Klotho is an antiaging hormone present in the kidney that extends the lifespan, regulates kidney function, and modulates cellular responses to oxidative stress. We investigated whether Klotho levels and signaling modulate inflammation in diabetic kidneys.
Renal Klotho expression was determined by quantitative real-time PCR and immunoblot analysis. Primary mouse tubular epithelial cells were treated with methylglyoxalated albumin, and Klotho expression and inflammatory cytokines were measured. Nuclear factor (NF)-κB activation was assessed by treating human embryonic kidney (HEK) 293 and HK-2 cells with tumor necrosis factor (TNF)-α in the presence or absence of Klotho, followed by immunoblot analysis to evaluate inhibitor of κB (IκB)α degradation, IκB kinase (IKK) and p38 activation, RelA nuclear translocation, and phosphorylation. A chromatin immunoprecipitation assay was performed to analyze the effects of Klotho signaling on interleukin-8 and monocyte chemoattractant protein-1 promoter recruitment of RelA and RelA serine (Ser)536.
Renal Klotho mRNA and protein were significantly decreased in db/db mice, and a similar decline was observed in the primary cultures of mouse tubule epithelial cells treated with methylglyoxal-modified albumin. The exogenous addition of soluble Klotho or overexpression of membranous Klotho in tissue culture suppressed NF-κB activation and subsequent production of inflammatory cytokines in response to TNF-α stimulation. Klotho specifically inhibited RelA Ser536 phosphorylation as well as promoter DNA binding of this phosphorylated form of RelA without affecting IKK-mediated IκBα degradation, total RelA nuclear translocation, and total RelA DNA binding.
These findings suggest that Klotho serves as an anti-inflammatory modulator, negatively regulating the production of NF-κB–linked inflammatory proteins via a mechanism that involves phosphorylation of Ser536 in the transactivation domain of RelA.
It has long been recognized that diabetes accelerates aging, particularly in the subpopulation of diabetic subjects who are at risk for developing complications (1). Numerous mechanisms have been proposed, including increased production of advanced glycation end products (AGEs), increased oxidative stress, DNA damage, and enhanced inflammation; it is noteworthy that all of these mechanisms have been implicated in the pathogenesis of diabetes complications. Tubular epithelium in the kidneys from type 2 diabetic patients with demonstrated nephropathy display accelerated senescence, characterized by decreased telomere length and an increased expression of senescence markers (2).
The recent characterization of the Klotho protein as an antiaging hormone that modulates the expression level of antioxidant enzymes (3,4), as well as its high expression level in the kidney (5–7), suggest that Klotho plays a role in accelerated aging and cellular senescence observed in diabetes. Klotho overexpression extends the mouse lifespan by 20–30% (8). More striking, Klotho-deficient mice exhibit multiple age-related phenotypes and succumb to early, premature death (7,9). Klotho is predominantly expressed in the brain and kidney of normal subjects, and a significant decline in Klotho gene and protein expression has been reported in kidneys of patients with chronic renal failure (10). Klotho expression is significantly suppressed after the induction of renal ischemia-reperfusion injury, whereas Klotho overexpression prevented the development of acute renal failure (11). Also noteworthy, Klotho overexpression suppressed glomerulonephritis-induced accelerated cellular senescence and apoptosis and preserved renal function (12). Despite these observations, the role of Klotho in diabetes remains unexplored, even though accelerated aging is associated with this disease.
We investigated potential links between Klotho expression and diabetes-induced inflammation. Our data show that Klotho suppresses nuclear factor (NF)-κB activation and the subsequent production of inflammatory cytokines in response to tumor necrosis factor (TNF)-α stimulation in kidney cells, including primary cultures of mouse tubular epithelium, HK-2, and human embryonic kidney (HEK) 293 cells. We explored potential mechanism(s) for this inhibition and identified a novel and specific site of inhibition. Klotho inhibited p38 kinase and specifically blocked RelA serine (Ser)536 phosphorylation and its subsequent recruitment to NF-κB–dependent promoters of multiple cytokines, without affecting inhibitor of κB (IκB)α degradation or total RelA nuclear translocation and DNA binding. These findings indicate that Klotho serves as an anti-inflammatory modulator, regulating the production of NF-κB–linked inflammatory cytokines, chemokines, and growth factors via a noncanonical NF-κB activation pathway involving RelA phosphorylation in the transactivation domain (13–15). Our observations that Klotho can modulate NF-κB activation and inhibit the production of diabetes-induced inflammatory cytokines suggest that Klotho exerts a renoprotective effect by increasing the resistance to oxidative stress and inhibiting inflammatory cytokine/chemokine cascades induced by NF-κB activation. Our observations further suggest that Klotho is a potential therapeutic target linking oxidative stress to inflammation in type 2 diabetes.
