Signaling Pathways Underlying Anthracycline Cardiotoxicity

The ROS-driven hypothesis

The susceptibility of cardiomyocytes to ROS-induced damage is related to their limited antioxidant resources. As a consequence of their prevalent oxidative metabolism, the antioxidant systems of cardiomyocytes are almost completely saturated at steady state (36) and any further source of ROS, such as ANTs, can exhaust their antioxidant capacity more rapidly than in any other cell type. ANTs are responsible for ROS generation since they are reductively activated to a semiquinone radical. This, in turn, undergoes redox cycling, thereby producing superoxide (O²⁻) and hydrogen peroxide (H₂O₂). A further mechanism of ANT-mediated ROS production involves iron. In the presence of iron, ANTs can form Fe³⁺-ANT complexes, which catalyze the Fenton reaction, where H₂O₂ is converted to various ROS, including the cytotoxic hydroxyl radicals (OH⁻) (22, 121) (Fig. 1). The ensuing massive ROS generation ultimately leads to cardiac cell death (17, 115).

OPEN IN VIEWER

FIG. 1.

The ROS-driven hypothesis of ANT-induced injury. The classic “ROS-driven hypothesis” points to the generation of ROS by the quinone moiety of ANTs as the leading cause of ANT cardiotoxicity. ANT, anthracycline; ROS, reactive oxygen species. Color images are available online.

Although ROS have long been recognized as pivotal mediators of ANT cardiotoxicity, the initial ROS-driven hypothesis has been challenged by a series of studies showing modest effects of ROS scavengers and iron chelators in preventing ANT cardiac side effects in chronic settings. Importantly, N-acetyl cysteine or combinations of antioxidants have been tested, with variable outcomes, but did not protect patients from ANT cardiotoxicity (114). Intriguingly, none has matched the effectiveness of dexrazoxane (ICRF-187 or DXR), a bisdioxopiperazine agent structurally related to ethylenediaminetetraacetic acid but with more potent chelating/antioxidant activities (118) (Table 1). This finding led to the hypothesis that the mechanism of action of DXR might be other than iron binding.

Yet, the contribution of redox cycling and primary ROS production to ANT cardiotoxicity is debated, and the most recent view is that the enhanced ROS generation is a secondary phenomenon resulting from mitochondrial dysfunction (48) or endoplasmic reticulum stress caused by accumulation of misfolded proteins (83).

Topoisomerase 2 poisoning and inhibition

In 2012, Zhang et al. showed that the production of ROS could also be secondary to the interaction of ANTs with the beta isoform of topoisomerase 2 (Top2) (143). DNA topoisomerases 2 are responsible for the catalysis of the transient breaking and adjoining of DNA double helix during DNA replication and transcription (18). As an intermediate of this process, a DNA-Top2 covalent complex, the so-called cleavage complex, is formed. ANTs act like Top2 poisons in stabilizing the complex, thereby preventing the religation of DNA and leading to the formation of DNA double-strand breaks, which are potent inducers of the DNA damage response (Fig. 2).

OPEN IN VIEWER

FIG. 2.

New mechanisms of ANT-induced injury to cardiac cells. According to the most widely accepted view, ANT toxicity is caused by poisoning of Top2-β, which leads to accumulation of double-stranded DNA breaks. This results in the activation of the p53 tumor-suppressor pathway and the ensuing mitochondrial dysfunction, ROS generation, imbalanced prosurvival/proapoptotic signals (Bcl2/BAX and mTOR), and, finally, to cardiac cell death (left panel). DNA damage simultaneously impairs mitochondrial biogenesis by affecting PGC-1, NRF1, and TFAM (right panel). BAX, Bcl2-associated X; Bcl2, B-cell lymphoma 2; mTOR, mammalian target of rapamycin; NRF1, nuclear respiratory factor 1; PGC-1, PPARγ coactivator 1; PPARγ, peroxisome proliferator-activated receptor-γ; TFAM, mitochondrial transcription factor A; Top2, type 2 topoisomerase. Color images are available online.

The DNA damage response to genotoxic stress is mainly regulated by three members of the phosphatidylinositol 3-kinase related kinase family: DNA-dependent protein kinase, ataxia-telangiectasia-mutated (ATM), and ATM and Rad3-related (ATR) (62, 66). The initial processing of double-strand DNA breaks, before repair, is mediated by the heterotrimeric Mre11-Rad50-Nbs1 (MRN) complex. Upon recognition of double-strand breaks by the MRN complex, ATM kinase is activated, leading to the phosphorylation of key regulators of cell cycle progression and DNA repair (78). However, if double-strand breaks are not properly repaired, DNA damage activates a set of apoptotic signaling routes, including the ATM/ATR-p53, as well as necroptosis (78, 102).

