Pharmacogenomics in Cancer Care: Adding Some Science to the Art of Medicine

Navin Pinto, MD

Uncategorized

Despite increasing publicity, “personalized medicine” is not a new phenomenon in cancer care. Oncologists have long used criteria such as body size, perfor­mance status, comorbid conditions, organ function, lifestyle, and a patient’s goals of care to individualize treatment decisions and drug doses. Additionally, dose adjustments of 1 or more agents during the course of treatment based on a physician’s knowledge of the toxicity profiles of chemotherapies is a common strategy to tailor a given patient’s treatment program. Clinically significant gene-drug interactions have been known for decades, and over the past several years, the FDA has altered the labeling of certain drugs, including several used in cancer care, to warn practitioners of potential gene-drug interactions and to recommend focused genetic testing prior to the initiation of therapy for several agents.¹ Since publication of the first drafts of the human genome in 2001,^2,3 our knowledge of the role of inherited, common genetic variation across the genome in both drug response and toxicity has increased exponentially. Coupled with the rapidly decreasing cost of genotyping and whole-genome analyses,4 the era of personalized genomics is upon us.

Oncologists are at the forefront of incorporating genetic testing into clinical care decisions. Sophisticated
molecular techniques to identify on­cogenic drivers and/or resistance mechanisms within tumors are becoming a part of the diagnostic workup and often help guide therapeutic decision making. For example, testing for somatic mutations, such as BRAF V600E, BCR-ABL, and EML4-ALK, is now commonplace in metastatic melanoma, chronic myeloid leukemia, and non–small cell lung cancer, respectively. The body of data derived from these tests has led to subdivisions of disease, which in turn has helped oncologists select the best therapies for their patients and enabled clinically significant improvements in survival for many of these disease subcategories.^5-7

Germline pharmacogenomics offers clinicians a mechanism to subclassify patients (rather than diseases) and personalize care, analogous to the approach of using tumor genomics to guide disease-specific therapy. In an era when adverse drug reactions are a major cause of morbidity and mortality,8 there is an increased patient and regulatory demand to incorporate pharmacogenomics into clinical care as a preventive strategy. Furthermore, when multiple treatment options are available for a given malignancy or for supportive care measures used in oncology, pharmacogenomics may enable selection of the least toxic treatment regimen without compromising overall treatment success.

Ever-increasing consumer demand and a growing body of clinical evidence have led to the rise of both commercial and academic organizations offering personalized genomic services. These services offer patients low-cost genomic screening, including customized reports about susceptibility to both disease and adverse drug reactions. These reports will inevitably land on the desks of treating oncologists along with patient requests to use these results to prescribe the safest and most effective treatment plans. Oncologists now have the ability to hone their sharp personalized medical skills with germline genomic information from their patients.

In this article, we will use 3 hypothetical case scenarios to illustrate the power of prospective genotyping in selecting the least toxic chemotherapy or supportive care plan, determining the right dose of cytotoxic chemotherapy, and identifying the agent responsible for a given toxicity in a combination chemotherapy plan with overlapping toxicities. We hope to illustrate that the much-lauded future of personalized medicine – incorporating genetic information into clinical care decisions – is today.

Case 1:

J.Y. is a 45-year-old female with recently diagnosed stage IA (1.3 cm) estrogen receptor–positive (ER+) breast cancer whose status is post lumpectomy and radiation. Her medical history is notable for menopause at age 42 and mild osteopenia. Family history is significant for an aunt diagnosed with ER+ breast cancer at age 57 who experienced severe musculoskeletal pain while on a clinical trial of exemestane. Her current oncologist is recommending an aromatase inhibitor (AI), but the patient is worried about out-of-pocket costs and has concerns about her family history of poor tolerance to aromatase inhibition. She is in your office to discuss the appropriate maintenance hormone therapy, given her family history, and wonders if genetic testing may help in deciding which drug is best.

