Age and oxidative stress regulate Nrf2 homeostasis in human articular chondrocytes

Abstract

Objective:

The purpose of this study was to investigate the effect of age and oxidative stress on regulation of Nrf2 in young, old, and osteoarthritic human articular chondrocytes.

Design:

Levels of Nrf2 in primary human chondrocytes isolated from young, old, and OA donors were measured by immunoblotting, qPCR, and immunohistochemistry. Effects on levels of Nrf2, antioxidant proteins regulated by Nrf2, as well as p65, and the anabolic response to IGF-1 were evaluated after induction of oxidative stress with menadione, Nrf2 knockdown with siRNA, and/or Nrf2 activation with RTA-408.

Results:

Nrf2 protein levels were significantly lower in older adult chondrocytes (~0.59 fold; p= 0.034) and OA chondrocytes (~0.50 fold; p=0.016) compared to younger cells. Menadione significantly increased Nrf2 protein levels in young chondrocytes by just under 4-fold without changes in old chondrocytes. Nrf2 knockdown and activation differentially regulated levels of anti-oxidant proteins including Srx and NQO1. Nrf2 activation with RTA-408 also decreased basal p65 phosphorylation, increased aggrecan and type II collagen gene expression, and increased production of proteoglycans in OA chondrocytes treated with IGF-1.

Conclusions:

Targeted therapeutic strategies aimed at maintaining Nrf2 activity could be useful in maintaining chondrocyte homeostasis through maintenance of intracellular antioxidant function and redox balance.

Introduction

Age is a major risk factor for the development of OA, however, the exact molecular mechanisms that are altered in articular joints during aging that contribute to OA are not clear. The redox theory of aging hypothesizes that increasing age leads to excessive accumulation of reactive oxygen species (ROS) resulting in the disruption of cell signaling, immune response, macromolecule trafficking, and physiologic regulation [1, 2]. Chondrocytes obtained from older adults and from OA joints have elevated ROS production and reduced antioxidant capacity, which can contribute to cartilage degradation as well as chondrocyte cell death [3]. One important regulator of antioxidant gene expression is nuclear factor erythroid-2-related factor 2 (Nrf2), which belongs to a class of proteins that binds to the antioxidant response elements (ARE) in the regulatory segments of specific antioxidant genes [4, 5].

Nrf2 is conserved among the vertebrates and is crucial in the regulation of antioxidant defense, oxidant signaling, and detoxification of drugs (phase II antioxidant response) [6]. Nrf2 regulates the expression of several antioxidant proteins. These include superoxide dismutase 3, glutathione peroxidase, peroxiredoxins (Prx), thioredoxin reductase, sulfiredoxin (Srx), NAD(P)H quinone dehydrogenase 1 (NQO1), and heme oxygenase-1 (HO-1) [6–9] . Nrf2 is also involved in the 23s proteasome degradation pathway and reprogramming of the cell based on its metabolism [4]. The target genes of Nrf2 are directly responsible for maintaining the reductive state of the cell by enhancing the glutathione (GSH) biosynthetic enzymes and glucose-6-phosphate dehydrogenase (G6PD) that generates NADPH.

One of the most studied and well characterized proinflammatory transcription factors is NFκB. Nrf2 and NFκB signaling pathways interact to help maintain cellular redox balance and regulate the cellular responses to oxidative stress [10]. A decrease in Nrf2 expression can indirectly increase NFκB activity, more specifically levels of the NFκB subunit p65, causing an increase in cytokine production leading to inflammation [11]. NFκB can also control both Nrf2 transcription and activity [12], while a study by Sivandzade shows that inhibition of NFκB was due to increases in HO-1 , a Nrf2 target gene [13]. Conversely, NFκB binds to the binding sequences in the Nrf2 gene and promotes the transcription of Nrf2 [11]. Interactions between Nrf2 and NFκB have not been studied in chondrocytes but may provide a better understanding of the proinflammatory milieu seen in OA.

