Development of Therapeutics That Induce Mitochondrial Biogenesis for the Treatment of Acute and Chronic Degenerative Diseases

Abstract

Mitochondria have various roles in cellular metabolism and homeostasis. Because mitochondrial dysfunction is associated with many acute and chronic degenerative diseases, mitochondrial biogenesis (MB) is a therapeutic target for treating such diseases. Here, we review the role of mitochondrial dysfunction in acute and chronic degenerative diseases and the cellular signaling pathways by which MB is induced. We then review existing work describing the development and application of drugs that induce MB in vitro and in vivo. In particular, we discuss natural products and modulators of transcription factors, kinases, cyclic nucleotides, and G protein-coupled receptors.

Graphical Abstract

Mitochondria, the metabolic powerhouses of the cell, have diverse functions including ATP production, biomolecule synthesis, ionic homeostasis and antioxidant defense. As cells age and accumulate damage, mitochondria less readily meet ATP demands, thereby diminishing the cells’ functions and regenerative capacity. After toxicant exposure or cell stress, mitochondria can be damaged, and increased free radical production may be followed by persistent mitochondrial dysfunction. Diminished ATP and increased free radicals propagate injury and subsequent tissue and organ dysfunction (Figure 1). Indeed, many acute and chronic degenerative diseases across multiple organ systems are associated with a degree of mitochondrial dysfunction, often with suppression of electron transport chain proteins and activities.^1–4

Figure 1.

Multiple insults converge upon the mitochondria, leading to mitochondrial dysfunction and subsequent organ injury and disease.

Because many diseases are associated with mitochondrial dysfunction, research is underway to develop therapeutics that target mitochondria to prevent disease progression. For example, numerous compounds have been studied that prevent cell death by interfering with the formation of the mitochondrial permeability transition pore (MPTP), reducing oxidative stress using mitochondrial-targeted antioxidants, or modulating mitochondrial dynamics by inhibiting mitochondrial fission or promoting mitochondrial networking.⁵ However, whereas many of these strategies are effective for preventing injury in animal models, they target events that occur early in cellular dysfunction and therefore may be less efficacious for facilitating recovery after an insult. To address this problem, some groups have investigated compounds that induce mitochondrial biogenesis (MB), or the generation of new, functional mitochondria within cells to promote repair and regeneration.¹

This perspective will describe the role of the peroxisomal proliferation activated receptor coactivator-1α (PGC-1α) in MB and the role of mitochondrial dysfunction in acute and chronic degenerative diseases. We will also describe existing compounds that induce MB, signaling pathways responsible for their effects, and finally, potential utility of these compounds for treating human acute and chronic degenerative diseases for which there are presently limited therapeutic options.

Regulation of MB

MB requires the activation of a complex transcriptional and translational program integrating both nuclear and mitochondrial genomes.^6–7 Nuclear encoded mitochondrial genes, such as the mitochondrial transcription factors and the mitochondrial DNA (mtDNA) replication complex, facilitate transcription, replication, and proofreading of the mitochondrial genome.⁶ Integrity of mtDNA replication is particularly important in aging and chronic degenerative diseases, where deleterious mtDNA mutations and deletions can lead to dysfunctional mitochondria.^8–9 For example, the nuclear transcription factors estrogen receptor (ER) and estrogen related receptor-α (ERRα), nuclear respiratory factors 1 and 2 (NRF-1 and NRF-2), peroxisome proliferator-activated receptor (PPAR) family of transcription factors, thyroid hormone receptor (TR), cAMP-responsive element binding protein (CREB), and yin yang-1 (YY-1)¹⁰ increase expression of genes of the electron transport chain, mitochondrial transporters, antioxidant proteins, and other mitochondrial transcription factors. However, these transcription factors are pleotropic with effects on genes unrelated to MB. Selective induction of MB is typically regulated through transcriptional co-activation proteins such as the PGC-1 family (Figure 2). PGC-1 proteins activate transcription and translation of mitochondrial genes and increase energy production in healthy cells, whereas in injured cells PGC-1 activation often normalizes overall mitochondrial function as measured by ATP production, mitochondrial membrane potential, and reactive oxygen species (ROS) generation.^11–13

Figure 2.

PGC-1α integrates extracellular and cytosolic signaling inputs to selectively upregulate mitochondrial biogenesis.

The PGC-1 family, composed of PGC-1α, PGC-1β, and PGC-1 related coactivator (PRC), facilitate the formation of complexes capable of activating the transcription of nuclear genes related to MB.¹⁴ PRC is thought to play a role in redox-sensitive inflammatory responses and MB during cellular proliferation, whereas PGC-1β appears to contribute more to maintenance of mitochondrial mass. In contrast, PGC-1α has been shown to drive MB in response to various environmental cues. Because PGC-1α tends to be the most inducible and responsive member of the PGC-1 family, its activation has emerged as a key therapeutic strategy for MB induction. However, it is important to note that PGC-1α-independent mechanisms of MB have been reported.^15–18 Such mechanisms include compensatory activation of PGC-1β or PRC and direct activation of transcription factors that induce mitochondrial genes.

Through activation of PGC-1α and its associated transcription factors, multiple signaling pathways have been shown to regulate MB. PGC-1α can be directly activated by silent mating type information regulation 2 homolog 1 (SIRT1)-mediated deacetylation,¹⁹ methylation by protein arginine methyltransferase 1 (PRMT1),²⁰ or phosphorylation by kinases such as p38,²¹ protein kinase A (PKA),²² and AMP-dependent kinase (AMPK).²³ Additionally, PGC-1α and other transcription factors associated with MB can be activated by NO/cGMP and calcium-dependent signaling.²⁴ In summary, these diverse signaling inputs allow exquisite control of mitochondrial homeostasis to meet cellular energy demands and to maintain proper cellular function.

The Importance of MB in Disease

Because mitochondria regulate many processes within cells, mitochondrial dysfunction or disruptions in mitochondrial homeostasis lead to severe deficits in cellular functions.^1–2 Injury to mitochondria following ischemia reperfusion injury, toxicant exposure, or severe inflammatory response leads to deficient ATP and disruption of ion homeostasis. Additionally, mitochondrial stress increases superoxide anion production and which causes damage to proteins and lipid membranes. These mitochondrial derangements disrupt cellular repair, proliferation, and differentiation status and increase cell death.

Mitochondrial dysfunction has been implicated in numerous acute and degenerative disease processes, such as myocardial infarction,²⁵ stroke,²⁶ and acute kidney injury (AKI).²⁷ These disease states may be attributed in part to the role of mitochondria and oxidative metabolism in cellular differentiation as observed in neurons,²⁸ myocytes,²⁹ and immune cells.³⁰ Chronic conditions causally linked to such acute insults (such as chronic kidney disease and heart failure) are also characterized by persistent mitochondrial dysfunction,^31–32 suggesting that the lack of mitochondrial recovery after an acute injury can also lead to chronic degenerative disease. For example, deficits in PGC-1 family proteins have been associated with the development heart failure in both animal models and human patients.^33–34 Interestingly, mice that overexpress PGC-1 proteins also exhibit abnormal cardiac function,³⁵ indicating that a tight control over mitochondrial content is necessary for normal organ function. Similarly, animal models of chronic kidney disease demonstrate diminished renal mitochondrial function,³⁶ and animal models of mitochondrial dysfunction demonstrate chronic kidney disease.³⁷ Finally, human patients with chronic kidney disease have decreased mtDNA in skeletal muscle and peripheral mononuclear blood cells,³⁶ suggesting that mitochondrial defects in a single organ can lead to global mitochondrial dysfunction.

