AMPK Potentiates Hypertonicity-induced Apoptosis by Suppressing NFκB/COX-2 in Medullary Interstitial Cells

Abstract

Cells residing in the hypertonic, hypoxic renal medulla depend on dynamic adaptation mechanisms to respond to changes in energy supply and demand. The serine/threonine kinase 5′-AMP protein kinase (AMPK) is a sensor of cellular energy status, but whether it contributes to the survival of cells in the renal medulla is unknown. Here, hypertonic conditions induced a decrease in AMPK phosphorylation within 12 hours in renal medullary interstitial cells (RMIC), followed by a gradual return to baseline levels. Activation of AMPK markedly increased hypertonicity-induced apoptosis of RMICs and suppressed both hypertonicity-induced NFκB nuclear translocation and cyclooxygenase-2 (COX-2) activation; overexpression of COX-2 significantly attenuated these effects. AMPK activation also markedly reduced generation of reactive oxygen species and nuclear expression of tonicity-responsive enhancer-binding protein, which prevented upregulation of osmoprotective genes. In vivo, pharmacologic activation of AMPK led to massive apoptosis of RMICs and renal dysfunction in the setting of water deprivation in mice. Taken together, these results identify a critical role for AMPK in the maintenance of RMIC viability and suggest that AMPK modulates the NFκB-COX-2 survival pathway in the renal medulla. Furthermore, this study raises safety concerns for the development of AMPK activators as anti-diabetic drugs, especially for patients prone to dehydration.

The renal medulla is a unique tissue in which cells are exposed to a hostile environment characterized by extremely high concentrations of sodium chloride and urea and low oxygen tension.^1,2 During antidiuresis, the renal medulla generates a tonicity up to four times the osmolality of plasma by the renal medulla-concentrating mechanism involving a countercurrent system of renal tubules and vessels. Low blood flow of the renal medulla (medullary hypoxia) has been found to be essential for the establishment of efficient renal concentration.³ With excessive blood flow, medullary osmolality gradients are disrupted; with too little blood flow, medulla hypoxic injury occurs. Therefore, an exact match of oxygen (energy) supply and demand is critical for the renal medulla to maintain the cell viability and function.¹

During antidiuresis, osmolality in the renal medulla significantly increases within a few hours, exceeding 1000 mOsm/kg H₂O in humans,⁴ and medullary cells initially shrink and then return to normal cell volume with the enhanced availability of extracellular Na⁺ by increasing Na⁺/K⁺/ATPase activity.² After prolonged exposure of hypertonicity, medullary cells activate various protective mechanisms that allow them to survive in this inhospitable environment.² In addition to tonicity-responsive enhancer-binding protein/osmotic response element-binding protein (TonEBP/OREBP) and its target genes, including aldose reductase (AR), sodium myo-inositol transporter (SMIT), betaine/γ-aminobutyric acid transporter (BGT1), and heat shock protein 70,^5–7 up-regulation of Na⁺/K⁺/ATPase expression represents an early protective mechanism; the up-regulation maintains the function of transmembrane transport pathways indispensable for cell homeostasis and viability.^8–10 However, activation of Na⁺/K⁺/ATPase may further deplete intracellular ATP storage and increase energy demand in medullary cells, thus predisposing the renal medulla to hypoxia damage.

5′-AMP protein kinase (AMPK) is a major cellular energy sensor and is sensitive to the cellular ratio of AMP to ATP. A high AMP or low ATP level activates AMPK, which inhibits energy-consuming processes and enhances energy-producing processes to restore the energy homeostasis.¹¹ AMPK is regulated by various stimuli including energy depletion, exercise, hypoxia, and hypertonicity in many cell types,^12–15 where it inhibits cell viability and induces apoptosis by up-regulating p53 expression, activating the caspase cascade and reducing the activation of the cytoprotector cyclooxygenase 2 (COX-2).^16–18 Because renal medullary interstitial cells (RMICs) are normally subjected to hypoxia and hypertonic challenges and NFκB-driven COX-2 expression is critical to the capability of RMICs to survive under hyperosmotic stress,¹⁹ we aimed to investigate whether AMPK plays an important role in regulating cell survival of RMICs and whether its activation results in massive apoptosis under hypertonic conditions.

RESULTS

Hypertonicity Suppressed AMPK Phosphorylation in RMICs

To evaluate the effect of hypertonicity on AMPK activity, rabbit RMICs were treated with NaCl-mediated hypertonicity (500 mOsm) for various times, and phosphorylated AMPKα protein level was analyzed by immunoblotting. In cultured RMICs, hypertonicity decreased AMPKα phosphorylation 1 hour after exposure to hypertonic media and decreased further thereafter (Figure 1A). At 12 hours, phosphorylated AMPKα decreased to 38.6 ± 0.8% of the baseline level (Figure 1B). After exposure to hypertonicity for 24 hours, the phosphorylated AMPKα levels returned to the baseline level. In addition, we examined the effect of hyperosmolality on AMPK in RMICs at earlier time points from 0 to 60 minutes. As shown in Supplemental Figure 2A, AMPK activity was significantly increased at 5 minutes and then reduced in RMICs 15 minutes after hypertonic treatment and remained at low level for 12 hours (Figure 1). Interestingly, AMPK activity significantly increased in cultured inner medullary collecting duct cells (IMCDs) (Supplemental Figure 2B), which suggests that hypertonicity regulates AMPK activity in a cell type-specific manner.

Figure 1.

Effect of hypertonicity on AMP-activated protein kinase (AMPK) phosphorylation. (A) Renal medullary interstitial cells (RMICs) were subjected to hypertonic stress (500 mOsm) for the indicated times. Phospho-AMPKα (Thr172) and total AMPKα levels in the cells were examined by Western blot analysis. The results are representative of at least three independent experiments. (B) Densitometry quantification of phosphorylated AMPKα (p-AMPKα) and AMPKα levels during each time (**P < 0.01 versus 300 mOsm, n = 3).

