Elucidation of cGMP-dependent induction of mitochondrial biogenesis through PKG and p38 MAPK in the kidney

Abstract

Previous studies have shown that cGMP increases mitochondrial biogenesis (MB). Our laboratory has determined that formoterol and LY344864, agonists of the β₂-adrenergic receptor and 5-HT_1F receptor, respectively, signal MB in a soluble guanylyl cyclase (sGC)-dependent manner. However, the pathway between cGMP and MB produced by these pharmacological agents in renal proximal tubule cells (RPTCs) and the kidney has not been determined. In the present study, we showed that treatment of RPTCs with formoterol, LY344864, or riociguat, a sGC stimulator, induces MB through protein kinase G (PKG), a target of cGMP, and p38, an associated downstream target of PKG and a regulator of peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) expression in RPTCs. We also examined if p38 plays a role in PGC-1α phosphorylation in vivo. Administration of l-skepinone, a potent and specific inhibitor of p38α and p38β, to naïve mice inhibited phosphorylated PGC-1α localization in the nuclear fraction of the renal cortex. Taken together, we demonstrated a pathway, sGC/cGMP/PKG/p38/PGC-1α, for pharmacological induction of MB and the importance of p38 in this pathway.

INTRODUCTION

In the presence of nitric oxide (NO), soluble guanylyl cyclase (sGC) produces cGMP from GTP. The resulting cGMP can bind to cGMP-gated ion channels, phosphodiesterases (PDEs), and protein kinase G (PKG) (9). As such, cGMP plays a role in a variety of processes in the cell, including mitochondrial biogenesis (MB) (21, 22).

PKG is a serine/threonine kinase that exists in two forms: PKG1 and PKG2 (14). In renal tubular cells, PKG1 activity and expression decreased when cells were exposed to cisplatin (19). Increasing PKG1 activity protected mitochondrial function and prevented cell apoptosis (19). In brown adipose tissue, natriuretic peptides activate guanylyl cyclase, resulting in activated PKG (1, 5) and the induction of MB (12). Adipocytes exposed to lipoamide also undergo MB through PKG (28).

Activated PKG leads to the phosphorylation of p38 in human adipocytes when they are stimulated with natriuretic peptides (1, 13). Browning et al. (3) showed the importance of activated PKG in NO-induced p38 phosphorylation in 293T fibroblasts. It was also shown in human platelets stimulated by thrombin that p38 activation is necessary for integrin activation and that activated PKG plays an important role in this mechanism (18). These studies showed that activated PKG plays a role in activating p38 in certain cell types.

Peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) is thought to be the master regulator of MB (15, 25). Phosphorylated (p-)p38 can directly phosphorylate PGC-1α at three sites: Thr²⁹⁸, Thr²⁶², and Ser²⁶⁵ (8). Phosphorylation at these sites can increase the stability of PGC-1α, promote its translocation into the nucleus, and induce transcription of PGC-1α target genes (12). Puigserver et al. (24) showed that cultured muscle cells treated with LPS resulted in p38-mediated PGC-1α phosphorylation.

Our laboratory showed that maximal mitochondrial respiration (i.e., uncoupled respiration), a marker of MB, increases in renal proximal tubule cells (RPTCs) treated with the membrane-soluble cGMP analog 8-Br-cGMP but not 8-Br-cAMP, suggesting that cGMP is responsible for inducing MB rather than cAMP in RPTCs (31). Moreover, we have shown that cGMP is a key player in inducing MB by two G protein-coupled receptor agonists. Formoterol, a β₂-adrenergic receptor agonist, and LY344864, a 5-HT_1F receptor agonist, induced MB through the G_β/γ subunit, Akt, sGC, and cGMP (4, 10). Our studies are consistent with reports for other G protein-coupled receptors, such as the cannabinoid type 1 receptor, inducing MB through increased NO production (30).

However, the pathway from cGMP to MB after pharmacological stimulation in highly oxidative renal epithelial cells is not clear. We propose that formoterol, LY344864, and riociguat, a sGC stimulator that increases cGMP, activate PKG, which leads to the phosphorylation of p38. In turn, p38 phosphorylates PGC-1α to facilitate translocation to the nucleus and produce MB.

MATERIALS AND METHODS

Reagents.

l-Skepinone was purchased from Selleckchem (Houston, TX). LY344864 and KT5823 were purchased from Tocris (Minneapolis, MN). Riociguat was purchased from Biovision (Milpitas, CA). Formoterol fumarate was purchased from Sigma (St. Louis, MO).

