Identification of dual mechanisms mediating 5-hydroxytryptamine receptor 1F-induced mitochondrial biogenesis

Abstract

Our laboratory recently made the novel observation that 5-hydroxytryptamine 1F (5-HT_1F) receptor activation induces mitochondrial biogenesis (MB), the production of new, functional mitochondria, in vitro and in vivo. We sought to determine the mechanism linking the 5-HT_1F receptor to MB in renal proximal tubule cells. Using LY344864, a selective 5-HT_1F receptor agonist, we determined that the 5-HT_1F receptor is coupled to Gα_i/o and induces MB through Gβγ-dependent activation of Akt, endothelial nitric oxide synthase (eNOS), cyclic guanosine-monophosphate (cGMP), protein kinase G (PKG), and peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α). We also report that the 5-HT_1F receptor signals through a second, Gβγ-dependent pathway that is linked by Akt phosphorylation of Raf. In contrast to the activated Akt pathway, Raf phosphorylation reduced extracellular signal regulated kinases (ERK1/2) and foxhead box O3a (FOXO3a) phosphorylation, suppressing an inhibitory MB pathway. These results demonstrate that the 5-HT_1F receptor regulates MB through Gβγ-dependent dual mechanisms that activate a stimulatory MB pathway, Akt/eNOS/cGMP/PKG/PGC-1α, while simultaneously repressing an inhibitory MB pathway, Raf/MEK/ERK/FOXO3a. Novel mechanisms of MB provide the foundation for new chemicals that induce MB to treat acute and chronic organ injuries.

INTRODUCTION

In addition to its roles as a hormone and neurotransmitter in the central nervous system, 5-HT (serotonin, 5-hydroxytryptamine) mediates vascular contraction and relaxation, gastrointestinal motility, apoptosis, and platelet aggregation through peripheral receptors (5, 17, 22, 35, 48, 54). The biological roles of 5-HT are mediated by a family of G protein-coupled receptors (GPCR). Recently, our group revealed a novel role for 5-HT receptors in mitochondrial biogenesis (MB) or generation of new, functional mitochondria (14, 19, 42).

MB is an attractive target for pharmacological intervention following acute organ injuries such as ischemia-reperfusion (IR) injury (11, 55). Rapid and persistent loss of mitochondrial homeostasis is a major contributor to the pathology of IR-induced renal injury, and IR suppresses peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α), the master regulator of MB, and its downstream targets (11, 55). Increasing PGC-1α promotes the transcription of genes necessary for mitochondrial function during repair and restoration following oxidant injury in renal proximal tubule cells (RPTCs) and IR-induced acute kidney injury (AKI) (21, 41, 43, 57).

Despite the promise of MB as a therapeutic target, few nontoxic pharmacological agents stimulate PGC-1α expression and activity. Our laboratory developed a phenotypic assay to measure MB and identified several pharmacological targets that activate PGC-1α and induce MB, including the 5-HT_1F receptor (3, 14). The selective 5-HT_1F receptor agonist LY344864 (LY) is a potent and efficacious inducer of MB in vitro and in vivo as demonstrated by increased electron transport chain gene and protein levels (14). LY exhibits high affinity for the 5-HT_1F receptor with a reported pK_d of 8.2 and is ~100 times more selective for the 5-HT_1F receptor compared with 5-HT_1A,_1B,_1D,_1E (59). Additionally, LY-induced 5-HT_1F receptor activation in a mouse model of IR-AKI restored mitochondrial DNA (mtDNA) copy number and accelerated recovery of renal function (14), providing evidence that the 5-HT_1F receptor is a good therapeutic target to stimulate MB and promote recovery from acute organ failure. While we have identified a novel role for this receptor, however, the signaling mechanism of 5-HT_1F receptor-induced MB remains unknown. In this article, we identify the signaling pathways responsible for MB by 5-HT_1F receptor activation.

METHODS

Reagents.

