From Molecular Mechanisms to Clinical Management of Antineoplastic Drug-Induced Cardiovascular Toxicity: A Translational Overview

A. Adverse effects

ANT-induced CTX can manifest as a sort of cardiomyopathy, referred to as ANT-induced cardiomyopathy, leading to HF, which limits the usability of the drugs, with important consequences for managing malignancies. ANT-related CTX represents a significant clinical burden, producing LV dysfunction in a maximum of 9% of cases in a recent large prospective study (35).

B. Mechanisms of adverse effects

ANT-related CTX mechanisms are derived from a combination of several cellular and molecular alterations, leading to myocardial damage and dysfunction (Fig. 1). ANT-induced CTX is still not completely understood, though new insights on molecular mechanisms have been elucidated in recent years (373, 432, 435).

OPEN IN VIEWER

FIG. 1.

Concentric representation of cardiotoxic effects of anthracyclines. ROS, reactive oxygen species.

The role of redox stress, resulting from an overproduction of ROS/RNS, in ANT-related CTX is undisputed (246, 247, 353, 434), although mechanisms are much less clear (see section X for further details). ROS and RNS may be not only either direct or indirect inducers of the cardiac injury but also triggers, byproducts of the injury, or end-products arising from the injury itself. In addition, severity of oxidative/nitrosative stress often does not correspond well with the severity of the injury. This is connected to acceptance and explanation of poor efficacy of different antioxidants under relevant conditions. Recent data suggest that ANT-induced CTX is also strictly related to ANT interaction with topoisomerase 2 (Top2)β in cardiomyocytes (229). Two types of Top2 enzymes are targeted by ANTs: Top2-α and Top2β. The former is present in rapidly dividing cells, such as cancer cells, and forms the ternary Top2-doxorubicin-DNA cleavage complex, inducing cell apoptosis, whereas Top2β is present in human cardiomyocytes, forming the Top2β-doxorubicin-DNA complex, which causes DNA double-strand breaks and transcriptome modifications, also leading to cell apoptosis (434). A seminal paper confirmed the crucial role of the Top2β gene in an animal model, demonstrating that cardiomyocyte-specific deletion of the Top2β gene is protective against ANT-induced damage (434). The DNA break caused by doxorubicin binding to Top2β can, therefore, stimulate the DNA injury response. Consequently, the tumor suppressor protein p53, a fundamental enzyme for activating DNA repair proteins, can be induced. Unfortunately, p53 can also cause defective mitochondria biogenesis and metabolic failure by suppressing genes involved in organelle biogenesis, such as peroxisome proliferator-activated receptor gamma coactivator 1-alpha, and alteration of oxidative phosphorylation (434). Intriguingly, the metabolic perturbations induced by doxorubicin-activated p53 are responsible for altered autophagy, a process that is necessary for the normal recycling of dysfunctional mitochondria. Consequently, doxorubicin-damaged mitochondria accumulate in the cardiomyocytes, resulting in enhanced ROS/RNS generation and, ultimately, cell death. Recent observations in p53-null mice found a smaller impairment in cardiac functional reserve after ANT treatment, supporting this hypothesis (157). Interestingly, in these mice, mitochondrial and LV function were maintained with increasing age, suggesting that p53-mediated inhibition of autophagy may play a role in all forms of cardiac dysfunction, not just doxorubicin-induced cardiomyopathy (157).

Apart from p53, doxorubicin may also induce the mitogen-activated protein kinase (MAPK) pathway via ROS- and Ca²⁺-dependent mechanisms (437). Importantly, extracellular signal-regulated kinases (ERKs), members of the MAPK family, may protect myocytes from apoptosis, whereas p38 MAPK induces death of cardiomyocytes (437). More studies are needed to elucidate the role of such kinases and of other less-characterized signaling pathways in ANT-induced cardiotoxicity. However, these data confirm that oxidative reactions, at the basis of ANT-induced LV dysfunction, are involved in most types of HF. Therefore, timely innovative pharmacological strategies that interfere with specific molecules involved in heart dysfunction (e.g., p53) may represent a potential common approach in limiting HF occurrence (250, 341).

ANT-alcohol metabolites also play a pivotal role in inducing cellular injury and CTX via iron-dependent and -independent mechanisms. In fact, these metabolites disrupt iron and calcium homeostasis and, ultimately, lead to intracellular Ca²⁺ overload. Calcium overload has also been related to increased calpain proteolytic activity, which leads to cellular disarray and sarcomere disruption, resulting in sarcopenia (220). In addition, the interaction of ANTs with critical signaling pathways and with the activity of transcription factors may also explain sarcopenia, which derives from the limitation of sarcomere protein synthesis (165).

Mitochondrial activity has a central role in ANT-induced CTX (257, 258). The presence of doxorubicin in the mitochondrion, due to a high affinity for the mitochondrial phospholipid cardiolipin, negatively affects its function, stimulating ROS/RNS production, inhibiting oxidative phosphorylation, and causing mitochondrial DNA damage (300), with a consequent progressive reduction of energy production, leading to cell dysfunction (210). ANTs also appear to be responsible for mitochondrial calcium accumulation (300), leading to mitochondrial membrane injury. The mitochondrial pathway is an additional mechanism that is responsible for cellular intrinsic apoptosis. This involves elements of the outer mitochondrial membrane, including Bax and Bak, and activates cytochrome C and caspase, among others (257).

Some researchers have hypothesized that the loss of iron homeostasis and the Ca²⁺ overload caused by ANT alcohol metabolites, impairing cardiomyocyte energy and redox balance, could be sufficient to induce significant myocardial dysfunction, and that, in addition, cardiac injury could be increased by an apoptotic loss of cardiomyocytes triggered by ROS and Fe²⁺ (257, 258). ANTs can also affect cardiac progenitor cells, hampering the regeneration capabilities of cardiac tissues after myocardial damage (158, 289). The controversy on the role of cardiac stem cells is discussed in the section XII.

Finally, it should be noted that several factors can favor ANT-induced CTX, and this may explain the individual variability in CTX occurrence (246, 247) (Fig. 2). The so-called multiple-hit hypothesis considers a late onset of CTX due to pharmacological and nonpharmacological subsequent injury. Therefore, strategies favoring cardiac adaptation to various stressors are crucial after ANT therapy (244). Of course, a better understanding of the molecular mechanisms of ANT-related CTX is essential to choose the best strategies to prevent and treat CTX (33, 231, 232, 345, 408).

OPEN IN VIEWER

FIG. 2.

Simplified algorithm showing the factors related to anthracycline therapy or patient characteristics that may determine cardiac damage.

C. Ways to reduce CV toxicity

Several approaches have been proposed to reduce ANT cardiotoxicity, including angiotensin-converting-enzyme inhibitors (ACE-Is), β blockers, Doxrazoxane, phosphodiesterase-5 (PDE-5) inhibitors, Ranolazine and Statins, as well as nutritional supplementation and exercise training. The way the various approaches may reduce CV toxicity by ANT is treated in sections X and XI.

A. Adverse effects

Trastuzumab is a humanized monoclonal antibody that binds the extracellular domain IV of HER/ErbB2 (107, 377). It is the prototypical anti-ErbB2 agent, and the first developed and most widely used type II cardiotoxic drug. Trastuzumab is particularly useful in treating ErbB2⁺ breast and gastric cancers. Unfortunately, it can also cause CTX in a substantial number of patients, peaking at 28% with concomitant administration of trastuzumab and ANTs (262, 362, 377). Indeed, as said earlier, ANTs are responsible for type I CTX, with permanent cardiac damage. Therefore, reduced left ventricle ejection fraction (LVEF) results from the association of trastuzumab and doxorubicin: Trastuzumab enhances or even induces doxorubicin toxicity. Once anti-ErbB2 agents block the protective mechanisms of ErbB2, the oxidative damage induced by doxorubicin increases (91). This co-administration is now avoided. As a class II cardiac dysfunction (93, 94), trastuzumab CTX appears to be elicited by the impairment of contractility rather than the loss of cardiomyocytes, and previous chemotherapy seems to be responsible for the troponin release observed in sequential treatment (91).

Pertuzumab is a more recent anti-HER2 antibody that binds the receptor's domain II. Lapatinib is a different anti-ErbB2 agent, a small-molecule inhibitor of the intracellular tyrosine kinase domain of ErbB2. Of note, trastuzumab only disrupts ligand-independent ErbB2 signaling, whereas pertuzumab interferes with the formation of ligand-induced ErbB2 heterodimers. In contrast, lapatinib inhibits both ligand-induced and ligand-independent ErbB2 signaling (69). Interestingly, lapatinib seems to be less toxic than trastuzumab. Data regarding the toxicity of pertuzumab are more limited (262).

