glucose is also used to generate substrates for antibody glycosylation⁴⁶. Aberrant IgG glycosylation in children with lupus nephritis has been associated with alterations in podocyte metabolism, which renders these cells vulnerable to injury⁴⁷. Glucose uptake is also required by LLPCs, which exclusively use glucose-derived pyruvate to fuel their TCA cycle and generate ATP⁴⁶. However, SLE-associated autoantibodies are secreted by both SLPCs and LLPCs⁴⁸ and therefore both types of antibody-secreting cells must be targeted to reduce the production of immune complexes that initiate and perpetuate lupus nephritis. Although not formally tested owing to technical limitations, it is likely that curbing glucose uptake or the downstream production of pyruvate in B cells would achieve both, by limiting the expansion of extrafollicular B cells and the resultant SLPCs and by limiting the survival of LLPCs. Moreover, as autoreactive B cells seem to be more glycolytic than bystander B cells, inhibition of glycolysis would probably preserve the ability of B cells to respond to foreign antigen challenges.

Glutamine is another carbon source to feed the TCA cycle as well as a substrate for epigenetic modifications⁴⁹. Treatment of lupus-prone mice with 6-diazo-5-oxo-L-norleucine (DON) — a glutamine analogue that inhibits the first step of glutaminolysis — reduced the production of autoantibodies and the frequency of germinal centre B cells⁴¹. Contrary to 2DG, DON also reduced the production of antibodies and the expansion of germinal centre B cells in response to immunization with a foreign antigen, suggesting a global requirement for glutamine by B cells for humoral responses. However, the differentiation of isolated lupus B cells into antibody-secreting cells under conditions that simulated T cell-dependent activation was not affected by DON³⁹, suggesting that the requirement for glutamine may not be B cell-intrinsic.

T cell metabolism

CD4⁺ T cells from lupus-prone mice and patients with SLE are also more glycolytic and have a higher mitochondrial respiration rate than CD4⁺ T cells from healthy control mice⁵⁰. Treatment of lupus-prone mice with 2DG reduced the frequency of autoreactive T_FH cells but did not affect the expansion of T_FH cells in response to foreign antigens⁴¹. Notably, 2DG treatment normalized the expression of a cluster of genes expressed by T_FH cells from lupus-prone mice and patients with SLE⁵¹. 2DG also inhibited both the in vitro polarization of T_FH cells from lupus-prone mice, their ability to help B cells⁵², and the differentiation of purified lupus-associated B cells into antibody-secreting cells under conditions that simulate T cell-dependent activation³⁹, suggesting that both lupus T_FH cells and germinal centre B cells are independently glycolytic. The role of enhanced glycolysis in functionally sustaining T_FH cells and their ability to help B cells has been validated in patients with SLE⁵³. Pharmacological inhibition of glutaminolysis decreased disease activity in lupus-prone mice and reduced the frequency of pro-inflammatory T helper 17 (T_H17) cells⁵⁴. More specifically related to autoantibody production, inhibition of glutaminolysis impaired T_FH cell function in lupus-prone mice by reducing the expression of inducible T cell costimulator — a costimulatory receptor that is required for the differentiation and maintenance of T_FH cells⁵⁵. T cell-specific deletion of Gls1 — which encodes the main target of DON, glutaminase — reproduced these phenotypes, demonstrating that glutaminolysis is required in a cell-intrinsic manner by lupus T_FH cells to exert their B cell helper function.

Mitochondrial metabolism

The TCA cycle-derived metabolite, itaconic acid, is synthesized by myeloid cells in response to inflammation, type I interferon (IFN) and oxidative stress. Treatment of lupus-prone mice with an itaconate derivative reduced autoantibody production and kidney injury⁵⁶. However, the extent to which the renoprotective effect was due to the reduction in autoantibody production is unclear. Itaconate inhibits the differentiation of T_H17 cells by suppressing glycolysis and oxidative phosphorylation, and reducing chromatin accessibility, as demonstrated in a mouse model of multiple sclerosis⁵⁷. The pathogenic role of T_H17 cells in lupus nephritis suggests that itaconate derivatives may have potential therapeutic benefits.

Mitochondrial alterations in bone-marrow-derived cells also contribute to the production of autoantibodies in SLE. Disrupted mitochondrial homeostasis and redox imbalance induce the release of oxidized mitochondrial DNA (mtDNA) from neutrophils. Oxidized mtDNA induces type I IFN through the cGAS–STING pathway, leading to the production of autoantibodies^58,59. Platelet-derived mitochondria also represent a major source of autoantigens in patients with SLE⁶⁰. Specifically, immune complexes in the circulation of patients with SLE can bind to the FcγRIIA receptor on platelets, which promotes their recruitment to the kidney and their extrusion of interferogenic mtDNA in situ⁶⁰. Erythrocytes that fail to switch from glycolysis to oxidative phosphorylation during their terminal differentiation represent an unexpected additional source of mtDNA in patients with SLE⁶¹. Glycolysis increases levels of lactate in lupus-associated erythrocytes, which leads to proteasome lactylation, impaired mitophagy and consequent mitochondrial retention. Erythrocytes are normally phagocytosed by the splenic reticuloendothelial macrophages. However, most erythrocytes in patients with SLE contain mtDNA, which activates the cGAS–STING pathway in splenic macrophages, inducing them to produce large amounts of type I IFN, and ultimately increasing autoantibody production. Defective mitochondria have also been documented in CD4⁺ T cells from patients with SLE⁶². A 2025 study demonstrated that sensing of cytosolic mtDNA by ectonucleotide pyrophosphatase–phosphodiesterase 1 (ENPP1) intrinsically activates SLE-associated CD4⁺ T cells⁶³. Specifically, binding of ENPP1 to mtDNA increases transcription of the glucose transporter GLUT1, which enhances glycolysis and mTOR activation. Inhibition of ENPP1 was as effective as inhibiting either glycolysis or mTOR in reducing the production of autoantibodies and attenuating nephritis in a humanized mouse model of SLE. Together these findings suggest that the release of mtDNA from defective mitochondria drives the metabolic activation of T cells and myeloid cells to trigger the production of type I IFN in SLE. Although not all patients with SLE present with high type I IFN activity, it correlates with disease activity⁶⁴.

In addition to inducing immune activation, mitochondrial nucleic acids are also a target of elevated autoantibody responses in patients with SLE^65,66. Although anti-mtDNA antibodies are associated with anti-dsDNA antibodies, one study found a negative correlation between levels of anti-mtDNA antibodies and lupus nephritis disease activity⁶⁷, whereas another study found a positive correlation with disease activity⁶⁶. Positive correlations have also been found between antibodies against whole mitochondria and disease activity in paediatric patients with SLE⁶⁸, and with higher mortality in patients with lupus nephritis⁶⁶. Moreover, a further study found a correlation between the number of cell-free mitochondria in serum and cell-free mtDNA in the urine and urine protein-to-creatinine ratio in patients with lupus nephritis⁶⁹. Whether these markers are secondary to increased cell death and decreased clearance of cellular debris, or to intrinsic cellular properties such as glycolysis, which might impair mitochondria processing, is unclear. One study found that serum markers of mitochondrial dysfunction, including a reduced mtDNA-to-nuclear DNA ratio, were elevated in patients with lupus nephritis compared with patients with SLE without renal involvement⁷⁰, supporting the notion that mitochondrial dysfunction contributes to lupus nephritis. However, available data clearly show that mitochondria and their DNA have key roles as both initiators and targets, in the process that leads to autoantibody production in the context of SLE.