RESEARCH DESIGN AND METHODS
Animal and surgical protocols.
Male Leprdb (db/db) and control mice with the same genetic background (C57BLKS/J) were purchased from The Jackson Laboratories (Bar Harbor, ME) and housed in the University of Texas Medical Branch Animal Resource Center in a room with a 12-h light cycle and with free access to standard diet and water. The animals were cared for in accordance with the University of Texas Medical Branch Institutional Animal Care and Use Committee policies and the Public Health Service Guide for the Care and Use of Laboratory Animals. At 20 weeks of age, the mice were anesthetized (70/10 mg/kg i.p. ketamine/xylazine), anticoagulated (5 units heparin), then exsanguinated prior to rapid aortic perfusion with ice-cold PBS to quickly rinse kidneys free of blood and to deliver a protease inhibitor cocktail (P8340; Sigma), phosphatase (1 mmol/L orthovanadate and 30 mmol/L sodium fluoride) inhibitors, and dithiothreitol (0.5 mmol/L). Both kidneys were removed, decapsulated, flash frozen in liquid nitrogen, then stored at −80°C until they were processed for protein or RNA extraction. A coronal section through the midline of one kidney at the level of the renal pelvis was fixed in 4% paraformaldehyde for 24 h, transferred to Hanks’ balanced salt solution, and stored at 4°C until it was processed for immunohistochemistry.
Isolation and primary culture of renal tubular epithelial cells.
Renal cortical tubular epithelial cells were isolated and cultured using published techniques with modifications (16). In brief, 12- to 16-week-old C57BL/6 male mice were used to prepare primary renal proximal tubule cell cultures. Kidneys were sliced into coronal sections, and the renal cortex was dissected from the medulla, minced, and then washed three times in ice-cold Dulbecco’s modified Eagle’s medium (DMEM)/F12 media containing 0.1% BSA, followed by enzymatic digestion using 1% Worthington collagenase type II and 0.25% soybean trypsin inhibitor. Cell suspensions were passed through 200 mesh filters followed by 325 mesh filters then resuspended in 45% Percoll and centrifuged at 26,891g for 15 min at 4°C. Proximal tubule cells were sedimented to a layer immediately above the erythrocyte pellet. Proximal tubule cells were removed, centrifuged, washed to remove the remaining Percoll, and then resuspended in DMEM/F-12 containing 50 units/mL penicillin, 50 µg/mL streptomycin, 10 ng/mL epidermal growth factor, 0.5 μmol/L hydrocortisone, 0.87 μmol/L bovine insulin, 50 μmol/L prostaglandin E1, 50 nmol/L sodium selenite, 50 µg/mL human transferrin, and 5 pmol/L 3,3′,5-triiodo-l-thyronine. Cells were plated on Matrigel-coated cover slips, or plastic cell-culture dishes coated with Matrigel, and maintained in an incubator at 37°C in 5% CO2. Cultures were left undisturbed for 48 h, after which culture media was replaced every 2 days until cells achieved confluence. For all experiments, cells were used within five passages, as described (16).
Human kidney cortex proximal tubular cells, HK-2 (American Type Culture Collection), were grown and maintained in the recommended medium, DMEM/Ham’s F12 (50/50) supplemented with 10% FBS, l-glutamine, insulin, transferrin, sodium selenite, epidermal growth factor (2.5 ng/mL), and pituitary extract (1.5 μg/mL) in a humidified atmosphere of 5% CO2. HEK epithelial cells, (HEK293) were cultured in MEM media with 10% FBS, nonessential amino acids, sodium pyruvate, and antibiotics in a humidified atmosphere of 5% CO2.