B. Mechanisms of adverse effects

Cardiotoxicity of anti-ErbB2 drugs has been attributed to the inhibition of fundamental actions of neuregulin 1 in the heart (262, 286). In brief, in response to various stimuli, including mechanical strain, adult cardiac microvascular endothelial cells may release neuregulin 1 (NRG1, especially the NRG1β isoform) (221). Thus, NRG1 acts on cardiomyocytes in a paracrine manner, triggering ErbB4/ErbB4 homodimerization and ErbB4/ErbB2 heterodimerization to stimulate protective pathways in response to stress (221, 286). The ErbB2 pathway mediates cell survival and functionality, with recent data focusing on its role in mammalian heart regeneration (63), and appears to be stimulated when the heart experiences adverse hemodynamics or other stress, such as ANT therapies (Fig. 3) (110). It has been postulated that by interfering with the NRG1/ErbB4/ErbB2 axis in the myocardium, anti-ErbB2 agents can cause cardiomyocyte damage and, eventually, HF, and that this is more likely to occur if myocytes are concomitantly exposed to other stressors, such as hypertension or doxorubicin (70, 91). Supporting such a hypothesis, ErbB2 cardiac KO mice exhibited dilated cardiomyopathy, with increased susceptibility to ANT-induced damage to cardiac myocytes (61, 291). Conversely, overexpression of ErbB2 in the heart resulted in lower levels of ROS in mitochondria, with reduced ROS levels and less cell death in neonatal myocytes isolated from ErbB2(tg) hearts after doxorubicin treatment, due to enhanced levels of glutathione peroxidase 1 (GPx1) protein and GPx activity, with higher levels of two known GPx activators, c-Abl and Arg. These data suggest new mechanisms by which ErbB2 blockers can damage heart structure and function (19).

OPEN IN VIEWER

FIG. 3.

ErbB receptor homodimerization or heterodimerization is induced by stressors, including anthracycline therapy, with consequent complex intracellular pathway activation. Cascade effects can be avoided by treatment with monoclonal antibodies and TKIs. See the text for further explanation. EGFR, epidermal growth factor receptor; ErbB2 (or HER2), human epidermal growth factor receptor 2; ERK, extracellular signal-regulated kinase; TKIs, tyrosine kinase inhibitors.

These data have led to further studies on NRG1/ErbB4/ErbB2 that have focused away from chemotherapy-induced HF to heart disease from any cause, with implications for new therapeutic perspectives. For example, in mice subjected to pressure overload, ErbB4 and ErbB2 (both mRNA and protein) decreased significantly with the progression of the disease from compensated cardiac hypertrophy to overt HF (250, 325). Consistently, ErbB2 and ErbB4 receptor expression and activation/phosphorylation were observed to be lower in human-failing myocardia, compared with organ donors (326). Interestingly, LV unloading by implantation of an LV assist device restored the levels of ErbB4 and ErbB2 (326, 400). In an apparent contrast with these results, dogs with pacing-induced HF showed increased phosphorylation of ErbB4 and ErbB2 (78). Inactivation of the intracellular downstream effectors of ErbB4 and ErbB2, ERK1/2 and Akt, was observed, suggesting a disabled NRG1/ErbB4/ErbB2 signaling. Actually, NRG1 expression is increased in HF compared with control conditions in most studies (78, 250, 326). All in all, these data hint that deranged NRG1/ErbB4/ErbB2 activity is involved in the pathophysiology of HF in at least two manners: (i) HF may derive from the use of anti-ErbB2 drugs, such as trastuzumab; (ii) ErbB4/ErbB2 is downregulated and/or uncoupled from intracellular signaling despite normal or increased NRG1, possibly leading to cardiac decompensation (250). Furthermore, novel observations suggest that levels of catecholamines, which usually increase with the occurrence of LV dysfunction and with ANT administration (176, 250, 282), can stimulate ErbB2 expression in myocytes, thus making these cells particularly susceptible to the effects of trastuzumab, resulting in myocardial toxicity (382).

C. Ways to reduce CV toxicity

The aforementioned experimental results may support the use of β blockers in the prevention of trastuzumab CTX (see section X.A.1) (281), in line with a retrospective study that found that continuous use of β blockers was associated with a lower risk of new HF events in subjects on trastuzumab, ANTs, or both (347). Prevention with β blockers is currently being assessed in clinical trials (180, 281, 347, 349) with bisoprolol (MANTICORE 101-Breast) (305), NCT01009918 (carvedilol), and NCT01434134/NCT00806390 (metoprolol) to cure or prevent trastuzumab-induced LV dysfunction (250, 281). Interestingly, from the recent PRADA (prevention of cardiac dysfunction during adjuvant breast cancer therapy) trial, we can infer that blocking only β1 with metoprolol may not produce an adequate and sufficient cardioprotection (128), thus supporting the use of nonselective β1 and β2 blockers (382). In the clinical setting of trastuzumab-induced cardiac dysfunction, when trastuzumab is discontinued, normal ErbB2 signaling is restored, and the reduced LVEF can increase back to normal levels. Indeed, trastuzumab re-administration after discontinuation is considered relatively safe on LVEF recovery (93, 94).

Animal studies have demonstrated that NRG1 regulates doxorubicin injury in rat myocytes (390). Owing to the aforementioned cardioprotective properties of NRG1 via ErbB4/ErbB2, and as the activity of these receptors is altered in HF, the axis neuregulin-ERB is now being intensively investigated in clinical trials for HF treatment (111, 112, 221). It has been hypothesized that NRG1 and NRG1 analogs can be used as therapeutic agents in HF. Intravenous administration of recombinant human NRG1 and of the glial growth factor 2 isoform of NRG1β enhanced heart function and reduced LV dimensions in experimental failing hearts (111, 112, 218, 227, 250). Since NRG1 exerted positive effects in animal models of ischemia-induced HF even when administered after acute myocardial infarction, it can be speculated that it is able to produce beneficial reverse remodeling of the damaged heart, and it does not simply limit cardiac dilation (111, 112, 227). Also, it appears that NRG1 exerts an antifibrotic effect, directly inhibiting cardiac fibroblasts and, thus, preventing fibrosis (111, 112).

Clinical studies have demonstrated that recombinant human NRG1 is well tolerated by patients, and it ameliorates cardiac dimensions and LVEF until a maximum of 3 months after treatment (116, 169). Nevertheless, NRG1 may be considered a growth factor for tumor cells, particularly when administered systemically. Hopefully, additional experimental and clinical studies can assess this fundamental safety concern, producing novel data regarding the effects of NRG1 in HF (221, 250).

A. Adverse events

Anti-angiogenic drugs disrupt the VEGF signaling cascade (Fig. 4) and may induce type II CTX (Fig 5). VEGF is a modulator of myocardial function and growth, while, at the same time, modulating the integrity and expansion of the coronary and systemic blood vessels (43, 44, 62, 88, 103, 104, 107, 240, 392). VEGF antagonists may, therefore, produce different forms of CTX, mainly hypertension, thromboembolism (TE), LV dysfunction, and HF (126, 344, 421).

OPEN IN VIEWER

FIG. 4.

VEGFR activation triggers a complex intracellular pathway. Cascade effects can be avoided by treatment with anti-angiogenic drugs acting at various levels of the cascade. See the text for further explanation. MAPK, mitogen-activated protein kinase; NO, nitric oxide; NOS, nitric oxide synthase; VEGFR, vascular endothelial growth factor receptor.

OPEN IN VIEWER

FIG. 5.

Schematic illustration of how TKI antagonists induce cardiac damage; these drugs can increase arterial pressure and ventricular-vascular coupling, leading to heart failure. PDGFR, platelet-derived growth factor receptor; VEGF, vascular endothelial growth factor.

In particular, bevacizumab, sorafenib, and sunitinib are now widely used for the treatment of different cancers. More recently introduced tyrosine kinase inhibitors (TKIs) can also induce CTX. Regorafenib is a multi-target TKI whose targets include VEGF receptor (VEGFR) 1–3, endothelial-specific receptor tyrosine kinase, platelet-derived growth factor receptor (PDGFR), fibroblast growth factor receptor, cKIT, RE arranged during transfection (RET), and rapidly accelerated fibrosarcoma kinase. Regorafenib is used in the treatment of colorectal tumors and gastrointestinal tumors (28). It may be responsible for arterial hypertension; less frequently, it can produce cardiac ischemia and myocardial infarction (30).

Pazopanib and axitinib, used in the treatment of metastatic renal tumors, are also associated with a high rate of arterial hypertension. Pazopanib is an orally administered multi-targeted TKI, targeting VEGFR 1–3, PDGFA and PDGFB receptors, and c-KIT. In a recent study, the frequency of pazopanib-associated hypertension varied between 36% and 46% (268). Axitinib is a third-generation VEGFR inhibitor used in metastatic renal cancer after failure of previous treatments, and it is very selective. In a study comparing axitinib and sorafenib, the frequency of hypertension was 29% for sorafenib and 40% for axitinib (162). New anti-angiogenic drugs not yet approved for clinical use are vatalanib and nintedanib. Preliminary evidence indicates a potential risk of arterial hypertension and congestive HF and, for vatalanib, of pulmonary embolism, although more rarely (318, 407).

B. Mechanisms of adverse effects

The myocardium requires appropriate perfusion to function properly (43, 44, 62, 88, 103, 104, 107, 240, 392), and it depends on HIF-1 and VEGF pathways, similar to tumors. Of note, inhibition of HIF-1 by p53 may produce HF during chronic pressure overload (336). Moreover, conditional expression of a VEGF scavenger may cause myocardial hibernation and microvessel rarefaction, which can be reversed by suppressing the expression of the scavenger, even months after its induction (244, 417). Such findings show that the myocardium is very sensitive to anti-angiogenic treatments, especially with hypertension-related pressure overload.