In vitro experiments.

RPTCs were isolated from female New Zealand White rabbit (2 kg) kidneys using the iron oxide perfusion method (23). Cells were plated and grown on 35-mm tissue culture dishes under conditions similar to physiological conditions in vivo. Confluent RPTCs were treated with riociguat (10 μM), LY344864 (10 nM), formoterol (30 nM), or vehicle consisting of DMSO (<0.5%). For inhibitor experiments, RPTCs were pretreated with 100 nM KT5823 or 100 nM l-skepinone for 30 min. Riociguat, formoterol, LY344864, or DMSO was added to RPTCs, and RPTCs were incubated for 2 h and harvested for further analysis. The concentrations of formoterol and LY344864 have previously be shown to induce MB in RPTCs (4, 10).

In vivo experiments.

Skepinone was dissolved in DMSO and diluted in sterile saline. The final DMSO concentration was 2%. Eight- to nine-week old male C57BL/6 mice (20–25 g, Charles River Laboratories) were injected intraperitoneally with skepinone at 1, 3, or 10 mg/kg or vehicle. After 6 h, mice were euthanized and kidneys were removed. Part of the kidney was flash frozen or processed for subcellular fractionation. Animal experiments and animal use was approved by the Institutional Animal Care and Use Committee of the University of Arizona.

Subcellular fractionation.

RPTCs were harvested in sucrose isolation buffer containing 250 mM sucrose, 1 mM EGTA, 10 mM HEPES, and 1 mg/ml fatty acid-free BSA at a pH of 7.4. Cells were homogenized using a dounce homogenizer and centrifuged at 700 g for 5 min. The cytosolic supernatant was stored in phosphatase inhibitors (1:100), 1 mM sodium orthovanadate, and 1 mM sodium fluoride, and Triton X-100 and SDS at 4%. The pellet was washed twice in isolation buffer and centrifuged at 1,000 g for 5 min. The pellet/nuclear fraction was resuspended in RIPA buffer containing 50 mM Tris·HCl, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, and 1% Triton X-100 (pH 7.4) with phosphatase inhibitors (1:100), 1 mM sodium orthovanadate, and 1 mM sodium fluoride. The purity of the cytosolic and nuclear fractions was determined using immunoblot analysis. Histone H3 and/or lamin B1 were used as nuclear markers, and α-tubulin was used as a cytosolic marker. Antibodies for histone H3 were purchased from Cell Signaling Technology (Danvers, MA). Antibodies for lamin B1 and α-tubulin were purchased from Abcam (Cambridge, MA).

Immunoblot analysis.

RPTCs or the snapped frozen mouse kidney cortex was added to RIPA buffer. Cells were sonicated for ~10 s and centrifuged at 7,500 g for 5 min at 4°C. Supernatants were removed, and protein was measured using a BCA assay. Equal amounts of protein were loaded onto 4–15% SDS-PAGE gels and separated by gel electrophoresis. Proteins were transferred onto nitrocellulose membranes and blocked in 5% milk or 5% BSA dissolved in Tris-buffered saline with Tween 20 (TBST). Membranes were incubated with primary antibodies overnight. Membranes were washed in TBST three times for 5 min, incubated with horseradish peroxidase-conjugated secondary antibody, and visualized using enhanced chemiluminescence (Thermo Scientific) and GE ImageQuant LAS4000 (GE Life Sciences). Optical density was determined using ImageJ software (National Institutes of Health). Primary antibodies [p-vasodilator-stimulated phosphoprotein (VASP) Ser²³⁹ (1:1,000), p-p38 MAPK (Thr¹⁸⁰/Tyr¹⁸²) (1:1,000), and p38 MAPK (1:1,000)] were purchased from Cell Signaling Technology.

Immunoprecipitation.

