TEC Assessment Index

Special Report: Pharmacogenomics of Cancer–Candidate Genes

Assessment Program
Volume 22, No. 5
November 2007

Executive Summary

Background

Pharmacogenomics describes the relationship between variation in the human genome, such as differences in DNA sequence, copy number, or transcriptional perturbations, and individual variation in response to or adverse effects from drug therapy. The benefits and harms of drug therapy for cancer may be influenced by pharmacodynamic variability (genetic variability in drug target effector molecules or downstream products) or pharmacokinetic variability (genetic variability in molecular pathways involved in drug uptake, distribution, and metabolism). In cancer, pharmacodynamic variability often involves somatic (non-germline) genetic changes in tumor tissue that are characteristic of tumor type or stage; or variability in gene expression reflecting alterations in genetic regulation, gene copy number, or tumor interactions with local tissue. Pharmacokinetic variability generally reflects inherited (germline) changes in gene coding, for example, enzymes that metabolize drugs.

The goal of pharmacogenetic research is "personalized medicine," or the ability to detect key genetic variation in individual patients in order to predict the most effective treatment, or avoid severe adverse effects. Such individualized therapy could help select the best treatment or dose early, avoiding trial and error management based on averages from clinical trials in large populations. Drug safety is an important concern, and an additional hope of individualized medicine is to identify patients with a high likelihood of severe reaction to a particular drug, and either modify the initial dose or choose an alternative treatment, avoiding extended monitoring and the extensive medical support required for severe toxicity reactions.

Relevant genes for pharmacogenomic study may be chosen via two different methods. The first is selecting "candidate" genes, based on known molecular interactions associated with the drug. The second involves large microarray analysis to define genetic marker patterns that correlate with drug response or adverse events. Either way, the goal is to develop genetic tests for clinical use in predicting response to therapy, or adverse reactions prior to treatment initiation. Genetic tests are already commercially available that purport to aid in selecting therapies that are most effective or that avoid adverse events.

Objective

The objective of this Special Report is to catalog genetic tests and measures of gene activity that are currently under study for pharmacogenomic applications. Pharmacogenomics will be broadly interpreted to include detecting the specific molecular targets of the newer, "targeted" therapies to select patients likely to respond, as well as detecting other pharmacodynamic markers that influence targeted therapy response. Molecular indicators of variability in response or toxicity to standard chemotherapy drugs will also be reviewed. This Report will also provide a more detailed summary of the supporting data for a few illustrative examples. The tests identified in this Report are not comprehensive, but rather reflect the greatest research activity and in most cases are commercially available. Further, only tests developed from candidate genes will be examined; microarray pattern analysis and the development of predictive panels composed of many genetic markers are beyond the scope of this Report.

Search Strategy

A MEDLINE search of relevant review articles was completed for 2006 and 2007 (through June). The search strategy included the text words "pharmacogenetics OR pharmacogenomics" combined with the MeSH® term "neoplasms." The bibliographies of these review articles were also examined for other relevant articles.

Selection Criteria

Articles were reviewed and a list was constructed of pharmacogenomic tests that were most often described and that were recommended for active use in clinical situations. The test menus of several laboratories known to offer molecular tests were also reviewed and a list of tests constructed. A final list of tests to be described in this report was limited primarily to those recommended in several articles for active use and available from CLIA-licensed clinical laboratories.

In addition to information culled on each test from the review articles, separate searches were conducted for articles on each test. These searches consisted of the term for the marker detected by the test, combined with terms such as "therapy," "chemotherapy," "response," "toxicity," or "adverse effects" [MeSH®], depending on the intent of the test. Relevant articles that provided examples or supportive data on test clinical use were selected for review.

Discussion

Targeted Therapy

Ideal drugs would interfere with a molecular target that is critical to and restricted to specific cancer types; presence of the target in the tumors of individual patients would vastly increase the likelihood of a meaningful clinical response and thus would determine eligibility for the targeted treatment. Such targets may be normal proteins that are found in abundance in cancerous tissue (e.g., CD20 protein expressed on the surface of normal B cells and on B-cell non-Hodgkin’s lymphoma cells); or may be the expression of somatically acquired genetic variants found only in malignant cells (e.g., BCR/ABL gene rearrangement [Philadelphia chromosome] characteristic of most chronic lymphocytic leukemias). This section describes types of targeted therapies, discusses a particular example (trastuzumab [Herceptin®]), describes the ideal co-development of targeted drug and target assay, and the impact of multiple targets. A table lists several examples of targeted therapy and specific laboratory tests that may (or may not) be used to determine eligibility for therapy.

Targeted Therapy: Pharmacodynamic Predictors of Response

Detection of the target molecule of a targeted drug is the first step in determining the likelihood of a response to therapy. However, in some cases presence of the target may not be sufficient; additional genetic alterations that differ among patients may affect response. Detection of these additional genetic changes may be important for selecting initial therapy to obtain the optimal response. In addition, initial response to a targeted drug may not last due to the emergence of drug-resistant clones. The table in this section lists some examples of genetic changes that are under study as additional determinants of response and/or resistance to targeted therapies. Genetic testing for imatinib resistance in chronic myelogenous leukemia and in gastrointestinal stromal tumors is discussed in greater detail.

Predicting Response to Non-"Targeted" Chemotherapy

Chemotherapy drugs that were not originally developed to "target" a specific molecule and favorably modify disease have historically been tested, dosed, and incorporated into treatment protocols based on "trial and error" approaches resulting in a single or a range of recommended dosages based on studies of populations. However, germline interindividual variability in rates of metabolism of these drugs results in sometimes large interpatient differences in systemic exposure, leading to toxicity for some, lack of efficacy for others, and a satisfactory response mainly for those close to population average metabolism. Chemotherapy drugs commonly have a narrow therapeutic index that may overlap with the range of systemic exposure, resulting in severe toxicity for some patients.

Many studies have been published describing associations between germline genetic variants of metabolizing enzymes and the toxicity of or the response to various chemotherapy drugs. The table in this section provides limited detail on a few examples, most of which have commercially available genetic tests. In addition, two examples are discussed in more detail: dihydropyrimidine dehydrogenase deficiency and its importance in predicting 5-fluorouracil and capecitabine toxicity; and tamoxifen efficacy in cytochrome p450 2D6 poor metabolizers.

TEC Assessment Index

KEYWORDS: ABL; acute myeloid leukemia; adverse effects; AML; analytic validity; BCR; bevacizumab; breast cancer; cancer; candidate; catalog; CD33; cetuximab; chemotherapy; chronic myelogenous leukemia; CLIA; clinical utility; clinical validity; colorectal cancer; EGAPP; EGFR; epidermal growth factor receptor; erlotinib; FISH; gefitinib; gene testing; genetic variation; genetics; genomics; germline; GIST; HER2; Herceptin; IHC; imatinib; KIT; leukaemia; lung cancer; MAb; microarray; molecular; monoclonal antibody; mutation; neoplasms; non-Hodgkin’s lymphoma; oncogene; oncology; PCR; personalized medicine; pharmacodynamics; pharmacogenetics; pharmacogenomics; pharmacokinetics; Philadelphia; polymorphism; predictive; proto-oncogene; side effects; small molecule; sorafenib; stromal tumor; target; temsirolimus; transcriptional; trastuzumab; VEGF; wild type;