Effect of Hypertonicity on the Ratio of AMP to ATP and the Expression of LKB1, calmodulin kinase II (CaMKII), and PP2A

To elucidate the mechanisms by which hypertonicity inhibited AMPK activation, we measured AMP and ATP levels in RMICs exposed to hypertonic stress. As shown in Supplemental Figure 3A, the ratios of AMP to ATP rapidly decreased after treatment of cells with hypertonicity for 5 minutes, and the reduced AMP/ATP ratios remained for at least 1 hour. However, the expression level of several key upstream kinases/phosphatases of AMPK including LKB1, CaMKII, and PP2A did not change in hypertonicity-treated RMICs (Supplemental Figure 3B). These findings suggest that the reduction of the ratio of AMP to ATP, but not the change of upstream AMPK kinase/phosphatase activity, may be responsible for hypertonicity-mediated suppression of AMPK activity in RMICs.

AMPK Activation Facilitated Hypertonicity-induced Apoptosis of RMICs

To investigate the role of early AMPK inactivation in RMIC survival under hypertonic stress, we treated cells with 5-aminoimidazole-4-carboxyamide ribonucleoside (AICAR), an activator of AMPK, to prevent the AMPK inactivation and determined the cell viability by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. RMICs were treated with different extents of hyperosmotic challenge (300 to 800 mOsm/Kg H₂O) in the presence of 1 mM AICAR for 12 hours. Aberrant AMPK activation by AICAR treatment aggravated the cell death of RMICs induced by hypertonic stress, with little effect on RMIC viability under the isotonic condition (Figure 2A). AICAR treatment decreased cell viability from 88.4 ± 2.2% to 70.9 ± 0.5% in 400 mOsm, from 72.3 ± 1.0% to 35.1 ± 1.3% in 500 mOsm, from 19.0 ± 0.7% to 6.0 ± 0.2% in 600 mOsm, and from 8.7 ± 0.3% to 2.5 ± 0.5% in 800 mOsm, respectively. Flow cytometry was further performed to examine hypertonicity-induced apoptosis of RMICs. AMPK activation enhanced hypertonicity (500 mOsm)-induced cell apoptosis by approximately four-fold (from 9.9 ± 1.7% to 42.6 ± 4.3%), but AICAR failed to affect cell survival in the isotonic condition (Figure 2B). RMICs exposed to AICAR treatment showed increased cleavage of procaspase-3 to caspase-3. As shown in Figure 2C, AICAR treatment significantly augmented caspase-3 activation triggered by hypertonic stress. These findings suggest that AMPK activation facilitates hypertonicity-induced RMIC apoptosis, concomitant with caspase-3 activation.

Figure 2.

Effect of 5′-AMP protein kinase (AMPK) activation on the survival of cultured renal medullary interstitial cells (RMICs) after hypertonic stress. RMICs were subjected to the indicated osmolalities for 12 hours with or without 5-aminoimidazole-4-carboxyamide ribonucleoside (AICAR) treatment (1 mM). (A) Cell viability was determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (**P < 0.01 versus cells under the same osmolality without AICAR treatment, n = 4). (B) Quantification of apoptotic cells by flow cytometry analysis (*P < 0.05, **P < 0.01 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 6). (C) Quantification of the ratio of cleaved caspase-3 to total caspase-3 in RMICs with hypertonicity (500 mOsm) for 12 hours in the presence or absence of AICAR (1 mM) (*P < 0.05, **P < 0.01 versus 300 mOsm; ^#P < 0.05 versus 500 mOsm, n = 5).

To confirm that AICAR accelerates hypertonicity-induced cell death in RMICs via activating AMPK, we tested another selective AMPK activator, A-769662. As shown in Supplemental Figure A4, A-769662 significantly increased hyperosmolality-induced cell apoptosis, which was assessed by MTT and cytometry assays (Supplemental Figure 4, A and B). Treatment of RMICs with A-769662 also increased hypertonicity-induced activation of caspase-3 (Supplemental Figure 4C).

AMPK Activation Inhibited Hypertonicity-elicited Activation of the NFκB Pathway

Hypertonicity-induced NFκB activation contributes to increased COX-2 expression, which is critical for the survival of RMICs under hypertonic stress.¹⁹ Increasing evidence suggests that AMPK activation may inhibit NFκB activity.²⁰ Therefore, we determined the effect of AMPK activation on the NFκB–COX-2 signaling pathway. Treating RMICs with AICAR for 12 hours markedly blocked the hypertonicity-induced decrease in cytoplasm IκB protein level (Figure 3A) and the increase in nuclear NFκB p65 protein level (Figure 3B), which suggests that AICAR inhibits the translocation of NFκB into the nucleus. Immunofluorescent staining further confirmed that hypertonicity-induced NFκB p65 nuclear translocation was abolished by AICAR treatment (Figure 3C). Consistently, AICAR treatment almost completely blocked the hypertonicity-induced NFκB luciferase reporter activity (from 13.0 ± 1.9- to 1.6 ± 0.4-fold as compared with the control group) (Figure 3D). We also noticed that NFκB activity was markedly increased 4 hours after hypertonicity treatment, which was also blocked by AICAR (Supplemental Figure 5). These observations were further supported by the finding that AICAR treatment prevented hypertonicity-induced NFκB-DNA binding activity as assessed by electrophoretic mobility shift assay (EMSA) (Figure 3E).

Figure 3.

5′-AMP protein kinase (AMPK) activation inhibits hypertonicity-induced NFκB nuclear translocation and activity. Renal medullary interstitial cells (RMICs) were subjected to 500 mOsm for 12 hours in the presence or absence of AICAR (1 mM). (A) Western blot assay of cytoplasm IκBα protein expression (densitometry, *P < 0.05 versus 300 mOsm, ^#P < 0.05 versus 500 mOsm alone, n = 3). (B) Western blot analysis of nuclear NFκB p65 protein levels (densitometry, **P < 0.01 versus 300 mOsm; ^#P < 0.05 versus 500 mOsm alone, n = 3). (C) Confocal immunofluorescence analysis was used to detect NFκB nuclear translocation. (D) Luciferase reporter assay for NFκB transcriptional activity. RMICs were transfected with a luciferase reporter gene that contains putative binding sites for NFκB. Effect of hypertonicity with or without AICAR on NFκB-luc reporter activity was determined and then normalized to β-gal luciferase activity (**P < 0.01 versus 300 mOsm, ^##P < 0.01 versus 500 mOsm, n = 7). (E) Effect of hypertonicity on DNA-binding activity of NFκB in RMICs. EMSA involved incubating nuclear extracts (5 μg of protein) with biotin-labeled NFκB consensus oligonucleotide. A representative gel from three independent experiments is shown. The arrow indicates a specific NFκB DNA probe-protein complex.

