Beyond Redox Regulation

Abstract

Cellular stress, such as oxidative and endoplasmic reticulum (ER) stresses, contributes to the development of various kidney diseases. Oxidative stress is prompted by reactive oxygen species accumulation and delicately mitigated by glutathione and thioredoxin (Trx) antioxidant systems. Initially identified as a Trx-binding partner, Trx-interacting protein (TXNIP) is significantly up-regulated and activated by oxidative and ER stresses. The function of TXNIP is closely linked to its subcellular localizations. Under normal physiological conditions, TXNIP primarily localizes to the nucleus. When exposed to reactive oxygen species or ER stress, TXNIP relocates to mitochondria and binds to mitochondrial Trx2, which releases Trx-tethered apoptosis signal-regulating kinase 1 and activates apoptosis signal-regulating kinase 1–mediated apoptosis. Oxidative and ER stresses are also closely associated with autophagy. TXNIP can promote or inhibit autophagy depending on context. Although recent studies have highlighted the indispensable role of TXNIP in the etiology and progression of kidney disease, TXNIP-targeted therapy is still missing. This review focuses on the following: i) oxidative and ER stresses; ii) regulation and function of TXNIP during cellular stress; iii) TXNIP in stress-regulated autophagy; iv) TXNIP in kidney diseases (nephrotic syndrome, diabetic nephropathy and chronic kidney disease, acute kidney injury, and kidney aging); and v) novel treatment agents targeting TXNIP in kidney disease. Current advances in chemical compounds and RNA-based therapy suppressing TXNIP are also reviewed.

Cellular stress, including oxidative and endoplasmic reticulum (ER) stresses, is widely observed in human disease. Thioredoxin-interacting protein (TXNIP), also named as vitamin D₃ up-regulated protein 1, is a key regulator of both oxidative and ER stresses; it exhibits aberrant expression and function in a variety of disorders, including kidney disease, diabetes, and neurodegenerative disease. The kidney faces external and internal stresses as the specific organ for blood filtration, and kidney function hinges on the balance between adaptive and terminal stress responses. Thus, the delicate regulation of cellular stress response determines the fate of renal cells and is crucial for maintaining kidney homeostasis.

The aim of the current review was to provide an updated overview of the novel roles of TXNIP under oxidative and ER stresses and discuss the potential therapeutic benefits of targeting TXNIP for the treatment of kidney diseases.

Oxidative and ER Stresses

Oxidative stress emerges from disturbance of cellular redox homeostasis. Once redox homeostasis is disrupted, the excessive accumulation of reactive oxygen species (ROS) causes organelle injury, oxidative stress, and DNA damage, subsequently leading to cellular senescence, abnormal cell proliferation, or apoptosis. In response to such circumstances, eukaryotic cells evolve a thiol-dependent antioxidant system comprising glutathione and thioredoxin (Trx).¹ Glutathione is a small thiol-containing tripeptide found in all the eukaryotes and some prokaryotes, existing in either a thiol-reduced state [reduced glutathione (GSH)] or an oxidized state [oxidized glutathione (GSSG)] consisting of two GSH molecules connected by a disulfide bond. The designated ratio of GSH:GSSG is pivotal for maintaining the redox balance of the cytosol and subcellular compartments. Although the majority of glutathione is in the form of GSH, with a typical molecular ratio of 50:1 (GSH to GSSG) at the whole-cell level, recent advances reveal varying GSH:GSSG ratios in different subcellular organelles. The conversion from GSH to GSSG provides a reducing equivalent (H⁺+ e⁻) to neutralize ROS or preserve a reduced form of other targets.²

Trx constitutes another major thiol-dependent antioxidant system in mammalian cells, with isoforms Trx1 and Trx2 located in the cytoplasm and mitochondria, respectively. Both cytosolic Trx1 and mitochondrial Trx2 exhibit a typical Trx-fold structure containing a Cys-Gly-Pro-Cys motif in the active site. Within the active core of reduced Trx, two cysteine residues form an intermolecular disulfide with target proteins through a thiol–disulfide exchange reaction, whereas oxidized Trx is restored to its reduced form after capturing electrons from NADPH. Trx plays a crucial role in regulating cellular signaling through receptors, as well as different biological processes within specific cellular compartments, such as controlling cell growth and angiogenesis within the nucleus, modulating the redox state in the cytosol, and suppressing apoptosis signal-regulating kinase 1 (ASK1) activity in the cytoplasm and mitochondria. The increased ROS converts reduced Trx to its oxidized form and disassembles the Trx–ASK1 interaction, resulting in ASK1 activation and cell death.³

ER stress, also known as the unfolded protein response (UPR), is the imbalance between the load of misfolded proteins and the folding capacity of the ER. The overload of misfolded and unfolded proteins is recognized and bound by an ER chaperone, binding immunoglobulin protein. Once dissociated from binding immunoglobulin protein, the two ER membrane sensors, inositol-requiring enzyme 1α (IRE1α) and protein kinase R-like endoplasmic reticulum kinase (PERK), undergo activation through homodimerization and autophosphorylation of their ER luminal domains. Subsequently, active IRE1α processes mRNA of X-box binding protein 1 (XBP1) to produce the spliced form of XBP1. Concurrently, phosphorylation of eukaryotic initiation factor 2α by PERK leads to selective translation of activating transcription factor 4 (ATF4). In addition, during ER stress, activating transcription factor 6α (ATF6α) moves from the ER to the Golgi apparatus, where it releases its p50 N-terminal fragment via proteolytic cleavage by site I and II proteases.⁴ The spliced form of XBP1, ATF4, or p50 ATF6α then translocates to the nucleus to regulate expression of different downstream target genes, including ER chaperones, cellular stress modulators and cell death executors, depending on the severity and duration of stress conditions, which determine the outcome of the ER stress response.

