Anti-Inflammatory Effect of Dimethyl Fumarate Associates with the Inhibition of Thioredoxin Reductase 1 in RAW 264.7 Cells

Macrophages secrete a variety of pro-inflammatory cytokines in response to pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) but abnormal release of cytokines unfortunately promotes cytokine storms. Dimethyl fumarate (DMF), an FDA-approved drug for multiple sclerosis (MS) treatment, has been found as an effective therapeutic agent for resolution. In this study, the anti-inflammatory effect of DMF was found to correlate to selenoprotein thioredoxin reductase 1 (TXNRD1). DMF irreversibly modified the Sec498 residue and C-terminal catalytic cysteine residues of TXNRD1 in a time- and dose-dependent manner. In LPS-stimulated RAW 264.7 cells, cellular TXNRD activity was increased through up-regulation of the protein level and DMF inhibited TXNRD activity and the nitric oxide (NO) production of RAW 264.7 cells. Meanwhile, the inhibition of TXNRD1 by DMF would contribute to the redox regulation of inflammation and promote the nuclear factor erythroid 2-related factor 2 (NRF2) activation. Notably, inhibition of cellular TXNRD1 by auranofin or TRi-1 showed anti-inflammatory effect in RAW 264.7 cells. This finding demonstrated that targeting TXNRD1 is a potential mechanism of using immunometabolites for dousing inflammation in response to pathogens and highlights the potential of TXNRD1 inhibitors in immune regulation.


Introduction
Selenoprotein thioredoxin reductase 1 (TXNRD1, encoded by txnrd1) is a basal antioxidant enzyme that participates in multiple cellular processes, such as antioxidant defense, programmed cell death, and regulation of cell fate [1,2]. TXNRD1 can scavenge excessive reactive oxygen species (ROS) by reducing and recycling thioredoxin (TXN, encoded by txn) or 14 kDa thioredoxin-related protein (TRP14, encoded by txndc17), or directly quenching free radicals through reacting with compounds or biomolecules [3][4][5]. Since the TXN system was known to modulate gene expression by promoting the activity of transcription factors such as NF-κB and HIF-1α, TXNRD1 was heretofore considered as a negative regulator of KEAP1-NRF2-ARE antioxidant pathway [6,7].

DMF Is Not a Substrate but an Inhibitor of TXNRD1
Monomethyl fumarate (MMF) and DMF are derivatives of metabolite fumaric acid (FA), as shown in Figure 1A. Regarding the possible interactions between TXNRD1 and DMF, we first questioned whether the three compounds are substrates of TXNRD1, able to accept electrons from NADPH. Two canonical TXNRD1 substrates, 9,10-PQ and juglone, were used in this experiment and NADPH consumption was observed using recombinant rat TXNRD1. However, FA or MMF or DMF failed to display any NADPH consumption, indicating that FA and its derivatives MMF and DMF were not substrates of the enzyme ( Figure 1B). Then we tested the inhibitory effect of these three molecules on recombinant rat TXNRD1. After incubation for the indicated time, DMF but not MMF and FA, significantly inhibited the TXNRD1 based on the DTNB reducing activity assay; the IC 50 value was approx. 30 µM, depending on the duration of incubation ( Figure 1C). This result showed that DMF is an inhibitor of TXNRD1 and not involved in the redox recycling of the enzyme.

DMF Irreversibly Inhibits TXNRD1 Activity but Shows Much Less Inhibition on GSR
We then investigated the inhibition mechanism of DMF on TNXRD1. In addition to the DTNB assay, the 9,10-PQ reducing activity, a highly Sec-dependent substrate of (C) Dose-and time-dependent inhibition of DMF on TXNRD1. Rat TXNRD1 was incubated with FA, MMF and DMF for the indicated time. The TXNRD1 activity was assayed by classic DTNB reducing activity assay. Data are presented as mean ± SEM with three independent experiments.

