Mitochondria are the cytoplasmic organelles of eukaryotic cells historically recognized as the mainly responsible for the production of metabolic energy as they house the oxidative phosphorylation (OXPHOS) [1]. Different metabolic pathways, such as the tricarboxylic acid (TCA) cycle, $\beta$-oxidation, urea cycle, gluconeogenesis, and synthesis and degradation of ketone bodies, among others, occur completely or partially in the mitochondria, depending on the tissue [2] (Fig. 1).

Mitochondria represent a significant source of reactive oxygen species (ROS) and are involved in calcium ions (Ca${}^{2+}$) homeostasis modulation [3]. Any disturbance in Na${}^{+}$, K${}^{+}$-ATPase function leads to Ca${}^{2+}$ overload, which can affect mitochondrial dehydrogenases and cellular energetics, hence cell death [4]. Mitochondria are crucial for storing intracellular Mg${}^{2+}$, that impacts mitochondrial function, in particular energy metabolism, mitochondrial Ca${}^{2+}$ handling, and apoptosis [5]. Furthermore, mitochondria play crucial roles in cell signaling, cell death, immune response [6, 7], aging processes, and the onset of degenerative disease [8, 9].

Mitochondria are double-membrane surrounded organelles: the outer mitochondrial membrane (OMM), freely permeable due to the presence of a large number of porins [10], and the inner mitochondrial membrane (IMM), structured in cristae, so highly impermeable that almost all ions and molecules require membrane transporters to cross it [11]. The IMM is the primary site of adenosine triphosphate (ATP) synthesis, and it has a high percentage content of proteins involved in OXPHOS and transport of metabolites between the cytosol and the mitochondria [12]. The double-membrane system demarcates two independent compartments: the intermembrane space and the mitochondrial matrix. The internal matrix contains numerous enzymes involved in oxidative metabolism as well as the mitochondrial DNA (mtDNA) [13]. Mitochondria are unique among cytoplasmic organelles since their own maternally inherited, circular, and double-stranded DNA, closely resembles a prokaryotic genome [14]. The mtDNA encodes some proteins involved in the OXPHOS system, ribosomal RNAs and most of the transfer RNAs required for protein translation within mitochondria [15].

The TCA cycle is the central hub of metabolism where molecules derived from different metabolic pathways, among which glycolysis and oxidation of fatty acids, are oxidized and the energy obtained is stored in nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FADH${}_2$). The high-energy electrons derived from NADH and FADH${}_2$ are transferred to molecular oxygen (O${}_2$) through the respiratory chain, localized in the IMM, and formed by the complexes I (NADH dehydrogenase), II (succinate dehydrogenase), III (ubiquinol-cytochrome c reductase), and IV (cytochrome c oxidase) [16]. The energy arising from the flux of electron is converted to potential energy, stored in an electrochemical gradient across the membrane, and used by the complex V (ATP synthase) to drive ATP synthesis [16]. TCA cycle intermediates serve as substrates for the biosynthesis of various molecules, including heme and macromolecules [17], and act as signaling molecules in epigenetics, immunity, hypoxic response, and cell homeostasis [18, 19, 20, 21, 22, 23].

Since O${}_2$ is the final electron acceptor of the respiratory chain, mitochondria are the main oxygen consumers (more than 90% of cell O${}_2$) and, consequently, the main producers of ROS including the radical anion superoxide (O${}_2$${}^{\mathrm{•}-}$), hydrogen peroxide (H${}_2$O${}_2$) and hydroxyl radical (OH${}^{\mathrm{•}}$), by-product of the respiratory chain [13, 24, 25]. Manganese superoxide dismutase (Mn-SOD) catalyzes the conversion of O${}_2$${}^{\mathrm{•}-}$ to H${}_2$O${}_2$, which can further be converted by mitochondrial aconitase to the highly reactive OH${}^{\mathrm{•}}$ [26]. However, the respiratory chain is not the only one source of ROS in the mitochondria: notably, cytochrome P450 (CYP) enzymes release O${}_2$${}^{\mathrm{•}-}$ and H${}_2$O${}_2$ as by-products of metabolism of different organic substrates (xenobiotics, steroid hormones and others) [27], and the enzyme monoamine oxidase (MAO) produces H${}_2$O${}_2$in the process of monoamine neurotransmitters degradation [28]. ROS are necessary for cell survival since they participate in cellular signaling processes [29], but the overproduction can cause OXPHOS dysfunction and damage to nucleic acids (DNA mutations), lipids (peroxidation of membrane lipids), proteins (glycation, oxidation, nitration of proteins, and inactivation of enzymes), therefore being associated with the development of cardiovascular, neurodegenerative and metabolic disorders, cancer, inflammation, and aging [30]. Multiple enzymatic and non-enzymatic antioxidant systems scavenge the ROS and limit their toxicity, notably glutathione peroxidase, catalase, glutathione (GSH) and vitamin E [31, 32, 33].