Metabolic regulation of immune infiltrates

The kidneys of patients with lupus nephritis and mouse models of lupus nephritis show complex and abundant immune cell infiltrates composed of myeloid cells, CD4⁺ T cells, CD8⁺ T cells and B cells¹⁸ (Fig. 2). Little is known about the metabolic programmes that regulate the immune cell infiltrates in the nephritic kidney. We can assume that the corresponding immune cell subsets in kidney draining lymph nodes or other lymphoid organs maintain their metabolic profile following their infiltration in the kidney; however — and as described below — this assumption is difficult to verify⁷¹. In addition, the tissue microenvironment shapes immune cell phenotypes, and most probably the metabolism of immune cells, as has been shown for tissue-resident macrophages⁷². The healthy kidney is low in oxygen, and hypoxia is further exacerbated in lupus nephritis in association with activation of the HIF1α pathway⁷³. This altered environment has consequences for T cell survival and function, and most probably the pathogenic contribution of T cells to lupus nephritis. In addition to low oxygen, the kidney has a high concentration of sodium to which infiltrating cells must adapt⁷⁴.

A comprehensive gene expression analysis of kidney samples from patients and mouse models of lupus nephritis revealed a global downregulation of metabolic genes in these samples compared with those from healthy controls⁷⁵. Why hypermetabolic immune cells would simultaneously shut down major energy-producing pathways when entering the inflamed kidneys is unclear. The researchers involved in that study proposed that hypoxia may have a role; however, hypoxia activates HIF1α, which would be expected to induce glycolysis⁷⁶, and therefore does not support this hypothesis. The bulk nature of the analysis as well as the lack of protein validation limits the interpretation of these results. As described below, another study identified exhausted CD8⁺ T cells in the lupus kidney⁷¹; however, direct and indirect evidence from other studies suggests that individual cell types found in the nephritic kidney are actually hypermetabolic.

Myeloid cells

NETosis is a process through which neutrophils participate in host defence by releasing their chromatin into the extracellular space, forming NETs that trap and destroy pathogens. Neutrophils — and in particular a distinct subset called low-density granulocytes (LDGs) that are expanded in the peripheral blood of patients with SLE — accumulate in the kidneys of patients with lupus nephritis;. Increased NETosis by neutrophils within the kidney and the impaired clearance of their products contribute to kidney pathology⁷⁷. As mentioned above, neutrophils are a major source of oxidized mtDNA, which triggers the production of pathogenic autoantibodies^58,59. NETs within the kidneys of patients with lupus nephritis also contain mtDNA, which might trigger the production of type I IFN by tubulointerstitial plasmacytoid dendritic cells (DCs)⁷⁸ or other cells. Lupus-associated neutrophils also demonstrate various metabolic alterations that may contribute to both systemic and renal pathogenesis. Healthy neutrophils rely on glycolysis to generate the energy required to sustain their activation. However, glycolysis is further enhanced in neutrophils from patients with another autoimmune disease — antiphospholipid syndrome — where it contributes to microvasculature damage⁷⁹. It is likely that increased glycolysis also sustains effector functions in kidney neutrophils in the context of lupus nephritis. However, the most consistent metabolic alterations in lupus-associated neutrophils, specifically LDGs, centre around mitochondrial dysfunction and the increased production of mitochondrial reactive oxygen species (ROS), which stabilizes HIF1α and increases glycolysis⁷. Metformin — a glucose-lowering drug that mildly inhibits Complex I in the electron transport chain — reduced NETosis in vitro⁶⁵. Similarly, idebenone, which is a synthetic analogue of the antioxidant coenzyme Q10 (CoQ10), restored mitochondrial function and reduced the formation of NETs by neutrophils isolated from patients with SLE⁸⁰, confirming a role for mitochondrial respiration in NETosis. The use of metformin as an add-on treatment to standard care was beneficial in patients with SLE and low disease activity⁶⁵ whereas idebenone ameliorated renal pathology in mouse models of lupus⁸⁰, although how much of this therapeutic benefit can be attributed to the targeting of neutrophils is unclear.

Both MDMs and KRMs are important components of the immune infiltrates in lupus nephritis, in terms of both their size and their pathogenic contribution¹⁸. MDMs are mainly responsible for the uptake of immune complexes and the resulting inflammatory response, whereas KRMs recruit monocytes, which can differentiate into MDMs and produce factors that support B cell aggregates²⁸. In a 2020 study, MDMs from the kidneys of mice with lupus nephritis demonstrated a glycolytic phenotype that was induced by their FcγR-mediated uptake of immune complexes⁸¹. Given that the differentiation of macrophages into a pro-inflammatory M1 phenotype is strongly dependent on glycolysis⁸², this finding is consistent with a glycolytic switch being necessary for MDMs to secrete pro-inflammatory factors upon immune complex uptake. Therefore, in addition to eliminating the production of autoantibodies, inhibition of glycolysis could limit the response of renal MDMs to immune complexes that contain autoantibodies, enabling dual systemic and tissue therapeutic outcomes. A 2023 study of gene expression showed that the metabolism of tissue-resident macrophages varies between tissues, probably as an adaptation to variations in nutrients and oxygen levels, with additional alterations induced by their activation⁷². However, KRMs were not evaluated in that study. A more recent, large-scale, single-cell RNA sequencing (scRNA-seq) study of intrarenal myeloid cells in which four strains of lupus-prone mice and a large cohort of patients with lupus nephritis were analysed found consistent expansion of a subset of KRMs and a subset of MDMs (called RM0 and C2, respectively)⁸³. In lupus-prone mice, both of these subsets shared the highest ATP production and mitochondrial membrane potentials, starting in prenephritic state and increasing with disease progression. These results suggest a correlation between hyper-mitochondrial metabolism and pathogenicity. Interestingly, the mouse and human C2 subsets shared transcriptional programs that were predicted to support oxidative phosphorylation and lipid metabolism, whereas MDM C1 subsets that did not expand with disease expressed a glycolytic signature⁸³. The apparent discrepancies between studies that have linked glycolysis⁸¹ or oxidative phosphorylation⁸³ with pathogenic macrophages could have originated from differences in the definition of macrophage subsets.

Finally, DCs are also part of the renal myeloid infiltrate in lupus nephritis. Biopsy studies have identified different DC subsets, including FcRγ⁺ DCs⁸⁴ and CD163⁺ DC3s⁸⁵, both of which are likely to contribute to CD4⁺ T cell expansion and polarization towards inflammatory phenotypes. However, the above-mentioned scRNA-seq study of intrarenal myeloid cells found low numbers of DCs with no significant differences observed during disease progression in mice, or between lupus nephritis and control human biopsy samples, suggesting that DCs may have a relatively minor role compared with that of macrophages⁸³. The metabolic requirements of these DCs are unknown. In general, glycolysis is upregulated in activated, pro-inflammatory DCs; however, additional metabolic processes, such as fatty acid synthesis, can also contribute to DC function⁸⁶. The metabolic programming of lupus nephritis-associated DCs cannot be inferred from studies of bone-marrow-derived DCs.

