Work

Pharmacogenetic association with neurotoxicity in Hispanic children with acute lymphoblastic leukaemia

Year: 2017

Type: article

Abstract: Research shows up to 10-fold variability in the pharmacokinetics of vincristine for patients with differences in genetic make-up and demarcations along ethnic lines (Renbarger et al, 2008). These differences may help to explain not only disparities in the rates of vincristine-related neurotoxicity between ethnicities, but may also play a role in the divergent rates of survival seen in paediatric acute lymphoblastic leukaemia (ALL) (Bhatia et al, 2002). Despite the fact that Hispanics have both the highest incidence and one of the lowest survival rates of childhood ALL compared to other races and ethnicities, there is limited literature on the occurrence of vincristine-related neurotoxicities and the possible pharmacogenomic mechanisms of such events in this population. Disparate small studies which include Hispanic paediatric ALL patients have noted that grades 2–4 neurotoxicity rates are approximately 20–30%, while examination of the most well-established polymorphic genes critical in diminishing vincristine metabolism have shown that Cytochrome P450 (CYP) 3A5*3 and Centrosomal Protein (CEP) 72 T allelic frequencies are approximately 70–80% and 35% respectively (Blanco et al, 2002; Anghelescu et al, 2011; Diouf et al, 2015). To date, no study has examined a large Hispanic population or explored rates of neurotoxicities and well-described pharmacogenomic mechanisms. We report the results of a large, single-institution, retrospective cohort study aimed to better explore the phenomenon of pharmacogenomic-influenced vincristine toxicities among Hispanic children with ALL. Allele discrimination of CEP72 and at the CYP3A5*3, *6, and *7 polymorphic loci was performed using TaqMan® assays. Patient medical records were electronically and manually searched for evidence of neuropathic events. When identified, events were categorized as motor or sensory and graded using the Modified ('Balis') Paediatric Scale of Peripheral Neuropathies (Table SII). Multivariate analysis modelled the influences of CYP3A5 and CEP72 genotypes on the development and time to development of ≥ grade 3 neuropathies. Descriptive statistics were used to identify the prevalence of CYP3A5 and CEP72 genotypes and the associated phenotypes. Two hundred and thirty-nine self-identified Hispanic patients <18 years old treated or followed for a diagnosis of ALL at Texas Children's Hospital, between 1991 and 2015, who received at least one dose of vincristine, had available data in the medical record and had germline DNA available, were included (Fig S1 and Table SI). Based on the CYP3A5 diplotypes, 5% of our Hispanic patients were extensive metabolizers of vincristine (*1/*1), 33% were intermediate metabolizers (*1/*3, *1/*6, *1/*7) and 62% were poor metabolizers (*3/*3, *3/*6, *3/*7) (Table 1a). The CEP72 TT high-risk genotype was a rare finding in our cohort (9·3%; Table 1b). Clinically, 18·4% of our patients experienced ≥ grade 3 neurotoxicity during ALL therapy. Assessing the influence of being both a CYP3A5 intermediate or poor metabolizer and having the TT risk CEP72 genotype was limited by the paucity of patients who shared both genotypes (n = 25). We did note that CYP3A5 poor metabolizers experienced neurotoxicity more often than intermediate or extensive metabolizers, although this phenomenon did not reach statistical significance (Table 1c). However, there was a statistically significant difference in time to development of neurotoxicity within the first 100 days of treatment between intermediate and poor CYP3A5 metabolizers (P = 0·036), with poor metabolizers experiencing ≥ grade 3 neurotoxicity sooner (Fig 1). *1/*3 *1/*6 *1/*7 *3/*3 *3/*6 *3/*7 The incidence of specific CYP3A5 polymorphisms and CEP72 genotype within our patient population was consistent with previous studies. The finding of 18·4% of our patients experiencing ≥ grade 3 neurotoxicity also appears to be consistent with the literature pertaining to paediatric Hispanic ALL patients. While reported rates of grades 2–4 neuropathy in Hispanic patients are roughly 20–30% (14/44 patients in a recent study), one can extrapolate that a focus on ≥ grade 3 neurotoxicity would slightly depress these totals and be in line with our findings (Diouf et al, 2015). Understanding that exact comparisons could not be made with the current published literature, we chose to focus our study on ≥ grade 3 toxicities because clinical modifications for vincristine toxicities are generally made only for patients experiencing ≥ grade 3 neuropathies (Abrahm, 2014). The rate of vincristine-related neuropathy in our Hispanic cohort is less than that reported for Caucasian patients (35–80%) and is significantly greater than what has been reported for African American patients (5%) (Renbarger et al, 2008; Sims, 2016). Our novel finding that genotypic risk portends, not only an absolute risk of neurotoxicity, but a temporal one as well, is important, partly because vincristine is dose-restricted by toxicities and the inability to maximize vincristine exposure may help explain poorer ALL outcomes (Lew et al, 2014). We show the mean cumulative vincristine dose at time of event for intermediate and poor metabolizers was 13·4 [standard deviation (SD) 13·55] mg/m2 and 10·4 (SD 8·89) mg/m2, respectively (Table 1c). This has significant implications, as Hispanic children with ALL who are poor metabolizers could be predicted to have a clinically actionable vincristine-associated neuropathic event between the first and twelfth doses of vincristine, assuming dosing based on most current paediatric ALL protocols (Table 1c). Because vincristine toxicity tends to be reversible and additive, earlier recognition of vincristine toxicity could allow for more tempered and shorter dose reductions, ultimately leading to better morbidity and mortality outcomes. In lieu of such steps, patients at high risk for vincristine neurotoxicities and resultant significant dose reductions could be considered candidates for emerging vincristine alternatives shown to be less neurotoxic, such as vincristine sulfate liposomal injection (Shah et al, 2016). Although our study is one of the largest evaluations of Hispanic ALL patients to date, the results of our study must be interpreted in the context of several limitations, most of which are intrinsic to the study design being a retrospective cohort analysis, and are discussed in detail in the supplementary material. Ultimately, further testing in a larger, more inclusive prospective cohort should be performed, but this study reinforces the significance of CYP3A5 metabolism of vincristine in leading to significant neurotoxicity and suggests that, at least in Hispanic patients, vigilance in detection of early neurotoxicity should be performed by practitioners to assess the need for more tempered dose reductions within the early treatment period in order to promote better outcomes. This study was funded, in part, through the Killian Owen Curing Kids Cancer Scholarship in Developmental Therapeutics (ESS). CAM designed the project, performed research, analysed data and wrote the paper; MBB designed the project, performed research, analysed data and reviewed the paper; AB designed the project and reviewed the paper; JCB designed the project and reviewed the paper; DMD performed research and reviewed the paper; RW performed research and reviewed the paper; MES designed the project, performed research, analysed data and reviewed the paper; ESS designed the project, performed research, analysed data and reviewed the paper. Fig S1. Consolidated Standards for Reporting of Trials (CONSORT) diagram of total patients included and excluded in each analysis. Table SI. Patient characteristics. Table SII. Neuropathies were categorized as motor or sensory and graded using the Modified ('Balis') Pediatric Scale of Peripheral Neuropathies. Data S1. Supplementary material. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Source: British Journal of Haematology

DOI: 10.1111/bjh.14678