The Goal of Pharmacogenomics
The branch of science dealing with genetic heterogeneity and patient drug response is known as pharmacogenomics (2). It is a field of science that aims to optimize drug therapy and minimize drug toxicity with respect to a patient’s genotype (1). Moreover, it is a tool that physicians can utilize to design patient-tailored treatment (1), and therefore, is a fundamental basis of personalized medicine. The important concepts involved in pharmacogenomics are the idea of genetic heterogeneity, particularly in the context of single nucleotide polymorphisms (SNPs), which then can result in diverse responses to drugs, possibly leading to adverse drug reactions (ADR).
Single Nucleotide Polymorphisms (SNPs)
One of the key findings of the Human Genome Project was that the human genome varied greatly between individuals. Often variations were observed to be at the single nucleotide level, also known as single nucleotide polymorphisms, or SNPs (3). SNPs account for over 90% of an individual’s genetic variations and may be defined as variations in DNA that occur at a frequency of greater than 1% in the population (20). SNPs occur in all areas of the genome whether they are non-coding or coding regions (3). According to the National Center for Biotechnology Information (NCBI) there have been 187,852,828 SNPs identified as of June 2012 (4). The NCBI’s Single Nucleotide Polymorphism database (dsSNP) lists over 40 million SNPs (as of March 2015) that have been validated and assigned an official RefSNP identification (rsID) (18, 20). The clinical significance of SNPs lies in the fact that they may be associated with particular phenotypes resulting from a variation in gene expression, function, or activity (9). By altering key amino acid sequences, RNA splice sites, transcriptional promoters, enhancers, or susceptibility to epigenetic mechanisms, SNPs may significantly alter an individual’s response upon exposure to a certain drug (12, 19). Classic examples of SNPs affecting a patient’s drug response are those belonging to genes coding for drug-metabolizing enzymes (7); SNPs may also lead to receptor or transporter polymorphisms that could affect a particular drug’s absorption, distribution, metabolism or excretion (19). Here we will describe two cases of common chemotherapeutic drugs to illustrate the importance of SNPs in pharmacogenomics.
Adverse Drug Reactions (ADRs)
Within the last decade, a large portion of medical research has been oriented towards advancements in personalized therapeutic approaches. These efforts have grown primarily in response to issues surrounding ADRs which is defined as any undesired consequence associated with the normal use of medication (8). ADRs cannot be defined as side-effects as side-effects may be beneficial whereas ADRs are only of a negative nature (20). Primarily, these issues revolve around the consequences of patient drug-induced toxicity, which include patient mortality and rising healthcare costs (8). Several researchers have sought out to quantify these issues. According to Ross et al., ADRs claim the lives of 100,000 to 218,000 lives in the US annually, resulting in a cost of over $100 billion (8) and is the 5th leading cause of death in the USA (20).
The high prevalence of ADRs is due to individual variations in patient drug response (10). The current paradigm of modern drug development revolves around a paradox where clinical trials determine a normalized dosage of drugs for the safety and efficacy of the whole population whereas physicians prescribe these normalized doses to individual patients who may respond quite differently to the drug with either no response, an ADR, or hopefully, a safe and effective response (20). This phenomenon has persisted in medical history since around 400 BC, the time of Hippocrates, known as the ‘Father of Western Medicine’ (11). In one of his several works, he discusses the personalization of therapy by “[providing] different [drugs] to different patients, for the sweet ones do not benefit everyone” (11). As Hippocrates hints, variations in drug response are due to several factors (see Table 9.4.1) such as a drug’s pharmacokinetic properties (i.e. how the drug is metabolized) and a patient’s physiological and lifestyle (i.e. age, gender, and diet) (9). However, one of the most influential factors causing variations in drug response remain to be the wide genetic heterogeneity amongst individuals (9).
How do scientists identify SNPs to prevent ADR?
The discovery of genetic variants has shown to be a benefit in patient prognosis and a key step in ADR prevention (9). This has been particularly prevalent in cancer therapy, where treatments are often burdened with narrow therapeutic windows. A common approach utilized to discover associations between genetic variants, such as SNPs, and a particular trait (i.e. increased drug sensitivity), is the genome wide association study (GWAS) (5).
A GWAS compares the genotypic profiles of cases to controls, in which cases are individuals with a specific trait of interest and controls are individuals who share similar characteristics to the cases cohort but not the trait of interest (5). Typically, GWAS are based on genotypic arrays, such as those highlighted in our genotyping section. A clear benefit of GWAS is that it can identify SNPs that are associated with a particular trait. However, a common critique of this method is that it cannot specify a causal gene (5). Nevertheless, GWAS can identify trait-associated SNPs and quantify the strength of association, or effect size, in the form of odds ratios (5).
In order for a case-control GWAS to be bias free, there are several thing to keep in mind. Firstly, the cases must represent all those who develop the disease of interest, because a survival bias can arise when different severities of the disease are missed, especially for fatal diseases with rapid progression. Secondly, the control groups must describe everyone that are at risk for the disease of interest, otherwise the population is not properly represented rendering the study invalid. Additionally, the data collection method of risk factor and exposure must be equal between controls and samples. Finally, the ancestral geographical location and major environmental exposures of controls should relatively equal to the cases (14). Case-control studies are prone to many types of biases and these requirements need to be met to increase relevance of human genome research.