Mitochondria regulate different types of cell death, with an essential role in apoptosis [34]. Various cellular stresses, among which DNA damage, hypoxia, viral infection, induce the intrinsic pathway of apoptosis initiated by the opening of the mitochondrial permeability transition pore (mPTP) located in the IMM. The aperture of the mPTP leads to the IMM swelling and the rupture of the OMM, followed by the leakage of cytochrome c, apoptosis-inducing factor (AIF) and endonuclease G [34]. Cytochrome c binds and activates apoptotic protease activating factor (Apaf-1), forming a complex known as apoptosome [35]. Apoptosome binds to and activates procaspase-9 to become the initiator caspase-9, which subsequently cleaves and activates the downstream effector caspases 3, 6, and 7, inducing apoptosis. Caspase-independent nuclear translocation of AIF and endonuclease G triggers chromatin condensation and degradation. Members of the B-cell lymphoma 2 (Bcl-2) family control these events exhibiting pro- and/or anti-apoptotic properties [36]. Mitochondria take part also in other cell death mechanisms [37], among which pyroptosis [38], necroptosis, and ferroptosis [39].

Mitochondrial biogenesis and mitochondria-selective autophagy (mitophagy), which fosters the elimination of damaged mitochondria, are two opposing processes that must be finely regulated and balanced to preserve mitochondrial homeostasis and ensure cellular adaptation to metabolic shifts, stress signals and environmental stimuli, such as aerobic exercise or caloric restriction [40]. Imbalance between mitochondrial biogenesis and mitophagy leads to development of countless pathologic conditions [41].

Mitochondrial biogenesis is the physiological response to adapt the number of mitochondria to energy demands by which new organelles are formed by the growth (increase in mitochondrial mass) and division of preexisting mitochondria [42, 43]. Mitochondrial biogenesis is controlled by both nuclear and mitochondrial genomes [43]. The master regulator in mitochondrial biogenesis is peroxisome proliferator-activated receptor-gamma coactivator-1 alpha (PGC-1$\alpha$), that induces the activation of transcription factors nuclear respiratory factor 1 (NRF1) and GA-binding protein (GABP), also known as nuclear respiratory factor 2 [44, 45], leading in turn to the upregulation of mitochondrial transcription factor A (TFAM) as well as other nuclear genes encoded mitochondrial subunits of electron transport chain complexes [46]. TFAM translocates to the mitochondria, where it stimulates the replication and transcription of mtDNA [47, 48]. TFAM is also involved in mtDNA repair [49]. PGC-1$\alpha$ is deacetylated by nicotinamide adenine dinucleotide (NAD${}^{+}$)-dependent deacetylase sirtuin 1 (SIRT1), when cellular NAD${}^{+}$ levels increase in an adenosine monophosphate (AMP)-activated protein kinase (AMPK)-dependent manner [50]. Thus, the modulation of the axes AMPK/SIRT1/PGC-1$\alpha$ represents a pharmacological target to modulate the mitochondrial biogenesis [50]. Impaired mitochondrial biogenesis contributes in pathogenesis of cardiovascular [51] and neurodegenerative diseases, including Alzheimer’s disease (AD) [52], Parkinson’s disease (PD), Huntington’s disease (HD) and amyotrophic lateral sclerosis (ALS) [8, 9].

Mitophagy is the process of selective degradation of damaged and dysfunctional mitochondria through autophagy [53]. Mitochondria are enclosed by a membrane, forming autophagosomes which fuse with lysosomes for hydrolytic degradation. Phosphatase and tensin homolog (PTEN)-induced putative protein kinase 1 (PINK1)/Parkin pathway is the main molecular mechanisms regulating mitophagy [54]. PINK1 is a serine/threonine kinase usually present in low levels. When any damage occurs to mitochondria like mutations in mtDNA, accumulation of misfolded proteins, increased ROS, or depolarization, PINK1 accumulates at the OMM. PINK1 phosphorylates ubiquitin on Ser65 and recruits Parkin from the cytosol, phosphorylating and activating it. Parkin is an E3-ubiquitin ligase that, when activated, triggers the ubiquitination of a wide range of OMM proteins. These events result in the recruitment of autophagy receptors, such as p62/sequestosome 1 (SQSTM1), that facilitate the sequestering of damaged mitochondria within autophagosomes [55]. SQSTM1 comprises a domain interacting with polyubiquitinated proteins on the mitochondrial surface and a motif binding to microtubule-associated protein 1A/1B-light chain 3-II (LC3-II) present on autophagosomes [56]. Mitophagy can occur independently from functional PINK1, allowing ubiquitin-independent recognition of damaged mitochondria. In mammals the known receptors for receptor-mediated mitophagy include Bcl-2 interacting protein 3 (BNIP3), its analog NIX (BNIP3L) and FUN14 domain-containing protein 1 (FUNDC1) [54, 57]. These adapters detect mitochondrial damage and guide dysfunctional mitochondria to the autophagosome by changing their subcellular location or the proteins they interact with. FUNDC1 is a mitochondrial protein located in the OMM protein sensitive to hypoxia. The affinity of FUNDC1 for LC3 is regulated by phosphorylation at Ser17 and Ser13 under different stresses. While phosphorylation at Ser17 activates mitophagy, phosphorylation at Ser13 inhibits this process [55, 58, 59]. Mitophagy declines with aging determining accumulation of defective mitochondria, increased oxidative stress, and higher cell apoptosis [60].

