Thioredoxin 1 Suppresses TXNIP-Driven Control of Glucose Metabolism in Human Cells

Abstract

Aims:

Cytosolic thioredoxin 1 (Trx1, TXN, TRX) is a central player in redox control. Thioredoxin interacting protein (TXNIP), an α-arrestin regulating glucose metabolism and inflammation, is widely regarded to inhibit TRX activity. However, the interactions between the two proteins across various cellular contexts remain poorly understood; in addition, only a limited number of studies have yet been conducted in human primary cells. We thus aimed here to investigate the functional relationship between TRX and TXNIP in human primary cells. We studied whether TXNIP inhibits TRX cellular activity in these primary cells and how this interaction influences cellular redox biology or glucose metabolism.

Results:

In primary cells, TXNIP deficiency did not increase cellular TRX activity. Instead, TXNIP deficiency elevated PGC-1α and PDK4 transcripts, increased PDHA1 Ser293 phosphorylation, and raised basal GLUT4, consistent with enhanced glucose uptake and restrained flux through the pyruvate dehydrogenase complex. Conversely, lowering TRX expression levels triggered higher TXNIP levels. This in turn correlated with suppressed transcripts for PGC-1α and PDK4, a lower extent of PDHA1 phosphorylation at Ser293, and decreased glucose uptake.

Innovation:

Our findings suggest that TXNIP, against common belief, may not necessarily be an endogenous inhibitor of TRX but, rather, that TRX can be an inhibitor of TXNIP.

Conclusion:

This study reveals that the key intracellular redox protein TRX inversely regulates TXNIP, suggesting that modulation of the TRX system may provide a previously unrecognized therapeutic avenue for modulation of glucose metabolism. Antioxid. Redox Signal. 44, 643–660.

Keywords

thioredoxin thioredoxin interacting protein peroxiredoxins pyruvate dehydrogenase glucose uptake glucose transporter

Introduction

The thioredoxin and glutathione systems provide the two major enzymatic antioxidant pathways defending mammalian cells from oxidative damage (Couto et al., 2016; Lu and Holmgren, 2014). A key cytosolic player in these systems is thioredoxin 1 (Trx1, TXN, or TRX¹), a small (∼12 kDa) highly conserved protein that plays a pivotal role in maintaining cellular redox homeostasis, functioning as a general disulfide reductase. It is kept reduced and thus active by the selenium-dependent thioredoxin reductase 1 (TXNRD1, or TrxR1) using NADPH as reducing power (Arner and Holmgren, 2000). TRX catalyzes the reduction of disulfides in a wide range of downstream substrate enzymes, including ribonucleotide reductase, methionine sulfoxide reductases, and peroxiredoxins (PRDXs), thereby supporting key cellular processes such as DNA synthesis, repair of oxidized methionine residues, and detoxification of hydrogen peroxide (H₂O₂) (Arner and Holmgren, 2000). Among the mammalian PRDXs, PRDX1 and PRDX2 are cytosolic and thus dependent on TRX, while PRDX3 is localized in mitochondria and thus reduced by mitochondrial TRX2 to maintain its catalytic activity (Rhee et al., 2012).

Innovation

The prevailing model holds that TXNIP is an endogenous inhibitor of TRX. While TXNIP is widely studied for controlling glucose metabolism, the functional implications of its interaction with TRX remain unclear. Here, using primary cells from TXNIP-deficient patients together with a doxycycline-inducible TRX knockdown system, we found that the cellular TRX activity was not changed in these TXNIP-deficient cells. Instead, our studies showed TRX suppresses TXNIP-driven glucose metabolism. Our work challenges the current model, provides evidence from human primary cells, connects redox control to glucose metabolism, and indicates the TRX–TXNIP axis as a potential therapeutic target for diabetes and insulin resistance.

Some proteins, such as apoptosis signal-regulating kinase 1 (ASK1), can bind reduced TRX without being proper disulfide substrates; upon TRX oxidation, ASK1 is released from TRX and thereby becomes activated, triggering stress- and cytokine-induced apoptosis (Saitoh et al., 1998). The noncovalent TRX binding to ASK1 occurs via protein–protein interaction surfaces near the redox-active disulfide/dithiol site of TRX, although without forming an intermolecular disulfide bond between the two proteins (Kosek et al., 2014). Thioredoxin interacting protein (TXNIP; previously named TBP-2 for thioredoxin binding protein-2) can also bind to TRX but has in this case been proposed to directly inhibit TRX activity and functions, thereby modulating the cellular redox balance (Nishiyama et al., 1999; Patwari et al., 2006). TXNIP is an α-arrestin protein well-studied for its roles in regulating glucose metabolism by inhibiting glucose influx and promoting glycolysis (Chutkow et al., 2008; Hui et al., 2008). The widely recognized role of TXNIP as an endogenous direct inhibitor of TRX activity is a concept Nishiyama et al early introduced in experiments using assays based upon TRX-catalyzed insulin disulfide reduction coupled to yeast-derived thioredoxin reductase (Nishiyama et al., 1999). However, the low-molecular-weight (∼35 kDa) type of thioredoxin reductases found in bacteria, yeasts, and plants are very different from the mammalian orthologues. Human TXNRD1 is a high-molecular-weight (∼55 kDa) enzyme that, being a selenoprotein, depends upon a selenocysteine residue for its activity and has another catalytic mechanism than the yeast enzyme (Williams et al., 2000). Importantly, the yeast enzyme cannot efficiently reduce human TRX (Oliveira et al., 2010), meaning that an assay for human TRX enzymatic activity when coupled to yeast thioredoxin reductase becomes highly unreliable. Those early findings suggesting that the TXNIP could directly inhibit human TRX activity, based upon results using a yeast thioredoxin reductase-coupled enzyme assay (Nishiyama et al., 1999; Oliveira et al., 2010), may thus not be conclusively showing a direct TRX-inhibitory function of TXNIP. In addition, genetic deletion of TXNIP in animal models does not overtly appear to increase TRX-dependent activity or availability, although TXNIP overexpression clearly suppresses glucose uptake and metabolic responsiveness of tissues to insulin (Chutkow et al., 2008; Sheth et al., 2005, 2006; Yoshioka et al., 2007). Importantly, overexpression of TXNIP can also suppress cellular TRX expression levels, thus lowering total cellular TRX activity, and thereby promote ASK1 activation (Junn et al., 2000). Such effects potentially involve the TXNIP–NRF2 (also named NFE2L2, for NFE2-like bZIP transcription factor 2) axis, as TXNIP has been shown to negatively regulate NRF2 signaling (He and Ma, 2012; Katsu-Jimenez et al., 2019; Maimaiti et al., 2025). Consequently, TXNIP overexpression lowering the cellular NRF2 activity can thus indirectly decrease cellular TRX levels, as the TXN gene is typically activated by NRF2 (Tonelli et al., 2018). Importantly, such effects do not necessarily involve the commonly postulated direct enzymatic inhibition of TRX by TXNIP. Thus, these earlier findings can collectively suggest that TXNIP may affect TRX indirectly and/or primarily under overexpression conditions, rather than under physiological states. This possibility prompted us to conduct the present study, aimed at further investigating the cellular consequences of deficiency in either TXNIP or TRX.

TXNIP has well-established functions in modulating glucose homeostasis, as well as promoting inflammation, and the protein has thereby been implicated in various pathological processes, including diabetes, cancer, and neurodevelopmental diseases (Choi and Park, 2023; Masutani, 2022). TXNIP acts as a negative feedback regulator of glucose metabolism. Its expression is strongly induced upon conditions of high glucose availability, a response mediated by the MondoA:MIx transcription factor complex. MondoA is activated by elevated glucose-6-phosphate levels and thereby translocates from the outer mitochondrial membrane to the nucleus, where it heterodimerizes with MIx and binds to carbohydrate response elements (ChoRE) within target gene promoter regions, including the TXNIP promoter, thus driving robust transcriptional activation (Stoltzman et al., 2008; Stoltzman et al., 2011). When upregulated, TXNIP inhibits glucose influx by downregulating glucose transporters (GLUTs) (Kim et al., 2024). Conversely, TXNIP deficiency leads to enhanced glucose uptake with a metabolic shift toward aerobic glycolysis, rather than complete oxidation via the tricarboxylic acid (TCA) cycle, accompanied by elevated lactate production, as observed in both mice and humans (Hui et al., 2008; Kahlhofer et al., 2024; Katsu-Jimenez et al., 2019). Notably, loss of TXNIP results in increased phosphorylation of the pyruvate dehydrogenase complex (PDHC), more specifically at serine-293 of its PDHA1 subunit, leading to PDHC inactivation and impaired pyruvate entry into the TCA cycle, as was shown physiologically in the soleus muscle of fasted mice (Andres et al., 2011). PDHC serves as a central metabolic node, connecting glycolysis to the TCA cycle by converting pyruvate—the final product of glycolysis—into acetyl-CoA, which enters the TCA cycle. PDHC activity is tightly regulated by reversible phosphorylation, with pyruvate dehydrogenase kinases (PDKs) inhibiting PDHC via phosphorylation, whereas pyruvate dehydrogenase phosphatases restore its activity through dephosphorylation (Park et al., 2018). The transcription of PDKs is positively regulated by peroxisome proliferator-activated receptor gamma coactivator-1α (PGC-1α) and peroxisome proliferator-activated receptors (Jeong et al., 2012). Interestingly, PGC-1α expression is typically upregulated in TXNIP-deficient mice, which may account for the enhanced PDK activities and consequent inactivation of PDHC seen in that context (Oka et al., 2009; Sheth et al., 2005).