LY344864, pertussis toxin, gallein, LY294002, nitro-l-arginine methyl ester (l-NAME), 1H-[1,2,4]oxadiazolo[4,3-α]quinoxalin-1-one (ODQ), and KT5823 were purchased from Tocris (Ellisville, MO). GDC0068 was purchased from Selleckchem (Houston, TX).

Isolation and culture of RPTCs and oxygen consumption.

Female White New Zealand rabbits (1.5–2.0 kg) were purchased from Charles River Laboratories (Wilmington, MA). RPTCs were isolated using the iron perfusion method previously described (38). RPTCs were plated and cultured in 96-well respiratory plates or 35-mm dishes in media previously described (7). Experiments were performed on the third or sixth day after plating when cells had formed a confluent monolayer. The oxygen consumption rate (OCR) of RPTCs was measured using the Seahorse Bioscience XF-96 Extracellular Flux Analyzer as previously described (3). Each 96-well assay plate was treated with vehicle (DMSO; <0.5%) or the experimental compounds. Basal OCR was measured before injection of carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP; 0.5 μM) to measure the uncoupled OCR (FCCP-OCR), a marker of MB (3). All studies conducted were approved by the Institutional Animal Care and Use Committee at the Medical University of South Carolina and the University of Arizona.

Analysis of mtDNA content.

MtDNA content was determined by quantitative real-time PCR analysis. Total DNA was isolated from RPTCs using the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA) as described in the manufacturer’s protocol. Extracted DNA was quantified and 5 ng was used for PCR. Relative mtDNA content was assessed by the mitochondrial encoded NADH dehydrogenase 1 (ND1) and was normalized to nuclear-encoded β-actin. Primer sequences for ND1 and β-actin were as follows: ND1 sense: 5′-TAGAACGCAAAATCTTAGGG-3′; ND1 antisense: 5′-TGCTAGTGTGAGTGATAGGG-3′; β-actin sense: 5′-GGGATGTTTGCTCCAACCAA-3′; and β-actin antisense: 5′-GCGCTTTTGACTCAGGATTTAA-3′.

Protein isolation, immunoblot analysis, and immunoprecipitation.

Protein isolation and immunoblot analysis were performed as previously described (7). For immunoprecipitation experiments, protein (500 μg) was precleared by being incubated with Pierce Protein A/G Plus Agarose beads for 30 min and then centrifuged at 14,000 g for 10 min at 4°C. The supernatant was collected and incubated with an anti-PGC-1α antibody (3 μg) overnight at 4°C. Pierce Protein A/G Plus Agarose beads were washed and incubated with PGC-1α-protein lysates for 2 h at 4°C. Lysates were then washed in immunoprecipitation buffer (25 mM Tris and 150 mM NaCl, pH 7.4) followed by centrifugation at 2,000 g for 1 min. Laemmli buffer and β-mercaptoethanol (1:50) were added to collected supernatant and incubated at 95°C for 5 min. Following a brief centrifugation, the supernatant was collected and analyzed by immunoblotting using antibodies against phosphoserine/threonine (1:1,000) from Abcam (Cambridge, MA) and PGC-1α (1:1,000) from EMD Millipore (Billerica, MA). Primary antibodies p-AKT Ser473 (1:1,000), total AKT (1:1,000), p-endothelial nitric oxide synthase (eNOS) Ser1177 (1:500), p-vasodilator-stimulated phosphoprotein (VASP) Ser239 (1:1,000), total VASP (1:1,000), p-ERK p44/42 (1:1,000), total ERK (1:1,000), p-foxhead box O3a (FOXO3a) Ser294, and total FOXO3a were purchased from Cell Signaling Technology (Danvers, MA), total eNOS (1:1,000) was purchased from Abcam, and GAPDH (1:10,000) from Fitzgerald (Acton, MA). Secondary antibodies include horseradish peroxidase-labeled anti-rabbit and mouse and were from Abcam (Cambridge, MA).

cAMP and cGMP enzyme-linked immunosorbent assay.