In keeping with the multiple functions, mitochondria can change rapidly both form (size, shape, and position) and number to meet the physiological needs of cells undergoing the two opposing processes of fission and fusion [61, 62]. Mitochondrial homeostasis is strictly dependent on the equilibrium between fusion and fission dynamics which are important for mitochondrial inheritance and maintenance of functions (e.g., efficient OXPHOS, transport and regulation of mitophagy). Fission involves the separation of both OMM and IMM and their subsequent rejoining in the correct orientation, besides the properly partitioning of mitochondrial proteins and mtDNA so that each daughter organelle can function normally. In physiological condition, fission happens prior to cell division to guarantee that organelles are equally distributed between daughter cells [62]. Fission is also required as a preliminary event to mitophagy, to allow for segregation of damaged organelles, their inclusion in autophagosomes and finally their degradation [63, 64]. Mitochondrial dynamics require specialized nuclear-encoded protein, mainly represented by GTP-hydrolyzing proteins (GTPases) belonging to the dynamin superfamily. Three central players are (1) mitofusin 1 and mitofusin 2 (OMM fusion) [65], (2) optic atrophy 1 (OPA1) (IMM fusion) [66], and (3) dynamin-related protein 1 (Drp1) (division of IMM and OMM) [67]. Perturbations in mitochondrial dynamics have detrimental consequences on bioenergetic supply and represent a common feature of different neurogenerative disorders such as AD, PD, and HD [9].

CA exerts cytoprotection in several in vitro experimental models involving brain cells [89, 144, 145]. CA induces antioxidant effects mainly by a mechanism associated with the upregulation of both non-enzymatic and enzymatic antioxidant defenses in neurons [144] (Fig. 4). Satoh et al. [146] reported that CA at 10 µM for 6–48 h stimulated the antioxidant-responsive element (ARE)/nuclear factor erythroid 2-related factor 2 (Nrf2), upregulating the enzymes heme oxygenase-1 (HO-1), NAD(P)H quinone oxidoreductase-1 (NQO-1), and also the catalytic (GCLC) and modifier (GCLM) subunits of $\gamma$-glutamate-cysteine ligase ($\gamma$-GCL, a major enzyme involved in the production of GSH, the major non-enzymatic antioxidant in animal cells) in the rat pheochromocytoma subclone PC12h cells. Moreover, Kosaka et al. [147] observed that CA promoted neurite outgrowth in the PC12h cells by a mechanism dependent on Nrf2, suggesting that this diterpene would be able to modulate neuronal differentiation and plasticity. CA at 10 µM also upregulated GSH production in the mouse hippocampal HT22 cell line exposed to glutamate [148], suggesting that the effects promoted by CA are not limited to just one cell type. Similarly, Chen et al. [149] reported that CA at 1 µM for 6–24 h stimulated the production of GSH and upregulated the GCLC and GCLM subunits in the human dopaminergic SH-SY5Y cells. The authors also observed that CA stimulated the translocation of Nrf2 to the cell nucleus, enhancing the activity of ARE. Moreover, CA suppressed the mitochondria-related apoptosis triggered by 6-hydroxydopamine (6-OHDA) by suppressing caspase-3 activation and poly (ADP-ribose) polymerase (PARP) cleavage in SH-SY5Y cells. Interestingly, blockade of the synthesis of GSH by using L-buthionine sulfoximine (BSO) abrogated the anti-apoptotic effects induced by CA in that experimental model. Fig. 4 summarizes these important findings related to the metabolism of GSH in CA-treated cells. In agreement with these data, Lin et al. [150] showed that CA at 1 µM for 18 h stimulated the phosphatidylinositol 3-kinase (PI3K)/Akt/NF-$\kappa$B signaling pathway and upregulated the Pi form of glutathione-S-transferase (GSTP) in the SH-SY5Y cells. Silencing of GSTP abrogated the mitochondria-related anti-apoptotic effect (i.e., suppression of caspase-3 activation) caused by CA in the 6-OHDA-challenged SH-SY5Y cells. GSTP depends on GSH to mediate phase II detoxification reactions [151]. Thus, it seems that the CA-induced upregulation of the production of GSH prevents cells collapse by a mechanism that does not depend only on the antioxidant role of that tripeptide. The same research group reported that CA at 1 µM attenuated the 6-OHDA-induced increase in the immunocontent of Bcl-2-associated X protein (Bax) (which mediates apoptosis by interacting with the mitochondria) and the decrease in the levels of Bcl-2 (which blocks apoptosis after interacting with the mitochondria) in the SH-SY5Y cells [152]. Bax activator molecule 7 (BAM7), an activator of Bax, abrogated the CA-induced anti-apoptotic effects in the 6-OHDA-treated SH-SY5Y cells, indicating that the mitochondria are central in the mechanism by which CA promotes anti-apoptotic actions.