AMPK Activation Blocked Hypertonicity-induced COX-2 Expression

To examine the role of AMPK in hypertonicity-induced COX-2 expression, RMICs. Exposing RMICs to hypertonic stress significantly increased the expression of COX-2 by 3.7-fold at the mRNA level (Figure 4A) and 3.2-fold at the protein level (Figure 4C). This effect was completely blocked by AICAR treatment at both mRNA and protein levels (Figure 4, A and C). Similarly, AICAR suppressed the effect of hypertonicity on COX-2 promoter activity (Figure 4B). To further confirm the inhibitory effect of AMPK activation on COX-2 activity, prostaglandin (PG) E₂ and PGI₂ production was measured. Hypertonicity significantly increased the production of PGE₂ and PGI₂ in RMICs (PGE₂: from 1127 ± 136.4 to 3870 ± 316.3 ng/mg protein; PGI₂: from 401.2 ± 23.3 to 1034.7 ± 65.5 ng/mg protein); and both were blocked by AICAR treatment (PGE₂: from 3870 ± 316.3 to 1086 ± 138.2 ng/mg protein; PGI₂: from 1034.7 ± 65.5 to 139.9 ± 13.6 ng/mg protein) (Figure 4, D and E). In addition, AICAR treatment for 4 hours also blocked hypertonicity-induced COX-2 expression (Supplemental Figure 5C). Furthermore, we found that hypertonicity appeared to increase PPARδ expression at the 6-hour time point in RMICs, and AICAR treatment had little effect on PPARδ expression (Supplemental Figure 6A). Collectively, these findings reveal an inhibitory role for AMPK in hypertonicity-stimulated activation of the NFκB-COX-2 pathway.

Figure 4.

5′-AMP protein kinase (AMPK) activation inhibits hypertonicity-induced cyclooxygenase-2 (COX-2) expression and activity. Renal medullary interstitial cells (RMICs) were treated with hypertonic stress (500 mOsm) for 12 hours in the presence or absence of AICAR (1 mM). (A) COX-2 mRNA level was analyzed by real-time PCR (*P < 0.05 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 3). (B) Effect of AICAR on hypertonicity-induced COX-2 promoter luciferase activity (**P < 0.01 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 4). (C) Immunoblot analysis showing that AICAR treatment attenuated hypertonicity-induced COX-2 protein expression (**P < 0.01 versus 300 mOsm; ^#P < 0.05 versus 500 mOsm, n = 3). (D) Effect of AICAR treatment on the production of PGE₂ by ELISA (**P < 0.01 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 4). (E) Effect of AICAR treatment on the production of PGI₂ by ELISA (**P < 0.01 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 4).

Overexpression of AMPKα Increased Apoptosis and Attenuated COX-2 Expression in RMICs Treated with Hypertonicity

AICAR might exert a cytotoxic effect via an AMPK-independent pathway. To confirm whether the effect of AICAR on facilitating RMIC apoptosis under hypertonic stress depends on AMPK activation, RMICs were infected with a recombinant adenovirus expressing a constitutively active form of AMPKα (Ad-AMPK-CA) for 36 hours. As shown in Figure 5, Ad-AMPK-CA significantly increased apoptosis (Figure 5A) and procaspase-3 cleavage and inhibited COX-2 expression in RMICs treated with hypertonicity (Figure 5B), which strongly supports the possibility that AICAR-induced apoptosis of RMICs under high osmolality is through AMPK activation.

Figure 5.

Overexpression of a constitutively active form of 5′-AMP protein kinase (AMPK)α aggravates hypertonicity-induced apoptosis and attenuates cyclooxygenase-2 (COX-2) expression in renal medullary interstitial cells (RMICs). Cultured RMICs were infected with adenoviruses expressing a constitutively active form of AMPKα (Ad-AMPK-CA) or a full-length green fluorescence protein (Ad-GFP). After 36 hours of infection, the cells were challenged with hypertonic stress (500 mOsm) for 12 hours. (A) Effect of Ad-AMPK-CA infection on cell apoptosis of RMICs and quantification of apoptotic cells by FACS (**P < 0.01 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 6). (B) Immunoblot assay examining the effect of Ad-AMPK-CA infection on caspase-3 activation and COX-2 expression. The results are representative of at least three independent experiments.

Overexpression of COX-2 Significantly Attenuated AMPK Activation-associated RMIC Apoptosis under Hypertonic Stress

To examine whether AICAR potentiated the apoptosis of RMICs via inhibition of the COX-2 pathway, we overexpressed COX-2 using a recombinant adenovirus (Ad-COX-2) in RMICs for 36 hours and determined its effect on RMIC survival. Ad-COX-2 infection markedly increased COX-2 expression in RMICs under both isotonic and hypertonic conditions (Supplemental Figure 7A). As shown in Figure 6A, overexpression of COX-2 significantly attenuated AICAR-induced cell death in the hypertonic condition (cell viability: from 40.5 ± 1.6% to 62.4 ± 3.8%). Consistently, overexpression of COX-2 significantly blocked AICAR-induced procaspase-3 cleavage (Figure 6B and Supplemental Figure 7B). These effects were associated with a marked increase in PGE₂ and PGI₂ production (Figure 6 C and D).

Figure 6.

Overexpression of cyclooxygenase-2 (COX-2) significantly blocks 5′-AMP protein kinase (AMPK) activation-induced renal medullary interstitial cell (RMIC) apoptosis under hypertonic stress. Cultured RMICs were infected with an adenovirus expressing a full-length rabbit COX-2 (Ad-COX-2) or its vector virus (Ad-Null). After 36 hours of infection, the cells were challenged with hypertonic stress (500 mOsm) in the presence or absence of AICAR (1 mM) for 12 hours. (A) 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay of cell viability (*P < 0.05, **P < 0.01 versus cells infected with Ad-Null under the same osmotic pressure (500 mOsm), n = 4). (B) Effect of COX-2 overexpression on caspase-3 activation. The results are representative of at least three independent experiments. (C and D) Effect of COX-2 overexpression on biosynthesis of PGE₂ (C) and PGI₂ (D). The medium of cultured cells was collected for determination of PGE₂ (C) and PGI₂ (D) level by ELISA (*P < 0.05, **P < 0.01 versus their control group with Ad-Null, n = 4).