Although TXNIP is widely regarded as an endogenous inhibitor of TRX, as outlined above, growing evidence from several studies suggest that many of the cellular functions of TXNIP can be independent from a potential inhibition of TRX. For example, a mutant version of TXNIP with a cysteine-to-serine substitution at position 247 (C247S), which no longer can bind to TRX, still inhibits glucose uptake as effectively as wild-type TXNIP, suggesting that TRX binding is not needed for the glucose uptake modulation of TXNIP (Patwari et al., 2009). Arrestin domain containing protein 4, another α-arrestin closely related to TXNIP, but which does not bind to TRX, can also inhibit glucose uptake just as TXNIP, further showing that TRX binding is not necessarily needed for this function (Patwari et al., 2009). However, some interaction with TRX may regulate TXNIP degradation during adipogenesis, as suggested by overexpression of TXNIP that suppresses adipogenesis, whereas overexpression of mutated TXNIP C247S, unable to bind TRX, exhibits a decreased inhibitory effect (Chutkow and Lee, 2011). In line with that concept, a recent study demonstrated that overexpression of TRX inhibits activation of the NOD-like receptor family pyrin domain containing 3 (NLRP3) inflammasome by downregulating TXNIP expression, thus disrupting the role of TXNIP in the inflammasome; overexpression of TRX here resulted in protective effects in Alzheimer’s disease models, possibly supporting a role of TRX as being a negative regulator of TXNIP (Jia et al., 2025).

In our previous work, we identified patients having TXNIP deficiency presenting with lactic acidosis and low serum methionine; we also found that primary myoblasts isolated from these patients exhibited increased NRF2 activity compared with control myoblasts (Katsu-Jimenez et al., 2019). Given that NRF2 is typically induced by oxidative or electrophilic stress, its activation in TXNIP-deficient cells was unexpected, as the loss of an endogenous TRX inhibitor would rather have been anticipated to enhance TRX activity, leading to increased antioxidant capacity and less oxidative stress, and thus instead suppressing NRF2 activation. Recently, we showed that the unexpected NRF2 activation in TXNIP-deficient cells correlates with an accumulation of the electrophilic glucose metabolite methylglyoxal (MGO), the production of which is driven by increased glucose influx triggered by the loss of TXNIP (Maimaiti et al., 2025). These prior observations prompted us to further investigate the cellular interplay between TXNIP and TRX, using the same primary patient-derived TXNIP-deficient cells, as well as a HeLa cell model allowing doxycycline-inducible knockdown of TRX. The results presented here suggest that TXNIP deletion in primary human cells does not overtly stimulate the cellular TRX-dependent functions. TRX downregulation, on the contrary, stimulates the TXNIP-driven effects on glucose metabolism.

Results

Cellular specific activity of TRX is not increased in patient-derived TXNIP-deficient fibroblasts or myoblasts compared with control cells

Given the presumed function of TXNIP as an endogenous direct inhibitor of TRX, we hypothesized that the absence of TXNIP would result in an elevated cellular enzymatic specific activity of TRX, defined as an increased ratio of total cellular TRX-catalyzed enzymatic activity over cellular TRX protein abundance. Surprisingly, when we performed the TXNRD1-driven insulin disulfide reduction assays, in both fibroblasts (Fig. 1A–F) and myoblasts (Fig. 1G–L), the total TRX protein levels were markedly increased (Fig. 1A, F, G, and L), yet no significant difference in absolute total cellular specific enzyme activity of TRX was found (Fig. 1B and H). Consequently, the specific activity of TRX in these cells, quantified as the ratio of active TRX over total TRX protein, was significantly downregulated in the TXNIP-deficient cells compared with controls (Fig. 1D and J). Previous findings from our group demonstrated that TRP32 (thioredoxin-like protein 32; also named TXNL1 for thioredoxin-like protein 1) can reduce insulin disulfides similarly to TRX in a TXNRD1-dependent manner (Andor et al., 2023). Therefore, we assessed TRP32 levels to determine whether the observed decrease in total activity was related to a downregulation of TRP32. However, TRP32 was upregulated three- to fourfold in TXNIP-deficient cells (Fig. 1E, F, K, and L). It is hence not clear why the specific activity of TRX in these TXNIP-deficient cell lysates was suppressed. However, the results clearly indicate that a loss of TXNIP does not increase the cellular TRX-specific activity in primary human cells, hence casting doubts on the notion that TXNIP would act as an endogenous direct TRX inhibitor in these cells.

FIG. 1.

Thioredoxin interacting protein (TXNIP) deficiency does not alter the insulin-reducing activity of TRX. (A) Quantification of TRX protein expression levels in control and TXNIP-null fibroblasts. Values are mean ± SEM, n = 4. (B) TRX activity was analyzed using the endpoint insulin reduction assay with the same fibroblast lysates tested for Western blot. The levels of active TRX were inferred from a standard curve with serial dilutions of recombinant human TRX protein run in parallel. Values are mean ± SEM, n = 4. Ns = not significant. (C) Quantification of TXNIP expression in control and TXNIP-null fibroblasts. Values are mean ± SEM, n = 4. ND indicates nondetectable. (D) The ratio of TRX activity over TRX protein levels. Values are mean ± SEM, n = 4. (E) Quantification of TRP32 expression in control and TXNIP-null fibroblasts. Values are mean ± SEM, n = 4. (F) Representative Western blot showing TRX, TXNIP, TRP32, and β-actin in control and TXNIP-null fibroblasts. (G) Quantifications of TRX protein expression levels in control and TXNIP-null myoblasts. Values are mean ± SEM, n = 4. (H) TRX activity in myoblast lysates was measured using the endpoint insulin reduction assay with the same samples used for Western blotting. Values are mean ± SEM, n = 4. (I) Quantifications of TXNIP expression levels in control cells and TXNIP-null myoblasts. Values are mean ± SEM, n = 4. ND indicates nondetectable. (J) The ratio of TRX activity over TRX protein levels in myoblasts. Values are mean ± SEM, n = 4. (K) Quantification of TRP32 expression levels in control cells and TXNIP-null myoblasts. Values are mean ± SEM, n = 3. (L) Representative Western blot showing TRX, TXNIP, TRP32, and β-actin in control cells and TXNIP-null myoblasts. (M) Quantification of TRX protein expression levels in Ctrl shRNA and TRX shRNA HeLa cells, with or without 1 μg/mL doxycycline for 7 days. Values are mean ± SEM, n = 4. (N) TRX activity in HeLa cells was analyzed using the endpoint insulin reduction assay, utilizing the same lysates tested for Western blot. Values are mean ± SEM, n = 3. (O) Quantifications of TXNIP expression levels in the indicated HeLa cells. Values are mean ± SEM, n = 5. (P) The ratio of TRX activity to TRX protein levels in the indicated HeLa cells. Values are mean ± SEM, n = 3. (Q) Quantification of TRP32 expression in the indicated HeLa cells. N = 3. (R) Representative Western blot showing TRX, TXNIP, TRP 32, and β-actin in the indicated HeLa cells with or without 1 μg/mL doxycycline for 7 days. Where statistical differences between groups were evaluated these were determined using two-tailed unpaired Student’s t-test, as indicated in the figure panels (ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001).

Validation of HeLa cell knockdown system

To further investigate the interplay between TXNIP and TRX, we analyzed the effects on TRX, TXNIP, and TRP32 levels as triggered by a knockdown of TRX levels, utilizing doxycycline-inducible TRX-knockdown HeLa cells described previously (Schwertassek et al., 2014) (Fig. 1M–Q and R). The TRX shRNA cells displayed lower TRX protein levels, which were completely downregulated upon treatment with 1 μg/mL doxycycline (Fig. 1M and R), with the cellular TRX activity consequently being significantly lower, although not absent due to the presence of TRP32 (Fig. 1N, Q, and R). Interestingly, the TXNIP expression was upregulated in these cells, both at baseline conditions and after doxycycline treatment (Fig. 1O and R). Due to the very low TRX levels in the knockdown cells with maintained TRP32 expression, the calculated cellular-specific activity of TRX appeared misleadingly elevated—almost 60-fold higher than baseline (Fig. 1P).

Together, these results show no evidence of TXNIP acting as an endogenous TRX inhibitor in fibroblasts and myoblasts, while the data from the TRX-knockdown HeLa cells remain inconclusive. Therefore, we next evaluated the oxidation state of PRDXs as another independent assessment of cellular enzymatic TRX function in the presence or absence of TXNIP.

The oxidation of PRDX1, PRDX2, and PRDX3 remained unchanged in TXNIP-deficient primary cells upon treatment with H₂O₂, but increased in HeLa cells upon knockdown of TRX

PRDXs are the main cellular enzymes for reduction of H₂O₂, thus protecting cells from oxidative damage, with their functions being dependent upon the enzymatic activity of TRX (Chae et al., 1999). It should be noted that the mammalian two-cysteine containing PRDXs are sensitive to overoxidation by H₂O₂ (Rhee et al., 2012). In the normal catalytic cycle, the peroxidatic cysteine of PRDX forms sulfenic acid (-SOH) with H₂O₂ and subsequently a disulfide bond with a resolving cysteine, which finally becomes reduced by TRX for recycling of the system (Fig. 2A, left part). Under conditions of excessive H₂O₂, however, the peroxidatic cysteine can be further oxidized (Peskin et al., 2013), forming nonfunctional hyperoxidized monomers or dimers (Fig. 2A, right part). Considering these forms of PRDXs, we found in our experimental setting that the oxidation states of PRDX1, PRDX2, and PRDX3 remained equal in TXNIP-deficient fibroblasts compared with control cells, both at baseline conditions and when challenged with 100 μM or 800 μM of H₂O₂ (Fig. 2B–H). Moreover, the abundance of the hyperoxidized monomeric PRDX-SO₃ forms increased progressively with the increasing H₂O₂ exposure (0, 100, and 800 μM), again with no differences between control TXNIP-deficient fibroblasts (Fig. 2E and I). We also noted that the extent of PRDX oxidation states varied across experiments (Supplementary Figs. S9, S10, S11, S12, S13, and S14), perhaps due to differences in cellular conditions such as varying cell densities at the time of treatment and/or speed of alkylation upon harvest.

FIG. 2.