DMF Irreversibly Inhibits TXNRD1 Activity but Shows Much Less Inhibition on GSR
We then investigated the inhibition mechanism of DMF on TNXRD1. In addition to the DTNB assay, the 9,10-PQ reducing activity, a highly Sec-dependent substrate of TXNRD1, was also inhibited by DMF (Figure 2A), whereas the juglone reduction activity of TXNRD1 was not so greatly inhibited ( Figure 2B). After removing the dissociative DMF from the incubation system using a NAP TM -5 desalting column, the residual activity was not significantly rescued, indicating an irreversible inhibition of DMF on TXNRD1 ( Figure 2C). Glutathione reductase (GSR, encoded by gsr) is a flavoprotein akin to TXNRD1, but lacking selenocysteine at its C-terminus. Using a high concentration (up to 500 µM), DMF showed irreversible inhibition towards the GSSG reducing activity of GSR ( Figure 2D). glone and 30 μM 9,10-PQ were used as positive controls in this experiment. FA, MMF and DMF were included for testing. The final mixture solution contained 30 nM TXNRD1, 200 μM NADPH and 200 μM compounds. (C) Dose-and time-dependent inhibition of DMF on TXNRD1. Rat TXNRD1 was incubated with FA, MMF and DMF for the indicated time. The TXNRD1 activity was assayed by classic DTNB reducing activity assay. Data are presented as mean ± SEM with three independent experiments.

DMF Irreversibly Inhibits TXNRD1 Activity but Shows Much Less Inhibition on GSR
We then investigated the inhibition mechanism of DMF on TNXRD1. In addition to the DTNB assay, the 9,10-PQ reducing activity, a highly Sec-dependent substrate of TXNRD1, was also inhibited by DMF (Figure 2A), whereas the juglone reduction activity of TXNRD1 was not so greatly inhibited ( Figure 2B). After removing the dissociative DMF from the incubation system using a NAP TM -5 desalting column, the residual activity was not significantly rescued, indicating an irreversible inhibition of DMF on TXNRD1 ( Figure  2C). Glutathione reductase (GSR, encoded by gsr) is a flavoprotein akin to TXNRD1, but lacking selenocysteine at its C-terminus. Using a high concentration (up to 500 μM), DMF showed irreversible inhibition towards the GSSG reducing activity of GSR ( Figure 2D).  Inhibition of 9,10-PQ and juglone reducing activity of TXNRD1 by DMF. Rat TXNRD1 (0.3 µM) was treated with DMF for 1 h, and then the 9,10-PQ and juglone reducing activities were assayed, respectively. (C) Irreversible inhibition of DMF on TXNRD1. TXNRD1 was treated with 100 µM DMF for 2 h and then desalted using NAP TM -5 desalting column. The residual TXNRD1 activity was measured by DTNB reducing activity. (D) Less inhibition of DMF on GSR. Yeast GSR was incubated with 500 µM DMF for 2 h and then desalted using NAP TM -5 desalting column. The activity of GSR was determined in presence of 200 µM NADPH by reducing GSSG. Data are presented as mean ± SEM with three independent experiments. n.s., not significant.

DMF Modifies the Selenocysteine and Cysteine Residues of TXNRD1
TXNRD1 contains a N-terminal GSR-like "-CVNVGC-" motif and a C-terminal Seccontaining "-GCUG" redox motif. To reveal the DMF inhibition in detail and identify the key residues of DMF targeting TXNRD1, we prepared five mutant variants of rat TXNRD1 and further analyzed the inhibition. Compared to the wild-type enzyme (-GCUG), the Sec-deficient mutation of TXNRD1 (-GCCG, -GSCG, -GCSG, -GC and -GSSG) conferred resistance to DMF, suggesting DMF mainly targets the Sec 498 residue of TXNRD1 ( Figure 3A). However, the Sec-deficient TXNRD1 mutant variants were inhibited by DMF, indicating that DMF can modify the C-terminal cysteine residues or even the N-terminal cysteine residues of TXNRD1 ( Figure 3A), which is consistent with the result that DMF inhibited GSR activity. These results elucidated that DMF inhibits both the C-terminal reducing activity and N-terminal NADPH oxidase activity of TXNRD1.
To reveal the inhibitory details of DMF on TXNRD1, we determined the k inact of DMF on wild-type TXNRD1 and its GCCG mutant (Sec 498 to Cys 498 , U498C). The k inact of the two variants was 1.426 × 10 −4 and 2.334 × 10 −5 µM −1 min −1 , respectively. The six-fold difference in k inact indicated the preferentially targeted residue of TXNRD1 by DMF ( Figure 3B,C).