Case 1: Azathioprine
As described by Sim et al., one such enzyme involved in drug metabolism is thiopurine methyltransferase (TPMT), which is responsible for the metabolism of thiopurine drugs such as 6-mercaptopurine (6-MP), also known as azathioprine in its pro-drug form. Azathioprine is a purine synthesis inhibitor commonly used for treating leukemia (10). Like many cancer drugs it has a narrow therapeutic window, in which a limited range of dosage is considered as both safe and effective. A common ADR resulting from an overdose of azathioprine is myelosuppression-a decrease in myeloid derived cells such as white and red blood cells (10). As a result, several vital processes, such as immunity and blood-oxygen transportation, are severely impaired and may become fatal (5). Thus, patients treated with azathioprine are constantly monitored by means of analyzing serum-drug metabolites to ensure that the appropriate dosage has been administered (10).
Interestingly, individuals differ in their ability to metabolize, or inactivate, azathioprine (10). This is primarily due to the presence of two SNP variants of TPMT, known as TPMT*2 and TPMT*3C. These variants confer low TPMT activity and as a result, a decreased ability to inactivate azathioprine (10). Since each individual receives two copies of a gene (i.e. alleles), one from both their mother and father, it is possible to have up to two copies of these variants. If an individual carries two copies of a low activity allele their ability to metabolize azathioprine will be significantly lower compared to those who carry normal activity alleles causing the azathioprine to accumulate. Consequently, these patients will require a much lower dose than normally administered or they will suffer from severe, and possibly lethal, azathioprine toxicity (10, 12). Even if a patient carries just one copy of a low activity variant, along with a normal copy, their ability to inactivate azathioprine will still be much less than an individual who carries two normal activity copies (10). These patients are also at risk for overdose, and adjustments from the normal dose must be made.
Case 2: Warfarin
The importance of genotyping for genetic variation when determining drug dosage is also seen by the example of the anticoagulant warfarin. This drug inhibits blood coagulation by inhibiting the enzyme Vitamin K Epoxide Reductase (VKORC1), which in turn inhibits generation of vitamin K and activation of coagulation factors resulting in a decrease in blood clotting (15). CYP2C9 is an enzyme that breaks down and inactivates warfarin and leads to excessive bleeding (15). The metabolic pathway of warfarin interactions with these two enzymes is depicted below.
Figure 6.4.1. Warfarin interactions with CYP2C9 and VKORCI enzymes
Variations in the VKORC1 and CYP2C9 genes accounted for 46-68% of variability of warfarin dosage among individuals (15). It is essential for oncologists to determine the appropriate dose of warfarin to give to patients because too low of a dose will result in increased blood clotting and too large a dose will result in excessive hemorrhaging. It is paramount that a delicate balance of dosing is found for each individual patient. Polymorphisms in the CYP2C9*2 and *3 alleles lead to functionally defective variants with low activity, meaning that these patients need a lower warfarin dose because it is harder to metabolize it (15). Additionally, patients that present with allelic variants in the VKORC1 gene also require a lower dosage of warfarin to prevent excessive warfarin function (15). The anticoagulation effect of warfarin is not only attributable to genetic factors and can also be influenced by factors such as age, sex, body weight and diet (15). Research has also shown that there may be differences in the frequency of allelic variation between ethnic groups (15). For instance, functionally defective CYP2C9*2 and *3 alleles are much less common in Asian populations than in Europeans, while VKORC1 has varying haplotype diversity in various ethnic populations (15).
Case 3: Anthracyclines
Anthracyclines such as Doxorubicin and Daunorubicin are highly effective as chemotherapy reagents. As so, they are administered to 60-70% of childhood cancer patients for the treatment of leukemias and solid tumors, as well as for the treatment of adult cancers such as breast cancer, sarcoma, lymphoma, and leukemia to name a few. However in 1967 anthracyclines were discovered to cause fatal cardiac toxicity (15). This extreme toxicity is only seen in a small subset of patients and it’s evident that the tolerable doses vary greatly among patients. While some patients experience toxicity at less than 300 mg/m2, others can tolerate over 1000 mg/m2. What’s even more concerning is that heart failure can occur years after the drug is given.
So what is the mechanism behind this adverse drug reaction? SNP analysis in patients that received anthracyclines showed a significant association in the ADR to the presence of a particular allele of the gene SLC28A3. This gene encodes a sodium-coupled nucleoside transporter expressed throughout the whole body including the heart and is capable of transporting pyrimidines, purines, and anti-cancer drugs such as anthracycline across the cell membrane (16). The protective allele of SLC28A3 is L461L, which has been shown to be associated with reduced RNA expression of this transporter, meaning less of the drug is pumped into the cells conferring reduced toxicity and higher tolerance due to reduced toxic metabolites. Another gene that is associated to this anthracycline-induced cardiotoxicity is UGT1A6, a UDP glucuronosyltransferase (17).
By understanding which patients are at high risk compared to intermediate or low risk, we can tailor their anthracycline therapy to their specific genetic needs. For example those patients that are at low risk for the ADR might simply undergo an echocardiogram follow-up as a precautionary measure after treatment. And for those at intermediate risk these echocardiogram follow-ups may be intensified. On the other hand, patients that undergo anthracycline treatment that possess the allele for high risk may be given alternative medications or lower doses as well as additional cardioprotectant. Another safety measure to prevent heart failure in these high-risk patients would be to start anthracycline treatment with ACE-inhibitors or beta-blockers to prevent further damage. By using genetic markers for predicting the risk of anthracycline-induced cardiotoxicity, pharmacogenomics allows personalized medicine with safer treatment options during cancer therapy.