RPTCs in 35-mm dishes were treated with DMSO or LY for 1 h. RPTCs were then harvested according to the manufacturer’s protocol, and cAMP or cGMP levels were measured using an ELISA kit (Cayman Chemical, Ann Arbor, MI). cAMP and cGMP values (pmol) were normalized to protein (mg) as quantified by a bicinchoninic acid assay followed by normalization to vehicle control for each biological replicate.

Statistical analysis.

Data are presented as means ± SE. Single comparisons were performed using Student’s t-test. Multiple comparisons were subjected to one-way ANOVA followed by Tukey’s post hoc test, with P < 0.05 considered to be a statistically significant difference between means. In the figures, different superscripts indicate statistical differences. RPTCs isolated from a single rabbit represented an individual experiment (n = 1) and were repeated until n = 4–5 were obtained.

RESULTS

5-HT_1F receptor agonism decreases cAMP formation and induces FCCP-OCR and mtDNA copy number in a Gβγ-Akt-NOS-PKG-dependent manner.

Because the 5-HT_1F receptor has been reported to be negatively coupled to adenylyl cyclase via Gα_i/o (2), we first sought to explore the roles of the Gα_i/o and Gβγ in RPTCs (23). To verify that the renal 5-HT_1F receptor is a Gα_i-coupled GPCR, we measured cAMP in LY-treated RPTCs in the presence and absence of pertussis toxin (PTX). PTX catalyzes the ADP-ribosylation of the Gα_i/o subunits locking the α-subunits into an inactive state and inhibiting adenylate cyclase activity, leading to increased cellular cAMP (30). RPTCs were pretreated with 100 ng/ml PTX or DMSO for 24 h, followed by a 20-min exposure to 10 nM LY or DMSO. A 10-nM concentration of LY was previously determined to induce MB in RPTCs (14). Treatment with PTX alone increased cAMP levels (Fig. 1A). LY reduced cAMP formation by 70%, and this reduction was blocked by PTX pretreatment (Fig. 1A), verifying that LY-induced 5-HT_1F receptor signaling in RPTCs is mediated by Gα_i/o, PTX-sensitive G proteins.

Fig. 1.

5-HT_1F receptor agonism decreases cAMP formation and induces FCCP-uncoupled oxygen consumption rate (OCR) and mitochondrial (mt) DNA copy number in a Gβγ-Akt-NOS-PKG-dependent manner. A: cAMP levels were measured by ELISA after a 20-min treatment with <0.5% DMSO or 10 nM LY344864 or a 24-h pretreatment of pertussis toxin (PTX; 100 ng/ml) followed by DMSO or LY. B: renal proximal tubule cells (RTPCs) were pretreated with DMSO or the pharmacological inhibitors gallein (Gal; 100 nM), GDC0068 (GDC; 100 nM), nitro-l-arginine methyl ester (l-NAME; 10 µM), 1H-[1,2,4]oxadiazolo[4,3-α]quinoxalin-1-one (ODQ; 100; nM), or KT5823 (KT; 100 nM) for 30 min. DMSO or LY was then added, and FCCP-OCR uncoupled mitochondrial respiration was measured using Seahorse XF 96 analyzer 24 h later. C: RTPCs were pretreated with DMSO or the pharmacological inhibitors Gal, GDC, l-NAME, ODQ, or KT for 30 min followed by 24-h exposure of DMSO or LY and mtDNA copy number was assessed. Data are reported as means ± SE; n = 4–5. Bar with different superscripts are significantly different from one another (P < 0.05).