This case illustrates one of the most common pharmacogenomic inquiries in oncology. The selective ER modulator tamoxifen is FDA approved for the treatment and prevention of breast cancer in women of all ages and is the most commonly prescribed therapy for the treatment of ER+ breast cancer worldwide.^9,10 Tamoxifen is converted to active metabolites by several cytochrome P450 enzymes, with CYP2D6 playing a dominant role in the creation of 2 of the most potent ER-blocking metabolites, 4-hydroxytamoxifen and endoxifen.^10,11 More than 100 genetic variants of CYP2D6 have been reported, some of which can lead to altered enzymatic function. Algorithms exist to label patients ultrarapid, extensive, intermediate, and poor CYP2D6 metabolizers based on the number of functional alleles, with ultrarapid metabolizers having numerous copies of CYP2D6 and poor metabolizers having 2 nonfunctional alleles.¹² Variant CYP2D6 allele frequencies differ by ethnicity,¹³ but between 7% and 21% of the population are intermediate or poor metabolizers of CYP2D6 substrates.¹⁴

Although a paucity of prospective data exists, there is retrospective evidence that decreased CYP2D6 function (either due to genetics or concomitant CYP2D6 inhibitors) is associated with increased rates of breast cancer recurrence.^15-18 In 2006, based in part on these results, the FDA recommended that the labeling of tamoxifen be changed to include language that CYP2D6 poor metabolizers may have inferior disease control compared with extensive metabolizers. Although prospective genotyping was not recommended in the FDA label, several companies are now offering CYP2D6 genetic testing to patients, and patients are turning to their oncologists for pharmacogenetic advice.

Although recent large retrospective analyses have thrown the association between CYP2D6 genotype and tamoxifen response into question,^19,20 controversy surrounds the methods used in at least 1 of these analyses,²¹ and the preponderance of evidence supports the use of genotyping to assess which patients might best respond to tamoxifen therapy. Because of lower cost and less toxicity relative to AIs, tamoxifen remains an important choice for women with breast cancer.²² While AIs are usually well tolerated, arthralgias occur in about 50% of patients and likely contribute to suboptimal adherence associated with this class of medications.²³ Recent genome-wide analyses in patients treated with AIs for early-stage breast cancer have identified variants (single-nucleotide polymorphisms) on chromosome 14 associated with the nearby gene TCL1A, whose expression was associated with severe musculoskeletal adverse events (AEs).²⁴

One could easily take the patient’s concerns about her family history of AEs and increased costs associated with AIs into account and simply prescribe tamoxifen. However, genotyping to determine her CYP2D6 status (ultrarapid, extensive, intermediate, or poor metabolizer) and/or her risk of musculoskeletal AEs from an AI would help to make a more informed decision about the most tolerable regimen.

Case 2:

I.S. is a 73-year-old Caucasian male concert pianist with metastatic colorectal cancer. Given the patient’s profession, FOLFIRI (folinic acid, 5-fluorouracil [5-FU], irinotecan) is chosen over FOLFOX (folinic acid, 5-FU, oxaliplatin) chemotherapy because of the risk of oxaliplatin-induced neuropathy. Cycle 1 is complicated by acute diarrhea responsive to atropine and loperamide, grade IV neutropenia complicated by fever and hospital admission, and grade 1 hand-foot syndrome. Since it is unclear which drug is causative for the severe neutropenia, you prescribe a 25% dose reduction of both 5-FU and irinotecan for cycle 2.

Although the practice of oncology is primarily based on randomized clinical trials, there are many situations where there are 2 or more appropriate treatment approaches. Thus, oncologists are often left to practice the art of medicine and use patient factors such as lifestyle, performance status, and comorbidities as well as the side effect profile when choosing the right treatment plan for their patients. FOLFIRI and FOLFOX have emerged as the leading options for the treatment of locally advanced or metastatic colorectal cancer, with head-to-head comparisons demonstrating similar response rates.²⁵ Despite this, the 2 regimens have distinct toxicity profiles, with peripheral neuropathy attributable to oxaliplatin being one of the most worrisome side effects for FOLFOX and diarrhea attributable to irinotecan as a leading side effect of FOLFIRI.

For Case 2, FOLFIRI appears to be the better choice, given our patient’s profession as a concert pianist. The patient is experiencing mild toxicities from both irinotecan (diarrhea) and 5-FU (hand-foot syndrome), but the more severe toxicity of neutropenia is not clearly attributable to 1 drug, given the overlapping toxicities of the agents. Existing pharmacogenetic knowledge on both of these drugs may help in determining the agent responsible for the neutropenia and lead to a more informed dose-reduction strategy.