PGI₂ Attenuated AMPK-elicited Apoptosis in RMICs under Hypertonic Stress

To determine whether PGI₂ could attenuate AMPK-induced apoptosis of RMICs, we treated RMICs with exogenous PGI₂. We found that AMPK reduced hypertonicity-induced PGI₂ and PGE2 biosynthesis (Figure 6 C and D) via blocking of the NFkB/COX-2 pathway (Figures 3 and 4) and that of PGI₂, but not PGE₂, protected AICAR-treated RMICs from apoptosis (Supplemental Figure 6, B and C). This finding suggests that AMPK may induce RMIC apoptosis, at least in part, via blocking PGI₂/PPARδ-dependent pathway.

Inhibition of Reactive Oxygen Species (ROS) Production Increased Hypertonicity-induced Caspase-3 Activation and Reduced Osmoprotective Gene Expression

Hypertonic stress resulted in a two-fold increase in ROS production in a time-dependent manner (Figure 7A), which was completely blocked by treatment with antioxidant N-acetyl-L-cysteine (NAC) (Figure 7B). Reduced ROS level in hypertonicity-treated RMICs was associated with a marked increase in caspase-3 activation (Figure 7C) and a significant reduction in osmoprotective AR and SMIT mRNA expression (Figure 7D). These findings suggest that hypertonicity-induced ROS production is essential to the survival of RMICs. In support of this, treating RMICs with H₂O₂ induced the expression of COX-2 in a time- and dose-dependent manner (Supplemental Figure 8, A and B). As expected, NAC treatment markedly attenuated hypertonicity-induced COX-2 expression in RMICs (Supplemental Figure 8C).

Figure 7.

Oxidative stress mediates hypertonicity-induced apoptosis and osmoprotective gene up-regulation. (A) Renal medullary interstitial cells (RMICs) were exposed to hypertonic stress (500 mOsm) for the indicated times, and reactive oxygen species (ROS) production was measured by dichlorofluorescein (DCF) (*P < 0.05, **P < 0.01 versus the baseline, n = 7). (B) RMICs were treated with hypertonicity (500 mOsm) for 1 hour in the presence or absence of N-acetyl-L-cysteine (NAC) (1 mM), and ROS production was measured (**P < 0.01 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 3). (C) RMICs were subjected to hypertonic stress (500 mOsm) for 12 hours in the presence or absence of NAC (1 mM), and then total cell protein underwent immunoblotting analysis for procaspase-3 and cleaved caspase-3 levels. Densitometric analysis of caspase-3 cleavage (*P < 0.05 versus300 mOsm; ^#P < 0.05 versus 500 mOsm, n = 3). (D) Real-time PCR analysis of mRNA expression of osmoprotective gene aldose reductase (AR) and sodium myo-inositol transporter (SMIT) (*P < 0.05, **P < 0.01 versus 300 mOsm; ^#P < 0.05 versus 500 mOsm, n = 6).

AMPK Activation Prevented Hypertonicity-induced ROS Production and TonEBP Transcriptional Activity

Hypertonic stress has been reported to activate TonEBP, a nuclear transcription factor that controls the transcription of many osmoprotective genes including AR, BGT1, and SMIT, thus leading to the accumulation of intracellular organic osmolytes.²¹ A previous study demonstrated that inhibition of COX-2 activity was associated with impaired hypertonicity-induced accumulation of osmolytes.²² To examine whether AMPK activation affects these processes by disturbing ROS production under hypertonic conditions, we analyzed the effect of AICAR on ROS level and TonEBP transcriptional activity. AICAR treatment markedly lowered ROS production in RMICs challenged by hypertonic stress (Figure 8A). AICAR also inhibited high NaCl-induced nuclear translocation of TonEBP (Figure 8B), concomitant with a reduced expression of its target genes AR and SMIT (Figure 8C).

Figure 8.

5′-AMP protein kinase (AMPK) activation prevents hypertonicity-induced reactive oxygen species (ROS) production and osmoprotective gene expression. (A) Renal medullary interstitial cells (RMICs) were exposed to hypertonic stress (500 mOsm) for 1 hour in the presence or absence of 1 mM AICAR, and ROS production was measured (**P < 0.01 versus 300 mOsm; ^#P < 0.05 versus 500 mOsm, n = 3). (B) RMICs were treated with hypertonicity (500 mOsm) for 12 hours in the presence or absence of AICAR (1 mM). Nuclear protein extracts (50 μg) were used for Western blot analysis of tonicity-responsive enhancer-binding protein (TonEBP) levels. The results are representative of three independent experiments. (C) Effect of AMPK activation on osmoprotective gene expression in RMICs exposed to hypertonic stress. RMICs were subjected to hypertonic stress (500 mOsm) for 12 hours in the presence or absence of 1 mM AICAR. Real-time PCR analysis of AR (left) and SMIT (right) mRNA levels (**P < 0.01 versus 300 mOsm; ^##P < 0.01 versus 500 mOsm, n = 3).

Aberrant AMPK Activation Induced Apoptosis of RMICs in Dehydrated Mice in Vivo

Mice were treated with AICAR for 2 days and then were deprived of water for 6, 12, and 24 hours. Although water deprivation alone appeared to have little effect, AICAR treatment significantly increased AMPK activation in both renal medulla (Supplemental Figure 9, A and B) and cortex (Supplemental Figure 9, C and D) as assessed by immunoblotting and immunohistochemistry. After water deprivation, apoptosis of RMICs was analyzed. Water deprivation alone for 6 to 24 hours failed to induce cell apoptosis in the renal medulla of mouse kidneys (Figure 9, A through C). In contrast, a significant amount of apoptotic cells appeared after 6 hours (Figure 9D) and was further increased after 12 and 24 hours of water deprivation with AMPK activation (Figure 9, E and F). To identify the cell type of apoptotic cells, collecting ducts were counterstained with an aquaporin 2 (AQP2) antibody, with AQP2-positive cells shown in dark blue. Most of the apoptotic cells were medullary interstitial cells, and a small amount appeared to be AQP2-positive collecting duct cells (Figure 9, D through F). AICAR treatment resulted in a significant increase in apoptosis of both RMICs (Figure 9G) and medullary collecting duct cells (data not shown).

Figure 9.

