*4.4. Systems Approach*

While genetic variation plays a distinct role, this does not completely account for interindividual differences in CYP450 enzyme concentration level and activity [7,16]. The contribution of variation in transcriptional regulators and (more recently) posttranscriptional protein modifications (especially alterations in microRNA) affecting CYP450 expression are gaining increasing recognition [15]. Further, ontogenic changes to CYP450 activity, due to xenobiotic exposure (e.g., phenobarbital) during early development and the early postnatal period, have been shown to have long-lasting effects in mouse models. Importantly, efforts to achieve both early diagnosis and optimal treatment of disease as well as prevent and mitigate cardiac adverse effects in oncology patients has led to study and potential application of multi-omic disciplines to the field of Cardio-Oncology [7,77]. Genomics, epigenomics, transcriptomics, proteomics, miRNAomics, metabolomics and microbiomics have the potential to more precisely guide the clinical management of these patients [7,78]. The integration of multi-omics with systems biology and incorporating disciplines such as mobile health (mHealth), pharmacogenomics, mathematical and computational modeling is depicted in Figures 2 and 3 and has been well described [7,77,78].

**Figure 2.** Multi-level systems-based approach to CYP450 expression, activity and regulation in Precision Cardio-Oncology. These multi-omics and additionally epigenomics, metabolomics, microbiomics and other personalization tools will likely be integrated in the future with mobile health, informatics and other emerging technologies for precision patient care relevant to Cardiology, Oncology and Cardio-Oncology.

**Figure 3.** The P\*3 Precision Medicine approach to individualizing therapy involving P450 enzymes. P1, Pre-empt: Pre-emption encompasses incorporation of elements of systems-based medicine, such as patients' at-risk genetic polymorphisms, into the EHR seamlessly alongside clinical data. P2, Predict: Prediction describes the process of utilizing systems medicine data to estimate patients' risk of developing cardiotoxicity. P3, Prevent: Prevention is proactively adjusting patients' treatment plans based on their cardiovascular toxicity risk to prevent de novo damage or mitigate further damage from antineoplastic therapy. Adapted with permission [78]. CYP450 = cytochrome P450; EHR = electronic health record.

#### 4.4.1. Transcriptomics

Transcriptomics or gene expression profiling, is the study of the complement of genes expressed, that is, mRNA [79]. Transcribed gene expression profiles can exhibit wide variability, depending on the physiologic stimuli. Genetic variation in promotor regions and transcriptional regulators of P450 enzymes can result in altered rates of gene expression and in turn, drug metabolism [16,80]. Approximately 10% of variation in target P450 enzyme expression is explained by shared hepatic P450s transcriptional regulators (e.g., hepatocyte nuclear factor (HNF) 4β, forkhead box proteins (FOX) A2 and A3 and pregnane X receptor (PXR) polymorphisms) [16,80].

#### 4.4.2. Epigenomics

Epigenomics is the study of physical modifications made to DNA molecules without alterations in DNA sequence. These modifications are most commonly but not limited to, acetylation and histone modification, methylation and transcription factor binding of DNA [7]. Both intrinsic and extrinsic factors can lead to epigenomic modifications of genes encoding CYP450 enzymes, thereby contributing to interindividual variability of therapeutic response [16,81]. Intrinsic factors include old age, congestive heart failure, cancer or pregnancy [7,82–84], while extrinsic factors include xenobiotics, tobacco smoke and air pollution [79]. Drugs can act as ligands to activate transcription factors, consequently inducing the expression of P450 genes. Phenobarbital and rifampin, for example, induce CYP2B6, CYP2C9 or CYP3A4 via activation of the nuclear receptors PXR or constitutive androstane receptor (CAR) [16]. It is hypothesized that epigenetic modifications in the promotor or coding region, in response to certain physiologic or pathophysiologic states, can either activate or silence certain CYP450 enzymes. For example, methylation of CpG motifs (CpG11, CpG12, CpG13) in the promoter region for CYP2C19 was associated with HTPR and increased risk of ischemic events on clopidogrel [7,85,86]. If an allele is preferentially silenced via epigenetic modifications in the coding region, the complementary allele will be predominantly expressed. Notably, DNA methylation and histone acetylation modification patterns have been found to be heritable in certain cases [79].

#### 4.4.3. Proteomics

Proteomics is the study of the complement of proteins and can also sometimes refer to chemical modifications (e.g., phosphorylation, acetylation, nitrosylation, etc.) undergone by proteins during and after translation, due to specific (patho)physiological states and their effects on protein structure and function [7,79]. These changes can occur hours following an exposure (e.g., drug administration). In one study, a decrease in peroxiredoxin-4 (a molecule associated with HTPR) was observed just 24 h following the administration of a loading dose of clopidogrel [7,87].