Irinotecan and 5-FU have marked interpatient variability in toxicity, including neutropenia, and have been the subject of intense pharmacogenetic investigation.^26,27 Irinotecan is a prodrug hydrolyzed by carboxylesterases to the more potent topoisomerase inhibitor, SN-38, which is in turn inactivated via glucuronidation by UGT1A1.²⁶ Genetic variation within UGT1A1 that reduces enzymatic activity, most notably a 7-TA repeat within the gene promoter (UGT1A1*28), has been associated with interpatient variability in susceptibility to neutropenia at all dose levels of irinotecan.²⁸ Dihydropyrimidine dehydrogenase (DPD, encoded by the gene DPYD) is responsible for more than 80% of 5-FU catabolism, and toxic deaths from 5-FU due to DPD deficiency were first reported nearly 30 years ago.²⁹ The most well-characterized reduced-function variant, DPYD*2A, is found in up to 50% of Europeans with DPD deficiency, but efforts to link additional DPYD genetic variants to functional DPD deficiency have been problematic.³⁰ The FDA has updated the labels of both irinotecan and 5-FU to urge caution in dosing in patients with reduced function of UGT1A1 or DPD, respectively.

Had our patient been prospectively genotyped, we may have discovered him to be a heterozygote in DPYD (DPYD*1/*2A) and wild-type at UGT1A1. One could make an educated assumption that poor irinotecan (SN-38) metabolism is likely not playing a role in the patient’s neutropenia and that reduced DPD function may be to blame. Cycle 2 could proceed with a dose reduction only in the 5-FU, with full-dose irinotecan, allowing the patient to receive maximal therapy.

Case 3:

J.N. is a 16-year-old Asian American male with B-precursor acute lymphoblastic leukemia (ALL). During induction therapy the patient develops a desquamating rash consistent with Stevens-Johnson syndrome (SJS). This is attributed to allopurinol given as tumor lysis prophylaxis. He requires total parenteral nutrition and a 2-week delay in the resumption of his induction therapy. During consolidation therapy, the patient develops high-spiking fevers, vomiting, and lethargy. He is found to be hypotensive and severely pancytopenic in the emergency room. Despite rapid initiation of antibiotic therapy, adequate transfusion support, and aggressive hypotension management, the patient dies 3 days later in the pediatric ICU. Several months later, his family angrily calls, stating that they read that genetics could have predicted the severe side effects he encountered during his therapy.

In oncology, where the therapeutic index of the majority of our active agents is quite low, severe adverse drug reactions are not uncommon. They are particularly troublesome in scenarios like Case 3, where the underlying disease is highly curable and the toxicities could have been accurately predicted and prevented with more information.

A growing body of evidence is highlighting the role of common genetic variation within the human leukocyte antigen (HLA) loci and susceptibility to severe drug hypersensitivity reactions. Drugs such as carbamazepine and abacavir now carry boxed warnings from the FDA of an extremely increased risk for severe adverse drug reactions (such as SJS) in patients with the variant HLA-B alleles HLAB*1502 and HLAB*5701, respectively.^31,32 Similar data exist for one of the most commonly used supportive care agents in hematologic malignancies, allopurinol, which is used to prevent and treat tumor lysis syndrome–associated hyperuricemia. Hypersensitivity reactions, although rare, are a potentially fatal side effect from this medication. Recently, the variant HLAB*5801, seen in <1% of Western Europeans but in up to 8% of Southeast Asians, was found to be strongly linked to the development of allopurinol-induced SJS, with a nearly 100-fold increase in the incidence of SJS compared with controls.33 Alternative tumor lysis syndrome management strategies, such as the recombinant urate oxidase (rasburicase), exist and should be utilized in patients found to be at high risk for allopurinol sensitivity.