DNA damage-induced cell death is the key mechanism behind the anticancer action of ANTs and, at the same time, is responsible for their cardiotoxicity. Notably, the effects of ANTs in proliferating (tumor) and nonproliferating (cardiomyocytes) cells are mediated by two distinct isoforms of Top2 (Fig. 3). While the alpha isoform of Top2 (Top2-α) is overexpressed in proliferating cells in S and G2 phase, and it is almost undetectable in quiescent tissues, Top2-β is the unique isoform expressed by adult mammalian cardiomyocytes (12) and mature induced pluripotent stem cell (iPSC)-derived cardiomyocytes (4, 15). The interaction of doxorubicin (Doxo) with Top2-α generates a ternary Top2-Doxo-DNA cleavage complex that triggers tumor cell death. Similarly, the Top2-β-Doxo-DNA complex induces DNA double-strand breaks and activates a DNA damage response that is responsible for cardiomyocyte death. In cardiomyocytes, the activation of p53 stimulates DNA repair processes, but concomitantly represses genes involved in mitochondrial biogenesis (such as peroxisome proliferator-activated receptor-γ [PPARγ] coactivator [PGC]-1) and oxidative phosphorylation, impairs the ratio between beta-cell lymphoma 2 (Bcl2) and Bcl2-associated X (BAX), culminating in cardiac cell death (86, 143). Consistently, p53-knockout mice and adult mouse hearts expressing cardiac myocyte-restricted dominant-interfering p53 are partially protected against Doxo-induced cell death and myocardial dysfunction (147). The observed protection in these models likely stems from both inhibition of apoptosis and preserved mammalian target of rapamycin (mTOR) signaling, which may contribute to prevent cardiac mass reduction (147).

OPEN IN VIEWER

FIG. 3.

Mechanism of ANT-mediated cell injury and death in cancer cells and cardiomyocytes. The anticancer action of ANTs is linked to the poisoning of Top2-α. In cardiomyocytes, the Top2-β isoform is prevalent and represents the major target of ANTs. DXR protects against ANT cardiotoxicity by interfering with the binding of ANTs to Top2-β, thus preventing the poisoning and the loss of function of Top2. DXR, dexrazoxane. Color images are available online.

In further agreement with the role of Top2-β in ANT cardiotoxicity, murine embryonic fibroblasts (72) and cardiomyocytes (12) lacking Top2-β are protected from Doxo-induced damage. Of note, Top2-β appears to be a key player in the cardiotoxicity elicited by ANTs not only when they are used in monotherapy but also in the setting of combinatorial regimens, such as with monoclonal antibodies. For instance, both Doxo and trastuzumab alone induce Top2-β protein downregulation in cardiomyocytes, which is exacerbated in combined treatments. This further reduction in Top2-β activity leads to enhanced ROS production and apoptosis and might be one of the reasons of the aggravated cardiotoxicity of combination therapies compared with single agents (53). Of note, Atwal et al. showed that, in vitro, 10 μM Doxo (or higher) can interfere with the action of other drugs that act as Top2 poisons, such as etoposide, thus hindering the overall anticancer efficacy of combined treatments (4).

Given the selective expression of the beta isoform in cardiomyocytes, the development of Top2-α-specific drugs may provide a valuable means to reduce cardiotoxicity, while maintaining the anticancer efficacy of the drug. This is a key aspect to be considered in the development of cardioprotective agents against ANT toxicity. On these grounds, the increased risk of secondary malignancies in patients treated with DXR has limited its application as cardioprotective agent in the past, although its use in children and adolescents has been recently reconsidered (98). Mechanistically, the cardioprotective action of DXR has been suggested to be independent of its antioxidant properties, rather it has been ascribed to its ability to compete with ATP binding on Top2-β, thereby producing a configuration change that prevents the formation of a complex with ANTs (72). In accord with this view, the DXR analogue JR-311 shows significant cardioprotective effects, which depend on Top2-β modulation, while being independent of its iron-chelating properties (10). Despite its cardioprotective function, caution is still used while prescribing DXR, and this drug is nowadays not recommended for children aged 0–18 years who are expected to receive a cumulative dose of Doxo (or other ANT) lower than 300 mg/m².

Considering the isoform selectivity and opposing effects of Top2 poisoning versus inhibition, it is reasonable to assume that all patients receiving ANTs would benefit from the development of new therapeutic regimens including Top2-α-specific poisons and Top2-β-specific inhibitors.

