*2.2. Phase II Enzymes*

Phase II enzymes are non-CYP450 proteins that can indirectly exert influence on CYP450 enzyme activity. The most commonly occurring phase II reactions are glucoronidation and sulfonation (or sulfurylation), which are catalyzed by the enzymes uridine diphosphate glucoronysltransferase (UGTs) and sulfotransferases (SULTs), respectively. These enzymes are located throughout the gastrointestinal and genitourinary tracts.

The UGTs catalyze transfer of glucuronic acid to onto oxygen, nitrogen or sulfur on substrate drugs. Substrates range from endogenous substances (e.g., bilirubin, estradiol, serotonin) to exogenous substances (e.g., propofol, morphine). Interindividual variability in levels of UGT, stemming from a patient's age, sex, presence of enzyme inhibitors/inducers can contribute to drug-induced toxicity (slow metabolism) or ineffectual drug levels (rapid metabolism) [20].

Toxic drug metabolites of UGTs levels can lead to additive toxicity with P450 (phase I) drug-drug interactions. For example, the glucoronated products of gemfibrozil can inactivate CYP2C8, causing toxic levels of statins and significant rhabdomyolysis [21]. Likewise, clopidogrel can also inactivate CYP2C8, resulting in toxic levels of gemfibrozil [22]. Finally, UGT1 levels have been shown to inversely associate with development of a number of cancers (i.e., colon cancer, breast, bladder and biliary) in conditional UGT1 knockout mice [23].

Sulfonation reactions result in increased hydrophilicity and (usually) decreased pharmacological activity or inactivation of certain endogenous substances, such as thyroid hormones, steroids and monoamine transmitters. Inhibition of sulfonation by some compounds or metabolites can increase the toxicity of these substances [24,25]. Conversely, sulfonation can also bioactivate some substrates. This can result in a benign, more metabolically active form (e.g., minoxidil, morphine) [25] or can produce certain toxic metabolites, thereby increasing drug toxicity (e.g., tamoxifen) [24,26]. There are three main SULTs supergene families in humans—SULT1, SULT2, SULT4 [24–26]. SULT1A is most concentrated in the liver and has also been found in the kidney, lung, brain and gastrointestinal and genitourinary systems. The extensive expression of SULT1A1 and SULT1A3 in the intestines and lungs suggest they may play a role in extrahepatic drug detoxification and metabolism. SULT1B functions in thyroid hormone metabolism. SULT2A and SULT2B family are active in the metabolism of steroids and bile acids and are present throughout the body. Notably, products of sulfonation reactions catalyzed by various SULT enzymes (e.g., 1A1, 1A2, 1A3, 1C2, 1C4 and 2A1) can result in chemically reactive intermediate compounds that bind DNA, eliciting mutagenicity and carcinogenicity [25]. Interindividual variation in human sulfotransferase activity varies from 5- to 36-fold, largely explained by single nucleotide polymorphisms (SNPs) in the coding regions of SULT genes. This variation can play a complementary role to phase I reactions (largely catalyzed by P450 enzymes) in determining an individual's response to therapeutics.

#### *2.3. Substrates, Inducers and Inhibitors*

Drugs that interact with the CYP450 enzymes can be divided into three categories: substrates, inhibitors and inducers. Substrates are drugs upon which specific CYP450 enzyme acts. Inducers are drugs that increase enzyme activity. Inhibitors are drugs that decrease enzyme activity. Inhibitors compete with other drugs (typically substrates) for enzyme active sites, therefore altering the optimal level of a given substrate drug in the plasma. This alters the intended drug pharmacokinetics, rendering many prodrugs ineffective or conversely, potentially raising other drugs' plasma concentrations to toxic levels. A strong inhibitor is defined as one that increases plasma AUC substrate values greater than 5-fold or decreases substrate clearance to more than 80% of normal levels. A moderate inhibitor causes a greater than 2-fold increase in the plasma AUC values or a 50–80% decrease in drug clearance. A weak inhibitor causes a greater than 1.25-fold increase in plasma AUC values or a 20–50% decrease in drug clearance. Drugs commonly used in cardiology fall into all three categories (substrates, inducers and inhibitors) (Table 2) and can potentially interact with oncology drug substrates; the converse is also true. Being mindful of drug-drug interactions due to CYP450 activity related to substrates, inducers and inhibitors may help protect the hearts of patients undergoing cancer therapies.






