Progress in Cardiovascular Diseases 53 (2010) 114–120www.onlinepcd.com

Why do Kinase Inhibitors Cause Cardiotoxicity and Whatcan be Done About It?Hui Cheng, Thomas Force⁎

Center for Translational Medicine and Cardiology Division, Thomas Jefferson University, Philadelphia, PA

Abstract Cancer growth and metastasis are often driven by activating mutations in, or gene amplications of,

Statement of Conf⁎ Address reprint

Medicine, Thomas Je1025 Walnut St., Phil

E-mail address: th

0033-0620/$ – see frodoi:10.1016/j.pcad.20

specific tyrosine or serine/threonine kinases. Kinase inhibitors (KIs) promised to provide targetedtherapy—specifically inhibiting the causal or contributory kinases driving tumor progressionwhile leaving function of other kinases intact. These inhibitors are of 2 general classes: (1)monoclonal antibodies that are typically directed against receptor tyrosine kinases or their ligandsand (2) small molecules targeting specific kinases. The latter will be the focus of this review. Thisclass of therapeutics has had some remarkable successes, including revolutionizing the treatmentof some malignancies (eg, imatinib [Gleevec] in the management of chronic myeloid leukemia)and adding significantly to the management of other difficult to treat cancers (eg, sunitinib[Sutent] and sorafenib [Nexavar] in the management of renal cell carcinoma). But in someinstances, cardiotoxicity, often manifest as left ventricular dysfunction and/or heart failure, hasensued after the use of KIs in patients. Herein we will explore the mechanisms underlying thecardiotoxicity of small-molecule KIs, hoping to explain how and why this happens, and willfurther examine strategies to deal with the problem. (Prog Cardiovasc Dis 2010;53:114-120)

© 2010 Elsevier Inc. All rights reserved.

Keywords: Kinase inhibitors; Cardiotoxicity; Heart failure

Protein kinases and cancer

Protein kinases are enzymes that catalyze transfer of aphosphate residue from ATP to tyrosine, serine, orthreonine residues in their substrate proteins. Thisphosphorylation of substrates results in changes insubstrate activity, subcellular location, stability, etc.Although the gain-of-function mutations, gene amplifica-tions, and/or overexpression that drive tumorigenesis canoccur in a variety of different gene classes, these genes, inmany cases, encode protein kinases, typically tyrosinekinases (TKs).1 Approximately 90 of the 518 kinases in thehuman kinome are TKs.2 Tyrosine kinases play central

lict of Interest: see page 119.requests to: Thomas Force, Center for Translationalfferson University, College Building, Suite 316,adelphia, PA 19107.omas.force@jefferson.edu (T. Force).

nt matter © 2010 Elsevier Inc. All rights reserved.10.06.006

roles in transducing extracellular signals (ie, growth factorsand cytokines) into activation of signaling pathways thatregulate cell growth, differentiation, metabolism, migra-tion, apoptosis, etc. Kinases that are mutated or over-expressed in cancers typically activate cellular pathwaysthat lead to promotion of cell cycle entry (proliferation),inhibition of proapoptotic factors, activation of antiapop-totic factors, and/or promotion of angiogenesis.

Data from tumor sequencing projects have foundremarkable mutation rates in protein kinases. One studyfound that mutations in as many as 120 kinases(approximately 25% of the kinome) were present in somecancers.3 Furthermore, many of these mutations were notjust “bystanders” but were so-called driver mutations (ie,playing a role in tumor progression). Given this complexity,it seems inconceivable that inhibiting individual kinases incancer would be effective, except for the relatively raremalignancies that are truly “oncogene-addicted” to aspecific mutated kinase (eg, chronic myeloid leukemia

114

Table 1FDA-approved KIs for cancer therapy

Agent (Trade Name) Primary

Imatinib (Gleevec) Bcr-AblNilotinib (Tasigna) Bcr-Abl and mostDasatinib (Sprycel) Bcr-Abl and mostSunitinib (Sutent) VEGFRs, PDGFRSorafenib (Nexavar) Raf-1/B-Raf, VEGLapatinib (Tykerb) EGFR (ERBB1),Gefitinib (Iressa) EGFRErlotinib (Tarceva) EGFRPazopanib (Votrient) VEGFR, PDGFRTemsirolimus (Torisel) mTOREverolimus (Afinitor) mTORSirolimus (Rapamune) mTOR