Lymphocytes

Kidney-infiltrating B cells in lupus nephritis demonstrate adaptations to the high sodium environment that are unique among renal immune cells but shared with renal epithelial cells⁷⁴. The expression of sodium potassium adenosine triphosphatase (Na⁺–K⁺ ATPase) correlates with the ability of B cells to survive in the kidney. Indeed, biopsy analyses demonstrate that B cells from patients with lupus nephritis — particularly those with class V lupus nephritis — express high levels of Na⁺–K⁺ ATPase. Pharmacological inhibition of Na⁺–K⁺ ATPase with ouabain or genetic knockout of the Na⁺–K⁺ ATPase γ-subunit reduced B cell numbers in the kidneys of lupus-prone mice, and proteinuria. Interestingly, plasma cells were not affected by ouabain treatment, suggesting that the pathogenic effects of Na⁺–K⁺ ATPase-expressing B cells are independent of autoantibody production. These findings indicate that treatment with FDA-approved ouabain or small molecules that target the Na⁺–K⁺ ATPase could represent a therapeutic option for patients with a proteinuria-dominant form of the disease, such as those with class V lupus nephritis⁷⁴. Plasma cells are also found in nephritic kidneys and the number of TLR4⁺CXCR4⁺ plasma cells correlates with the severity of renal pathology⁸⁷. These plasma cells, which are non-proliferating and most probably LLPCs, are an in situ potent source of anti-dsDNA IgG that induces local inflammation. They probably rely on glucose uptake to generate pyruvate, as has been described for LLPCs in the bone marrow⁴⁶.

Expansion of CD11c⁺ B cells, which correspond to murine ABC cells, has been reported in the kidney and urine of patients with class IV lupus nephritis⁸⁸. These cells overexpress the m⁶A demethylase FTO, which promotes TLR7-mediated differentiation of ABCs in mice. Mechanistically, FTO induces the expression of ATP6V1G, a subunit of the vacuolar H⁺-ATPase, which supports mitochondrial autophagy and the elimination of damaged, oxidized mitochondria. B1a cells — an innate-like, tissue-resident type of B cells — have also been found in nephritic kidneys⁸⁹. The metabolism of B1a cells in the peritoneal cavity, where they are the most abundant, at least in mice, differs from that of follicular conventional B cells. Their homeostatic maintenance requires a combination of glycolysis, fatty acid synthesis and autophagy⁹⁰. This unique metabolic profile might be linked to their reliance on self-renewal rather than rapid proliferation in response to BCR signalling, as per conventional B cells. Whether the expanded B1a cell subset in nephritic kidneys shares the metabolic profile of cells in the peritoneal cavity is unknown. We propose that treatment with 2DG, which greatly reduces the numbers of homeostatic B1a cells⁹⁰, would also eliminate B1a cells in the kidneys of lupus-prone mice⁹¹. Of note, kidney B1a cells are a major source of IL-10, which promotes the polarization of KRMs into anti-inflammatory M2 cells, characterized by an oxidative phosphorylation-dominant metabolism⁹². Patients with SLE have high numbers of systemic B1a cells, which contribute to kidney injury through the production of antiphosphatidylserine antibodies, both systemically and locally in the kidney⁸⁹. Whether these antiphosphatidylserine-producing renal B1a cells produce IL-10 and polarize KRMs or other renal macrophages is unknown, and future investigation is warranted.

A number of studies have investigated the metabolic phenotype of CD4⁺ and CD8⁺ T cells that infiltrate the kidneys of lupus-prone mice and patients with lupus nephritis. T cells within the kidney produce IFNγ and granzyme B, but are transcriptionally distinct from their splenic counterparts⁷³. The hypoxia-induced expression of HIF1α in T cells from the kidneys of lupus-prone mice promotes the production of NAD⁺ and NADP⁺ as well as proline metabolism. Glycolysis is also increased through the expression of the pyruvate dehydrogenase kinase PDK2, which prevents pyruvate from converting into acetyl-CoA and entering the TCA cycle. HIF1α also promotes T cell survival through the expression of BNIP3 by promoting mitophagy. The differential expression of these genes and pathways was also observed in T cells in biopsy samples from patients with lupus nephritis. T cell-specific deletion of Hif1α was sufficient to reduce the number of T cells infiltrating the kidneys of lupus-prone mice without affecting the production of autoantibodies or systemic autoimmunity. Consequently, immune complex-induced glomerulonephritis was unaffected, although tubulointerstitial pathology scores were lower and survival increased in Hif1α-knockout mice, suggesting a direct effect of HIF1α in tubular pathology. These results also demonstrate that adaptation to hypoxia enables pathogenic T cells to expand and contribute to a type of kidney injury associated with poor outcomes in lupus nephritis⁹³.

In contrast to the above studies, which suggest that infiltrating T cells have a pathogenic role in lupus nephritis, one study has shown in three lupus-prone mouse models that kidney-infiltrating T cells present an exhausted phenotype, characterized by reduced cytokine production and a suppressed metabolism with highly dysfunctional mitochondria⁷¹. The reason for the discrepant findings, despite use of the same mouse models⁷³ is unclear. However, one potential explanation is that the study that reported the exhausted phenotype assessed glucose uptake with 2NBDG — a dye that may have a limited reliability for the assessment of glucose transport in T cells⁹⁴ — rather than glycolysis itself⁷¹. Another study found that exhausted CD8⁺ T cells are not major contributors to the gene signature in biopsy samples from patients with lupus nephritis, although the expression of metabolic genes was suppressed globally⁷⁵. However, a further scRNA-seq analysis of T cells in lupus nephritis, IgAN and membranous nephropathy also found profound metabolic alterations with an overall lower metabolic flux compared with that of healthy T cells⁹⁵. These metabolic alterations correlated with clinical features, and reduced N-linked glycan synthesis, glutamate metabolism, glycolysis and TCA cycle activity were particularly prominent. Moreover, another scRNA-seq study of biopsy samples from patients with lupus nephritis showed high levels of granzyme gene expression in CD8⁺ T cells⁹⁶, which are considered to be tissue-resident T cells that contribute actively to organ damage⁹⁷. Interestingly, peripheral CD8⁺ T cells with reduced cytotoxicity have been reported in association with recurring infections in patients with SLE⁹⁸. CD38 is expressed at high levels by these hypofunctional CD8⁺ T cells, which results in low levels of NAD⁺ and inhibition of the deacetylase sirtuin 1. The discordance in functional activity between peripheral and kidney CD8⁺ T cells in lupus nephritis has been attributed to metabolic differences, such as oxygen availability, between the two tissue environments. However, further analyses are needed to understand the link between the expression of exhaustion markers, hypoxia and pathogenicity in renal T cells in lupus nephritis. Crosstalk between immune cells and renal tubular epithelial cells could also be involved. Activation of IL-23 receptor and calcium–calmodulin-dependent protein kinase type IV (CaMK4) signalling in tubular epithelial cells reduces the expression of arginase 1, increasing local levels of free arginine, which promotes T cell proliferation and pathogenesis⁹⁹. Although unclear, discrepancies in the reported metabolic status of T cells in lupus nephritis most probably reflect the complexity of the infiltrates and differences in experimental design.