The antibody bevacizumab binds circulating VEGF-A, which triggers signaling in endothelial cells, and is used as a therapy for advanced lung, breast, and colon/rectum cancers (160, 334). It has been observed that bevacizumab can induce cardiac dysfunction in 1% of chemotherapy-naïve patients and in 3% of patients with previous chemotherapy (256). Sunitinib and sorafenib are small-molecule TKIs, and they are approved for treating metastatic renal cancer and imatinib-resistant gastrointestinal stromal tumors (45, 126). Importantly, they are not highly selective, and they can inhibit kinases other than VEGF (43). In particular, sunitinib interferes with >30 other tyrosine kinases, including PDGFR alpha and beta, the RET proto-oncogene, FMS-related tyrosine kinase 3 (FLT3), c-Kit, and colony-stimulating factor 1 receptor (CSF1R) (43, 107, 135, 230). This explains why sunitinib is considered more cardiotoxic than other anti-angiogenic drugs, with an incidence of LV dysfunction in a maximum of 28% of patients (51, 187, 269, 387) (Fig. 5). Indeed, all these kinases play a role in the maintenance of CV function (9, 168, 217, 228). The high CTX of sunitinib is also due to interference with off-target kinases, such as ribosomal S6 kinase, with consequent activation of the intrinsic apoptotic pathway, and 5′ AMP-activated protein kinase (AMPK, involved in the response to energy stress), with exacerbation of ATP depletion (107, 186). Interestingly, Hasinoff and Patel (135) not only confirmed the inhibition of AMPK in isolated cardiomyocytes but also suggested that this is likely not essential for the development of sunitinib cardiotoxicity. Creatine kinase may be also involved in the modulation of sunitinib contractile actions (393). A study (51) in mice treated with sunitinib also showed a prolonged opening of the nonselective mitochondrial permeability transition pore (mPTP) and a significant mitochondrial swelling with deformation of the normal mitochondrial architecture in myocytes from hearts subjected to pressure overload. Other studies (422) showed that sunitinib likely does not induce significant impairment of oxidative phosphorylation; thus, the impact of sunitinib on energy metabolism is still controversial, with the incidence of sunitinib-induced myocardial dysfunction being perhaps lower than initially suggested.

In addition, a 2013 manuscript (46) pointed to attention from cardiomyocytes to microvascular dysfunction in the heart and pericyte damage as another important mechanism of cardiotoxicity induced by sunitinib.

Sorafenib appears to inhibit at least 15 kinases, including the VEGFR, PDGFR, Raf-1/B-Raf, FLT3, and c-Kit (43, 107, 392). The incidence of sorafenib-induced CTX is not yet fully characterized. In two meta-analyses that included almost 7000 patients treated with sunitinib and 900 patients treated with sorafenib, rates of 4.1% for sunitinib-induced HF and 1% for sorafenib-associated cardiac dysfunction were shown (77, 321), but most of these observations were derived from retrospective analyses; only a few trials have assessed LV dysfunction prospectively. The work of Schmidinger et al. (344) showed that 3 out of 14 patients who were administered sorafenib and had cardiac events exhibited abnormal LVEF.

Recently, Feng et al. (98) observed that the axis between the brain-derived neurotrophic factor (BDNF) and a tyrosine kinase receptor (TrkB) is a key player in the regulation of myocardial excitation-contraction coupling, independently and in parallel to G protein-coupled receptor (GPCR) signaling. These data further support the concept that tyrosine kinase blockade with antineoplastic treatments can interfere with fundamental signaling, thus impairing cardiac mechanical work that may produce LV dysfunction (106).

In summary, several mechanisms may lead to cardiac dysfunction in patients treated with anti-VEGF therapy: cKit inhibition, altered activation of MAP with consequent alteration of energy production and mitochondrial function, PDGFR inhibition, and inhibition of signaling favoring cell survival, as well as angiogenesis inhibition and induction of arterial hypertension.

Importantly, hypertension is a main side-effect of all three major anti-VEGF drugs (272). Mechanisms of arterial hypertension include both functional (inactivation of endothelial nitric oxide synthase and production of vasoconstrictors such as endothelin-1) and anatomic (capillary rarefaction) modifications, which lead to vasoconstriction and an increase in peripheral vascular resistance (272, 277, 353). The main mechanisms of capillary rarefaction are considered the loss of pericytes due to PDGFR inhibition, along with inhibition of angiogenesis by VEGFR inhibition (353). Moreover, renal dysfunction due to angiogenesis inhibition can play a role in inducing and maintaining arterial hypertension. All said, the mechanisms of hypertension induced by antiangiogenic drugs have not yet been fully elucidated, and they are still an unresolved issue.

Bevacizumab produces severe hypertension that, at least in some circumstances, is not reversed by discontinuing the drug. Hypertension from anti-angiogenic agents has been suggested to be a biomarker of anti-cancer drug efficacy, because some patients who developed hypertension were observed to survive longer than those who did not (122). It has been shown that in cases of metastatic colorectal cancer, 20% of patients showed grade 2–3 hypertension. A partial remission was observed in 75% of subjects with bevacizumab-induced hypertension, and in only 32% of those with no hypertension. Moreover, patients who showed grade 2–3 hypertension had significantly longer progression-free survival than nonhypertensive patients (342).

Sunitinib can also cause hypertension, with an incidence ranging from 5% to 47% in different studies. Sunitinib was shown to induce hypertension (≥150/100 mmHg) in 47% of the subjects studied, with grade 3 hypertension seen in 17%. High blood pressure was observed within the first 4 weeks of therapy (32, 270).

Inhibition of VEGF signaling can also cause arterial and venous thrombosis. Arterial thromboembolic events (ATE) are mainly related to VEGF inhibition and consequent impairment of vascular homeostasis due to a reduction of nitric oxide (NO) synthesis, endothelial dysfunction, and production of vasoactive mediators that favor plaque instability and thrombus formation (212). Bevacizumab causes ATE more frequently in patients contemporaneously treated with other chemotherapeutic agents, in older patients, and in those who have experienced previous thrombotic events (315). The real burden of venous TE related to bevacizumab is less clear (161). The pathogenesis of venous events is always related to impairment of vascular homeostasis due to VEGF inhibition and to production of pro-inflammatory cytokines that favor activation of the coagulation cascade (144).

Sorafenib, pazopanib, and axitinib may also increase the risk of ATE. In a meta-analysis of 10,000 patients, the incidence of ATE was 1.7% in sorafenib-treated patients (49). Venous thrombotic events were reported in 3% of axitinib-treated patients, with ATE in 2% and transient ischemic attack in 1%. However, HF was observed in <1% of subjects treated with sorafenib and in 3% of patients treated with pazopanib (324).

Cardiac ischemia or myocardial infarction can also occur with sorafenib (3.8%), pazopanib (2%), and axitinib (<1%) (30). This is mainly related to the inhibition of VEGF signaling, leading to endothelial dysfunction, vasoconstriction, and microvascular rarefaction. Moreover, a reduction in VEGF levels creates an imbalance between cell survival and apoptosis. Endothelial cell apoptosis favors an exposure of sub-endothelial collagen and activation of the coagulation cascade, whereas reduced VEGF levels also interfere with the platelet-mediated fibrinolytic cascade (212).

Finally, patients treated with sorafenib, pazopanib, and vandetanib can present an electrocardiogram (ECG) with QT interval prolongation, which predisposes them to an enhanced risk of ventricular arrhythmias (194, 214). In particular, sorafenib can frequently induce QT prolongation (40.5%), thus greatly increasing the risk of ventricular arrhythmias (30). Vandetanib has also been related to a moderate risk of QT prolongation (16%), and more rarely, to life-threatening arrhythmias; in addition, it is associated with an enhanced risk of arterial hypertension (24%) and HF (21%) (434). In patients treated with pazopanib, the risk of QT prolongation and torsade de pointes, a dangerous polymorphic ventricular tachycardia leading to sudden cardiac death, is <2% (303).

In patients receiving drugs potentially interfering with QT duration, electrocardiographic monitoring of QT interval and periodic control of electrolytes are recommended. A reduction in the drug's dose can be warranted, but rarely discontinuation. Particular attention should be paid to patients with a personal or familial history of QT prolongation or to those who are under concomitant medication that can influence QT interval. For further details on the approaches to reduce CV toxicity, see sections X and XI.

A. Imatinib adverse events

A 5-year follow-up study on patients with newly diagnosed CML treated with imatinib showed an excellent CV safety record for the drug (285). However, Kerkelä et al. reported clinical data from 10 patients with normal ventricular function who developed severe cardiac dysfunction and New York Heart Association (NYHA) Class 3–4 HF after imatinib treatment, and they demonstrated that mice treated with imatinib develop a severe form of LV contractile dysfunction (185). The publication by Kerkelä et al. prompted many hematologists who had used imatinib to treat multiple patients to review their clinical database; the rates of CTX found were extremely low (16, 113, 138, 329). Further clinical studies then confirmed the excellent CV risk profile of imatinib (54, 60, 64, 80, 90, 148).

B. Mechanisms of imatinib-induced adverse events

Several studies investigated the mechanisms responsible for cardiac toxicity, and they found that imatinib-induced ABL inhibition leads to alterations in the endoplasmic reticulum (ER) stress response by activating the IRE1 kinase arm of the response. IRE1 activates apoptosis signal-regulating kinase 1 and c-Jun N-terminal kinase, which, in turn, causes inhibition of antiapoptotic B cell leukemia/lymphoma and cytochrome c release from mitochondria, thus leading to mitochondrial function impairment and cell death (43, 44). The inhibition of Abl kinase by imatinib is also responsible for PRKR-like endoplasmic reticulum kinase activation, which results in impaired protein import in the mitochondrial matrix, inducing mitochondrial death pathways (409).

