Pharmacogenomics in Cardio-Oncology

Introduction

The cardiotoxicity of certain anticancer agents is now well-established and has led to increased cardiac screening of patients with cancer and limitation of patients' maximum cumulative chemotherapeutic dose. The effect of anticancer regimes on the heart largely involves cardiomyocyte death, leading to cardiomyopathy and heart failure, or the induction of arrhythmias. Of these cardiotoxic drugs, those resulting in clinical cardiotoxicity can range from 8-26% for doxorubicin,1,2 7-28% for trastuzumab,3,4 or 1-19% for 5-Fluorouracil.5 Regarding tyrosine kinase inhibitors (TKIs), 25 of the 27 oncology TKIs that are currently approved by the Food and Drug Administration (FDA) have some type of cardiovascular toxicity-related warning in their package inserts. For TKIs, QT prolongation and arrhythmia, ischemia, and hypertension have been reported in 2-35% of patients.6 Furthermore, newly introduced anticancer agents are commonly used as part of changed combinational regimens with significantly increased incidence of cardiotoxicity, exemplified by the combination of doxorubicin and paclitaxel.7 The mechanism of action of these drugs is often independent of their cardiotoxicity, and the basis for why these drugs specifically affect the heart has yet to be established. The genetic rationale for why certain patients experience cardiotoxicity while other patients can tolerate high anticancer drug doses has proven elusive. This has led to significant genomic efforts using targeted and genome-wide association studies (GWAS) to divine the pharmacogenomic cause of this predilection.

Patient-Specific Toxicity: Pharmacogenomics and Personalized Medicine

Achieving a tolerable balance between efficacy and toxicity is the most important challenge facing effective anticancer treatment. An individual patient's response to therapy is dependent on the plasma and target site concentration of the anticancer drugs, which are controlled by pharmacokinetic (absorption, distribution, metabolism, and excretion) and pharmacodynamic factors. Inherited polymorphisms in drug metabolizing enzymes and transporters can alter their expression and/or activity, influencing pharmacokinetics. Genetic alterations in target enzymes, transporters, ion channels, and receptors may influence drug pharmacodynamics.8 Thus, a realistic option to improve management and outcome of anticancer drug-induced cardiotoxicity is the development of individualized treatment strategies including the use of predictive genetic host factors. Extensive efforts in pharmacogenomics research have been conducted in an attempt to uncover the genetic variants associated with anticancer agents' clinical outcome. Despite this enormous effort, few biomarkers are routinely used in clinical practice, which reflects the complexity of identifying causal variants. Currently there are more than 150 drugs with FDA-approved pharmacogenetic testing information in their drug labels, most of which are anticancer agents (Figure 1).

Figure 1

Figure 1
FDA-approved pharmacogenomics biomarker in drug labeling. Bar plot diagram showing number of drugs that contain pharmacogenetic testing information in their package insert and their distribution across different therapeutic areas (n = 158) (from Magdy et al., 2016).

Although the terms "pharmacogenetic" and "genetic" testing are used interchangeably, there is a significant difference in their target population and the manner in which each test is used in clinical investigation. Pharmacogenetic testing targets subjects who are experiencing a specific disease. This method is used to provide guidance in selecting the appropriate therapeutic agent, and, in some instances with the presence of sufficient clinical data, for individualized dosing selection. On the other hand, genetic testing is utilized when assessing a relative risk of a target population to develop a certain disease, as well as when predicting patients' prognoses. Similarly, somatic (tumor) and germline (individual) mutations are two types of genetic mutations involved in predicting cancer outcome. Somatic mutations are genetic variations in the tumor tissue that affect tumor microenvironment and determine the cancer profile, including prognosis, metastasis, and aggressiveness. Studying somatic mutations will be beneficial not only in predicting disease prognosis but also in developing tumor-specific therapeutics that are capable of targeting particular oncogenic aberrations. Germline mutations are genetic variants in a patients' genome. Inherited mutations in drug transporters and/or drug metabolizing enzymes determine the concentration of drug at the target site, subsequently tuning the efficacy and toxicity of anticancer therapeutics. Additionally, germline mutations in certain signaling pathways (e.g., genes controlling DNA repair machinery, cell division, and reprogramming) may cause a predisposition to cancer. Therefore, the study of germline aberrations has significant prognostic value. Accordingly, obtaining informative genetic information about both germline and somatic polymorphisms will ideally allow us to draw conclusive decisions about disease prognoses and adequate therapeutics.9

Pharmacogenomics Study: Design and Limitations

Pharmacogenomic studies principally adapt a case control study-based design, in which frequencies of genetic variants, mainly single nucleotide polymorphisms (SNPs), are detected and compared in cases (subjects with the investigated phenotype) and controls (subjects without the investigated phenotype). Genomic research has accommodated two main approaches:

  1. Candidate gene studies in which a single gene or a list of well-founded, preselected genes are investigated
  2. GWAS in which genetic variations across the whole genome are analyzed and linked to the investigated phenotype

In terms of the number of SNPs investigated, both genomic study approaches are quite different. Candidate gene studies investigate anywhere from one SNP to a complete gene sequence. GWAS analyze a range of several hundred thousand to millions of SNPs. The momentous advances in the field of next-generation sequencing, analysis algorithms, and data storage capacity, coupled with the experimental evidences revealing the role of genetic variation in various diseases, have shifted the paradigm toward whole genome studies to help identify SNPs that protect against or predispose individuals to different clinical conditions and phenotypic traits. The number of GWAS published reports has dramatically increased over the last decade from fewer than 50 studies in 2006 to about 2,000 studies in 2013.10 GWAS are based on the principle of linkage disequilibrium (LD), which exists when two or more SNPs at discrete loci are found together more frequently than would likely happen by chance. Accordingly, analyzing only a selected set of tag-SNPs across the genome to act as surrogates for several other linked SNPs gives complete information about the un-typed SNPs. The LD-based approach is very useful as it significantly decreases the number of genotyped SNPs while providing information about the descent number of genetic variants. Nevertheless, this methodology raises the question of whether or not the identified SNP is the causal one. Even though the LD-based genome-wide study is an appropriate tool for mapping Mendelian traits that are predisposed due to the segregation of risk alleles within a single gene, it is not as efficient when it comes to polygenic traits like doxorubicin-induced cardiotoxicity. Multiple genes are implicated in doxorubicin-induced cardiotoxicity, and it thus becomes nearly impossible to identify causal SNPs with just a single association study.11 Population stratification constitutes a major limitation for GWAS because heterogeneous subject recruitment significantly affects the output of pharmacogenomic studies. Ethnically diverse populations have different LD profiles caused by distinct recombination rates. Thus, SNPs having significantly different minor allele frequencies exist among diverse populations. An exemplary African population has very short LD haplotypes because of cumulative recombination events, which make it even more difficult to capture the causal polymorphisms.12 Because minor differences in ethnicity between cases and controls can result in false positives even after exclusion of extreme outliers, an odds ratio of at least 2-3 is required for an association to be robust enough to overcome cryptic population stratification. Odds ratios <1.5 are questionable regardless of the p value.13 Failure to identify large insertions and deletions is considered another GWAS limitation because GWAS primarily focus on single base pair alterations rather than larger genetic mutations. Importantly, the majority of identified GWAS SNPs are located in intergenic or intronic regions and, in many instances, in genes that are irrelevant to the studied phenotype, where the biological relevance of identified polymorphisms is far from being well-described.

Despite the efforts and partial successes of many research groups to identify genetic polymorphisms associated with doxorubicin-induced cardiotoxicity, these studies were hampered by small sample sizes, inhomogeneous patient cohorts, and nonsystematic genetic analysis and mostly lacked any functional validation. Furthermore, doxorubicin-induced cardiotoxicity appears to be a polygenic trait, and single SNP-based association tests ignore synergistic and antagonistic effects between different genes polymorphisms. Most pharmacogenomics studies lack any downstream mechanistic validation; thus, the impact of SNPs on the biological system and the relationships between identified SNPs and doxorubicin-induced cardiotoxicity are poorly understood. Importantly, elucidation of causal mechanisms leading to SNP-associated doxorubicin-induced cardiotoxicity and functional changes are important for potential future anticancer therapeutics dosing recommendations. Testing for the causal variants will guarantee that the best possible clinical associations will be detected; however, identification of causal variants can be a challenging task. These observations taken together, coupled with the fact that multiple neglected candidate genes need to be systematically examined in relation to doxorubicin-induced cardiotoxicity, emphasize the need for a comprehensive genetic approach to address these issues. It is necessary to validate significant associations in large independent cohorts and conduct proper patient-specific functional studies for validation of SNPs implicated in doxorubicin-induced cardiotoxicity.

Conclusion

Genetic background and mechanistic explanation for doxorubicin-induced cardiotoxicity, as well as intra-individual variability across the population in susceptibility to cardiotoxic events, have yet to be determined. We believe that a comprehensive whole genome platform based on whole genome genotyping, patient-derived hiPSC-CMs, and utilization of CRISPR technology will help pinpoint robust genotype/phenotype associations and provide functional mechanistic validation for the involvement of candidate genes/SNP(s)/haplotypes in doxorubicin-induced cardiotoxicity (Table 1 and Figure 2).14 This methodology will generate a set of validated SNPs that are predictive for cardiotoxicity and can be directly used in a clinical cardiotoxicity algorithm that can classify patients who are more susceptible to doxorubicin-induced cardiotoxicity. Furthermore, this platform would provide cardio-oncologists with an invaluable tool to individualize patient-specific anticancer therapies before beginning treatment, rather than experience undesirable cardiotoxicity retrospectively. This methodology will help achieve maximal benefit and minimal side-effects from evolving anticancer therapeutics, thus significantly improving cancer treatment.