Abbreviations: IRMs indicates imatinib-releukemia; CMML, chronic myelomonoclymphocytic leukemia; RCC, renal cell caduring transfection; mTOR, mammalian ta

Abbreviations and Acronyms

TK = tyrosine kinase

KI = kinase inhibitors

TKI = tyrosine kinaseinhibitors

CML = chronic myeloidleukemia

GIST = gastrointestinal stromaltumors

LV = left ventricle

AMPK = AMP-activatedprotein kinase

VEGFR = vascular endothelialgrowth factor receptor

PDGFR = platelet-derivedgrowth factor receptor

PI3K = phosphatidylinositol3-kinase

115H. Cheng, T. Force / Progress in Cardiovascular Diseases 53 (2010) 114–120

[CML] and Bcr-Abl).4

Surprisingly, “targetedtherapeutics” have radi-cally transformed thetreatment of some he-matologic malignanciesand solid tumors.

Kinase inhibitors

The identification ofmutated or amplifiedkinases has allowed thedevelopment of thera-peutics specifically tar-geting the oncogenickinases. Although on-cogenic mutations com-monly occur in otherclasses of proteins inaddition to kinases,such as cell cycle reg-

ulators and proapoptotic or antiapoptotic factors, kinaseshave become favorite targets of the pharmaceutical industrydue not only to their importance in tumor initiation andprogression but also to the relative ease with whichinhibitors can be made (see below). This has led to anexplosion in drug development targeting TKs (TKinhibitors) and, to a lesser but increasing extent, serine/threonine kinases. At present, 12 small-molecule kinaseinhibitors (KIs) are Food and Drug Administration (FDA)–approved for cancer therapy (Table 1), with several moreseeking approval over the next 2 years7 and many more(N100) in various phases of development. The first KI to

Targets

Other

Abl, c-Kit, PDGFRIRMs Abl, c-Kit, PDGFRIRMs Abl, c-Kit, PDGFRs, c-Kit CSF-1R, FLT3, REFR2, PDGFRβ c-Kit, FLT3, etcHER2 (ERBB2) NI

NINI

, c-Kit NINININI

sistant Abl mutants; NI, none idenytic leukemia; HES, hypereosinoprcinoma; CSF-1R, colony-stimulatinrget of rapamycin. Please see text fo

reach market was imatinib in 2001.8 It is still the mostcommercially successful KI, with sales close to $4 billion in2009. Imatinib revolutionized the treatment of CML.Before the introduction of imatinib, CML was uniformlyfatal within 5 years, whereas now, ≈90% of patients arealive 5 years after diagnosis. Indeed, this and other drugshave changed our thinking about some cancers that can nowbe viewed as a group of diseases that, even if not curable,can be managed effectively for years, similar to many otherchronic diseases.

Mechanisms of action of KIs

Small-molecule KIs typically compete with ATP forbinding to the ATP pocket of the kinase. If ATP cannotbind, phosphotransferase activity is blocked and down-stream substrates cannot be phosphorylated, even if thekinase is fully activated. In the cell, ATP is present inmillimolar concentrations, but KIs will be present innanomolar to very low micromolar concentrations. Thus,the KIs must bind with very high affinity. Because thestructure of the ATP pocket is known for many kinasesand is highly conserved across the human genome, it isrelatively easy to make an inhibitor that blocks the ATPpocket of a kinase of interest. These inhibitors are termedtype I inhibitors. Given the degree of conservation, it is notsurprising that lack of selectivity is an issue with most typeI inhibitors.5 Type II inhibitors (eg, imatinib and therelated nilotinib) not only bind the ATP pocket but alsointeract with a site adjacent to the pocket, generallymaking them more selective.9 Furthermore, unlike type Iinhibitors, which only bind to an active kinase (becausethe ATP pocket is only accessible when the kinase isactivated), type II inhibitors can also bind to the kinase