Other than the potential proapoptotic effect on cardiomyocytes, imatinib has been shown to exert a number of cardiologically relevant effects. First, it causes fluid retention and edema that may contribute to a mistaken diagnosis of HF. This is probably due to imatinib's inhibition of PDGFR (356). Paradoxically, imatinib treatment can also protect against brain and lung edema during stroke and lung injury, respectively (39, 374). It has been demonstrated that in human endothelial cells pretreated with a variety of barrier-disruptive agents, imatinib reduces endothelial barrier dysfunction via inhibition of the abl-related gene (7, 47). Imatinib also has favorable effects on glucose control by improving both β cell function and peripheral insulin sensitivity (101, 130). Imatinib improves hemodynamics and reverses vascular remodeling and cor pulmonale in pulmonary arterial hypertension (PAH) by PDGF inhibition (343). Finally, imatinib was found to attenuate myocardial remodeling and to improve LV diastolic dysfunction in spontaneously hypertensive rats by affecting the PDGFR pathway without the blood pressure-lowering effect (170).

Nilotinib is a structural analog of imatinib that binds and potentially inhibits PDGFR and KIT family kinases, but not kinases of the SRC group (124). It is approved for the therapy of subjects with newly diagnosed Ph⁺ CML in any phase. Nilotinib is also approved for CML patients who are resistant to or intolerant of previous therapies, including imatinib.

C. Nilotinib adverese events

Early in the development of Nilotinib, there was evidence of QT prolongation, leading to a warning to perform serial ECGs, prevent and correct electrolyte abnormalities, and avoid concomitant drugs potentially responsible for QT interval prolongation (182). However, subsequent trials found no alarming signals of QT-related ventricular arrhythmias (267). Clinical trials, as well as studies analyzing real-life settings, found that nilotinib exacerbates the glycometabolic and lipid profile (26, 166, 167, 267, 313, 317). Over the past 5 years, several clinical trials, retrospective studies, meta-analyses, and postmarketing surveillance studies have shown that nilotinib-treated patients are at an increased risk of CV events, especially those related to peripheral artery disease (PAD) (2, 54, 64, 80, 118, 148, 190, 209, 267, 403). Oddly, these vascular adverse events were not reported during the first clinical studies.

D. Mechanisms of nilotinib-induced adverse events

Recent studies have documented a number of nilotinib-induced effects on endothelium, platelets, and coagulation, which, together with the metabolic effects, enhance the risk of vascular adverse events (4, 22).

As for myocardial toxicity, clinical studies have shown that nilotinib does not cause ventricular dysfunction. This finding concurs with experimental data showing that nilotinib and imatinib produce little or no damage to cardiomyocytes, in comparison with dasatinib, bosutinib, and ponatinib (137).

Dasatinib is a second-generation TKI, and it is more potent than imatinib. Unlike imatinib and nilotinib, dasatinib binds to TKs, in both their active and inactive conformational state. It is classified as a dual Abl/Src inhibitor, although it inhibits a broad spectrum of receptor kinases, including c-Kit and PDGFR (124).

Dasatinib is approved as a frontline therapy of Ph⁺ CML patients in the chronic phase, and for salvage treatment of Ph⁺ CML or Ph⁺ acute lymphoblastic leukemia (ALL) patients who are intolerant of or resistant to other TKIs.

E. Dasatinib adverse events

For years, the CV safety profile of dasatinib was considered similar to that of imatinib, with the exception of pleural effusion, which has been observed since the first use of the drug (383). Clinical studies with a long follow-up period found that pleural effusion occurs in one of every three patients, and that half of pleural events occur within the first 2 years of therapy. Pleural effusion requiring drug discontinuation occurs in <10% of the cases (351).

PAH is a rare but serious complication of dasatinib treatment. A transient significant increase of pulmonary arterial pressure detected by echocardiography was observed for the first time in 2017 among patients with pleural effusion (312). Subsequently, other reports emerged of PAH not associated with pleural effusion (263).

In clinical studies, PAH has been observed in 5% of subjects on dasatinib as a frontline therapy for 5 years and in 2.4% of patients treated with dasatinib for 7 years, who were resistant to or intolerant of other treatments (351). A recent assessment of all causes of dasatinib-related PAH confirmed by catheterization (41 patients) showed that a unique feature of dasatinib-induced PAH is different complete or partial reversibility on drug discontinuation (352). However, the French PH Registry that identified nine cases of dasatinib-associated PAH reported that most patients did not experience complete recovery, and it recorded two cases of cardiac death after dasatinib withdrawal (263).

Analysis of the FDA database designed to support postmarketing surveillance found that PAH was associated with dasatinib and not with other TKIs, and that it often occurred in patients with CV risk factors or a medical history of cardiopulmonary events (54).

F. Mechanisms of dasatinib-induced adverse events

Pleural effusion is considered the result of PDGFR inhibition, particularly of the PDGFR-β isoform, of which dasatinib is the most potent inhibitor; however, other mechanisms have been suggested (184). Due to the high frequency of lymphocytes in pleural fluid and tissue and the association with skin rash or history of autoimmunity, an autoimunne mechanism has been hypothesized (297). It has been suggested that an immune-mediated mechanism underlying the occurrence of pleural effusion could help to promote tumor regression. Patients who develop lymphocytosis and increased levels of specific lymphocyte subsets in association with pleural effusion have a higher major molecular response rate and better progression-free and overall survival than patients who do not (89).

The underlining molecular mechanism responsible for PAH is still poorly understood. Since dasatinib is a potent pan-Src inhibitor, it has been suggested that Src inhibition is pivotal in the development of PAH (264). The evidence that CML patients treated with dasatinib may display high circulating levels of E-selectin, intercellular adhesion molecule, and vascular cell adhesion molecule, and the observation that cultured pulmonary endothelial cells administered with dasatinib display an increased mitochondrial ROS production and apoptosis support the notion that dasatinib-induced PAH is caused by endothelial dysfunction (127).

Because vascular events have been described in patients treated with other TKIs, recent clinical trials have evaluated the incidence of coronary, cerebral, and peripheral thrombotic events in dasatinib-treated patients; a slightly higher risk was found, compared with imatinib. However, a recent meta-analysis showed that the use of dasatinib is associated with a significant increase in the risk of CV events. In a population-based cohort study, the incidence of myocardial infarction was 2.4 times higher in patients treated with dasatinib than in those treated with imatinib.

Bosutinib is a second-generation, dual Src/Abl TKI lacking significant PDGFR or c-KIT binding properties (124). Bosutinib is currently approved only for patients with Ph1 chronic-phase CML who were resistant to or intolerant of previous TKI therapy.

G. Bosutinib adverse events

Studies that evaluated bosutinib enrolled patients newly diagnosed with CML, or patients resistant to or intolerant of other TKIs; these studies documented a low incidence of cardiac and vascular events (31, 55, 57, 58). Similar to other TKIs, the incidence of adverse events was higher in second- or later-line treatment compared with first-line treatment. HF and coronary or PAD-related events were rare, suggesting that the cardiotoxic profile of bosutinib differs from that of other second- or third-generation TKIs. Most events occurred within the first year of therapy, and very few patients discontinued treatment because of these events, which were mostly managed with concomitant medications (55).

It should be noted that in a first-line setting, the overall incidence of events and dose reduction or interruption is similar for imatinib and bosutinib, even though pericardial disorders occurred more often in bosutinib-treated patients than in those treated with imatinib.

Ponatinib, a third-generation BCR-ABL TKI, was originally developed for the inhibition of pan-resistant ABL TK, and it is clinically effective in imatinib-, nilotinib-, and dasatinib-resistant CML. Ponatinib targets BCR-ABL in its inactive conformation and inhibits T315I and other BCR-ABL mutants. It also targets a broad spectrum of TK receptors, including epidermal growth factor receptor (EGFR), PDGFR, VEGFR, SRC kinase, c-KIT, FLT3, and RET (124).

H. Ponatinib adverse events

Ponatinib was tested in Ph⁺ patients with refractory CML or ALL enrolled in phase I (65 patients) and in phase II (449 patients) of the PACE trial. Both studies documented a high efficacy of the drug, accompanied by an increased risk of arterial thrombotic events (56, 59). For these reasons, ponatinib was approved with a warning of the risk of serious vascular events. Findings from these studies provided the rationale to perform a phase III study (EPIC) to test the hypothesis that ponatinib may lead to a deep molecular response in patients newly diagnosed with CML, and to assess its safety in patients with no previous exposure to other TKIs (226). The extended follow-up periods of the phase I and II studies documented that cardiac, cerebral, and peripheral thrombotic events continued to occur; this finding determined the early termination of the phase III study and led ARIAD Pharmaceuticals and the FDA to withdraw ponatinib from the market in October 2013. In January 2014, after a reassessment of the benefit/risk ratio of ponatinib, the FDA allowed its reintroduction into the market, with a recommendation to take appropriate measures to mitigate adverse effects in patients. The mechanisms of ponatinib CV toxicity are still poorly understood. Diagnostic and pharmacological measures aimed at reducing the CV risk were not included in the protocols of the phase I–III trials. The implementation of these preventive measures may help to improve the therapeutic index of ponatinib in the clinical setting (27). For further details on the approaches to reduce TKIs-induced CTX, please refer to sections X and XI.