Studies in mice have revealed that germ-line deletion of Nrf2 results in more severe OA in the destabilization of the medial meniscus (DMM) model of OA [14] and more severe spontaneous OA in mice at 90 weeks of age [15]. The objective of the present study was to measure the levels of Nrf2 in human chondrocytes derived from young, old, and OA cartilage tissues and examine the effects of Nrf2 knockdown and activation on Nrf2 target genes, NFκB signaling, and chondrocyte anabolic activity.

Materials and Methods

Antibodies and reagents

Antibodies to Nrf2 (#62352), β-actin (#8226), and Prx1(#109506) were from Abcam. Anti-Srx1 (#14273–1) was from ProteinTech and anti-LDH from Fitzgerald (#20-LG22). Antibodies to HO-1 (#5061), NQO1 (#62262), thioredoxin interacting protein (TXNIP) (#14715), total p65 (#8242), phosphorylated-p65 (#3031), β-tubulin (#2146), and lamin B (#12586) were from Cell Signaling Technology. Secondary antibodies used were anti-rabbit IgG (#7074V) or anti-mouse IgG (#7076S) from Cell Signaling Technology. Primary antibodies were used at a 1:1000 dilution except for β-actin (1:2000), LDH (1:10,000) and lamin B (1:10,000) with secondary antibodies at 1:2000.

Human articular chondrocyte isolation and culture

Normal human articular chondrocytes were isolated from the talar tissue of young (n=46) and old (n=56) tissue donors. The human talar tissues were obtained through the Gift of Hope Organ and Tissue Donor Network (Itasca, IL) in collaboration with Rush University Medical Center (Chicago, IL). The donors ≤45 years old were defined as young while those ≥50 years old were defined as old to be consistent with the ages chosen for young and old in our previous studies [16, 17]. The osteoarthritic (OA) chondrocytes were isolated from the knee joint tissue from patients (n=51) that underwent knee arthroplasty. The OA knee joint tissue was obtained in collaboration with the Department of Orthopedic Surgery, University of North Carolina Medical Center (Chapel Hill, NC). The chondrocytes were isolated using an enzymatic method and cultured to confluency in monolayers as described previously [18]. In some experiments, isolated chondrocytes were treated with a redox cycling agent, menadione, to induce oxidative stress conditions as we have previously described [19–22]. Briefly, confluent cultures were serum-starved overnight to dampen the cell signaling induced by growth factors in the serum. Cells were then treated with 25 μM menadione for the indicated times and the cell lysates were collected and blotted for Nrf2 and p65. For pellet cultures, 4 × 10⁵ isolated cells were centrifuged for 5 min at 150g in 15 mL polystyrene tubes for each pellet. Dulbecco’s Modified Eagle Medium (DMEM) was changed every day without disruption to the pellet and pellets were re-centrifuged for 5 min at 150g after media changes to help maintain the pellet. After 3 days, pellet cultures were treated every 48hrs with 20nM RTA-408 and 100ng/ml IGF-1.

Immunoblotting

Total cell extracts were collected by lysing confluent chondrocyte monolayers with cell lysis buffer (Cell Signaling Technology) supplemented with 1 mM phenylmethylsulfonylfluoride (PMSF) and 1X protease inhibitor cocktail. The supernatant from each sample was stored at −20°C until immunoblotting was performed. Protein content in cell lysates was determined using the Pierce^™ BCA protein assay kit (ThermoFisher Scientific) and equal protein was loaded onto each well in an SDS-polyacrylamide gel. For cytosolic and nuclear fractions, 15ug were loaded for per sample in each lane. The proteins were separated, transferred to a nitrocellulose membrane, and probed using antibodies against the target proteins. Membranes were incubated with a horse radish peroxide (HRP) conjugated secondary antibody (1:2,000) for 1 hour at room temperature. The chemiluminescence was captured using an Azure Biosystems imager and bands were analyzed with image J software. For basal protein levels, β-actin was used as loading control. Although β-actin can be a good house-keeping protein in most conditions, it is susceptible to oxidative modifications making it unsuitable for redox experiments. Hence, β-tubulin was used as the loading control in those experiments.