Transient transfection and luciferase activity assay.
Transient transfection using lipofectamine PLUS reagent (Invitrogen) was performed in triplicate plates according to the manufacturer’s instructions. Cells were plated into six-well plates and transfected with 2 µg of construct with a NF-κB–LUC reporter gene. A total of 48 h after transfection, cells were pretreated with different concentrations of Klotho (50–200 pmol/L) for 45 min prior to TNF-α (20 ng/mL) stimulation for 6 h. Cells were lysed, and luciferase activity was measured using a luminometer. Values were normalized to untreated empty vector–transfected cells and expressed as fold change over control.
Preparation of methylglyoxal-modified serum albumin.
Human serum albumin (HSA) minimally modified by methylglyoxal was prepared by incubation of the protein (100 μmol/L) in sodium phosphate buffer (100 mmol/L [pH 7.4] and 37°C) with 500 μmol/L methylglyoxal for 24 h, followed by dialysis of the modified protein against ammonium bicarbonate buffer (30 mmol/L [pH 7.9] and 4°C) for 24 h, and then the dialysis buffer was changed to PBS for another 24 h. The modified albumin was sterilized by filtration (0.22 µm) before aliquoting and storage at −80°C. Unmodified protein was processed similarly for control experiments by excluding the 500 μmol/L methylglyoxal. All reagents used in the preparation of methylglyoxal-modified human serum albumin (MG-HSA) were endotoxin free. The extent of modification was determined by amino acid analysis, and <15% modification was used in this study, as reported earlier (17). Negligible levels of endotoxin were detected in methyglyoxal-modified human serum albumin (MG-HSA) (0.0019 endotoxin units [EU]/mL), and control HSA (0.0029 EU/mL) was determined using a commercially available kit (GenScript, Piscataway, NJ).
Multiple cytokines and chemokines were measured on aliquots of culture media or cell extracts collected after MG-HSA or TNF treatment using the Bio-Plex system, which is a multiplex bead-based assay used with the Luminex xMAP technology, according to the manufacturer’s instructions (Bio-Rad Laboratories, Hercules, CA). Eight-point standard curves were performed for each cytokine using the same Luminex bead technology.
Preparation of subcellular extracts.
Cells were harvested in PBS, centrifuged, and the pellets resuspended sequentially in low-salt, sucrose, and high-salt solutions to obtain cytosolic and highly purified nuclear extracts as previously described (18). Protein concentrations were measured by Coomassie dye binding (protein reagent; Bio-Rad Laboratories). Efficiency of separation of cytosolic and nuclear proteins was assayed by Western blotting and probing both samples for cytosolic (β-tubulin and glyceraldehyde-3-phosphate dehydrogenase [GAPDH]) and nuclear (lamin B) proteins.
Proteins were fractionated by SDS-PAGE and transferred to nitrocellulose or polyvinylidene difluoride membranes (Millipore, Bedford, MA). Membranes were blocked in 5% milk or 5% BSA for 0.5–1 h and then incubated with the indicated primary antibody at 4°C overnight. Membranes were washed in Tris-buffered saline, 0.1% Tween 20, and incubated with secondary antibody at 20°C for 1 h. Signals were visualized by the Odyssey Infrared Imaging System using fluorescent secondary antibodies or with an enhanced chemiluminescent (ECL) system onto film. β-Actin was used as a loading control.
Quantitative real-time PCR.
Total cellular RNA was extracted using Tri Reagent (Sigma). A total of 2 µg RNA was used for reverse transcription using the SuperScript III First-Strand Synthesis System from Invitrogen (Carlsbad, CA). A total of 2 µL cDNA products were amplified in a 20-µL reaction system containing 10 µL iQ SYBR Green Supermix (Bio-Rad) and 400 nmol/L primer mixture. Relevant primers were purchased from SA Bioscience (Frederick, MD). All reactions were processed in a MyiQ Single Color Real-Time PCR thermocycler using a two-step–plus–melting curve program. Results were analyzed by the iQ5 program (Bio-Rad), and the data were analyzed using the ΔCT method in reference to GAPDH (19).
Chromatin immunoprecipitation assay.