Table 1: Glossary of Terms

Abbreviation

Term

Description

hiPSC-CM

Human induced pluripotent stem cell-derived cardiomyocytes

 

CRISPR

Clustered regularly interspaced short palindromic repeats

Type II system is a bacterial immune system that has been modified for genome engineering. CRISPR could be used to overexpress, knockout target genes, and for genome editing using a DNA repair template containing the desired sequence to be inserted in the genome.

OR

Odds ratio

In pharmacogenomics studies, a measure of association between a marker (such as SNP) and outcome (such as cardiotoxicity phenotype)
OR = 1 existence of tested marker does not affect odds of outcome
OR > 1 existence of tested marker is associated with higher odds of outcome
OR < 1 existence of tested marker with lower odds of outcome

 

Haplotype

Group of alleles at different loci that are inherited together from one generation to another.

Figure 2

Figure 2
Schematic of the process for elucidating the role of genetic mutations in anticancer drug-induced cardiotoxicity (from Magdy et al., 2016).

References

  1. Swain SM, Whaley FS, Ewer MS. Congestive heart failure in patients treated with doxorubicin: A retrospective analysis of three trials. Cancer 2003;97:2869-79.
  2. Wouters KA, Kremer LC, Miller TL, Herman EH, Lipshultz SE. Protecting against anthracycline-induced myocardial damage: a review of the most promising strategies. Br J Haematol 2005;131:561-78.
  3. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, et al. Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med 2005;353:1659-72.
  4. Guarneri V, Lenihan DJ, Valero V, et al. Long-term cardiac tolerability of trastuzumab in metastatic breast cancer: the M.D. Anderson Cancer Center experience. J Clin Oncol 2006;24:4107-15.
  5. Saif MW, Shah MM, Shah AR. Fluoropyrimidine-associated cardiotoxicity: revisited. Expert Opin Drug Saf 2009;8:191-202.
  6. Moslehi JJ, Deininger M. Tyrosine Kinase Inhibitor-Associated Cardiovascular Toxicity in Chronic Myeloid Leukemia. J Clin Oncol 2015;33:4210-8.
  7. Giordano SH, Booser DJ, Murray JL, et al. A detailed evaluation of cardiac toxicity: a phase II study of doxorubicin and one- or three-hour-infusion paclitaxel in patients with metastatic breast cancer. Clin Cancer Res 2002;8:3360-8.
  8. Evans WE, McLeod HL. Pharmacogenomics--drug disposition, drug targets, and side effects N Engl J Med 2003;348:538-49.
  9. Hertz DL, McLeod HL. Use of pharmacogenetics for predicting cancer prognosis and treatment exposure, response and toxicity. J Hum Genet 2013;58:346-52.
  10. Welter D, MacArthur J, Morales J, et al. The NHGRI GWAS Catalog, a curated resource of SNP-trait associations. Nucleic Acids Res 2014;42:D1001-6
  11. Botstein D, Risch N. Discovering genotypes underlying human phenotypes: past successes for mendelian disease, future approaches for complex disease. Nat Genet 2003;33(Suppl):228-37.
  12. Reich DE, Cargill M, Bolk S, et al. Linkage disequilibrium in the human genome. Nature 2001;411:199-204.
  13. McClellan J, King MC. Genetic heterogeneity in human disease. Cell 2010;141:210-7.
  14. Magdy T, Burmeister BT, Burridge PW. Validating the pharmacogenomics of chemotherapy-induced cardiotoxicity: What is missing? Pharmacol Ther 2016;168:113-125.

Clinical Topics: Arrhythmias and Clinical EP, Cardio-Oncology, Heart Failure and Cardiomyopathies, Prevention, Implantable Devices, EP Basic Science, Genetic Arrhythmic Conditions, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias, Statins, Acute Heart Failure, Heart Failure and Cardiac Biomarkers, Hypertension

Keywords: Alleles, Antineoplastic Agents, Arrhythmias, Cardiac, Biological Markers, Cardiomyopathies, Cardiotoxicity, Clustered Regularly Interspaced Short Palindromic Repeats, Doxorubicin, Fluorouracil, Genome-Wide Association Study, Genomics, Genotype, Haplotypes, Heart Failure, Hypertension, Immune System, Induced Pluripotent Stem Cells, Multifactorial Inheritance, Myocytes, Cardiac, Paclitaxel, Pharmacogenetics, Polymorphism, Single Nucleotide, Prognosis, Protein-Tyrosine Kinases, Recombination, Genetic, Risk, Tumor Microenvironment


< Back to Listings