Representative Malignancies

s, DDR1, etc5,6 CML, Ph+ ALL, CMML, HES, GISTs Imatinib-resistant CML, ALL, GISTs, DDR1, etc5,6 Imatinib-resistant CML, ALL, GISTT, etc RCC, GIST

RCC, hepatocellular carcinomaHER2+ breast cancer, ovarian cancer, gliomas, NSCLCNSCLC, gliomasNSCLC, pancreatic cancer, gliomasRCCRCCRCCRCC

tified; Ph+ ALL, Philadelphia chromosome–positive acute lymphocytichilic syndrome; NSCLC, non–small-cell lung cancer; CLL, chronicg factor 1 receptor; FLT3, FMS-like tyrosine kinase 3; RET, rearrangedr additional abbreviations.

116 H. Cheng, T. Force / Progress in Cardiovascular Diseases 53 (2010) 114–120

when it is in the inactive conformation. Therefore, theseagents possess enhanced selectivity and are typically(though not always) more potent. Type III inhibitors (eg,the families of mitogen-activated protein/extracellularsignal-regulated kinase kinase inhibitors includingPD98059 and UO126) bind to regions remote from theATP pocket. These regions are typically not highlyconserved, accounting for the excellent selectivity of theaforementioned mitogen-activated protein/extracellularsignal-regulated kinase kinase inhibitors.10 Althoughtype III agents are more selective, they are a smallminority of agents in development because they are moredifficult to design and not as predictably effective. Withthat said, there is intense interest in these types ofcompounds, particularly for the treatment of imatinib-resistant CML because the mutations leading to resistanceare, to date, all in the kinase domain, and therefore, themutated kinases should still be inhibited by type III agents.

KIs: a concern for the heart

Against the successes in cancer with the small-molecule inhibitors and the belief that these targetedtherapeutics would be far less toxic than traditionalchemotherapy, it was something of a surprise whencardiotoxicity was detected. The first report of cardio-toxicity with a small-molecule KI was a case series of10 patients who developed congestive heart failure whilereceiving imatinib.11 Subsequently, much more serioustoxicity was identified in the first study specificallyfocused on cardiotoxicity of a small-molecule KI.12 Inthis study, serial evaluations of left ventricle (LV)ejection fraction and biomarker determinations (troponinI) were performed in patients with gastrointestinalstromal tumors (GISTs) receiving sunitinib. Eighteenpercent of patients fully developed either congestiveheart failure or a decline in LV ejection fraction of morethan 15 percentage points. In addition, cardiotoxicitywith sorafenib treatment has also been reported thoughoverall risk is unclear.13 It is critical to note, however,that cardiotoxicity is not a class effect of KIs because therisk of significant cardiotoxicity seems to be low for mostof the approved agents. Only those that target essentialkinases expressed in the heart and vasculature will likelyhave associated cardiotoxicity.

Currently, small-molecule KIs account for ≈20% of allmoney spent in drug development and the majority of that(≈80%) is in cancer (with a small percentage ininflammatory and other diseases). There seems to be noslowing down in this important and increasingly lucrativearea, and we will likely see many more of these agents onthe market over the next several years. However, otherthan screening for QT prolongation, it is quite rare forthese agents to be screened for cardiotoxicity, and very

few prospective studies have been performed to carefullyexamine the issue. Thus, in most cases, the practicingcardiologist and oncologist will not know whether a newagent will be problematic. Furthermore, long-term follow-up of patients is not available because of the recentintroduction of these drugs. As opposed to traditionalchemotherapeutics, KIs are often taken for life becausewithdrawal can lead to reemergence of the malignancy.Thus, long-term follow-up of patients will be particularlyimportant. Furthermore, it must be realized that the typicalpreapproval trial enrolls the healthiest of the cancerpatients, and patients with comorbidities, particularlycardiovascular comorbidities, are usually excluded.Thus, confirmation of lack of toxicity must await approvalwhen the agent will be used in a broader population ofcancer patients with cardiovascular comorbidities. Thissupports the need for patient registries, which allowcaregivers and researchers to identify critical problemspost-FDA approval.