Rac1 as a master regulator of DNA damage and ROS production

As Top2 emerged as a good pharmacological target to prevent ANT cardiotoxicity, new Top2 regulators have started to be investigated. Among these, Rac1 has been shown to contribute to Doxo-mediated cardiomyocyte death [for a complete review, see Ref. (38)], in both ROS-dependent and ROS-independent manners (73) (Fig. 4). Rac1 is a small guanosine triphosphate-bound protein, belonging to the Rho family, a group of molecular switches and signal transducers that transmit extracellular stimuli from the inner plasma membrane to intracellular signaling pathways. Rac1 is involved in a variety of cellular processes, such as cell–cell and cell–substrate adhesion, polarization, migration, invasion, proliferation, transcription, vesicle formation, and apoptosis, which depend on its spatiotemporal distribution (94). Several studies have identified the presence of a nuclear pool of Rac1 and have unraveled the mechanisms underlying its nuclear localization (94), which include DNA damage itself (41). This is in agreement with the finding that oxidized DNA bases can act as Rac1 guanine nucleotide exchange factors (35). In the nucleus, Rac1 was recently found to be associated with Top2 (108), where it is responsible for the activation of a DNA damage-ATM-p53-apoptosis pathway (139).

OPEN IN VIEWER

FIG. 4.

Geno-protective effect of pharmacological inhibition of Rac1 signaling. Rac1 signaling contributes to ANT-induced cardiotoxicity by regulating Nox2 NADPH oxidase complex and via a ROS-independent nuclear mechanism, leading to the formation of DNA double-stranded breaks and subsequent activation of the ATM/ATR-p53 pathway. In the nucleus, Rac1 binds to and regulates the activity of Top2. Inhibition of the HMG-CoA reductase (mevalonate pathway) by statins causes depletion of the cellular pool of isoprene precursors, which are essential for membrane localization of Rac1. Pharmacological targeting of Rac1 by statins or by the Rac1 inhibitor (NSC23766) attenuates ROS production and DNA damage response. ATM, ataxia-telangiectasia-mutated; ATR, ATM and Rad3-mutated related kinase; NADPH, nicotinamide adenine dinucleotide phosphate. Color images are available online.

Pharmacological inhibition of Rac1 signaling with NSC23766 interferes with Doxo genotoxicity by preventing the formation of the cleavage complex, thereby reducing the generation of double-strand breaks and the activation of the DNA damage response (46). Of note, the effect of NSC23766 can be further improved by the lipid-lowering drug lovastatin (46). Inhibition of the HMG-CoA reductase by statins causes the depletion of the cellular pool of isoprene precursors, which are essential for the C-terminal prenylation, and the subsequent membrane localization, of Rho proteins (26). Thus, prevention of Rac1 prenylation by statins promotes Rac1 translocation to the nucleus, suggesting that the therapeutic effect of statins might be linked to the upregulation of the nuclear pool of Rac1 versus the cytosolic counterpart. Nevertheless, the mechanism by which nuclear Rac1 regulates Top2 activity is still unclear and deserves further experimental efforts. It was shown that in acute (73) and subacute (39) models of ANT cardiotoxicity, cardiac Rac1 knockout confers cardioprotection by means of ROS-independent as well as ROS-dependent mechanisms. In fact, Rac1 contributes to the assembly and activation of the phagocyte nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complex (42), and cardiomyocyte-specific deletion of Rac1 was shown to impair Doxo-induced NADPH oxidase activation, ROS production, DNA fragmentation, cardiomyocyte apoptosis, and finally improve cardiac function. Consistently, mice with impaired Nox2 NADPH oxidase complex are less sensitive to ANTs (145).

Overall, statins interfere with two of the main mechanisms underlying ANTs cardiotoxicity, that is, ROS production and Top2-mediated DNA damage. Their beneficial effect in models of ANT cardiotoxicity and their favorable tolerability make them promising candidates for the prevention and/or treatment of ANT cardiotoxicity, as confirmed in small-scale clinical studies (1). More importantly, statins may sensitize certain tumors to chemotherapy, thereby enhancing the anticancer efficiency of ANT regimens and at the same time protecting the heart (25).

mtDNA and mitochondrial homeostasis

Another interesting aspect related to Top2 poisoning by ANTs is that both Top2-α and Top2-β isoforms were found in mitochondria (70, 141). Top2-β is likely the most relevant isoform in cardiac mitochondria (141). In cardiomyocytes, ANT-mediated Top2-β poisoning might affect not only nuclear DNA but also mtDNA, further contributing to increased ROS production and cytotoxicity. Accordingly, ANT-related chronic heart failure is characterized by imbalanced mitochondrial mass and reduced nuclear expression of genes pivotal to mitochondrial biogenesis and homeostasis (54). These include the PGC-1α and β, nuclear respiratory factor 1 (NRF1), and mitochondrial transcription factor A (TFAM). Accordingly, the protective effect exerted by the histone deacetylase and PGC-1α activator, Sirtuin 1 (SIRT1), against ANT cardiotoxicity (11) is likely due, at least in part, to the regulation of mitochondrial biogenesis.