Other chronic diseases also have been associated with disruption of mitochondrial homeostasis. Type II diabetes mellitus and metabolic syndrome are characterized by mitochondrial dysfunction associated with insulin resistance.³⁸ In metabolic syndrome, pancreatic beta cells exhibit increases in UCP2, decreased ATP synthesis, and increased levels of ROS.^39–40 Additionally, reductions in complex IV of the electron transport chain have been associated with the development of diabetes in obese mice and patients.⁴¹ Furthermore, epigenetic silencing of electron transport chain genes and mtDNA,^42–44 along with genes associated with MB such as PGC-1α and TFAM,^45–46 lead to decreased mitochondrial content and a greater proportion of dysfunctional mitochondria, thereby causing sustained deficiencies in cellular respiration.

Multiple neurodegenerative diseases also have been associated with decreased mitochondrial mass, altered mitochondrial dynamics, and dysregulation of MB. Parkinson disease has been linked to a panoply of mutations that lead to mitochondrial dysfunction. Defects in PINK1 and Parkin disrupt clearance of damaged mitochondria, permitting accumulation of oxidative damage in dopaminergic neurons and suppression of PGC-1α and decreased cellular respiration.^47–50 Mutations in DJ-1 increase ROS while decreasing anti-oxidant defenses,⁵¹ leading to decreases in mitochondrial membrane potential, poor mitochondrial quality control, and altered mitochondrial morphology. Similarly, mutations in mTDNA,^52–55 TFAM,⁵⁶ mortalin,⁵⁷ and α-synuclein⁵⁸ lead to increased susceptibility to ROS and subsequent mitochondrial dysfunction. Additionally, huntingtin mutants associated with Huntington’s disease bind to the PGC-1α promoter and prevent its transcription and the transcription of other nuclear transcription factors associated with MB, including CREB.^59–60 Huntingtin mutations also cause impaired mitochondrial calcium handling,⁶¹ reduced respiration,^62–63 and disrupted mitochondrial dynamics.^64–65 Finally, genetic and toxicant-induced models of Alzheimer disease and samples from human patients confirm the suppression of mitochondrial proteins and the MB transcriptome in Alzheimer disease,^66–67 along with mtDNA damage and disruptions in mitophagy and mitochondrial morphology.^68–70 Thus, compounds that induce MB may alleviate cellular dysfunction associated with acute and chronic degenerative diseases and promote organ repair and recovery that leads to improvements in patient health.⁷¹

Natural Products

Because mitochondria and oxidative stress are associated with aging, populations with longer lifespans have been studied to identify a potential means for preventing deleterious effects of aging. These studies have identified multiple chemicals capable of inducing MB (Figure 3), and these compounds have shown efficacy in multiple disease models by modulation of multiple signaling axes. Nonetheless, their therapeutic applicability in many cases is limited by poor absorption and low oral bioavailability.

Figure 3.

Naturally occurring polyphenols capable of inducing MB.

Resveratrol

A widely studied nutritional activator of MB is the polyphenol resveratrol (1).⁷² Compound 1 has been shown to induce MB by activating SIRT1 directly or indirectly through AMPK.⁷³ SIRT1 in turn deacetylates PGC-1α and allows it to exert its transcriptional effects. In particular, 1 activates AMPK by inhibiting components of the electron transport chain such as complex I and F1/F0 ATPase.^74–75 Docking studies with complex I suggest that resveratrol binds to the NAD⁺ binding site of complex I through pi stacking interactions with its aromatic components and by hydrogen bond interactions through its hydroxyl group.⁷⁵ When binding F1/F0 ATPase, 1 prevents rotation of the ATP synthase complex through a network of hydrophilic and hydrophobic interactions.⁷⁴ Compound 1 can also directly activate PPARα via interactions with the 4′-hydroxyl group.⁷⁶ It also activates PPARγ by interactions between R280 and its 4′-hydroxyl group near the opening of the ligand binding pocket as well as Van der Waals interactions with F264, H266, and R288.⁷⁷ Together, protein-ligand interactions trigger signals that induce MB.

In models of diabetic cardiovascular disease, 1 induces MB and restores vascular reactivity in vitro and in vivo.⁷⁸ In cellular and animal models of neuronal radiation damage,⁷⁹ Alzheimer disease,⁸⁰ Parkinson disease,⁸¹ and Huntington’s disease,⁸² 1 normalizes mitochondrial function and rescue cellular viability and function. Compound 1 also attenuates oxidative stress in fibroblasts from patients with Complex I deficiency by increasing SOD2 in a SIRT3-dependent manner.⁸³ Human clinical trials using 1 demonstrated improved lipid profiles, antioxidant defenses, and vascular reactivity in diabetic and obese subjects;^84–89 however, there are conflicting data regarding the effect of 1 on insulin sensitivity,^{84, 88, 90} and 1 had no effect in non-obese subjects.⁹¹

Epicatechins

(−)-Epicatechin (2),⁹² primarily found in cocoa, has been shown to induce MB through multiple signaling pathways, including Akt-dependent nitric oxide (NO) generation,^93–94 CREB phosphorylation, and δ-opioid receptor activation.⁹⁵ The epicatechin epigallocatechin-3-gallate (3),⁹⁶ promotes cAMP-dependent signaling to increase SIRT1 and PGC-1α.⁹⁷ Although there are limited data regarding the structural basis for 2 activation of cAMP-dependent signaling, Akt-dependent signaling is mediated by the 3″-, 3′-, and 4′-hydroxyl groups.⁹⁸ Following oxygen-glucose deprivation, neuronal viability is rescued by 2 via the Akt-eNOS pathway and CREB activation.⁹⁴ In a mouse model of diabetes, 2 reduces oxidative stress in cardiac tissue by inducing MB.⁹⁹ Similarly, in mouse models of cardiovascular disease, 2 acts through the δ-opioid receptor to prevent mitochondrial swelling and to increase respiration;^{95, 100} it can also decrease cardiac ischemia-reperfusion injury through NO and cGMP generation. Even in aged mice, epicatechin increases expression of mitochondrial and antioxidant proteins.¹⁰¹ Through its cAMP-dependent activation of SIRT1 and PGC-1α, 3 enhances MB in Down’s syndrome patient fibroblasts and enhances mitochondrial calcium handling by modulating mitochondrial tethering to the rough endoplasmic reticulum.⁹⁷ Compound 2 also induces MB in human diabetic patients to improve skeletal muscle metabolism.¹⁰²