Molecular mechanisms of cardiotoxicity

In 2002, Hoshijima and Chien14 drew largely theoret-ical parallels between the dysregulation of the signalingpathways driving cancer and those driving cardiachypertrophy. Indeed, there are numerous parallels betweensignaling pathways that drive tumorigenesis and those thatregulate not only hypertrophy but also survival ofcardiomyocytes, especially in the stressed heart. Hence,inhibition of the “key” kinases that drive tumorigenesiscould potentially compromise the survival of cardiomyo-cytes. Indeed, this seems to be at the core of thecardiotoxicity of KIs. The 2 types of toxicity will bediscussed here to elucidate the underlying molecularmechanisms of KI-derived cardiotoxicity.

On-target toxicity

With on-target toxicity (also known as mechanism-based or target-related), the kinase that is targeted in thecancer also provides an important maintenance function inthe heart and vasculature (Cheng and Force7 andreferences therein). Thus, inhibiting it leads to adverseconsequences in the heart. A classic example is the cardiactoxicity caused by trastuzumab (Herceptin), a monoclonalantibody against the ERBB2 receptor (also known asHER2). The HER2 receptor is overexpressed in approx-imately 20% of breast cancers and is critical for drivingprogression. Inhibiting HER2 with trastuzumab provides asurvival advantage in patients, but LV dysfunction canoccur. Although the precise mechanism of trastuzumabcardiotoxicity is still being debated, it is apparent thatHER2 plays a critical role in cardiomyocyte proliferation(during development) and survival (during adulthood).In addition, HER2 protects against anthracycline

117H. Cheng, T. Force / Progress in Cardiovascular Diseases 53 (2010) 114–120

cardiotoxicity,15 probably accounting for the markedincrease in LV dysfunction seen when the two wereadministered together. Trastuzumab-related cardiotoxicitywill be discussed by Suter et al in this issue. Imatinib-induced cardiotoxicity is another example of on-targettoxicity, mechanisms of which have been extensivelyreviewed recently.7

Off-target toxicity

With off-target toxicity, a kinase that was not intendedto be inhibited by a drug is inhibited, and if this kinase playsa key role in the heart, this inhibition will lead tocardiotoxicity. Off-target toxicity is inherently related tothe limited selectivity of most KIs (especially type I

Fig 1. Adverse effects of sunitinib on energy-responsive signaling pathways ininduces loss of mitochondrial membrane potential and energy stress (increase inphosphorylation of T172 of AMPK, activates AMPK. This produces a number ophosphofructokinase (PFK), leading to decreased fatty acid synthesis, increasmTORC1 signaling, leading to increased eEF2 phosphorylation (mediated by etranslation (a major energy-consuming process in cardiomyocytes) and proteinHowever, in the presence of sunitinib, ATP cannot bind to AMPK, and thereforeenergy-conserving mechanisms are not recruited and energy depletion is exacinhibition of receptor tyrosine kinase (RTK) signaling, can lead to inhibitionincomplete and energy rundown continues.

inhibitors) and the size of the human kinome. For example,sunitinib, which was initially developed primarily to inhibitvascular endothelial growth factor receptors (VEGFRs) 1-3, platelet-derived growth factor receptors (PDGFRs) α/β,and c-Kit, has been predicted to inhibit at least 50 kinases.16

In cultured cardiomyocytes, sunitinib induced loss ofmitochondrial membrane potential and energy rundown.17

Typically, when energy stores drop in the cardiomyocyte, akinase called AMP-activated protein kinase (AMPK) isactivated, leading to increased energy generation anddecreased energy utilization. However, AMPK was notactivated in the energy-stressed cardiomyocytes. In fact,AMPK activity was reduced in hearts of sunitinib-treatedmice and cardiomyocytes in culture, and this was due topotent and direct inhibition of AMPK by sunitinib. Thus,

the heart. In cultured cardiomyocytes, sunitinib, via unclear mechanisms,AMP/ATP ratio), which, together with CaMKK- and/or LKB1-mediatedf relatively rapid responses including (1) phosphorylation of ACC1/2 anded fatty acid oxidation, and increased glycolysis, and (2) inhibition ofEF2Kinase) and inhibition of eEF2.18 This leads to inhibition of proteinsynthesis. Together, these responses help to restore energy homeostasis., AMPK cannot transfer phosphate from ATP to the substrates. Thus, theerbated. Although multiple other AMPK-independent inputs, includingof mTORC1, inhibition of protein translation via these mechanisms is