In addition, ANTs can induce mitochondrial damage through the uncoupling of the electron transport chain, disruption of mitochondrial membrane potential, and generation of ROS, especially in conjunction with mitochondrial iron metabolism (32). However, the relative contribution of direct Top2 poisoning and of secondary oxidation of nucleotides by ROS on mtDNA depletion has yet to be clarified.

Recently, Jean et al. (51) showed that mice treated with a modified form of Doxo, which exclusively targets mitochondria but not nuclei, are resistant to ANT cardiotoxicity, downsizing the contribution of mtDNA damage to the disease. In contrast, Khiati et al. (56) showed that mice lacking mitochondrial Top1 are hypersensitive to ANTs and display severe mitochondrial damage. Doxo-induced mitochondrial defects, including mitochondrial cristae disorganization and impaired production of respiratory chain proteins, ultimately result in decreased O₂ consumption, increased ROS production, and cardiac damage leading to lethality.

Cytosolic Ca²⁺ impairment

Isolated cardiac myocytes perfused with Doxo show increased Ca²⁺ sparks occurrence and decreased Ca²⁺ transients (104). Of note, the Doxo metabolite, doxorubicinol, has been shown to target calsequestrin type 2, thus increasing cytoplasmic Ca²⁺ concentration and favoring arrhythmic conditions (1) (Fig. 5). Doxo also inhibits the transcription of the sarcoplasmic reticulum Ca²⁺-ATPase 2a (SERCA2a) and therefore reduces Ca²⁺ uptake in the sarcoplasmic reticulum (SR) (2). Moreover, Doxo and its metabolite doxorubicinol modify ryanodine receptor (RyR) and SERCA2a activity by binding to the proteins via thiol oxidation (37) and induce calcium/calmodulin-dependent protein kinase-II (CaMKII)-dependent Ca²⁺ leakage from the SR (104, 124). Overall, the effects on RyR, SERCA2a, and phospholamban (PLN), mediated by the activation of protein kinase A (PKA) and CaMKII, result in transient intracellular Ca²⁺ alterations. These, in turn, lead to a temporary depression of Ca²⁺ transients that precede structural defects and overt cardiac dysfunction in models of cardiotoxicity induced by chronic low-dose Doxo (69). Although it is known that ROS activate both PKA and CaMKII (8, 141), the two main regulators of the excitation-contraction coupling, the contribution of β-adrenergic signaling in the setting of ANT cardiotoxicity has been postulated but still requires experimental confirmation (69).

OPEN IN VIEWER

FIG. 5.

Calcium overload as a central event in ANT-induced cardiotoxicity. Doxo and its metabolite, doxorubicinol, affect the activity of different Ca²⁺-handling proteins of cardiomyocytes. Doxorubicinol targets CSQ2 in the SR, thus increasing cytoplasmic Ca²⁺ concentration, whereas Doxo inhibits the transcription of the SERCA2a, thereby reducing Ca²⁺ reuptake in the SR. Doxo and its metabolite doxorubicinol modify the activity of RyR and SERCA2a and induce CaMKII-dependent Ca²⁺ leakage from the SR. ANT-mediated deregulation of CaMKII also impact on mitochondrial Ca²⁺ regulation. CAMKII is responsible for the activation of NF-κB and p53 signaling and leads to increased Ca²⁺ influx in mitochondria through MCU. Mitochondrial Ca²⁺ overload, in turn, triggers MTP opening and release of proapoptotic cytochrome c. Pharmacological targeting of key regulators of Ca²⁺ signaling, such as PDE3 by levosimendan or LTCC by LTCC blockers, as well as chaperone-mediated alleviation of SR stress (GRP78 overexpression) are emerging therapies against ANT cardiotoxicity. CaMKII, calcium/calmodulin-dependent protein kinase-II; CSQ2, calsequestrin type 2; Doxo, doxorubicin; LTCC, L-type Ca²⁺ channels; MCU, mitochondrial calcium uniporter; MTP, mitochondrial permeability transition pore; NF-κB, nuclear factor-kappa B; PDE3, phosphodiesterase 3; RyR, ryanodine receptor; SERCA2a, sarcoplasmic reticulum Ca²⁺-ATPase 2a; SR, sarcoplasmic reticulum. Color images are available online.

The ensuing disruption of sarcoplasmic Ca²⁺ handling machinery leads to intracellular Ca²⁺ overload and, finally, to sarcomeric disarray and myofibril deterioration, which is further aggravated by Doxo-activated Ca²⁺-dependent calpain proteases (68).