Numerous studies have shown that T_H17 cells promote kidney injury in lupus nephritis through various mechanisms, including direct damage to podocytes and tubular epithelial cells, and by promoting the secretion of pro-inflammatory cytokines and chemokines that recruit further immune cells to the kidney¹⁰⁰. The differentiation and function of T_H17 cells is an energy-consuming process and depends on glycolysis¹⁰¹, glutaminolysis¹⁰² and fatty acid synthesis¹⁰³. As mentioned above, inhibition of either glycolysis or glutaminolysis reduces disease severity in mouse models of lupus; however, the effect of these treatments on renal T_H17 cells has not been studied. The inhibition of fatty acid synthesis, which ameliorated disease in a mouse model of multiple sclerosis¹⁰³, has not been tested in lupus. Finally, kidneys from lupus-prone mice and patients with lupus nephritis accumulate double-negative (that is, CD4⁻CD8⁻) T cells that adopt a pathogenic T_H17 phenotype¹⁰⁴. Little is known about the metabolic profile of double-negative T cells, except that their expansion is linked to mTORC1 activation, a hallmark of T cells in SLE¹⁰⁵.

Two 2025 studies have shown complex mechanisms by which kidney cells promote the expansion of pathogenic T cells within the kidney in the context of lupus nephritis. One of these studies showed that circulating DNA activates the STING–cGAS pathway in glomerular endothelial cells, inducing the expression of CD38, which is exported to T cells in exosomes. There, CD38 depletes NAD⁺ levels, impairing the ADP ribosylation of early growth response protein 1, which in turns enhances the transcription of PRDM1 (which encodes the transcription factor B lymphocyte-induced maturation protein 1) to promote the differentiation of CD103⁺ T cells¹⁰⁶. The second study showed that circulating DNA induces mesangial cells to produce lactate, which is taken up by infiltrating CD4⁺ T cells where it promotes lactylation of the T_FH cell regulator, BCL6, thereby preventing its proteasomal degradation¹⁰⁷. Therefore, metabolic alterations in kidney cells induced by systemic autoimmune processes amplify the pathogenicity of immune cells in situ.

Finally, innate lymphoid cells (ILCs) have been also implicated in the pathogenesis of lupus nephritis. Specifically, type 1 NKp46⁺ ILCs (ILC1s) contribute to kidney injury independently from autoantibodies by recruiting inflammatory myeloid DCs¹⁰⁸. In addition, IL-22-producing ILC3s accumulate in the kidneys of mice with lupus nephritis where they promote macrophage infiltration¹⁰⁹. Very little is known about ILC metabolism, except that both ILC1s and ILC3s seem to rely primarily on glycolysis for their activation with involvement of the mTOR–HIF1α pathway¹¹⁰. Whether inhibition of glycolysis or mTOR signalling, which ameliorates kidney pathology in mouse models of lupus^91,111, have any effects on renal ILCs is unknown.

Metabolic regulation of kidney cells

A single-cell transcriptional analysis has revealed minor changes in metabolic genes of resident and infiltrating immune cells in biopsy samples from patients with lupus nephritis compared with biopsies of kidneys from healthy living donors⁹⁶. An independent study confirmed that the presence of T cells contributed minimally to changes in renal metabolic gene expression⁷⁵. Collectively these studies suggest that renal metabolic changes during the progression of lupus nephritis predominantly occur in parenchymal cells.

As described earlier, the loss of immune tolerance in SLE results in the production of autoantibodies that form circulating and in situ immune complexes that disproportionately affect the vascular bed of the kidneys, particularly the glomeruli^112,113. Glomerular immune complexes can deposit in the subendothelial (class III and IV proliferative disease), subepithelial (class V disease, injured podocytes) or mesangial space¹¹⁴. The deposition of immune complexes activates the complement cascade or resident glomerular cells, inducing the release of inflammatory chemokines and cytokines that recruit immune effectors including neutrophils, monocytes and T cells to perpetuate glomerular injury.

Although glomerular immune complex deposition is the dominant feature of lupus nephritis, tubulointerstitial pathology predicts outcomes independent of glomerular damage^31,32. PTECs are more polarized and more metabolically active than other cells of the kidney, including distal tubular epithelial cells. They reabsorb 80% of the glomerular filtrate, including glucose, ions and nutrients, and contain more mitochondria than any other cells in the kidney¹¹⁵. Exposure of human PTECs to immune complexes containing anti-dsDNA IgG increased cellular levels of hydrogen peroxide, which activated the ROS-sensitive transcription factor NF-κB, ERK, MAPK and downstream JNK signalling pathways³⁰. This activation induced the secretion of IL-6, CXCL8, CCL2 and soluble fibronectin and a downstream increase in profibrotic TGFβ1 and collagen synthesis^116,117. Inflammatory cytokines and chemokines produced by injured PTECs also recruit immune effector cells into the renal tubulointerstitium¹¹⁸. Thus, glomerular and tubular injury, the altered metabolism of glomerular and tubular cells, and consequent immune infiltrates, combined with alterations in the metabolism of immune cells, as discussed above, collectively exacerbate lupus nephritis^119,120 (Fig. 3).

Fig. 3 |. Metabolism of renal parenchymal cells in lupus nephritis.

The renal pathology in lupus nephritis is mediated by initial injury to the glomerulus and perpetuated by tubular cell injury, which can change the metabolic profile of affected cells. Within the glomeruli, podocytes, glomerular endothelial cells and mesangial cells can be injured following the deposition of immune complexes and by infiltrating immune cells. The energy metabolism of podocytes is complex and can change depending upon conditions and insult. Exposure to IgG attenuates the rate of glycolysis, whereas the secretory phospholipase A2 group IB increases glycolysis and mTOR activation in the context of lupus nephritis. The renal tubules constitute the majority of the renal protein mass. A decreased glycolytic signature is observed in endothelial cells, suggesting that they are metabolically altered. Mesangial cells are early responders to immune complex deposits where the deposition of anti-DNA immune complexes triggers a glycolytic signature. Proximal tubule epithelial cells are in particular very metabolically active cells and contain a lot of mitochondria. These cells rely on fatty acid oxidation (FAO) rather than glucose to sustain their metabolic requirements; however, their inability to switch to glycolysis in the hypoxic lupus kidney leads to cell death, and might perpetuate renal injury.

Glomerular cells

Glomeruli are composed of endothelial cells, podocytes and mesangial cells. In contrast to the substantial number of studies that have investigated immune cell metabolism as a potential target to alleviate SLE, a limited number of studies have investigated changes in the metabolic prolife of renal structural cells during the evolution of lupus nephritis. However, increased expression of glomerular genes involved in FAO and a decrease in endothelial cell glycolysis has been reported⁷⁵. As healthy endothelial cells are highly glycolytic¹²¹, these findings suggest that endothelial cells may be metabolically altered. However, these observations are based on gene expression and need validation at the protein level.