Even though the existence of findings indicating that CA would be capable to promote neuroprotective actions, it was not demonstrated whether those effects would be linked to the ability of this diterpene in modulating mitochondrial physiology (which involves bioenergetics, redox biology, dynamics, and biogenesis of the organelles). Furthermore, it was not clearly demonstrated whether Nrf2 (or its downstream targets) would present a role in the mechanism promoted by CA with regard to the mitochondria. Thus, it was investigated whether CA would prevent mitochondrial dysfunction in different in vitro experimental designs related to neurotoxicity. We found that CA at 1 µM for 12 h attenuated the mitochondrial dysfunction induced by the agrochemical paraquat (a mitochondrial toxicant) on the human dopaminergic SH-SY5Y cells [153]. CA blocked the Bax upregulation induced by paraquat, consequently decreasing the release of cytochrome c from the mitochondria to the cytosol. Moreover, CA attenuated the paraquat-induced upregulation in the activity of the pro-apoptotic caspases-9 and -3, resulting in a decrease in the fragmentation of DNA. CA significantly reduced the impact caused by paraquat on the activity of the Complexes I and V, as well as on the mitochondrial membrane potential (MMP) and production of ATP, indicating a role for CA in rescuing mitochondrial function (Fig. 5). In the same work, we demonstrated that CA prevented the paraquat-induced increase in the mitochondrial levels of malondialdehyde (an index of lipid peroxidation) and protein carbonylation. These effects may be linked to the ability of CA in upregulating the levels of GSH and Mn-SOD in the mitochondria of the SH-SY5Y cells (Fig. 6). Nonetheless, it was not examined in that work. It was found that blockade of PI3K/Akt or knocking down of the transcription factor Nrf2 suppressed the mitochondria-related cytoprotection caused by CA in that experimental model, suggesting that the PI3K/Akt/Nrf2 axis would be involved in mediating the effects promoted by CA in the SH-SY5Y cells. Similarly, de Oliveira et al. [154] reported that blockade of the PI3K/Akt/Nrf2 axis attenuated the effects caused by CA on the mitochondria of the SH-SY5Y cells challenged with methylglyoxal, a dicarbonyl commonly associated with diabetes mellitus and neurodegeneration [155] (Fig. 5).

CA is a potent activator of Nrf2, leading to the expression of several enzymes associated with cytoprotection in different cell types [146, 149]. The enzyme HO-1, one of the most studied downstream targets of Nrf2, mediates heme degradation generating biliverdin, carbon monoxide (CO), and free iron and promoting cytoprotection, mainly by antioxidant and anti-inflammatory-related mechanisms, in different human cell types [156]. Biliverdin generated by HO-1 is further converted into bilirubin by the enzyme biliverdin reductase (BVR) [157]. Since it was showed that CA upregulated HO-1 in previous works [146, 153], we investigated whether this enzyme would exert a role in mediating the protection of mitochondria in CA-treated SH-SY5Y cells exposed to paraquat [158]. It was found that the inhibition of HO-1 by using zinc protoporphyrin-IX (ZnPP-IX) blocked the preventive effects caused by CA on the mitochondrial function and redox parameters (please, see Table 2 (Ref. [149, 150, 152, 153, 154, 158, 159, 160, 161, 162, 163, 164, 165, 166]) for detailed information). Similarly, inhibition of HO-1 abrogated the effects stimulated by CA in SH-SY5Y cells challenged with glutamate [167]. The authors also found that a treatment with either bilirubin or carbon monoxide-releasing molecule-2 (CORM-2, which generates CO) alleviated the alterations caused by glutamate in the mitochondria, suggesting a link between the products generated by the HO-1/BVR axis in mediating the protection of the organelles (Fig. 7).

Activation of Nrf2 by CA also can lead to increased production of GSH, as previously reported by different research groups [146, 148, 149]. Thus, we investigated whether the CA-induced upregulation in the Nrf2/GSH axis would be associated with the benefits caused by CA in the mitochondria of SH-SY5Y cells exposed to methylglyoxal, whose biotransformation depends on the availability of GSH [155]. It was found that GSH production blockade, by using BSO, suppressed the actions promoted by CA on the mitochondria of the methylglyoxal-treated cells [159]. CA failed to prevent the redox impairment in the membranes of mitochondria extracted from the methylglyoxal-treated cells exposed to BSO. BSO also suppressed the CA-stimulated effects on the function of mitochondria (which was assessed through the quantification of the activity of the Complexes I and V, of MMP and of the levels of ATP) in the cells challenged with methylglyoxal. Using a different approach, Tamaki et al. [148] showed that CA activated the metabolism of GSH in the HT22 cells, leading to cytoprotection against glutamate. Similarly, Nishimoto et al. [168] reported that upregulation of GSH attenuated the pro-apoptotic stimuli caused by methylglyoxal in the SH-SY5Y cells. Since GSH also exerts signaling functions in mammalian cells [169], it should be further examined whether other mechanisms, in addition to those related to antioxidant and detoxifying defenses, would be associated with the mitochondria-related benefits promoted by GSH in brain cells (Fig. 8).