Regulation and Function of TXNIP during Cellular Stress

TXNIP was first identified as a Trx-binding protein through yeast two-hybrid screening.⁵ Analysis of human TXNIP protein sequence reveals its homology to β-arrestins, which typically associate with the plasma membrane, where they promote endocytosis of G-protein–coupled receptors and diminish the signaling. TXNIP is one of at least six members of the novel α-arrestin family but is unique in that its active cysteine residues required for binding Trx are not present in other members of the family. Briefly, in the protein sequence of TXNIP, cysteine 247 largely mediates the redox-sensitive functions of TXNIP by forming an intramolecular mixed disulfide bond with the Trx catalytic site (Figure 1). Conserved Rous sarcoma oncogene homology 3 and Proline-Proline-x-Tyrosine motifs are shared among α-arrestin members and mediate protein–protein interactions. The N-terminal conserved Rous sarcoma oncogene homology 3 domains of TXNIP bind to conserved Rous sarcoma oncogene, and Proline-Proline-x-Tyrosine in the C-terminus interacts with an E3 ubiquitin ligase, Itch, that facilitates TXNIP degradation by the proteasome⁶ (Figure 1). In addition, TXNIP contains both an immunoreceptor tyrosine-based inhibition motif and chromatin maintenance region 1. The immune-receptor tyrosine-based inhibition motif domain interacts with tyrosine phosphatases and regulates important cellular responses. The chromatin maintenance region 1 domain associates with the β-domain of von Hippel-Lindau protein and enhances the interaction between von Hippel-Lindau protein and transcriptional factor hypoxia-inducible factor-1α to promote the nuclear export and cytosolic degradation of hypoxia-inducible factor-1α⁷ (Figure 1). Thus, beyond Trx regulation, additional roles of TXNIP in cell survival, inflammatory response, metabolism, and transcriptional regulation mediated by these various TXNIP-binding partners have been unveiled.

Figure 1.

Structural analysis of thioredoxin-binding protein (TXNIP)-binding domains. CRM1, chromatin maintenance region 1; HIF-1α, hypoxia-inducible factor-1α; ITIM, immunoreceptor tyrosine-based inhibition motif; PPxY, Proline-Proline-x-Tyrosine; ROS, reactive oxygen species; SH3, Src homology 3; Src, conserved Rous sarcoma oncogene; Trx, thioredoxin; VHL, von Hippel-Lindau.

TXNIP and Oxidative Stress

TXNIP represses the reductase activity of Trx by interacting with its catalytic core and thus exacerbates oxidative stress.⁵ The redox active Cys32 and Cys35 residues of Trx and the reduced form of Trx are critically important for the TXNIP–Trx interaction. When TXNIP is induced by 1,25-dihydroxyvitamin D₃, the reducing activity and expression level of Trx are diminished in the HL60 leukemia cell line.⁵ Moreover, once TXNIP binds to Trx, Trx-tethered substrates such as ASK1 are subsequently released and exert their biological functions.

The function of TXNIP during oxidative stress is closely linked to its subcellular localizations in different cells.⁶ For example, in pancreatic β cells, under normal physiological conditions, TXNIP primarily localizes to the nucleus, whereas hydrogen peroxide treatment results in translocation of TXNIP from the nucleus to the mitochondria. In the mitochondria, TXNIP binds to and oxidizes Trx2, leading to dissociation between Trx2 and ASK1, and subsequent activation of ASK1-induced cell death.⁸ In endothelial cells, TXNIP relocation from the nucleus to the plasma membrane protects against oxidative stress–induced apoptosis.⁹ It has been shown that poly-ADP-ribose polymerase 1 (PARP1) regulates the TXNIP subcellular localization. Under basal conditions, PARP1 acts as a scaffold to retain TXNIP in the nucleus. In response to hydrogen peroxide, ADP-ribose transfer activity of PARP1 increases, and poly-ADP ribosylation of PARP1 maintains TXNIP in the nucleus. Conversely, PARP1 inhibition and the decrease in PARP1 poly-ADP ribosylation dissociates TXNIP from PARP1, allowing it to travel to the plasma membrane. At the plasma membrane, TXNIP promotes tyrosine phosphorylation of vascular endothelial growth factor receptor 2 and enhances the survival of endothelial cells.⁹