Thioredoxin interacting protein (TXNIP) deficiency does not alter the oxidation states of PRDX1, PRDX2, and PRDX3 in fibroblasts. (A) Schematic representation of peroxiredoxin (PRDX) oxidation and reduction. Step 1: Reduced monomeric form of PRDX is oxidized upon exposure to H₂O₂, leading to the oxidation of peroxidatic cysteine (S_pH). Step 2: The oxidized monomer can undergo further oxidation, leading to the formation of hyperoxidized monomers. Step 3: Oxidized monomers can also form disulfide-linked dimers or hyperoxidized dimers. These oxidized dimers are typically reduced back to reduced monomers by the TRX system. Thus, if TXNIP deficiency enhanced TRX activity, less PRDX oxidation would be expected. In TRX-knockdown systems, the PRDX reduction is impaired, leading to the accumulation of oxidized dimers. S_pH: peroxidatic cysteine; S_rH: resolving cysteine. (B–D) Control and TXNIP-null fibroblasts were left untreated (Ctr) or treated with 100 or 800 μM H₂O₂, followed by Western blotting for PRDX1 (B), PRDX2 (C), and PRDX3 (D) under nonreducing conditions. (E) The same samples of control and TXNIP-null fibroblasts were analyzed by Western blotting for hyperoxidized PRDX-SO₃ under reducing SDS-PAGE conditions. The arrow indicates the hyperoxidized monomeric form of the PRDXs. (F–I) Quantification of dimer-to-total (monomer + dimer) ratios of PRDX1, PRDX2, and PRDX3 from (B–D), as well as the hyperoxidized form PRDX-SO₃, was analyzed using two-tailed unpaired Student’s t-test. Data are presented as mean ± SEM, n = 3–5. (J and K) Ctrl shRNA and TRX shRNA HeLa cells were cultured with or without 1 μg/mL doxycycline for 7 days, then left untreated (0) or treated with 100 or 800 μM H₂O₂. Lysates were analyzed under nonreducing SDS-PAGE conditions for PRDX1 (J), PRDX2 (K), and PRDX3 (L) using Western blot. (M–O) Quantification of dimer-to-total ratios from (J–L) for PRDX1 (M), PRDX2 (N), and PRDX3 (O). Statistical analysis was performed using two-tailed unpaired Student’s t-test. Values are presented as mean ± SEM, n = 4 (ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001).

Similarly to the case with fibroblasts, no significant differences were observed in the oxidation states of PRDX1 and PRDX3 comparing TXNIP-deficient myoblasts with controls, either at baseline or upon H₂O₂ treatment, while in this case, some of the dimer formations of PRDX2 were slightly lower upon H₂O₂ treatment (Supplementary Fig. S1). As comparison, using the TRX-knockdown HeLa cells, upon 7 days of doxycycline treatment that significantly lowered the TRX levels (Fig. 1M and R), this led to the expected increase in PRDX dimer formation. The differences to control cells were in this case statistically significant for all three PRDXs (PRDX1, PRDX2, and PRDX3) upon exposure to 800 μM of H₂O₂ (Fig. 2J–O), showing that TRX impairment can indeed hinder regeneration of PRDX monomers from dimers during H₂O₂ clearance. Comparing these results using the HeLa cell model, validating this assay as a probe for cellular TRX functionality, with the results using the TXNIP-deficient primary cells, we again find no compelling evidence that TXNIP would act as an endogenous TRX inhibitor. We therefore focused next on the established role of TXNIP in the regulation of glucose metabolism, in relation to the cellular TRX and TXNIP expression levels.

Inverse relationship between TXNIP and TRX levels in relation to PDHC phosphorylation

Compared with control myoblasts, patient-derived TXNIP-deficient myoblasts exhibited significantly increased phosphorylation of the PDHA1 subunit of PDHC at its serine-293 (Ser293) residue (Fig. 3A and B). Since malate treatment previously was shown to restore the TCA cycle function in these cells (Katsu-Jimenez et al., 2019), we tested the effects of malate on PDHA1 phosphorylation but observed no changes (Fig. 3A and B). In contrast to myoblasts, TXNIP-deficient fibroblasts exhibited no alteration in PDHA1 Ser293 phosphorylation (Supplementary Fig. S2), thus revealing cell type-specific effects on PDHA1 phosphorylation upon TXNIP loss in human cells.

FIG. 3.

Thioredoxin interacting protein (TXNIP) deficiency enhances PDHA1 Ser293 phosphorylation in primary myoblasts, while TRX knockdown in HeLa cells suppresses this phosphorylation and increases TXNIP levels. (A) Representative Western blot showing levels of phosphorylated serine 293 on PDHA1 and total PDHA1 in control and TXNIP-null myoblasts treated with different concentrations of malate (0, 0.5, and 5 mM) for 3 days. (B) Results shown in (A) were quantified using densitometry and analyzed using two-tailed unpaired Student’s t-test. Values are presented as mean ± SEM, n = 3. (C) The mRNA expression levels of PGC-1α and PDK4 in TXNIP-null and control myoblasts, with or without 5 mM of pretreatment with malate for 3 days. Data are presented as mean ± SEM from 3 to 4 independent experiments. Statistical analysis is performed using two-tailed unpaired Student’s t-test (*p < 0.05; **p < 0.01). (D) TRX and TXNIP protein levels, normalized to β-actin, were quantified by densitometry in Ctr shRNA and TRX shRNA HeLa cells after 1 μg/mL doxycycline treatment for 3, 5, and 7 days and plotted to assess changes. (E) Protein expression levels of TRX, phosphorylation of serine 293 on PDHCE1-α, and total PDHA1 were analyzed in HeLa cells treated with 1 μg/mL doxycycline for 3, 5, and 7 days. (F) Quantification and analysis of phospho and total PDHC expression in TRX-knockdown cells with and without doxycycline (1 μg/mL) treatment for 7 days. Data are presented as mean ± SEM, n = 3. Statistical significance was observed (p = 0.0006 and p = 0.00004). (G) RT-PCR analysis of PGC-1α and PDK4 expression in HeLa Ctrl shRNA and TRX shRNA treated with or without doxycycline for 3 days. Data are presented as mean ± SEM from 3 to 4 independent experiments. Statistical analysis was performed using a two-tailed unpaired Student’s t-test (ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001).

A previous report showed that PGC-1α transcriptionally regulates PDK4, being a key modulator of PDHC phosphorylation (Ma et al., 2005). In line with this, we also observed a significant upregulation of PGC-1α mRNA expression in the TXNIP-null myoblasts relative to controls, along with a modest increase in PDK4 transcripts (Fig. 3C). Interestingly, treatment with 5 mM of malate significantly decreased both the PGC-1α and PDK4 mRNA levels in the TXNIP-null myoblasts and also decreased the PGC-1α mRNA expression in the control myoblasts (Fig. 3C). The elevated phosphorylation of PDHA1 correlated with increased PGC-1α and PDK4 mRNA expression (Fig. 3C). These results suggest that TXNIP deficiency promotes PGC-1α–mediated upregulation of PDK4, contributing to enhanced PDHA1 phosphorylation, which in turn can be modulated by malate treatment.

Using the TRX-knockdown HeLa cells, we observed a significantly elevated TXNIP expression when TRX was downregulated (Fig. 3D). Although PDHA1 Ser293 phosphorylation varied across time points at baseline conditions and 1 μg/mL doxycycline treatment (Fig. 3E), phosphorylation normalized to the total PDHA1 level was significantly decreased both at baseline and upon 7 days after doxycycline addition (Fig. 3F). We also investigated whether the degree of phosphorylation of PDHA1 correlated with the expression of PGC-1α and PDK4, whereby we analyzed their transcript levels in the HeLa cells at both baseline and after 3–7 days of treatment with 1 μg/mL doxycycline. The transcript levels of PGC-1α and PDK4 were significantly downregulated upon 3 days of doxycycline-induced TRX knockdown compared with controls (Ctrl shRNA) (Fig. 3G and Supplementary Fig. S3). These results show that TRX-knockdown affects PGC-1α and PDK4 transcription. Although mRNA levels returned to basal condition after 7 days of doxycycline treatment, PDHA1 phosphorylation at Ser293 remained significantly lower (Fig. 3E–G).

Inverse relationship between TXNIP and TRX with cellular glucose uptake

It is well established that TXNIP can suppress glucose uptake into cells by targeting GLUTs to endocytosis, especially class 1 GLUTs such as GLUT1, GLUT3, and GLUT4 (Kim et al., 2024; Patwari et al., 2009; Yoshihara et al., 2010; Yoshioka et al., 2007). Using immunocytochemistry analyses with the primary patient-derived cells, we consequently found a significant increase in GLUT4 levels in the TXNIP-deficient fibroblasts compared with controls (Fig. 4A, first row, red, and Fig. 4B, black and white bars). Upon insulin stimulation, GLUT4 expression increased in the control cells but remained unchanged in TXNIP-deficient fibroblasts, where it remained significantly higher than in the insulin-stimulated controls (Fig. 4A, second row, red; Fig. 4B, yellow and orange bars). These results demonstrate that TXNIP deficiency leads to a constitutive enhancement of GLUT4 levels also in primary human fibroblasts.

FIG. 4.

Glucose uptake is downregulated in TRX-knockdown HeLa cells. (A) Representative immunofluorescence microscopy images showing GLUT4 (red) localization in human thioredoxin interacting protein (TXNIP)-null and control fibroblasts. Nuclei were counterstained with DAPI (blue). Images were acquired after a 4 h starvation (top row: untreated) and following insulin stimulation (bottom row: insulin). (B) Quantification of GLUT4 immunofluorescence intensity shown in (A) Statistical analysis was performed using two-tailed unpaired Student’s t-test. Values are presented as mean ± SEM (n = 3, from three independent staining experiments). (C) Glucose uptake assay in HeLa cells expressing control shRNA and TRX shRNA, treated with or without doxycycline (1 μg/mL) for 7 days. Glucose uptake was assessed using the fluorescent glucose analog, 2-NBDG (ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001).

We also quantified the uptake of a fluorescent derivative of d-glucose (2-NBDG) in the TRX-knockdown HeLa cell model. This revealed that the TRX shRNA cells had a suppressed glucose uptake compared with the control cells under both baseline conditions and upon doxycycline treatment (Fig. 4C). This should be compared with the results showing that both conditions had higher relative ratios of the TXNIP over TRX expression levels (Fig. 1M and O). Collectively, these results thus suggest that modulation of TRX levels can be linked to regulation of TXNIP-dependent control of glucose metabolism.