DMF Inhibits Three Species of TXNRD1
Rat TXNRD1 is highly homologous to human and mouse species, regarding the amino acid sequence. To validate the inhibitory effect and targeting priority of DMF, we recombinantly expressed and purified human TXNRD1 and mouse TXNRD1 and performed the SDS-PAGE analysis and Native-PAGE analysis together with rat TXNRD1 ( Figure 4A). In parallel, we tested the inhibition of DMF on TXNRD1 from three species. DMF showed inhibition towards the three TXNRD1s. The specific activity of the enzymes was 8 U/mg for human TXNRD1, 11 U/mg for mouse TXNRD1 and 22.7 U/mg for rat TXNRD1 ( Figure 4B). teine residues of TXNRD1 ( Figure 3A), which is consistent with the result that DMF inhibited GSR activity. These results elucidated that DMF inhibits both the C-terminal reducing activity and N-terminal NADPH oxidase activity of TXNRD1.
To reveal the inhibitory details of DMF on TXNRD1, we determined the kinact of DMF on wild-type TXNRD1 and its GCCG mutant (Sec 498 to Cys 498 , U498C). The kinact of the two variants was 1.426 × 10 −4 and 2.334 × 10 −5 μM −1 min −1 , respectively. The six-fold difference in kinact indicated the preferentially targeted residue of TXNRD1 by DMF ( Figure 3B,C). Figure 3. The C-terminal redox motif of TXNRD1 is more prone to be targeted by DMF than Nterminal domains. (A) Inhibition of TXNRD1 variants by DMF. Wild-type TXNRD1 (0.1 μM) and its mutant variants (1 μM) were pre-reduced by 100 μM NADPH, and then treated with DMF for 2 h; the remained activity was determined by DTNB reducing activity. The final reaction mixture contained 10 nM wild-type TXNRD1 or 100 nM TXNRD1 mutant variants. (B,C) Time-dependent inhibition of DMF on wild-type rat TXNRD1 and its GCCG mutant. 0.2 μM wild-type TXNRD1 and 2 μM its GCCG mutant were pre-reduced by NADPH and then treated with DMF for the indicated time. The residual enzyme activity was assayed by DTNB reducing assay. Data are presented as mean ± SEM with three independent experiments. Figure 3. The C-terminal redox motif of TXNRD1 is more prone to be targeted by DMF than Nterminal domains. (A) Inhibition of TXNRD1 variants by DMF. Wild-type TXNRD1 (0.1 µM) and its mutant variants (1 µM) were pre-reduced by 100 µM NADPH, and then treated with DMF for 2 h; the remained activity was determined by DTNB reducing activity. The final reaction mixture contained 10 nM wild-type TXNRD1 or 100 nM TXNRD1 mutant variants. (B,C) Time-dependent inhibition of DMF on wild-type rat TXNRD1 and its GCCG mutant. 0.2 µM wild-type TXNRD1 and 2 µM its GCCG mutant were pre-reduced by NADPH and then treated with DMF for the indicated time. The residual enzyme activity was assayed by DTNB reducing assay. Data are presented as mean ± SEM with three independent experiments.

DMF Inhibits Three Species of TXNRD1
Rat TXNRD1 is highly homologous to human and mouse species, regarding the amino acid sequence. To validate the inhibitory effect and targeting priority of DMF, we recombinantly expressed and purified human TXNRD1 and mouse TXNRD1 and performed the SDS-PAGE analysis and Native-PAGE analysis together with rat TXNRD1 ( Figure 4A). In parallel, we tested the inhibition of DMF on TXNRD1 from three species. DMF showed inhibition towards the three TXNRD1s. The specific activity of the enzymes was 8 U/mg for human TXNRD1, 11 U/mg for mouse TXNRD1 and 22.7 U/mg for rat TXNRD1 ( Figure 4B).

LPS Stimulation Increases Cellular TXNRD1 Activity through up-Regulated TXNRD1 Level
We were curious whether the anti-inflammation effect of DMF is correlated with the inhibition on TXNRD1. Upon LPS treatment, the cellular morphology of RAW 264.7 cells was changed from round morphology into spindly morphology and the NO production

LPS Stimulation Increases Cellular TXNRD1 Activity through Up-Regulated TXNRD1 Level
We were curious whether the anti-inflammation effect of DMF is correlated with the inhibition on TXNRD1. Upon LPS treatment, the cellular morphology of RAW 264.7 cells was changed from round morphology into spindly morphology and the NO production was significantly increased, indicating the polarization of macrophages into M1 type ( Figure 5A,B). Meanwhile, the cellular TXNRD1 activity was significantly increased, and the protein level of TXNRD1 was up-regulated ( Figure 5C,D).