5′-AMP protein kinase (AMPK) activation induces renal medullary interstitial cell (RMIC) apoptosis in water-deprived mice. C57BL/6 mice were treated with AICAR at 0.25 mg/g body weight by intraperitoneal injection for 2 days, and then water was deprived for 6, 12, and 24 hours. Apoptotic cells were identified as brown staining by terminal deoxynucleotidyl transferase–mediated digoxigenin-deoxyuridine nick-end labeling (TUNEL) assay. Aquaporin 2 (AQP2) was counterstained to indicate the position of the collecting duct cells, shown in dark blue. (A through C) TUNEL assay of apoptosis in renal medulla of mice with water deprivation for 6 (A), 12 (B), and 24 hours (C). Few apoptotic cells were evident. (D through F) Apoptosis of RMICs after AMPK activation by AICAR treatment in mice with water restriction for 6 (D), 12 (E), and 24 hours (F). CD, collecting duct; Magnification, 1000×. (G) Quantitative analysis of apoptotic RMICs in renal medulla of mice with AICAR treatment (*P < 0.05; **P < 0.01 versus water deprivation without AICAR treatment, n = 3).

To further validate that the effect of AICAR on RMICs in mice with water restriction was AMPK dependent, we used an additional AMPK activator, A-769662 to examine its in vivo effects on RMIC apoptosis after water deprivation (Supplemental Figure 10). We found that treatment of water-restricted mice with A-769662 for 12 hours resulted in a significant increase in medullary interstitial cell apoptosis.

DISCUSSION

This study demonstrates that hypertonic stress transiently suppresses AMPK activation in RMICs, which is essential for the survival of RMICs in this hostile condition. Activation of AMPK results in significant cell apoptosis of RMICs under hypertonicity in vitro and after water deprivation in mice in vivo. The underlying mechanism may involve the blockade of hypertonicity-elicited NFκB activation and COX-2 expression in RMICs. AMPK may represent a key determinant of the function of the NFκB-COX-2 survival pathway in RMICs. However, whether AMPK is also critical for the survival of other cell types in the renal medulla remains to be determined.

AMPK is an important energy sensor and a master metabolic regulator. AMPK is sensitive to the cellular ratio of AMP to ATP. A high AMP or low ATP level activates AMPK, which inhibits energy-consuming processes and enhances energy-producing processes to restore the energy homeostasis.¹¹ Activation of AMPK ameliorates insulin resistance by enhancing glucose uptake, increasing fatty acid oxidation, and inhibiting fatty acid synthesis in various tissues, including liver, skeletal muscle, and adipose tissue.^23–26 In addition, AMPK has been proposed as a tumor suppressor because of its ability to inhibit cell viability and induce apoptosis by upregulating p53 and p21 expression, activating the caspase cascade and reducing COX-2 expression.^16–18 Because of these favorable effects, AMPK is an attractive therapeutic target for the treatment of type 2 diabetes, metabolic syndrome, and tumors.²⁷

Evidence has emerged that AMPK also has an important role in renal physiology and pathophysiology.²⁸ Several studies have demonstrated that energy sensing by AMPK may be a physiologically relevant mechanism by which renal tubular cells maintain tight coupling between energy metabolism and tubular transport.²⁹ Also, AMPK is renoprotective in attenuating renal hypertrophy in diabetic rats³⁰ and oxidative stress in podocytes.³¹ Thus, AMPK may be involved in energy homeostasis regulation in many cell types in the kidney. In fact, almost all subunits of AMPK can be found in the kidney, and activated AMPK was found strongly expressed at the apical surface of cortical thick ascending limbs and the macula densa.¹⁵ However, the detailed expression profile of AMPK and its activity in each cell type of the kidney is still under investigation. This study demonstrates constitutive expression of AMPK in RMICs, where its activity is essential for determining the survival of RMICs under the hypertonic condition both in vitro and in vivo.

RMICs are normally subjected to hypoxia and hypertonic challenges, both of which stimulate AMPK activation.^14,15 Surprisingly, we found that AMPK phosphorylation in RMICs was transiently increased at 5 minutes, then rapidly decreased by hypertonicity within 12 hours, and then gradually returned to baseline levels, which is in sharp contrast to a previous report suggesting that hyperosmolality activated AMPK in MMDD1 cells, a clonally derived macula densa cell line, within 30 minutes.¹⁵ This inhibitory effect on AMPK activity appears to be specific for RMICs, because hypertonicity significantly increased AMPK activity in IMCD-3 cells, which is consistent with a previous report in MDCK cells.³² The mechanism by which hypertonicity suppresses AMPK phosphorylation in RMICs is currently unknown. One possibility is that hypertonic stress may affect AMPK activity in a cell type-specific manner within the kidney. AMPK activity is regulated by the ratio of AMP to ATP and the activity of upstream kinases/phosphatases including LKB1, CaMKII, and PP2A.¹¹ When RMICs were exposed to hypertonic stress, the ratios of AMP to ATP rapidly decreased after treatment of cells with hypertonicity for 5 minutes, and the reduced AMP/ATP ratios remained for at least 1 hour. However, the expression levels of LKB1, CaMKII, and PP2A remained unchanged in hypertonicity-treated RMICs. These findings suggest that the reduction of the ratio of AMP to ATP, but not the change of upstream AMPK kinase/phosphatase activity, may be responsible for hypertonicity-mediated suppression of AMPK activity in RMICs.

Increasing evidence suggests that hypertonicity can trigger energy-consuming actions of cells including up-regulation of Na⁺/K⁺/ATPase for cell homeostasis and viability,^8–10 which leads to AMP accumulation and ATP depletion. A high AMP or low ATP level activates AMPK, which in turn inhibits energy-consuming processes and enhances energy-producing processes to restore the energy homeostasis. This may explain the initial transient activation of AMPK in hypertonicity-treated RMICs. Strikingly different from many other cell types, including IMCD-3 cells under hyperosmolality challenge,^15,32 RMICs exhibited very transient activation (5 minutes) followed by a marked reduction in AMPK activity over 12 hours under hypertonic stress. This phenomenon is puzzling, and the underlying mechanisms remain unclear. One possibility is that hyperosmolality-elicited prostaglandin production via constitutively expressed COX-2 may prevent further ATP depletion and keep the AMP/ATP ratio at low level in RMICs, because it has been previously reported that nonsteroidal anti-inflammatory drugs and selective COX-2 inhibitors can uncouple mitochondrial respiration and inhibit ATP synthesis.³³ Another possibility is that hyperosomolality-induced ROS production may inhibit AMPK. It has been previously shown that hypertonicity treatment rapidly increased ROS production within 2 minutes in mIMCD-K2 cells.³⁴ Although it is frequently reported that ROS activates AMPK,³⁵ increasing evidence also shows an opposite effect.³⁶ Therefore, it is possible that hypertonicity-induced rapid induction of ROS is associated with AMPK inhibition in RMICs. However, the precise mechanisms by which hyperosmolality suppresses AMPK activity in RMICs warrant further investigation.