Curcumin

Curcumin (4),¹⁰³ a diarylheptanoid found in turmeric, has shown promise for promoting MB and improved function in several disease models. By activating multiple signaling molecules, including p38, PKA, AMPK, SIRT1, and NRF2, 4 can induce MB and protect cells against injury.^104–106 The o-methoxy group in compound 4 is important for increasing p38-mediated HO-1 expression, which confers cytoprotection in endothelial cells.¹⁰⁴ The unsubstituted 5′- and 5″-positions and its olefinic system allow 4 to inhibit NF-κB and activate the NRF2 pathway.¹⁰⁷ In cellular models of metabolic syndrome, 4 rescues hepatic mtDNA, NRF1, and TFAM and reduces inflammation and NFκB activity.¹⁰⁸ In white adipose tissue, 4 increases browning and markers of MB via increases in norepinephrine and β₃ adrenergic receptor expression.¹⁰⁹ Pretreatment with 4 improves mitochondrial membrane potential, oxygen consumption rates, and survival in cellular models of Parkinson disease.¹¹⁰ Compound 4 attenuates neuronal death and reduces infarct size following cerebral ischemia-reperfusion injury with concomitant increases in mitochondria and improvements in neurological function.¹¹¹ In animal models of metabolic syndrome, 4 restores hepatocyte mitochondrial function to reduce hepatosteatosis.¹¹² Following gentamicin-induced nephrotoxicity, 4 can increase PGC-1α and NRF2, thereby elevating mitochondrial protein expression and improving mitochondrial structure.¹⁰⁵ In rat skeletal muscle, 4 increases mtDNA content and mitochondrial protein expression following endurance training via PKA-dependent activation of AMPK, SIRT1, and PGC-1α.¹⁰⁶

Phytoestrogens

Phytoestrogens, such as genistein (5),¹¹³ daidzein (6),¹¹⁴ pyrroloquinoline quinone(7),¹¹⁵ coumestrol (8),¹¹⁶ and equol (9),¹¹⁷ are natural products often found in legumes such as soybeans. They have been shown to exert their effects in part by modulation of estrogen receptors and partly via activation of SIRT1.^118–120 5-hydroxyl groups prevent SIRT1 activation, whereas 7-hydroxyl groups are necessary for SIRT1 activation. Similarly, a 3-phenyl group appears to drive increased SIRT1 expression.¹²⁰ Compounds 5–8 have been shown to induce MB in vitro.^120–122 Additionally, through their biogenic effects, 5 and 6 rescued cultured renal proximal tubule cells from oxidant injury.¹²⁰ In vivo, 5 and 9 induce MB to improve bioenergetics in ovariectomized mice.^123–124 Both 5 and 6 increase mitochondrial markers with associated improvements in insulin sensitivity and glucose metabolism in diabetic mice.^125–126 Compound 5 also reduces the size of a myocardial infarct in mice by rescuing mitochondrial function.¹¹⁸ Finally, 7 stimulates MB in both wild type mice and transgenic models of Alzheimer disease;^127–128 in the latter model, improvements in synaptosomal bioenergetics are correlated with cognitive improvement.

Transcription Factor Modulators

Although natural products have been useful in identifying biological targets for MB, their poor pharmacokinetic parameters limit their therapeutic potential. Modulators of the transcriptional machinery responsible for MB can potently and efficaciously induce MB; however, because they activate transcriptional programs other than MB, these compounds can have severe side effects that limit their clinical utility. Thiazolidinediones

The thiazolidinediones (TZDs) are a class of hypoglycemic drugs used to treat diabetes mellitus that includes rosiglitazone (10),¹²⁹ pioglitazone (11),¹³⁰ troglitazone (12),¹³¹ and ciglitazone (13) (Figure 4).¹³² Classically, they act as agonists of the transcription factor peroxisome PPARγ, leading to increased insulin sensitivity. These effects are primarily mediated by the acidic head group, which engages in necessary hydrogen bonding interactions with PPARγ to stabilize its active conformation.^133–134 More recently, acute PPARγ-independent effects of TZDs have been discovered, including inhibition of the electron transport chain, which reduces the ATP/AMP ratio, leading to AMPK activation and subsequent MB.^135–137 TZDs have also been shown to exert anti-inflammatory effects and to upregulate the mitochondrial stress-response, leading to increased anti-oxidant defenses.¹³⁵ Although they upregulate multiple signaling pathways, the capacity of TZDs to sensitize tissues to the effects of insulin has been shown to correlate with increased expression of mitochondrial proteins, suggesting that induction of MB may be central to the clinical efficacy of these drugs.¹³⁸

Figure 4.

Thiazolidinedione inducers of MB.

In vitro, 10–13 increase cell viability and improve neuronal function in models of ischemic injury,¹³⁹ Alzheimer disease,¹⁴⁰ Huntington’s disease,^141–142 and multiple sclerosis.¹⁴³ Similarly, in animal models of neurodegenerative diseases, 10 and 11 improve both cellular and behavioral markers of neurological function.^144–145 In animal models of cardiac disease, 10 can rescue cardiac mitochondrial function following septic injury;¹⁴⁶ however, other studies indicate that 10 increases cardiac ROS and can be arrhythmogenic.^147–148 In models of metabolic syndrome, 10–13 induce MB in adipose tissue,^{15, 149} pancreatic beta cells,¹⁵⁰ and skeletal muscle^{137, 151} to enhance insulin sensitivity. In humans, 11 induces MB in subcutaneous adipose tissue,¹⁵² and 10 can do so in skeletal muscle.¹⁵³

Estrogens

To understand the underlying processes responsible for sex-dependent differences in lifespan and oxidative stress, multiple groups reported that estrogens can be protective in various tissues. Furthermore, reduced levels of estrogens, such as in ovariectomized mice, lead to increased ROS production.¹⁵⁴ Estrogens (Figure 5) can bind to the transcription factors estrogen receptor α (ERα) and estrogen receptor β (ERβ) to directly influence gene expression. 17β-Estradiol (14)¹⁵⁵ and progesterone (15)¹⁵⁶ are the principle biologically active estrogens. 14 and 15 interact with nuclear estrogen receptors by hydrogen bonding interactions between the ligands’ hydroxyl groups and the receptors’ polar residues and by hydrophobic interactions with the receptors’ binding pockets.¹⁵⁷ ERα-selectivity, such as by the selective ligand 4,4′,4″-(4-propyl-[1H]-pyrazole-1,3,5-triyl)trisphenol (16),¹⁵⁸ is mediated by steric bulk to interact with a residue found in ERα but not ERβ.¹⁵⁸ Selectivity for ERβ by diarylpropionitrile (17)¹⁵⁹ is mediated by phenolic groups, while its efficacy is improved by its nitrile group.¹⁵⁹ Recently, it has also been shown that estrogens activate plasma membrane-bound estrogen receptors such as the G protein-coupled estrogen receptor (GPER). The GPER-selective ligand G-1(18)¹⁶⁰ is structurally similar to 14 but is unable to form hydrogen bonds in the nuclear estrogen receptors;¹⁶⁰ however, 18’s acetyl group and pseudosymmetry allows engagement of specific residues of the GPER to stabilize the active conformation.^161–162

Figure 5.

Estrogen inducers of MB.

