Enhanced expression of thioredoxin‐interacting‐protein regulates oxidative DNA damage and aging

The “free radical theory of aging” suggests that reactive oxygen species (ROS) are responsible for age‐related loss of cellular functions and, therefore, represent the main cause of aging. Redox regulation by thioredoxin‐1 (TRX) plays a crucial role in responses to oxidative stress. We show that thioredoxin‐interacting protein (TXNIP), a negative regulator of TRX, plays a major role in maintaining the redox status and, thereby, influences aging processes. This role of TXNIP is conserved from flies to humans. Age‐dependent upregulation of TXNIP results in decreased stress resistance to oxidative challenge in primary human cells and in Drosophila. Experimental overexpression of TXNIP in flies shortens lifespan due to elevated oxidative DNA damage, whereas downregulation of TXNIP enhances oxidative stress resistance and extends lifespan.

the major factor contributing to aging and organ malfunction is accumulation of oxidative damage of macromolecules (e.g. lipids, proteins and DNA) over time [1][2][3][4]. One well accepted determinator of aging is accumulation of genetic damage throughout life [5]. Moreover, various premature aging diseases are the consequence of accumulation of DNA damage [6]. The integrity and stability of DNA are continuously challenged by ROS [7]. Interestingly, age-dependent dysregulation of the redox system and accumulation of DNA damage is connected by the cellular anti-oxidant TRX. It has been shown that TRX is involved in DNA replication and DNA repair [8,9]. Further studies demonstrated that redox regulation by TRX protected against aging and age-related diseases [10]. Overexpression of TRX in transgenic mice or in transgenic Drosophila melanogaster leads to lifespan extension [11][12][13]. Otherwise, loss of TRX leads to a decrease in lifespan as shown in C. elegans as well as in flies [14,15]. Down regulation of TRX in mice showed no beneficial effect on lifespan. However, these studies demonstrate that reduced levels of TRX may be more important for tumor development than aging [10]. Since, TRX levels remain constant during life we speculated that the activity of TRX is regulated by its natural inhibitor TXNIP during aging. TXNIP, also known as Vitamin-D 3 -Upregulated-Protein 1 (VDUP1), is a member of the a-arrestin family [16]. Here, we show for the first time that the TRX inhibitor TXNIP is upregulated during aging in primary human cells and Drosophila melanogaster. Thus, we elucidate a novel mechanism conserved from fly to man showing that age-dependent upregulation of TXNIP induces a perturbation of the intracellular redox equilibrium. TXNIP upregulation leads to accumulation of ROS and, concordantly, to an increase in oxidative DNA damage, both crucial hallmarks of aging. We demonstrate that in Drosophila increased TXNIP expression leads to induction of DNA damage and, therefore, to a significant reduction in median lifespan, whereas decreased TXNIP expression results in prolonged median lifespan due to lower oxidative DNA damage.

Chemicals
Chemicals were obtained from Sigma-Aldrich unless otherwise indicated. Hygromycin B was obtained from GERBU.
Primary human T cells were cultured at a concentration of 2 9 10 6 cellsÁmL À1 in RPMI 1640 supplemented with 10% FCS.

Blood donors
T cells were isolated from the blood of healthy human donors at the age of 20-25 years (n = 7) and above 55 years old (n = 16). Informed consent was obtained from all subjects before inclusion in the study. The study was conducted according to the ethical guidelines of the German Cancer Research Center and the Helsinki Declaration, and it was approved by the ethics committee II of the Ruprecht-Karls-University of Heidelberg, Germany.

Isolation of human peripheral T cells
Primary human T cells were purified as described [17]. Purity of the prepared T cells was verified by staining with PE-conjugated anti-CD3 antibodies followed by fluorescence-activated cell scanning (FACS) analysis.
Gene expression analysis in human hematopoietic progenitor cells CD34 + cells were isolated from cord blood or mobilized peripheral blood of 15 healthy donors between 27 and 73 years and analyzed by Affymetrix technology as described [18].

Generation of stable TXNIP knockdown
For production of lentiviral particles, HEK293T cells, pretreated with 25 lM chloroquine for 1 h, were transfected with vectors containing the shRNA against TXNIP (Open Biosystems, Heidelberg) and a plasmid mixture for gag, pol, env and VSV-G for pseudotyping. 8 h post transfection medium was replaced from packaging cells. After 2 days, the supernatant was passed through a 0.45 lm filter, supplemented with Polybrene (8 lgÁmL À1 ). 1x10 5 target cells were infected by spin occulation with 1 mL of viral supernatant. Stably transduced Jurkat cells were selected by puromycin (1 lgÁmL À1 ) and Doxycycline (Dox)-dependent shRNA expression was checked by Western blot analysis.