The thiopurines 6-mercaptopurine (6-MP) and 6-thioguanine (6-TG) are backbones of curative treatment of ALL. One of the earliest pharmacogenetic associations linked the variability in enzymatic activity of thiopurine methyltransferase (TPMT) to the observed variability in both toxicity and response to the thio­purines.³⁴ Later, genetic variation within TPMT was found to account for the majority of enzymatic variation.³⁵ In patients with 1 reduced-function TPMT allele, 60% will not tolerate full-dose 6-MP, and all patients with 2 reduced-function TPMT alleles will develop severe, life-threatening myelosuppression when exposed to continuous standard-dose 6-MP.36 Based on these findings, the FDA encourages TPMT function testing or TPMT genotyping for patients needing thiopurines, and formal genotype-based dosing guidelines exist.36 Despite these guidelines, most patients with ALL are not typically assessed for TPMT status.

This case illustrates the danger of ignoring well-validated pharmacogenetic associations linked to life-threatening adverse drug reactions. Prospective genotyping could have revealed variant HLAB and TPMT genotypes in this patient and would likely have saved his life by predicting his risk of either allopurinol hypersensitivity or thiopurine-induced severe myelosuppression. Alternative supportive care measures and chemotherapy dosing strategies based on genotype would have reduced this patient’s risk of AEs without impacting his chance of cure.

Conclusions and Future Directions

Oncologists practice a form of personalized medicine that attempts to give the right dose to the right patient largely by using their clinical knowledge of the best evidence and a patient’s goals of therapy to guide the principles of management. Pharmacogenomics will likely never replace this clinical judgment, and while each of the genetic variants described in this article impacts only 1 aspect of pharmacokinetics or pharmacodyna­mics, pharmacogenetic algorithms that incorporate multiple variants are now emerging in medicine and will inevitably make their way into oncology. Several well-characterized – and many less well-characterized – gene-drug interactions are becoming more and more evident to the increasingly well-informed patient, and personalized genomic testing is now in the hands of thousands of consumers. As highlighted in this article, oncologists can harness the power of pharmacogenetics in several ways to help their patients select the most appropriate treatment plan, to determine the offending agent in a combination chemotherapy regimen with overlapping toxicity, and to avoid potentially life-threatening adverse drug reactions.

Germline variation and its impact on pharmaco­kinetics, pharmacodynamics, and immunogenicity can help us understand both interpatient variability in toxicity and, in some cases, response to a given drug, just as understanding the tumor’s driving mutation(s) helps clarify the patient’s disease. As consumer awareness and scientific inquiry expand in pharmacogenomics, the list of important gene-drug interactions is likely to grow, and expert opinion on drug and dose selection based on a patient’s genetic makeup will be of the utmost importance. To date, routine clinical pharmacogenetic implementation has been hindered by lack of physician and patient knowledge on the subject. Specifically, physicians need to know how and where to order pharmacogenetic or genomic testing, as well as be aware of test turnaround timing. With rapid and inexpensive genomic screens accessible to most patients, physicians should educate themselves on the well-validated pharmacogenetic associations in their field and be prepared to make treatment recommendations based not only on a physical examination, but also on a genetic examination. Several centers, including ours, are incorporating prospective genotyping into routine medical care, as well as providing expert opinion portals for providers to access pharmacogenetic information for their patients.³⁷

In oncology, one of the barriers to pharmacogenetic implementation has been the physicians’ desire to rapidly initiate treatment. As evidenced by the cases outlined in this article, prospective genotyping can aid in selecting the appropriate therapy and, in some cases, the appropriate dose. Oncologists should therefore be leading the charge in the call for prospective genotyping so this information is at the ready when we are faced with a new patient. Because randomized, double-blind trials to prospectively test the clinical utility of single variants will be difficult to conduct, evidenced-based pharmacogenetic dosing strategies are likely to be considered only noninferior to conventional approaches, yet still appropriate for clinical implementation.38 Although further validation is needed for many preliminary pharmacogenetic associations, well-performed studies in additional cohorts are coming at a rapid pace, and the list of gene-drug associations with actionable strength of evidence is likely to grow in the coming years. With this ever-growing body of evidence and the rapidly decreasing cost of genomic assessments, the future of prospective genotyping is “today,” and routine clinical implementation will likely lead to improvements in drug safety as well as efficacy.