Podocytes synthesize large quantities of ATP to maintain their cytoskeleton and glomerular filtration barrier, and to clear basement membranes of adhered albumin and immune complexes¹²². Pharmacological inhibition of glycolysis, but not mitochondrial respiration, suppressed the formation of lamellipodia — actin filament structures that are important for cell migration — and induced apoptosis in differentiated podocytes in vitro, indicating that podocytes depend on glycolysis as a source of ATP¹²³. Human podocytes cultured with IgG derived from patients with lupus nephritis had a lower rate of glycolysis than podocytes cultured with deglycosylated IgG derived from patients with lupus nephritis or with IgG from healthy volunteers⁴⁷, implicating IgG glycosylation in the metabolic alteration of podocytes. Mechanistically, glycosylated IgG induces the activation of CaMK4 in podocytes, which downregulates the expression of the slit diaphragm protein, nephrin, in an NF-κB–SNAIL-dependent pathway. Paradoxically, CaMK4 induces glycolysis in T cells^124,125, suggesting that CaMK4 activation by glycosylated IgG could also potentially increase glycolysis in podocytes, and thereby alter their function. In support of this notion, another study identified glycolysis as the most enriched metabolic pathway following exposure of podocytes to IgG derived from patients with lupus nephritis.

In addition, the secretory phospholipase A2 group IB (sPLA2-IB), which has been linked to proteinuric kidney disease, increased glycolysis and mTOR activation in podocytes and was associated with the development of podocyte pathology¹²⁶. Moreover, levels of sPLA2-IB were increased in the glomeruli of rats subjected to anti-Thy1-induced glomerulonephritis¹²⁷ and in patients with idiopathic membranous nephropathy, in which it induces podocyte pathology by activating p38MAPK–mTOR–ULK1ser757 signalling¹²⁸. In a model of diabetic kidney disease, podocyte-specific overexpression of pyruvate kinase M2 (PKM2) provided renoprotection by normalizing diabetes-induced alterations in glycolytic rate and mitochondrial function throughout the glomeruli¹²⁹. In its tetrameric form, PKM2 catalyses the final step of glycolysis to produce ATP, whereas in its dimeric form, PKM2 translocates to the nucleus to promote the expression of glycolytic genes¹³⁰. Mechanistically, the tetrameric structure and enzymatic activity of PKM2 in podocytes preserves mitochondrial function and VEGF expression and prevents the development of diabetic nephropathy. These findings suggest that podocyte energy metabolism, especially glycolysis, could represent a therapeutic target in a time-dependent and context-dependent manner, and further investigation in the setting of lupus nephritis is therefore warranted.

The mesangium and mesangial cells provide structural support for the capillary loops and are important for the production of glomerular matrix¹³¹. Mesangial cells respond early to the glomerular deposition of immune complexes and exhibit diverse immunological functions, contributing to inflammation and fibrosis in lupus nephritis¹³². Anti-DNA immune complex deposits in the mesangium trigger a glycolytic response in mesangial cells¹³³, which enhances the lactylation and subsequent degradation of the transcription factor, PBX1, leading to excessive mesangial cell proliferation. Accordingly, pharmacological inhibition of lactate production in humanized mice carrying immune cells from patients with lupus nephritis effectively alleviated renal inflammation and progression of fibrosis. Additional evidence for increased glycolysis in mesangial cells was reported in the Fcgr2b^−/− mouse model of SLE¹³⁴. Fcgr2b^−/− mesangial cells overexpress the transcription factor, DEC2, which activates glycolysis in mesangial cells by increasing the expression of TLR4 and the glucose transporter GLUT1 (ref. 134).

Renal tubules

The renal tubule comprises at least 14 segments and contains at least 16 distinct epithelial cell types¹³⁵, although PTECs are the predominant cell type. As described above, the high workload and mitochondrial content of PTECs underlie their very high energy requirements, which are largely met by FAO under healthy conditions¹³⁶. Although PTECs reabsorb glucose, they do not typically use glucose as their energy source¹³⁷. Healthy urine is glucose-free by the time it reaches the distal tubular segments, which lack apical glucose transporters. The reliance of PTECs on oxidative mitochondrial metabolism combined with their limited glycolytic capacity renders them uniquely susceptible to damage, particularly in response to hypoxia¹³⁶. Anaerobic glycolysis is critical for the response of PTECs to injury as it generates small amounts of ATP, which is necessary for the maintenance of cell viability and later recovery¹³⁸. However, glycolysis is limited in fully differentiated PTECs, as they have low levels of glycolytic enzymes, including hexokinase¹³⁹. This limited capacity for anaerobic glycolysis is especially critical in the context of lupus nephritis, in which the kidneys can become overtly hypoxic. Under these conditions, insufficient levels of oxygen for the generation of ATP from FAO leads to cell death^73,140. Therefore, contrary to other cell types in which glycolysis is associated with a pathogenic response, the relative inability of PTECs to switch from FAO to glycolysis may contribute to renal injury in lupus nephritis.

The lipid profile of kidneys in lupus nephritis

FAO is critical for PTEC respiration and the fatty acids used for oxidation are derived from lipids¹⁴¹. Although the cellular source of lipids within the kidney is not known, lipid species and abundance vary during the evolution of lupus nephritis. Nephritic kidneys from MRL/lpr mice have substantially higher levels of the glycosphingolipids glucosylceramide (GlcCer) and lactosylceramide (LacCer) than in kidneys from non-nephritic and healthy control mice¹⁴². The increase in these glycosphingolipids occurred secondary to an increase in the activity of neuraminidase 1, an enzyme involved in sphingolipid synthesis. Notably, elevated levels of urinary LacCer were detected prior to the onset of proteinuria in MRL/lpr mice, and were also higher in the urine and kidneys of patients with lupus nephritis than in patients with SLE without nephritis or healthy individuals. Kidney glycosphingolipids are also elevated early in diabetic nephropathy, in which they mediate mesangial cell hypertrophy¹⁴³. Moreover, inhibition of GlcCer accumulation in a mouse model of polycystic kidney disease blocked disease progression¹⁴⁴. Similarly, administration of the GlcCer synthase inhibitor, eliglustat, reversed disease phenotypes in a model of type 1 Gaucher disease¹⁴⁵. Collectively, these studies suggest that blocking glycosphingolipid synthesis should be explored as a potential treatment strategy in lupus nephritis.

Aberrant lipid metabolism in pre-disease and diseased kidneys from MRL/lpr mice was also found in a comprehensive lipidomics study¹⁴⁶. Although the cellular source of those lipids was not identified, ectopic fat deposits and lipid species relevant to oxidative stress (for example, 4-hydroxyalkenal, ceramide and lysophospholipids) were associated with the development of lupus nephritis. Furthermore, increased levels of multiple species of anti-inflammatory, fatty acid ester of hydroxyl fatty acids (FAHFA) in the kidneys reflected the severity of lupus nephritis, possibly in response to tissue injury. Although treatment with the corticosteroid prednisone delayed the progression of lupus nephritis, it aggravated the aberrant metabolism of the lipids, particularly when used for a long time. Finally, in line with their high dependency on FAO, acute and chronic PTEC pathology is associated with lipid accumulation, which correlates with disease severity^147,148. Lipid peroxidation — a process that ultimately leads to cell death — is predominantly observed in the tubules of patients with lupus nephritis and diseased MRL/lpr mice¹⁴⁹. It would be interesting to know whether changes in FAHFAs and other lipid species occur predominantly in this segment of the nephron, particularly in the PTECS.