Compound 14 has been shown to induce MB in immortalized cell lines and in a cellular model of Leber hereditary optic neuropathy, a mitochondrial disease. ^163–164 In animal models, 14 normalizes ROS production, increases antioxidant defenses, and enhances respiratory capacity in the heart and brain.^{154, 165} Furthermore, 15 and synthetic estrogen receptor agonists such as 16 and 17 have been shown to enhance respiratory capacity in the brain and promote clearance of lipid peroxidation products.¹⁶⁶ Of note, the use of receptor subtype selective agonists suggests that ERα and ERβ differentially regulate the expression of electron transport chain proteins. Additionally, at least a portion of the cardioprotective effects of estrogen are mediated through the GPER, as shown by stimulation with the GPER-selective agonist 18.¹⁵⁴ Despite the clear protective potential of estrogens, their proliferative and endocrine effects limit their use as a long-term therapy for chronic degenerative diseases. However, the development of selective ER and GPER ligands that drive specific signaling and transcriptional programs may improve the utility of such therapeutics.

SIRT1 activators

The identification of SIRT1 as a common target of natural product-induced increases in PGC-1α led to the development of multiple SIRT1 activators, such as SRT1720 (19),¹⁶⁷ SRT1460 (20),¹⁶⁷ SRT2183 (21),¹⁶⁷ and SRT2104 (22) (Figure 6).¹⁶⁸ In the initial synthesis of SIRT1 activators,¹⁶⁹ the basic methylamino ring at C-3 of the imidazothiazole ring of 19 and 20 enhanced water solubility, while derivatization of the amide group (such as with the 2-quinoxaline group of 19) improved potency and efficacy. Interestingly, both 19 and 20 share a methylamino ring and have greater efficacy, whereas 19 and 21 have a 2-quinoxaline group and more potency,¹⁶⁷ suggesting that the two groups may play distinct roles in the pharmacodynamic qualities of these compounds. The direct mechanisms of action for the sirtuin class have been controversial. Assays with isolated fluorescent peptides were used for optimization, but direct proteomic assays indicate that 19–21 do not directly activate SIRT1 and, rather, act promiscuously to activate or inhibit numerous targets;¹⁷⁰ however, other work has shown that these compounds directly activate SIRT1 by binding to amino acid E230.¹⁷¹

Figure 6.

Activators of SIRT1 that induce MB.

Due to numerous SIRT1 targets, these activators can affect various cellular processes, including inflammation, lysosomal trafficking, and metabolism. Among its targets, SIRT1 deacetylates PGC-1α, facilitating nuclear import of and transcriptional regulation by PGC-1α, leading to MB. In models of type II diabetes mellitus, SIRT1 activators have been shown to improve lifespan, normalize pancreatic morphology, improve insulin, glucose, and fatty acid metabolism and increase mitochondrial markers;^{167, 172–174} however, other studies have shown a lack of efficacy in diabetic mice, calling into question the beneficial effects of these compounds.¹⁷⁰ With respect to neurodegenerative diseases, SIRT1 activators prevent neurodegeneration and restore MB in animal models of Huntington’s disease and multiple sclerosis.^175–176 SIRT1 activation has shown promise in renal disease, restoring renal function after AKI and preventing renal medullary damage in obstructive nephropathy.^177–179 In models of cardiovascular disease, 19 reduces the size of myocardial infarction and preserves contractility,¹⁸⁰ as well as reducing ROS and improving contractility in mice with enhanced ALDH2 activity.¹⁸¹ Compound 19 also preserves endothelial function in aged mice.¹⁸² Even in healthy animals, 19 and other SIRT1 activators have been shown to extend lifespan and “healthspan” by preventing the development of age-associated diseases in multiple organ systems.¹⁸³ In human trials, 22 improved lipid profiles in diabetic patients but did not affect plasma glucose or insulin, likely due to large pharmacokinetic variability.¹⁸⁴ Additionally, 22 reduces cholesterol, LDL, and triglycerides in otherwise healthy smokers,¹⁸⁵ suggesting that SIRT1 activation is important to the human healthspan.

Kinase Modulators

Kinases either phosphorylate target proteins or function as scaffolds to co-localize other kinases and targets to regulate cellular signaling. Phosphorylation of specific targets can either activate or inhibit cellular signaling pathways in response to environmental cues. Because they are central signaling molecules, kinases are attractive therapeutic targets. In particular, activators of kinases that induce MB, such as AMPK, can be useful in multiple diseases. Unfortunately, inhibitors are easier to develop, and most kinase modulators are inhibitors. However, inhibitors of kinases that negatively regulate MB, such as extracellular signal-regulated kinases 1/2 (ERK1/2), also provide promise as therapeutics.

AMPK

AMPK is an energy sensing kinase involved in the modulation of metabolism through the cellular AMP/ATP ratio. AMPK activation is increased during exercise and induces MB, and it is decreased with aging and during multiple chronic degenerative diseases.¹⁸⁶ AMPK activation has been shown to be an upstream regulator of sirtuins and therefore PGC-1α.¹⁸⁷ Furthermore, pharmacologic activation of AMPK has been observed with multiple natural products that induce MB. Activators of AMPK (Figure 7), including the indirect activators AICAR (23),¹⁸⁸ metformin (24),¹⁸⁹ phenformin (25),¹⁹⁰ R419 (26),¹⁹¹ and C24 (27),¹⁹² and the direct activator A769662 (28),¹⁹³ have been developed and induce MB in multiple cell lines. Additionally, 23 has been shown to enhance proliferation and increase ATP in models of complex I deficiency and MELAS.^194–195 Compound 23 is biotransformed via phosphorylation within the cell and acts as an AMP mimetic to activate AMPK and other AMP-dependent processes.¹⁸⁸ The biguanides 24 and 25 activate AMPK in a LKB1-dependent manner and through inhibition of complex I;^{191, 196} by inhibiting the electron transport chain, the AMP/ATP ratio is increased, leading to AMPK activation. Compound 26 also indirectly activates AMPK via complex I inhibition,¹⁹¹ and 28 activates AMPK by binding to an allosteric site between the alpha and beta subunits of AMPK. 28 both allosterically activates and prevents Thr172 dephosphorylation.¹⁹⁷

Figure 7.

Kinase modulators that induce MB.

In models of diabetes and metabolic syndrome, 23 mimics high intensity exercise in skeletal muscle with accompanying increases in SIRT1 activation and PGC-1α activity. These improvements in MB decrease oxidative stress in both renal and endothelial cells,^198–200 preventing common comorbidities such as diabetic nephropathy and poor wound healing. Compound 23 can also improve pancreatic morphology via AMPK activation to enhance insulin sensitivity and GLUT4 expression,²⁰¹ thereby decreasing plasma glucose. In hepatic cells, 27 reduces lipid biosynthesis to prevent lipid accumulation and preserve hepatic function.¹⁹² In humans with gestational or type II diabetes, 23 and 25 prevents insulin resistance in multiple tissues.^202–204 In the heart, 23 reduces oxidative stress and improves contractility,¹⁸¹ and it is associated with improvements in insulin sensitivity in diabetic mice as well as reductions in cold ischemic injury in mouse models of heart transplant.²⁰⁵