A. Adverse effects

Although the potential of taxanes to induce cardiac dysfunction was likely overestimated in the past, they deserve a brief note. The most commonly used taxanes are docetaxel and paclitaxel, which are still used in breast, nonsmall cell lung, and ovarian cancers. Among the adverse effects of these drugs, a high incidence of arrhythmias has been observed, including asymptomatic sinus bradycardia (330), which usually is not a major problem for patients. A large range of conduction blocks and cardiac ischemia have been also observed in a few patients participating in trials (330). Ventricular arrhythmias are far less common: Only 0.26% of patients experienced ventricular tachycardia and ventricular fibrillation, and many of them were also co-treated with cisplatin (245, 330). In a clinical trial (EORTC 24971/TAX 323 Study Group), patients treated with docetaxel displayed a 1.7% incidence of cardiac ischemia (411). In about 3500 patients treated with paclitaxel followed by the Cancer Therapy Evaluation Program's Adverse Drug Reaction database, the overall incidence of severe (grade 4 and 5) cardiac events was <0.3% (12). These severe events were observed till 2 weeks after the beginning of paclitaxel treatment. Of note, many of these subjects had important CV risk factors, including coronary disease and hypertension.

B. Mechanisms of adverse effects

Taxanes exert anticancer effects, acting as anti-microtubule agents and fostering polymerization of tubulin. In this way, these agents induce microtubule dysfunction and alter cell division. Moreover, they may affect histamine release (330). These taxane-induced effects may perturb cardiac function. Brady-arrhythmias may be due to stimulation of histamine H₁ and H₂ receptors. The latter receptors may increase the possibility of reentry of depolarization, which is at the basis of ventricular ectopy (245, 330). Ischemia and increased myocardial oxygen demands render the heart more susceptible to arrhythmias. Of note, paclitaxel in combination with ANTs increases the incidence of CTX, which appears at lower cumulative doses with respect to ANTs alone (120). In particular, plasma levels of doxorubicin and the formation of the toxic metabolite doxorubicinol in myocytes can be enhanced by taxanes.

C. Ways to reduce CV toxicity

Cardiotoxic symptoms are frequently resolved on discontinuation of taxanes. Since histamine is considered a relevant player in taxane CTX, treatment with corticosteroids and antihistamines may reduce the incidence of clinically significant arrhythmias (245, 330).

A. Immune checkpoint inhibitors

Tumors exploit immune checkpoints to evade antitumor immunity. The success of cancer immunotherapy over the past decades has revolutionized the management of a wide array of malignancies. At the forefront of immunotherapy are immune checkpoint blockers, which have seen unprecedented success in cancer treatments because of their broad bioactivity across many tumor types. The two most prominent to date are the targeting of cytotoxic T lymphocyte-associated protein 4 (CTLA-4) and the interaction between the proteins programmed cell death 1 (PD-1) and programmed death-ligand 1 (PD-L1).

B. CTLA-4

It has structural homology to the co-stimulatory molecule CD28, and it can also bind B7 molecules on antigen-presenting cells (APCs) with a higher affinity than CD28. CTLA-4 is a competitive antagonist of the CD28-B7 interaction in that it blocks co-stimulation at the T cell-APC interaction, thus suppressing activation of its cell response (200) (Fig. 6). Ipilimumab and tremelimumab are monoclonal antibodies for clinical use against CTLA-4 (117, 354, 355, 431).

OPEN IN VIEWER

FIG. 6.

Role of CTLA-4 and its interaction with ipilimumab in tumor immunity. (A) Dying tumor cells release tumor neo-antigens that are taken up by APCs, which then present B7 co-stimulatory molecules to T cells. T cells recognize tumor neo-antigens on APCs and are activated. Activated T cells upregulate inhibitory checkpoints, such as CTLA-4 and PD-1, which block T cell activation and attack of tumor cells. (B) Ipilimumab blocks CTLA-4, resulting in T cell activation and destruction of tumor cells. See the text for further explanation and acronyms. APCs, antigen-presenting cells; CTLA-4, cytotoxic T lymphocyte-associated protein 4; PD-1, programmed cell death 1; TCR, T cell receptor.

C. PD-1/PD-L1 pathway

PD-1 is a checkpoint receptor that, by binding with its ligands (either PD-L1 or PD-L2), inhibits T effector functions by dampening signaling downstream of the T cell receptor (TCR) (397). Therefore, the expression of PD-L1 in the tumor micro-environment protects cancer cells from immune-mediated destruction. Several monoclonal antibodies against the PD-1/PD-L1 axis have been developed and/or approved as anti-cancer agents (396, 397).

PD-1, similar to CTLA-4, is expressed in activated cells, and its activation downregulates signaling that is mediated on antigen recognition by the TCR (354). PD-1 has two ligands, PD-L1 and PD-L2. PD-L1 can be expressed in many cell types, including tumor cells, immune cells, epithelial cells, and endothelial cells (354). PD-L2 is predominantly expressed in APCs. Pembrolizumab and nivolumab are PD-1-blocking monoclonal antibodies that are approved for treating different types of cancers. Atezolizumab is a PD-L1 targeted monoclonal antibody recently approved for treatment of certain tumors (354). Other inhibitory molecules (e.g., LAG3, OX40, TIM3, TIGIT) are being assessed for their potential utility as immunotherapeutic targets (15, 211, 279, 399).

D. Combination of checkpoint inhibitors

CTLA-4 or PD-1/PD-L1 blockade as a monotherapy has shown impressive tumor regression in some patients with different solid and hematological cancers (95, 225, 241); however, the overall long-term survival rate should be improved (304). Ongoing clinical trials combine different checkpoint inhibitors. In addition, some clinical trials combine checkpoint inhibitors with angiogenesis inhibitors (i.e., bevacizumab), EGFR, TKIs or BRAF inhibitors, and chemotherapy or radiation therapy (396).

E. Checkpoint inhibitors and immune-related adverse events

Checkpoint inhibitors can cause a novel spectrum of adverse events called immune-related adverse events (IRAEs) (254, 420), because they can induce autoimmune manifestations (75, 202). The most common IRAEs after treatment with CTLA-4 inhibitors include rash, diarrhea, colitis, hepatotoxicity, and endocrinopathies (115, 149). Immune-mediated CTX produced by these compounds has generally been observed in individual cases with variable presentations, comprising pericarditis and myocarditis (117, 141, 414, 431). The largest case series to date describing CTX after treatment with checkpoint inhibitors reported seven cases (141).

Monoclonal antibodies targeting PD-1 (such as pembrolizumab, nivolumab, and pidilizumab) or PD-L1 (atezolizumab, durvalumab) have shown a favorable toxicity profile in several trials (82). However, occasional reports of myocarditis after nivolumab (151) or pembrolizumab (207) treatment have been reported.

Although so far, there have been few case reports on cardiac adverse effects in patients treated with checkpoint inhibitors, several notes of caution should be considered. First, experimental data in PD-1-deficient mice are associated with dilated cardiomyopathy and sudden death by congestive HF (280, 288). In addition, PD-1 protects against inflammation in T cell-mediated myocarditis (385). Finally, PD-1 deficiency leads to the development of fatal myocarditis in mice genetically predisposed to autoimmunity (418). The latter observation is particularly important, because patients with underlying autoimmune disorders have always been excluded from clinical trials with checkpoint inhibitors. Therefore, the real-life risk of cardiac adverse events in patients with autoimmune disorders treated with checkpoint inhibitors is so far unknown.

Checkpoint inhibitors have only recently been introduced, and there are limited data on late-onset IRAEs after several years of initiating treatment. Careful follow-up of cardiac adverse events in long-term treated patients is warranted. Finally, and importantly, treatment with checkpoint inhibitors or sequential therapy with different checkpoint inhibitors, in association with chemotherapy or radiation therapy, is on the rise. Therefore, a combination or sequential therapies may contribute to some of the adverse cardiac effects caused by chemotherapy or radiation therapy.

A. Adverse effects

Although well tolerated, both capecitabine and its main compound can induce CTX. The onset of capecitabine CTX is generally rapid. It usually occurs within 4 days from the start of the first cycle of chemotherapy (174, 175). Prospective and retrospective studies have shown that previous events of cardiac ischemia are the strongest risk factors for fluoropyrimidine-induced CTX (195, 309, 310). ECG abnormalities have been found in patients treated with a 24-h 5-FU infusion in combination with capecitabine. Palpitations and chest pain are the most common clinical manifestations, accompanied by new-onset ECG alterations, such as ST segment deviation, sinus tachycardia, QT prolongation, and signs of myocardial ischemia and acute coronary syndromes. Less frequent are arterial hypertension, ventricular arrhythmias, LV dysfunction, HF, and cardiogenic shock. Cardiac arrest and sudden death are very rare events, and they account for <1% of clinical manifestations (5, 198, 309, 408).

The incidence of CTX produced by 5-FU varies between 0% and 35%, with a mortality rate ranging between 2.2% and 13.3%, depending on the dose, cardiac comorbidities, and route of administration (198, 309, 310). Larger trials and meta-analyses on breast and colorectal cancer patients, however, have reported an incidence of symptomatic CTX of 1.2%–4.3% during treatment (205, 407). A recent study found silent ischemia due to the cardiac stress test in 6%–7% of 5-FU-treated patients (216). Unfortunately, patients without overt CV risk factors are often treated as outpatients. As a result, symptomatic and asymptomatic cardiac events remain undiagnosed because of inadequate cardiac monitoring. This can lead to a discrepancy between observed and actual CTX, in terms of both symptomatic and asymptomatic patients.