Nrf2 and NFκB, both being transcription factors, shuttle in and out of the nucleus depending on their activation status. In select experiments, we used NE-PER Nuclear and cytoplasmic Extraction Reagents (ThermoFisher Scientific) to separate the nuclear and cytoplasmic (non-nuclear) proteins prior to immunoblotting. The cell fractionation procedure was performed according to the manufacturers’ protocol. Lamin B was used as a nuclear marker and LDH was used as a cytoplasmic marker.

Nrf2 knockdown and activation

To identify the genes that were regulated by Nrf2 in cultured chondrocytes, we either knocked down Nrf2 using small interfering RNA against Nrf2 (Dharmacon, human NFE2L2 Smart Pool) or non-specific control siRNA (#D-001810–0X) or pharmacologically activated Nrf2 using RTA-408 (MedChem Express). For siRNA experiments, normal chondrocytes were plated at 1×10^6 cells/ mL and the nucleofection method (Lonza) was used to transfect the siRNA (1μM and 2μM) into the chondrocytes. Cells were harvested for nucleofection using collagnease and pronase, centrifuged, and resuspended in Amaxa primary cell nucleofection reagent. The siRNA nucleofection protocol was modified from the plasmid nucleofection method we previously described [23]. For the activation of Nrf2, chondrocyte monolayers were treated with RTA-408 at 10nM or 20nM for 48 hours and the cell extracts were collected for immunoblotting. For gene expression studies, normal and OA chondrocytes were pre-treated with or without RTA-408 for 48 hours followed by a media change and then treated with or without 100 ng/mL IGF-1 (recombinant human from Austral Biologicals #GF050–8) for an additional 24 hours.

Quantitative real-time polymerase chain reaction

The gene expression of Nrf2, aggrecan, and Col2a1 was measured using RNA isolated from chondrocytes obtained from the young, old, and OA donors, as described [24]. RNA was isolated using the RNeasy mini kit (Qiagen). The reverse transcription of the RNA to cDNA was achieved using ImProm-II^™ Reverse Transcriptase system (Promega). QuantStudio^™ 6 Flex machine (Applied Biosystems) was used for the quantitation of gene expression of Nrf2, aggrecan, and Col2a1 and normalized to TBP. Primers are provided in Supplementary Table 1.

Proteoglycan assay

Cell pellets from OA chondrocytes (ages 65 to 72, n=3) were digested overnight in a solution containing 0.05 M phosphate buffered saline (PBS), pH 6.5, 5 mM cysteine, 5 mM EDTA and 125 μg/mL papain. Proteoglycan concentrations were assessed by using reagents from a glycosaminoglycan (GAG) Assay Kit (Chondrex Inc).

Immunohistochemistry

Cartilage explants from normal (ages 35 to 48, n=5), old (ages 55 to 65, n=5), and OA (ages 63 to 75, n=4) were fixed in formalin for 48 hours, decalcified in EDTA, embedded in paraffin, and sectioned coronally at a thickness of 5 microns. Heat induced epitope retrieval was performed using sodium citrate buffer at 60 °C for 1 hour. Immunohistochemistry was performed with antibodies (diluted in 1% bovine serum albumin in Tris-buffered saline) for Nrf2 (1:250) at room temperature overnight. Sections were then incubated with a biotinylated secondary antibody (1:1000) for 1 hour at room temperature. Chromogens were detected with a VECTASTAIN Elite ABC HRP Kit (Peroxidase, Standard) Kit and 3,3-diaminobenzidine (DAB). Sections were counterstained with hematoxylin.

Statistical analysis

The data were analyzed using GraphPad Prism Version 7. All the experiments were performed in at least three independent donors. Exact independent sample numbers are shown in the figures for each experiment. For comparisons between the menadione treatment times in the young and the old donors, a two-way ANOVA was used with appropriate post-hoc corrections. For siNrf2 and RTA-408 experiments, a one-way ANOVA with post-hoc corrections or repeated measures ANOVA (where indicated in the figure legends) were performed to determine the statistical significance. Results shown are mean ± standard deviation. A p-value <0.05 was considered significant.