AMPK activators have also shown promise for treating neurodegenerative diseases. Neuronal activity has been shown to drive PGC-1α and NRF-1 expression in an AMPK-dependent manner,²⁰⁶ leading to MB, and pharmacologic activation of AMPK has been shown to mimic these effects. Compound 23 has also been shown to impact neuronal development by promoting mitochondrial accumulation at axonal branch points, thereby facilitating branch formation and retention.²⁰⁷ In models of Alzheimer disease, 23 ameliorated mitochondrial dysfunction and prevented neurotoxicity and tau hyperphosphorylation.^208–209 Compound 23 decreased amyloid beta, a protein implicated in Alzheimer disease, in a PPARγ dependent manner.²¹⁰ Compound 23 has also been shown to decrease inflammation in models of multiple sclerosis, attenuating pathological and behavioral changes. Furthermore, in models of ischemic brain injury, 23 diminishes ischemic neuronal damage.²¹¹

ERK1/2

Another means of inducing MB is the inhibition of negative regulators of MB, such as ERK1/2. Following its activation by MEK1/2, ERK1/2 regulates a variety of cellular processes, including differentiation, apoptosis, survival, proliferation, and motility.²¹² Inhibition of MEK by U0126 (29)²¹³ or trametinib (30)²¹⁴ leads to a rapid suppression of ERK1/2 phosphorylation (Figure 7). Compound 29 can exist in the (Z,Z) or (Z,E) isomer; however, the (Z,Z) isomer provides better MEK inhibition, as does the presence of electron donating amino groups at o-positions of its phenyl groups.²¹³ The iodo- and cyclopropyl groups of Compound 30 improve potency for cancer cell growth inhibitory activity over its lead compound JTP-70902 (31)²¹⁴, while its methyl groups improve stability and its acetamide group improves solubility.²¹⁴ ERK1/2 has been shown to suppress PGC-1α in melanoma cells.²¹⁵ Additionally, in models of Parkinson disease ERK1/2 activation leads to phosphorylation of TFAM, impairing its ability to bind to mitochondrial DNA.²¹⁶ MEK1/2 inhibitors, such as 29 and 30, have been developed for cancer chemotherapy. In vitro models of renal oxidative stress indicate that ERK1/2 is a mediator of oxidative damage in proximal tubule cells, and that its inhibition by 29 prevents oxidative damage.²¹⁷ Our laboratory has shown that ERK1/2 activation increases after AKI and that pre-treatment with the MEK1/2 inhibitor 30 rescues mitochondrial function and restores renal function in a mouse model of AKI.²¹⁸ These data indicate that inhibition of suppressors of MB can induce MB and restore organ function following injury.

Cyclic Nucleotide Modulators

The cyclic nucleotides cGMP and cAMP are cellular second messengers that are generated in response to extracellular signals. They activate downstream kinases or are hydrolyzed by phosphodiesterases (PDE). NO increases cGMP synthesis by binding to a heme group on soluble guanylate cyclase (sGC), while cAMP is increased through activation of adenylate cyclase by the stimulatory G-protein Gα_s. Because cyclic nucleotide generation is disrupted in multiple pathological states, cyclic nucleotide modulators are attractive targeted therapies for the induction of MB in various diseases.

NO-cGMP-PKG Axis

The NO-cGMP-PKG pathway can be modulated by: 1) nitric oxide (NO) donors, such as sodium nitroprusside (32), (±)S-nitroso-N-acetylpenicillamine (SNAP, 33),²¹⁹ diethylamine NONOate (DEA-NONOate, 34),²²⁰ and diethylenetriamine-NONOate (DETA-NONOate, 35)²²⁰ which increase cellular NO (Figure 8); 2) sGC stimulators and activators, such as cinaciguat (36),²²¹ riociguat (37),²²² and BAY 41-2272 (38)²²³ which directly increase cGMP production (Figure 8); and 3) phosphodiesterase (PDE) inhibitors, such as zaprinast (39),²²⁴ sildenafil (40),²²⁵ udenafil (41),²²⁶ tadalafil (42),²²⁷ and vardenafil (43)²²⁸ which increase cGMP by preventing its hydrolysis (Figure 9). Clinically, these compounds are used to induce vasodilation to treat hypertension or erectile dysfunction. Activation of this pathway has been shown to increase PGC-1α and stimulate MB both through the activation of PKG and nitrosylation of transcription factors to increase their binding to the PGC-1α promoter.^229–230

Figure 8.

Activators and stimulators of the NO/cGMP pathway.

Figure 9.

Phosphodiesterase (PDE) inhibitors associated with MB.

As their name implies, all NO donors have a group, usually a nitrate or a furoxan group, that can be liberated to form NO. Because the NO donating group is small, NO donors can be “fine-tuned” for multiple clinical uses and to slow the rate of NO release.^231–232 However, because NO generation causes such a dramatic drop in blood pressure, NO donors are of limited clinical use. However, these compounds readily confirm the importance of NO for preventing metabolic derangements and cell death, particularly in skeletal muscle. In hypoxia, dietary nitrate (a natural NO donor) prevents PGC-1α suppression, leading to increases in fatty acid oxidation and respiration. Even under normoxic conditions, nitrate stimulates MB in a cGMP/PKG-dependent manner.²³³ Compound 33 has also been shown to induce MB in myoblasts and reduce the effects of caspase-dependent and –independent apoptotic molecules,²³⁴ and 34 also improves synaptic conduction in models of Alzheimer disease in a cGMP-dependent manner.²³⁵

sGC activators and stimulators increase the activity of sGC in the absence of NO. Stimulators such as 37 and 38 increase sGC activity with a non-oxidized heme group, whereas activators increase sGC activity even if the heme prosthetic group is oxidized. Both classes of compounds have been approved for clinical use to treat pulmonary hypertension. Compound 38 was optimized for vasorelaxation through the addition of a 2-fluoro-phenyl group, a pyrazolo[3,4-b]pyridine ring, and a cyclopropyl group.²²³ Compound 37 was optimized to increase oral bioavailability and half-life, and to reduce clearance via amino and N-methylcarbamate substitutions on the pyrimidine group.²²² On the other hand, sGC activators have shown greater utility beyond blood pressure control, likely due to their capacity to activate sGC even under high oxidative stress. Compound 36 was identified using a high-throughput screen and was confirmed to displace the heme of sGC by interacting with its YXSXR motif through carboxylic acid moieties.²³⁶ In pre-clinical studies, compounds 36–38 improve cardiac, renal, and neurological function across multiple disease models including ischemia reperfusion injury, sepsis, diabetes, and Alzheimer disease.^237–241 However, despite the efficacy of cGMP in promoting MB, few studies have examined the role of MB in these functional improvements. Compound 36 protects against myocardial infarction by increasing H₂S, a known inducer of MB,²⁴² suggesting that further investigation is warranted into the role of MB in these compounds’ protective effects.

Inhibition of cGMP-selective PDEs prevents cGMP hydrolysis, promoting its accumulation in the cell and facilitating stimulation of MB. Compound 40 was designed from 39 by mimicking the guanosine dipole moment, adding an ethoxy group to improve potency, and adding a piperazine sulfonamide to improve solubility, selectivity, and potency.²⁴³ However, both 40 and 41 discriminate poorly between PDE5 and PDE6, leading to visual side effects.²²⁶ Compound 42 has better selectivity for PDE5 over PDE6 with the addition of more electron donating groups; however, relative to 40 and 43, 42 is less selective for PDE11.^{227, 244–245} Although these compounds have been extensively developed for treating pulmonary hypertension and erectile dysfunction, they also have been tested for treating other diseases.