Discussion

In this study, we investigated the interplay between TRX and TXNIP to examine how modulation of the expression levels of these two proteins may influence key pathways in either cellular redox biology or glucose metabolism, using the physiologically relevant model of patient-derived TXNIP-deficient primary cells, complemented with TRX knockdown in HeLa cells. Our results do not support the view that TXNIP is an endogenous inhibitor of TRX activity; instead, we suggest that TRX can modulate TXNIP-mediated regulation of glucose metabolism.

First, we evaluated whether the absence of TXNIP in the human primary cells would enhance two canonical functions of TRX, its insulin disulfide reduction capacity in cell lysates, and its role in intracellular recycling of PRDXs. Contrary to the hypothesis that TXNIP inhibits TRX activity, we found no obvious increase in either of these two TRX functions in TXNIP-deficient primary cells compared with control cells. However, a forced TRX knockdown in HeLa cells clearly impaired the total insulin disulfide reduction capacity in lysates, as well as suppressed recycling of oxidized PRDXs, confirming the validity of the assays. These findings argue against TXNIP acting as an endogenous TRX inhibitor under baseline conditions, at least in these human cell types.

We also examined the previously described roles of TXNIP in regulating PDHC activity through phosphorylation of PDHA1 at its Ser293 residue (Yoshioka et al., 2012), as well as preventing glucose influx (Parikh et al., 2007), using either the primary TXNIP-deficient cells or shRNA-mediated TRX knockdown in HeLa cells. We observed elevated phosphorylation of PDHA1 at Ser293, along with elevated mRNA levels of PGC-1α and PDK4 in TXNIP-deficient myoblasts. This agrees with previous studies showing that TXNIP deficiency enhances PGC-1α expression (Oka et al., 2009; Sheth et al., 2005), which upregulates PDKs (Araki and Motojima, 2006; Ma et al., 2005), leading to PDHA1 phosphorylation and PDHC inhibition. Lower PDHC enzymatic activity impairs mitochondrial pyruvate oxidation and promotes a shift toward anaerobic glycolysis, which promotes lactate accumulation, thus providing a mechanistic explanation for the lactic acidosis found in the TXNIP-deficient patients (Katsu-Jimenez et al., 2019). These findings are also in accordance with previous studies reporting increased lactate levels following genetic TXNIP deletion in mice (Bodnar et al., 2002; Oka et al., 2006; Sheth et al., 2005; Yoshioka et al., 2012) and provide mechanistic insights into how TXNIP can regulate glycolytic flux through modulation of the PDK-PDHC axis. Interestingly, we found that phosphorylation of PDHA1 at Ser293 was not altered at baseline in the TXNIP-deficient fibroblasts compared with controls. It is in this context important to note that the PDHC subunits contain three regulatory phosphorylation sites, Ser232, Ser293, and Ser300, each of which can independently contribute to inhibition of PDHC activity (Park et al., 2018). As we here focused exclusively on Ser293 of PDHA1, the potential involvement of the other sites of phosphorylation being TXNIP-regulated in the fibroblasts remains to be investigated in future studies. It is also possible that the protein levels of PDHA1 are substantially lower in fibroblasts than in myoblasts, as we were also unable to detect the mRNA expression of PGC-1α and PDK4 in these cells, which may hinder detection of differences by Western blot. Importantly, TRX knockdown in HeLa cells suppressed PDHA1 Ser293 phosphorylation and upregulated TXNIP under both baseline conditions and after 7 days of doxycycline treatment. Interestingly, the mRNA levels of PGC-1α and PDK4 were, however, downregulated only after 3 days of doxycycline-induced TRX knockdown, but not at baseline, suggesting that TRX might regulate their transcriptions via TXNIP earlier than it affects PDHA1 phosphorylation.

Our finding that GLUT4 expression was upregulated in TXNIP-deficient fibroblasts, seemingly being mainly localized toward the cell membrane, agrees with prior findings suggesting that TXNIP negatively regulates GLUTs (Kim et al., 2024), thereby inhibiting glucose uptake. Consistent with this, silencing TXNIP has been shown to improve glucose uptake in human adipocytes, skeletal muscle, and mouse embryonic fibroblasts (Huy et al., 2018; Parikh et al., 2007). Importantly, we also found that the forced TRX knockdown in HeLa cells triggered downregulation of glucose uptake, which occurred in correlation with elevated TXNIP expression levels. That finding was consistent with previous reports showing that increased TXNIP expression levels suppress glucose uptake (Parikh et al., 2007; Patwari et al., 2009), and suggests that regulation of TRX expression can modulate this function of TXNIP. The lysine acetyltransferase GCN5 has been reported under high-glucose conditions to enhance TXNIP transcription and suppress glucose uptake through the TXNIP–GLUT4/2 axis (Liao et al., 2022). Consistent with that regulatory pathway, we also found that GCN5 expression was significantly upregulated after 3 days of doxycycline-induced TRX knockdown, but not at baseline (Supplementary Fig. S6). We speculate that lowering the TRX expression levels may disrupt redox homeostasis and indirectly lead to altered nutrient sensing; in response, cells can upregulate GCN5, thereby increasing TXNIP expression and promoting GLUT4/2 endocytosis, ultimately resulting in suppressed glucose uptake. Our findings herein are the first to suggest that this effect can be triggered by downregulating cellular TRX levels. That finding is potentially important, as it implies a novel mechanism by which the cellular thioredoxin system can have a role in controlling glucose metabolism, supporting the notion that the interplay between TRX and TXNIP can be critical for integration of metabolic flux and redox signaling pathways. These unexpected findings, to some extent, challenge existing paradigms, as they propose a novel regulatory relationship between TRX and TXNIP.

There is, in fact, some support in the literature for the concept of TRX suppressing TXNIP functions. For example, TRX overexpression was reported to suppress the NLRP3 inflammasome by downregulating TXNIP levels and disrupting its interaction with NLRP3, thereby exerting anti-inflammatory effects (Jia et al., 2025). In addition, during adipogenesis, TRX appears to regulate TXNIP stability, as the overexpression of TRX suppressed the inhibitory effects of TXNIP on differentiation, with a non-TRX-binding mutant of TXNIP having diminished regulatory capacity (Chutkow and Lee, 2011). These studies, along with our current data, suggest that TRX may limit TXNIP activity by modulating its expression, stability, or interaction, thereby attenuating TXNIP-driven metabolic and inflammatory signaling.

One advantage with our study is the use of patient-derived primary cells deficient in TXNIP because of a natural mutation (Katsu-Jimenez et al., 2019), thus being “normal” cells, although any cell culture model cannot recapitulate physiological context. Nonetheless, we argue that our results, complemented with the TRX-knockdown HeLa cell model, together suggest that TRX can suppress TXNIP-driven glucose metabolism, at least in these cells. This proposed function should be scrutinized in additional studies, in additional cell types, and in a physiological context. Many investigations thus far regarding the TRX and TXNIP interactions have either relied solely on in vitro systems or on overexpression of TXNIP (Hwang et al., 2014; Yoshihara et al., 2014), alternatively on TXNIP knockout models in mice (Abdelsaid et al., 2013; Chutkow et al., 2010). Those studies suggest that TRX inhibition is primarily observed under forced TXNIP overexpression, which may not accurately represent the interplay between TXNIP and TRX under physiological conditions. It should here be noted that the knockout of TRX in mice is embryonically lethal at the early blastocyst stage (Matsui et al., 1996), meaning that mouse studies of TXNP functions in the absence of TRX can be done only using conditional or organ-specific knockouts of TRX.

With our data suggesting a context-dependent dynamic role of TRX suppressing TXNIP-mediated glucose metabolism, this revised model offers a new framework linking redox and metabolic pathways via a TRX–TXNIP axis. We propose that TRX has the capacity to suppress TXNIP expression and cellular functions, thereby regulating PGC1-α and PDK4, and ultimately influencing mitochondrial function and glucose metabolism. This is schematically summarized in Figure 5 and further discussed as follows.

FIG. 5.

Summary of the main findings and conclusions of this study. Top panel: In wild-type cells, thioredoxin interacting protein (TXNIP) helps maintain metabolic balance by sensing glucose influx, suppressing PGC-1α and PDK4, and limiting excessive GLUT expression (yellow). Our findings suggest that TRX counteracts excessive TXNIP activity while, naturally, supporting its own reductive functions, including peroxiredoxin (PRDX) recycling. Middle panel (TXNIP^−/−): TXNIP-deficiency increases phosphorylation of the E1 subunit PDAH1 Ser293 via derepression of PGC-1α and PDK4, while loss of GLUT suppression elevates GLUT levels at the cell membrane and enhances glucose uptake. This impairs mitochondrial pyruvate entry, elevates methylglyoxal (MGO) production, and activates NRF2, driving TRX upregulation, as previously shown (Katsu-Jimenez et al., 2019; Maimaiti et al., 2025). Bottom panel (shRNA TRX): TRX knockdown prevents its inhibition of TXNIP, leading to elevated TXNIP activity, less glucose uptake, and suppression of PGC-1α and PDK4, thereby enhancing PDHC activity. Concurrently, TRX-dependent functions, such as PRDX recycling, become impaired.