Generation of S2 overexpressing TXNIP
To generate S2 cells overexpressing TXNIP (OE-TXNIP), TXNIP was amplified by PCR from the cDNA clone (RE 65531, DGRC). The 5 0 -primer was modified to introduce an EcoRV restriction site, whereas the 3 0 -primer was modified to reconstitute a stop codon and to introduce an XhoI restriction site. The PCR product was cloned into pAc 5.1/ V5-HisA (Life Technologies). The corresponding plasmid pAc 5.1/V5-HisA without cDNA served as empty vector control (EV).

Quantitative PCR (qPCR)
Total RNA was isolated from cells, with the RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA from fly heads was isolated with TRIzol reagent (Life Technologies) according to the manufacturer's instructions. RNA was reverse transcribed and qPCR performed as described previously [19]. The housekeeping gene GAPDH (human samples) and rp49 (Drosophila samples) was used as control gene for normalization. For primer sequences Table S1.

Determination of ROS generation
Cells were stained with 5 lM H 2 DCF-DA (Life Technologies) for 10 min for human samples and 30 min for S2 cells. Cells were washed twice with ice-cold PBS and ROS generation was determined by FACS analysis. ROS generation was quantified as the increase in mean fluorescence intensity (MFI) calculated as reported previously [20].

Thioredoxin (TRX) activity assay
Analysis of TRX activity was assessed in protein lysates of primary T cells as described previously [21,22]. The fluorescence-based activity assay FkTRX-02 (BIOZOL Diagnostica) was performed according to manufacturer's instructions. For S2 as well as for whole fly lysates the Thioredoxin Activity Fluorescent Assay Kit (Cayman) was performed according to manufacturer's instructions. Here the method is based on the reduction of insulin by reduced TRX.

Assessment of cell death
Cell death of human cells was assessed as the decrease in the forward-to-side scatter profile compared to living cells and recalculated to "specific cell death," as described previously [17]. Cell death of S2 was analyzed by flow cytometry (FACS-Canto II, Becton Dickinson) by determining percentage of propidium-iodide (PI) positive cells (ex488/em617). 5 9 10 5 cells were collected and once washed with PBS. After washing, cells were stained with PI (2 lgÁmL À1 ) and incubated for 15 min at RT.

Fly strains and husbandry
The following strains of Drosophila melanogaster were used: isogenic w 1118 (here referred to as wild type WT, Bloomington Stock Center), yw; +;P[UAS-TXNIP] [25],

Lifespan studies
Age-matched flies were collected as previously described [27]. In brief, w 1118 ; P[UAS-TXNIP-RNAi]; + or w 1118 ; +; P[UAS-TXNIP] and w 1118 ; +; + females were crossed to w 1118 ; +; P[tub-GAL4]/TM6B males. Flies of the F1 generation were mated for 3 days and, thereafter, females were collected and used for analysis. Flies were transferred to fresh vials and dead flies were counted every 2nd or 3rd day. The three genotypes were scored in parallel.
Paraquat treatment (analysis of resistance towards oxidative stress) 200 female flies per genotype were used to determine oxidative stress resistance. Flies were transferred to empty vials and starved for 3 h before paraquat (PQ, a redox-cycling agent) treatment was started. Paraquat (150 lL, 15 mM in 5% sucrose or 5% sucrose for control flies) was supplied on filter paper. Dead flies were counted every 6 to 12 h. The three genotypes were scored in parallel.

Starvation stress treatment
For starvation stress determinations 150 female flies per genotype were used. Flies were transferred into vials containing a filter paper soaked with 150 lL of water. Dead flies were counted every 6-12 h. The three genotypes were scored in parallel.

Statistics
Unpaired student's t-tests were used to compare samples derived from young and aged donors as well as empty vector (EV) control transfected and OE-TXNIP S2 cells. Welch's t-tests were used to calculate statistics for TRX activity and Log-Rank test for lifespan analysis. P values were calculated using GraphPad Prism 4 or the Sigma Plot 13 software package.

TXNIP is upregulated in various human tissues during aging
To analyze age-related changes in humans we isolated primary T cells from aged (> 55 years old) and young (20 -25 years old) individuals. Since accumulation of oxidative damage has been linked to aging we asked whether expression of TXNIP is altered during aging. We observed increased TXNIP expression levels in T cells of aged donors in comparison to T cells isolated from young donors (Fig. 1A,B and Fig. S1A). Analysis of other primary human cells of the hematopoietic system (e.g. hematopoietic progenitor cells [HPC] and monocytes) also showed enhanced TXNIP levels in aged individuals (Fig. 1D,E). Non-hematopoietic tissues such as liver, showed a marked tendency for upregulation of TXNIP levels in aged individuals (Fig. 1F). Similar effects have also been described in the brain cortices of aged rats [22]. Thus, these data suggest a general role for TXNIP in aging. Of note, increased expression of other a-arrestins during aging was not observed (Fig. S1B,C). Since TXNIP is a negative regulator of TRX-1 [28][29][30] we analyzed TRXactivity. We found that increased TXNIP expression correlated with reduced TRX activity in cells from older individuals (Fig. 1C). This suggests that increased TXNIP expression may increase the pro-oxidative intracellular status during aging. To further test this idea, Jurkat T cells were stably transfected with an inducible TXNIP shRNA. TXNIP knock-down ( Fig. 2A,B) improved resistance to oxidative stress induced by hydrogen peroxide (H 2 O 2 ), further substantiating a role for TXNIP in redox regulation (Fig. 2C).

