This article is more than 5 years old. Information may no longer be current.
Host genetic variants as predictors of chemotherapy-related toxicity in children with ALL: Personalizing the benefit–risk balance
Click Here to Manage Email Alerts
Acute lymphoblastic leukemia is the most common childhood cancer.
Intense rounds of chemotherapy — induction, consolidation and maintenance — have resulted in cure rates of 85% to 90%.
Prolonged daily exposure to dose intense 6-mercaptopurine (6-MP) is the basis of most ALL regimens and critical for the cure of the disease. Nevertheless, dose-limiting toxicities exist, including severe myelosuppression, which may compromise clinical benefit. Similarly, vincristine is highly effective and widely used as part of nearly every child’s ALL regimen.
Jai N. Patel
The dose-limiting toxicity of vincristine is peripheral neuropathy, which can result in significant morbidity and disrupt curative treatment. Given these therapies are indispensable for survival but also may adversely affect quality of life in a large proportion of children, better methods of dose individualization upfront are critical. This column highlights advancements in the discovery and validation of inherited genetic variants as predictors of chemotherapy-related toxicity in children with ALL.
6-MP pharmacogenomics
Thiopurine S-methyltransferase (TPMT) catalyzes the S-methylation and inactivation of 6-MP. It is well known that the TPMT locus is subject to genetic polymorphisms, where approximately 6% to 11% of the population are carriers of the TPMT*2, *3A or *3C alleles, resulting in deficient TPMT activity and increased circulating thioguanine nucleotides. In one study, the percentages of wild-type, heterozygous and homozygous deficient patients able to tolerate full dose 6-MP were 84%, 65% and 7%, respectively. The OR for severe myelosuppression in children with at least one variant allele who receive standard 6-MP dosages is approximately 4, whereas, the OR for children who receive a 30% to 70% dose reduction based on genotype is approximately 1.3.
As such, 6-MP and TPMT genotype-guided dose adjustment represents one of the first examples of cancer pharmacogenomic implementation in clinical practice (see Clinical Pharmacogenetics Implementation Consortium guidelines at www.pharmgkb.org/guideline/PA166104945). However, inter-patient tolerability of 6-MP is not fully explained by TPMT variations alone, as a substantial proportion of patients with sufficient TPMT still experience excessive toxicity that may disrupt treatment and compromise therapeutic activity. In fact, children of East Asian descent are more intolerant to full dose 6-MP, even though TPMT deficiency is less common in East Asians compared with children of European descent.
Yang and colleagues performed a genome-wide association study (GWAS) to further identify the genetic basis for 6-MP intolerance in children. Investigators genotyped a discovery cohort of 657 children with ALL — as well as a replication cohort of 371 children — for more than 600,000 single nucleotide polymorphisms (SNPs). The results affirmed that variations in TPMT were a major predictor of 6-MP toxicity (rs1142345; P = 8.6 x 10-9). Importantly, a novel variant in the NUDT15 gene also was significantly associated with 6-MP intolerance (rs116855232; P = 8.8 x 10-9). NUDT15 is shown to act as a safeguard by degrading oxidized purine nucleoside triphosphates to prevent incorporation into DNA. Patients with the TT, TC and CC genotypes had average dose intensities of 8.3%, 63% and 83.5%, respectively. Of those who were homozygous for either TPMT or NUDT15, or heterozygous for both, 100% required at least a 50% dose reduction, compared with 7.7% of other children.
Importantly, the NUDT15 variant was found to be prevalent in the East Asian population (9.8%) and Hispanic population (3.9%), compared with 0.2% in Europeans. This is of particular interest, as previous discovery studies primarily have been conducted in patients of European descent. The study by Yang and colleagues provided large enough sample sizes with different ethnicities to adequately power and identify novel variants not previously elucidated. However, East Asian descent remained associated with 6-MP dose intensity after adjustment for the NUDT15 variant, suggesting there may be additional factors contributing to 6-MP intolerance in this population. As such, additional discovery and validation studies are needed to both replicate the findings by Yang and colleagues and to discover other novel variants that contribute to 6-MP intolerance.
Vincristine pharmacogenomics
Vincristine is preferentially metabolized via cytochrome P450 3A5 (CYP3A5). The CYP3A5*1 allele is required for functional activity, whereas the most common variants resulting in little to no enzyme activity are CYP3A5*3, *6, and *7. CYP3A5*1 demonstrates significant inter-ethnic variation between black patients (70% prevalence) and white patients (20% prevalence). In fact, the incidence (35% vs. 5%, respectively; P = .007) and severity (2.72 vs. 1, respectively; P < .0001) of neurotoxicity was found to be significantly higher in white patients compared with black patients. A subsequent study identified that children with precursor B-cell ALL expressing CYP3A5*1 experienced significantly less vincristine-induced peripheral neuropathy (VIPN).