118 H. Cheng, T. Force / Progress in Cardiovascular Diseases 53 (2010) 114–120

the findings suggested that off-target inhibition by sunitinibof AMPK, a kinase that plays key roles in maintainingmetabolic homeostasis in the heart, especially in the settingof energy stress, accounts, at least in part, for sunitinib-induced toxicity seen in cardiomyocytes. This representedthe first example of off-target inhibition of a kinase by a KIleading to cardiotoxicity (Fig 1).17

Selectivity versus multitargeting

One obvious question to address is: why not make moreselective inhibitors, thereby reducing off-target effects? Onereason, noted above, is that ATP competitive inhibitors arerelatively easy to make and more predictably effective. Theother reason is more complex but important to understand.Tumor growth in most cases is driven by mutations in morethan one kinase. The clearest example of this is the necessityof tumor neoangiogenesis for a tumor to grow beyond acertain size. Thus, one clear direction in drug design is“multi-targeting,” in which a single drug targets both VEGFreceptors and kinases that specifically drive tumor growth.Examples of this trend are sunitinib and sorafenib, which areapproved for several solid tumors and are in development formany more, and although multitargeting can enhance tumorcell killing, the flip side is that this leads to inherentnonselectivity and potential increased risk of cardiacdysfunction.One final note onmechanisms of cardiotoxicityof VEGFR inhibitors is that studies in mouse models havesuggested that angiogenesis is key to maintaining cardiachomeostasis in response to pressure overload,19,20 which,taken together with the significant hypertension induced byVEGFR inhibitors, might explain, in part, the LVdysfunction seen with sunitinib. More recently, PDGFRβ,another target of sunitinib, was also found to be critical forangiogenesis, and deletion of the gene in mice led to heartfailurewhen themicewere exposed to high pressure loads.21

A variant of multitargeting is using multiple drugs totarget multiple components of a single pathway. Thisstrategy is gaining popularity particularly for targeting thegrowth factor receptor/PI3K/Akt pathway. Components ofthis pathway are mutated or overexpressed in a host ofcancers, and typically, multiple components are mutated ineach tumor, making this pathway an ideal target in cancertherapy.22 Furthermore, Yuan and Cantley23 proposed thatredundant activation of the PI3K pathway by multiplemutations or amplifications of pathway components,combined with activation of multiple nonoverlappingpathways will, with the exception of the rare oncogene-addicted cancer, require combination therapy. The caution-ary note here is that this pathway is critical to cardiomyocytesurvival in the setting of stress, in insulin sensitivity, and inphysiologic growth (driven by p110α).24,25 Inhibition of oneor two components of the pathwaymight allow for sufficientresidual signaling down the pathway to preserve cardio-myocyte integrity, yet inhibiting multiple components could

shut down signaling more completely, jeopardizing thecardiomyocyte. Again, hypertension might be a significantexacerbating factor for cardiac dysfunction when compo-nents of the PI3K pathway are targeted.25

Heart failure with KIs: cardiomyocyte lossand/or dysfunction?

Although it has been possible to implicate criticalkinases in KI-induced cardiotoxicity, it remains unclearwhether LV dysfunction with KIs is attributable tomyocyte loss (and therefore largely irreversible) ormyocyte dysfunction (potentially reversible). Neuronsand cardiomyocytes seem to be quite resistant to apoptosisinduced by cytochrome c release and caspase activation.Contributing to this is the decreased expression of Apaf1that is directly linked to the tight regulation of caspaseactivation by endogenous XIAP.26 Consistent with this,we did not see an increase in apoptosis in mice treatedwith sunitinib until we exposed the mice to phenyleph-rine-induced pressure load, and even then the increase incell death was modest.12 In contrast, we saw clearevidence of opening of the mitochondrial permeabilitytransition pore in the mice as evidenced by markedmitochondrial swelling and destruction of the normalmitochondrial architecture.12 Strikingly, we saw a verysimilar picture in transmission electron micrographs of anendomyocardial biopsy obtained from a patient whopresented with advanced heart failure while receivingsunitinib.12,17 This picture could be consistent with LVdysfunction being secondary to impaired energy genera-tion, with or without cell death by necrosis. Furthermore,sunitinib-induced LV dysfunction can normalize afterwithdrawal of the drug and/or institution of standard heartfailure therapy,12 but this is not a universal response.27