The Dynabeads Protein G Immunoprecipitation Kit (ThermoFisher Scientific) was used to immunoprecipitate PGC-1α in cytosolic and nuclear fractions. Proteins (200 µg) from cytosolic and nuclear fractions were precleared with Pierce Protein A/G Plus Agarose beads (ThermoFisher Scientific) for 2 h and then centrifuged at 14,000 g for 10 min at 4°C. Dynabeads magnetic beads were incubated with 10 µg PGC-1α antibody for 4 h at room temperature. Supernatants from the precleared cytosolic and nuclear fractions were added to the magnetic beads and PGC-1α antibody and incubated with rotation overnight at 4°C. PGC-1α was immunoprecipitated and eluted (denaturing elution) based on the manufacturer’s instructions. The resulting supernatant was directly loaded on to a 4–15% SDS-PAGE gel and separated by gel electrophoresis. After protein transfer, one nitrocellulose membrane was blocked in 5% milk to measure total PGC-1α and the other was blocked in 5% BSA to measure phosphorylated serine/threonine residues for 1 h. Membranes were incubated with phosphoserine/threonine (1:1,000) antibody from Abcam or PGC-1α antibody (1:1,000) from EMD Millipore (Billerica, MA) overnight at 4°C. After membranes had been washed in TBST three times, they were incubated in horseradish peroxidase-conjugated secondary antibodies from Abcam for 2 h at room temperature. Membranes were visualized as described above.

Statistics.

Data are presented as means ± SE. RPTCs isolated from a single rabbit represent n = 1. For a single comparison, Student’s t test was performed. Multiple comparisons were subjected to one-way ANOVA and Tukey’s post hoc test with P < 0.05 being statistically significant between means.

RESULTS

Riociguat activates PKG and is blocked by KT5823.

To elucidate the signaling pathway from sGC to nuclear PGC-1α phosphorylation, RPTCs were treated with riociguat (10 μM), a sGC stimulator that targets the reduced form of sGC (2, 16, 20, 29). The reduced form of sGC predominates in naïve cell types and increases cGMP directly. cGMP production can activate PKG, and PKG activation was determined by measuring the phosphorylation of VASP at Ser²³⁹, a specific target of PKG (6).

RPTCs were treated with riociguat for 2 h, resulting in a 1.3-fold increase in VASP phosphorylation compared with controls (Fig. 1). Pretreatment with KT5823, a PKG inhibitor, for 30 min and subsequent treatment with vehicle or riociguat for 2 h resulted in a decrease in VASP phosphorylation to control levels, demonstrating that KT5823 inhibits PKG. Pretreatment with KT5823 alone lowered phosphorylation of VASP in RPTCs to below control levels.

Fig. 1.

Riociguat phosphorylates vasodilator-stimulated phosphoprotein (VASP) and is blocked by KT5823 (KT). A: representative immunoblot for phosphorylated (p-)VASP and tubulin after treatment. B: densitometric analysis for p-VASP protein. Data are presented as means ± SE; n = 6–7. *Significance compared with vehicle (Veh) treatment (P < 0.05).

Riociguat activates p38 through PKG.

We hypothesized that p38 is the mediator between PKG activation and PGC-1α phosphorylation in RPTCs. Riociguat treatment of RPTCs increased p-p38 2.1-fold at 2 h. Pretreatment with KT5823 for 30 min with subsequent treatment with vehicle or riociguat for 2 h resulted in a decrease in p-p38 to control levels (Fig. 2, A and B).

Fig. 2.

KT5823 (KT) and skepinone (SK) inhibit phosphorylation of p38. A and C: representative blots for phosphorylated (p-)p38, total p38, and tubulin after treatment. B and D: densitometric analysis for p-p38 protein. Data are presented as means ± SE; n = 6–7. *Significance compared with vehicle (Veh) treatment (P < 0.05); #significance compared with riociguat treatment (P < 0.05).

To validate skepinone inhibition of p38 phosphorylation at 2 h in the presence and absence of riociguat, RPTCs were treated with skepinone for 30 min and then treated with vehicle or riociguat for 2 h. Skepinone inhibited p-p38 when RPTCs were treated with riociguat (90%) or vehicle (80%) at 2 h (Fig. 2, C and D).

LY344864 and formoterol activate p38.

Based on the above results, we examined if p38 plays a role in the signaling pathways of formoterol and LY344864. We treated RPTCs with vehicle, formoterol, or LY344864 for 2 h. p38 phosphorylation increased 1.6-fold and 1.9-fold when RPTCs were exposed to formoterol and LY344864, respectively (Fig. 3). RPTCs were pretreated with skepinone for 30 min and exposed to vehicle, formoterol, and LY344864 for 2 h. Skepinone alone inhibited p-p38 by 70% compared with vehicle control. Skepinone pretreatment inhibited p-p38 by 80% after formoterol and LY344864 treatment.

Fig. 3.