In addition to promoting antioxidant and detoxifying defenses that lead to mitochondrial protection, CA also affords cytoprotection by modulating other aspects related to mitochondrial physiology. Lin et al. [160] published that CA prevented the pro-apoptotic effects caused by 6-OHDA in SH-SY5Y cells by a mechanism involving upregulation of PINK1 and parkin. The authors also found that silencing of parkin suppressed the mitochondria-related anti-apoptotic effects caused by CA in the 6-OHDA-challegned cells. PINK1, a serine/threonine kinase associated with the IMM, is responsible for recruiting parkin for damaged mitochondria leading to the removal of the dysfunctional organelles from the cells by mitophagy [170]. Proteins linked to ubiquitin are digested by the ubiquitin-proteasome system (UPS), allowing clearance of damaged proteins [171]. Mutations in PINK1 or parkin may lead to parkinsonism due to a failure to maintain mitophagy, among other consequences, such as redox impairment and bioenergetics collapse [172]. Lin and Tsai [161] found that CA stimulated mitophagy by activation of the PINK1/parkin axis, preventing the mitochondria-related apoptosis in the 6-OHDA-treated SH-SY5Y cells (Fig. 9). Silencing of PINK1 abrogated the upregulation of parkin and of voltage-dependent anion channel 1 (VDAC1, which is present in the outer membrane of mitochondria and serve as a target of parkin after mitochondrial depolarization) observed in the cells pretreated with CA and challenged with 6-OHDA. Thus, CA promoted neuroprotection by depending on the stimulation of the PINK1/parkin-dependent mitophagy.

Lin et al. [162] reported that CA stimulated PINK1 leading to Mic 60 phosphorylation at a threonine residue in SH-SY5Y cells administrated with 6-OHDA. Mic60 is part of the mitochondrial contact site and cristae junction organizing system (MICOS), which is responsible for the maintenance of mitochondrial architecture regarding cristae plasticity [173]. Moreover, Mic60 also binds to PINK1 and stabilizes this protein on the surface of damaged mitochondria, favoring the binding to parkin and the triggering of mitophagy [174]. Blockade of this interaction leads to mitochondrial impairment and cell death [174]. Furthermore, Mic60 is a target for protein kinase A (PKA), that phosphorylates Mic60 at a serine residue, affecting the interaction of this protein with PINK1 and promoting several negative consequences, including the suppression of the translocation of parkin to the injured mitochondria and the triggering of cytochrome c release to the cytosol [175]. The effects seen by the authors in the cells administrated with CA indicate that this diterpene downregulated PKA, stimulating the Mic60/PINK1-mediated mitophagy and rescuing neurons from the mitochondria-dependent cell death triggered by 6-OHDA (Fig. 10).

Parkin participates in the control of autophagy also by interacting with Beclin1, which is central for the formation of the autophagosome [176]. For example, silencing of Beclin1 blocked the degradation and consequent removal of $\alpha$-synuclein in the dopaminergic PC12 cell line [177]. In that context, Lin and Tsai [163] have reported that the silencing of parkin abrogated the CA-dependent upregulation in Beclin1 in the 6-OHDA-treated SH-SY5Y cells. Similarly, administration of bafilomycin A1 (an inhibitor of the fusion between the autophagosome and the lysosome) or wortmannin (a compound that blocks the early stage of the formation of the autophagosome) suppressed the anti-apoptotic actions related to the mitochondria promoted by CA in the cells challenged with 6-OHDA. Then, it may be suggested that CA depends on the stimulation of autophagy to prevent cell death triggered by 6-OHDA.

Interestingly, Sun et al. [178] demonstrated that Beclin1 can improve mitochondria-associated membranes (MAMs) in the heart of mice submitted to an experimental model of endotoxemia. MAMs may be characterized as regions in which it is observed a close physical connection involving the outer mitochondrial membrane and other biomembranas intracellularly [179]. MAMs are important to the maintenance of communication between the mitochondria and the endoplasmic reticulum necessary to the transport of Ca${}^{2+}$ and lipids, for example [180, 181]. Nonetheless, it remains to be examined whether CA would be capable to modulate the MAMs-related physiology in brain cells.

Parkin has been considered as a cytoprotective protein also due to its ability to control the apoptosis-related protein in the transforming growth factor-$\beta$ (TGF-$\beta$) signaling pathway (ARTS)/X-liked inhibitor of apoptosis protein (XIAP) pathway [182]. The protein ARTS is inserted in the outer mitochondrial membrane and, in response to pro-apoptotic stimuli, migrates to the cytosol, binding and blocking XIAP, which, in turn, exerts an anti-apoptotic action by inhibiting caspases-3, -7, and -9 [183]. With regard to CA, it was examined whether this diterpene would be capable to counteract the pro-apoptotic effect triggered by 6-OHDA in the SH-SY5Y cells by a mechanism dependent on the parkin/ARTS/XIAP axis. Fu et al. [164] observed that a pretreatment with CA attenuated the 6-OHDA-induced ARTS upregulation and XIAP downregulation in the cells treated with 6-OHDA. Moreover, CA prevented the activation of the caspases-9 and -7 in that work. Silencing of parkin blocked the effects promoted by CA on levels and activation of ARTS, XIAP, and caspases-9 and -7 in the cells challenged with 6-OHDA, leading to cell death (Fig. 11). Thus, in addition to the role of parkin in modulating mitophagy, this protein is associated with the mitochondria-dependent ARTS/XIAP axis involved in the control of apoptosis in CA-treated cells. Together, these data indicate a complex mechanism by which CA promotes neuroprotection by both mitochondria-dependent and independent manners.