More recently, Diouf and colleagues performed a GWAS in two independent cohorts of patients enrolled in two ALL treatment protocols (n = 222 and n = 99) to identify genetic variants associated with the occurrence or severity of VIPN. Investigators identified a common SNP in the promoter region of CEP72, which encodes a centrosomal protein involved in microtubule formation, as significantly associated with VIPN (P = 6.3 x 10-9). Roughly 56% of TT patients and 21% of CC/CT patients experienced at least one episode of grade 2 to grade 4 neuropathy (P = 2.4 x 10-6). The severity of neuropathy was 2.7-fold greater in TT patients compared with CC/CT patients. The HR for the risk for developing neuropathy was 3.58 (95% CI 2.1-6.1) after adjustments for vincristine dose and genetically determined ancestry (P < .001). Lastly, investigators determined that those with the TT genotype had significantly lower CEP72 mRNA expression compared with CC/CT patients (P = .03), which resulted in increased sensitivity to vincristine in neurons and leukemia cell lines.
These data suggest that leukemia cells of patients with the TT genotype are more sensitive to vincristine, and thus lower doses of vincristine may be administered to decrease the risk for or severity of neuropathy without compromising the anticancer effect; however, future trials addressing this question, in addition to the combined influence of CYP3A5 variations, are warranted.
Conclusion
The primary focus of cancer drug development and research has been on enhancing tumor control, which has resulted in prolonged life for patients with cancer. However, as therapies continue to extend survival, addressing toxicity and quality of life becomes critical. Further, reducing adverse effects to optimize dose intensity of chemotherapy allows for a greater chance of clinical benefit and cure.
The recent studies by Yang and colleagues and Diouf and colleagues highlight the importance of pharmacogenomic translational research, as well as the need to identify and validate genetic predictors of toxicity to allow for better drug and dose selection. Although vincristine and 6-MP are key components of ALL regimens, these studies suggest dose reductions may be possible in the right population, without compromising efficacy. Ideally, prospective clinical trials addressing this question are warranted to prove whether routine genotype-guided dosing should be implemented in clinical practice. However, as more patients are undergoing sequencing of both tumor and germline DNA, pharmacogenetic results may soon be readily available. Clinicians and researchers must be aware of pharmacogenetic variants as potential predictors of toxicity and use the data available to personalize the benefit–risk balance for every patient.
References:
Childhood ALL Collaborative Group. Lancet. 1996;347:1783-1788.
Diouf B, et al. JAMA. 2015;doi:10.1001/jama.2015.0894.
Egbelakin A, et al. Pediatr Blood Cancer. 2011;doi:10.1002/pbc.22845.
Kham SK, et al. Eur J Clin Pharmacol. 2008;doi:10.1007/s00228-007-0426-x.
Knepper TC and McLeod HL. J Clin Oncol. 2015;doi:10.1200/JCO.2014.60.1740.
Koren G, et al. N Engl J Med. 1990;323:17-21.
McLeod HL, et al. Leukemia. 2000;14:567-572.
Patel JN. Pharmacogenet Genomics. 2015;doi:10.1097/FPC.0000000000000134.
Pui CH, et al. N Engl J Med. 2006;354:166-178.
Relling MV, et al. Blood. 1999;93:2817-2823.
Relling MV, et al. Clin Pharmacol Ther. 2013;doi:10.1038/clpt.2013.4.
Relling MV, et al. J Natl Cancer Inst. 1999;91:2001-2008.
Renbarger JL, et al. Pediatr Blood Cancer. 2008;50:769-771.
Stanulla M, et al. JAMA. 2005;293:1485-1489.
Stocco G, et al. Clin Pharmacol Ther. 2009;doi:10.1038/clpt.2008.154.
Yang JJ, et al. J Clin Oncol. 2015;doi:10.1200/JCO.2014.59.4671.
For more information:
Jai N. Patel, PharmD, is chief of pharmacology research and phase 1 trials at Levine Cancer Institute at Carolinas HealthCare System, as well as adjunct assistant professor at UNC Eshelman School of Pharmacy. He also is a HemOnc Today Editorial Board member. He can be reached at jai.patel@carolinashealthcare.org.