References

1. Drugs: table of pharmacogenomic biomarkers in drug labels. FDA Web site. www.fda.gov/drugs/scienceresearch/researchareas/pharmacogenetics/ucm083378.htm. Updated August 3, 2012. Accessed September 9, 2012.
2. Lander ES, Linton LM, Birren B, et al; International Human Genome Sequencing Consortium. Initial sequencing and analysis of the human genome. Nature. 2001;409:860-921.
3. Venter JC, Adams MD, Myers EW, et al. The sequence of the human genome. Science. 2001;291:1304-1351.
4. Wetterstrand KA. DNA sequencing costs: data from the NHGRI Large-Scale Genome Sequencing Program. National Human Genome Research Institute Web site. www.genome.gov/sequencingcosts. Updated May 21, 2012. Accessed September 9, 2012.
5. Sosman JA, Kim KB, Schuchter L, et al. Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib. N Engl J Med. 2012;366:707-714.
6. O’Brien SG, Guilhot F, Larson RA, et al; IRIS Investigators. Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med. 2003;348:994-1004.
7. Kwak EL, Bang YJ, Camidge DR, et al. Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. N Engl J Med. 2010;363:1693-1703.
8. Shepherd G, Mohorn P, Yacoub K, et al. Adverse drug reaction deaths reported in United States vital statistics, 1999-2006. Ann Pharmacother. 2012;46:169-175.
9. de Souza JA, Olopade OI. CYP2D6 genotyping and tamoxifen: an unfinished story in the quest for personalized medicine. Semin Oncol. 2011;38:263-273.
10. Wu X, Hawse JR, Subramaniam M, et al. The tamoxifen metabolite, endoxifen, is a potent antiestrogen that targets estrogen receptor alpha for degradation in breast cancer cells. Cancer Res. 2009;69:1722-1727.
11. Desta Z, Ward BA, Soukhova NV, et al. Comprehensive evaluation of tamoxifen sequential biotransformation by the human cytochrome P450 system in vitro: prominent roles for CYP3A and CYP2D6. J Pharmacol Exp Ther. 2004;310:1062-1075.
12. Owen RP, Sangkuhl K, Klein TE, et al. Cytochrome P450 2D6. Pharmacogenet Genomics. 2009;19:559-562.
13. Bradford LD. CYP2D6 allele frequency in European Caucasians, Asians, Africans and their descendants. Pharmacogenomics. 2002;3:229-243.
14. Crews KR, Gaedigk A, Dunnenberger HM, et al; Clinical Pharmacogenetics Implementation Consortium. Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for codeine therapy in the context of cytochrome P450 2D6 (CYP2D6) genotype. Clin Pharmacol Ther. 2012;91:321-326.
15. Kelly CM, Juurlink DN, Gomes T, et al. Selective serotonin reuptake inhibitors and breast cancer mortality in women receiving tamoxifen: a population based cohort study. BMJ. 2010;340:c693.
16. 16. Dezentjé VO, van Blijderveen NJ, Gelderblom H, et al. Effect of concomitant CYP2D6 inhibitor use and tamoxifen adherence on breast cancer recurrence in early-stage breast cancer. J Clin Oncol. 2010;28:2423-2429.
17. Lim HS, Ju Lee H, Seok Lee K, et al. Clinical implications of CYP2D6 genotypes predictive of tamoxifen pharmacokinetics in metastatic breast cancer. J Clin Oncol. 2007;25:3837-3845.
18. Schroth W, Goetz MP, Hamann U, et al. Association between CYP2D6 polymorphisms and outcomes among women with early stage breast cancer treated with tamoxifen. JAMA. 2009;302:1429-1436.
19. Regan MM, Leyland-Jones B, Bouzyk M, et al; Breast International Group (BIG) 1-98 Collaborative Group. CYP2D6 genotype and tamoxifen response in postmenopausal women with endocrine-responsive breast cancer: the Breast International Group 1-98 trial. J Natl Cancer Inst. 2012;104:441-451.
20. Rae JM, Drury S, Hayes DF, et al; ATAC trialists. CYP2D6 and UGT2B7 genotype and risk of recurrence in tamoxifen-treated breast cancer patients. J Natl Cancer Inst. 2012;104:452-460.