Our laboratory developed a high-throughput screening assay to assess MB by measuring FCCP-OCR in RPTCs, a marker of MB (3). Previous studies demonstrated that LY-induced FCCP-OCR resulted from increased mtDNA number and electron transport chain proteins at 24 h (14). To understand the signaling pathways leading to increased FCCP-OCR, inhibitors of Gβγ and other pathways were analyzed for their ability to block LY-induced FCCP-OCR. RPTCs were pretreated for 30 min with DMSO or gallein, GDC0068 (GDC), l-NAME, ODQ, and KT5823 (KT), inhibitors of Gβγ, Akt, NOS, soluble guanylyl cyclase, and PKG, respectively (13, 15, 24, 26), and then exposed to DMSO or LY for 24 h. Pretreatment with these inhibitors prevented LY-induced FCCP-OCR (Fig. 1B). In addition, pretreatment with gallein, GDC, l-NAME, ODQ, and KT prevented LY-induced mtDNA copy number (Fig. 1C), suggesting that Gβγ, Akt, NOS, cGMP, and PKG are key regulators of the 5-HT_1F receptor-induced MB pathway.

LY-induced Akt and eNOS phosphorylation is blocked by inhibitors of Gβγ, phosphoinositide 3-kinase, and Akt.

To examine Akt phosphorylation following LY treatment, RPTCs were treated with LY or DMSO for 15 min, 30 min, 1 h, 2 h, and 3 h. Immunoblot analyses revealed a twofold increase in p-Akt at Ser473 at 15 min and a 1.5-fold increase at 30 min (Fig. 2, A and B). RPTCs were pretreated with gallein (100 nM, 30 min) and exposed to LY for 15 min. Immunoblot analyses revealed that gallein inhibited LY-induced Akt phosphorylation (Fig. 2, C and D). Because Akt is a substrate of phosphoinositide 3-kinase (PI3K), RPTCs were pretreated with the PI3K inhibitor LY294002 (10 μM, 30 min) and then treated with LY. Blockade of PI3K inhibited Akt phosphorylation after treatment (Fig. 2, E and F). Next, we examined eNOS phosphorylation at Ser1177, a direct phosphorylation target of Akt. RPTCs were treated with DMSO or LY for 1, 2, 3, and 4 h. eNOS phosphorylation was elevated 2.5- and 3-fold at 1 and 2 h, respectively (Fig. 2, G and H). To confirm that Gβγ and Akt are upstream of eNOS phosphorylation, RPTCs were pretreated with gallein and GDC (100 nM, 30 min), then exposed to LY for 1 h (Fig. 2, I and L). Inhibiting Gβγ and Akt also prevented LY-induced eNOS phosphorylation. These data demonstrate that LY activates the Gβγ-PI3K-Akt-eNOS pathway.

Fig. 2.

LY induced Akt and endothelial nitric oxide synthase (eNOS) phosphorylation is blocked by inhibitors of Gβγ, phosphoinositide 3-kinase (PI3K), and Akt. Phosphorylated Akt (Ser473) was measured by immunoblot analysis 15 min, 30 min, 1 h, or 3 h after treatment with DMSO or LY (A and B), after 30-min pretreatment with DMSO or gallein (Gal; 100 nM) followed by exposure to DMSO or LY for 15 min (C and D), and after 30-min pretreatment with DMSO or LY29004 (LY29; 10 µM) followed by exposure to DMSO or LY for 15 min (E and F). Phosphorylated eNOS (Ser1177) was measured by immunoblot analysis after 1, 2, 3, or 4 h treatment with DMSO or LY (G and H), after 30-min treatment with DMSO or Gal (100 nM) followed by exposure to DMSO or LY for 1 h (I and J), and after 30-min pretreatment with DMSO or GDC0068 (GDC; 100 nM) followed by exposure to DMSO or LY for 1 h (K and L). Data are reported as means ± SE; n ≥ 5. Bars with different superscripts are significantly different from one another (P < 0.05).

LY increases cGMP formation and cGMP- and PKG-dependent VASP phosphorylation.