TXNIP and ER Stress

TXNIP is modulated by ER stress and plays a crucial role in ER stress response as well. In pancreatic β cells, the TXNIP level is elevated by the PERK and IRE1α arms of the UPR.^10,11 At the transcriptional level, TXNIP expression is promoted by PERK-regulated carbohydrate response element-binding protein (ChREBP) and ATF5.¹¹ Meanwhile, Lerner et al¹⁰ elucidated that IRE1α modulates TXNIP expression posttranscriptionally, which is potentially supplementary to the transcriptional modulation by PERK. Their study showed that the stability of TXNIP mRNA was maintained by IRE1α by diminishing the level of miRNA-17 that directly bound to 3′-untranslated regions of TXNIP mRNA and caused its degradation.¹⁰ Hyperactivated IRE1α elevated TXNIP, which induced Nod-like receptor protein 3 (NLRP3) inflammasome activation and IL-1β secretion, ultimately leading to apoptosis of pancreatic β cells.^10,11

A recent study involved mouse models of genetic nephrotic syndrome (NS) caused by deletion of the glomerular basement membrane component laminin β2 (Lamb2 ^−/− mice) and NS acquired after injection with adriamycin.¹² In this study, TXNIP exhibited increased abundance in glomeruli in a podocyte-distribution pattern, as well as in renal tubules, associated with albuminuria-induced selective activation of p50 ATF6 and induction of C/EBP homolog protein (CHOP), a transcription factor mediating ER stress–induced proapoptotic signaling. Using the Assay for Transposase-Accessible Chromatin with high-throughput sequencing and the JASPAR database (https://jaspar.genereg.net, last accessed December 10, 2024), our bioinformatics analysis revealed putative CHOP binding sites in the promoter of Txnip gene.¹² Furthermore, CHOP up-regulation induced by albuminuria drove TXNIP shuttling from the nucleus to the mitochondria, where it is required for the induction of mitochondrial ROS (mROS). The increased mROS accumulation oxidized mitochondrial Trx2, thus liberating TXNIP to associate with mitochondrial NLRP3 to activate inflammasome; it also released mitochondrial ASK1 to induce mitochondria-dependent apoptosis (Figure 2).¹²

Figure 2.

Scheme depicting that the C/EBP homolog protein (CHOP)–thioredoxin-binding protein (TXNIP) signaling axis modulates mitochondria-dependent Nod-like receptor protein 3 (NLRP3) inflammasome activation and mitochondrial phosphorylation (P) apoptosis signal-regulating kinase 1 (ASK1)-mediated intrinsic apoptosis in nephrotic syndrome. ATF6, activating transcription factor 6; ER, endoplasmic reticulum; ROS, reactive oxygen species; Trx2, thioredoxin 2. Image is reprinted with permission from Park et al.¹²

Conversely, TXNIP expression can be suppressed by UPR branches. For instance, in breast cancer cells, estrogen activates the IRE1α pathway, which down-regulates TXNIP expression.¹³ Knockdown of TXNIP enhances glucose uptake and facilitates aerobic glycolysis (the Warburg effect) to support rapid proliferation of breast cancer cells, whereas forced overexpression of TXNIP suppresses glucose uptake and the Warburg effect. Thus, TXNIP links UPR to reprogrammed glucose metabolism by estrogen in breast cancer cells.¹³

Reciprocally, TXNIP exerts a regulatory control over the ER stress response. For example, TXNIP interacts with protein disulfide isomerase in the ER, enhancing its enzymatic activity.¹⁴ It has also been shown that TXNIP deletion in cells and animals increases XBP1 splicing under both basal and ER stress conditions. Thus, TXNIP can inhibit ER stress. Conversely, TXNIP can also promote ER stress. In a mouse model of fasting-induced liver steatosis, fasting prompted up-regulation of TXNIP and activation of ER stress, such as XBP1 splicing and IRE1α phosphorylation, and TXNIP knockout in the mouse liver attenuated activated ER stress.¹⁵ Similarly, in multiple special diet–induced nonalcoholic steatohepatitis mouse models, TXNIP and CHOP proteins accumulated in the liver. Mechanistically, the C-terminus of TXNIP associates with the N-terminus of the α-helix domain of CHOP and decreases CHOP ubiquitination, thus increasing the stability of CHOP protein.¹⁶

TXNIP and Stress-Regulated Autophagy

Autophagy is a critical cellular stress response. It is a conserved lysosome-dependent process that removes damaged and dysfunctional cellular components. During autophagy, a portion of the cytoplasm or specific cellular components is enclosed into a double-layer membrane vesicle known as the autophagosome.¹⁷ Autophagosome formation involves several key steps, including: phagophore initiation and nucleation that requires Unc-51-like kinase (ULK)1/2/FIP200/Atg13 and phosphoinositide-3-kinase (PI3K)/Vps34/Vps15/beclin 1, respectively; vesicle elongation forming autophagosomes that requires microtubule-associated proteins 1A/1B light chain 3B (LC3B) and Atg5-Atg12-Atg16 ubiquitin-conjugation systems; autophagosome maturation; and autophagosome-lysosome fusion forming autolysosomes where cargo is degraded and recycled.¹⁷ The autophagy receptor p62 binds to both polyubiquitinated substrate proteins and LC3B attached to phagophores, which enables sequestration of the ubiquitinated proteins/p62 in the autophagosomes and being degraded. Adenosine-monophosphate activated-protein kinase (AMPK) and mammalian target of rapamycin (mTOR) are important positive and negative regulators of autophagy, respectively.