Iron and cellular metabolism

Iron is essential for physiological processes and iron-containing proteins catalyse critical reactions involved in oxygen sensing^150,151, energy metabolism¹⁵² and DNA synthesis¹⁵³. However, bioactive or labile Fe (Fe²⁺) participates in the Fenton reaction and catalyses the generation of ROS, which induce oxidative stress and lipid oxidation^154,155. The role of iron in the pathogenesis of SLE is well-documented^156–158. Iron also contributes to the pathophysiology of lupus nephritis by accumulating within renal tubules, where it drives lipid peroxidation^{149,159–163}. Iron sulfur clusters are key components in several complexes within the electron transport chain where they are essential for the efficiency of oxidative phosphorylation¹⁶⁴. As discussed above, PTECs rely predominantly on oxidative phosphorylation for their energy metabolism; thus, iron closely intertwines with PTEC metabolism. Although iron is necessary for metabolism, both iron deficiency and iron overload have been associated with dysregulated glucose metabolism and FAO inhibition in a cell-dependent manner. Modulating iron levels can alter the expression of enzymes involved in glycolysis and the TCA cycle¹⁶⁵. In the cardiac muscles of mice, iron overload attenuated glucose oxidation and electron transport chain activity¹⁶⁶. By contrast, iron deprivation enhanced glycolysis and blocked oxidative phosphorylation in human macrophages¹⁶⁷. The role of iron overload in the metabolism of renal cells in lupus nephritis has not been specifically investigated. However, the direct and indirect metabolic effects of iron overload in kidney diseases and its known contribution to lupus pathogenesis highlight iron and iron-regulated proteins as potential therapeutic targets. In addition to the effect of iron on energy metabolism, iron-induced oxidative stress and lipid peroxidation are key features of ferroptosis — an iron-dependent, regulated form of cell death driven by excessive lipid peroxidation, resulting in disrupted permeability and structure of the membrane bilayers^168,169. Lipid peroxidation can occur via selective enzymatic or random non-enzymatic free radical chain reactions^170,171, which represent potential targets for intervention. Neutrophils¹⁶², T cells¹⁷² and B cells¹⁶³ undergo ferroptosis in SLE and lupus nephritis, which is thought to contribute to disease pathogenesis.

Cell crosstalk

Alterations in cellular metabolism can contribute to the dysregulation of cell-to-cell crosstalk networks, which can exacerbate the progression of kidney disease¹⁷³. Within glomeruli, podocytes, endothelial cells and mesangial cells engage in tridirectional crosstalk¹³¹. For example, injury to podocytes induces mesangial cell proliferation, and, conversely, mesangial cell injury can lead to podocyte effacement and fusion. Similarly, the secretion of endothelin 1 by podocytes induces glomerular endothelial cell dysfunction¹⁷⁴, whereas endothelial cell-derived PDGFβ is critical for mesangial cell survival¹⁷⁵. As the metabolic signature of these three types of resident glomerular cell is altered in lupus nephritis, it is likely that injury to cells of one type influences the metabolic signature of neighbouring cells. The tubular compartment also engages in a crosstalk with cells of different types via the local release of signalling molecules, cytokines and growth factors¹⁷³. Although this crosstalk plays out in normal physiology, whether injury that changes the secretome of a particular tubule segment subsequently influences the metabolic profile of adjacent cells in the context of lupus nephritis remains unknown.

Similarly, crosstalk between infiltrating immune cells and kidney parenchymal cells might alter the metabolism of kidney parenchymal cells and contribute to tissue injury in lupus nephritis¹⁷⁶. Cytokines secreted by immune cells in the glomerular and tubulointerstitial compartments induce cellular damage. In addition, these cytokines might also directly affect renal cell metabolism. For example, IFNγ-producing CD4⁺ T cells localize in close proximity to renal tubules in patients with lupus nephritis¹⁷⁷. IFNγ profoundly alters the metabolic profile, basal respiration and ATP production of human PTECs¹⁷⁸. Other cytokines secreted by the infiltrating immune cells might also alter renal cell metabolism and influence the outcomes of lupus nephritis.

Other autoantibody-mediated kidney diseases

IgAN is a chronic glomerular disease that results from the deposition of IgA immune complexes and the induction of mesangial cell proliferation¹⁷⁹. IgA class switching depends on HIF1α-induced glycolysis, which produces acetyl-CoA for histone H3K27 acetylation at the Sα region¹⁸⁰. We would expect this pathway to be hyperactivated in patients with IgAN. IgA is known to shape the microbiome and its metabolic output. Microbial dysbiosis has been reported in patients with IgAN, including an increased abundance of Akkermansia muciniphila¹⁸¹. This mucin-degrading bacterium deglycosylates IgA1 (Gd-IgA1), which facilitates Gd-IgA1 retrotranscytosis out of the gut into the circulation and promotes its deposition in glomeruli. This mechanistic link between the gut–kidney axis and IgAN may be just one example, given that microbial metabolites — specifically short-chain fatty acids (SCFAs) — can reprogram B cell metabolism and modulate host antibody production¹⁸². This proposal is supported by the finding that the gut microbiome of patients with IgAN have a reduced abundance of bacterial taxa that produce butyrate — a SCFA with strong anti-inflammatory activity¹⁸³. Once in the kidney, Gd-IgA1 promotes mesangial cell proliferation by inducing glycolysis through a complex mechanism that involves a circular RNA and a microRNA, and promotes the translation of the glycolytic enzyme PKM2 (ref. 184). A scRNA-seq analysis of renal biopsy samples from patients with IgAN showed profound metabolic shifts in renal cell types other than glomerular cells, most predominantly in tubular cells, with disease progression to kidney failure¹⁸⁵. These findings require functional validation and further investigation to determine the extent to which they are specific to IgAN. Of note, a 2025 transcriptomics study found that T cells in biopsy samples from patients with IgAN demonstrate a metabolic phenotype similar to that of T cells from kidneys of patients with lupus nephritis⁹⁵; again, further analyses are necessary to validate this finding.

AAV encompasses a group of autoimmune conditions that affect the microvasculature, including in the kidney¹⁸⁶. ANCA autoantibodies, specifically anti-MPO antibodies, target neutrophils, which are abundant in the circulation. As with IgAN, it is unknown whether the production of anti-MPO IgG involves the metabolic reprogramming of B cells and T_FH cells that leads to the production of autoantibodies in lupus nephritis. However, increased oxidative stress and mTOR activation has been reported in dysfunctional peripheral FOXP3⁺CD4⁺ T regulatory (T_reg) cells in patients with AAV¹⁸⁷. Although the impaired suppressive activity of these T_reg cells could contribute to autoantibody production, it is also possible that their metabolic and functional alterations are a result, rather than a cause, of immune dysfunction. Given the central role of neutrophils in AAV and the strong evidence that metabolic programming influences neutrophil function, it is very likely, but not proven, that the metabolism of ANCA-stimulated neutrophils is altered. Interestingly, a 2020 study found that anti-MPO IgG increases glycolysis and oxidative phosphorylation in monocytes, and that these metabolic changes are required for the production of pro-inflammatory IL-1β¹⁸⁸. These findings were obtained with human cells in vitro and need to be functionally validated; however, they could have physiological relevance given the known contribution of monocytes to the pathogenesis of AAV, especially ANCA-induced glomerulonephritis¹⁸⁹. Whether metabolic alterations occur in kidney cells in ANCA-induced glomerulonephritis is also unknown. However, lipid peroxidation, which as discussed above is a driver of ferroptosis, was observed in kidneys and endothelial cells of a mouse model of AAV. Inhibiting ferroptosis prevented endothelial cell death induced by ANCA-activated neutrophils¹⁹⁰, and thus provides a potential therapeutic intervention.