Results

Nrf2 is decreased in human chondrocytes with respect to age and OA

Articular chondrocytes isolated from normal human tali and osteoarthritic knee joints were used to determine the basal levels of Nrf2 in young, old, and OA populations. In prior work [20], we directly compared normal knee to normal talar cartilage and found that the IGF-1 response and redox signaling does not differ between chondrocytes isolated from the two joints, supporting the use of normal tali to compare to OA knee for the present studies. Total cell lysates were immunoblotted to measure the basal levels of Nrf2 protein. A significant decrease in basal Nrf2 protein levels was observed in chondrocytes derived from old (66.5± 1.0 yrs, n=14) and OA (62.7± 3.4 yrs, n=12) donors when compared to younger donors (28.8± 2.1 yrs, n=13) (Fig. 1A). To investigate whether the differences in Nrf2 protein levels were due to differences in transcription, we performed qRT-PCR on chondrocyte RNA samples from young (ages; 30 to 45, n=5), old (ages: 55 to 71, n=9), and OA (ages: 45 to 71, n=9) donors. There was no significant difference between young and old donors, while a significant decrease in Nrf2 expression was seen in OA chondrocytes compared to young (Fig. 1B). We next examined Nrf2 levels using immunohistochemistry in cartilage from normal (ages 35 to 45, n=5), old (ages 55 to 65, n=5), and OA (ages 63 to 75, n=4) donors. We didn’t obverse a significant difference between young and old donors, but there was a significant decrease in Nrf2 levels when comparing OA to both young and old donors (Fig. 1D, Supplementary Fig. S1).

Figure 1. Effects of age and OA on Nrf2 levels in human chondrocytes.

(A) Lysates from young (ages 17 to 38, n=13), old (ages 58 to 72, n=14), and OA (ages 37 to 81, n=12) chondrocytes were immunoblotted with an antibody to Nrf2 and then blots were stripped and reprobed for β-actin as a control. (B) Immunoblot results of Nrf2 normalized to β-actin (mean ± SD). (C) RNA was collected from human articular chondrocytes isolated from young (ages 30 to 45, n=5) and old (ages 55 to 71, n=9) tissue donors, and from OA (ages 45 to 71, n=9) tissue. Nrf2 gene expression was analyzed using qRT-PCR and was normalized to TBP expression. (D) Cells positively stained for Nrf2 in cartilage measured by immunohistochemistry. Normal young (ages 35 to 48, n=5), old (ages 55 to 65, n=5), and OA (ages 63 to 75, n=4). Immunohistochemical images are shown in Supplementary Figure S1.

Chondrocytes from older adults have a lower Nrf2 response to an acute ROS insult

To evaluate the functional significance of decreased Nrf2 in articular cartilage from older donors, we treated the chondrocytes from both young and old donors with menadione, a redox cycling agent that generates H₂O₂ in the mitochondria in a dose-dependent manner. Our previous studies have shown that 25 μM menadione generates sufficient H₂O₂ to induce acute oxidative stress that alters chondrocyte cell signaling [19, 20, 25]. Increases in the level of cellular ROS is a well-known trigger for Nrf2 stabilization and nuclear translocation [6]. Treatment of chondrocytes isolated from young or old donors with 25 μM menadione for different time points showed that Nrf2 protein levels significantly increased in young donors at 10 mins (3.9-fold) and 30 mins (3.75 fold) and were maintained out to 60 mins (2.7-fold) while no significant change was noted in cells from the older donors (Fig. 2A–C).

Figure 2. Nrf2 levels in young and old human chondrocytes in response to menadione.