Autophagy can be regulated by oxidative and ER stresses. ROS accumulation causes oxidative damage of cellular components, such as DNA, protein, or lipid, triggering the autophagic process. Excessive ROS can also oxidize critical cysteine residues in several key autophagy-related proteins. The first is AMPK. Oxidation of cysteine residues in the α and β subunits of 5′-AMPK by ROS leads to AMPK activation. Active AMPK promotes assembly of ULK1 complex by phosphorylating ULK1 and inhibiting mTOR complex 1 (mTORC1).¹⁸ The second protein is p62. Redox-sensitive Cys103 and Cys105 residues in p62 can be oxidized by ROS, thus disrupting p62 oligomerization and p62–LC3 interaction to affect vesicle elongation.¹⁹ The third protein is Atg4, which cleaves and primes Atg8 homologues (LC3, GATE-16, and GABARAP in mammals) for conjugation with phosphatidylethanolamine, allowing for lipidated Atg8 incorporation into the autophagosome membrane. When Atg8-phosphatidylethanolamine is deconjugated by Atg4, the Atg8 is recycled. Oxidation of the Cys81 residue of Atg4 by hydrogen peroxide inhibits its cysteine protease activity, thereby promoting lipidation of Atg8 and autophagy.²⁰

ER stress and autophagy are also dynamically interconnected, and the three UPR branches can regulate autophagy through Akt-mTOR, Atgs, beclin-1, AMPK, and so forth.²¹ First, after PERK activation, ATF4 up-regulates transcription of CHOP, sestrin 2, and REDD1 (regulated in development and DNA damage response 1). CHOP facilitates subsequent tribbles pseudokinase 3 expression, which reduces Akt phosphorylation and downstream mTOR activation.²¹ Sestrin 2 and REDD1 bind to mTOR and inhibit its activity. ATF4 and CHOP also control transcription of several Atgs.²² Second, IRE1α activation by tumor necrosis factor associated factor 2 leads to Bcl-2 phosphorylation and subsequent release of beclin-1, which is involved in the pre-autophagosomal assembly and autophagy initiation.²³ Third, ATF6α activation in the heart during cardiac hypertrophy increases transcription of Ras homolog enriched in brain, leading to mTORC1 activation and inhibition of the ULK1 complex assembly.²⁴ In addition, ER stress integrates all UPR branches at AMPK, which is activated by IRE1α-dependent up-regulation of ribosomal protein S6 kinase, PERK/ATF6-CHOP–mediated ATP depletion, and ER calcium leakage-prompted activation of calcium/calmodulin-dependent protein kinase kinase 2.²¹

TXNIP can positively regulate autophagy through interaction with the stress-responsive protein REDD1.²⁵ REDD1 expression can be induced by multiple stress conditions and their downstream transcription factors.²⁶ Transcription factor hypoxia-inducible factor-1α, activated by energy stress, hypoxia, or ischemia, binds directly to the promoter of REDD1 gene and promotes its transcription.²⁷ REDD1 expression is also up-regulated by ATF4, which responds to the oxidative stress inducer, hydrogen peroxide, and ER stress inducers, homocysteine and a non-metabolizable synthetic glucose analog, 2-deoxyglucose, that disrupts glycolysis and depletes energy.^28,29 REDD1 is a negative regulator of mTORC1 through binding to 14-3-3 protein, which dissociates 14-3-3 from tuberous sclerosis complex 2 to suppress mTOR activity. The C-terminus of TXNIP directly interacts with REDD1, which inhibits REDD1 protein degradation during treatment with 2-deoxyglucose. Thus, TXNIP potentiates REDD1-induced mTOR suppression through stabilization of REDD1 in response to metabolic stress.²⁸ Similarly, under hypoxic/ischemic stress, TXNIP/REDD1 are induced and form a robust physical complex, which increases autophagosome formation, and excessive accumulation of autophagosomes in the cytosol contributes to autophagic cardiomyocyte death in mouse myocardial ischemia/reperfusion injury.³⁰ It has also been shown that the TXNIP/REDD1 complex is both necessary and sufficient for ROS induction in response to stress, which suppresses Atg4-mediated LC3 delipidation and thus promotes autophagosome maturation and autophagy.³¹

Besides REDD1, TXNIP can inhibit mTOR signaling through other mediators in response to metabolic stress. In diabetic retinopathy, TXNIP overexpression enhances autophagic flux, as evidenced by increased expression of beclin-1 and Atg12-Atg5, by deactivating PI3K/Akt/mTOR signaling in rat retina Müller glial cells.³² The excessive autophagy in the presence of hyperglycemia is associated with activation of apoptosis in Müller cells. Meanwhile, in nonalcoholic steatohepatitis caused by ectopic fat accumulation in the liver owing to increased fat uptake, de novo lipid synthesis, and reduced fatty acid β-oxidation, elevated TXNIP directly interacts with and stimulates AMPK to inactivate mTORC1, as well as promotes cytosol to nuclear translocation of transcription factor EB to induce autophagy in hepatocytes.³³ TXNIP deletion impairs autophagy induction, along with attenuated fatty acid oxidation due to increased activity of mTORC1, thus exacerbating hepatic steatosis, inflammation, and fibrosis.³³

The effect of TXNIP on autophagy is disease-context dependent. As detailed in TXNIP in DN and CKD, TXNIP inhibition enhances autophagy in the murine model of diabetic nephropathy (DN).