In this study, we found that a long period of suppression after an initial transient activation of AMPK activity was critical for the survival of RMICs under the hypertonic condition. Activation of AMPK with AICAR and A-769662 treatment resulted in a marked activation of caspase-3 and massive apoptosis in RMICs in response to hypertonic stress. This effect appeared to be AMPK dependent, because overexpression of a constitutively active form of AMPK by transfection with an adenovirus (Ad-AMPK-CA) led to an identical result. Furthermore, both AICAR treatment and Ad-AMPK-CA transfection significantly blocked COX-2 expression in RMICs, which was critical for the survival of RMICs in hypertonic stress.¹⁹

NFκB activation is well known to be involved in hypertonicity-induced COX-2 expression. Normally, NFκB protein binds to its inhibitory protein IκB and is retained in the cytoplasm as an inactive form. After cells are stimulated by hyperosmotic stress, and IκB is phosphorylated and then degraded via an ubiquitination-dependent mechanism, thus resulting in the release of NFκB from the inactive complex and translocation into the nucleus. Nuclear NFκB finally initiates the transcription of its target genes, including COX-2.³⁷ COX-2 is constitutively expressed in RMICs and exerts its osmoprotective effect mainly through the signaling of two major COX-2-derived bioactive prostanoids, PGI₂ and PGE₂. PGI₂ promotes RMIC viability via activation of its nuclear receptor PPARβ/δ,³⁸ whereas PGE₂ enhances cell survival in parallel with increased expression of osmoprotective genes³⁹ and potentiates tonicity-induced COX-2 expression in a positive feedback loop involving EP₂-cAMP-PKA signaling.⁴⁰ Our finding that AMPK activation inhibited the NFκB-COX-2-PG survival pathway demonstrates that AMPK activity represents a key determinant in the fate of RMICs exposed to hypertonic stress. This conclusion was further supported by the adenoviral overexpression of COX-2 attenuating AICAR-induced RMIC apoptosis under the hyperosmotic condition. In this study, we provided evidence that further supports the critical role of PGI₂/PPARβ/δ system in the survival of RMICs exposed to hypertonicity, because exogenous administration of PGI₂ protected RMICs not only from hypertonic stress-induced cell death but also from AMPK activation-elicited apoptosis.

ROS is usually thought to be destructive, because it could provoke DNA mutation by forming 8-oxoguanine⁴¹ and carbonylate proteins, thereby affecting protein stability and function.⁴² However, in certain types of cells, ROS is reported to play an important role in osmoprotective responses. For example, antioxidants could inhibit high NaCl-induced TonEBP/OREBP transcriptional activity and osmoprotective gene BGT1 mRNA expression⁴³; the antioxidant NAC could inhibit high urea-induced Gadd153/CHOP mRNA and protein level⁴⁴ and high NaCl-induced COX-2 protein level.³⁴ Our results showed that in cultured RMICs, high NaCl-induced ROS production, at least in part, mediated COX-2 induction and osmoprotective gene expression. AMPK activation markedly attenuated hypertonicity-induced ROS production, thus inhibiting the induction of COX-2 in RMICs. AMPK activation also resulted in a reduction in nuclear levels of TonEBP/OREBP and in the expression of osmoprotective genes. Collectively, these findings suggest that activation of AMPK may suppress the cell viability of RMICs by decreasing ROS production and TonEBP/OREBP transcriptional activity.

The effect of AMPK on apoptosis is variable in different situations. Liu et al.⁴⁵ demonstrated that activation of AMPKα1 alleviates endothelial cell apoptosis by increasing the expression of anti-apoptotic proteins Bcl-2 and survivin. However, the study by Zhang et al.⁴⁶ showed that activation of AMPK contributes to apoptosis in glioblastoma cells via p53 activation and mTORC1 inhibition. Therefore, the effect of AMPK on apoptosis appears to be variable under different situations and in different cell types. Our in vitro studies reveal a role for AMPK in attenuating the ability of RMICs to survive hypertonic stress. Consistent with the findings in cultured RMICs, in vivo studies using both AICAR and A-769662 in mice demonstrated that cell viability of RMICs is tightly correlated with AMPK activity after water deprivation. A significant amount of apoptotic cells were observed only in dehydrated animals pretreated with an AMPK activator. Most of the apoptotic cells were RMICs, which further reinforces the link between AMPK activation, COX-2 function, and the survival of RMICs. Interestingly, some apoptotic cells appeared to be AQP-2-positive medullary collecting duct cells. Although the role of AMPK activation in the viability of these cells was not the focus of this study, the presence of relatively high levels of phosphorylated AMPKα suggests that AMPK may play an important role in regulating collecting duct function by sensing the energy stress.¹⁵ Therefore, examining the effect of AMPK activation on collecting duct cell viability under the hypertonic condition would be equally important.

The renal medulla consists of two major cell types: RMICs and medullary collecting duct cells (MCDs). As discussed above, the inhibitory effect of hypertonicity on AMPK appears to be specific for RMICs, because hyperosmolality rapidly activated AMPK in MCDs. Because water deprivation may reduce AMPK in RMICs but increase it in MCDs in the renal medulla, renal medullary AMPK activity remained unchanged after water restriction. However, AICAR-induced massive medullary interstitial cell apoptosis has a functional effect on the kidney. AICAR treatment significantly damaged the renal concentrating ability in mice with water deprivation for 24 hours (Supplemental Table D2). Compared with the control mice, AICAR-treated mice exhibited increased 24-hour urine output and electrolyte (Na⁺, K⁺, and Cl⁻) excretion and reduced urine osmotic pressure. However, the serum osmolality, sodium, creatinine, and urea levels remained unaltered. The reason may be the short period of treatment with AICAR. The mice were treated with AICAR for only 2 days, which may not be sufficient for causing irreversible permanent damage to the renal medulla. In addition, animals with free access to water and food may compensate for the loss of water and sodium in the urine, therefore keeping the serum osmolality and sodium levels unchanged. Because AICAR treatment only damages urinary concentrating ability, we do not anticipate a significant change in GFR and serum creatinine and blood urea nitrogen levels. It would be clinically important to study the long-term effect of AMPK activators, including metformin, on renal function and fluid and sodium homeostasis. Collectively, these observations suggest that AMPK activation results in renal concentrating function change and raise safety concerns for the use of AMPK activators in patients with dehydration.