Anti-Inflammation Effect of DMF Is Correlated with TXNRD1 Activity
Previous studies showed that DMF inhibited the inflammation of macrophages by altering NRF2 and NF-κB activity [20,22]. In this study, we found that DMF treatment impaired the cell viability of RAW 264.7 cells ( Figure 6A), whereas the cellular TXNRD activity was not strongly affected ( Figure 6B). However, upon LPS-stimulation, DMF showed no cytotoxicity towards RAW 264.7 cells ( Figure 6C) but could decrease the NO production and alleviate the cellular morphology changes, suggesting an anti-inflammation effect of DMF ( Figure 6D,E). Notably, the cellular TXNRD1 activity was decreased ( Figure 6F) without observed down-regulation of TXNRD1, indicating the directly inhibitory effect of DMF on TXNRD1 ( Figure 6G). What is more, after being treated with LPS, immunoresponsive gene 1 (IRG1), a gene that is highly expressed in mitochondria of macrophages under pro-inflammatory states, was increased and the genes, such as hemeoxygenase-1 (HO-1) and GSR, which are controlled by the KEAP1-NRF2-ARE antioxidant system, were significantly up-regulated, suggesting that DMF activates the NRF2 transcription activity ( Figure 6G). Interestingly, the cystine/glutamate transporter SLC7A11 was decreased, which inconsistent with other NRF2 targets. Previous reports showed that DMF induces ferroptosis in cancer cells, which may explain the phenomenon, as SLC7A11 is down-regulated along with the accumulation of lipid peroxidation and ferroptosis process [26,27]. Taken together, these results suggested that DMF directly inhibited cellular TXNRD activity and increased the NRF2 transcription activity.
We then used a potent TXNRD1 inhibitor, auranofin (AF) and TRi-1 to verify the role of TXNRD1 in LPS-stimulated macrophages. The NO production in LPS-induced macrophages was significantly decreased (Figure 6H), indicating that targeting TXNRD1 is a potential means for dousing inflammation.

Anti-Inflammation Effect of DMF Is Correlated with TXNRD1 Activity
Previous studies showed that DMF inhibited the inflammation of macrophages by altering NRF2 and NF-κB activity [20,22]. In this study, we found that DMF treatment impaired the cell viability of RAW 264.7 cells (Figure 6A), whereas the cellular TXNRD activity was not strongly affected ( Figure 6B). However, upon LPS-stimulation, DMF showed no cytotoxicity towards RAW 264.7 cells ( Figure 6C) but could decrease the NO production and alleviate the cellular morphology changes, suggesting an anti-inflammation effect of DMF ( Figure 6D,E). Notably, the cellular TXNRD1 activity was decreased ( Figure 6F) without observed down-regulation of TXNRD1, indicating the directly inhibitory effect of DMF on TXNRD1 ( Figure 6G). What is more, after being treated with LPS, immunoresponsive gene 1 (IRG1), a gene that is highly expressed in mitochondria of macrophages under proinflammatory states, was increased and the genes, such as hemeoxygenase-1 (HO-1) and GSR, which are controlled by the KEAP1-NRF2-ARE antioxidant system, were significantly up-regulated, suggesting that DMF activates the NRF2 transcription activity ( Figure 6G). Interestingly, the cystine/glutamate transporter SLC7A11 was decreased, which inconsistent with other NRF2 targets. Previous reports showed that DMF induces ferroptosis in cancer cells, which may explain the phenomenon, as SLC7A11 is down-regulated along with the accumulation of lipid peroxidation and ferroptosis process [26,27]. Taken together, these results suggested that DMF directly inhibited cellular TXNRD activity and increased the NRF2 transcription activity.
We then used a potent TXNRD1 inhibitor, auranofin (AF) and TRi-1 to verify the role of TXNRD1 in LPS-stimulated macrophages. The NO production in LPS-induced macrophages was significantly decreased ( Figure 6H), indicating that targeting TXNRD1 is a potential means for dousing inflammation.  Taken together, our results showed that DMF inhibits TXNRD1's activity by modifying the Sec and Cys residues in the catalytic motif of TXNRD1. In LPS-stimulated RAW 264.7 cells, the TXNRD1 was up-regulated but DMF directly inhibits the cellular TXNRD1 activity, which may contribute to the anti-inflammatory effect of DMF. We also observed the upregulation of NRF2 target genes in DMF-treated cells, such as HO-1 and GSR. We speculate that the inhibition of TXNRD1 would promote the NRF2 antioxidant system, which further promotes the anti-inflammation process (Figure 7). Taken together, our results showed that DMF inhibits TXNRD1's activity by modifying the Sec and Cys residues in the catalytic motif of TXNRD1. In LPS-stimulated RAW 264.7 cells, the TXNRD1 was up-regulated but DMF directly inhibits the cellular TXNRD1 activity, which may contribute to the anti-inflammatory effect of DMF. We also observed the up-regulation of NRF2 target genes in DMF-treated cells, such as HO-1 and GSR. We speculate that the inhibition of TXNRD1 would promote the NRF2 antioxidant system, which further promotes the anti-inflammation process (Figure 7).