Because cGMP-selective PDE inhibitors were designed to reduce blood pressure via increased vasodilation, it is reasonable that they have been tested for conditions characterized by endothelial dysfunction, such as diabetes. As expected, in models of diabetes, 40 improves endothelial function as measured by flow-mediated dilation.^246–247 In addition to their effects on vascular reactivity, 40, 42, and 43 reduce plasma markers of diabetes, such as lipids, serum glucose, and HbA_1c, and are associated with improvements in mitochondrial content.^248–251 In adipocytes and hepatocytes, 40 enhances lipid oxidation and increases insulin tolerance and cellular morphology.²⁴⁸ cGMP-selective PDE inhibitors also reduce diabetic complications in other organs, such as the kidney and heart. In models of diabetic nephropathy, 40 reduces microalbuminuria, a predictor of renal and cardiac dysfunction.²⁴⁹ Additionally, in diabetic mice, 42 rescues the expression of cardiac cytoskeletal and redox proteins to improve cardiac morphology and function.^251–252

In addition to beneficial reductions in the development of diabetic cardiomyopathy, cGMP-selective PDE inhibitors also ameliorate non-diabetic cardiac dysfunction. In ischemic cardiomyopathy and myocardial infarction, 40, 42, and 43 increase survival and decrease infarct size by reducing cell death and preserving mitochondrial function.^253–255 Compound 42 also prevents cardiac remodeling and hypertrophy, stabilizing contractility rather than allowing progression to heart failure and pulmonary edema.²⁵⁶ Similarly, in models of mitral regurgitation and doxorubicin toxicity, 40 inhibits cell death and preserves mitochondrial function by upregulating anti-apoptotic proteins and maintaining the mitochondrial membrane potential.^257–258

cAMP-PKA-CREB axis

CREB regulates PGC-1α activity and expression to promote MB and is down-regulated in multiple disease states characterized by mitochondrial dysfunction. In Alzheimer disease, CREB phosphorylation is diminished due to impaired activation by PKA. This loss of activity leads to a downregulation of PGC-1α and an imbalance in tau protein, a driver of Alzheimer disease.²⁵⁹ A similar decrease in CREB activity has been observed in Huntington’s disease.²⁶⁰ Additionally, ethanol decreases cellular cAMP, thereby reducing CREB activity to suppress PGC-1α and thereby exert its toxic effects.²⁶¹ Taken together, these data indicate that activation of the cAMP-PKA-CREB signaling pathway can promote MB and protect against neurodegenerative diseases.

The primary therapeutic approach for activating this signaling axis is with phosphodiesterase (PDE) inhibitors such as rolipram (44)²⁶² and cilostazol (45)²⁶³ (Figure 9). Compound 44 inhibits PDE4, a cAMP-selective PDE, whereas 45 inhibits PDE3, a PDE capable of hydrolyzing both cAMP and cGMP; however, PDE3’s V_max for cAMP is substantially higher than that of cGMP. Compound 44’s selectivity arises in part from its optimized potency for PDE4 and the unfavorable orientation of a conserved glutamate residue in other PDEs.²⁶⁴ In contrast, the lactam group of 45 engages in hydrogen bonding interactions with multiple receptor residues to promote PDE3 selectivity.²⁶⁵ Both 44 and 45 can increase PGC-1α in vitro, indicating that they induce MB,²⁶⁶ and both have shown potential for therapeutic use in pre-clinical disease models. However, in humans, 44’s narrow therapeutic window limits its application, whereas 45 is approved for clinical use in the treatment of diabetic vascular complications.

Restoration of the cAMP-PKA-CREB pathway substantially reduces the effects of neurodegenerative diseases. In animal models of Huntington’s disease, 44 improves neuronal function, morphology, and survival and decreases neurological impairment.^{260, 267} Compound 44 also reduces synaptic conduction abnormalities associated with Alzheimer disease, improving cognition.^268–269 Interestingly, these effects and increased CREB phosphorylation lasted beyond the cessation of treatment. In ischemic brain injury, 45 reduces neuroinflammation, reducing infarction size and decreasing apoptosis and free radical production.^270–271 In models of Alzheimer disease, 45 increases SIRT1 expression, reducing symptoms and improving cognitition.²⁷² Furthermore, in a retrospective study, 45 improved cognition in human patients,²⁷³ suggesting that PDE3 inhibition holds promise for treating Alzheimer disease.

Used clinically to treat claudication, the beneficial effects of 45 in models of diabetic cardiovascular disease are well studied. In models of limb ischemia, 45 increases angiogenesis by rescuing PPARγ, increasing angiogenic factors vascular endothelial growth factor (VEGF) and hepatocyte growth factor (HGF);^274–275 this normalization of PPARγ also occurs in other tissues, such as the retina and the kidney.²⁷⁶ Compound 45 also prevents endothelial cell senescence by increasing cAMP, leading to SIRT1 activation. In the heart, 45 reduces oxidant-induced mitochondrial dysfunction and significantly reduces myocardial infarction size.^277–279 Furthermore, 45 improves insulin sensitivity and reduces blood glucose and HbA_1c in diabetic mice and human subjects,^280–282 as well as reducing the urinary excretion of albumin and renal inflammation, indicating that 45 improves diabetic nephropathy.

Despite these promising data, controversy exists regarding use of cAMP-selective PDEs in chronic degenerative diseases of the liver and kidney. On the one hand, 45 improves hepatic function after ischemic insult by inducing MB;²⁸³ however, in models of lipotoxicity, increased cAMP acts synergistically to induce cell death despite concurrent stimulation of MB.²⁸⁴ Additionally, despite the promising work in diabetic nephropathy described previously, we found that cAMP-selective PDE inhibitors do not induce MB in proximal tubule cells,²⁸⁵ suggesting they are poor therapeutic options for treating AKI.

GPCR Ligands

G protein-coupled receptors (GPCRs) are well characterized plasma membrane receptors that are the target of a substantial portion of currently available drugs. By coupling to G proteins, GPCRs can modulate cAMP, calcium, and NO and activate various kinases and signaling pathways. Additionally, different ligands of the same receptor can cause activation of distinct signaling programs, a phenomenon known as “functional selectivity” or “biased agonism.”²⁸⁶ By stabilizing different receptor conformations, different ligands can alter receptor interactions with G proteins, G protein-coupled receptor kinases (GRKs), and scaffolding proteins such as arrestins. One such scaffolding protein, GRK interacting protein 1 (GIT1), regulates MB in the heart, likely in an eNOS-dependent manner.^287–288 Biased agonism allows for the development of ligands that selectively stimulate signaling pathways that lead to MB while inhibiting negative regulators of MB. Many GPCRs are modulated by endogenous molecules, a fact which has facilitated the development of potent and selective agonists and antagonists for various receptors. Despite the potential of GPCRs to activate pathways known to induce MB and the availability of clinically approved GPCR ligands, little investigation has occurred to explore the potential of such compounds to induce MB.