The previously established model suggests that TXNIP binds to TRX via disulfide exchange, forming an intermolecular disulfide bond between Cys32 of TRX and Cys247 of TXNIP under reducing conditions, thereby inhibiting TRX activity. Under oxidative stress, this disulfide bond was proposed to be disrupted, leading to the release and reactivation of TRX (Hwang et al., 2014). However, such a model conflicts somewhat with some basic cellular redox principles. The cytosol is a strongly reducing environment that disfavors spontaneous or stable formation of intermolecular disulfide bonds (Hansen et al., 2009; Linke and Jakob, 2003). Stable protein disulfides typically form in oxidizing environments, such as the endoplasmic reticulum or extracellular space, where cysteine oxidation is favored (Linke and Jakob, 2003; Yi and Khosla, 2016). Moreover, oxidative stress typically promotes, not disrupts, disulfide bond formation (Sen, 1998). Also, if cellular TRX should be inhibited by TXNIP under reducing cytosolic conditions, the essential reductive functions of TRX, such as supporting ribonucleotide reductase or PRDXs, would be compromised. We believe that a more feasible model would be that the TRX–TXNIP interaction resembles the TRX–ASK1 system, where reduced TRX binds to and inhibits ASK1, while oxidized TRX releases ASK1 to permit its downstream signaling capacity (Matsuzawa and Ichijo, 2008; Saitoh et al., 1998). Previous studies could show that TXNIP overexpression only modestly suppressed (by <20%) the total insulin disulfide reduction activity of cellular TRX (Hwang et al., 2014), which should be considered in line with multiple studies—including our own work reported here using TXNIP-deficient primary cells—showing that a genetic loss of TXNIP does not enhance the cellular TRX activity (Chutkow et al., 2008; Sheth et al., 2005, 2006; Yoshioka et al., 2007). Early studies furthermore showed that vitamin D-induced TXNIP upregulation lowered the cellular TRX levels and activated ASK-1, with the ASK-1 activation thereby possibly having been due to lower cellular TRX levels rather than due to a direct inhibition of TRX by TXNIP (Junn et al., 2000). Similarly, decreased TXNIP expression has been linked to enhanced macrophage growth and increased protection against reactive oxygen species (Hu et al., 2020), which has been interpreted as indicative of increased TRX function. Perhaps those beneficial outcomes could more plausibly be attributed to NRF2 activation, rather than less direct TXNIP-mediated TRX inhibition. Such an alternative explanation would fit with the widely studied tissue protective functions of NRF2 (Anandhan et al., 2023; Copple et al., 2008; Higgins and Hayes, 2011; Mann et al., 2007; Surh et al., 2008; Suzuki et al., 2023) and our own recent observations that TXNIP deletion triggers NRF2 activation due to the formation of MGO (Katsu-Jimenez et al., 2019; Maimaiti et al., 2025). Therefore, cellular TRX levels can increase upon TXNIP deletion since TRX is a downstream target of NRF2 (Hawkes et al., 2014). Moreover, in vivo studies that suggest direct TRX inhibition by TXNIP typically do not consider the TRX protein levels in cells and do not determine the cellular TRX-specific activity, as we have done herein. Finally, deletion of TXNIP in mice confers cardioprotective effects, such as less cardiac hypertrophy, prevention of cardiac dysfunction, and improved recovery of cardiac function following ischemia–reperfusion injury, without any observable changes in TRX activity (Yoshioka et al., 2007, 2012).

TXNIP function is also potentially related to the well-known Warburg effect, that is, increased aerobic glycolysis in cancer cells; the understanding of its mechanisms having major implications for therapeutic cancer strategies (Pelicano et al., 2006; Stine et al., 2022). Elevated expression of TXNIP is associated with improved survival rates, more favorable clinical outcomes, and enhanced responsiveness to anticancer therapies across multiple cancer types (Meylan et al., 2021; Miligy et al., 2018; Woolston et al., 2012). One study has also shown that lactic acidosis drives cellular metabolism toward oxidative respiration and suppresses glycolysis through metabolic reprogramming (Chen et al., 2008). That metabolic shift involves upregulation of the transcription factor MondoA, which binds to the promoter region of the TXNIP gene, leading to increased TXNIP expression under conditions of lactic acidosis (Chen et al., 2010). Our own findings with TXNIP-deficient patient cells as well as prior TXNIP-knockout models in mice have consistently demonstrated that TXNIP deficiency promotes glycolysis and increases lactate production (Beg et al., 2021; Hui et al., 2008; Katsu-Jimenez et al., 2019). Cells may restore TXNIP expression via metabolic reprogramming and MondoA activation under lactic acidosis. Our findings suggest that TXNIP loss activates glycolysis through PGC-1α-mediated PDK4 induction, leading to PDHA1 phosphorylation, PDHC inhibition, and suppressed pyruvate flux into the TCA cycle.

Although the TRX protein levels here were found to be increased in the TXNIP-deficient cells, the total TRX activity remained unchanged, and the explanation of that discrepancy remains to be clarified. As this phenomenon was furthermore accompanied by a significant upregulation of TRP32, possibly as part of a broader metabolic adaptation, these effects clearly need to be further studied. Interestingly, TRP32 was not upregulated in the HeLa cells upon TRX knockdown, despite a diminished total cellular TRX activity. That may indicate that TRP32 induction may be more tightly regulated by metabolic stress or glucose handling rather than by redox imbalances or the total lack of TRX activity.

TRX is known to protect cells from oxidative stress and is frequently overexpressed in cancer cells, contributing to tumor progression, resistance to chemotherapy, and enhanced cell survival (Bhatia et al., 2016; Cha et al., 2009; Karlenius and Tonissen, 2010; Lin et al., 2017; Shang et al., 2019). Conversely, TXNIP is often downregulated in tumors, and its lower expression has been associated with cancer progression and treatment resistance (Meylan et al., 2021; Miligy et al., 2018; Woolston et al., 2012). These observations suggest an inverse expression relationship between TRX and TXNIP in cancer. TRX overexpression seems to suppress TXNIP expression, thereby increasing glucose utilization and aerobic glycolysis, ultimately contributing to the Warburg effect. Supporting that notion, overexpression of the calcium signaling protein S100P in colorectal cancer cell lines was shown to increase the cellular TRX levels while decreasing TXNIP expression (Lin et al., 2017), again suggesting that TRX overexpression may actively suppress TXNIP expression.

Our findings, together with the herein discussed previous reports, indicate that the TRX–TXNIP interaction is more complex than previously recognized and can be bidirectional. Specifically, our data provide novel insights into the interplay between TRX and TXNIP, suggesting a reciprocal relationship whereby TRX can play a role in regulating TXNIP-driven glucose metabolic signaling, potentially supporting cellular glucose uptake and a metabolic shift toward aerobic glycolysis. Further studies are required to dissect the molecular mechanisms underlying this interaction, including validation in additional cell or animal models under diverse physiological and pathophysiological conditions. Overall, these findings challenge the conventional view of TXNIP solely as an inhibitor of TRX activity, suggesting instead that the metabolic functions of TXNIP can be suppressed by TRX.

Materials and Methods

Ethical considerations

Studies with the cells obtained from both patients and healthy control subjects were approved by the Stockholm Regional Ethics Committee, and for the underaged patients, informed consents were obtained from both parents, as detailed earlier (Katsu-Jimenez et al., 2019). The electronic laboratory notebook system of Karolinska Institutet was used to document the experiments of this study.

Primary cell culture

Isolation and culture conditions of the primary myoblasts and fibroblasts were performed as described in the previous studies from our group (Katsu-Jimenez et al., 2019; Maimaiti et al., 2025). Briefly, primary myoblasts were maintained in Ham’s F-10 nutrient mix with GlutaMAX (Thermo Fisher Scientific, catalog number: 41550-021) supplemented with 2.5 ng/mL basic fibroblast growth factor (Thermo Fisher Scientific), 10% FBS, 1% penicillin/streptomycin, and 100 nM sodium selenite. The primary fibroblasts were cultured in minimal essential medium (MEM) with GlutaMAX (Thermo Fisher Scientific, catalog number: 41090-028), enriched with 10% FBS, 1% penicillin/streptomycin and 100 nM sodium selenite.

Immunoblot analyses

Cells were washed twice with ice-cold phosphate-buffered saline (PBS) to remove residual media and then lysed using a lysis buffer containing 20 mM Tris-HCI (pH 7.5), 10 mM EDTA, 150 mM NaCI, 0.5% Triton X-100, 0.5% sodium deoxycholate, and 30 mM sodium pyrophosphate. The lysis buffer was supplemented with a 1× complete EDTA-free protease inhibitor cocktail (Roche Applied Science) and phosphatase inhibitors, including 0.2 mM phenylmethylsulfonyl fluoride, 0.2 mM sodium orthovanadate, and 1 mM sodium fluoride (NaF). Cell lysates were incubated on ice for about 20 min, followed by centrifugation at 14,000 g for 10 min at 4°C to pellet cellular debris, and supernatants were collected. Protein concentration was subsequently determined using the Pierce bicinchoninic acid protein assay kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. For SDS-PAGE gel analyses, LDS sample buffer (4×, Thermo Fisher Scientific) and 10 mM DTT were first added to cell lysates in microcentrifuge tubes (20 µg protein per sample), then they were boiled at 95°C for 5 min, followed by a quick centrifugation to spin down condensed drops in the tube. After centrifugation, samples were carefully loaded into wells of Bolt™ 4%–12% Bis-Tris Plus gels (Thermo Fisher Scientific) and electrophorized using the program of 165 V for 40 min. Then, proteins were electrotransferred to an iBlot 2 NC transfer stack (Thermo Fisher Scientific), using iBlot Dry Blotting system (Thermo Fisher Scientific), at 25 V for 8 min. Ponceau S staining (Sigma) was performed to validate equal loading between lanes, before proceeding to block membranes with 5% nonfat skim milk prepared in Tris-buffered saline with 0.1% Tween-20. After blocking the membranes at room temperature (RT, ≈20°C) for 1 h, they were incubated overnight with the described primary antibodies at 4°C. The next day, membranes were incubated with the secondary antibodies at RT for 1 h. Signals were detected with the ECL detection reagent (Cytiva Amersham) using a Bio-Rad ChemiDoc XRS scanner, and detected bands were quantified using Bio-Rad Quantity One software. Results were normalized to β-actin for each sample. The antibodies used, with their dilutions, are listed in Supplementary Table S1.

TRX activity as assessed by TXNRD-driven insulin disulfide reduction capacity

The endpoint insulin disulfide reduction assay was used to measure the thioredoxin activity in cell lysates, being the most specific TRX activity assay for crude lysates, as it is coupled with the TXNRD1 selenoprotein, thus specifically defining total enzymatic TRX activity, as elsewhere described and discussed in detail (Arner and Holmgren, 2005; Arnér, 2018), here with slight modification for use with 96-well plates. In short, total cellular protein (5, 10, or 15 µg) was incubated in a reaction mixture containing 50 nM recombinant human TXRND1, 297 µM insulin, and 1.3 mM NADPH in 50 mM Tris-HCI buffer (pH 7.5) supplemented with 2 mM EDTA, in a final volume of 50 µL. To improve the accuracy of activity measurements, identical cell lysates were assayed for 15, 30, 45, and 60 min at 37°C. Each reaction was measured in duplicate. Reactions were terminated by the addition of 200 µL of 7.2 M guanidine-HCl (pH 8.0) containing 1 mM DTNB. Absorbance was monitored at 412 nm. Control reactions containing all assay components except TXNRD1 were included for each sample to correct background activity. The activity was determined as the average inactivity over time using the different time points, with serial dilutions of recombinant human TRX prepared separately to generate a standard curve. TRX activity in each sample was subsequently calculated based on the standard curve. Recombinant human TXNRD1 used in the assay was expressed and purified as described elsewhere (Cheng and Arner, 2017).