Discussion
DMF is an electrophilic molecule and a well-known immunomodulatory drug, harboring anti-inflammatory effects via activating NRF2 or inducing electrophile stress through GSH conjugation and Cys-residue modification in enzymes. In the present study, we showed that DMF inhibits the activity of selenoprotein TXNRD1, which further promotes the dissociation of NRF2 from KEAP1, and increases expression of antioxidant enzymes and detoxifying enzymes, such as GSR and HO-1 [28]. As expected, the inhibitory action of DMF on TXNRD1 will enhance its anti-inflammatory effect, precisely as exactly as later we showed in cultured RAW 264.7 cells in vitro.
The discovery of TXNRD inhibitors attracted considerable interest, due to the vital roles of TXNRD1 in tumor progression and metastasis [29,30]. It is quite a big surprise that DMF exhibits an inhibition effect on TXNRD1, given that the profound influences of DMF and such inhibition on cellular function have not been fully understood. Differing from MMF and FA, the action of DMF on TXNRD1 indicates that the electrophilicity of the molecule is vital for TXNRD1 inhibition. Compared with other TXNRD1 inhibitors such as quinones [31,32] or gold-based compounds such as auranofin [33], DMF is not a potent inhibitor of TXNRD1, which also could be inferred by the electrophilicity of DMF. However, it is worth noting that DMF inhibits both the N-terminal and the C-terminal redox active motifs of TXNRD1, distinct from some Sec-targeting electrophilic drugs. Previous works showed that electrophilic molecules modified the Sec 498 residue of TXNRD1 at the C-terminal producing selenium compromised thioredoxin reductase-derived apoptotic proteins (SecTRAPs) which are devoid of TXNRD1's activity but can induce rapid cell death by directly generating ROS through its N-terminal NADPH oxidation domain [31,[34][35][36][37]. We inferred that the inhibitory effect of DMF on TXNRD1 works more on the cellular function of TXNRD1 rather than directly generating ROS. Meanwhile, the critical