B. Mechanism of adverse effects

Although 5-FU has a brief half-life, active metabolites are retained in all tissues, including cardiac and cancer tissues, resulting in a prolonged exposure of cells to the drug (198, 215, 261). Thymidine phosphorylase, the pivotal enzyme involved in the conversion of capecitabine to 5-FU, is highly expressed in both atherosclerotic plaques and cancer tissues, explaining the higher prevalence of capecitabine-induced CTX in subjects with coronary artery disease (CAD). Nevetheless, the pathophysiological mechanisms underlying CTX from antimetabolite drugs remain poorly understood. Several mechanisms have been proposed (Fig. 7): (i) Reduction in synthesis of NO, with consequent coronary spasms and endothelium-independent vasoconstriction via the protein kinase pathway (52, 296, 357); (ii) increased intracellular ROS/RNS, with consequent oxidative stress and cardiomyocyte apoptosis (205); (iii) increased endothelial thrombogenicity, with consequent induction of acute coronary syndromes (180); and (iv) interference with DNA and RNA growth by substituting for the normal building blocks of RNA and DNA.

OPEN IN VIEWER

FIG. 7.

Main cardiotoxic mechanism and effects induced by antimetabolite drugs. See the text for further explanation. 5-FU, 5-fluorouracil; HF, heart failure; RNS, reactive nitrogen species; ROS, reactive oxygen species.

The existence of a synergy between capecitabine and other anti-cancer drugs has also been suggested. Several studies have shown that CTX is more frequent in subjects administered a combined therapy of capecitabine and either taxanes or lapatinib than in patients treated with capecitabine alone (105, 174, 175, 278).

Studies in rabbits have shown that a single high dose of capecitabine caused hemorrhagic infarction of the ventricle walls, proximal spasms of the coronary arteries, and death within a few hours from intravenous injection. In contrast, repeated lower doses produced LV hypertrophy, concentric fibrous thickening of the coronary intima, and foci of necrotic myocardial cells (398). Due to the importance of redox stress in antimetabolite-induced CTX, please refer to sections X and XI for ways to reduce this toxicity.

A. Adverse effects

Experimental data suggest that chronic (over 12 weeks) proteasome inhibition is associated with systolic dysfunction and increased atherosclerosis (143, 284). A meta-analysis conducted by Xiao et al. reported that the incidence of CTX in subjects treated with bortezomib varied with tumor type and treatment regimen, in the range of 2.3%–3.8%, with a mortality rate of 3.0% (427).

Treatment with proteasome inhibitors may be associated with hypertension, HF, myocardial infarction, and cardiac arrest, especially if such therapies are associated with concomitant use of doxorubicin, lenalidomide, and steroids, or in patients with a history of cardiac events (17, 136, 151, 373, 423, 424). The incidence of HF in subjects receiving bortezomib is relatively low (up to 4%) compared with carfilzomib (up to 25%), a more potent and irreversible proteasomal inhibitor; patients treated with higher doses (≥36 mg/m²) of carfilzomib have a higher risk of cardiac toxicity (65, 328, 331).

B. Mechanisms of adverse effects

Proteasome inhibitors are compounds that block the activity of proteasomes, protein complexes that play a key role in degrading dysfunctional or unneeded proteins; these cellular complexes that break down proteins are particularly important for the functional maintenance of cardiomyocytes. Therefore, cardiac dysfunction may be expected if the function of these complexes is impaired.

Many mechanisms of chemotherapy-induced cardiotoxicity still remain to be clarified. It has been suggested (284) that bortezomib alters the function of cardiomyocytes through the impairment of mitochondrial energetics. The cardiomyocytes are contractile cells with a very high demand for ATP and may be particularly sensitive to agents that disrupt mitochondrial activity, such as the proteasome inhibitor bortezomib. Furthermore, a reduced synthesis of ATP could trigger the capillary tunneling, as revealed by a histopathological examination of heart sections of rats treated with bortezomib (284). Chronic uptake inhibitor of the proteasome is associated with increased oxidative stress at the level of the intima of the epicardial coronary arteries, resulting in thickening of the vessel wall, which can trigger premature atherosclerosis (143).

The addition of other chemotherapeutic agents such as ANT with proteasome inhibitors improves the effectiveness of antineoplastic therapy; however, this combination of drugs can cause cardiotoxicity. Spur et al. (368) have analyzed the function of the proteasome in primary cardiomyocytes treated with doxorubicin in the presence of proteasome inhibitors. Interestingly, the authors concluded that, contrary to carfilzomib, which targets both the β5 standard proteasome and the LMP7 immunoproteasome subunit, immunoproteasome-specific inhibitors with known anti-tumor capabilities for specific cancer cells, such as multiple myeloma, may be advantageous to reduce the mortality of cardiomyocytes, when there is a combination therapy, and, therefore, may be envisioned as a way to reduce CV toxicity, when compared with traditional proteasome inhibitors.

C. Ways to reduce CV toxicity

The initial step in CV management of subjects treated with proteasome inhibitors is to assess their baseline risk for CTX, by taking a clinical history and conducting an examination. More frequent surveillance may be warranted for subjects with higher baseline clinical risk or abnormal cardiac baseline function.

CTX induced by proteasome inhibitors may be reversible in some patients with prompt cessation of these therapies and initiation of traditional HF treatments (123). ACE-Is or angiotensin II receptor blockers (ARBs) in combination with β blockers are recommended in patients with symptomatic HF, similar to the general HF population and in asymptomatic cardiac dysfunction, to prevent further cardiac dysfunction or the development of symptomatic HF in patients at a high risk (432).

Single-center experiences have shown a decrease in the occurrence of cardiac events when infusion time of carfilzomib was settled to 30 min and a cardioprotective activity of dexrazoxane was exhibited (17, 136).

In some cases, re-administration of carfilzomib with dose modification is possible. Long-term surveillance should be considered for patients who developed CTX during these therapies and for those in whom cardioprotective treatments have been started, to confirm recovery or to detect irreversible cardiac dysfunction.

A. Drugs with antioxidant properties

The cardiotoxic mechanisms involving oxidative stress provide the rationale for using drugs with antioxidant properties (29, 34, 73, 219, 224, 242, 350, 366, 379, 381) (Fig. 11).

OPEN IN VIEWER

FIG. 11.

Drugs and cardioprotective strategies against antineoplastic-induced cardiotoxicity. The main mechanism(s) of protection are reported. See the text for further explanation. ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; NRG1, neuregulin 1; Top2, topoisomerase 2.

The iron-chelating drug dexrazoxane has been identified as the most promising cardioprotective strategy in the past years and, up to now, it is the only cardioprotective drug approved by the FDA for ANT-induced CTX. Dexrazoxane is a pro-drug that, once it enters the cardiomyocyte and is rapidly turned into its active form, counteracts the formation of ANT-iron complexes and the subsequent formation of ROS (358) (Table 2). Its efficacy has been extensively tested in multiple randomized trials and two pooled analyses (350, 379), and it is confirmed in different subsets of cancer patients (222, 381).

It is likely that dexrazoxane possesses unique pharmacodynamic protective mechanisms, because no other iron chelator has shown any cardioprotective effects (358). At the molecular level, dexrazoxane has been shown to modify the Top2β configuration, preventing its interface with ANTs, thereby impeding the realization of the Top2-DNA cleavage complexes (213, 229). However, dexrazoxane has been underused in recent years, despite strong evidence of its cardioprotection (367). Although CTX mitigation induced by new-generation β blockers could be due more to their antioxidant properties than to their β-adrenergic receptor-blocking action (243, 366, 430), it appears that the cardioprotective effect of iron chelators does not stem from their antioxidant properties, but rather from their ability to affect iron regulatory proteins and to reduce iron accumulation (42, 114). Interestingly, close derivatives of dexrazoxane lacking Top2β were found not to be protective in relevant chronic ANT CTX models (178, 242), which also suggest Top2β importance in the cardioprotective mechanism. These observations indicate that ANT-mediated damage relies on complicated alterations of the intracellular metabolic pathways or Top2β interaction. Moreover, recent preliminary data from Stěrba and coworkers suggest that cardioprotective effects of dexrazoxane against chronic ANT cardiotoxicity are due to its interaction with Top2β, rather than to its metal-chelating metabolite (213, 372). Nevertheless, the discordant results underline the need for further investigation to ascertain the exact biological interaction of anti-cancer drugs with the cellular redox balance.

B. Novel strategies in cardioprotection against ANT-induced CTX

Several CV drugs have been tested as potentially effective in counteracting ANT-induced CTX, and particularly statins, ranolazine, and phosphodiesterase-5 inhibitors.

Lovastatin and fluvastatin, hydroxymethylglutaryl-coenzyme A reductase inhibitors, showed a cardioprotective effect in experimental studies (159, 320). Of note, these data derive from acute, subacute experimental cardiotoxicity studies and cell-based studies performed on proliferating H9c2 cell lines, not cardiomyocytes. Therefore, effects of ANTs in this setting deserve caution since these drugs possess strong antiproliferative action besides cardiotoxic effects. The effects of the statins on the Top2β have not been directly demonstrated in the studies, just indirectly suggested to be Top2β related. Besides these caveats, LV systolic dysfunction induced by doxorubicin was not affected by lovastatin (142). Only small clinical studies have evaluated the effects of statins in patients treated with ANTs (48, 348), reporting only minor positive results. Unfortunately, the study from the Marwick group (348) was a retrospective observational analysis, and even Chotenimitkhun et al. (48) do not appear to have conducted a standard controlled randomized trial. Therefore, larger studies are, therefore, required to confirm these findings.