Skepinone (SK) inhibits LY344864 (LY)- and formoterol (F)-mediated p38 phosphorylation. A: representative blots for phosphorylated (p-)p38, total p38, and tubulin after treatment. B: densitometric analysis for p-p38 protein. Data are presented as means ± SE; n = 8. *Significance compared with vehicle (Veh) treatment (P < 0.05); #significance compared with LY344864 treatment; $significance compared with formoterol treatment.

Pretreatment with KT5823 or skepinone decreases nuclear p-PGC-1α at serine and threonine residues in the presence of riociguat.

Phosphorylation of PGC-1α at serine and threonine sites by p38 can increase the stability of PGC-1α, resulting in its translocation to the nucleus and transcription of PGC-1α. RPTCs were pretreated with KT5823 for 30 min and then exposed to vehicle or riociguat for 2 h. RPTCs were subjected to subcellular fractionation and tested for the purity of nuclear and cytosolic fractions by measuring α-tubulin, a cytosolic marker, and lamin B, a nuclear marker (7, 11). PGC-1α was immunoprecipitated from nuclear and cytosolic fractions and immunoblotted for phosphorylated serine and threonine residues (Fig. 4, A and C). p-PGC-1α in the nuclear fraction was increased by 1.64-fold after riociguat treatment (Fig. 4B) and decreased to control levels in the presence of KT5823. There were no changes in PGC-1α phosphorylation in the cytosolic fraction compared with vehicle control (Fig. 4, C and D).

Fig. 4.

Pretreatment with KT5823 (KT) decreases nuclear phosphorylated (p-)peroxisome proliferator-activated receptor-γ coactivator (PGC)-1α in the presence of riociguat. A and C: phosphorylated serine and threonine residues were measured after immunoprecipitation (IP) of PGC-1α by immunoblot (IB) analysis in nuclear and cytosol fractions after 30 min of pretreatment with vehicle (Veh) or KT5823 followed by exposure to DMSO and riociguat for 2 h. Total PGC-1α expression was measured after immunoprecipitation. B and D: densitometric analysis for phosphorylated serine and threonine residues in the nuclear and cytosolic fractions. Data are presented as means ± SE; n = 6. *Significance compared with Veh control treatment (P < 0.05); #significance compared with riociguat treatment.

RPTCs were pretreated with skepinone for 30 min and then exposed to vehicle or riociguat for 2 h. p-PGC-1α was measured in nuclear and cytosolic fractions. Skepinone alone had no effect on p-PGC-1α in the nucleus (Fig. 5, A and B). Skepinone decreased p-PGC-1α in the nucleus to control levels when RPTCs were exposed to riociguat. p-PGC-1α did not change in the cytosolic fraction compared with vehicle control (Fig. 5, C and D).

Fig. 5.

Pretreatment with skepinone (SK) decreases nuclear phosphorylated (p-)peroxisome proliferator-activated receptor-γ coactivator (PGC)-1α at serine and threonine sites in the nucleus. A and C: phosphorylated serine and threonine residues were measured after immunoprecipitation (IP) of PGC-1α by immunoblot (IB) analysis after 30 min of treatment with DMSO or 100 nM skepinone followed by exposure to DMSO and riociguat (Rio) for 2 h in the cytosol and nuclear fractions. Total PGC-1α expression was measured after immunoprecipitation. B and D: densitometric analysis for phosphorylated serine and threonine residues in the cytosol and nuclear fractions. Data are presented as means ± SE; n = 6–7. *Significance compared with vehicle (Veh) control treatment (P < 0.05); #significance compared with riociguat.

Skepinone decreases nuclear p-PGC-1α in the renal cortex.

Based on the above cellular data, we determined if p38 inhibition alters p-PGC-1α in the kidney. We performed a dose-response experiment using skepinone and measured p38 phosphorylation in the renal cortex. Naïve mice were injected with 1, 3, and 10 mg/kg skepinone and then euthanized 6 h after injection. At all three doses, skepinone equally blocked p38 phosphorylation by >50% (Fig. 6, A and B). Using the lowest dose of skepinone (1 mg/kg), the number of mice was increased (n = 6–7; Fig. 6C), and skepinone decreased p38 phosphorylation by 75% (Fig. 6D).

Fig. 6.