#### 4.4.4. Metabolomics

Proteins with enzymatic activity produce metabolites. Metabolomics is the study of all measurable metabolites resulting from biotransformation of compounds, often by CYP450 enzymes, corresponding to specific physiologic or pathophysiologic states [7,88]. Homeostasis of an individual's metabolome is impacted by both intrinsic (e.g., resting metabolic rate, age, genotype, etc.) and extrinsic (e.g., xenobiotics, microbiome, surrounding environment, etc.) factors [7,84,87,88]. In cardioncology patients, metabolic analysis can identify a specific metabolic pattern—a metabotype—corresponding to specific (patho)physiologic states (e.g., heart failure, myocardial infarction, myocarditis, etc.) and may serve as a 'yellow-brick road,' of sorts, offering insights into the present state and potentially, the future arc of disease [7,77]. Indeed, changes in a patient's metabolic profile over time may help identify predictive biomarkers associated with different clinical trajectories, while helping to identify an individual's response to therapy (or lack thereof), preceding phenotypic changes, helping to guide therapy [7,77].

#### 4.4.5. Microbiomics

Microbiomics studies the impact of the human microbiome on disease states. As 70% of the millions of bacteria inhabiting the human body reside in the gut, the metabolic products of these bacteria play a role in modulating the body's inflammatory response and impacting chronic disease [7]. For example, trimethylamine N-oxide (TMAO) is a product of dietary phosphatidylcholine and gut microbe metabolism of carnitine and betaine [7,88]. High levels of TMAO are independently associated with platelet hyperactivity and thrombosis of atherosclerotic plaques [7,89–91], increasing risk of cardiovascular events [7,89] and predicting all-cause mortality [7,91]. Additionally, aspirin has been observed to alter the composition of the gut microbiota. This likely accounts for its ability to abate the effect of TMAO on platelets [7,91,92].

#### 4.4.6. MicroRNAomics

The field of miRNAomics is quickly emerging and encompasses a diverse array of applications ranging from regulation of translation by noncoding miRNAs to small RNA therapies for challenging drug targets. Micro-RNAs (miRNAs) are a class of noncoding RNAs that play an important role in selective gene silencing, through which miRNAs regulate the diverse functions of target genes ranging from proliferation and apoptosis to differentiation. Specifically, mature miRNAs function by binding to complementary regions of transcripts (via the RNA induced silencing complex or RISC) to either (mechanically) prevent translation by the ribosome or cut target mRNAs into fragments, thereby precipitating their degradation inside the cell [15,61]. The former regulatory method of translational prevention is the predominant one in mammals (i.e., humans), while the latter method resulting in mRNA degradation is used in plants [15]. Over 1900 miRNAs have been identified in humans [15,93]. As a class, miRNAs are predicted to regulate up to 60% of the human genome [15,94,95], with miRNAs acting on one or multiple genes and single genes often targeted by one or more miRNAs [15,96].

The role of miRNAs in regulating CYP450 enzymes was first demonstrated nearly 15 years ago [61]. Since that time, the number of known miRNAs has nearly doubled and innovative techniques such as computer modeling (i.e., *in silico*) studies are being used to discover more miRNAs and determine their associated function at an exponential rate [15,80]. Data from such computer modeling, integrated with in vitro and in vivo techniques, revealed that translation of CYP3A4 mRNA is repressed by multiple miRNAs, including hsa-miR-577, hsa-miR-1, hsa-miR-532-3p and hsa-miR-627 [15]. Notably, another study found that specific miRNA profiles affected response to chemotherapy, with hsa-miR-577 and hsa-miR-1 significantly improving chemosensitivity following administration of chemotherapy for gastric cancer [15,97]. More studies are needed to confirm whether this increase in chemosensitivity is related to an inhibitory effect of hsa-miR-577 and hsa-miR-1 on CYP3A4 transcripts.

Various miRNAs are generally downregulated in various cancers, often to limit translation and promulgate the action of CYP450 enzymes promoting metabolic production of pro-carcinogens, facilitating cancer growth [61]. For example, expression of CYP1B1 is post-transcriptionally regulated by miR-27b. Studies have found that the downregulation [61] or actual deletion [98] of the 9q22.1 gene locus coding for miR-27b allows for the high expression of CYP1B1, which can facilitate tumor growth. While this inverse relationship of miR-27b downregulation and CYP1B1 upregulation expression mainly occurs in estrogen-sensitive malignancies, it is also found in urothelial or bladder cancers [61,98,99].

In addition, miR-34a has an inverse relationship with expression of CYP3A4 and CYP2C19 proteins, as well as transcription factors and other proteins that indirectly affect CYP450 levels or activity (e.g., AKR1D1 and SLC10A1) and is generally found at higher levels in females than males [80]. Females are also known to have higher CYP3A4 levels than males, suggesting miR-34a as a possible etiology for intersex differences in CYP3A4 activity [80]. The miR-34a is upregulated in several cancers, including hepatocellular cancer [80,100]. Given the inhibition of CYP3A4 and CYP2C19 by miR-34a, patients with high levels of miR-34a are potentially at an increased risk of drug toxicity. Conversely, miR148a increases levels of CYP3A4 and CYP2C19 and is downregulated in a number of cancers, including most of the lower gastrointestinal (GI) tumors, along with head & neck cancers, breast cancer, lung cancer and melanoma [80].