Mitochondrial Ca²⁺ impairment

Besides excessive accumulation of Ca²⁺ in the cytoplasm, impaired Ca²⁺ fluxes have been reported in mitochondria in response to Doxo treatment (19). Ca²⁺ enters mitochondria through the mitochondrial calcium uniporter and, above a physiological threshold, mitochondrial Ca²⁺ triggers the opening of the mitochondrial permeability transition pore (MTP), resulting in dissipation of transmembrane potential, mitochondrial swelling, and increased permeability of the outer membrane to apoptotic factors, such as cytochrome c (144). In turn, in the cytosol, cytochrome c forms a complex with the adaptor protein apoptosis protease activator protein-1 (Apaf-1) and caspase-9, the so-called apoptosome, which finally activates the apoptotic cascade (144).

Emerging therapeutic strategies to rescue Ca²⁺ homeostasis

As the relevance of SR Ca²⁺ leakage in ANT cardiotoxicity emerged, new therapeutic approaches have been explored, including cardiac gene therapy with the GRP78 chaperone, which is responsible for amelioration of SR stress (124). Intriguingly, adeno-associated virus-mediated GRP78 overexpression partly protects cardiomyocytes from Doxo-induced cell death by modulating Ca²⁺/CAMKII-dependent pathways and preventing p53 accumulation (124). This study suggests that SR stress is a valuable therapeutic target in the context of Doxo-induced, and possibly other, cardiac disease.

Further highlighting the central role of CaMKII in this process, it has been recently demonstrated that pharmacological or genetic blockade of L-type Ca²⁺ channel (LTCC) attenuates Doxo-induced cardiomyocyte apoptosis by suppressing intracellular Ca²⁺ accumulation and the ensuing activation of CaMKII-nuclear factor-kappa B (NF-κB) pathway (49). Hence, LTCC blockers might be potential therapeutic agents against ANT-induced cardiomyopathy.

Recently, Efentakis et al. (23) showed that targeting both impaired cytosolic Ca²⁺ homeostasis and mitochondrial alterations is a promising strategy. A single-dose of levosimendan, an inotropic vasodilator (92) that is clinically used for the treatment of decompensated heart failure (24), prevented subchronic Doxo-induced cardiotoxicity, as indicated by preserved left ventricle systolic function and structure, by acting on cAMP/PKA and PLN pathway (23). Intriguingly, levosimendan and its active metabolite OR-1896 act as PDE3 inhibitors, with whom they share structural similarity (90). Of note, besides counteracting cytoplasmic Ca²⁺ accumulation, levosimendan displays anti-inflammatory and antioxidant activities, sensitizes Troponin C to Ca²⁺ (inotropic effect), and activates the ATP-sensitive K⁺ channel (24), events that all converge into a cardioprotective effect.

As increased intracellular Ca²⁺ levels emerged as key triggers of ROS production, which, in turn, further disrupts intracellular and mitochondrial Ca²⁺ homeostasis, new therapeutic tools aimed at preventing or restoring Ca²⁺ balance may prove effective against ANT-induce toxicity.

Adenosine monophosphate-activated protein kinase/mTOR signaling pathway

Autophagy is the major cellular recycling process, essential for maintaining cellular homeostasis (113). The initiation of autophagy is under the control of autophagy-related genes (ATG)1, also known as activated unc-51-like autophagy activating kinase 1 (Ulk-1), which phosphorylates Beclin-1 (5, 85), thus triggering the formation of autophagosomes. The activity of Ulk-1 is under the control of adenosine monophosphate-activated protein kinase (AMPK) and mTOR signaling pathways. AMPK and mTOR promote and inhibit autophagy, respectively, by finely regulating the kinase activity of Ulk-1. Consistently, ANTs have been found to interfere with mTOR and AMPK regulatory pathways leading to autophagy dysregulation.

ANTs inhibit autophagy

ANTs activate mTOR that, in turn, induces the inhibitory phosphorylation of Ulk-1 at Ser757, thereby preventing Ulk-1-mediated autophagosome initiation (132). However, ANTs inhibit the phosphorylation of AMPK, an upstream positive regulator of autophagy initiation, further contributing to the block of the autophagic process. Interestingly, Li et al. (65) recently identified another kinase, which is involved in autophagy regulation, and whose activity is critically affected by ANTs. The authors show that, in cardiomyocytes, Doxo engages phosphatidylinositol 3-kinase (PI3K)γ signaling, converging on autophagy inhibition and maladaptive metabolic reprogramming. In particular, Doxo activates PI3Kγ/Akt signaling downstream of Toll-like receptor (TLR) 9, leading to increased phosphorylation of mTOR and of its downstream target Ulk-1 and, in turn, to autophagy inhibition. Accordingly, genetic or pharmacological blockade of PI3Kγ reactivates autophagy by releasing the inhibitory brake of mTOR and protects the heart against ANT-induced cardiotoxicity (65). Notably, the enhanced autophagy triggered by PI3Kγ inhibition protects cardiomyocytes against multiple Doxo-induced toxic effects, such as DNA damage, Ca²⁺ dysregulation, and ROS production. Indeed, in cardiomyocytes, Doxo treatment results in ROS accumulation, which is further enhanced by the addition of the autophagy blocker BafA1, while being reduced by PI3Kγ inhibition. Of note, BafA1 completely abrogated the protective effect of PI3Kγ blockade, thus corroborating the view that PI3Kγ inhibition protects against Doxo-induced oxidative stress and cardiotoxicity via activation of autophagy (65).