However, whether any reversibility of LV dysfunction isaccompanied by reversibility of injuries at the myocytelevel, and whether it will be long-lasting, remains unclear.That said, in the patient noted above, there was verystriking reversibility at the myocyte level with markedrestoration of mitochondrial number and integrity afterwithdrawal of sunitinib.12,17

Three drug development approaches to dealingwith toxicity

(1) Obtain the complete kinase selectivity profile of adrug and use that to predict cardiotoxicity

Given the fact that a nonselective agent may have betteranticancer activity and can be used inmore cancers, it seemssafe to say that for the near future, we will be forced toidentify mechanisms of cardiotoxicity of relatively nonse-lective drugs. This will make it essential to know the full-

119H. Cheng, T. Force / Progress in Cardiovascular Diseases 53 (2010) 114–120

selectivity profile of any agent against the entire kinome tobe able to define mechanisms of toxicity. At present,available commercial platforms for kinase selectivityprofiling typically cover approximately half of thekinome (250-300 kinases; eg, www.millipore.com, www.proqinase.com). KINOMEscan (a division of AmbitBiosciences, http://www.ambitbio.com/technology/) hasrecently developed a high-throughput, active site-dependentcompetition-binding assay platform that provides scientistswith access to a broader panel of 442 kinase assays. Thus,with the advancement of profiling technologies, lack of afull-selectivity profile is a deficit that should be correctedshortly. Based on the full kinase selectivity profile of a KI,we will, in some cases, be able to predict toxicities of drugs.Added to this, many kinases are simply not expressed in theheart, and although they may lead to other organ toxicities,the heart should be spared.

The glaring limitation of this “best guess” approach isthat we simply do not know the roles played by manykinases that are expressed in the heart and may be targetsof KIs. Although prior studies in various models haveidentified key roles that some kinases play in the heart andvasculature,7 much more work needs to be done tounderstand the roles of those novel kinase targets.Predicting problematic agents based on studies withgene-targeted mice can give valuable information, espe-cially when combined with pathway analysis programssuch as Jubilant (http://www.jubilantbiosys.com), ToxWiz(http://www.camcellnet.com), and Ingenuity (http://www.ingenuity.com). As a caveat, because drugs never inhibittheir targets 100%, knockouts would likely have moresevere phenotypes than any drug targeting the kinase ofinterest. With those caveats in mind, we refer the reader torecent reviews examining how mouse models can predictcardiotoxicity.7,28

(2) Redesign a drug to avoid the cardiotoxicity

If one can identify the mechanisms of cardiotoxicity ofa specific KI, the compound could be redesigned to avoidthe target. For example, Kerkela et al11 determined thatimatinib-mediated cardiotoxicity was due to the inhibitionof c-Abl in the cardiomyocyte, resulting in endoplasmicreticulum stress and cell death. Motivated by this finding,Fernandez et al29 and Demetri30 redesigned imatinib to nolonger inhibit Abl (but still inhibited c-Kit), andcardiotoxicity was significantly reduced. Although theredesigned drug was obviously ineffective in treatingCML (driven by Bcr-Abl), it was equally effective toimatinib in treating GIST models driven by c-Kitmutations. Thus, by knowing the mechanism of toxicityand redesigning the drug accordingly, one could theoret-ically reduce cardiotoxicity. Other examples of thisstrategy can be found in recent reports by Crespo et al31

and Fernandez et al.32 If proved to be effective, this

redesign strategy would potentially curb cardiotoxicitiescaused by multitargeted drugs and enhance their safety.