Mitochondrial dynamics, i.e., fusion (which is the generation of one mitochondrion by the fusion of two mitochondria) and fission (when one mitochondrion divides generating two mitochondria), alterations have been seen in several disorders, including neurodegeneration and cardiovascular diseases [184, 185]. The proteins Drp1 and fission 1 (Fis1) mediates fission, whereas the proteins mitofusin 1 and 2 (MFN1 and MFN2, respectively) and OPA1 modulate mitochondrial fusion [186]. OPA1, which is located in the IMM, has been reported as a target of CA through the stimulation of the parkin/I$\kappa$B kinase (IKK)/NF-$\kappa$B signaling pathway [165]. The authors have shown that CA depended on both parkin and OPA1 to block the mitochondria-related pro-apoptotic effects elicited by 6-OHDA in the SH-SY5Y cells. CA upregulated the expression of OPA1 by a mechanism associated with the stimulation of NF-$\kappa$B, a major transcription factor involved with the maintenance of mitochondrial dynamics [187] (Fig. 12). It is worthy of mention that OPA1 is not associated only with the regulation of protein dynamics, since this protein has been viewed as a neuroprotectant agent in several experimental models [188, 189]. Actually, downregulation of OPA1 leads to loss of mitochondrial architecture and, consequently, mitochondrial dysfunction, cytochrome c release and apoptosis in cultured cells [190]. Therefore, CA is a potential neuroprotectant also due to its ability in controlling mitochondrial dynamics and causing pro-survival effects.

In addition to the ability to modulate mitochondrial dynamics, CA also promoted mitochondrial biogenesis in in vitro experimental models. Lin et al. [166] observed that a pretreatment with CA attenuated the downregulation induced by 6-OHDA on the proteins NRF1 and TFAM, that modulates mitochondrial biogenesis [43, 191]. CA also upregulated PGC-1$\alpha$, a transcription factor that controls the expression of NRF1 and TFAM [43]. In turn, PGC-1$\alpha$ expression is controlled by parkin-interacting substrate (PARIS), a peptide that, under physiological conditions, is ubiquitinated by parkin and follows to degradation by the proteasome [192]. In that context, the authors found that CA attenuated the 6-OHDA-induced upregulation of PARIS by a mechanism associated with parkin. Thus, CA stimulated the parkin/PARIS/PGC-1$\alpha$ pathway leading to mitochondrial biogenesis and mitochondria-related anti-apoptotic actions in the SH-SY5Y cells challenged with 6-OHDA (Fig. 13). Indeed, mitochondrial biogenesis has been viewed as a potential pharmacological target to either prevent or manage neurodegenerative disorders [193].

The ability of CA in modulating mitochondria-related parameters in mammalian brain was tested in different experimental models. Prior studies have proved that CA affects the function and dynamics of mitochondria and, therefore, is a potential agent to be utilized as a mitochondria-related protectant in brain disorders.

CA exhibits a potent antioxidant action by modulating both non-enzymatic and enzymatic antioxidant defenses in mammalian cells [89, 156]. In a study conducted in aged rats, it was found that supplementation with rosemary extract containing different concentrations of CA improved antioxidant defenses in the rat brain. Rats exposed to rosemary extract showed decreased levels of ROS, lipid peroxidation (LPO), catalase (CAT), and nitric oxide synthase (NOS), indicating a better redox balance in comparison with the control [194]. In another study, the authors reported that an extract of rosemary presenting 60% CA decreased the levels of 4-hydroxynonenal (4-HNE, a LPO indicator) in the cerebral cortex of the senescence accelerated mice (SAMP8) [195]. The authors also demonstrated that rosemary extract combined with 10% CA significantly attenuated the levels of protein carbonylation in the mice hippocampus.

CA presents a catechol loop that can be oxidized to form an electrophilic quinone [146]. This quinone can interact with particular cysteine bonds on target proteins through thiol S-alkylation. The primary response against certain reactive agents, such as ROS and electrophiles, is the Kelch-like ECH-associated protein 1 (Keap1)/Nrf2 pathway [196]. Satoh et al. [146] provided the first description of the role of the Nrf2 signaling in mediating the neuroprotective action of CA against oxidative damage and excitotoxicity in an experimental model of middle cerebral artery occlusion (MCAO). The authors demonstrated that the crucial cysteine thiol of the Keap1 protein was alkylated by the electrophilic quinone-type of CA, which triggered the activation of the transcription element Nrf2, and upregulated several cytoprotective enzymes, as commented before.