Another recent study (131) confirms that Doxo impairs the autophagic flux and the transcription of ATG by increasing the expression and phosphorylation of mTOR. Activated mTOR, indeed, retains in the cytoplasm the transcriptional regulator transcription factor EB (TFEB), which normally drives the expression of autophagy and lysosomal genes, such as LAMP1 (57, 109, 110, 132, 148). The transcriptional activity of TFEB is inhibited by Doxo-induced phosphorylation of mTOR, whereas the impaired nuclear localization of TFEB causes the decline of lysosomal biogenesis and function (77, 82), resulting in increased levels of Beclin-1, LC3-II, and P62. In line with these findings, inhibition of mTOR activity with Torin-1 in H9c2 cardiomyoblasts restores the nuclear localization of TFEB and autophagy, preventing Doxo-induced apoptosis and improving cell viability (6).

In contrast, Zhu et al. report that mice with cardiomyocyte-specific and constitutively active mTOR are protected against ANT-induced acute cardiac dysfunction (147). In this work, the authors observed that Doxo treatment in vivo decreases mTOR activity, as evidenced by decreased mTOR phosphorylation, inducing cardiomyocyte death and mass reduction, whereas the activation of the mTOR pathway effectively retains normal cardiac mass and function.

ANTs induce autophagy

In line with the study of Zhu et al., other studies have reported that ANTs increase autophagy, leading to maladaptive cardiac remodeling. A first study of 2009 by Lu et al. (71) demonstrated that ANTs induce an upregulation of the autophagy-related marker Beclin-1 and autophagic vacuole formation. In agreement with this study, Kobayashi et al. (58) showed that Doxo markedly increases LC3-II protein expression, leading to cardiomyocyte death via inhibition of antiapoptotic protein B cell lymphoma 2 (Bcl2), a negative regulator of Beclin-1. Conversely, Bcl2 is directly upregulated by the transcription factor GATA4 (61), which appears to participate to autophagy regulation (58) also by binding to and blocking Beclin-1 (93). Of note, the preservation of GATA4 activity attenuates Doxo cardiotoxicity by inhibiting autophagy through modulation of the expression of Bcl2 and ATG.

Li et al. (64) further investigated the role of Beclin-1 in ANT cardiotoxicity in vivo using Beclin-1 heterozygous-deficient mice (Beclin-1^+/−) since Beclin-1 knockout is embryonic lethal. Beclin-1^+/− hearts emerged to be protected against ANT-induced damage. Moreover, Beclin-1 inhibition limits the initiation of autophagy, thus preventing the accumulation of damaged cellular components within autophagosomes and reducing cardiomyocyte apoptosis (58, 64). Consistently, later studies confirmed the activation of autophagy as a primary cause of cardiomyocyte programmed cell death in response to ANTs and suggested inhibition of autophagy as a valuable tool to significantly improve cardiac function and prevent ANT-induced cardiotoxicity (71, 135, 142).

Although autophagy has been revealed as an important mechanism underlying ANT-induced cardiotoxicity, whether ANTs induce or inhibit autophagy (Fig. 6), and if autophagy is beneficial or harmful in this context is still a matter of debate. Future studies are awaited to clarify the scenario.

OPEN IN VIEWER

FIG. 6.

Overview of ANT-induced dysregulation of autophagy. Autophagy deregulation is a key mechanism underlying ANT-induced cardiotoxicity, although its beneficial versus maladaptive role is still controversial. Cardiomyocyte exposure to ANTs perturbs AMPK, Akt, and mTOR signaling pathways, which are all upstream regulators of autophagy. ANTs inhibit the autophagic process by activating mTOR or by blocking AMPK. Furthermore, Doxo engages PI3Kγ signaling downstream of TLR9, converging on inhibitory phosphorylation of Ulk-1 at Ser757 and preventing autophagosome initiation. Additionally, activated mTOR restricts TFEB in the cytoplasm and inhibits the transcription of autophagy-related genes, such as LAMP1. Conversely, Doxo blocks the activity of GATA4, thereby inhibiting the transcription of anti-apoptotic protein Bcl2, a negative regulator of Beclin-1. As a result, the autophagy-related marker Beclin-1 and LC3-II are upregulated and lead to cardiomyocyte death. Restoration of GATA4 attenuates Doxo cardiotoxicity by upregulating Bcl2 and in turn inhibiting autophagy. AMPK, adenosine monophosphate-activated protein kinase; LC3, microtubule-associated protein 1A/1B-light chain 3; mTOR, mammalian target of rapamycin; PI3K, phosphatidylinositol 3-kinase; TFEB, transcription factor EB; TLR, Toll-like receptor; Ulk-1, Unc-51-like autophagy activating kinase 1. Color images are available online.