21. Nakamura Y, Ratain MJ, Cox NJ, et al. Re: CYP2D6 genotype and tamoxifen response in postmenopausal women with endocrine-responsive breast cancer: the Breast International Group 1-98 trial. J Natl Cancer Inst. 2012;104:1264.
22. Winer EP, Hudis C, Burstein HJ, et al. American Society of Clinical Oncology technology assessment on the use of aromatase inhibitors as adjuvant therapy for postmenopausal women with hormone receptor-positive breast cancer: status report 2004. J Clin Oncol. 2005;23:619-629.
23. Gaillard S, Stearns V. Aromatase inhibitor-associated bone and musculoskeletal effects: new evidence defining etiology and strategies for management. Breast Cancer Res. 2011;13:205.
24. Ingle JN, Schaid DJ, Goss PE, et al. Genome-wide associations and functional genomic studies of musculoskeletal adverse events in women receiving aromatase inhibitors. J Clin Oncol. 2010;28:4674-4682.
25. Colucci G, Gebbia V, Paoletti G, et al; Gruppo Oncologico Dell’Italia Meridionale. Phase III randomized trial of FOLFIRI versus FOLFOX4 in the treatment of advanced colorectal cancer: a multicenter study of the Gruppo Oncologico Dell’Italia Meridionale. J Clin Oncol. 2005;23:4866-4875.
26. Innocenti F, Ratain MJ. Pharmacogenetics of irinotecan: clinical perspectives on the utility of genotyping. Pharmacogenomics. 2006;7:1211-1221.
27. van Kuilenburg AB. Dihydropyrimidine dehydrogenase and the efficacy and toxicity of 5-fluorouracil. Eur J Cancer. 2004;40:939-950.
28. Hu ZY, Yu Q, Pei Q, et al. Dose-dependent association between UGT1A1*28 genotype and irinotecan-induced neutropenia: low doses also increase risk. Clin Cancer Res. 2010;16:3832-3842.
29. Tuchman M, Stoeckeler JS, Kiang DT, et al. Familial pyrimidinemia and pyrimidinuria associated with severe fluorouracil toxicity. N Engl J Med. 1985;313:245-249.
30. Yen JL, McLeod HL. Should DPD analysis be required prior to prescribing fluoropyrimidines? Eur J Cancer. 2007;43:1011-1016.
31. Ferrell PB Jr, McLeod HL. Carbamazepine, HLA-B*1502 and risk of Stevens-Johnson syndrome and toxic epidermal necrolysis: US FDA recommendations. Pharmacogenomics. 2008;9:1543-1546.
32. Martin MA, Klein TE, Dong BJ, et al. Clinical Pharmacogenetics Implementation Consortium guidelines for HLA-B genotype and abacavir dosing. Clin Pharmacol Ther. 2012;91:734-738.
33. Somkrua R, Eickman EE, Saokaew S, et al. Association of HLA-B*5801 allele and allopurinol-induced Stevens Johnson syndrome and toxic epidermal necrolysis: a systematic review and meta-analysis. BMC Med Genet. 2011;12:118.
34. Lennard L, Van Loon JA, Lilleyman JS, et al. Thiopurine pharmacogenetics in leukemia: correlation of erythrocyte thiopurine methyltransferase activity and 6-thioguanine nucleotide concentrations. Clin Pharmacol Ther. 1987;41:18-25.
35. Otterness D, Szumlanski C, Lennard L, et al. Human thiopurine methyltransferase pharmacogenetics: gene sequence polymorphisms. Clin Pharmacol Ther. 1997;62:60-73.
36. Relling MV, Gardner EE, Sandborn WJ, et al; Clinical Pharmacogenetics Implementation Consortium. Clinical Pharmacogenetics Implementation Consortium guidelines for thiopurine methyltransferase genotype and thiopurine dosing. Clin Pharmacol Ther. 2011;89:387-391.
37. O’Donnell PH, Bush A, Spitz J, et al. The 1200 Patients Project: creating a new medical model system for clinical implementation of pharmacogenomics [published online ahead of print August 29, 2012]. Clin Pharmacol Ther.
38. Altman RB. Pharmacogenomics: “noninferiority” is sufficient for initial implementation. Clin Pharmacol Ther. 2011;89:348-350.

Related Articles