Cannabinoid-1 receptor

Cannabinoid-1 receptor (CB1R) antagonists such as taranabant (46)²⁸⁹ and rimonobant (47)²⁹⁰ were studied for anorectic effects (Figure 10). Despite the lack of a cyclic linker, 46 binds in a similar mode to 47; however, the amide group on 46 is able to engage in an extra hydrogen bonding interaction, leading to its enhanced affinity for the CB1R.^{289, 291} By inhibiting CB1R activity in the brain, these compounds can suppress appetite and cause weight loss with concomitant improvements in plasma lipid profiles. Both 46 and 47 were efficacious for inducing weight loss in wild type mice, mice fed a high fat diet, and ob/ob mice.^292–293 Inhibition of CB1R by 47 or by genetic ablation induces MB in adipose tissue and MB in a cAMP- and eNOS-dependent manner, leading to decreases in body weight and fat content.²⁹² Interestingly, 47 increased mitochondrial energy consumption did not increase mitochondrial mass in rat livers, indicating improved mitochondrial efficiency.²⁹⁴ Although both 46 and 47 were efficacious in animal models, investigation of 46 was halted in Phase III trials, and 46 was withdrawn from the market in the U.S. after initial approval as an anti-obesity drug. In humans, 47 reduced food intake and increased energy consumption to promote weight loss but caused serious side effects such as suicidal ideation and severe depression.^295,296

Figure 10.

Cannabinoid-1 Receptor antagonists.

5-Hydroxytryptamine receptors

Endogenous serotonin binds to the 5-hydroxytryptamine (5-HT) class of receptors (48, Figure 11)²⁹⁷. 5-HT receptors are primarily GPCRs that have been identified as therapeutic targets for neuropsychiatric, neurologic, and cardiac diseases. The synthetic ligand alpha-methyl-5-hydroxytryptamine (49)²⁹⁸ possesses an extra methyl group that prevents its metabolism by monoamine oxidase.²⁹⁹ The 5-HT₂ receptor agonist DOI (50)³⁰⁰ has enhanced selectivity due to its primary amine, with the iodo-group adding to its potency.³⁰¹ Much work has been done to identify and characterize the pharmacophore of 5-HT_2C receptor agonists (e.g., CP809101, 51)³⁰² and antagonists (e.g. SB242084, 52)³⁰³ and optimize their selectivity.^301–303 5-HT_2C receptor agonists stabilize the TM6 domain of the receptor through its aromatic group, whereas antagonists interact with Asn331, Val354, and Ser334 through a positively ionizable group.³⁰⁴

Figure 11.

5-Hydroxytryptamine receptor modulators that induce MB.

In addition to direct 5-HT receptor antagonists, serotonin reuptake inhibitors such as fluoxetine (53)³⁰⁵ prevent the uptake and degradation of 48 and prolong its actions at its receptors. The p-trifluoromethyl group of 53 confers selectivity for the serotonin reuptake transporter by binding to I172 in its transmembrane domain.^306–307 Treating rat pups with 53 improves mitochondrial membrane potential, respiratory capacity, and antioxidant defense in the heart, implicating 48 in mitochondrial health during development.³⁰⁸

Our laboratory identified multiple ligands that induce MB through various 5-HT receptors. In renal proximal tubule cells, we have shown that the non-selective 5-HT receptor agonist 49 induces MB.³⁰⁹ The 5-HT₂ receptor agonist 50 increased cellular respiration in vitro and improved recovery from oxidant injury by tert-butyl hydrogen peroxide (TBHP); interestingly, induction of MB did not reduce initial injury by TBHP.³¹⁰ The 5-HT_2C selective ligands 51 and 52 induce MB in vitro and in naïve mice; interestingly, siRNA studies and work in knockout mice indicate that the ligands exert these effects through the 5-HT_2A receptor.³¹¹

In contrast to 5-HT₂ receptors, the 5-HT_1F receptor has few selective ligands-namely, LY334370 (54) and LY344864 (55) and limited data regarding its pharmacophore. Nevertheless, the selective 5-HT_1F agonists 54 and 55 induced MB in vitro, and 55 also improved recovery from ischemia-reperfusion-induced AKI in vivo.³⁰⁹ Additionally, preliminary data suggest that 55 stimulates MB through the Gβγ-dependent activation of Akt and eNOS (Gibbs, W.; Beeson, C.C.; Schnellmann, R.G., unpublished results). These data indicate that the induction of MB by 5-HT agonists could be clinically useful for treating AKI and other acute organ injuries as they effectively promote recovery and regeneration even after initial injury.

Beta adrenergic receptors

The beta adrenergic receptor family is activated by endogenous stress hormones epinephrine (56)³¹² and norepinephrine (57, Figure 12)³¹² and the family comprises three receptors. First, the beta-1 adrenergic receptor, primarily expressed in the heart, is targeted by drugs that affect cardiac contractility and heart rate. The beta-2 adrenergic receptor, which is ubiquitously expressed, is a target of bronchodilators to treat asthma and COPD. The beta-3 adrenergic receptor, which is primarily expressed in adipose tissue and the urinary bladder and is targeted to treat overactive bladder.³¹³

Figure 12.

Beta adrenergic receptor modulators tested for the induction of MB.

Beta-adrenergic agonists contain distinct structural features, specifically a catechol or phenethanolamine core, whereas antagonists have a 3-aminophenoxypropan-2-ol core. However, while beta-adrenergic agonists have been extensively studied to optimize pharmacodynamics and pharmacokinetic parameters, there are few studies relating structural features to the induction of MB. Compounds 56, 57, and the non-selective beta adrenergic receptor agonist isoproterenol (58)³¹⁴ increase PGC-1α in brown adipose of naïve mice and in models of obesity in a cAMP- and p62-dependent manner.³¹⁵ Interestingly, in models of cardiac dysfunction, beta-1 adrenergic receptor stimulation by dobutamine (59)³¹⁶ increases cell death and inflammation,³¹⁷ but its blockade by the beta-1 selective antagonist metoprolol (60)³¹⁸ enhances PGC-1α activation and improves cardiac metabolism and function.^319–320 Our laboratory has studied beta-2 adrenergic receptor selective agonists in renal MB. In particular, formoterol (61),³²¹ fenoterol (62),³²² and procaterol (63)³²³ induced MB in vitro at pharmacologically relevant doses.^324–325 Compound 61 has been confirmed to induce MB in vivo in naïve mice as well as in mice subjected to AKI,³²⁶ and this was associated with improvements in renal function, indicating that formoterol has therapeutic promise for treating AKI. However, other beta-2 adrenergic receptor agonists such as clenbuterol (64)³²⁷ and isoetharine (65)³²⁸ did not induce MB in vitro,³²⁴ suggesting that biased agonism can be exploited to develop more effective mitochondrial biogenic beta-2 adrenergic receptor agonists. Because both MB-inducing and non-MB-inducing beta-2 adrenergic receptor agonists increase cAMP, we suggest that the classical Gα_s-signaling pathway is not responsible for beta-2 adrenergic receptor-induced MB in the kidney. Preliminary data suggests that 61 but not 64 activates the Akt-eNOS pathway in a Gβγ-dependent manner (Cameron, R.B.; Beeson, C.C.; Schnellmann, R.G., unpublished results). In addition to its renal effects, 61 induces MB in multiple other tissues, including the heart and skeletal muscle.^{325, 329} Together, these data indicate that certain beta-2 adrenergic receptor agonists such as 61 can be used to treat multiple diseases and improve mitochondrial function and ameliorate symptoms.