Doxycycline-inducible TRX knockdown HeLa cells

HeLa cells constructed for doxycycline-inducible knockdown of TRX were kindly provided by Prof. Tobias Dick, Heidelberg, Germany, and the method for the generation and use of these cells was previously reported (Schwertassek et al., 2014). Two constructs from that study were used here: TRX-281, which shows a minimal change in TRX expression after treatment with 1 μg/mL of doxycycline for 7 days, and TRX-252, in which the same treatment results in ∼95% reduction in TRX expression levels. These constructs are here referred to as Ctrl shRNA (control) and TRX shRNA, respectively. These cells were maintained in Dulbecco’s modified essential medium (DMEM) (Thermo Fisher Scientific, catalog number: 40965-039), supplemented with 10% FBS, 1% penicillin/streptomycin, and 100 nM sodium selenite.

Glucose uptake assay with 2-NBDG

The Ctrl shRNA and TRX shRNA cells were seeded in quadruplicates in 96-well plates and cultured as specified in the text for 3 or 7 days in the presence or absence of doxycycline (1 μg/mL). The method has been described in detail (Stein et al., 2025). Briefly, on the day of measurement, cells were starved for 4 h in DMEM without glucose (Gibco, 11966025) that also contained 1% penicillin/streptomycin, 1% sodium pyruvate, and 0.25% BSA, and continuously stimulated with or without 1 μg/mL of doxycycline. Subsequently, the starved cells were treated with a fluorescent glucose homologue, 2-NBDG (catalog number: N13195, Invitrogen), at a final concentration of 60 μM, and incubated for 30 min. DMSO treatment served as control as 2-NBDG was dissolved in DMSO. After incubation, cells were washed twice with warm PBS, 100 µL of PBS was added to each well before measuring the fluorescence intensity using a plate reader (Tecan) with excitation/emission wavelengths of 485/535 nm. Starvation and incubation procedures were carried out in an incubator maintained at 37°C with 5% CO₂. At the end, normalization of the signals for the number of live cells was performed using a crystal violet assay, as previously described (Feoktistova et al., 2016). The fluorescence readings from the control (without 2-NBDG) were subtracted from the experimental samples (with 2-NBDG).

Assessing PRDX (PRDX1, PRDX2, and PRDX3) oxidation using H₂O₂ treatment

Cells were cultured as described above and seeded in 35 mm cell culture dishes until they reached ∼90% confluency. On the day of the experiment, cells were either left untreated (control) or were treated with freshly prepared hydrogen peroxide (H₂O₂) at final concentrations of 100 μM for 2 min or 800 μM for 5 min at RT. After exposure to H₂O₂, the medium was removed by vacuum aspiration, immediately followed by blocking unreacted thiols with 20 mM N-ethylmaleimide (NEM) for 15 min at RT using a wash buffer containing 50 mM Tris-HCI (pH 7.5), 150 mM NaCI, and 10 μg/mL catalase. After blocking thiols, cells were washed once with the same buffer to remove residual NEM. Cells were then detached by trypsinization and collected at 3000 g for 1 min at RT, and the supernatant was discarded. The resulting pellet was washed twice with a wash buffer. Finally, pellets were lysed in the wash buffer supplemented with 1% IGEPAL and 1× complete EDTA-free protease inhibitor cocktail (Roche Applied Science). Lysates were incubated on ice for about 20 min and then followed by centrifugation at 14,000 g for 10 min at 4°C to pellet cellular debris. The resulting supernatants were collected and analyzed by immunoblotting, under nonreducing or reducing conditions, as indicated, to evaluate the oxidation states of PRDX1, PRDX2, and PRDX3.

Transcript determinations using RT-PCR

Total RNA was extracted from six-well plates with cells grown until they reached around 90% confluency using the TRIzol reagent (Invitrogen, catalog number: 15596018) according to the manufacturer’s instructions. RNA concentration and purity were evaluated using a Nanodrop spectrophotometer (Nanodrop 2000, Thermo Scientific). Complementary DNA (cDNA) was synthesized from 1 μg of total RNA using the Maxima First Strand cDNA Synthesis Kit (Thermo Scientific, catalog number: K1641) according to the manufacturer’s protocol, and reactions were carried out in a Biometra T3000 Thermocycler.

Quantitative real-time PCR (qRT-PCR) was performed using the Luminaris Color HiGreen qPCR Master Mix Kit (Thermo Scientific, catalog number: K0391), which includes Hot Start Taq DNA polymerase, uracil-DNA glycosylase (UDG), double strand-DNA binding SYBR Green I dye, and dNTPs. Reactions were run on the PikoReal 96 Real-Time PCR System (Thermo Scientific), with the following cycling conditions: 3 min at 50°C for UDG activation, 10 min at 95°C for initial denaturation and enzyme activation, followed by 40 cycles of 15 s at 95°C for denaturation, 30 s at 60°C for annealing, and 30 s at 60°C for extension. The primers used for each gene were synthesized by Integrated DNA Technology, and their sequences are provided in Supplementary Table S2. Primer efficiencies were calculated in both myoblasts and HeLa cells using the formula $E = {(10}^{(1 / - slope}) - 1) * 100 %$ , as previously described (Racz et al., 2021). A no-template control was included for each primer pair to assess potential primer–dimer formation, which was found to be minimal as no product was made without a template. Relative gene expression levels were determined using the 2^ (-ΔΔCt) method and normalized against ACTB and SNW1 as housekeeping genes. Primer efficiency validation in both myoblasts and HeLa cells is presented in Supplementary Figure S4A and B.

Immunocytochemistry

The cells were seeded onto sterile coverslips placed in a 12-well plate with culture medium, and the following day, cells were starved in glucose-free medium for 4 h, then either left untreated or stimulated with 100 nM insulin for 15 min at 37°C. After stimulation, the media were removed, and the cells were washed three times with PBS. Fixation was performed using 4% paraformaldehyde for 15 min at RT, followed by additional PBS washes. The cells were then permeabilized with 0.2% Triton X-100 for 10 min and subsequently blocked for 1 h using a blocking buffer (5% BSA in PBS). For GLUT4 staining, cells were incubated with a primary anti-Glut4 antibody (ABclonal, A7637) followed by an Alexa Fluor 555-conjugated donkey anti-rabbit secondary antibody (Thermo Fisher, A31572). Nuclei were counterstained with DAPI. Phalloidin-Alexa Fluor 488 (Thermo Fisher) was applied for 15 min to stain actin filaments, as shown in Supplementary Figure S5. Imaging and visualization were performed using a Leica DMi8 microscope.

Statistical analyses

Analyses of data were performed in GraphPad Prism version 10. Western blot densitometry was quantified using Fiji (ImageJ). Statistical comparisons were made using two-tailed unpaired Student’s t-tests and correlation analyses using Pearson r, as indicated. p-Values are reported as follows: ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001.

Footnotes

Author Disclosure Statement

The authors declare that they have no conflicts of interest with the contents of this article.

Funding Information

Funding is acknowledged from Karolinska Institutet, the Knut and Alice Wallenberg Foundations (KAW 2019.0059), the Swedish Cancer Society (21 1463 Pj), the Swedish Research Council (2021-02214), the Cayman Biomedical Research Institute (CABRI), the Hungarian Thematic Excellence Program (TKP2021-EGA-44), the Hungarian National Research, Development and Innovation Office (ED_18-1-2019-0025), and the Hungarian National Tumor Biology Laboratory (02022-2.1.1-NL-2022-00010).

Supplemental Material

Abbreviations

References

Abdelsaid

, Matragoon

, El-Remessy

. Thioredoxin-interacting protein expression is required for VEGF-mediated angiogenic signal in endothelial cells. Antioxid Redox Signal, 2013; 19(18):2199–2212; doi: 10.1089/ars.2012.4761

Anandhan

, Dodson

, Shakya

, et al. NRF2 controls iron homeostasis and ferroptosis through HERC2 and VAMP8. Sci Adv, 2023; 9(5):eade9585; doi: 10.1126/sciadv.ade9585

Andor

, Mohanraj

, Pató

et al. TXNL1 has dual functions as a redox active thioredoxin-like protein as well as an ATP- and redox-independent chaperone. Redox Biol, 2023; 67:102897; doi: 10.1016/j.redox.2023.102897

Andres

, Ratliff

, Sachithanantham

, et al. Diminished AMPK signaling response to fasting in thioredoxin-interacting protein knockout mice. FEBS Lett, 2011; 585(8):1223–1230; doi: 10.1016/j.febslet.2011.03.042

Araki

, Motojima

. Identification of ERRalpha as a specific partner of PGC-1alpha for the activation of PDK4 gene expression in muscle. FEBS J, 2006; 273(8):1669–1680; doi: 10.1111/j.1742-4658.2006.05183.x

Arner

, Holmgren

. Measurement of thioredoxin and thioredoxin reductase. CP Toxicology, 2005; 24(1):Unit 7 4; doi: 10.1002/0471140856.tx0704s05

Arner

ESJ

, Holmgren

. Physiological functions of thioredoxin and thioredoxin reductase. Eur J Biochem, 2000; 267(20):6102–6109; doi: 10.1046/j.1432-1327.2000.01701.x

Arnér

ESJ

. Selective evaluation of thioredoxin reductase enzymatic activities. Methods Mol Biol, 2018; 1661:301–309; doi: 10.1007/978-1-4939-7258-6_21

Beg

, Zhang

, McCourt

, et al. ATGL activity regulates GLUT1-mediated glucose uptake and lactate production via TXNIP stability in adipocytes. J Biol Chem, 2021; 296:100332; doi: 10.1016/j.jbc.2021.100332

10.