(3) Targeted delivery of drugs specifically to cancer cells

Compared with the traditional drug delivery systems,targeted delivery seeks to concentrate the agent in thetissue of interest (eg, malignant tumors), while reducingthe relative concentration of the drug in other tissues. Thisimproves drug efficacy while minimizing the adverseeffects and thus has been proposed to deal with theunavoidable on-target toxicities of cancer drugs. Duringthe last two decades, nanotechnology strategies havedeveloped a number of targeted nanoparticle deliverysystems for cancer therapy, including those against themost difficult challenges such of drug resistance andmetastasis.33 Several liposomal, polymer-drug conjugates,and micellar formulations are in the clinic, and an evengreater number of nanoparticle platforms are in preclinicaldevelopment. In a recent study, nanoparticle-mediateddelivery of a selective mitogen-activated protein kinaseinhibitor was shown to optimize cancer chemotherapy.34

Interestingly, nanoparticles loaded with a PI3K inhibitordemonstrated efficacy in inhibiting tumor angiogenesis ina zebrafish tumor model.35 Thus, targeted delivery of KIswith nanoparticles offers an attractive strategy to tackle theissue of cardiotoxicity.

Concluding remarks

Currently, preclinical testing is unable to accuratelyidentify agents that will have associated cardiotoxicity.Full-selectivity profiles of agents, coupled with a betterunderstanding of the role played by kinases in thecardiomyocyte, and better preclinical models to detectcardiotoxicity are needed. It is critical for basic cardio-vascular researchers and clinicians to understand theseagents, their mechanisms of action (including mechanismsby which they kill cancer cells), and how these actions canlead to cardiotoxicity. We believe that with (1) greaterawareness of the problem, (2) advances in drug design anddelivery, (3) better preclinical screening approaches, and(4) close cooperation among clinical and basic cardiolo-gists, oncologists, and the pharmaceutical industry in theearly phases of drug development, cardiotoxicity of cancertherapeutics will become a very manageable problem.

Statement of Conflict of Interest

All authors declare that there are no conflicts of interest.

References

1. Krause DS, Van Etten RA: Tyrosine kinases as targets for cancertherapy. N Engl J Med 2005;353:172-187.

120 H. Cheng, T. Force / Progress in Cardiovascular Diseases 53 (2010) 114–120

2. Manning G, Whyte DB, Martinez R, et al: The protein kinasecomplement of the human genome. Science 2002;298:1912-1934.

3. Haber DA, Settleman J: Cancer: drivers and passengers. Nature2007;446:145-146.

4. Luo J, Solimini NL, Elledge SJ: Principles of cancer therapy:oncogene and non-oncogene addiction. Cell 2009;136:823-837.

5. Bantscheff M, Eberhard D, Abraham Y, et al: Quantitative chemicalproteomics reveals mechanisms of action of clinical ABL kinaseinhibitors. Nat Biotechnol 2007;25:1035-1044.

6. Quintas-Cardama A, Kantarjian H, Cortes J: Flying under the radar:the new wave of BCR-ABL inhibitors. Nat Rev Drug Discov 2007;6:834-848.

7. Cheng H, Force T: Molecular mechanisms of cardiovascular toxicityof targeted cancer therapeutics. Circ Res 2010;106:21-34.

8. Sherbenou DW, Druker BJ: Applying the discovery of thePhiladelphia chromosome. J Clin Invest 2007;117:2067-2074.

9. Okram B, Nagle A, Adrian FJ, et al: A general strategy for creating“inactive-conformation” abl inhibitors. Chem Biol 2006;13:779-786.

10. Ohren JF, Chen H, Pavlovsky A, et al: Structures of human MAPkinase kinase 1 (MEK1) and MEK2 describe novel noncompetitivekinase inhibition. Nat Struct Mol Biol 2004;11:1192-1197.

11. Kerkela R, Grazette L, Yacobi R, et al: Cardiotoxicity of the cancertherapeutic agent imatinib mesylate. Nat Med 2006;12:908-916.

12. Chu TF, Rupnick MA, Kerkela R, et al: Cardiotoxicity associatedwith tyrosine kinase inhibitor sunitinib. Lancet 2007;370:2011-2019.

13. Schmidinger M, Zielinski CC, Vogl UM, et al: Cardiac toxicity ofsunitinib and sorafenib in patients with metastatic renal cellcarcinoma. J Clin Oncol 2008;26:5204-5212.