TXNIP in Kidney Diseases

Due to the important role of cellular stress in renal injury and the critical function of TXNIP as a key mediator of cellular stress responses, TXNIP has been investigated in various renal diseases, including NS, DN and chronic kidney disease (CKD), acute kidney injury (AKI), and kidney aging. We discuss here the novel roles of TXNIP in these kidney diseases.

TXNIP in NS

Characterized by pronounced albuminuria/proteinuria stemming from various etiologies of glomerular disease, NS leads to glomerulosclerosis, interstitial fibrosis, and the gradual deterioration of renal function.³⁴ NS also stands out as one of the leading causes of CKD, which affects approximately 800 million people worldwide, and is associated with substantial morbidity and mortality rates.³⁵ Primary NS is caused by focal segmental glomerulosclerosis, membranous nephropathy, and minimal change disease.

To demonstrate the relevance of TXNIP in human proteinuric kidney disease, we mined gene expression profiles in Nephroseq (https://www.nephroseq.org, last accessed December 10, 2024), which showed that TXNIP expression was significantly up-regulated in either glomeruli or tubulointerstitium in various kidney diseases, including CKD, glomerulosclerosis, and DN. Notably, TXNIP mRNA levels in glomeruli and tubules exhibited a positive correlation with severity of proteinuria in human focal segmental glomerulosclerosis kidney samples. Similarly, the tubular TXNIP transcript level showed a strong positive correlation with proteinuria in human membranous nephropathy patients. More importantly, in patients with NS, a strong negative association was identified between tubulointerstitial TXNIP gene expression and a kidney function marker, estimated glomerular filtration rate.¹²

Furthermore, by using the Lamb2 and Chop double knockout mouse model, we proved that CHOP deletion abolished TXNIP up-regulation and its shuttling to mitochondria in response to albuminuria in renal tubular cells.¹² Subsequently, abrogation of CHOP-TXNIP signaling mitigated mROS overproduction and mROS-dependent NLRP3 inflammasome activation, as well as dampened ASK1-mediated mitochondrial apoptosis in albuminuria. In this study, for the first time, we used noninvasive, sensitive, and quantitative positron emission tomography/computed tomography molecular imaging to detect mROS in live mice by employing the newly developed mROS radiotracer ⁶⁸Ga-Galuminox.¹² Consequently, CHOP depletion mitigated albuminuria, improved renal function, and increased the lifespan of Lamb2 knockout mice.¹² Thus, we have identified CHOP-TXNIP as critical molecular connections between ER dysfunction and mitochondrial dyshomeostasis in NS.

TXNIP in DN and CKD

DN is one of the most prevalent microvascular complications of diabetes mellitus and a leading cause of CKD and end-stage renal disease. DN is primarily driven by sustained hyperglycemia and accumulation of advanced glycation end-products, which trigger a cascade of pathologic changes in different cell types of the kidneys. Podocytes are specialized cells in the glomerulus, and podocyte slit diaphragm is a major component of the glomerular filtration barrier that filters blood and prevents albuminuria. Mesangial cells support the structural integrity of the glomerulus. Renal tubular cells are essential for the reabsorption of nutrients, electrolytes, and water. In DN, chronic hyperglycemia, proteinuria, hypoxia, and inflammation cause metabolic and oxidative stresses, which lead to podocyte injury and loss,³⁶ mesangial cell hyperplasia and expansion,³⁷ and maladaptive responses of tubular cells that contribute to the progressive decline in kidney function and the development of CKD.³⁸

TXNIP induction is implicated in the pathogenesis of DN. In normal human and rat kidneys, expression of TXNIP mRNA and protein is most abundant in the glomeruli and distal nephron (distal convoluted tubules and collecting ducts), whereas Trx mRNA and protein expression is confined to the renal cortex, particularly within the proximal tubules.³⁹ In patients with DN, TXNIP levels in the kidneys are significantly increased. In contrast, kidney Trx levels do not differ between DN kidney biopsies and normal controls. Notably, TXNIP is the most significantly up-regulated gene in the transcriptional profiles of proximal tubular HK-2 cells exposed to high glucose (HG),⁴⁰ and kidney TXNIP induction precedes any structural and functional manifestations of DN.⁴¹ Moreover, TXNIP mRNA elevation in the urinary sediment of patients with type 1 DN is associated with renal function decline.⁴² Importantly, in the mouse model of streptozotocin-induced DN, TXNIP deficiency protects against podocyte foot process effacement and albuminuria, as well as attenuates renal fibrosis and inflammation.⁴³ Mechanistically, TXNIP deletion abolishes augmented renal ROS levels in diabetic mice.