Discussion
DMF is an electrophilic molecule and a well-known immunomodulatory drug, harboring anti-inflammatory effects via activating NRF2 or inducing electrophile stress through GSH conjugation and Cys-residue modification in enzymes. In the present study, we showed that DMF inhibits the activity of selenoprotein TXNRD1, which further promotes the dissociation of NRF2 from KEAP1, and increases expression of antioxidant enzymes and detoxifying enzymes, such as GSR and HO-1 [28]. As expected, the inhibitory action of DMF on TXNRD1 will enhance its anti-inflammatory effect, precisely as exactly as later we showed in cultured RAW 264.7 cells in vitro.
The discovery of TXNRD inhibitors attracted considerable interest, due to the vital roles of TXNRD1 in tumor progression and metastasis [29,30]. It is quite a big surprise that DMF exhibits an inhibition effect on TXNRD1, given that the profound influences of DMF and such inhibition on cellular function have not been fully understood. Differing from MMF and FA, the action of DMF on TXNRD1 indicates that the electrophilicity of the molecule is vital for TXNRD1 inhibition. Compared with other TXNRD1 inhibitors such as quinones [31,32] or gold-based compounds such as auranofin [33], DMF is not a potent inhibitor of TXNRD1, which also could be inferred by the electrophilicity of DMF. However, it is worth noting that DMF inhibits both the N-terminal and the C-terminal redox active motifs of TXNRD1, distinct from some Sec-targeting electrophilic drugs. Previous works showed that electrophilic molecules modified the Sec 498 residue of TXNRD1 at the C-terminal producing selenium compromised thioredoxin reductase-derived apoptotic proteins (SecTRAPs) which are devoid of TXNRD1's activity but can induce rapid cell death by directly generating ROS through its N-terminal NADPH oxidation domain [31,[34][35][36][37]. We inferred that the inhibitory effect of DMF on TXNRD1 works more on the cellular function of TXNRD1 rather than directly generating ROS. Meanwhile, the critical role of TXNRD1 implies that activity loss of TXNRD1 will rewire some cellular processes and alter the tolerance of cells to some stress [38,39]. This may be considered means of DMF for inducing cancer cell death and regulating cellular inflammation. TXNRD1 and GRX maintain the reduced states of TXN, working together with the GSH system for regulating cell redox balance [40][41][42][43]. We confirmed a major function of TXNRD1 in the current study; it can serve as a negative regulator of the KEAP1-NRF2-ARE antioxidant system as previously reported [44,45]. KEAP1 is a sensor of oxidative stress and electrophile stress [28]. In the canonical pathway, ROS-mediated oxidation or electrophilesinduced modification of the cysteine residue of KEAP1 causes its dissociation from NRF2, which interrupts the ubiquitin-dependent proteasomal degradation and thereby promotes NRF2 activation [46]. However, based on the high chemical reactivity of the Sec residue on TXNRD1, the electrophiles are prone to attack the TXNRD1, and affect the balance of multiple downstream redox signaling and cellular redox activities [40,47]. We propose that the induction of NRF2 by DMF is associated with the inhibition of TXNRD1, which further promotes the anti-inflammation activity of DMF.

Cultured Cell Lines
RAW 264.7 cell line was purchased from Procell Corp. (Wuhan, China) and cultured in DMEM (Procell, Wuhan, China). All the cultured cells were supplemented with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin in a humidified incubator (Heal Force, Shanghai, China) with an atmosphere of 5% CO 2 at 37 • C.

Cell Viability Assay
MTT assay was used to measure the cell viability. In brief, RAW 264.7 cells (10,000 per well) were seeded into 96-well plates and incubated at 37 • C overnight. Cells were then incubated with the different concentrations of DMF for the indicated time. Afterwards, the media were discarded, and fresh media containing 0.5 mg/mL MTT were added and continually incubated for an additional 4 h. Finally, the formed formazan crystals were dissolved in 100 µL DMSO and the absorbance was detected using an Infinite 200 PRO plate reader (Tecan, Männedorf, Switzerland) at 570 nm, with 630 nm as a reference.

Expression and Purification of TXNRD1
Human, rat and mouse TXNRD1 were expressed in BL21(DE3) gor strain with bacterial SECIS element located downstream of the UGA codon. Plasmid pSUABC, which contains selA, selB and selC genes, was co-transformed with the plasmid pET-TRS TER to the strain to improve the Sec insertion efficiency. The recombinant enzyme was induced according to "2.4/24/24" protocol based on the previous report [48,49]. Purification of the enzyme was performed by affinity chromatography using ADP-sepharose followed by a gel filtration using Hiprep 16/60 Sepharcryl S-300. All the enzymes were stored in a TE buffer, pH = 7.5. The specific activities of the human, rat and mouse TXNRD1 are 8.0 U/mg, 22.7 U/mg and 11.0 U/mg, respectively, using DTNB as the substrate. TXNRD1 variants used in this study include Sec 498 to Cys 498 (-GCCG, U498C), Cys 497 to Ser 497 and Sec 498 to Cys 498 (-GSCG, C497S/U498C), Sec 498 to Ser 498 (-GCSG, U498S), Truncated form (-GC, 2) and Cys 497 to Ser 497 and Sec 498 to Ser 498 (-GSSG, C497S/U498S).