Skepinone (SK) decreases nuclear phosphorylated (p-)peroxisome proliferator-activated receptor-γ coactivator (PGC)-1α in the renal cortex. A: representative blots for p-p38, total p38, and tubulin in naïve mice 6 h after treatment with 1, 3, or 10 mg/kg skepinone. C: representative blots for p-p38, total p38, and tubulin with 1 mg/kg skepinone. B and D: densitometric analysis for p38 protein. Data are presented as means ± SE; n = 3–7. E: representative blots for tubulin and lamin B1 in the nuclear and cytosol fractions. F and H: phosphorylated serine and threonine residues were measured after immunoprecipitation (IP) of PGC-1α in the cytosol and nuclear fractions by immunoblot (IB) analysis. G and I: densitometry analysis for phosphorylated serine and threonine residues in the cytosol and nuclear fractions. Data are presented as means ± SE; n = 7. *Significance compared with vehicle (Veh) treatment (P < 0.05). The black lines in the center of the blots are used for dividing experimental groups only and do not alter the information contained therein.

The next experiment measured p-PGC-1α in cytosolic and nuclear fractions of the renal cortex (Fig. 6, F and H). Mice were treated with skepinone (1 mg/kg) for 6 h, and the renal cortex was subcellular fractionated. The purity of cytosolic and nuclear fractions was tested (Fig. 6E). In the cytosolic fraction, skepinone had no affect on p-PGC-1α; however, in the nuclear fraction, skepinone decreased p-PGC-1α by 50% (Fig. 6, G and I).

DISCUSSION

Previous studies from our laboratory have shown that cGMP, not cAMP, induces MB in RPTCs (31). PDE3 inhibitors such as cilostamide and trequinsin, compounds that prevent the degradation of cGMP, also induce MB in RPTCs (31). In vivo, sildenafil, a PDE5 inhibitor, also induces MB in the renal cortex of naïve mice (31). However, the mechanisms by which cGMP can induce MB are still under investigation for different cell types.

We have shown the induction of MB in RPTCs treated with formoterol, a β₂-adrenergic receptor agonist, and LY344864, a 5-HT_1F agonist. Moreover, we have reported that sGC and cGMP are important components for signaling MB in RPTCs (4, 10). In the present study, we propose a signaling pathway between sGC and MB, sGC/cGMP/PKG/p38/PGC-1α, in RPTCs. In addition, we pharmacologically targeted sGC by administering riociguat, a sGC stimulator that targets the reduced/heme-dependent form (2, 16, 20, 29).

To elucidate the role of p38 in the sGC-dependent induction of MB, skepinone, a potent and selective ATP-competitive inhibitor for p38α and p38β inhibitor, was used (17). Previous studies have shown that p38 can directly phosphorylate PGC-1α at Thr²⁹⁸, Thr²⁶², and Ser²⁶⁵ in the cytosol, causing its translocation to the nucleus to sustain PGC-1α transcription and therefore MB (24). We show that inhibition of PKG activation and p38 phosphorylation decreased nuclear p-PGC-1α when RPTCs were exposed to the pharmacological compounds riociguat, formoterol, and LY344864, demonstrating that PKG activation and p38 are needed in the pathway from sGC to MB. Further support for this pathway is that administration of skepinone in vivo decreased nuclear p-PGC-1α in the renal cortex.

Previous literature shows that PKG activation can lead to p38 phosphorylation and that p38 is a downstream associated target of PKG (1, 3, 18, 26). Based on this literature, we did not explore the effect of skepinone on PKG activation in the presence of riociguat. Although the literature provides evidence for other downstream effectors such as ERK1/2, glycogen synthase kinase-3β, and Akt, we chose to focus on p38 for its role in NO signaling, cGMP/PKG-dependent signaling, and regulation of PGC-1α (3, 8, 18, 26, 27).

In summary, we have elucidated a pathway for sGC-dependent induction of MB by investigating the importance of PKG and subsequent activation of p38 leading to an increase in nuclear p-PGC-1α by pharmacological compounds. This study supports our previous work with potent inducers of MB and shows the importance of sGC/cGMP in the signaling pathways leading to the induction of MB in the kidney.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

P.B., J.J., and R.G.S. conceived and designed research; P.B. and J.J. analyzed data; P.B., J.J., and R.G.S. interpreted results of experiments; P.B. and J.J. prepared figures; P.B. drafted manuscript; P.B., J.J., and R.G.S. edited and revised manuscript; J.J. performed experiments; R.G.S. approved final version of manuscript.