Bhatia

, McGrath

, Di Trapani

, et al. The thioredoxin system in breast cancer cell invasion and migration. Redox Biol, 2016; 8:68–78; doi: 10.1016/j.redox.2015.12.004

11.

Bodnar

, Chatterjee

, Castellani

, et al. Positional cloning of the combined hyperlipidemia gene Hyplip1. Nat Genet, 2002; 30(1):110–116; doi: 10.1038/ng811

12.

Cha

, Suh

, Kim

. Overexpression of peroxiredoxin I and thioredoxin1 in human breast carcinoma. J Exp Clin Cancer Res, 2009; 28(1):93; doi: 10.1186/1756-9966-28-93

13.

Chae

, Kim

, Kang

, et al. Characterization of three isoforms of mammalian peroxiredoxin that reduce peroxides in the presence of thioredoxin. Diabetes Res Clin Pract, 1999; 45(2–3):101–112; doi: 10.1016/s0168-8227(99)00037-6

14.

Chen

, Lucas

, Schroeder

, et al. The genomic analysis of lactic acidosis and acidosis response in human cancers. PLoS Genet, 2008; 4(12):e1000293; doi: 10.1371/journal.pgen.1000293

15.

Chen

, Merl

, Peterson

, et al. Lactic acidosis triggers starvation response with paradoxical induction of TXNIP through MondoA. PLoS Genet, 2010; 6(9):e1001093; doi: 10.1371/journal.pgen.1001093

16.

Cheng

, Arner

. Selenocysteine insertion at a predefined UAG codon in a release factor 1 (RF1) depleted Escherichia coli host strain bypasses species barriers in recombinant selenoprotein translation. J Biol Chem, 2017; 292(13):5476–5487; doi: 10.1074/jbc.M117.776310

17.

Choi

, Park

. TXNIP: A key protein in the cellular stress response pathway and a potential therapeutic target. Exp Mol Med, 2023; 55(7):1348–1356; doi: 10.1038/s12276-023-01019-8

18.

Chutkow

, Birkenfeld

, Brown

, et al. Deletion of the alpha-arrestin protein Txnip in mice promotes adiposity and adipogenesis while preserving insulin sensitivity. Diabetes, 2010; 59(6):1424–1434; doi: 10.2337/db09-1212

19.

Chutkow

and Lee

. Thioredoxin regulates adipogenesis through thioredoxin-interacting protein (Txnip) protein stability. J Biol Chem, 2011; 286(33):29139–29145; doi: 10.1074/jbc.M111.267666

20.

Chutkow

, Patwari

, Yoshioka

, et al. Thioredoxin-interacting protein (Txnip) is a critical regulator of hepatic glucose production. J Biol Chem, 2008; 283(4):2397–2406; doi: 10.1074/jbc.M708169200

21.

Copple

, Goldring

, Kitteringham

, et al. The Nrf2-Keap1 defence pathway: Role in protection against drug-induced toxicity. Toxicology, 2008; 246(1):24–33; doi: 10.1016/j.tox.2007.10.029

22.

Couto

, Wood

, Barber

. The role of glutathione reductase and related enzymes on cellular redox homoeostasis network. Free Radic Biol Med, 2016; 95:27–42; doi: 10.1016/j.freeradbiomed.2016.02.028

23.

Feoktistova

, Geserick

, Leverkus

. Crystal violet assay for determining viability of cultured cells. Cold Spring Harb Protoc, 2016; 2016(4); doi: 10.1101/pdb.prot087379

24.

Hansen

, Roth

, Winther

. Quantifying the global cellular thiol–disulfide status. Proc Natl Acad Sci U S A, 2009; 106(2):422–427.

25.

Hawkes

, Karlenius

, Tonissen

. Regulation of the human thioredoxin gene promoter and its key substrates: A study of functional and putative regulatory elements. Biochim Biophys Acta, 2014; 1840(1):303–314; doi: 10.1016/j.bbagen.2013.09.013

26.

, Ma

. Redox regulation by nuclear factor erythroid 2-related factor 2: Gatekeeping for the basal and diabetes-induced expression of thioredoxin-interacting protein. Mol Pharmacol, 2012; 82(5):887–897; doi: 10.1124/mol.112.081133

27.

Higgins

, Hayes

. The cap’n’collar transcription factor Nrf2 mediates both intrinsic resistance to environmental stressors and an adaptive response elicited by chemopreventive agents that determines susceptibility to electrophilic xenobiotics. Chem Biol Interact, 2011; 192(1–2):37–45; doi: 10.1016/j.cbi.2010.09.025

28.

, Tian

, Li

, et al. Brucella infection regulates thioredoxin-interacting protein expression to facilitate intracellular survival by reducing the production of nitric oxide and reactive oxygen species. J Immunol, 2020; 204(3):632–643; doi: 10.4049/jimmunol.1801550

29.

Hui

, Andres

, Miller

, et al. Txnip balances metabolic and growth signaling via PTEN disulfide reduction. Proc Natl Acad Sci U S A, 2008; 105(10):3921–3926; doi: 10.1073/pnas.0800293105

30.

Huy

, Song

, Kim

, et al. TXNIP regulates AKT-mediated cellular senescence by direct interaction under glucose-mediated metabolic stress. Aging Cell, 2018; 17(6):e12836; doi: 10.1111/acel.12836

31.

Hwang

, Suh

, Jeon

, et al. The structural basis for the negative regulation of thioredoxin by thioredoxin-interacting protein. Nat Commun, 2014; 5:2958; doi: 10.1038/ncomms3958

32.

Jeong

, Jeoung

, Park

, et al. Transcriptional regulation of pyruvate dehydrogenase kinase. Diabetes Metab J, 2012; 36(5):328–335; doi: 10.4093/dmj.2012.36.5.328

33.

Jia

, Sheng

, Zhang

, et al. Thioredoxin-1 inhibits NLRP3-mediated pyroptosis by regulating TXNIP in models of Alzheimer’s disease. Sci Rep, 2025; 15(1):16551; doi: 10.1038/s41598-025-01636-5

34.

Junn

, Han

, Im

, et al. Vitamin D3 up-regulated protein 1 mediates oxidative stress via suppressing the thioredoxin function. J Immunol, 2000; 164(12):6287–6295.

35.

Kahlhofer

, Marchet

, Seifert

, et al. TXNIP mediates LAT1/SLC7A5 endocytosis to reduce amino acid uptake in cells entering quiescence. bioRxiv, 2024; doi: 10.1101/2024.10.29.620655

36.

Karlenius

, Tonissen

. Thioredoxin and cancer: A role for thioredoxin in all states of tumor oxygenation. Cancers (Basel), 2010; 2(2):209–232; doi: 10.3390/cancers2020209

37.

Katsu-Jimenez

, Vazquez-Calvo

, Maffezzini

, et al. Absence of TXNIP in humans leads to lactic acidosis and low serum methionine linked to deficient respiration on pyruvate. Diabetes, 2019; 68(4):709–723; doi: 10.2337/db18-0557

38.

Kim

, Ge

, Kim

, et al. TXNIP-mediated crosstalk between oxidative stress and glucose metabolism. PLoS One, 2024; 19(2):e0292655; doi: 10.1371/journal.pone.0292655

39.

Kosek

, Kylarova

, Psenakova

, et al. Biophysical and structural characterization of the thioredoxin-binding domain of protein kinase ASK1 and its interaction with reduced thioredoxin. J Biol Chem, 2014; 289(35):24463–24474; doi: 10.1074/jbc.M114.583807

40.

Liao

, Xu

, Zhang

, et al. Persistent high glucose induced EPB41L4A-AS1 inhibits glucose uptake via GCN5 mediating crotonylation and acetylation of histones and non-histones. Clin Transl Med, 2022; 12(2):e699; doi: 10.1002/ctm2.699

41.

Lin

, Zhang

, Zuo

, et al. Thioredoxin-1 promotes colorectal cancer invasion and metastasis through crosstalk with S100P. Cancer Lett, 2017; 401:1–10; doi: 10.1016/j.canlet.2017.04.036

42.

Linke

, Jakob

. Not every disulfide lasts forever: Disulfide bond formation as a redox switch. Antioxid Redox Signal, 2003; 5(4):425–434; doi: 10.1089/152308603768295168

43.

, Holmgren

. The thioredoxin antioxidant system. Free Radic Biol Med, 2014; 66:75–87; doi: 10.1016/j.freeradbiomed.2013.07.036

44.

, Zhang

, Elam

, et al. Cloning of the rat pyruvate dehydrogenase kinase 4 gene promoter: Activation of pyruvate dehydrogenase kinase 4 by the peroxisome proliferator-activated receptor gamma coactivator. J Biol Chem, 2005; 280(33):29525–29532; doi: 10.1074/jbc.M502236200

45.

Maimaiti

, Dagnell

, Coppo

, et al. Patient-derived TXNIP-deficient primary cells exhibit NRF2 activation linked to upregulation of glyoxalase 1 (GLO1). Free Radic Biol Med, 2025; 239:230–241; doi: 10.1016/j.freeradbiomed.2025.07.035

46.

Mann

, Niehueser-Saran

, Watson

, et al. Nrf2/ARE regulated antioxidant gene expression in endothelial and smooth muscle cells in oxidative stress: Implications for atherosclerosis and preeclampsia. Sheng Li Xue Bao, 2007; 59(2):117–127.

47.

Masutani

. Thioredoxin-interacting protein in cancer and diabetes. Antioxid Redox Signal, 2022; 36(13–15):1001–1022; doi: 10.1089/ars.2021.0038

48.

Matsui

, Oshima

, et al. Early embryonic lethality caused by targeted disruption of the mouse thioredoxin gene. Dev Biol, 1996; 178(1):179–185.

49.

Matsuzawa

, Ichijo

. Redox control of cell fate by MAP kinase: Physiological roles of ASK1-MAP kinase pathway in stress signaling. Biochim Biophys Acta, 2008; 1780(11):1325–1336; doi: 10.1016/j.bbagen.2007.12.011

50.