Numerous studies showed that CA stimulates Nrf2, thus promoting cytoprotective benefits. It was observed that an in vivo treatment with CA at 1 mg/kg attenuated the mitochondrial impairment caused by the in vitro exposure of rat cerebral cortex to 4-HNE in an ex vivo experimental model [197]. The authors also reported that administering CA in vivo, reduced the levels of 4-HNE associated with mitochondrial proteins probably by a mechanism related to the stimulation of the Nrf2/ARE/HO-1 pathway. It was also shown that a combination of CA and sulforaphane considerably reduced the induced-4-HNE decline in the respiration rates of mitochondria. CA also attenuated mitochondrial dysfunction in an experimental model of controlled cortical impact (CCI) in mice. Post-administration of 3 mg/kg CA preserved the respiratory action of mitochondria and alleviated oxidative damage by lowering the protein nitration levels, LPO, and cytoskeletal disintegration in the CCI mice [198].

The effects of CA therapy in a recent study included the stimulation of the HO-1, thioredoxin-1, Nrf2, and brain-derived neurotrophic factor (BDNF) expressions, as well as an increase in the brain serotonin levels in ovariectomized mice [199]. In the same study, CA attenuated oxidative stress, as well as caused a decline in the expression of tumor necrosis factor-$\alpha$ (TNF-$\alpha$), interleukin-1$\beta$ (IL-1$\beta$), and inducible NOS (iNOS). Even though Nrf2 and BDNF are linked and it has been documented, leading to antidepressant-like effects in rodents [200], it was not investigated yet regarding the effects caused by CA in the mammalian brain.

Additionally, CA reduced inflammation in mice subjected to the pesticide chlorpyrifos (CPF) [201]. A pretreatment with CA reduced the activity of IL-6, TNF-$\alpha$, and IL-1$\beta$ in the serum of the CPF-administered mice. CA also inhibited the CPF-induced decline in the levels of acetylcholinesterase (AChE) in the mice serum. Besides, CA-treated animals showed decreased amounts of NO${}^{\mathrm{•}}$ and malondialdehyde (MDA) along with increased GSH, glutathione peroxidase (GPx), superoxide dismutase (SOD), and CAT levels in the brain and eye tissues. Similarly, in another study it was observed that protein levels of IL-1$\beta$, IL-6, and TNF-$\alpha$ were significantly lowered in the brain of CA-treated mouse submitted to an experimental model of high-fat diet (HFD)-induced inflammation. Treatment with CA attenuated brain injury by upregulating the Bcl-2 protein and downregulating the Bax and matrix metallopeptidase 9 (MMP9) proteins [202]. The neuro-inflammatory markers such as glial fibrillary acidic protein (GFAP) and ionized calcium-binding adapter molecule 1 (Iba1) were found to be decreased with the post-administration of CA in an experimental model of traumatic brain injury (TBI) [203]. In the same study, authors also reported that CA did not affect the changes in mitochondrial respiration after TBI.

The most widely used drug for the development of PD model is 6-OHDA, a neurotoxic organic compound that specifically kills dopaminergic and noradrenergic neurons [204]. It was found that CA increased the GSH levels with decrease in lipid peroxidation levels along with enhanced protein expression of GCLC, GCLM, SOD, and glutathione reductase (GR) in rats exposed to 6-OHDA in PD experimental model. Simultaneously, CA deactivated c-Jun NH2-terminal kinase and p38, upregulated the Bcl-2/Bax ratio, downregulated the cleaved caspase 3/caspase 3 and cleaved PARP/PARP ratios, and upregulated tyrosine hydroxylase (TH) protein in the brain of 6-OHDA lesioned rats [152].

The mitochondria-related anti-apoptotic CA-induced effect has also been demonstrated in an animal model of neurotoxicity induced by acrylamide. Treatment with 40 mg/kg CA for 11 days enhanced the GSH content and lowered the levels of MDA, the Bax/Bcl-2 ratio, and the activation of caspase-3 in the brain of the rats exposed to acrylamide [205]. Another study showed that CA, in combination with Trolox and human chorionic gonadotropin (HCG), reduced the activity of caspase-3 and Bax in the hippocampus of rats administered to an experimental model of ischemia-reperfusion induced by occlusion of the carotid artery [206]. It was also reported that CA upregulated the levels of SIRT1, Mn-SOD, and Bcl-2, while downregulated the expression of p66shc and Bax and the activation of caspase-3 in an experimental model of subarachnoid haemorrhage (SAH) in rats [207].

Overall, it has been demonstrated both CA can modulate mitochondrial physiology by both indirect and direct mechanisms in the brain cells also in vivo (Fig. 14; Table 3, Ref. [146, 152, 160, 194, 195, 197, 198, 199, 201, 203, 205, 206, 207]). Further studies are necessary to confirm this ability in other contexts involving mitochondrial disturbances.

The constellation of effects induced by CA on the mitochondria clearly demonstrates that this diterpene is a potential candidate to be utilized as a drug in mitochondrial therapy. Even though some of the findings were confirmed in in vivo experimental models, further investigations are necessary to better understand CA bioavailability and to examine whether CA would exert toxic effects in brain cells (as well as in other organs), for example. Moreover, more research needs to be done on this diterpene’s potential to alter some signaling pathways connected to mitochondria in vivo. Overall, in various brain cells, CA has been shown to be a powerful agent that protects the mitochondria.