(A) Lysates were prepared from young (17 to 38 yrs of age; n=3) and old (58 to 72 yrs of age; n=3) chondrocytes after treatment with 25 μM menadione to induce ROS production for the indicated times and used for immunoblotting with antibodies to Nrf2 and β-tubulin as a control. (B) Nrf2 levels normalized to β-tubulin for young and old individually. (C) Relative change in Nrf2 levels from a baseline (0 mins) set to 1 in response to menadione in young and old donors. P values indicate differences between young and old at the indicated time points. (D) Nuclear and cytosolic fractioning of Nrf2 from lysates prepared from young (ages 25 to 38, n=3) and older adults (ages 58 to 70, n=3). LDH was used as a cytosolic marker and lamin B for the nucleus. (E) The nuclear Nrf2 was normalized to lamin B and the cytosolic Nrf2 was normalized to LDH. The relative intensity of Nrf2 levels was calculated by normalizing all the time points to baseline (0 mins) set to 1 in each age group. Repeated measures ANOVA was used for statistical analysis. Results shown in scatter graphs are mean±SD.

To understand the effect of age on Nrf2 nuclear translocation after acute ROS insult, we fractionated the nucleus and the cytosol in cell lysates and used immunoblotting to measure the Nrf2 levels after menadione treatment. After 15 mins of menadione treatment, Nrf2 translocation to the nucleus more than doubled (2.8-fold) and the cytoplasmic fraction increased by 11-fold in young donors (Fig. 2D). In contrast, the older donors showed no significant increase in nuclear Nrf2 levels while an increase (3-fold) in cytoplasmic Nrf2 was observed at 15 mins of exposure. At 30 mins of treatment, Nrf2 from the young chondrocytes showed maximal nuclear localization of 4-fold and a corresponding slight decline (1.2-fold) in the cytoplasmic Nrf2 pool. However, such a change in the nuclear to cytoplasmic distribution was not observed in the older chondrocytes (Fig. 2E). Taken together, the data suggest that Nrf2 levels in young adult chondrocytes are sensitive to an acute ROS insult while the older adult chondrocytes have a blunted response that when combined with lower basal Nrf2 would reduce their ability to counteract the increase in ROS due to reduced capacity of older cells to translocate Nrf2 to the nucleus.

Nrf2 controls HO-1, Srx, and NQO1 levels in human articular chondrocytes

Since the chondrocytes from older adult donors and OA tissue showed lower Nrf2 levels and older chondrocytes had a dramatic decrease in Nrf2 induction after a ROS trigger, we tested the protein profile of Nrf2 targets by immunoblotting to determine which of the potential Nrf2 target genes were regulated by Nrf2 in chondrocytes. After siRNA mediated knockdown of Nrf2 in normal talar chondrocytes, a significant decrease in basal protein levels of Srx, HO-1, and NQO1 was noted without changes in Prx1 or TXNIP levels (Fig. 3). RTA-408 activates Nrf2 by inhibiting the binding of Nrf2 to Keap1 which prevents Nrf2 degradation resulting in an increase in Nrf2 levels. We tested the effect of RTA-408 on Nrf2 knockdown cells. Results show that Nrf2 knockdown cells had no change in Nrf2, NQO1, or Srx1 levels when treated with RTA-408, indicating a lack of a Nrf-2-independent effect by RTA-408 (Fig. 3B). Treatment of normal (ages: 42 to 48) and OA (ages: 45 to 70) chondrocytes with 20 nM RTA-408 led to a significant increase in Nrf2, Srx, and NQO1 levels in both, while Prx1 levels were unchanged and TXNIP protein levels were decreased in OA but not normal chondrocytes (Fig. 4A and B). We also examined effects of Nrf2 activation on phosphophorylated-p65 levels since, depending on the cell type, NFκB and Nrf2 can negatively regulate each other [26]. We noted activation of Nrf2 with 20nM RTA-408 in OA, but not normal chondrocytes, resulted in a decrease in the active NFκB subunit phosphorylated-p65 (Fig. 4B).

Figure 3. Effects of Nrf2 knockdown on Nrf2-regulated proteins in human chondrocytes.

(A) siRNA mediated knockdown of Nrf2 in normal chondrocytes (ages 32 to 45, n=4) was performed by nucleofection yielding an ~89% depletion of Nrf2 compared to control siRNA. Nrf2 and its target protein levels were measured by immunoblotting and β-tubulin served as the loading control. The band intensity was normalized to β-tubulin. Results shown are mean ± SD. (B) siRNA knockdown of Nrf2 was performed in normal chondrocytes (ages 42 to 55, n=4) and then the cells were treated with RTA-408 to test for the specificity of Nrf2 activation. After nucleofection, the transduced cells were treated with RTA-408 for 48 hours. Repeated measures ANOVA was used for statistical analysis.