BNIP3 and BNIP3L/Nix

Both BNIP3 and BNIP3L/Nix act on the outer mitochondrial membrane as direct mitophagy receptors, as they have an LC3-II recognition motif, thereby allowing autophagosomal engulfment of mitochondria. ANTs lead to severe necrosis in cardiomyocytes, promoting upregulation and translocation of BNIP3 to the mitochondrial membrane, which, in turn, triggers the opening of the permeability transition pore and the increase in ROS production (Fig. 7) (20). Conversely, inhibition of BNIP3, with either shRNA against BNIP3 or through the expression of a mutant BNIP3 that fails to target mitochondria, rescues ANT-induced defects of the mitochondrial respiratory chain in vitro in cultured cardiomyocytes. Furthermore, mitochondria morphology and cardiac function are preserved in BNIP3-knockout mice (BNIP3^−/−) treated with ANTs, confirming a maladaptive role of BNIP3 in ANT cardiotoxicity (20). Additionally, BNIP3 activation in cardiomyocytes treated with ANTs is associated to loss of nuclear Hmgb1 and release of classical markers of necrosis (i.e., lactate dehydrogenase and cardiac troponin T), supporting a primary role of BNIP3 in necrosis (20).

OPEN IN VIEWER

FIG. 7.

Mechanism of ANT-mediated mitochondrial damage and cell death. Normally, mitochondrial homeostasis is regulated through a selective form of autophagy, namely mitophagy. Under stress conditions, PINK1 aggregates and recruits Parkin to the mitochondrial outer membrane mediating the ubiquitination of mitochondrial proteins and their recognition by p62, consequently leading to the initiation of mitophagy. However, Doxo-induced p53 accumulation inhibits mitophagy by suppressing Parkin translocation and interaction with SESN2 and p62, therefore inducing cardiomyocytes apoptosis. Conversely, in models of acute cardiotoxicity induced by high-dose ANTs, BNIP3-induced mitophagy has a maladaptive role. Upregulation and translocation of BNIP3 lead to the permeability transition pore opening and increased ROS production. BNIP3, BH3-only protein Bcl-2-like 19 kDa-interacting protein 3; PINK1, PTEN-induced kinase 1; SESN2, sestrin-2. Color images are available online.

New compelling evidence establishes that NF-κB signaling, which transcriptionally silences BNIP3 activation under basal conditions, is dramatically reduced in cardiac myocytes treated with Doxo (19). This, in turn, promotes BNIP3 gene activation, mitochondrial injury including calcium influx, permeability transition pore opening, and necrotic cell death. Interestingly, restoring NF-κB signaling suppresses BNIP3 expression, mitochondrial perturbations, and necrotic cell death, thus suggesting a new potential therapeutic approach for cancer patients treated with Doxo.

PINK1/Parkin pathway

However, PINK1 accumulates and recruits Parkin to the outer membrane of depolarized mitochondria, mediating the ubiquitination of mitochondrial proteins and their recognition by p62 and LC3, consequently leading to the initiation of mitochondrial autophagy (105).

In Doxo-induced cardiomyopathy, mitophagy is critically involved in the elimination of vacuolated and dysfunctional mitochondria (47). Hull et al. (47) demonstrated that PINK1/Parkin-dependent mitophagy is suppressed early after ANT treatment (2–8 days post-treatment), but is restored at a later stage, when PINK1 and Parkin are significantly upregulated. Interestingly, the mitophagy inhibitor peptide, Mdivi-1, prevents Doxo-induced cardiotoxicity, indicating that excessive mitophagy contributes to Doxo cardiotoxicity (29).

Nevertheless, the molecular mechanisms by which ANTs inhibit mitophagy have still to be uncovered. A recent study provides novel and compelling evidence that overexpression of SESN2, a member of the sestrins family, protects cardiomyocytes against Doxo-induced cardiomyopathy via the improvement of mitochondrial function and rescue of mitophagy (129). Of note, the expression of SESN2 is significantly reduced following Doxo stimulation, in vitro and in vivo. Mechanistically, SESN2 interacts with Parkin and p62 and promotes Parkin recruitment to mitochondria, subsequently activating mitophagy. These results establish SESN2 as a new player in the regulation of mitochondrial function and provide a potential therapeutic approach to prevent Doxo-induced cardiomyopathy.