Meylan

, Pich

, Winkler

, et al. Low expression of the PPARgamma-regulated gene thioredoxin-interacting protein accompanies human melanoma progression and promotes experimental lung metastases. Sci Rep, 2021; 11(1):7847; doi: 10.1038/s41598-021-86329-5

51.

Miligy

, Gorringe

, Toss

, et al. Thioredoxin-interacting protein is an independent risk stratifier for breast ductal carcinoma in situ. Mod Pathol, 2018; 31(12):1807–1815; doi: 10.1038/s41379-018-0086-7

52.

Nishiyama

, Matsui

, Iwata

, et al. Identification of thioredoxin-binding protein-2/vitamin D3 up-regulated protein 1 as a negative regulator of thioredoxin function and expression. J Biol Chem, 1999; 274(31):21645–21650; doi: 10.1074/jbc.274.31.21645

53.

Oka

, Liu

, Masutani

, et al. Impaired fatty acid utilization in thioredoxin binding protein-2 (TBP-2)-deficient mice: A unique animal model of Reye syndrome. FASEB J, 2006; 20(1):121–123.

54.

Oka

, Yoshihara

, Bizen-Abe

, et al. Thioredoxin binding protein-2/thioredoxin-interacting protein is a critical regulator of insulin secretion and peroxisome proliferator-activated receptor function. Endocrinology, 2009; 150(3):1225–1234; doi: 10.1210/en.2008-0646

55.

Oliveira

, Discola

, Alves

, et al. Insights into the specificity of thioredoxin reductase-thioredoxin interactions. A structural and functional investigation of the yeast thioredoxin system. Biochemistry, 2010; 49(15):3317–3326; doi: 10.1021/bi901962p

56.

Parikh

, Carlsson

, Chutkow

, et al. TXNIP regulates peripheral glucose metabolism in humans. PLoS Med, 2007; 4(5):e158; doi: 10.1371/journal.pmed.0040158

57.

Park

, Jeon

J-H

, Min

B-K

, et al. Role of the pyruvate dehydrogenase complex in metabolic remodeling: Differential pyruvate dehydrogenase complex functions in metabolism. Diabetes Metab J, 2018; 42(4):270–281; doi: 10.4093/dmj.2018.0101

58.

Patwari

, Chutkow

, Cummings

, et al. Thioredoxin-independent regulation of metabolism by the alpha-arrestin proteins. J Biol Chem, 2009; 284(37):24996–25003; doi: 10.1074/jbc.M109.018093

59.

Patwari

, Higgins

, Chutkow

, et al. The interaction of thioredoxin with Txnip. Evidence for formation of a mixed disulfide by disulfide exchange. J Biol Chem, 2006; 281(31):21884–21891; doi: 10.1074/jbc.M600427200

60.

Pelicano

, Martin

, Xu

, et al. Glycolysis inhibition for anticancer treatment. Oncogene, 2006; 25(34):4633–4646; doi: 10.1038/sj.onc.1209597

61.

Peskin

, Dickerhof

, Poynton

, et al. Hyperoxidation of peroxiredoxins 2 and 3: Rate constants for the reactions of the sulfenic acid of the peroxidatic cysteine. J Biol Chem, 2013; 288(20):14170–14177; doi: 10.1074/jbc.M113.460881

62.

Racz

, Nagy

, Tovari

, et al. Identification of new reference genes with stable expression patterns for gene expression studies using human cancer and normal cell lines. Sci Rep, 2021; 11(1):19459; doi: 10.1038/s41598-021-98869-x

63.

Rhee

, Woo

, Kil

, et al. Peroxiredoxin functions as a peroxidase and a regulator and sensor of local peroxides. J Biol Chem, 2012; 287(7):4403–4410; doi: 10.1074/jbc.R111.283432

64.

Saitoh

MNH

, Fujii

, Takeda

, et al. Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK). EMBO J, 1998; 17(9):2596–2606; doi: 10.1093/emboj/17.9.2596

65.

Schwertassek

, Haque

, Krishnan

, et al. Reactivation of oxidized PTP1B and PTEN by thioredoxin 1. FEBS J, 2014; 281(16):3545–3558; doi: 10.1111/febs.12898

66.

Sen

. Redox signaling and the emerging therapeutic potential of thiol antioxidants. Biochem Pharmacol, 1998; 55(11):1747–1758; doi: 10.1016/s0006-2952(97)00672-2

67.

Shang

, Xie

, Lu

, et al. Increased thioredoxin-1 expression promotes cancer progression and predicts poor prognosis in patients with gastric cancer. Oxid Med Cell Longev, 2019; 2019:9291683; doi: 10.1155/2019/9291683

68.

Sheth

, Bodnar

, Ghazalpour

, et al. Hepatocellular carcinoma in Txnip-deficient mice. Oncogene, 2006; 25(25):3528–3536; doi: 10.1038/sj.onc.1209394

69.

Sheth

, Castellani

, Chari

, et al. Thioredoxin-interacting protein deficiency disrupts the fasting-feeding metabolic transition. J Lipid Res, 2005; 46(1):123–134; doi: 10.1194/jlr.M400341-JLR200

70.

Stein

, Geserick

, Friedrich

, et al. Pro-inflammatory cytokines as regulators of protein tyrosine phosphatases and insulin signaling in murine skeletal muscle cells. Cell Signal, 2025; 135:112037; doi: 10.1016/j.cellsig.2025.112037

71.

Stine

, Schug

, Salvino

, et al. Targeting cancer metabolism in the era of precision oncology. Nat Rev Drug Discov, 2022; 21(2):141–162; doi: 10.1038/s41573-021-00339-6

72.

Stoltzman

, Kaadige

, Peterson

, et al. MondoA senses non-glucose sugars: Regulation of thioredoxin-interacting protein (TXNIP) and the hexose transport curb. J Biol Chem, 2011; 286(44):38027–38034; doi: 10.1074/jbc.M111.275503

73.

Stoltzman

, Peterson

, Breen

, et al. Glucose sensing by MondoA:Mlx complexes: A role for hexokinases and direct regulation of thioredoxin-interacting protein expression. Proc Natl Acad Sci U S A, 2008; 105(19):6912–6917.

74.

Surh

, Kundu

, Na

. Nrf2 as a master redox switch in turning on the cellular signaling involved in the induction of cytoprotective genes by some chemopreventive phytochemicals. Planta Med, 2008; 74(13):1526–1539; doi: 10.1055/s-0028-1088302

75.

Suzuki

, Takahashi

, Yamamoto

. Molecular basis of the KEAP1-NRF2 signaling pathway. Mol Cells, 2023; 46(3):133–141; doi: 10.14348/molcells.2023.0028

76.

Tonelli

, Chio

IIC

, Tuveson

. Transcriptional regulation by Nrf2. Antioxid Redox Signal, 2018; 29(17):1727–1745; doi: 10.1089/ars.2017.7342

77.

Williams

, Arscott

, Muller

, et al. Thioredoxin reductase two modes of catalysis have evolved. Eur J Biochem, 2000; 267(20):6110–6117; doi: 10.1046/j.1432-1327.2000.01702.x

78.

Woolston

, Zhang

, Storr

, et al. The prognostic and predictive power of redox protein expression for anthracycline-based chemotherapy response in locally advanced breast cancer. Mod Pathol, 2012; 25(8):1106–1116; doi: 10.1038/modpathol.2012.60

79.

, Khosla

. Thiol-disulfide exchange reactions in the Mammalian extracellular environment. Annu Rev Chem Biomol Eng, 2016; 7:197–222; doi: 10.1146/annurev-chembioeng-080615-033553

80.

Yoshihara

, Fujimoto

, Inagaki

, et al. Disruption of TBP-2 ameliorates insulin sensitivity and secretion without affecting obesity. Nat Commun, 2010; 1:127; doi: 10.1038/ncomms1127

81.

Yoshihara

, Masaki

, Matsuo

, et al. Thioredoxin/Txnip: Redoxisome, as a redox switch for the pathogenesis of diseases. Front Immunol, 2014; 4:514; doi: 10.3389/fimmu.2013.00514

82.

Yoshioka

, Chutkow

, Lee

, et al. Deletion of thioredoxin-interacting protein in mice impairs mitochondrial function but protects the myocardium from ischemia-reperfusion injury. J Clin Invest, 2012; 122(1):267–279; doi: 10.1172/JCI44927

83.

Yoshioka

, Imahashi

, Gabel

, et al. Targeted deletion of thioredoxin-interacting protein regulates cardiac dysfunction in response to pressure overload. Circ Res, 2007; 101(12):1328–1338; doi: 10.1161/CIRCRESAHA.106.160515

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

3.87 MB

Thioredoxin 1 Suppresses TXNIP-Driven Control of Glucose Metabolism in Human Cells

Abstract

Aims:

Results:

Innovation:

Conclusion:

Keywords

Introduction

Innovation

Results

Cellular specific activity of TRX is not increased in patient-derived TXNIP-deficient fibroblasts or myoblasts compared with control cells

Validation of HeLa cell knockdown system

The oxidation of PRDX1, PRDX2, and PRDX3 remained unchanged in TXNIP-deficient primary cells upon treatment with H2O2, but increased in HeLa cells upon knockdown of TRX

Inverse relationship between TXNIP and TRX levels in relation to PDHC phosphorylation

Inverse relationship between TXNIP and TRX with cellular glucose uptake

Discussion

Materials and Methods

Ethical considerations

Primary cell culture

Immunoblot analyses

TRX activity as assessed by TXNRD-driven insulin disulfide reduction capacity

Doxycycline-inducible TRX knockdown HeLa cells

Glucose uptake assay with 2-NBDG

Assessing PRDX (PRDX1, PRDX2, and PRDX3) oxidation using H2O2 treatment

Transcript determinations using RT-PCR

Immunocytochemistry

Statistical analyses

Footnotes

Author Disclosure Statement

Funding Information

Supplemental Material

Abbreviations

References

Supplementary Material

The oxidation of PRDX1, PRDX2, and PRDX3 remained unchanged in TXNIP-deficient primary cells upon treatment with H₂O₂, but increased in HeLa cells upon knockdown of TRX

Assessing PRDX (PRDX1, PRDX2, and PRDX3) oxidation using H₂O₂ treatment