The chromatin immunoprecipitation (ChIP) assay was performed as described (20). In brief, 4–6 × 106 cells were sequentially cross-linked with disuccinimidyl glutarate (Thermo Scientific, Rockford, IL) and 1% formaldehyde in PBS, solubilized in 500 μL SDS lysis buffer (1% SDS; 50 mmol/L Tris [pH 8.0]; and 10 mmol/L EDTA) with a protease inhibitor cocktail (Sigma-Aldrich), and chromatin sheared by sonication. Equal amounts of DNA were immunoprecipitated overnight at 4°C with 4 μg of indicated antibody. Anti-RelA (sc-372) or IgG (as negative control) were obtained from Santa Cruz (Santa Cruz, CA). Antiphospho-serine536 RelA antibodies were from Cell Signaling (catalog no. 3031; Danvers, MA). Immunoprecipitates were collected with protein A magnetic beads (Invitrogen). Eluted DNA was de–cross-linked and used as a template in real-time-PCR. Primers for amplifying the interleukin (IL)-8 promoter were AGGTTTGCCCTGAGGGGATG (F) and GGAGTGCTCCGGTGGCTTTT (R). For the monocyte chemoattractant protein (MCP)-1 promoter, the primer sequences were CTGCTTCCCTTTCCTACT (F) and ATCTTCCATGAGTGATAAGTG (R).
Data are reported as means ± SD. Because all data were normally distributed on the basis of the Kolmogorov-Smirnov test, both one-way and two-way ANOVA tests were performed to evaluate overall group differences. This was followed by the Tukey post hoc test to determine pairwise significance if the ANOVA test indicated that a significant difference was present in the dataset. In the case of only two group comparisons, a two-sample Student t test was performed after checking for variance distribution via the Levene test. In all cases, P < 0.05 was considered significant.
Diabetes reduces renal cortical Klotho transcript and protein levels.
As shown in Fig. 1, renal cortical Klotho protein (by immunoblot) and RNA message (by quantitative RT-PCR) levels in db/db mice were decreased significantly at 20 weeks of age (after ~12 weeks of hyperglycemia). Despite heterogeneity in protein expression levels within each group (Fig. 1A), we observed a 50% decrease in Klotho protein in the db/db group compared with age- and sex-matched controls (Fig. 1B). We observed a significant decline (P < 0.01) in Klotho message (mean ΔCT = 10.4) in the diabetic renal cortex compared with the control renal cortex (mean ΔCT = 4.5) (Fig. 1C), suggesting that the decline in Klotho protein expression resulted from suppressed Klotho gene expression. Klotho was predominantly expressed in tubules but not in the glomeruli of control mice, with less Klotho immunostaining evident in diabetic kidneys (Fig. 1D).
AGEs (MG-HSA) induce the loss of Klotho expression and increase cytokine expression in the primary cultures of renal tubular epithelium.
We previously have reported noncanonical NF-κB pathway activation in the renal cortex of db/db mice, with a several-fold increase in NF-κB–inducing kinase (NIK) expression, a key enzyme regulating activation of this particular NF-κB pathway (18). Because both Klotho and NIK are predominantly expressed in renal tubules (6,21), we explored the possibility that increased NF-κB activation and cytokine production in the renal tubules of db/db mice are linked to the decline of Klotho. Renal tubular epithelial cells were isolated from C57BL/6 mice, and immunofluorescence microscopy was used to check the purity of the enriched tubular epithelial cell preparation (see Supplementary Fig. 1). Primary cultures of enriched tubular epithelial cells were exposed to MG-HSA (5 μmol/L) for 1–24 h followed by RNA isolation with the trizol reagent. As shown in Fig. 2A and B, 24 h of MG-HSA treatment resulted in a fourfold increase in the expression of the receptor for AGEs (RAGEs) and a 50% decline in Klotho levels similar to the decline observed in vivo. MG-HSA treatment produced a significant increase in proinflammatory cytokine mRNA expression for IL-6 (Fig. 2C) and keratinocyte chemoattractant (KC), the mouse ortholog of IL-8 (Fig. 2D), as well as increased protein levels for these cytokines in culture media (Supplementary Fig. 2). The time course of KC production was similar to RAGEs, whereas IL-6 began to increase within 1 h. These experiments suggest that primary renal tubular cells respond to MG-HSA by inducing RAGE expression and increased cytokine production and that a diabetic stimulus, such as MG-HSA, can suppress Klotho expression. Taken together, these experiments indicate that the diabetic milieu decreases renal Klotho levels, which may be linked to diabetes-induced inflammation.