Numerous reports demonstrate that increased PGC-1α expression and other markers of MB in tissues, including the kidney, are elevated through an increase in cGMP production (36, 37, 58). To determine if cGMP is involved in LY-induced MB, we measured cGMP following LY treatment. RPTCs were treated with the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (100 μM) and DMSO or LY. LY elicited a transient 1.6-fold increase in cGMP at 1 h that returned to baseline levels at 2 h (Fig. 3A). The formation of cGMP induces a conformational change in cGMP-dependent PKG allowing activation of this serine/threonine protein kinase and the phosphorylation of substrate proteins (1). Vasodilator-stimulated phosphoprotein (VASP), a marker of PKG activation, was assessed to determine the role of PKG in LY-induced MB signaling (53). RPTCs expressed the PKG splice variant PKGI, which preferentially phosphorylates VASP at Ser239 (10). LY increased VASP phosphorylation 1.8- and 2-fold at 1and 2 h, respectively (Fig. 3, B and C). To determine if cGMP production and PKG activation are responsible for increased VASP phosphorylation, RPTCs were pretreated with ODQ (100 nM, 30 min) or KT (100 nM, 30 min) and then exposed to LY for 1 h. LY-induced p-VASP upregulation was inhibited by blockade of cGMP production (ODQ) and PKG activation (KT) (Fig. 3, D and G). Taken together these findings reveal agonist stimulation of the 5-HT_1F receptor mediates Gβγ-PI3K-AKT-eNOS-cGMP-PKG-VASP signaling.

Fig. 3.

LY increases cGMP formation and cGMP- and PKG-dependent VASP phosphorylation. A; cGMP levels were measured by ELISA after 1-h treatment with DMSO or LY in the presence of 3-isobutyl-1-methylxanthine (IBMX; 100 µM). Phosphorylated VASP (Ser239) was measured by immunoblot analysis 1 or 2 h after treatment with DMSO or LY (B and C), after 30-min treatment with DMSO or ODQ (100 nM) followed by exposure to DMSO or LY for 1 h (D and E), and after 30-min treatment with DMSO or KT5923 (KT, 100 nM) followed by exposure to DMSO or LY for 1 h (F and G). Data are reported as means ± SE; n ≥ 5. Bars with different superscripts are significantly different from one another (P < 0.05).

LY increases phosphorylated serine/threonine residues on PGC-1α.

PGC-1α is regulated by numerous posttranslational modifications such as phosphorylation, methylation, and acetylation (11, 27, 40, 56). Since we elucidated a number of activated kinases in the 5-HT_1F receptor signaling pathway, we examined the phosphorylation of PGC-1α in RPTCs exposed to LY or DMSO for 2 h. PGC-1α was immunoprecipitated from RPTCs and subjected to immunoblot analysis with antibodies against phosphoserine/threonine residues and PGC-1α. The ratio of phosphorylated serine/threonine to total PGC-1α was elevated 2.8-fold, and this increase was attenuated following Akt and PKG inhibition (Fig. 4, A and B), suggesting posttranslational activation during 5-HT_1F receptor signaling in RPTCs is dependent on Akt and PKG activity.

Fig. 4.

LY increases phosphorylated serine/threonine residues on peroxisome proliferator- activated receptor gamma-γ coactivaor-1α (PGC-1α). A and B: phosphorylated serine/threonine residues were measured following immunoprecipitation of PGC-1α by immunoblot analysis after 30-min treatment with DMSO, KT (100 nM), or GDC (100 nM) followed by exposure to DMSO or LY for 2-h treatment. Total PGC-1α expression was measured to verify presence of PGC-1α protein following immunoprecipitation (IP) and equal protein input. Twenty percent of total protein lysate was used to verify immunoprecipitation of PGC-1α was successful. Data are reported as means ± SE; n = 5. Bars with different superscripts are significantly different from one another (P < 0.05).

LY reduces ERK phosphorylation and inhibitors of Gβγ and Akt prevent the reduction in ERK phosphorylation.

ERK1/2 signaling has been implicated as a negative regulator of MB (6). We analyzed the phosphorylation of ERK1/2 in RPTCs at 1, 2, and 3 h after LY treatment. ERK1/2 phosphorylation decreased 50% at 1 h (Fig. 5, A and B). We hypothesized that this reduction was also mediated by Gβγ signaling. RPTCs were pretreated with DMSO and gallein (100 nM, 30 min), followed by 1-h exposure of DMSO or LY. Blockade of Gβγ prevented LY-mediated decrease in ERK1/2 phosphorylation elevating ERK1/2 phosphorylation above vehicle levels (Fig. 5, C and D).