Thus, patchy evidence supports the potential for cellular metabolic alterations in IgAN and AAV, not only in immune cells, but also in renal parenchymal cells in IgAN, and potentially in monocytes in AAV (Fig. 4). Despite the technical challenges in studying these two diseases, including the inadequacy of animal models, further investigations into the metabolic alterations of immune cells and target tissues should reveal whether they are disease-specific or follow similar patterns to those observed in lupus nephritis.

Fig. 4 |. Metabolic regulation in IgA nephropathy and ANCA-associated vasculitis.

a, IgA nephropathy (IgAN) is a chronic glomerular disease that results from the deposition of IgA-containing immune complexes in the mesangium, which induces mesangial cell proliferation by promoting glycolysis. IgA class switching itself depends on HIF1α-induced glycolysis in germinal centre (GC) B cells, leading to the accumulation of acetyl-CoA, which is essential for histone H3K27 acetylation at the Sα region. The increased abundance of Akkermansia muciniphila in the dysbiotic microbiome of patients with IgAN produces deglycosylated IgA1 (Gd-IgA1). This process facilitates the retrotranscytosis of Gd-IgA1 out of the gut into the circulation, leading to its deposition in glomeruli. In addition, the gut microbiome of individuals with IgAN have a reduced abundance of short-chain fatty acid (SCFA)-producing bacterial taxa, which is expected to promote inflammation. b, Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) damages the renal microvasculature. Circulating regulatory T (T_reg) cells in individuals with AAV demonstrate elevated levels of oxidative stress and mTOR activation, which is associated with impaired suppressive activity, and possibly contributes to autoantibody production. Antimyeloperoxidase (Anti-MPO) IgG increases glycolysis and oxidative phosphorylation in monocytes, leading to the production of pro-inflammatory IL-1β. Neutrophil metabolism might also undergo changes following stimulation with ANCA.

Cellular metabolism as therapeutic target

The most salient challenge of using cellular metabolism, especially glycolysis, for therapeutic purposes is the requirement to target pathogenic immune cells or parenchymal cells while preserving protective and homeostatic immune responses. This issue has been addressed directly in a model of multiple sclerosis, whereby pathogenic T_H17 cells were found to rely entirely on glycolysis whereas homeostatic microbiome-driven T_H17 cells in the gut compensated for the loss of glycolysis through mitochondrial respiration, enabling specific therapeutic targeting of the glycolytic, pathogenic T_H17 cells¹⁹¹. Whether T_H17 cells in nephritic kidneys and T_FH cells in secondary lymphoid organs in SLE could also be eliminated by inhibiting glycolysis without impairing gut health and/or the ability of the immune system to respond to microbial challenges remains to be investigated⁴¹.

Several therapeutic approaches that target metabolic pathways have demonstrated efficacy in preclinical models of lupus nephritis. Treatment of lupus-prone mice with an inhibitor of HIF1α attenuated kidney injury to a greater extent than T cell-specific deletion of Hif1a⁷³. This difference in outcome is because the inhibitor also reduced systemic autoimmunity, most probably due to its effect of germinal centre B cells and LLPCs in the bone marrow, both of which function in a hypoxic environment. The HIF1α pathway is also critical for cytotoxic T cells in cutaneous lupus erythematosus, and pharmacological inhibition of HIF1α also reduced skin lesions and systemic immunity in lupus-prone mice¹⁹². These two studies indicate that HIF1α is a valid target for reducing organ damage in patients with SLE, and further studies to evaluate the safety of treatment with these inhibitors in patients are therefore warranted.

Supplements such as fish oil, polyenylphosphatidylcholine, taurine, fumaric acid and salbutamol alter metabolic pathways that either inhibit the polarization of macrophages towards an M1 phenotype or promote M2 polarization¹⁹³, which might benefit patients with lupus nephritis. However, the classic M1–M2 macrophage dichotomy is probably an oversimplification, with available data suggesting that macrophages in the nephritic kidney may be more complex²⁸. Targeting mitochondrial oxidative stress might also represent a valuable approach in autoimmune diseases. Treatment of MRL/lpr lupus-prone mice with the mitochondria-targeted antioxidant MitoQ or CoQ10 analogue, idebenone, reduced ROS production, serum levels of type I IFN, immune complex deposition and renal pathology^80,194.

Therapeutic targeting of metabolic pathways in kidney resident cells represents a particular challenge, as under physiological settings, podocytes rely on glycolysis, whereas the PTECs are dependent on FAO. These different programs could be targeted by exploiting temporal differences in the involvement of different cell types. For example, glomerular injury typically precedes tubular injury and these different compartments could therefore potentially be targeted by timing early and late interventions with different metabolic inhibitors. The changes in lipid profile during the evolution of lupus nephritis could provide a mechanism with which to detect the progression of nephritis. Voclosporin — a second-generation clinical calcineurin inhibitor that is approved for use in combination with immunosuppressive therapy for the treatment of active lupus nephritis in adult patients¹⁹⁵ — reduces free fatty acids, acylcarnitine and complex lipids with the exception of phosphatidylinositol¹⁹⁶. Participants in the trial who achieved a complete renal response demonstrated a decrease in levels of acylcarnitine, triacylglycerols, phosphatidylcholines, phosphatidylethanolamines and sphingomyelins. However, whether lipid accumulation is a cause or a consequence of PTEC pathology remains unclear¹⁹⁷.

Sodium–glucose cotransporter 2 (SGLT2) inhibitors have also shown promising results in lupus nephritis. Treatment of MRL/lpr mice with the SGLT2 inhibitor empagliflozin for 10 weeks reduced levels of serum anti-dsDNA IgG, creatinine and proteinuria, along with attenuated glomerular and tubulointerstitial pathology¹⁹⁸. SGLT2 expression was increased in the podocytes of patients with lupus nephritis and in nephritic MRL/lpr mice. Mechanistically, empagliflozin alleviated podocyte injury by attenuating inflammation and enhancing autophagy by reducing mTORC1 activity. However, an independent study showed no therapeutic effect of empagliflozin in MRL/lpr mice treated for 4 weeks¹⁹⁹. Of note, retrospective analyses of SGLT2 inhibitors in lupus nephritis also suggest a beneficial effect. In one study, patients with SLE and type 2 diabetes who were treated with SGLT2 inhibitors had a lower risk of developing lupus nephritis and a lower risk of all-cause mortality compared with patients not using SGLT2 inhibitors²⁰⁰. This renoprotective effect of SGLT2 inhibitors was confirmed in a large cohort of patients with SLE²⁰¹. However, a degree of caution is warranted, as two studies have demonstrated that the beneficial effects of SGLT2 inhibitors in mouse models of heart failure might relate to off-target effects^202,203. Hence, large-scale, long-term studies are needed to confirm the effects of SGLT2 inhibitors in lupus nephritis, and identify the underlying mechanisms of protection.