Klotho inhibits NF-κB promoter activity.
Because the loss of Klotho in diabetic kidneys could lead to activation of NF-κB, we investigated whether Klotho negatively regulates NF-κB pathway activity. HEK293 cells were first transfected with the luciferase plasmid containing the consensus NF-κB sequence followed by 45 min of preincubation with different concentrations of exogenously added Klotho prior to TNF-α (20 ng/mL) treatment for 6 h, as previously described (4). Exogenous addition of Klotho inhibited NF-κB–dependent promoter activity in a dose-dependent manner, with an ~70% inhibition observed at 200 pmol/L (Fig. 3A). Klotho was used at this concentration in subsequent experiments. Klotho also had some inhibitory effects on basal NF-κB activation in the absence of TNF-α stimulation (Fig. 3B), suggesting that it might have a constitutive anti-inflammatory role in cells.
Klotho inhibits TNF-α–induced cytokine production.
To further ascertain the inhibitory effect of Klotho on NF-κB activation, HEK293 cells were preincubated with 200 pmol/L Klotho for 45 min prior to TNF-α (20 ng/mL) treatment for 1 h; then the cells were lysed in trizol reagent and the extracted RNA was used for quantitative RT-PCR. Klotho significantly inhibited TNF-α–induced IL-8, IL-6, regulated upon activation normal T-cell expressed, and presumably secreted (RANTES), and MCP-1 production (Fig. 4), although to different extents, which may reflect differences in their kinetics and efficacies of induction by TNF-α. To rule out the possibility that exogenously added Klotho might interfere with TNF-α binding to its receptor, thereby inhibiting NF-κB activation and cytokine production, Klotho was transiently overexpressed in HK-2 cells. These cells were treated with TNF-α (20 ng/mL) for different time intervals. Figure 5A and B shows that overexpressed Klotho completely inhibited MCP-1 expression at all time intervals and significantly inhibited IL-8 expression up to 6 h. MCP-1 and IL-8 levels also were measured in culture media (Fig. 5C and D) and cell lysates (Fig. 5E and F). For these experiments, cells expressing Klotho or vector were exposed to TNF-α for 6 h prior to cytokine measurement by enzyme-linked immunosorbent assay. Although Klotho overexpression significantly reduced IL-8 in both culture media and cell lysates, MCP-1 production was significantly reduced only in culture media and did not reach statistical significance in cell lysates (P = 0.08). Supplementary Fig. 3 shows the induced expression level of Klotho in these cells. Thus, Klotho-mediated inhibition of TNF-α–induced NF-κB activation is a cellular event and not the physical interference of TNF-α binding to its receptor.
Klotho inhibits TNF-α–induced RelA (Ser)536 phosphorylation.
To further explore mechanism(s) of Klotho-mediated inhibition of TNF-α–induced NF-κB activation, we investigated the impact of Klotho on cytoplasmic and nuclear events in NF-κB activation. Figure 6A and Supplementary Fig. 4 show similar TNF-α–induced cytosolic IκBα degradation in the presence and absence of Klotho, suggesting that Klotho has no effect on IκB kinase (IKK) activation of the NF-κB canonical pathway. Next, we quantified nuclear translocation of total RelA after TNF-α treatment for 30 min. This experiment demonstrated that increased RelA nuclear accumulation in response to TNF-α was unaffected by Klotho treatment (Fig. 6B, top panel and Supplementary Fig. 5). However, when immunoblotted for RelA (Ser)536 phosphorylation, Klotho blocked this phosphorylation event in the cytoplasm and to a lesser, but still significant, extent in the nucleus (Fig. 6B, middle panel). Figures 6C and D provide quantitation of RelA (Ser)536 immunoblots from three independent experiments. These results suggest that Klotho specifically blocked a noncanonical NF-κB activation pathway mediated by RelA (Ser)536 phosphorylation and did not inhibit the IKK-dependent classical NF-κB activation pathway.