Drosophila is an aging model comparable to humans
Since Drosophila melanogaster express a TXNIP homologue called VDUP1 and are a widely accepted aging model comparable to humans [25,[31][32][33] we used them to test whether TXNIP directly affects organismal aging. Importantly, as seen in humans, Drosophila TXNIP is upregulated during aging in Drosophila (Fig. 4A). To determine whether TXNIP is a pivotal regulator of aging, we investigated the role of Drosophila TXNIP in vitro and in vivo. We generated Drosophila Schneider-2 cells (S2) overexpressing Drosophila TXNIP (OE-TXNIP) (Fig. 3A,B). To elucidate the role of TXNIP in regulating redox equilibrium, we determined basal levels of reactive oxygen species (ROS). OE-TXNIP cells showed an increase of up to  68% in basal ROS levels compared to empty vector (EV) transfected control cells (Fig. 3C). In addition, TXNIP overexpression decreased TRX activity by 14% relative to EV transfected cells (Fig. 3D). Hence, TXNIP overexpression in S2 cells led to a shift to a pro-oxidative cellular status. Of note, overexpression of TXNIP did not result in changes in expression levels of Drosophila TRX homologues or other redoxrelated genes (Fig. S2). Impairment of the cellular redox balance has been shown to render cells more susceptible to oxidative stress [1,34]. Therefore, we tested resistance to oxidative challenge by assaying specific cell death after H 2 O 2 addition. Overexpression of TXNIP sensitized S2 cells towards oxidative stress (Fig. 3E). Increased ROS levels and an attenuated oxidative defense can lead to damage of macromolecules such as proteins, lipids and nucleic acids. Accumulation of DNA damage is known to be involved in aging [5,35,36]. An indicator for DNA double strand breaks and oxidative DNA damage is phosphorylation of histone H2A. To study the effect of TXNIP expression on the generation of oxidative DNA damage, we investigated cH2Av phosphorylation by immunofluorescence. While unchallenged S2 cells overexpressing TXNIP already showed an increase in cH2Av phosphorylation compared to EV transfected control cells (Fig. 3F, upper panel), H 2 O 2 treatment further increased DNA damage, especially in cells overexpressing TXNIP (Fig. 3F, middle panel). The ROS-scavengers N-acetyl cysteine (NAC) and Trolox both blocked cH2Av phosphorylation efficiently (Fig. 3F, lower panel). This clearly indicates that TXNIP-mediated DNA damage is ROS dependent.