Figure 4. Effects of Nrf2 activation on Nrf2-regulated proteins in human chondrocytes.

(A-B) Nrf2 was activated using RTA-408 at 10nM and 20nM (A, normal chondrocytes, ages 42 to 48, n=3) and (B, OA chondrocytes, ages 45 to 70, n=4). Cells were treated in monolayers for 48 hrs and cellular contents were extracted for immunoblotting. The band intensity of antioxidant proteins was normalized to β-tubulin and phosphorylated (p)-p65 subunit was normalized to total (t)-p65. (C) OA chondrocytes (ages 62–71, n=3) and normal chondrocytes (ages 48–59, n=3) were treated with RTA-408 for 48 hours. Nuclear and cytoplasmic fractions were isolated to analyze total (t) and phosphorylated (p) p65 after treatment. The phospho-p65 subunit was normalized to total (t)-p65.

p65 levels increased with age and under oxidative stress conditions

Because we noted that Nrf2 levels were decreased with age and this could potentially affect p65, we examined levels of p65 and expression of RelA, the gene that encodes p65, in chondrocytes from young and older adults. Both p65 protein and RelA expression were higher in chondrocytes derived from older adults (Fig. 5A and B). We also examined nuclear translocation of phosphorylated-p65 under oxidative stress conditions using menadione. Treatment of chondrocytes with menadione led to a significant increase in nuclear phosphorylated-p65 at 30 and 60 mins (ages 39 to 63, n=4) (Fig. 5C) demonstrating that NFκB can be activated by acute oxidative stress in chondrocytes isolated from young and older adult donors. Phosphorylated-p65 levels were higher in donors with osteoarthritis in both nuclear and cytoplasmic fractions, and treating with RTA-408 decreased the phosphorylation of p65 in both normal and OA donors (Fig. 4C).

Figure 5. Effects of age and oxidative stress on human chondrocyte p65.

(A) Total cell lysates isolated from young (ages 25 to 50, n=4), older adult (ages 60 to 77, n=5), and OA (ages 57–78 n=5) donor samples were immunoblotted for the p65 subunit of NFκB. The band intensity of total p65 was normalized to β-actin, and phosphorylated (p)-p65 normalized to total (t) p65 levels. (B) The cartilage tissue from young, old, and OA donors was used to extract RNA for gene expression profiling of the p65 subunit (RelA) of NFκB in young (ages 35 to 49, n=7), old (ages 52 to 75, n=5) and OA (ages 61 to 81, n=8) chondrocytes. The qRT-PCR data was normalized to TBP. (C) 25 μM menadione was used to induce oxidant stress in normal chondrocytes (ages 39 to 63, n=4). Lysates from the cytosol and nucleus were fractionated after menadione treatment at the times indicated and the relative band intensity was obtained by normalizing to LDH for the cytosolic fraction and lamin B for the nucleus.

Nrf2 activation increases aggrecan and type II collagen expression in normal and OA chondrocytes treated with IGF-1

Prior studies showed that compared to normal, OA chondrocytes have reduced anabolic activity including the anabolic response to IGF-1 [17, 27]. Given that RTA-408 increased the levels of chondrocyte Nrf2 and the antioxidant proteins Srx and NQO1 and, in OA chondrocytes, reduced the levels of the antioxidant inhibitor TXNIP, we measured the effect of RTA-408 on expression of Col2a1 and aggrecan in both normal (ages 35–59) and OA (ages 52–70) human chondrocytes. We found that RTA-408 alone had little to no effect on the expression of aggrecan or Col2a1 in either normal or OA chondrocytes (Fig. 6A, B). When normal chondrocytes were treated with IGF-1 alone, there was a ~3-fold increase in both aggrecan and Col2a1 expression, while OA chondrocytes treated with IGF-1 had no change in either aggrecan or Col2a1 gene expression. Pre-treating with RTA-408 prior to IGF-1 stimulation did not change the response to IGF-1 in normal chondrocytes, while in OA chondrocytes we observed a ~3.5-fold and ~4-fold increase in aggrecan and Col2a1, respectively, demonstrating that activating Nrf2 was able to improve the IGF-1 response in OA cells (Fig. 6).