Klotho inhibits p38 kinase.
Although several kinases have been reported to phosphorylate RelA at Ser536 in response to various stimuli (22), IKKβ and phosphoinositide 3-kinase (PI3K)/Akt seem to be the predominant kinases involved. Because Klotho had no effect on TNF-α–induced IκBα degradation, IKKβ may not be a site of action of Klotho. However, earlier reports have shown that the exogenous addition of Klotho significantly reduced Akt phosphorylation (4). Moreover, RelA (Ser)536 phosphorylation by Akt has been reported to be mediated via p38 activation (23). Therefore, we investigated if Klotho inhibits p38 phosphorylation. HEK293 cells were treated with TNF-α (20 ng/mL for 1 h), and p38 phosphorylation was examined by immunoblotting. TNF-α treatment increased p38 phosphorylation compared with untreated cells (Fig. 7A, upper panel), and pretreatment with Klotho significantly reduced the TNF-α–induced phosphorylation. Immunoblots from three independent experiments were quantitated and normalized to total p38 and expressed as fold change versus control cells (Fig. 7A, lower panel). We also investigated if Klotho inhibited IKKα, another RelA (Ser)536 kinase. Although 30 min of TNF-α treatment increased p-IKKα levels, as demonstrated by immunoblot, exogenously added Klotho failed to inhibit the phosphorylation (Fig. 7B and Supplementary Fig. 6).
Klotho inhibits TNF-α–induced recruitment of phospho-RelA (Ser)536 on IL-8 and MCP-1 promoters.
We investigated whether Klotho inhibition of RelA phosphorylation affected its recruitment onto NF-κB–dependent promoters, leading to inhibition of cytokine production. HEK293 cells were pretreated with Klotho (200 pmol/L) for 45 min and then stimulated with TNF-α (20 ng/mL) for 30 min. Recruitment of either total RelA or phospho–RelA (Ser)536 onto IL-8 and MCP-1 promoters was evaluated by the ChIP assay. Figure 8A shows a strong induction of total RelA as well as phospho–RelA (Ser)536 binding on the MCP-1 promoter after 30 min of TNF-α treatment, whereas Klotho pretreatment in the absence of TNF-α had no significant effect on their promoter binding. Klotho treatment significantly reduced phospho–RelA (Ser)536 binding induced by TNF-α (Fig. 8A, lower panel) without significantly affecting total RelA binding to the MCP-1 promoter. Similar results were obtained using the IL-8 promoter (Fig. 8B). Supplementary Fig. 7 shows a representative gel from these experiments depicting the similar changes in the RelA and RelA (Ser)536 recruitment on both the promoters. These results indicate that Klotho did not inhibit TNF-α–induced RelA nuclear translocation and relevant promoter binding but specifically blocked RelA (serine)536 phosphorylation and its binding to NF-κB–dependent promoters and further suggest that the anti-inflammatory effects of Klotho are mediated via inhibition of this NF-κB noncanonical activation pathway.
The recent characterization of Klotho signaling, an antiaging hormone that modulates the expression level of antioxidant enzymes, offers a potentially new focus for understanding the accelerated senescence and increased oxidative stress observed in diabetes. Klotho recently has been shown to have renoprotective effects primarily mediated by mitigating mitochondrial oxidative stress and apoptosis (12). In the current study, we have shown a significant decrease of renal Klotho mRNA as well as protein in db/db mice compared with age- and sex-matched controls, which is consistent with a recent report (24) of a similar decline of Klotho in kidneys of streptozotocin-induced diabetic rats. In the latter study, the decline in Klotho was linked to hyperglycemia because two pharmacologic approaches to attenuate the severity of hyperglycemia (insulin and phloridzin) restored Klotho levels. We also have reported a novel role for Klotho as a negative modulator of NF-κB activation via a specific noncanonical activation pathway involving RelA (Ser)536 phosphorylation. These results suggest that Klotho not only confers resistance to oxidative stress (4) but may provide protection against the aberrant activation of NF-κB pathways. Because Klotho knockout mice exhibit an enhanced aging phenotype, a diabetes-induced decline in Klotho could be an unappreciated mechanism linking increased oxidative stress, inflammation, and accelerated aging in the pathophysiology of diabetic nephropathy.