Several key transcription factors governing renal TXNIP expression have been identified. Krüppel-like factor 6 and peroxisome proliferator-activated receptor-γ have opposite effects on TXNIP promoter activity in response to HG.⁴¹ Krüppel-like factor 6 is up-regulated in diabetic kidneys, and it directly binds to TXNIP promoter and positively regulates TXNIP expression in HK-2 cells in vitro.⁴¹ Conversely, peroxisome proliferator-activated receptor-γ mRNA is slightly decreased in the kidneys of diabetic rats. Peroxisome proliferator-activated receptor-γ agonists, both pioglitazone and rosiglitazone, attenuate HG-induced TXNIP promoter activity, as well as TXNIP mRNA and protein expression in HK-2 cells. In line with this, peroxisome proliferator-activated receptor-γ silencing in the cells significantly increases TXNIP mRNA levels.⁴¹ In addition, fructose is a significant risk factor for the development of diabetes and up-regulates expression of ChREBP-β, a key transcription factor that regulates fructose metabolism. In renal tubular epithelial cells, ChREBP-β significantly induces TXNIP expression by binding to its promoter region. Furthermore, the ChREBP-β/TXNIP axis mediates high fructose–induced ferroptosis, a regulated cell death mode caused by iron-dependent lipid peroxidation. By decreasing ChREBP-β levels, metformin can inhibit TXNIP transcription and alleviate high fructose–induced ferroptosis and renal tubular epithelial degeneration.⁴⁴

Hyperglycemia-stimulated renal TXNIP expression also involves epigenetic mechanisms. A genetic DN model, Sur1-E1506K mouse, which harbors a knock-in mutation equivalent to the human mutation (E1506K) in the sulfonylurea receptor 1 to cause a reduction of K_ATP channel activity, was used to determine the epigenetic regulation of Txnip.⁴⁵ De Marinis et al⁴⁶ showed that in kidneys from homozygous mutant mice, which develop diabetes due to impaired insulin secretion with age, hyperglycemia-induced Txnip expression was linked with an increase in activating histone marks H3K9ac, H3K4me3, and H3K4me1, as well as a decrease in the repressive histone mark H3K27me3 at the promoter region of the gene. The same effect of glucose on TXNIP expression and histone modifications at the TXNIP promoter was observed in normal human mesangial cells. Moreover, in normal human mesangial cells, glucose-augmented histone acetylation and TXNIP up-regulation are reversed by inhibition of histone acetyltransferase at H3K9ac or enhanced by inhibition of histone deacetylase.⁴⁶ In agreement with these in vivo and in vitro findings, deacetylation of Txnip H3K9ac, mediated by a histone deacetylase sirtuin 1, is responsible for attenuation of DN induced by a high-fat diet after treatment with a glucagon-like peptide 1 receptor agonist, exenatide.⁴⁷ Meanwhile, up-regulation of the histone methyltransferase enzyme enhancer of zeste homolog 2 (EZH2) increases H3K27me3 levels, down-regulates TXNIP, and attenuates oxidative stress in podocytes exposed to HG. Furthermore, enhancer of zeste homolog 2 represses transcription factor Pax6 through H3K27 trimethylation at its promoter and, hence, Txnip expression in podocytes.⁴⁸ Conversely, pharmacologic or genetic depletion of enhancer of zeste homolog 2 enhances TXNIP expression and oxidative stress in cultured podocytes treated with HG and glomeruli isolated from diabetic rats.⁴⁸ In concert with these findings, in human DN, it has been found that aberrant proximal tubule DNA methylation is related to kidney dysfunction. In the micro-dissected proximal tubules from patients with DN, three CpG sites annotated to TXNIP are demethylated, and the DNA methylation level of TXNIP, which indicates silencing of the gene, is positively associated with the estimated glomerular filtration rate.⁴⁹

Besides regulating ROS, TXNIP is involved in the aberrant lipid accumulation in DN through activation of Akt/mTOR signaling.⁵⁰ Loss of TXNIP abrogates increased lipid accumulation and expression of key proteins in de novo lipogenesis, including acetyl-CoA carboxylase, fatty acid synthase, and sterol regulatory element binding protein-1 in diabetic kidneys.⁵⁰ Meanwhile, TXNIP knockout also significantly promotes fatty acid oxidation and increases expression of peroxisome proliferator-activated receptor-α, acyl-coenzyme A oxidase 1, and carnitine palmitoyltransferase 1 in diabetic kidneys.⁵⁰ Using HK-2 cells, in vitro experiments show that blockade of the Akt/mTOR signaling pathway with a specific PI3K inhibitor replicates the effects of TXNIP silencing on lipid metabolism.⁵⁰

In HG-exposed podocytes, up-regulated TXNIP activates NLRP3 inflammasome by directly interacting with NLRP3,⁵¹ and blocking NLRP3 inflammasome activation by caspase-1 inhibitor prevents IL-1β production and eventually attenuates podocyte and glomerular injury in DN.⁵² At the same time, NADPH oxidase is also involved in NLRP3 inflammasome activation in HG-stimulated podocytes, and TXNIP is required for expression of the membrane-bound gp91^phox subunit of NADPH oxidase in HG-treated podocytes.⁵²