Recombinant TXNRD1 Activity Assay
Activity of recombinant rat TXNRD1 was determined by using DTNB reducing assay according to the previous studies [32]. In brief, the reaction mixture contained 2.5 mM DTNB, 300 µM NADPH, and 10 nM W.T. TXNRD1 or 100 nM TXNRD1 mutants as indicated, in 50 mM TE buffer (pH 7.5), and the enzyme activity was calculated by the TNB − formation at 412 nm (εTNB − = 13,600 M −1 cm −1 ). In the 9,10-PQ reducing assay, the final mixture contained 30 nM W.T. TXNRD1, 30 µM 9,10-PQ, and 200 µM NADPH in 50 mM TE buffer (pH 7.5). The enzymatic activity of 9,10-PQ reduction was calculated by the NADPH oxidation at 340 nm (εNADPH = 6200 M −1 cm −1 ). All reactions were performed in an Infinite 200 PRO plate reader (Tecan, Männedorf, Switzerland) at 25 • C. The same reaction mixture lacking the enzyme was used as a control.

DMF Treatment on Cells
RAW 264.7 cells were seeded into 6-well plates at 400,000 cells per well and treated with different doses of DMF for 24 h. The cells were washed with ice-cold PBS buffer three times and all adherent cells were lysed by the RIPA buffer with 1 mM PMSF protease inhibitor (Solarbio, Beijing, China). Then, the cell lysate was centrifuged at 18,000 rpm at 4 • C for 20 min. Total protein contents were determined by BCA kit (Beyotime, Shanghai, China).

Cellular TXNRD Activity Assay
Cellular TXNRD activity was determined by using end-point insulin assay as previously described [50,51]. In brief, an appropriate amount of cell lysates was added into a master mixture containing 80 mM Hepes buffer (pH 7.5), 15 µM TXN1, 300 µM insulin, 660 µM NADPH, and 3 mM EDTA. A reaction mixture without TXN1 was used as a background control. Samples were incubated at 37 • C for 30 min. Subsequently, 6 M guanidine hydrochloride containing 1 mM DTNB, 10 mM EDTA was added to each well, and an endpoint at absorbance of 412 nm was measured. TXNRD activities of cell lysates were normalized to protein concentration for an accurate comparison.

Glutathione Reductase Activity
Recombinant yeast glutathione reductase (GSR) activity was determined using the oxidized glutathione (GSSG) as the substrate [52]. The final reaction system (200 µL) contained 1 mM GSSG, 2 nM yeast GSR, and 200 µM NADPH. GSR activity was calculated by following the NADPH oxidation at an absorbance of 340 nm (εNADPH = 6200 M −1 cm −1 ) using an Infinite 200 PRO plate reader (Tecan, Männedorf, Switzerland) at 25 • C. The same reaction mixture lacking the enzyme was used as a reference.

NO Production Determination
RAW 264.7 cells were seeded into 24-well plates with the density of 100,000 per well and cultured overnight. Next day, cells were incubated with LPS or DMF as indicated. Cell supernatants were collected and added to 96 well plates. 50 µL samples were co-incubated with 50 µL Griess Reagent I and 50 µL Griess Reagent II (Beyotime, Shanghai, China) for 15 min at room temperature (20 • C ± 1 • C). The amount of NO production was monitored at 540 nm by an Infinite 200 PRO plate reader.

Western Blot Assay
RAW 264.7 cells were lysed with RIPA buffer containing 1 mM PMSF for 20 min, and then centrifugated with 18,000 rpm at 4 • C. The equal mass samples were loaded onto SDS-PAGE gel and then transferred onto the PVDF membranes. The blots were blocked with skimmed milk for 1 h at room temperature and incubated with primary antibodies (TXNRD1 1:6000, GSR 1:3000, IRG1 1:5000, HO-1 1:2000, SLC7A11 1:3000) overnight at 4 • C, then incubated with secondary antibodies (1:7500) for 2 h at room temperature. The ECL solution was used to display the blots by image analyzer (Sagecreation, Beijing, China).

Statistical Analysis
All experiments were performed in triplicate and the data were presented as the mean ± SEM. Statistical differences between the two groups were analyzed by the Stu-dent's t-test. Comparisons among multiple groups were statistically assessed by one-way analysis of variance (ANOVA) and followed by a post hoc Scheffe test. The significant differences between groups were defined as * p < 0.05, ** p < 0.01, *** p < 0.001; n.s. means not significant.

Conclusions
In summary, our data revealed details of the inhibitory effect of DMF on TXNRD1, which contributes to both NRF2 activation activity and cell death induction activity of DMF. This work also highlighted that TXNRD1 may be a target of some electrophilic immunometabolite derivatives, such as itaconate. The role of TXNRD1 in immunoregulation should be further investigated.