Perspectives

Because MB can arise from diverse signaling pathways, a number of drug classes have been identified to induce MB. The earliest identified inducers of MB are natural products, such as 1–9, which are efficacious,^{73, 97, 111, 120} but MB induction often occurs through multiple signaling pathways and these compounds may activate signaling programs unrelated to MB. Such promiscuity means that these compounds are poor therapeutic agents, particularly for chronic degenerative diseases for which a more targeted approach may be required.

Transcription factor activators such as TZDs (10–13), estrogens (14–18), and SIRT1 activators (19–22) induce MB by activating transcription factors that act on mitochondrial genes.^{139, 163, 173} This selectivity facilitates the induction of relatively small gene sets. Furthermore, transcription factor modulation can drive the recruitment of a select set of transcriptional machinery, increasing the specificity of the resulting transcriptome. However, ligands with that degree of specificity, particularly for MB, have not yet been designed. Thus, currently, activation of these transcription factors upregulates unwanted genes and causes detrimental neurological and hyperproliferative effects.

Similar to transcription factor modulators, kinase modulators such as 23 and 30 have been developed with a high activity for their targets. Although some kinase activators are available, many kinase inhibitors have been developed and are utilized clinically. These inhibitors will be of particular use as more negative regulators of MB, such as ERK1/2, are identified. Kinase signaling is fairly well-characterized, so acute downstream effects of such modulators are usually predictable. Nonetheless, because kinases have central roles in cellular processes, predicting longer-term effects of such drugs is not straightforward.

Cyclic nucleotide modulators such as sGC stimulators and activators and PDE inhibitors have recently been shown to be efficacious inducers of MB.^{242, 266, 285, 330} However, as with kinase modulators, these drugs influence central signaling processes, often in a manner that prevents physiological feedback loops to prevent pathological signaling. Additionally, cyclic nucleotides can have tissue-specific effects that can give rise to either injurious or curative effects to different organ systems.

GPCR modulators are the most widely developed and prescribed drug class. Although few of these compounds have been tested for MB induction, several promising classes, have been identified to induce MB, such as cannabinoid, serotonergic, and adrenergic ligands.^{292, 309–311, 324} These compounds can act through a single target and activate a particular signaling program. ²⁸⁶ Unlike the above-mentioned compound classes, GPCR ligands act at surface receptors and can retain cellular feedback mechanisms to limit signaling if necessary, so GPCR ligands represent promising chemical space for the induction of MB.

In vitro, in vivo, and human studies indicate that induction of MB promotes recovery from disease states among many organ systems due to myriad roles played by mitochondria in both physiological and pathophysiological states. However, relatively few drugs have been identified to induce MB, and much chemical space remains untested for MB. One domain of chemical space that may be promising for phenotypic screens to identify lead compounds is the so-called “dark chemical space,” as compounds derived from this space tend to have high specificity for a given target.³³¹ As more chemical space is investigated for MB, we will gain a better understanding of the role of mitochondria in health and disease and will provide researchers and clinicians with better tools for treating debilitating acute and chronic degenerative diseases.

ABBREVIATIONS

ATP

adenosine triphosphate

AMP

adenosine monophosphate

MPTP

mitochondrial permeability transition pore

MB

mitochondrial biogenesis

PGC-1α

peroxisomal proliferation activated receptor coactivator-1α

mtDNA

mitochondrial DNA

ER

estrogen receptor

ERRα

estrogen related receptor-α

NRF-1

nuclear respiratory factor 1

NRF-2

nuclear respiratory factor 2

PPAR

peroxisome proliferator-activated receptor

TR

thyroid hormone

CREB

cAMP-responsive element binding protein

YY-1

yin yang-1

ROS

reactive oxygen species

PGC-1

peroxisomal proliferation activated receptor coactivator-1

PGC-1β

peroxisomal proliferation activated receptor coactivator-1β

PRC

PGC-1 related coactivator

PKA

protein kinase A

NO

nitric oxide

AMPK

growth stimulatory AMP-dependent kinase

SIRT1

silent mating type information regulation 2 homolog 1

AKI

acute kidney injury

Tfam

mitochondrial transcription factor A

UCP2

uncoupling protein 2

PINK1

PTEN-induced putative kinase 1

NAD⁺

nicotinamide adenine dinucleotide

SOD2

superoxide dismutase 2

SIRT3

silent mating type information regulation 2 homolog 3

eNOS

endothelial nitric oxide synthase

TZD

thiazolidinedione

PPARγ

peroxisomal proliferation activated receptor-γ

ERα

estrogen receptor α

ERβ

estrogen receptor β

GPER

G protein-coupled estrogen receptor

ALDH2

aldehyde dehydrogenase 2

LDL

low density lipoprotein

ERK 1/2

extracellular signal-related kinases 1/2

MEK 1/2

mitogen-activated protein kinase kinase 1/2

sGC

soluble guanylate cyclase

PDE

phosphodiesterase

PKG

Protein kinase G

VEGF

vascular endothelial growth factor

HGF

hepatocyte growth factor

HbA_1c

glycated hemoglobin

GPCR

G protein-coupled receptor

GRK

G protein-coupled receptor kinase

GIT1

GRK interacting protein 1

CB1R

cannabinoid-1 receptor

5-HT

5-hydroxytryptamine

TBHP

tert-butyl hydrogen peroxide

Biographies

Robert B. Cameron obtained his B.S. from Davidson College where he performed research on the synthesis and characterization of light-harvesting molecules under the direction of Dr. Durwin R. Striplin. He is currently an M.D./Ph.D. candidate at the Medical University of South Carolina studying the signaling mechanisms by which mitochondrial biogenesis occurs following G protein-coupled receptor stimulation under the direction of Rick G. Schnellmann at the University of Arizona.

Craig C. Beeson obtained his organic chemistry B.S. degree from CSU, Northridge, M.S. degree from San Diego State University and Ph.D. degree from U.C. Irvine. After studying the biophysics of T-cell activation under the direction of Harden M. McConnell at Stanford University, he started his academic career in the Chemistry Department at the University of Washington, Seattle and he is now a Professor in the Drug Discovery and Biomedical Sciences Department at the Medical University of South Carolina.

Rick G. Schnellmann obtained his B.S. in Pharmacy degree from the St. Louis College of Pharmacy, St. Louis, MO and his Ph.D. degree in pharmacology and toxicology from the University of Arizona, Tucson, AZ. After a postdoctoral fellowship at Duke University in mitochondrial biology and renal toxicity, he rose through the ranks at the University of Georgia and University of Arkansas for Medical Sciences, and became Eminent Scholar, Distinguished University Professor, and Chair in the Department of Drug Discovery and Biomedical Sciences at the Medical University of South Carolina. He currently serves as dean of the College of Pharmacy for the University of Arizona.