In conclusion, we report that AMPK activation facilitates RMIC apoptosis under hypertonic stress both in vitro and in vivo. AMPK activation results in reduced cell viability, possibly by blocking the NFκB-COX-2-PG survival pathway and decreasing intracellular osmolyte accumulation by attenuating TonEBP/OREBP transcriptional activity and osmoprotective gene expression under hypertonic stress. These results identify AMPK as a critical factor involved in the maintenance of RMIC viability and suggest that great caution should be exercised in prescribing medications with AMPK-activating properties to patients with dehydration.

CONCISE METHODS

Reagents

AICAR was purchased from Cell Signaling. Antioxidant NAC was obtained from Sigma (St. Louis, MO). All of the cell culture media and supplements were from Sigma. Antibodies against AMPKα, p-AMPKα, caspase-3, cleaved caspase-3, COX-1, COX-2, and TonEBP were from Cell Signaling, Cayman, and Abcam, respectively. Adenoviral vectors encoding a constitutively active AMPKα (Ad-AMPK-CA) and a full-length green fluorescence protein (Ad-GFP) were kindly provided by Prof. Yi Zhu, as described previously.⁴⁷ AMPKα2-CA was created by truncating a full-length Myc epitope–tagged rat AMPKα2 cDNA at residue 312, as described for the AMPKα1 isoform. AMPKα2-CA was amplified by PCR then subcloned into the pAdTrack shuttle vector, and the recombinant adenoviral vector was obtained as described previously.⁴⁸ Adenovirus were propagated in HEK-293 cells and purified by cesium chloride density centrifugation. An adenovirus expressing a full-length rabbit COX-2 (Ad-COX2) and its empty viral vector (Ad-Null) were constructed by SinoGenoMax (Beijing).

Cell Culture

Rabbit RMICs were cultured as previously reported.⁴⁹ Briefly, male Japanese White rabbits (weighing 2 to 2.5 Kg) were anesthetized (7 ml/Kg, 20% ethylcarbamate). The kidneys were removed, and the medulla was dissected and minced under sterile conditions in 6 ml of sterile RPMI 1640 supplemented with 10% FBS. The homogenate was injected subcutaneously in the abdominal wall of a male Japanese White rabbit via a 14-gauge needle. Twenty-five days later, the subcutaneous nodules were removed under sterile conditions, cut into 1-mm fragments, and plated in six-well plates. The cells were cultured in RPMI 1640 medium supplemented with 20% FBS, streptomycin, and penicillin (100 U/ml). These cells were characterized and exhibited abundant Oil-red O-positive lipid droplets (Supplemental Figure 1). Cells at passages 3 to 10 were used. Osmolality of the control medium was 300 mOsm/Kg H₂O, and the hypertonic medium was generated by adding NaCl. If not specifically indicated, the hypertonic medium was prepared by adding NaCl to achieve a tonicity of 500 mOsm/Kg H₂O.

Western Blot Analysis

Cultured RMICs were grown to 80% confluence in six-well plates and then subjected to hypertonic stress for the indicated times in the presence or absence of 1 mM AICAR, an AMPK activator, or a constitutively active AMPKα adenovirus (infecting the cells 36 hours before the hypertonic treatment). The cells were scraped off the plates and lysed in an ice-cold lysis buffer (20 mM Tris-Cl, pH 7.4, 1 mM EDTA, 0.1% SDS, 1% Triton X-100, 0.1 mM Na₃VO₄, and 25 mM NaF). The lysate was centrifuged at 12,000 × g for 5 minutes at 4°C, and the protein content in the supernatant was determined. The supernatant was treated with 2× SDS loading buffer at 95°C for 5 minutes. In total, 40 μg of each protein sample was separated by 12% SDS-PAGE and transferred to nitrocellulose membrane. The membrane was incubated at room temperature for 1 hour in TBS-T (Tris-buffered saline containing 0.1% Tween 20) containing 5% skimmed milk for blocking nonspecific binding sites. The blocked membrane was incubated with primary antibodies at 4°C overnight. The membrane was washed for 5 minutes with TBS-T buffer for 3 to 5 times and then incubated with 1: 5000 horseradish peroxide-conjugated secondary antibody at room temperature for 1 hour. The membrane was washed as above. Finally, the membrane was developed with ECL reagent (Vigorous Biotechnology, Beijing, China) and exposed to Kodak XBT-1 film. Densitometric analysis was performed, and protein expression level was quantified by Image J (National Institutes of Health) software.

Cell Viability

Cell viability was determined by MTT assay as described previously.⁵⁰ Briefly, rabbit RMICs were grown to 80% confluence in 24-well plates and subjected to hypertonic stress for the indicated times in the presence or absence of AICAR (1 mM). Then MTT solution (5 mg/ml) was added to reach a final concentration of 0.5 mg/ml. The cells were cultured for another 4 hours. The medium was removed, and the MTT formazan crystals were dissolved in 1 ml of isopropanol by incubating at room temperature in darkness for 1 hour. The absorbance was measured at 570 nm by use of a microplate reader (model 680; Bio-Rad).

Flow Cytometry of Apoptosis

Cultured RMICs were subjected to hypertonic stress for 12 hours in the presence or absence of 1 mM AICAR or Ad-AMPK-CA (infecting the cells 36 hours before the hypertonic treatment). Phosphatidylserine externalization analysis was performed as described.⁵¹ In brief, after the stress treatment, floating and adherent cells were collected and washed with ice-cold PBS twice, then resuspended in 200 μl of binding buffer (10 mM HEPES, pH 7.4, 140 mM NaCl, 1 mM MgCl₂, 5 mM KCl, and 2.5 mM CaCl₂), and then incubated with a final concentration of 0.5 μg/ml FITC-conjugated annexin V at room temperature in darkness for 15 minutes. The cells were washed, centrifuged, and resuspended in 500 μl of binding buffer. Cellular DNA was stained by the addition of 5 μl of propidium iodide (50 μg/ml) at room temperature for 5 minutes, and the cells were analyzed by FACScan analysis with Cellquest software (Becton Dickinson).