The concept that diabetes is a chronic inflammatory disease is supported by a rapidly growing array of clinical and experimental data indicating that inflammation, manifested as increases in TNF-α and IL-6, is a common denominator linking obesity, insulin resistance, atherosclerosis, dyslipidemia, and excessive glucose metabolism in diabetes (25–28). We recently have reported that chronic NF-κB activation in the renal cortex of db/db mice involves a novel mechanism of NF-κB activation with recruitment of both the canonical and noncanonical NF-κB pathways (18). In that study, we have shown that multiple proinflammatory cytokines relevant to diabetes, including TNF-α, are increased in the renal cortex of db/db mice. We have shown here that Klotho, either expressed transiently as a membranous form or added exogenously in its soluble form, suppresses TNF-α–induced NF-κB activation and subsequent production of proinflammatory cytokines such as RANTES, MCP-1, IL-6, and IL-8. Our rationale for selecting TNF-α as a diabetic milieu inflammatory stimulus was based on clinical and epidemiological data indicating a role for TNF-α in the initiation and progression of diabetes, including the following: 1) proinflammatory cytokines such as IL-6, TNF-α, IL-1β, and plasminogen activator inhibitor-1 are increased in diabetic patients and are independently correlated with duration of diabetes (2,29) and 2) acute hyperglycemia increases circulating levels of IL-6, TNF-α, and IL-18 (30,31). Although TNF-α expression in insulin-resistant and in diabetic subjects has been reported to be several-fold higher than in control subjects (28,32), these studies remain controversial. TNF-α plays a permissive role in the development of obesity-induced insulin resistance in mice (33), and numerous tissue culture experiments have demonstrated that glucose, AGEs, and a variety of RAGE ligands increase TNF-α in a variety of cell types (30,34–36). Taken together, these studies indicate that TNF-α is a product of numerous diabetic stimuli and suggest that TNF plays an important upstream role in the activation of NF-κB observed in diabetes.
There are various regulatory check points for TNF-α–induced canonical NF-κB pathway activation. Canonical NF-κB proteins (RelA and p50) remain sequestered in the cytoplasm by inhibitors of NF-κB (such as IκB) (37). In response to stimuli that activate NF-κB via the canonical pathway, such as TNF-α, there is activation of the IKK complex; this leads to IκBα phosphorylation at two specific NH2-terminal serine residues of the protein, targeting it for ubiquitination and subsequent proteosomal degradation (38,39). This process allows cytoplasmically sequestered RelA p50 dimers to enter the nucleus and bind to NF-κB promoter sequences to initiate transcription of specific genes. Although Klotho inhibited NF-κB activation, it did not interfere with TNF-α–induced IκBα degradation, total RelA nuclear translocation, or DNA binding, suggesting that Klotho might use a novel mechanism of inhibition. Potential possibilities could be inhibition by posttranslational modification of RelA or formation of an active enhanceosome on NF-κB promoters.
Our findings differ slightly from the recent observations of Maekawa et al. (40). Although both studies report a role for Klotho in modulating NF-κB activity, the latter study reported that Klotho added exogenously blocked expression of NF-κB–dependent adhesion molecules by inhibiting IκBα phosphorylation in response to TNF-α treatment. Several potential reasons for this discrepancy are apparent, including 1) differences in cell type (human umbilical vein endothelial cells vs. renal epithelial cells used in our study), 2) the unique expression of signaling molecules and distinctive environments of the cultured cells, 3) the source of Klotho, and 4) the relatively high constitutive phosphorylation of IκBα in the aortic endothelial cell control group. It was surprising that these investigators did not measure total IκBα degradation or inhibition of RelA nuclear localization by Klotho to buttress their suggestion of NF-κB canonical pathway inhibition by Klotho.