4-HNE, 4-hydroxynonenal; 6-OHDA, 6-hydroxydopamine; AChE, acetylcholinesterase; ACLY, ATP-citrate lyase; AD, Alzheimer’s disease; AIF, apoptosis-inducing factor; ALS, amyotrophic lateral sclerosis; AMP, adenosine monophosphate; AMPK, AMP-activated protein kinase; Apaf-1, apoptotic protease activating factor; ARE, antioxidant-responsive element; ARTS, apoptosis-related protein in the TGF-$\beta$ signaling pathway; ATP, adenosine triphosphate; BDNF, brain-derived neurotrophic factor; BNIP3, BCL2 interacting protein 3; BSO, L-buthionine sulfoximine; BVR, biliverdin reductase; CA, carnosic acid; CAT, catalase; CCI, controlled cortical impact; CO, carbon monoxide; CORM-2, carbon monoxide-releasing molecule-2; CPF, chlorpyrifos; CYP, cytochrome P450; Drp1, dynamin-related protein 1; FADH${}_2$, flavin adenine dinucleotide, reduced form; Fis1, fission 1; FUNDC1, FUN14 domain-containing protein 1; GABP, GA-binding protein; $\gamma$-GCL, $\gamma$-glutamate-cysteine ligase; GCLC, glutamate-cysteine ligase catalytic subunit; GCLM, glutamate-cysteine ligase modifier subunit; GPx, glutathione peroxidase; GR, glutathione reductase; GSH, glutathione, reduced form; GSTP, Pi form of glutathione-S-transferase; GTPase, GTP-hydrolyzing proteins; H${}_2$O${}_2$, hydrogen peroxide; HCG, human chorionic gonadotropin; HD, Huntington’s disease; HO-1, heme oxygenase-1; IKK, I$\kappa$B kinase; IMM, inner mitochondrial membrane; iNOS, inducible nitric oxide synthase; IL-1$\alpha$, interleukin-1$\alpha$; IL-6, interleukin-6; Keap1, Kelch-like ECH-associated protein 1; MAMs, mitochondria-associated membranes; MAO, monoamine oxidase; MCAO, middle cerebral artery occlusion; MDA, malondialdehyde; MICOS, mitochondrial contact site and cristae junction organizing system; MFN1, mitofusin 1; MFN2, mitofusin 2; MMP, mitochondrial membrane potential; MMP9, matrix metallopeptidase 9; Mn-SOD, manganese-dependent superoxide dismutase; mPTP, mitochondrial permeability transition pore; mtDNA, mitochondrial DNA; NADH, nicotinamide adenine dinucleotide; NF-$\kappa$B, nuclear factor-$\kappa$B; NO${}^{\mathrm{•}}$, nitric oxide; NOS, nitric oxide synthase; NQO1, NAD(P)H quinone oxidoreductase-1; NRF1, nuclear respiratory factor 1; Nrf2, nuclear factor erythroid 2-related factor 2; O${}_2$${}^{\mathrm{•}-}$, radical anion superoxide; OH${}^{\mathrm{•}}$, hydroxyl radical; OMM, outer mitochondrial membrane; OPA1, optic atrophy 1; OXPHOS, oxidative phosphorylation; PARIS, parkin-interacting substrate; Parkin, E3 ubiquitin ligase; PARP, poly (ADP-ribose) polymerase; PD, Parkinson’s disease; PGC-1$\alpha$, peroxisome proliferator-activated receptor-gamma coactivator-1 $\alpha$; PGE${}_2$, prostaglandin E${}_2$; PI3K, phosphatidylinositol 3-kinase; PINK1, PTEN-induced putative protein kinase 1; PKA, protein kinase A; PPAR$\gamma$, peroxisome proliferator-activated receptor $\gamma$; ROS, reactive oxygen species; SAH, subarachnoid haemorrhage; SAMP8, senescence accelerated mice; SIRT1, NAD${}^{+}$-dependent deacetylase sirtuin 1; SOD, superoxide dismutase; SQSTM1, sequestosome 1; TBI, traumatic brain injury; TCA, tricarboxylic acid; TFAM, mitochondrial transcription factor A; TH, tyrosine hydroxylase; TNF-$\alpha$, tumor necrosis factor-$\alpha$; UPS, ubiquitin-proteasome system; VDAC1, voltage-dependent anion channel 1; XIAP, X-liked inhibitor of apoptosis protein; ZnPP-IX, zinc protoporphyrin-IX.

Conceptualization: MRdeO, VI, IP, AS, ST, GS; writing—original draft preparation: VI, IP, AS, ST, GS, MRdeO; writing—review and editing: VI, IP, AS, ST, GS, MRdeO. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

Not applicable.

Not applicable.

MRdeO receives a “Bolsa de Produtividade em Pesquisa (Research Productivity Grant) 2—PQ2” fellow from the CNPq (protocol number 301273/2018-9).

The authors declare no conflict of interest. Marcos Roberto de Oliveira is serving as one of the Guest editors of this journal. We declare that Marcos Roberto de Oliveira had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to Gernot Riedel.