**Table 2.** *Cont.* Note: for medications not categorized as strong, moderate or weak inducers/inhibitors, there is insu fficient evidence to further categorize them. Medications denoted with an asterisk (\*) are not available in the US. Enzymes in bold denote the most commonly occurring CYP450 enzymes. Adapted from various sources [1,2,27–41]; used with permisson of the three primary sources [1,2,27].

#### **3. Drug-Drug Interactions**

Drug-drug interactions are fairly common in the oncologic patient with cardiac disease (Tables 2 and 3). In one study [42], 16% of patients receiving oral antineoplastic agents developed at least one major drug-drug interaction. This is of particular concern, given the narrow therapeutic window of many antineoplastic agents and some cardiology medications. In another study, a range of drug-drug interactions involving chemotherapeutic and common cardiac medications resulting from either pharmacokinetic (PK) interactions, pharmacodynamic (PD) interactions or a combination of the two was described [3,43]. The most common PK interactions in oncology involve the CYP450 enzymes and the efflux pump P-glycoprotein located in the intestine [43,44]. In essence, PK interactions describe the body's effect on a drug or substance, especially its absorption, distribution, metabolism or elimination [44,45]. On the other hand, PD interactions describe a drug's effect on the body. Drug-drug interactions in this arena are due to often unintentional additive effects of two agents with similar molecular targets, resulting in toxicity.

An example of a PD interaction in Cardio-Oncology occurs with concurrent use of beta-blockers (in Cardiology) and ceritinib/crizotinib (in Oncology); the latter is a combination chemotherapeutic drug used to treat metastatic (ALK-/ROS1-positive) non-small cell lung cancer. Co-administration of these medications can lead to symptomatic bradycardia, which can potentially be life-threatening [43]. Additionally, ceritinib/crizotinib can prolong the QT interval. Therefore, administration with other QT-prolonging medications that are often administered with chemotherapy, such as antiemetics, antibiotics and antidepressants, can potentially lead to malignant arrhythmias, including polymorphic ventricular tachycardia or 'torsades de pointes.' [43,46]. Consequently, beta-blockers or QT-prolonging medications should be used judiciously with ceritinib/crizotinib if co-administered with any drugs that inhibit CYP3A, as both ceritinib and crizotinib are extensively metabolized by CYP3A in the liver [47,48].

An example of a PK interaction in Cardio-Oncology involves the moderate inhibition of CYP3A4 by diltiazem/verapamil. When co-administered with chemotherapeutic agents metabolized by the same pathway, such as doxorubicin, imatinib or ibrutinib, this could lead to increased chemotherapy drug concentration. This can be accompanied by several adverse effects, including QT prolongation, gastrointestinal symptoms, shortness of breath, edema, chest pain, hepatotoxicity or bone marrow suppression [43]. This can be managed by using alternative medications for chronotropy or blood pressure control during the expected course of chemotherapy or appropriately decreasing the dose of administered chemotherapy if absolutely necessary [43,46].

Antiplatelet agents are a mainstay of atherosclerotic cardiovascular disease treatment and account for 40.4% of drug sales in cardiovascular disease [49]. PK interactions between antiplatelet agents and chemotherapeutics can alter the level of functioning of one or both drugs. For example, the chemotherapeutic combination agent enzalutamide/dasatinib can decrease the level of antiplatelet medication in the blood, causing disastrous/catastrophic consequences following cardiac catheterization [43]. When doxorubicin is administered with ticagrelor, CYP3A4 inhibition by ticagrelor can lead to an increase in doxorubicin exposure, placing patients at increased risk for known toxicities of the drug [43]. Based on the indicated chemotherapy regimen, the antiplatelet agent can usually be adjusted, taking into account patient characteristics.