Luciferase Reporter Assays

RMICs were transfected with a NFκB 3×-luc reporter⁵² or a mouse COX-2 promoter (−815)-luc reporter⁵³ with Lipofectamine 2000 (Invitrogen) under conditions recommended by the manufacturer. A β-galactosidase reporter gene (Promega) was cotransfected for normalization. At 24 hours after transfection, the cells were subjected to hypertonic stress for 12 hours in the presence or absence of 1 mM AICAR, and then the cells were washed twice with PBS and lysed with 1× luciferase lysis buffer (luciferase assay kit; Promega), and luciferase activity was quantified in a luminometer (Turner BioSystems). Luciferase levels in individual samples were normalized to β-galactosidase activity.

EMSA

NFκB DNA binding activity was determined by use of a commercial LightShift Chemiluminescent EMSA kit (Pierce), according to the manufacturer's instructions.⁵⁴ In brief, 5 μg of nuclear extract from treated or control rabbit RMICs was incubated with biotin-labeled NFκB consensus oligonucleotide (5′-AGT TGA GGG GAC TTT CCC AGG C-3′) in the reaction buffer at room temperature for 20 minutes and then loaded onto the 6% polyacrylamide gel and transferred to nylon membrane. The membrane was cross-linked at 120 mJ/cm² by use of a commercial UV-light cross-linker instrument equipped with 254-nm bulbs (45- to 60-second exposure with the auto cross-link function). The cross-linked membrane was blocked and incubated with streptavidin-horseradish peroxidase conjugate for 15 minutes and then washed five times (5 minutes each time) with wash buffer. The membrane was incubated with substrate solution for 5 minutes and exposed to Kodak XBT-1 film for 2 to 5 minutes.

Real-Time PCR

Total RNA was extracted from rabbit RMICs, and equal amounts (2 μg) were reverse-transcribed to cDNA. SYBR Green (Invitrogen) was used as fluorochrome according to the manufacturer's instructions. Primers for rabbit AR, SMIT, COX-2, and 18 S rRNA were designed and are listed in Supplemental Table S1. 18 S rRNA was used as an internal control. The PCRs were 94°C for 5 minutes; then 35 cycles of 94°C for 30 seconds, 54 to 56°C for 30 seconds, and 72°C for 30 seconds; and with a final extension at 72°C for 5 minutes. Quantitative values were obtained as threshold PCR cycle number (C_t) when the increase in fluorescence signal of PCR product became an exponential growth. Target gene mRNA level was normalized to that of 18 S rRNA in the same sample as detailed previously.⁵⁵ In brief, the relative expression of the target gene to that of 18 S rRNA was calculated as 2^−ΔC_t, where ΔC_t = C_{t target gene} − C_{t 18 S rRNA}). Each sample was measured in duplicate or triplicate in each experiment. Moreover, melting curves for each PCR product were analyzed to ensure the specificity of the amplification product.

Prostaglandin Measurement

PGE₂ production by cultured RMICs was determined by ELISA with a prostaglandin E2 EIA kit (Cayman). PGI₂ is nonenzymatically hydrated to 6-keto PGF_1α, and 6-keto PGF_1α is commonly measured using a 6-keto PGF_1α EIA kit (Cayman) in cell culture medium as an estimate of PGI₂ synthesis.³⁸ After 12 hours of hypertonic stress (500 mOsm) in the presence or absence of AICAR (1 mM), the medium of cultured cells was collected and centrifuged at 12,000 rpm at 4°C for 5 minutes. The PGE₂ and 6-keto PGF_1α levels in the supernatant were analyzed according to the manufacturer's instructions.

Reactive Oxygen Species (ROS) Production

ROS production was determined by use of a reactive oxygen species assay kit⁵⁶ (Beyotime, China), according to the manufacturer's instructions. In brief, cultured RMICs were subjected to hypertonic stress for the indicated times (30 minutes, 1 hour, 3 hours, and 6 hours), and 2′,7′-dichlorofluorescein diacetate (10 mM) was added in nonserum medium to a final concentration of 10 μM 20 minutes before cell collecting. At the end of the incubation period, the cells were trypsined, washed twice with PBS, and resuspended in 300 μl of PBS. Finally, fluorescence was determined (excitation: 488, emission: 525 nm) in cell suspension by FACScan analysis with Cellquest software (Becton Dickinson).

Animal Studies

To examine the effect of AMPK activation on the survival of RMICs in vivo, male C57BL/6 mice (8 to 10 weeks old) were purchased from the Experimental Animal Center at Peking University Health Science Center (Beijing, China) and treated with AICAR at 0.25 mg/g body weight by intraperitoneal injection for 2 days followed by water deprivation for 6 to 24 hours. The animals were allowed free access to food. After water deprivation, the mice were killed, and the left kidneys were perfused, fixed, and paraffin embedded.

Terminal Deoxynucleotidyl Transferase–mediated Digoxigenin-Deoxyuridine Nick-End Labeling Assay

Medullary apoptotic cells were examined by an in situ cell death detection kit, POD (Roche), according to the manufacturer's instructions.¹⁹ Briefly, after routine dewaxation and rehydration, the tissue sections were pretreated with proteinase K and then incubated with TdT labeling reaction mixture at 37°C for 1 hour. The samples were incubated in the Converter-POD at 37°C for 30 minutes, and apoptotic cells were detected by use of a diaminobenzidine (DAB) kit. The sections were then counterstained with hematoxylin. To identify collecting duct cells, the sections were incubated with an AQP2 antibody (Millipore) at 37°C for 1 hour. After being washed three times with PBS, the sections were incubated with an alkaline phosphatase-conjugated secondary antibody and then stained by use of the 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium kit.

Statistical Analyses

The data are presented as the means ± S.E.M. Differences between the two groups were analyzed by t test by use of Prism4 software. P < 0.05 was considered statistically significant.

DISCLOSURES

None.