TXNIP boosts production of extracellular matrix proteins, which is a hallmark of DN and has a pivotal role in the development of CKD in DN. In renal mesangial cells, TXNIP overexpression increases expression of fibronectin.⁵³ On the contrary, TXNIP silencing inhibits HG-promoted synthesis of transforming growth factor-β and fibronectin.^54,55 In HK-2 cells, knockdown of TXNIP ameliorates HG-induced expression of α-smooth muscle actin.⁵⁶ Studies have also shown that autophagy is a mechanism mediating the effect of TXNIP on DN and CKD development. In the rat model of DN, inhibition of TXNIP with TXNIP DNAzyme suppresses diabetes-induced mTOR activation, thus promoting autophagy and attenuating extracellular matrix deposition in the kidneys.⁵⁷ Consistent with these results, TXNIP knockout mitigates activation of mTORC1 and restores nuclear translocation of transcription factor EB, thereby stimulating autophagy and alleviating fibrosis in diabetic kidneys.⁵⁸ In vitro, when HK-2 cells are exposed to HG, silencing of TXNIP by siRNA accelerates autophagic clearance, as shown by reduced autophagic vacuoles and less accumulation of LC3-II and p62,⁵⁹ which concurs with the in vivo finding.

TXNIP in AKI

AKI is defined by a rapid increase in serum creatinine, decrease in urine output, or both. AKI is estimated to account for 2 million deaths worldwide annually and is a growing global health concern.⁶⁰ It has become clear that an episode of AKI is associated with an increased risk of developing CKD, as well as both short- and long-term mortality and other adverse outcomes. As a multifactorial clinical condition, AKI is caused by a variety of etiologies, including ischemia, sepsis, and nephrotoxic drugs.⁶¹ An imbalance between ROS generation and elimination leads to inflammation, cell death, and kidney damage in AKI and CKD.⁶² Meanwhile, unfolded and misfolded proteins are accumulated in the ER lumen of renal tubular epithelial cells in AKI, resulting in ER stress.

During ischemia/reperfusion–mediated AKI, NLRP3 inflammasome deletion mitigates ischemic kidney injury.⁶³ To understand the mechanism, HK-2 cells were subjected to in vitro simulated ischemia/reperfusion injury. It has been shown that mROS is required for NLRP3 inflammasome activation, and MitoTEMPO, a mitochondria-specific antioxidant, prevents NLRP3 inflammasome activation when it is given prior to simulated ischemia/reperfusion.⁶³ In addition, TXNIP-NLRP3 co-localization in the mitochondria mediates mROS-dependent NLRP3 inflammasome activation.⁶³ The increased kidney expression of TXNIP is also involved in the susceptibility to ischemic AKI in diabetes.⁶⁴ Preventative treatment with resveratrol, which inhibits TXNIP binding to NLRP3, significantly attenuates the kidney damage in streptozotocin-induced diabetic rats.⁶⁴

TXNIP in Kidney Aging

The kidney undergoes many structural and functional changes with aging, including glomerulosclerosis, tubular atrophy, interstitial fibrosis, and nephron loss. Functionally, glomerular filtration rate, the capacity to conserve and secrete sodium, as well as urine concentrating and diluting abilities, are all decreased with age. In addition, kidney aging is a major risk factor for disease.⁶⁵ Cellular senescence, a multi-faceted program, is involved in promoting and accelerating kidney aging. The most commonly used marker of senescence is senescence-associated β-galactosidase activity at pH 6.0, which reflects the increased lysosomal content of senescent cells. Key characteristics of senescence are the absence of proliferation markers such as Ki67 and increased levels of cyclin-dependent kinase inhibitors such as p21^CIP1 or p16^INK4a. The DNA damage response, detected with γ-histone H2AX, and senescence-associated heterochromatic foci are other important features of many senescence mechanisms. ⁶⁵

TXNIP expression increases with kidney aging. TXNIP deletion reduced age-related kidney function decline and renal pathologic changes such as age-associated renal fibrosis in aged mice (11 and 20 months).⁶⁶ Meanwhile, TXNIP knockout kidneys exhibited down-regulation of aging markers, such as p16^INK4a and γ-H2AX, as well as decreased activity of senescence-associated β-galactosidase in aged mice.⁶⁶ In contrast, TXNIP overexpression in renal tubular cells in vitro increases senescence markers and stimulates a profibrotic response. Mechanistically, TXNIP directly binds to and activates STAT3 signaling, and a STAT3 inhibitor can mitigate TXNIP-mediated profibrotic response.⁶⁶ Thus, these results show that TXNIP up-regulation contributes to age-related renal fibrosis.