Emerging evidence has also shown a link between diet and immune cell metabolism²⁰⁴, although most available studies have only established the dietary origin of individual metabolites that regulate immune cell functions rather than demonstrate a link between diet and/or systemic metabolism and the metabolic reprogramming of immune cells²⁰⁵. Whether the skewing of immune cell metabolism towards a glycolytic phenotype as observed in patients and mouse models of lupus nephritis is linked to high systemic levels of glucose metabolism is unknown. For example, the metabolic syndrome is highly prevalent among patients with SLE, but whether the systemic metabolic disturbances associated with metabolic syndrome contribute to autoimmune pathogenesis, beyond effects on inflammation, is unknown²⁰⁶. Also unknown is whether the hyperglycaemia and hyperlipidaemia in patients with SLE and the metabolic syndrome increase glucose or lipid utilization by their immune cells. However, a high consumption of ultraprocessed food, in particular sugar or artificially sweetened food, has been associated with a 50% increased risk of SLE in the Nurses’ Health Study cohort²⁰⁷. Conversely, it has been shown that dietary interventions, such as a Mediterranean style diet, could be beneficial in patients with SLE, at least in part through modulating the immune system²⁰⁸. These studies highlight the need for both mechanistic and interventional studies to understand the links between systemic and cellular metabolism in immune-related diseases, and opportunities for therapeutic targeting.

Novel approaches to treat lupus nephritis could reduce dose dependency on traditional immunosuppression-based therapies. The hepcidin–ferroportin axis primarily regulates systemic iron metabolism²⁰⁹; however, a growing body of evidence supports its involvement in lupus nephritis²¹⁰. In line with animal studies that have demonstrated an important role for iron in SLE, deferiprone, a clinically approved iron chelator, delayed the onset of albuminuria and reduced blood urea nitrogen in NZB/W lupus-prone mice, despite the presence of immune complex deposits in the kidneys and serum autoantibodies¹⁵⁸. Administration of hepcidin also showed prophylactic and therapeutic benefits in delaying the onset and severity of spontaneous lupus nephritis in MRL/lpr mice²¹¹. Like deferiprone, hepcidin treatment did not reduce the deposition of immune complexes or serum autoantibodies in MRL/lpr mice; however, it mitigated the production of cytokines in the kidney, immune cell infiltration with direct effects on macrophages, and tubular injury without worsening lupus-associated anaemia. These data support the concept that drugs that normalize iron metabolism could be repurposed as adjunct therapeutics for lupus nephritis and reduce dosage and/or dependency on toxic, non-specific immunosuppressants. Along these lines, the induction of ferroptosis in human PTECs following exposure to serum from patients with lupus nephritis was inhibited by administration of the next-generation ferroptosis inhibitor, liproxstatin 2 (ref. 149), identifying ferroptosis as a druggable target for immune and parenchymal cells to improve outcomes in lupus nephritis.

However, several outstanding questions remain about the mechanisms by which ferroptosis contributes to lupus nephritis. For example, proliferative, autoreactive B cells depend on iron, fatty acid uptake and the active breakdown of lipid droplets to fuel the TCA cycle for ATP synthesis⁹⁰. The formation of lipid droplets is dependent on the acyl-CoA synthetase long-chain family member (ACSL)-mediated activation of fatty acids²¹². ACSL4 enhances intracellular lipogenesis and lipid droplet accumulation to promote FAO and ATP production by upregulating the FAO rate-limiting enzyme CPT1A²¹³. However, ACSL4 also enriches cellular membranes with long polyunsaturated ω6 fatty acids and is an essential driver of ferroptosis²¹⁴. Thus, activation of ACSL4 during lipid droplet formation and lipolysis might collaterally facilitate ferroptosis in B cells. Along similar lines, PTECs predominantly depend on FAO for their cellular energetics under physiological conditions¹³⁶. As discussed above, this pathway also generates intermediate metabolites that are essential for ferroptosis. Improved understanding of how PTECs shunt iron and fatty acids for enhanced metabolic needs while simultaneously preventing ferroptosis might help to identify novel cellular mechanisms that contribute to the pathogenesis of lupus nephritis. New-generation ferroptosis inhibitors such as FerroLOXINs that block enzyme–protein complexes²¹⁵, unlike traditional ferroptosis inhibitors such as liproxstatin 1 that suppress ferroptosis by trapping lipid-derived free radicals²¹⁶, may also be of use as an adjunct therapy to reduce dependency on toxic immunosuppressants.

Finally, chimeric antigen receptor T (CAR T) cell therapies are developing at a rapid pace and may have applications in immune-mediated kidney diseases²¹⁷. CD19-targeted CAR T cells (ref. 218) and, more recently, B cell maturation antigen-targeted CAR T cells²¹⁹ have been shown to deplete B cells and induce lasting remission, including renal manifestations, in patients with SLE. Multiple other trials in lupus nephritis, IgAN and AAV are ongoing²¹⁷. Strategies have also been proposed to improve the function of CAR T cells by altering their metabolism. For example, the intrinsic metabolic flux of CAR T cells affects their proliferation capacity, and T cell metabolism can be modified by reprogramming their extracellular single-chain variable fragments²²⁰. Conditioning CAR T cells with metformin and rapamycin to improve their mitochondrial function before their transfer increased their efficacy in the hypoxic brain in a preclinical model of glioma²²¹. As metformin and rapamycin can modify the metabolic profile of T cells^50,222, this protocol could be considered as a potential option for targeting T cells within the hypoxic kidney in lupus nephritis.

Conclusions

A strong body of evidence supports the notion that the metabolic programming of immune cells and renal parenchymal cells sustains the pathogenesis of autoimmune kidney diseases. The metabolic profile of cells has been best characterized for lupus nephritis; however, metabolic skewing has also been demonstrated in AAV and IgAN. Increased glycolysis and mitochondrial oxidation are two common themes that have emerged among the processes involved immune-related renal pathogenesis. This finding is of therapeutic relevance as drugs that inhibit glycolysis or mitochondrial oxidation are predicted to be renoprotective, as demonstrated in mouse models. Multiple mitochondrial antioxidant drugs are available, but safely targeting glycolysis is a challenge that has not yet been achieved in humans, despite great interest from the field of oncology. Despite heightened glycolysis and mitochondrial oxidation, a substantial amount of heterogeneity exists between cell types, which represents another challenge in therapeutic targeting. The cell-specific targeting of metabolic enzymes that can be achieved using genetic approaches in studies in mice is obviously not a foreseeable therapeutic goal for patients. The complexity of the metabolic landscape in autoimmune kidney diseases therefore presents a limitation to its effective therapeutic targeting. However, advances in technological approaches, such as spatial transcriptomics and metabolomics, combined with a growing interest in the mechanisms by which cellular metabolism drives immune-mediated pathogenesis are providing new insights into the drivers of immune-mediated kidney disease. Such insights could ultimately lead to the identification of markers of disease progression and possibly novel therapeutic targets.

Key points.

• Immune cells and parenchymal cells within the kidney undergo numerous metabolic alterations in autoimmune renal diseases, such as lupus nephritis, IgA nephropathy and ANCA-associated vasculitis.

• Kidney-infiltrating immune cells alter their metabolism to adapt to the inflammatory, hypoxic and hypertonic environment.

• Autoimmunity alters the metabolism of kidney cells, which enhances the pathogenicity of infiltrating immune cells.

• Enhanced glycolysis and mitochondrial oxidation are shared by most immune and kidney cell types involved in autoimmune renal diseases.

• Alterations in iron metabolism induce ferroptosis in both immune cells and kidney cells in autoimmune renal diseases.

• Insights into metabolic alterations in autoimmune renal diseases may offer opportunities for therapeutic targeting.