TXNIP expression regulates lifespan of Drosophila
First, we analyzed endogenous TXNIP levels of young (12 days) and aged (52 days) WT flies (Fig. 4A). This increase in endogenous TXNIP expression also modulates the cellular redox balance, as shown by the reduction in endogenous TRX activity in aged (> 50 days) compared to young (12 days) WT flies (Fig. 4B). To analyze the role of TXNIP in stress resistance and lifespan, we generated TXNIP knockdown (TXNIP-RNAi) flies and flies overexpressing TXNIP (OE-TXNIP) (Fig. 4C). We excluded that TXNIP expression influenced body size and weight in these different genotypes, because these factors can impact lifespan in insects (Fig. S3). Since we had shown that TXNIP expression regulates TRX activity (Figs 1C, 3C and 4B), we investigated whether transgenic TXNIP levels also influence TRX activity in vivo.
Whole fly lysates were prepared and TRX activity was determined. We observed that enhanced TXNIP (levels OE-TXNIP) decreases TRX activity whereas low levels of TXNIP (TXNIP-RNAi) resulted in increased TRX activity in comparison to young wild type (WT) controls (Fig. 4D). Decreased TRX activity led to accumulation of oxidative damage in vitro (Fig. 3F). Therefore, we compared young (12 days) and old WT flies (> 52 days) to analyze if lower TRX activity also led to DNA damage in vivo. Western blot analysis revealed a strong increase in phosphorylated cH2Av in aged WT flies (Fig. 4E, left panel). Thereafter, DNA damage in young flies (12 days) was analyzed. Already young WT flies showed minor increased cH2Av phosphorylation indicating low but detectable levels of DNA damage. However, young flies overexpressing TXNIP showed a massive increase in cH2Av phosphorylation indicating severe DNA damage. Remarkably, nearly no cH2Av phosphorylation could be detected in young RNAi-TXNIP flies. Thus, downregulation of TXNIP results in a better protection to DNA damage even in comparison to normal (WT) expression levels of TXNIP (Fig. 4E, right panel). Next, we investigated the role of TXNIP expression in resistance to oxidative stress in vivo. We fed flies with the redox-cycling agent paraquat (PQ) and monitored survival under these oxidative stress conditions. OE-TXNIP flies showed significantly reduced survival upon PQ treatment (Fig. 4F). Interestingly, knock-down of TXNIP had only mild effects on oxidative stress resistance compared to WT (Fig. 4F). This is explained by downregulation of endogenous TXNIP expression upon exposure to oxidative stress (Fig. 4F, inset). Of note, survival under starvation conditions in flies (fed with water only) was not altered in OE-TXNIP or TXNIP-RNAi flies (Fig. S3B). Endogenous TXNIP is upregulated during aging in flies (Fig. 4A). Increased TXNIP expression led to downregulation of TRX activity (Fig. 4B,D), increased DNA damage (Fig. 4E) and impaired resistance (D) Thioredoxin (TRX) activity in lysates was normalized to EV. Statistical differences were determined by Welch's t-test (mean AE SEM, n = 3, *P < 0.05). (E) Cell death was induced by treatment with hydrogen peroxide (H 2 O 2) and analyzed by FACS. Statistical differences were determined by an unpaired student's t-test (mean AE SEM, n = 3, not significant (ns): P > 0.05, *P < 0.05 and **P < 0.01). (F) DNA damage was analyzed by immunostaining for phospho-cH2Av. cH2Av is shown in green and nuclei are shown in blue by Hoechst33342 staining. Scale bar, 10 lm. The bar diagram shows % of MFI of 10 nuclei and was normalized to EV. Statistical differences were determined by an unpaired student's t-test (mean AE SEM, n = 10, ***P < 0.001). towards oxidative stress (Fig. 4F). All these changes are important events that occur during aging [1,5,34]. Thus, TXNIP is a promising candidate to regulate lifespan in flies. To investigate whether TXNIP expression is directly involved in aging we analyzed the lifespan of the different fly strains. We found that overexpression of TXNIP significantly shortened median lifespan (Fig. 4G, Log-Rank-Test: P < 0.001), whereas depletion of TXNIP significantly extended median lifespan (Fig. 4G, Log-Rank-Test: P < 0.001). Lifespan shortening and the stress-resistance phenotypes noted above were observed in both sexes, but significant lifespan extension was observed only in females, the standard subject for Drosophila longevity assays [37,38]. Thus, the increased expression of TXNIP during aging correlates with a regulatory effect in which TXNIP levels negatively affect Drosophila melanogaster lifespan.

Discussion
In summary, we identified that the redox regulator TXNIP controls lifespan and resistance towards oxidative stress. We propose a novel mechanism in which TXNIP decreases TRX activity and, thereby, impairs cellular redox homeostasis during the aging process. Data from both humans and Drosophila melanogaster support this assumption. In agreement with the "free radical theory of aging" [1], our data from Drosophila show for the first time that the TXNIP-induced shift to a pro-oxidative cellular environment facilitates DNA damage (Figs 3,4). In accordance with the idea that oxidative stress and DNA damage are major factors influencing deterioration of cellular functions during aging [5,35,36,39], lowered expression of TXNIP improved cellular resistance to oxidative stress (Fig. 3) and extended lifespan (Fig. 4G). Our finding that TXNIP expression increases during aging is all the more interesting when one considers that increased expression of TXNIP is also observed in age-related diseases such as diabetes [40][41][42] and neuronal degeneration [43,44]. Aging of the population has major implications for health care resources and workforce productivity [45]. Therefore, improvements in healthy and productive aging are of major interest. Manipulation of TXNIP expression/activity could be a promising option to promote healthy aging. on the manuscript as well as L. Weingarten, D.

Author contributions
TO designed and performed experiments, analyzed the data, discussed results and wrote the paper. JB, SZ, AS, DR, LK, WW designed and performed experiments and analyzed data. BAE contributed to experimental design, analyzed data and edited the manuscript. KG and PHK initiated, designed, guided the research and edited the manuscript.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Fig. S1. Expression of arrestin family members is not changed upon aging. Fig. S2. Gene expression of Drosophila redox-related genes is not changed upon TXNIP overexpression. Fig. S3. TXNIP expression levels in female flies do not influence body size or weight and do not influence lifespan under starvation of the generated fly stains. Table S1. Summary of qPCR primer sequences.