Figure 6. Effects of Nrf2 activation on the IGF-1 response in normal and OA chondrocytes.

(A) Normal (ages 35 to 59, n=3) and (B) OA (ages 52 to 70, n=3) chondrocytes were treated with vehicle as control or RTA-408 (10nM and 20nM) for 48hrs and then media was changed, and cells were treated with IGF-1 for 24hrs in the fresh media. RNA was extracted and used for qPCR to measure Aggrecan and Col2a1 expression. Transcript levels were normalized to the reference gene TBP. Abundance and relative fold changes in transcript gene expression were quantified using the 2 ^−ΔΔCt method relative to the control. Results shown in scatter graphs are mean ± SD.

To further test the effects of RTA-408, OA chondrocytes were placed in 3D culture. When treating OA chondrocytes with IGF-1 alone there was no change in either aggrecan or Col2a1 gene expression levels. Pretreating with RTA-408 and then adding IGF-1 increased gene expression levels ~9 fold for aggrecan and ~5 fold for Col2a1 (Fig. 7A). To further understand the effects that RTA-408 and IGF-1 have on OA chondrocytes in pellet cultures we measured GAG levels as a readout for proteoglycan synthesis after 21 days of culture. Consistent with the aggrecan expression results, GAG production was significantly higher in chondrocytes treated with both RTA-408 and IGF-1 in 3D culture (Fig. 7B).

Figure 7. Effects of Nrf2 activation in 3D chondrocyte pellet cultures.

(A) OA (ages 51 to 61, n=4) chondrocytes were treated with vehicle as control or RTA-408 (20nM) for 48hrs followed by a media change and the addition of 100ng/ml IGF-1 for 24hrs. Col2a1 and aggrecan (Acan) levels were measured using q-PCR. (B) Glycosaminoglycan (GAG) production by OA chondrocytes (ages 63–77, n=3), treated every other day with 20 nM RTA-408 with or without 100ng/ml IGF-1, was measured using GAG assay after 21 days in culture. Results were analyzed using repeated measures ANOVA and graphs show mean ± SD.

Discussion

Studies have shown that Nrf2 contributes to controlling redox regulation by maintaining oxidant levels through regulating NQO1, Srx1, Prx1 and other antioxidants [6]. The role of Nrf2 in the oxidative stress response has been studied in many disease states, such as cancer [28, 29], neurodegeneration [13] and pulmonary disease [11, 30], but its role in osteoarthritis has yet to be fully deciphered. Nrf2 has been shown to regulate chitinase 3 like 1 causing a suppressed immune response to TNFα, IL-1β and IL-6, which led to decreased MMP13 expression in a post-traumatic osteoarthritis model [42]. In this study, we implicate age and OA as determinants of Nrf2 homeostasis in human chondrocytes and demonstrate that Nrf2 activation is feasible in both normal and OA cells. In OA cells, Nrf2 activation helped to overcome the resistance to IGF-1 stimulation. Using immunohistochemistry, Nrf2 protein levels were noted to be increased in human OA synovium and in the synovium of rats with surgically-induced OA [23]. We observed a significant decrease in Nrf2 protein levels by immunoblot in chondrocytes isolated from older adults and from OA cartilage compared to young chondrocytes which could alter chondrocyte redox balance. There was no observed difference in Nrf2 RNA levels between chondrocytes from young and old donors or in protein levels detected by immunohistochemistry. Nrf2 RNA levels and protein levels by immunohistochemistry were decreased in OA chondrocytes, suggesting that the mechanism for a reduction in Nrf2 protein levels may be different in aged chondrocytes relative to OA cells.