Late I_Na inhibition with ranolazine has been proposed as a cardioprotective strategy in several experimental models of cardiac dysfunction (316, 391, 394). Ranolazine was shown to preserve heart function in mice treated with ANTs by reducing oxidative stress. The inhibition of elevated [Na⁺]_i levels with ranolazine would prevent calcium overload and the occurrence of oxidative damage by suppressing ROS generation, offering an advantage over conventional antioxidant strategies that counteract ROS after their production (196). The INTERACT study (259), designed to assess whether ranolazine could prevent ANT-induced diastolic dysfunction, indicated that ranolazine was a promising cardio-oncologic drug, but more convincing data are still required.

Phosphodiesterase-5 inhibitors have also been tested as cardioprotective agents. Sildenafil demonstrated cardioprotective activity against ANT-induced LV dysfunction by opening mitochondrial K_ATP channels, preserving mitochondrial membrane potential and myofibrillar integrity, and preventing cardiomyocyte apoptosis (100). Tadalafil attenuated ANT-induced cardiomyopathy through NO-mediated increases in cGMP levels (177, 197).

Finally, redox mechanisms have also been proposed for the ability of NRG1 to limit antiblastic drug-induced damage in adult rat cardiomyocytes (390).

Another redox compound attracting the interest of researchers and cardio-oncologists is hydrogen sulfide (H₂S), whose role in cancer is still poorly understood. The discoveries that ROS-induced inhibition of cystathionine gamma-lyase, a key enzyme in the synthesis of H₂S, is involved in ANT-induced CTX in H9c2 cardiomyocytes, and that the administration of exogenous H₂S can confer protection against this CTX, have suggested a new tool with antioxidant properties in therapeutic strategies against antineoplastic drug-induced CTX (33, 247, 295). Further research in appropriate animal models is required, as well as multi-centric randomized trials that compare the efficacy of these new compounds with that of already approved cardioprotective drugs.

Experimental studies have shown that sunitinib inhibits AMPK activity, whereas restoration of this enzyme's activity inhibits cell death (186). It has been speculated that promoting AMPK activity with metformin may reduce sunitinib-induced CTX. Although metformin can counter heart dysfunction in different experimental models of cardiac injury (53, 129) (Table 3), an in vitro study that focused on sunitinib-related CTX did not find any beneficial effect of metformin (33). As mentioned earlier, a recent study (98) demonstrated that TrkB, a tyrosine kinase receptor, and its endogenous ligand BDNF could directly modulate the cardiac excitation-contraction coupling, independently or in parallel with GPCR-mediated signaling, supporting the concept that inhibition of the tyrosine kinase receptor with antineoplastic drugs can interfere with fundamental pathways, damaging cardiac contractility (106). At the same time, this study highlights BDNF as an alternative potential therapeutic strategy for HF.

C. Nonpharmacologic strategies

Apart from pharmacologic interventions, we should also consider that nutritional supplementation and exercise training may have important antioxidant properties (8, 139, 192, 345, 346, 359, 369).

Dietary supplementation of antioxidants proved to be able to alleviate doxorubicin cardiotoxicity in animal models (8, 426). However, there is still scant evidence on both quality and sample size, suggesting that some antioxidant supplementations may lower ANT-induced CTX (109). Nonpharmacologic strategies for countering ANT-induced CTX include lifestyle interventions and exercise “prehabilitation,” a type of preventive exercise rehabilitation. It has been speculated that aerobic exercise can decrease ROS generation, inhibiting pro-apoptotic signaling, restoring proper calcium cycling, and stimulating the AMPK pathway, thus improving myocardial energetics (345) (Table 3). In addition, exercise can have a positive impact on CV risk factors, such as hypertension, high cholesterol and lipids, overweight and obesity, and high blood glucose or diabetes (192). In cancer survivors, a short period of mixed aerobic and resistance exercise builds tolerance and flexibility in physical activity (369).

Although several studies have shown a role for aerobic exercise in countering ANT (343)- and trastuzumab (139)-related CTX, additional work is needed to better characterize the effects of training on antineoplastic drug-induced CTX (346).

D. Biomarkers of oxidative/nitrosative CTX

Since the importance of ROS/RNS production as early mediators of chemotherapy-related CTX has been ascertained, we should identify useful circulating biomarkers of oxidative/nitrosative CTX. Biomarkers with redox significance include the metabolomic identification of acetate and succinate (8). A decrease in NAD(P)H:quinone oxidoreductase 1 activity and an increase in ROS generation by NAD(P)H oxidases have been considered early biomarkers of ANT-induced CTX (283). A significant variation in IL-6, IL-6R, and ROS has been observed together with a change in the enzymatic activity of GPx, which was correlated with an early variation in longitudinal systolic function in patients after the administration of epirubicin (73, 74).

Other potential CTX markers with redox importance under investigation in cardiology are high-sensitivity C-reactive protein, heart-type fatty acid-binding protein (H-FABP), glycogen phosphorylase BB (GPBB), and circulating microRNAs, especially microRNA-146a (miR-146a). These are considered new potential markers for the diagnosis and risk stratification of acute coronary syndromes, and they may be helpful in the early detection of anti-cancer CTX; however, the clinical data available in the chemotherapy-induced CTX scenario remain insufficient (131, 152 –155, 371, 419). For a discussion on other biomarkers, see section XI.D.

A. Cardioncology: an expanding science for an expanding problem

Assessing CTX from antineoplastic drugs is crucial. Screening for CTX at a preclinical stage is essential when investigating the effects of TKIs. Many of these drugs cause an increased risk of arrhythmia; thus, most of the preclinical work focuses on hERG (K⁺ channel). A major limitation of in vitro work is that primary myocyte cultures of animal or human hearts dedifferentiate and die after a short while. Hence, assays based on stem cells and on the use of engineered heart tissue may better reflect in vivo settings. Such experimental studies could integrate assessment of several effects, including those on membrane action potential, calcium cycling, and myofilament function, as well as those on gene expression and cell survival (69, 88, 236, 376).

In clinical practice, most of the methods used to assess heart function are often inaccurate, and they may also mislead physicians to stop potentially lifesaving therapies. Of note, patients with advanced cancer may already exhibit potential confounder symptoms mimicking anti-cancer drugs' side-effects. Moreover, complications such as myocardial ischemia, arterial hypertension, and arrhythmias are readily recognized, but identifying mild LV dysfunction (especially at a subclinical stage) can be more challenging (376). In this context, a prompt and sensitive monitoring of CV function plays a pivotal role.

In addition, the time point when CTX can become clinically overt varies substantially. Several antineoplastic drugs (i.e., ANTs) can induce progressive cardiac remodeling and late cardiomyopathy, whereas others (i.e., trastuzumab) can produce transient LV dysfunction without known long-term consequences (376, 432). Furthermore, the inability to exactly predict the long-term consequences of tumors leads to a false diagnosis of CV complications. Owing to all these concerns, cardioncology, a new branch of CV sciences, recently developed, with the aim of preventing, identifying, and treating CV complications of cancer therapy, optimizing clinical status for cardiac patients who need potentially cardiotoxic treatment, facilitates in any way cancer therapy in patients with CV complications and surveilles for late side-effects in cancer survivors.

B. Diagnostic tools for CTX detection

The most common clinical diagnostic tools for the screening and detection of CTX are biomarkers and cardiac imaging, including echocardiography, multiple-gated acquisition (MUGA) scan, and cardiac magnetic resonance (CMR) (Table 4). Nevertheless, an ECG is recommended for all subjects before and during therapies, particularly for detecting arrhythmias and QT interval prolongation. Moreover, subjects on pyrimidine analogs should be monitored closely for cardiac ischemia by regular ECGs, and chemotherapeutic treatments should be withheld if ischemia is observed (432).

C. LVEF and emerging modalities for assessing myocardial damage: some claryfing points

LVEF is the most frequently used parameter for assessing myocardial function, and echocardiography and MUGA scans are largely employed in clinical practice for this purpose (40, 41, 62, 171, 206, 234, 299, 307, 376, 432, 438). An LVEF decrease (>10%) to a value <50% identifies patients with CTX in both imaging modalities (307, 432). Compared with echocardiography, MUGA provides the advantage of a high reproducibility, but it is hampered by significant disadvantages, such as cumulative radiation exposure and no data on other cardiac structures (299, 432). However, the choice of modalities for an early detection and monitoring of CTX depends on local expertise and availability, being careful to use the same imaging modality with the best reproducibility for continued screening throughout the treatment period (432). Imaging techniques that provide additional relevant clinical information (e.g., right ventricular function, pulmonary pressures, valvular function, pericardial evaluation) and high-quality radiation-free imaging are preferred (307, 432). Confirming LVEF decrease by repeated cardiac imaging (preferably by echocardiography) 2–3 weeks after the baseline initial decrease in LVEF is strongly recommended, as well as categorizing LVEF reduction as asymptomatic or symptomatic, and repeating echocardiography measurements during follow-up (432). Repeated MUGA is not recommended as a first-line option, and it should only be conducted if the radiation exposure is less than five mSv (432).

On the other hand, echocardiographic measurements, though simple, noninvasive, and readily available, are often variable and of a low sensitivity (290), because of the serious damage and myocyte loss, leading to an observable reduction in LVEF (92, 416). On these grounds, novel indices of ventricular function are being introduced (37, 146, 150, 206, 386). Speckle tracking echocardiography has shown the highest sensitivity in detecting early alterations in LV function associated with future development of cardiomyopathy (134, 146, 306, 307, 395, 438). In particular, a global longitudinal strain (GLS) >15% relative percentage reduction from baseline can identify patients who are at risk for CTX (97, 172, 307, 432).

