PROK1 locus variant identified as a genetic determinant of anthracycline-related cardiomyopathy in long-term childhood acute leukemia survivors

Abstract Introduction Survival rates for childhood and adolescent acute leukemia have significantly improved over the past few decades, reaching 60% for acute myeloid leukemia and 90% for acute lymphoblastic leukemia. However, childhood leukemia survivors face increased morbidity and mortality due to treatment-related late effects, which accumulate over time. Among these complications, anthracycline-related cardiomyopathy is a severe adverse effect that significantly impacts survivors' quality of life and can jeopardize their vital prognosis again. Despite known risk factors, such as second-line treatment after relapse or allogeneic hematopoietic stem cell transplantation (HSCT), substantial inter-individual variability in the development of cardiomyopathy persists, suggesting a critical role for genetic susceptibility. While several studies have investigated genetic factors underlying anthracycline-related cardiomyopathy in childhood cancer survivors, few have focused specifically on homogeneous cohorts of pediatric leukemia survivors. Notably, despite insights from candidate gene studies, no genome-wide association study (GWAS) was performed exclusively in this population. This study aims to identify genetic determinants associated with cardiomyopathy in a homogeneous cohort of pediatric leukemia survivors from the French LEA program. Methods LEA is the French national program dedicated to the study of long-term complications following treatment for childhood leukemia. It includes over 7,500 patients treated since 1980 across 19 French pediatric oncology centers. Cardiomyopathy was systematically assessed through clinical evaluation and echocardiography during dedicated long-term follow-up visits occurring every two to four years, and was defined by a left ventricular fractional shortening <28%, a left ventricular ejection fraction <55%, the initiation of cardiomyopathy treatment or cardiac transplantation. The LEA biobank was established in 2018 to collect biological samples. For this project, 794 samples with sufficient DNA quantity and quality were typed for genome-wide common single nucleotide polymorphisms (SNPs). Genotyping was performed with Illumina Global Screening Array Multi-disease (GSA-MD V3.0) at the French national center for human genome research (CNRGH). Genotype data were imputed with Minimac4 using the 1000 Genomes phase 3 version 5 reference panel. Individuals with at least one of the following criteria were excluded: discordant sex information, relatedness, genotyping call rate lower than 95%, heterozygosity rate higher/lower than the average rate ±3 standard deviation or of non-European ancestry. To identify common genetic variants associated with cardiomyopathy, we conducted a GWAS using a Cox proportional hazards model considering delay since leukemia diagnosis as time scale, implemented in R package survival, adjusted for sex, age at leukemia diagnosis, leukemia type, HSCT and the first four principal components derived from genetic data. Results We analyzed 7,015,704 SNPs with minor allele frequency (MAF) >5% and imputation quality score (Rsq) >0.3 in 723 leukemia survivors (44 with cardiomyopathy, 679 without cardiomyopathy). No variant reached genome-wide significance (p<5x10-8), but several loci showed suggestive associations (p<1x10⁻5) and were prioritized for further investigation. The variant rs74117657 was significantly associated with cardiomyopathy (Hazard Ratio (HR)=6.48, p=3.93x10-9) in the subgroup of patients who had not undergone HSCT (n=605, 30 with cardiomyopathy). This variant is located 7 kb downstream PROK1 gene and is not in linkage disequilibrium (r2>0.80) with other nearby SNPs. PROK1 encodes Prokineticin-1, a protein involved in angiogenesis and cardiovascular homeostasis, which has been involved in cardioprotective mechanisms in response to stress (Nebigil et al., 2009), including anthracycline-induced cardiotoxicity (Gasser et al., 2019). Interestingly, this variant was significantly associated with circulating levels of Prokineticin-1 (p=9.17x10-9) in the UK Biobank study (Sun et al., 2023). Ongoing work includes the measurement of Prokineticin-1 levels in patient plasma and gene-environment analysis to assess interactions between SNPs and anthracycline exposure. Results from these analyses will be available for presentation at the meeting. Conclusion This study suggests PROK1 as a new locus for anthracycline-related cardiomyopathy in long-term childhood acute leukemia survivors.