14. Hoshijima M, Chien KR: Mixed signals in heart failure: cancer rules.J Clin Invest 2002;109:849-855.

15. Crone SA, Zhao YY, Fan L, et al: ErbB2 is essential in the preventionof dilated cardiomyopathy. Nat Med 2002;8:459-465.

16. Ghoreschi K, Laurence A, O'Shea JJ: Selectivity and therapeuticinhibition of kinases: to be or not to be? Nat Immunol 2009;10:356-360.

17. Kerkela R, Woulfe KC, Durand JB, et al: Sunitinib-inducedcardiotoxicity is mediated by off-target inhibition of AMP-activatedprotein kinase. Clin Transl Sci 2009;2:15-25.

18. Ma XM, Blenis J: Molecular mechanisms of mTOR-mediatedtranslational control. Nat Rev Mol Cell Biol 2009;10:307-318.

19. Izumiya Y, Shiojima I, Sato K, et al: Vascular endothelial growthfactor blockade promotes the transition from compensatory cardiac

hypertrophy to failure in response to pressure overload. Hypertension2006;47:887-893.

20. Sano M, Minamino T, Toko H, et al: p53-induced inhibition of Hif-1causes cardiac dysfunction during pressure overload. Nature 2007;446:444-448.

21. Khakoo AY: Regulation of hypertrophic responses by PDGFRβ.American Heart Association, Basic Cardiovascular Sciences annualmeeting, Lake Las Vegas, NV; 2009 July 21st; 2009.

22. Garcia-Echeverria C, SellersWR:Drug discovery approaches targetingthe PI3K/Akt pathway in cancer. Oncogene 2008;27:5511-5526.

23. Yuan TL, Cantley LC: PI3K pathway alterations in cancer: variationson a theme. Oncogene 2008;27:5497-5510.

24. Muslin AJ, DeBosch B: Role of Akt in cardiac growth andmetabolism. Novartis Found Symp 2006;274:118-126 discussion26-31, 52-55, 272-276.

25. McMullen JR, Jay PY: PI3K(p110alpha) inhibitors as anti-canceragents: minding the heart. Cell Cycle 2007;6:910-913.

26. Potts MB, Vaugha AE, McDonough H, et al: Reduced Apaf-1 levelsin cardiomyocytes engage a strict regulation of apoptosis byendogenous XIAP. J Cell Biol 2005;171:925-930.

27. Khakoo AY, Kassiotis CM, Tannir N, et al: Heart failure associatedwith sunitinib malate: a multitargeted receptor tyrosine kinaseinhibitor. Cancer 2008;112:2500-2508.

28. Chen MH, Kerkela R, Force T: Mechanisms of cardiac dysfunctionassociated with tyrosine kinase inhibitor cancer therapeutics.Circulation 2008;118:84-95.

29. Fernandez A, Sanguino A, Peng Z, et al: An anticancer C-Kit kinaseinhibitor is reengineered to make it more active and less cardiotoxic.J Clin Invest 2007;117:4044-4054.

30. Demetri GD: Structural reengineering of imatinib to decrease cardiacrisk in cancer therapy. J Clin Invest 2007;117:3650-3653.

31. Crespo A, Zhang X, Fernandez A: Redesigning kinase inhibitors toenhance specificity. J Med Chem 2008;51:4890-4898.

32. Fernandez A, Crespo A, Tiwari A: Is there a case for selectivelypromiscuous anticancer drugs? Drug Discov Today 2009;14:1-5.

33. Alexis F, Pridgen EM, Langer R, et al: Nanoparticle technologies forcancer therapy. Handb Exp Pharmacol 2010:55-86.

34. Basu S, Harfouche R, Soni S, et al: Nanoparticle-mediated targetingof MAPK signaling predisposes tumor to chemotherapy. Proc NatlAcad Sci U S A 2009;106:7957-7961.

35. Harfouche R, Basu S, Soni S, et al: Nanoparticle-mediated targetingof phosphatidylinositol-3-kinase signaling inhibits angiogenesis.Angiogenesis 2009;12:325-338.