Other imaging modalities, such as CMR, appear very promising, although their availability is still limited (97, 238, 290, 299, 388, 415). CMR is the gold-standard technique for LVEF estimation and can identify myocardial edema, inflammation, capillary leak, and fibrosis, enabling an accurate diagnosis of early and late CTX (239, 388, 432). Other recent strategies include measuring uptake of iodine-124-metaiodobenzylguanidine, a radiolabeled analog of norepinephrine, which is reduced after myocardial insults (145, 239, 415). In addition, antimyosin antibodies that detect myosin exposed after cardiac damage could be useful (235, 239, 401, 402), whereas the uptake of radiolabeled chemotherapeutics could be used as a predictor of CTX (18, 239, 301).

D. Cardiac biomarkers

Troponins and natriuretic peptides can be correlated with findings from cardiac imaging for the early detection of CTX (283, 292). Natriuretic peptide-level elevations were demonstrated to predict the development of future CTX in patients with NHL and with breast cancer (99, 119). Persistent B-type natriuretic peptide (BNP) rises, after exposure to ANTs, were more predictive of following LV dysfunction than transient elevations (327, 335). However, the utility of NPs for predicting anticancer drug-induced CTX is still controversial. Indeed, several studies did not confirm a prognostic value for BNP levels, especially when this biomarker was used alone (66, 340, 378). It has been observed that troponins I and T can predict late ANT-induced CTX in children (223) and in adult patients undergoing high-dose chemotherapy. Not only the increase but also the pattern of the elevation, and particularly a persistent increase after 1 month since treatment, added prognostic information. Conversely, persistently negative TnI identified low-risk patients who did not need a strict cardiological follow-up (38). Patients with troponin increase may benefit from early cardioprotective treatment with ACE-Is (36). In patients treated with trastuzumab, especially after previous exposure to ANTs, troponin I elevation can identify those who will develop LV dysfunction and who will not recover in spite of treatments for HF (37). High-sensitivity troponin I is more sensitive in the evaluation of subclinical depletion of cardiac physiologic reserve, and a new elevation of this marker predicted subsequent LV dysfunction in patients receiving a conventional dose of ANTs and/or trastuzumab (203). The combination of high-sensitivity troponin with GLS predicted the risk to develop future cardiotoxicity with increased sensitivity (93%) and negative predictive value (91%) compared with the evaluation of a single parameter in patients with breast cancer (340). Therefore, “cardiac biomarkers plus cardiac imaging” should be encouraged to detect CTX early. There are still scant data supporting the use of biomarkers to predict future cardiac dysfunction due to immune and targeted cancer therapies. Few data indicated an increase of troponin in subjects administered sunitinib and sorafenib for metastatic renal cell carcinoma, who developed ventricular function impairment (344). Several other biomarkers have been studied or are under investigation to stratify the risk of cardiotoxicity. Myeloperoxidase, an index of oxidative stress, measured after ANT administration, seems to correlate with subsequent development of CTX (203). FABP and GPBB have also been proposed as novel biomarkers of chemotherapy-mediated cardiac damage, which are even more precocious of troponin and BNP (86, 152, 153). MicroRNAs have emerged as an attractive tool to detect myocardial injury. miR-146a may induce downregulation of ErbB4 and subsequent stimulation of apoptotic pathways. Preclinical studies have shown increased levels of miR-146a after doxorubicin administration (155). However, more trials with long-term follow-up are needed to validate these data.

The use of cardiac biomarkers in the prediction of cardiotoxicity has not only several advantages (accuracy, reproducibility, high sensitivity, wide availability, and relatively low cost) but also limitations, for example, inter-assay variations, insufficient evidence to establish specific cutoffs for enzyme elevations in this setting, and uncertainty in the timing of blood sampling (432). Standardization in the assessment of cardiac toxicity using biomarkers is needed, and multi-center trials are strongly required before establishing more clear recommendation (376). Table 5 summarizes the main biomarkers that are evaluated for the detection of cardiac toxicity.

E. Detection of vascular complications: old and new strategies

Structural endothelial damage, endothelial dysfunction, acceleration of atherosclerosis, prothrombotic actions, and vascular inflammation are some of the underlying mechanisms of cancer therapy-related changes in vessel function. VEGF inhibitors, TKIs (such as nilotinib, ponatinib, dasanitib), cisplatin, 5-FU, and ANTs can induce hypertension, ischemia, and TE through their endothelial toxicity. CAD, stroke, PAD, PAH, arterial/venous TE disease, and nephrotoxicity are the main cinical presentations of vascular damage by cancer therapy. In particular, nuclear cellular actions by ANTs may adversely modulate collagen turnover and vascular wall homeostasis, leading to acute damage (endothelial dysfunction, increase in smooth muscle tone) and chronic damage (atherosclerosis, increased collagen synthesis), and resulting in an impaired CV coupling (81, 84, 273). Age and CV risk factors may influence the extent of cardiac dysfunction (Fig. 12). VEGF inhibitors increase vascular tone due to a decrease in NO production, increased peripheral resistance from endothelial damage (sloughing) and dysfunction, and capillary rarefaction (96). Cisplatin can induce myocardial ischemia through an increased procoagulant status.

OPEN IN VIEWER

FIG. 12.

Simplified algorithm showing how cancer therapy alone or together with demographic and conventional risk factors may determine cardiac damage; both factors can increase myocardial and arterial stiffness, thus impairing ventricular-vascular coupling and, finally, leading to HF, with either decreased ejection fraction (EF) or preserved EF (HFPEF). CV, cardiovascular.

For an early detection of vascular complications, current reccomendations (432) simply suggest monitoring CV risk factors, assessing patient functional status, measuring blood pressure before cancer treatment, and optimizing anthypertensive therapy during treatment with a goal of <140/90 mmHg (particularly in patients treated with VEGF inhibitors). Moreover, the identification of patients with pre-existing CAD in those undergoing 5-FU is of paramount importance since this drug substantially increases the risk of developing myocardial ischemia with a vasospam-recognized mechanism. In addition, it should be considered that clinical factors (cancer-, treatment-, and patient related) could be related to an enhanced risk of venous TE in subjects administered TKIs. Nilotinib, ponatinib, and BCR-ABL TKIs used in treating CML are, instead, associated with PAD and stroke. The evaluation of PAD risk at baseline (risk factor assessment, clinical examination, ankle-brachial index measurement) is, therefore, recommended in these patients (432). Moreover, cerebrovascular ultrasound screening is indicated in subjects who are irradiated for head and neck cancer or lymphoma, especially beyond 5 years after irradiation (432).

However, to date, a common standard approach that is useful in clinical practice for early detection of vascular complications such as atherosclerosis is missing.

The assessment of arterial stiffness, which increases early in the atherosclerotic process, to provide vasculotoxic profiles of chemotherapeutic agents seems to be very proimising. These could be useful in the future in improving therapeutic strategies (84, 179, 273, 425). Drafts et al. found, by magnetic resonance imaging of the aorta, an early (∼3 months) and abrupt increase of pulse wave velocity, a marker of arterial stiffness, which was dose-, age-, and CV risk factor independent in patients treated with ANTs; furthermore, this increase was accompanied by a subclinical reduction in LVEF and strain (81). This is one of the first demostrations of early atherosclerosis with associated impaired ventricular-arterial coupling in cancer survivals. More recently, a marked stiffening of large elastic arteries was demonstrated by applanation tonometry also in patients undergoing VEGF inhibitors (265). Accordingly, ACE-Is should be preferred because of their anti-stiffening properties in the aorta in hypertensive patients treated with ANTs and VEGF inhibitors.

F. Timing of clinical evaluation and follow-up: the importance of risk factors assessment

To date, no algorithm has been established for evaluation and follow-up of patients before, during, and after cancer therapies. However, a strong collaboration between cardiologists and oncologists is required to monitor and treat oncology patients (3, 432). Of course, the cardio-oncology team must be equipped appropriately to address the most common questions (i.e., CTX prediction, prevention strategies, high-risk patient monitoring, management strategies) with an evidence-based approach (410, 425).

Long-term surveillance of cancer survivors with a potential for late-onset CV complications should involve cardio-oncology teams. According to recent clinical guidelines on cancer treatment and CV toxicity, the first step in cardioprotection should be the identification of subjects at a higher risk; this depends on several factors: cumulative dose, body mass index >30 (especially for anti-HER2 compounds), age (elderly and pediatric population treated with ANTs) (199, 376, 380), concomitant or previous radiation therapy, previous cardiotoxic cancer treatment, pre-existing conditions such as cardiovascular disease (CVD), demographic and other CV risk factors, and lifestyle and genetic risk factors that may predispose patients to CTX at lower ANT doses (37, 50, 432). Although recent experimental studies from Moulin and collaborators (271) seem to suggest the opposite, female patients appear to be more susceptible to doxorubicin damage (222 –224).

Gender-specific pharmacokinetic differences or even a degree of protection conferred by androgens could be responsible for this phenomenon. It has been observed that testosterone can reduce ANT-induced damage in cardiac myocytes in vitro (6).

G. Drug-specific timing for CTX monitoring and treatment

The exact timing and frequency of imaging and/or biomarker sampling will depend on the specific antineoplastic therapy, total cumulative dose, treatment protocol and duration, and baseline CV risk (432).