Both aging and OA are associated with oxidative stress [3] and it might be expected that this would increase Nrf2 levels, as noted above in synovium, through oxidation of the Nrf2 inhibitor Keap1 [31] as a cellular response to combat excessive ROS. However, we found that compared to chondrocytes from young donors, cells from older adults had a reduced Nrf2 response to menadione, used to increase ROS levels, suggesting that chondrocytes from older donors have a reduced ability to respond to oxidative stress conditions through Nrf2, further promoting redox imbalance. It is possible that an overall age-dependent increase in ROS might have saturated Nrf2 sensitivity in older chondrocytes through oxidation of the Nrf2 binding protein Keap1.

Antioxidant enzymes help to control redox levels to ensure that the response to ROS is sufficient. One such example is Srx, which is transcribed in response to Nrf2 nuclear translocation and binding of Nrf2 to the ARE element in the Srx promoter. Srx reduces hyperoxidized thiols in peroxiredoxins that are in the sulfinic acid form in an ATP- and GSH-dependent manner [8, 32]. We noted that knockdown of chondrocyte Nrf2 significantly decreased levels of Srx, as well as other antioxidant proteins including HO-1 and NQO1, while activation of Nrf2 with RTA-408 had the opposite effect, demonstrating that Nrf2 regulates the expression of these genes in human chondrocytes. As shown by Kubo et. al [14], Sox9 expression was decreased in the absence of Nrf2, but Keap1 knockdown increased the expression of Sox9, showing the relationship between Nrf2 activation and Sox9 expression. In OA chondrocytes, we noted activation of Nrf2 promoted the anabolic response to IGF-1 measured as Col2a1 and aggrecan expression demonstrating a potential beneficial effect of increasing Nrf2-mediated antioxidant gene expression in OA or perhaps through Nrf2 activation of Sox9.

NFκB can be activated in response to ROS [33] and so regulation of the redox status of the cell by Nrf2 may influence NFκB activity. We observed a basal increase in protein and RNA levels of the NFκB subunit p65/RelA in older and OA chondrocytes compared to younger chondrocytes and other studies have noted increased activity of NFκB family members in OA chondrocytes [34]. NFκB upregulates transcription of target genes including inflammatory cytokines, chemokines, mediators of apoptosis, and matrix degrading enzymes that promote OA [29, 35, 36]. We found that activation of Nrf2 with RTA-408 decreased the phosphorylation of p65 in OA chondrocytes in association with the increase in the antioxidant proteins Srx and NQO1, as well as a decrease in the antioxidant inhibitor TXNIP. This shows that Nrf2 activation in OA chondrocytes could potentially reduce the catabolic pathways associated with p65 activation. This finding is supported by a recent study using the compound Al-1, which like RTA-408 activates Nrf2 through inhibition of Keap1, and which was found to reduce expression of chondrocyte IL-6 in association with reduction in the NFκB target gene regulator IκB-ζ [37].

A limitation of the present study was that we examined Nrf2 function in human cells ex vivo. Prior in vivo studies found that deletion of Nrf2 in mice was associated with cartilage thinning and an increase in cartilage lesions with age [14] and young Nrf2 knockout mice had more cartilage damage when OA was induced using the DMM model [38]. We used human primary chondrocytes isolated from both normal and OA cartilage to examine Nrf2 functions that could not be done in vivo to provide additional support for the importance of Nrf2 in maintaining chondrocyte homeostasis.

Nrf2 activators like RTA-408, have the ability to mimic Nrf2 activation by covalent modification of cysteine groups in Keap1 to prevent Nrf2 ubiquitination and subsequent degradation [39]. Even though there are many genes associated with regulation of ROS metabolism by Nrf2 [40, 41], the broad mechanisms by which Nrf2 controls oxidant levels in chondrocytes have yet to be fully deciphered. Targeted therapeutic strategies aimed at maintaining Nrf2 activity could be useful in maintaining chondrocyte homeostasis through maintenance of intracellular antioxidant function and redox balance. Generally, the regulation of Nrf2 in chondrocytes could open new therapeutic opportunities for the treatment of osteoarthritis.