We have shown that Klotho specifically inhibited TNF-α–induced RelA phosphorylation at Ser536 and its subsequent recruitment on IL-8 and MCP-1 promoters, without affecting total RelA recruitment in response to TNF-α stimulation. RelA (Ser)536 phosphorylation has been shown to occur independently of IκBα degradation (13,14) and is suggested to be critical for increased NF-κB transactivation and subsequent activation of a subset of NF-κB–dependent genes (14). Although the induction of RelA nuclear translocation has been regarded as the main event in NF-κB pathway activation, more recently the phosphorylation of RelA at various critical Ser residues in its transactivation domain has been proposed as a key event in the activation of NF-κB signaling (22). The literature suggests that several kinases may phosphorylate the RelA transactivation domain at Ser536 in response to TNF-α and IL-1, including IKKs, PI3K/Akt, TANK-binding kinase-1, and NIK (15,41). NIK, which is induced several-fold in diabetic kidneys, has been reported to induce RelA (Ser)536 phosphorylation directly (13), through p38 mitogen-activated protein kinase (42), or in cooperation with activated Cot (a serine/threonine kinase) (43,44). In addition, overexpression of constitutively active Akt increases RelA transactivation domain phosphorylation in the absence of other stimuli, and pretreatment of cells with LY294002, a PI3K inhibitor, completely blocks this phosphorylation event in response to TNF-α and IL-1 (15,45). Taken together, these results clearly suggest involvement of PI3K/Akt in TNF-α/IL-1–induced RelA (Ser)536 phosphorylation and subsequent NF-κB activation. Akt has been proposed to mediate RelA phosphorylation by activating p38 and requiring both IKKs, IKKα and IKKβ (23). Because we did not observe an inhibitory effect of Klotho on TNF-α–induced IκBα degradation, we reasoned that IKKβ, which is a potent inducer of IκBα phosphorylation and its subsequent degradation, may not be a target of Klotho. Likewise, TANK-binding kinase-1, which is an IKK-related kinase but not a part of the IKK complex and not required for IκBα phosphorylation and degradation, can phosphorylate RelA at Ser536 in response to TNF-α treatment (46,47). Thus, it is possible that Klotho may unleash its anti-inflammatory effects by blocking NF-κB activation through inhibiting the PI3K/Akt, IKKα, or NIK activation normally seen after TNF-α treatment. Our finding that Klotho blocks p38 phosphorylation is consistent with earlier observations that Klotho blocks basal or insulin-induced Akt phosphorylation (4). It will be important to identify the RelA transactivation domain kinase, which is sensitive to Klotho to understand how Klotho suppresses this noncanonical RelA activation pathway and how its loss contributes to the renal inflammation associated with diabetic nephropathy.
In conclusion, we have established that Klotho has an anti-inflammatory function in the kidney in addition to its ability to engender resistance to oxidative stress. Thus, an adequate tissue level of Klotho may provide dual protection against both oxidative stress and inflammation, while loss of Klotho may be a common denominator linking increased oxidative stress, NF-κB activation, and accelerated aging in diabetes.
This work was supported by a grant to S.C. and R.G.T. from the National Institutes of Health (NIH), National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK-079053) and a grant to S.C. from the American Heart Association (0630100N). K.P.R. also was supported by an NIH National Heart, Lung, and Blood Institute Grant (NO1-HV-28184), a Welch Foundation Endowment in Chemistry and Related Science Grant (L-AU-0002), an American Diabetes Association Grant (1-08-CR-65), and an NIH Clinical and Translational Award (UL1-RR-024148-05).
No potential conflicts of interest relevant to this article were reported.
Y.Z., S.B., and N.D. performed the experiments. W.S.L. contributed to animal handling and mouse renal primary tubular epithelial cell isolation. P.S.S. researched data and contributed to discussion. R.B. prepared the expression plasmids and purified the Klotho protein. K.P.R. contributed to data analysis and discussion and reviewed and edited the manuscript. R.G.T. designed the animal studies, analyzed data, contributed to discussion, and reviewed and edited the manuscript. S.C. researched and analyzed data, contributed to discussion, and wrote the manuscript.
The authors thank Allan Brasier, University of Texas Medical Branch, Galveston, TX, for his helpful discussions concerning the approaches used in this study and his expertise in the field of NF-κB activation and inflammation. The authors also acknowledge the assistance of Heidi Sprat, University of Texas Medical Branch, Galveston, TX, with statistical analyses.