Influence Of Genetic Variation On Metabolic Outcomes Among Survivors Of Pediatric Hematopoietic Cell Transplantation

Childhood hematopoietic cell transplant (HCT) survivors have been shown to be at increased risk for a variety of long-term cardiometabolic sequelae. However, it is unclear if genetic factors can modify the levels of various quantitative cardiometabolic traits in this population.

Prospective cross-sectional study of ≥2-year survivors of pediatric allogeneic (n=47) or autologous (n=27) HCT recruited from two centers. Allogeneic HCT survivors had to be free of chronic GVHD and off immunosuppression at time of assessment. All participants underwent detailed physiologic measurements (euglycemic insulin clamp, body composition assessments, blood pressure, fasting lipids) and also provided DNA samples (both donor and host if allogeneic HCT recipient). A panel of carefully annotated genes and respective single nucleotide polymorphisms (SNPs) associated with each cardiometabolic trait (fat mass, body mass index, insulin resistance, triglyceride, HDL and LDL cholesterol) was assembled based on evidence from the general population using the Human Genome Epidemiology (HuGE) Navigator (accessed March 2013). For each trait, we estimated the least square mean values and determined the overall and pairwise p-values among genotypes. Multiple comparisons were controlled for using Benjamini and Hochberg’s false discovery rate. For initial validation, a separate dataset of 134 ≥2-year pediatric allogeneic HCT survivors without active chronic GVHD and retrospectively collected phenotypes was available.

Among the primary study participants (n=74), 73% were exposed to total body irradiation, and their median current age was 24 years (range 10-50), with median interval since HCT of 11.4 years (range 2.6-27.6). 95% of these survivors self-reported White non-Hispanic race/ethnicity. Host polymorphisms in LPL (rs319; p=0.0009) and ADRB2 (rs12654778; p=0.002) were found to be significantly associated with differences in triglyceride levels and body mass index, respectively. While several other genes (all host) had SNPs that were associated with various outcomes (e.g. ADIPOQ and insulin resistance, CETP with both HDL and LDL-cholesterol; all p<0.01), none met statistical significance as defined by the false discovery rate (Table). No donor polymorphisms was found to be associated. Among the validation sample (74% exposed to total body irradiation, 100% White non-Hispanic), host rs319 had no association with triglyceride levels (p=0.81). An effort to validate the other polymorphisms listed above is ongoing.

It is possible that underlying genetic factors may influence subsequent development of adverse metabolic traits, even among heavily treated childhood cancer survivors. However, any associations require further validation. Furthermore, the relatively small numbers of such patients at most centers limit the power of such genetic investigations. Future efforts should focus on developing appropriate consortia in which samples may be pooled. Application of rigorous phenotype definitions may also enhance the likelihood of identifying true genetic associations.

No relevant conflicts of interest to declare.