Fig. 5.

LY reduces ERK phosphorylation and inhibitors of Gβγ and Akt prevent the reduction in ERK phosphorylation. Phosphorylated ERK (p44/42) was measured by immunoblot analysis 1, 2, or 3 h after treatment with DMSO or LY (A and B), after 1-h treatment with DMSO or gallein (Gal; 100 nM) followed by exposure to DMSO or LY for 1 h (C and D), after 1-h treatment with DMSO or GDC0068 (GDC; 100 nM) followed by exposure to DMSO or LY for 1 h (E and F), and after 1-h pretreatment with DMSO or l-NAME (10 µM) followed by exposure to DMSO or LY for 1 h (G and H). Data are reported as means ± SE; n ≥ 5. Bars with different superscripts are significantly different from one another (P < 0.05).

To investigate cross talk between LY-mediated Akt and ERK signaling, RPTCs were pretreated with GDC (100 nM, 30 min) and then DMSO or LY. Akt inhibition prevented decreased LY-mediated ERK1/2 phosphorylation and increased ERK1/2 phosphorylation compared with vehicle (Fig. 5, E and F). RPTCs were also pretreated with l-NAME (10 µM, 30 min) before 1-h treatment with LY to determine if NOS is involved in modulating p-ERK1/2. ERK1/2 phosphorylation was unchanged compared with LY-treated RPTCs, indicating that NOS was not responsible for reduced ERK1/2 phosphorylation (Fig. 5, G and H). It is important to note that gallein and GDC alone increased ERK1/2 phosphorylation by 2.5- and 2-fold in RPTCs, respectively, suggesting that inhibition of Gβγ-AKT modulates ERK1/2 under physiological conditions, strengthening the premise that the AKT and ERK1/2 pathways collaborate to maintain cellular processes in RPTCs. Collectively, these data reveal that Gβγ and Akt activation is upstream of the suppression of ERK1/2 phosphorylation following LY treatment in RPTCs.

LY induced c-raf phosphorylation at site Ser259 and is Akt dependent.

c-Raf is known to regulate the activity of the ERK1/2 signaling cascade. Typically, c-raf phosphorylation of Ser338 activates the MEK-ERK1/2 pathway, while phosphorylation of site Ser259 is inhibitory (60, 61). c-Raf Ser259 is a target of active Akt, and there is evidence that Raf-MEK-ERK and PI3K-Akt pathways cross talk at the level of raf and Akt (44, 61). We detected phosphorylation of both Ser259 and Ser338 in DMSO and LY-treated RPTCs for 15 and 30 min. Ser259 phosphorylation was elevated 1.5- fold at 30 min in the presence of LY, while Ser338 phosphorylation remained at control levels (Fig. 6, A–C). Additionally, GDC (100 nM, 30 min) pretreatment, followed by 30-min LY exposure inhibited Ser259 phosphorylation (Fig. 6, D and E). In summary, LY exposure resulted in Akt-dependent c-raf inhibition, further evidence that ERK1/2 inhibition is regulated by Akt.

Fig. 6.

LY induced c-raf phosphorylation at site Ser259 and is Akt dependent. A: phosphorylated c-raf was measured by immunoblot analysis 15 or 30 min after treatment with DMSO or LY. B and C: densitometry analysis of Ser259 (B) and Ser338 (C) phosphorylation. D and E: phosphorylated c-raf (Ser259) was measured by immunoblot analysis after 30-min treatment with DMSO or GDC0068 (GDC; 100 nM) followed by exposure to DMSO or LY for 15 min. Data are reported as means ± SE; n ≥ 5. Bars with different superscripts are significantly different from one another (P < 0.05).

LY reduces FOXO3a phosphorylation.