Anticoagulants are indicated for patients with malignancies who develop deep venous thromboses (DVT) or pulmonary emboli (PE), which complicate the clinical course of approximately 5–10% of all cancer patients [50]. In patients with malignancies who also have cardiovascular disease requiring anticoagulation (e.g., atrial fibrillation, mechanical valves, mechanical support devices), the number of concurrent medications can increase the risk of drug-drug interaction. Historically, warfarin has been the most commonly used anticoagulant and remains in frequent use due to familiarity, cost and patient preference. Most drug-drug interactions involving chemotherapeutic agents and warfarin are due to a reduction in warfarin metabolism, often from CYP450 inhibition, leading to increased risk of bleeding (Table 2) [43]. As warfarin operates via vitamin K inhibition, oncologic patients with numerous reasons for vitamin K deficiency, such as diarrhea from chemotherapy or radiation or antibiotics for infections due to immunosuppression, face an additionally increased risk of bleeding [43]. The current standard of care for management of cancer-associated venous thromboembolism (VTE) is low molecular weight heparin [51–55]. Direct oral anticoagulants (DOACs) are emerging as potentially equally efficacious alternatives to low molecular weight heparin, with ideal bioavailability and mode of administration (orally) [29]. However, DOACs should be used cautiously in certain populations at increased risk for bleeding complications (e.g., gastrointestinal malignancies, advanced age and frailty) [43,50]. It is of note that DOACs are variably metabolized by CYP450 enzymes (dabigatran 0%, edoxaban < 4%, apixaban 15%, rivaroxaban 66%) [29]. Therefore, in cancer patients concurrently treated with strong CYP3A4 inducers or inhibitors (Table 2), dabigatran (or possibly edoxaban) may become the DOAC of choice; such decision-making may benefit from close collaboration with a clinical pharmacist [29,56]. Nevertheless, both LMWH and DOACs have fewer drug-drug interactions than warfarin [43,56–58].

#### **4. Precision Cardio-Oncology**

#### *4.1. Variability in Concentration and Activity*

There can be wide variation in the concentration levels and activity of CYP450 between and within populations, as illustrated in the following examples. In the general population, there is up to 50-fold variation in the levels of CYP3A4 among individuals [15,59]. In a small study investigating the metabolism of cyclophosphamide in patients with lung, breast and gastrointestinal malignancies, the level of CYP2C19 was lower in patients with cancer compared to the general population [60]. Conversely, in a study investigating the role of miRNA in the regulation of CYP1B1, higher levels of CYP1B1 were noted particularly among individuals with estrogen-sensitive cancers [60,61]. Indeed, some CYP450 enzymes are preferentially upregulated by cancerous cells. For example, CYP1B1, mainly expressed in the ovary, uterus and breast tissue [61,62], is upregulated in malignant cells [63] to catalyze the metabolic activation of pro-carcinogens such as polycyclic aromatic hydrocarbons, aryl nitrate and 4-hydroxyestradiol [61]; 4-hydroxyestradiol is a catechol metabolite of 17β-estradiol, which generates free radicals, resulting in DNA damage [64,65]. Additionally, among patients with cancer, there can be wide variation in CYP450 enzyme activity.

#### *4.2. Interindividual and Genetic Variability*

Genetic variability, for example, of promoter or coding regions, may in part explain why different individuals have varying responses to the same drugs. Single nucleotide polymorphisms among CYP450 enzymes affect metabolism and therefore bioavailability of substrate drugs. Polymorphisms in CYP1A2 and CYP2B6 can result in decreased nicotine metabolism in smokers and have been associated with increased susceptibility to cancers and possibly atherosclerotic cardiovascular disease [1]. In particular, different combinations of alleles of CYP450 enzymes resulting in absent, low or high levels of enzymatic activity can lead to differential responses (e.g., toxicity, underdosing) to drug regimens, due to variable drug clearance [1,42,43].

#### *4.3. Genomic Profiling*

Genomic profiling in general characterizes an individual's complement of genes [66]. Next generation sequencing and genome-wide association studies (GWAS), which correlate SNPs with disease phenotypes, have revolutionized the speed at which such information can be analyzed, investigated and translated into a component of clinical care [67]. Genomic profiling is most commonly utilized in oncology—particularly for breast, ovarian, colon and lung cancers [67]. The applicability in cardiology is increasing, from medication metabolism to treatment of cardiomyopathy and inherited arrhythmias [67]. Genomic profiling has a number of potential applications, including the study of genetic variations that influence individual response to drugs (i.e., pharmacogenomics), precision medicine and new modalities to diagnose and treat disease.