Novel Therapeutic Agents Targeting TXNIP in Kidney Disease

Considering its significance in the disease pathogenesis and progression, TXNIP has emerged as a potential therapeutic target in various diseases. In diabetes, a novel substituted quinazoline sulfonamide, SRI-37330, has been discovered to treat murine models of streptozotocin- and obesity-induced diabetes. SRI-37330 inhibits TXNIP expression in mouse and human islets through inhibition of polymerase II binding to the E-box motif region of the TXNIP promoter. In addition, SRI-37330 treatment inhibits glucagon secretion and function, blocks hepatic glucose output, and reverses hepatic steatosis. These findings suggest that targeting TXNIP with SRI-37330 may be an attractive treatment strategy for diabetes.⁶⁷

To treat DN, a variety of compounds and natural products have been developed to inhibit TXNIP expression (Table 1). For example, tranilast and calycosin, an isoflavone that constitutes the major constituent in Radix Astragali, exhibit therapeutic effects in streptozotocin-induced diabetic rats by suppressing TXNIP expression, thereby alleviating oxidative stress, inflammation, and renal fibrosis in DN.^68,69 Salidroside, the major gradient in Rhodiola rosea, decreases TXNIP levels and inhibits activation of TXNIP-NLRP3 inflammasome, hindering cell proliferation, oxidative stress, and extracellular matrix protein expression in rat mesangial cells in the presence of HG.⁷⁰ High dietary fructose is an important causative factor in the development of metabolic syndrome–associated glomerular podocyte oxidative stress and injury. Fructose feeding increased miRNA-377 as a biomarker of oxidative stress in renal cortex of fructose-fed rats, which promotes glomerular p38 mitogen-activated protein kinase phosphorylation and TXNIP expression, inducing podocyte NLRP3 inflammasome activation, injury, and albuminuria.⁷¹ The antioxidants pterostilbene and allopurinol have been found to ameliorate fructose-induced, miRNA-377/TXNIP-dependent activation of podocyte NLRP3 inflammasome and inflammation, thus attenuating albuminuria.⁷¹

Table 1.

Therapeutic Agents Targeting TXNIP in Renal Disease

Renal disease	Agent	Agent type	Mechanism of action	Reference
DN	Tranilast and calycosin	Natural product	Suppress kidney TXNIP expression, thus alleviating oxidative stress, inflammation, and fibrosis	68,69
	Salidroside	Natural product	Decrease TXNIP level and inhibit activation of TXNIP-NLRP3 inflammasome in mesangial cells	70
	Pterostilbene and allopurinol	Chemical compound	Ameliorate miR-377/TXNIP-dependent NLRP3 inflammasome activation in podocytes	71
AKI	Anisodamine	Chemical compound	Inhibit ER stress–associated TXNIP/NLRP3 inflammasome activation	72
	Dexmedetomidine	Chemical compound	Mitigate p38 MAPK/TXNIP signaling	73
	miR-30c-5p	miRNA	Repress TXNIP expression and alleviate activation of NLRP3/caspase-1	74
	miR-93-5p	miRNA	Down-regulate TXNIP expression and inhibit NLRP3-related pyroptosis	75

AKI, acute kidney injury; DN, diabetic nephropathy; ER, endoplasmic reticulum; MAPK, mitogen-activated protein kinase; NLRP3, Nod-like receptor protein 3; TXNIP, thioredoxin-binding protein.

Chemical compounds and miRNAs targeting TXNIP have also been identified to treat AKI (Table 1). Anisodamine, an antagonist of muscarinic acetylcholine receptor, inhibits ER stress–associated TXNIP/NLRP3 inflammasome activation and protects against rhabdomyolysis-induced AKI.⁷² Dexmedetomidine, a highly selective α₂-adrenergic receptor agonist, attenuates ischemic AKI in diabetic rats, partly through inhibition of p38 mitogen-activated protein kinase/TXNIP signaling.⁷³ Meanwhile, miRNAs have been used to alleviate septic AKI by targeting TXNIP. miR-30c-5p binds to 3′-untranslated regions of TXNIP and represses TXNIP expression. Kidney miR-30c-5p expression is decreased in the lipopolysaccharide-induced septic AKI model and associated with activation of NLRP3/caspase-1–mediated pyroptosis.⁷⁴ Up-regulation of miR-30c-5p in kidneys ameliorates expression of TXNIP, which inhibits NLRP3 and caspase-1 expression, as well as inflammatory cytokine secretion.⁷⁴ Studies have also shown that miRNAs from macrophage-derived exosomes, which are phospholipid bilayer vesicles derived from the endosome pathway, play important roles in AKI. miR-93-5p derived from anti-inflammatory M2 macrophages has been identified to directly bind to the 3′ untranslated regions of TXNIP, which negatively regulates TXNIP expression and inhibits NLRP3 inflammasome-related pyroptosis in renal epithelial cells in vitro.⁷⁵ Moreover, in the mouse model of septic AKI, M2-derived exosomes significantly mitigate the sepsis-induced kidney dysfunction, whereas exosomes from M2 knocked down of miR-93-5p almost lose their protective effect on kidney function.⁷⁵ Collectively, chemical compounds, natural products, and miRNAs that reduce TXNIP expression represent promising agents for the treatment of various kidney diseases.

Conclusion

This review focused on the regulation and function of TXNIP during oxidative and ER stresses, as well as stress-associated autophagy, which are implicated in a myriad of dysfunctional cellular processes. Aberrant expression of TXNIP affects function and survival of different cell types in a variety of kidney diseases, such as NS, DN and CKD, AKI, and kidney aging. Targeting TXNIP may change the therapeutic landscape for kidney disease.

Disclosure Statement

None declared.