A myriad of miRNAs are being considered for clinical application as biomarkers or therapeutics for a variety of diseases [101]. Perhaps miRNAs that repress CYP450 enzymes to enhance tumor growth could be used as biomarkers if homeostatically their levels vary with concentration or activity of CYP450 induced by exogenous methods. For example, hsa-miR-577, hsa-miR-1, hsa-miR-532-3p and hsa-miR-627, could be used as a biomarker to gauge response to any therapeutics involving CYP3A4. Similarly miR-27b could be used as a biomarker for therapeutics related to CYP1B1 and miR-34a for CYP3A4 and CYP2C19 in addition to transcription factors and other proteins that indirectly affect CYP450 levels or activity (e.g., AKR1D1 and SLC10A1). Conversely, perhaps miRNAs that are endogenously downregulated to increase levels of CYP450 enzymes (thereby enhancing tumor growth) could be used as therapeutic options in various cancers. For instance, exogenous miR148a could be administered to boost its levels and effect on CYP3A4 and CYP2C19 in GI tumors, head & neck cancers, breast cancer, lung cancer and melanoma [80].

#### 4.4.7. Small RNA Therapeutics

In addition to miRNA, the entities antisense oligonucleotides (ASOs), aptamers, siRNAs and synthetic mRNAs are collectively termed small RNAs and are used in the development of RNA interference therapeutics [102]. ASOs are single-stranded deoxyribonucleotides, which bind complementary mRNA targets [103]. This leads to cleavage of the mRNA-DNA heteroduplex by RNase H endonuclease, which prevents translation of the target mRNA and thereby downregulates expression of the corresponding target protein [103]. Aptamers are single-stranded oligonucleotides that bind target mRNA with high affinity and specificity, through physiochemical mechanisms such as hydrophobic, electrostatic, hydrogen bonding, van der Waals forces, base stacking and shape complementarity interactions [104,105]. Due to their desirable tissue penetrability and high affinity and specificity target-binding, aptamers are expected to become a widely used platform for delivery of therapeutic small RNAs [105]. Synthetic small interfering RNAs (siRNAs) are short double-stranded RNAs that contain a guiding strand that binds target mRNA more completely than miRNA and similarly limit translation of the target mRNA [106].

Small RNAs are currently being investigated for RNA-targeting therapeutics to treat (or prevent) illness by limiting translation of mRNA destined to become disease-relevant proteins [102]. Downregulation of disease-causing genes, that is, gene silencing, occurs via binding of the RISC complex to endogenous mRNA. Small RNA can be introduced into the cell via viral vectors or by directly insertion into cytoplasm, where the antisense siRNA-RISC complex then forms and blocks translation of target complementary mRNA. The direct interaction of these small RNA (in the RISC complex) with endogenous mRNA vastly expands the repertoire of possible therapeutics for previously 'undruggable' targets. Traditionally, drugs have been developed based on their interactions as ligands for proteins, particularly those with enzyme binding sites [107]. Expanding drug development to small RNAs opens up a new frontier in therapeutics targeting nucleic acids instead of proteins, with potential for overcoming barriers to treating previously intractable diseases [102]. The promise of small RNAs therefore may address close to 85% of the human proteome lacking ligand-binding domains or enzyme binding sites [108,109]. Development of small RNA therapeutics is not without challenges. For decades, development of ASOs and siRNAs has been tempered by immunogenicity, limited potency and poor focused delivery to the cytoplasm of the right cells and tissue [102]. Other setbacks for small RNA therapeutics have included unintended off-target effects on homologous RNA sequences and premature metabolism and excretion of the small RNA. Dozens of trials have been designed to addresses these limitations. While progress is being made to reduce these barriers, more research is needed for application of siRNA therapies to become more widespread [110]. The Food and Drug Administration (FDA) has begun to approve small RNAs as biomarkers to gauge response to therapy or as therapeutics specifically targeting previously undruggable targets, such as the RAS oncogene, with ramification for patients with lung and pancreatic cancer. This is only the beginning of a new era.

#### 4.4.8. Integration of 'Omics'

In summary, multiple 'omics' systems within the individual operate independently and synergistically to modulate various states of health and disease. While it has been known for some time that variations in genomics incompletely account for interindividual variability and genotype-phenotype discordance in P450 enzyme expression, the underlying mechanisms have been unclear. The effect of multiple 'omics' on various states of health and disease is slowly being elucidated with continued investigation. The application of systematic multi-omics approaches to precision medicine and systems biology has great potential to improve the care of patients in Cardio-Oncology.