We recently reported that ERK1/2 inhibition reduces phosphorylation of FOXO3a, allowing for increased nuclear FOXO3a to activate transcription of genes including PGC-1α (6). We observed that FOXO3a phosphorylation at Ser294 was reduced 40% by LY at 2 h (Fig. 7), suggesting that suppressed ERK1/2 phosphorylation upregulates PGC-1α through FOXO3a.

Fig. 7.

LY reduces forkhead box O3 (FOXO3a) phosphorylation. A and B: phosphorylated FOXO3a (Ser294) was measured by immunoblot analysis 1 or 2 h after treatment with DMSO or LY. Data are reported as means ± SE; n ≥ 5. Bars with different superscripts are significantly different from one another (P < 0.05).

DISCUSSION

Classically, the 5-HT_1F receptor has been characterized as a mediator of pain without vasoconstriction, which led to the development of 5HT_1F agonists such as LY for the treatment of migraines. However, the biological roles of peripherally expressed 5-HT_1F receptors have been understudied. We recently observed that LY induces MB and accelerates recovery of renal function following IR-AKI in mice (14). The goal of this study was to elucidate the signaling mechanism connecting the 5-HT_1F receptor to MB. Utilizing a high-throughput MB screening assay and immunoblot analyses, we determined that Gβγ, Akt, NOS, cGMP, and PKG are crucial components in MB following 5-HT_1F receptor stimulation. In addition, a second, parallel pathway was identified that negatively regulates PGC-1α and MB through ERK and FOXO3a phosphorylation. This is the first study to report that Gβγ initiates MB through dual mechanisms, increasing Akt/eNOS/cGMP/PKG/PGC-1α and decreasing Raf/MEKERK/FOXO3a pathways (Fig. 8).

Fig. 8.

Proposed 5-HT_1F receptor-mediated mitochondrial biogenesis pathway proceeds through dual mechanisms dependent on Gβγ in RPTCs. Gβγ initiates Akt phosphorylation and the subsequent phosphorylation of eNOS, both in a PI3K-dependent manner. LY increases cGMP production and induces PKG activation and PGC-1α phosphorylation in RPTCs. Simultaneously, Raf phosphorylation reduces ERK1/2 and FOXO3a phosphorylation, also a Gβγ-dependent process. Reduced FOXO3a phosphorylation promotes nuclear translocation of FOXO3a for transcription of genes such as PGC-1α. Orange phosphorylation sites indicates activation of downstream effectors. Red phosphorylation sites indicate reduction of downstream effectors. PI3K, phosphoinositide-3-kinase; Akt, protein kinase B; eNOS, endothelial nitric oxide synthase; cGMP, cyclic guanosine monophosphate; PKG, protein kinase G; PGC-1α, peroxisome proliferator-activated receptor-γ coactivaor-1α; Raf, rapidly accelerated fibrosarcoma; ERK, extracellular signal-regulated kinase; FOXO3a, forkhead box O3; MB, mitochondrial biogenesis; TFAM, transcription factor A; OPXHOS, oxidative phosphorylation; NTFs, nuclear transcription factors.

Following ligand binding to the Gα_i-coupled GPCR, Gα_i and Gβγ disassociate and activate downstream signaling. Gβγ has been shown to activate a variety of signal transduction pathways (9, 28, 45, 51). Through the use of gallein, we demonstrated that Gβγ activates the PI3K-Akt pathway in RPTCs following 5-HT_1F receptor stimulation. By recruiting the PI3K regulatory subunits p101 to the membrane, Gβγ is reported to be a direct activator of GPCR-induced PI3K activity. Gβγ has also been previously linked to Akt phosphorylation through PI3K stimulation (4).

