Variants in PVT1 are Associated with Susceptibility for Hodgkin Lymphoma in Children and Adolescents: A Children’s Oncology Group Study

Few studies have investigated the etiology of pediatric and adolescent Hodgkin lymphoma (HL), the 8^th most common cancer in U.S. children aged <15 years. In economically developed countries, HL in children has distinct demographic and clinical characteristics and risk factor profile from HL in older adolescents and young adults (15-39 years), implying a unique etiology. Established HL risk factors include congenital and acquired immunodeficiency, Epstein-Barr virus (EBV) infection, and family history of hematological cancer. The latter strongly supports a genetic component to HL etiology, although few susceptibility variants have been consistently identified. Linkage studies and genome-wide association studies (GWAS) in adult HL have implicated the major histocompatibility (MHC) region in susceptibility; however, these associations are thought to be insufficient to fully explain the strong familial clustering observed. Single nucleotide variants (SNVs) in PVT1, REL, GATA3, and EOMES have also been identified via GWAS. We hypothesized that these same SNVs would also be associated with risk for pediatric/adolescent HL.

HL cases (n=160) diagnosed in 1989-2003 at 3-17 years of age were enrolled through North American Children’s Oncology Group institutions and genotyped via the Illumina HumanCoreExome array. These data were merged with a convenience control sample of unrelated parents of pediatric/adolescent germ cell tumor cases genotyped using the same array. Genotype clusters were called together and manually reviewed for all SNVs of interest using GenomeStudio. Race/ethnicity was jointly determined by principal components analysis (PCA) using the EIGENSOFT package and HapMap2 and 1000 Genomes Project samples as anchors. Primary analyses were restricted to Hispanic and non-Hispanic white individuals (128 cases and 746 controls). Case genotypes were compared to controls for 18 SNVs previously identified in adult HL GWAS via unconditional logistic regression using an additive model and adjusting for sex and the first two principal components for race/ethnicity, resulting in odds ratios (ORs) and 95% confidence intervals (CIs) to estimate the risk per additional copy of the variant allele. A second control sample pulled from three Minnesota Center for Twin and Family Research studies was also used for comparison. Different arrays were used (Illumina 660W-Quad, HumanExome12 v1.0), so separate PCAs were performed to select only non-Hispanic white case and control samples (95 cases and 3,124 unrelated controls). These analyses were also performed using an additive logistic regression model adjusted only for sex.

The primary analyses revealed significant associations (Bonferroni alpha<0.003) with SNVs in/near PVT1 on chromosome 8 (rs2608053: OR=0.56, 95% CI: 0.41-0.75, P=0.0001; rs2019960: OR=1.49, 95% CI: 1.10-1.99, P=0.008). Comparable ORs were observed for the 3-9 and 10-17 year age groups examined separately. Restricting to non-Hispanic white individuals produced similar results (twin studies controls: rs2608053: OR=0.59, 95% CI: 0.44-0.80, P=0.0007; rs2019960: OR=1.71, 95% CI: 1.26-2.33, P=0.0007), indicating the observation is not attributable to population stratification. PVT1, a non-protein coding oncogene adjacent to and interacting with MYC, is a compelling candidate for further exploration, given that its transcripts are important in the generation of murine T-lineage lymphomas. Additionally, reciprocal translocations involving PVT1 are observed in variant Burkitt lymphoma.

SNVs in the MHC region, including rs6903608, rs204999, and rs2248462, trended toward significance (P<0.05), suggesting that MHC variants are likely associated with pediatric/adolescent HL risk. Unfortunately, we lacked sufficient statistical power to achieve significance for any single marker examined in this region. Importantly, rs204999 has been specifically associated with EBV EBNA-1 antibody levels, supporting a role for MHC variants to impact lymphocyte responses to EBV infection.

We found that some of the SNVs identified in adult HL GWAS are also associated with pediatric/adolescent HL. Additional strategies are needed to identify etiologically relevant variants to uncover the “missing heritability” of this unique malignancy and to understand the significance of MHC variants in imparting susceptibility.