Interestingly, mitophagy appears to be coregulated with mitochondrial biogenesis, through a shared signaling pathway involving nuclear translocation of NRFs and a functional interaction between PGC-1α and Parkin (137). Doxo inhibits the expression of PGC-1α and its downstream targets, NRF1 and TFAM, thus reducing the expression of mitochondrial proteins. Inhibition of mitophagy by mdivi-1 attenuates Doxo-mediated activation of the PINK1/Parkin pathway and rescues mitochondrial biogenesis by preserving the expression of PGC-1α (137).

Autophagy promotes apoptosis

Apoptosis is regulated by Bcl2 and p53 genes, which are closely associated with autophagy. In general, autophagy restrains apoptosis, while apoptosis-associated caspase activation shuts off the autophagic process (76). However, in particular cases, such as following Doxo stimulation, autophagy is associated with high levels of apoptotic cell death. Autophagy is dramatically increased in Doxo-treated cardiomyocytes and associates with increased cleavage of caspase 3 and PARP, which indicate apoptosis initiation (58). Similarly, in human umbilical vein endothelial cells. Doxo induces autophagy early after treatment (since 3 h), while the level of apoptosis as well as lysosomal membrane permeabilization (LMP) and mitochondrial outer membrane permeabilization (MOMP) increase at later stages (12 h) (52). This finding implies that Doxo-induced autophagy and the ensuing autophagy-dependent LMP are upstream events, followed by distal events including MOMP and cytochrome c release, which finally triggers caspase-dependent apoptosis. Consistently, inhibition of Doxo-induced autophagy, through the autophagy inhibitor 3-MA or Beclin-1 knockdown, results in significant attenuation of cell death (52, 58). Conversely, activation of autophagy by rapamycin or Beclin-1 expression exacerbate apoptosis in H9C2 cells and in mice hearts treated with Doxo (117, 134).

In addition, Doxo-induced cell death correlates with depletion of GATA4, a transcription factor essential for cardiomyocyte growth and survival (89). GATA4 is both sufficient and necessary to inhibit autophagy by controlling the gene expression of Bcl2, a survival factor with anti-autophagy activity (93). Concurrently, GATA4 reduction induces an upregulation of ATG genes through p53, leading to the excessive activation of autophagic flux, which ultimately contributes to Doxo-induced cardiomyocyte death.

Autophagy inhibits apoptosis

The above-presented findings are challenged by the observation that activation of autophagy, by stimuli such as AMPK activation by glucose-depletion or mTOR inhibition by rapamycin, results in enhanced cell viability and reduced apoptosis following Doxo treatment in H9C2 cardiomyoblasts and mice overexpressing GFP-LC3 (117). In this study, the induction of autophagy by rapamycin reduces the activity of cleaved caspase-3, thus supporting an inhibitory role of autophagy on apoptosis, and indicating a therapeutic potential of rapamycin against Doxo-induced myocardial damage (117).

Further confirming the beneficial role of autophagy against apoptosis, it has been shown that autophagy inhibition, by decreasing the expression of AMPK or knocking out AMPKα1 in mice (AMPKα1^−/−), increases the sensitivity to Doxo-induced apoptosis, while restoration of autophagy by adenovirus-mediated AMPK constitutive activation or 5-aminoimidazole-4-carboxamide ribonucleoside significantly limits the iatrogenic effects of Doxo (130). Consistently, genetic deletion of AMPKα1 (AMPKα1^−/−) enhances Doxo-related reduction of p53 phosphorylation, establishing a critical role of AMPK in Doxo-induced p53-dependent DNA damage and apoptosis (130).

Similarly, Doxo inhibits autophagy by activating E2F1/mTORC1 pathway and further induces apoptosis by activating E2F1/AMPKα2 pathway in starved H9C2 cardiomyoblasts (33). The same result was observed in a mouse model of Doxo-induced cardiotoxicity, in both myocardial ischemic and nonischemic conditions, in which resveratrol was able to repress E2F1, thereby activating autophagy and reducing apoptosis (33).

Finally, p53, that is mostly studied for its role as a transcription factor, displays a key role in cell death and autophagy regulation (13, 119). Accumulation of cytosolic p53 can directly activate BAX to permeabilize mitochondria, thereby inducing caspase activation and apoptosis (13). p53 also functions as an endogenous repressor of autophagy (80). An additional function of cytosolic p53 has been discovered in Doxo-treated cardiomyocytes (44), in that it inhibits mitophagy by blocking Parkin translocation to damaged mitochondria (Fig. 7). The inhibition of Parkin-mediated mitophagy compromises mitochondrial function, resulting in aging-related and Doxo-mediated decrease in cardiac contractility. In keeping with this mechanism, p53-deficient mice and Parkin-overexpressing cardiomyocytes show a less severe decline of mitochondrial integrity and cell death after treatment with Doxo (44).