This study not only elucidated the signaling mechanism of 5-HT_1F receptor-induced MB in RPTCs, but also the duration of this signaling, which is crucial in the roles of Akt and NO. Specifically, chronic activation of Akt and NO has been linked to defective mitochondrial function and mitophagy (25, 29, 50). We demonstrate that Akt and eNOS are sequentially and transiently activated to produce MB as opposed to the oxidative stress that can occur due to prolonged NO production (46). NO is a key molecule in PGC-1α regulation predominantly through induction of soluble guanylate cyclase and cGMP (57). Downstream effectors of cGMP, such as PKG, have also been linked to PGC-1α expression and activity (18, 20). Although the role of PKG in 5-HT_1F receptor induced MB remains unclear, we demonstrated that PKG is upstream of LY-induced PGC-1α phosphorylation. Further studies will elucidate the direct link between PKG and PGC-1α phosphorylation. One possibility is that the PKG substrate p38 MAPK is directly phosphorylating PGC-1α and increasing its activity. Several studies have reported p38 MAPK to phosphorylate and activate PGC-1α, specifically by disrupting the interaction between PGC-1α and the co-repressor p160MBP as observed in myoblasts (18, 40, 49). PGC-1α is also a substrate for a number of other kinases that regulates its activity. AMP-activated protein kinase (AMPK) phosphorylates and activates PGC-1α as well as increases its transcription (8); however, we did not observe any changes in AMPK phosphorylation following acute or chronic 5-HT_1F receptor stimulation (data not shown). Furthermore, Akt substrates such as glycogen synthase kinase-3β and mammalian target of rapamycin regulate PGC-1α activity (7, 8, 31), but the phosphorylation of these substrates was unchanged directly following Akt activation (data not shown). Collectively, these findings implicate the Akt/eNOS/cGMP/PKG signaling axis as a critical mechanism for LY-induced MB.

Interestingly, parallel to the activated Akt/eNOS/cGMP/PKG pathway, we observed decreased ERK/FOXO3a signaling. Collier et al. (6) recently demonstrated that reduced FOXO3a phosphorylation leads to increased nuclear FOXO3a expression that ultimately increased PGC-1α. We determined that both Gβγ and Akt mediate LY reduction of p-ERK1/2 and p-FOXO3a. Inhibition of Gβγ and Akt resulted in increased ERK1/2 phosphorylation, while activation of Gβγ and Akt following LY treatment reduced ERK1/2 phosphorylation. These exciting results support the hypothesis that there is cross-talk between Raf/MEK/ERK and PI3K/Akt pathways in MB. Cross talk between Raf/MEK/ERK and PI3K/Akt pathways has been reported in numerous cell types including renal cells (34, 47, 52, 61). Consistent with other findings, we determined the inhibitory 14-3-3 binding site on c-raf (Serine259) is directly phosphorylated by Akt, leading to reduced ERK1/2 signaling in RPTCs treated with LY (61).

Our group and others have demonstrated persistent disruption of mitochondrial homeostasis and inhibition of MB in myocardial infarction, spinal cord injury, stroke, drug- induced toxicities, and AKI (12, 16, 32, 33, 39). Restoration of mitochondrial number and function are necessary for normal cell and tissue function and is critical in ATP-dependent repair processes for the recovery of organ function. Despite strong evidence supporting mitochondria as a therapeutic target, there are very few drugs available to promote mitochondrial function or MB. Many of these available agents lack specificity, lack potency, or have toxic effects.

Our laboratory has identified potent inducers of MB that act on different targets. For example, the 5-HT_2A receptor agonist NBOH-2C-CN and the β₂-adrenergic receptor agonist formoterol are potent inducers of PGC-1α and MB in vitro and in vivo (19, 21, 42). Additionally, specific inhibition of phosphodiesterases 3 and 5 increased cGMP to induce MB and accelerate the recovery of renal function following AKI (58). These studies provide evidence that induction of MB to stimulate repair and recovery of organ dysfunction is an effective approach to treat a variety of acute and chronic disease.

GRANTS

This study was funded by National Institute of General Medical Sciences Grant R01-GM-084147 (to R. G. Schnellmann), Biomedical Laboratory Research and Development Program of the Department of Veterans Affairs Grant BX-000851 (to R. G. Schnellmann), and National Institute of Diabetes and Digestive and Kidney Diseases Ruth L. Kirschstein National Research Services Award Grant 5T32-DK-083262 (to W. S. Gibbs).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

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