Abstract
Background: Survivors of childhood cancer are at risk for multiple treatment related morbidities, including the early onset of chronic health conditions (CHC) that are late effects of cancer treatment. Epigenetic age acceleration (EAA), i.e., DNA methylation (DNAm)-predicted age in excess of chronologic age, characterizes survivors of childhood cancer vs. controls that have no cancer history and is suggestive of premature, accelerated aging in this population. Further, among these survivors, greater EAA is associated with cardiometabolic, pulmonary, and neurologic late effects of cancer treatment. We examined the relationship between EAA and CHC in a diverse cohort of survivors of childhood leukemia, to better understand the impact of DNAm on outcome disparities.
Approach: The study population included survivors of acute lymphoblastic leukemia (ALL) diagnosed at Texas Children's Hospital between 1992 and 2013 and seen at least once in the Texas Children's Cancer Center Long-Term Survivor Clinic. Demographics, diagnosis, treatment data, and date of onset and type of CHC were abstracted from the electronic health record. CHCs were defined as diagnoses attributed to the patient in the medical record and present for ≥ 1 year. CHCs were categorized by system. Germline DNA was obtained from survivors at least one year off therapy, and analyzed on an iScan System by Illumina MethylationEPIC BeadChip array in the Baylor College of Medicine Lab for Translational Genomics. Methylation data were processed using minfi, normalized with noob, and EAA estimated using the Horvath calculator. Clinical determinants of EAA were identified overall and at specified time periods after the completion of therapy using multivariable linear regression. The relationship between EAA and CHCs were assessed by multivariable Cox regression in a subset of Hispanics and non-Hispanic White (NHW) survivors. All analyses were adjusted for age at diagnosis, age at follow-up, Hispanic ethnicity yes/no, alkylator yes/no, anthracycline yes/no, and cranial radiation therapy (CRT) yes/no.
Results: Cases included 268 survivors of ALL: 174 Hispanic, 156 NHW, 29 Black, 6 Asian, and 3 multi-racial. Younger age at diagnosis was associated with endocrine CHC (OR 0.84, 95% CI 0.70, 1.00) and neuropsychiatric CHC, specifically communication disorder (OR 0.82, 95% CI 0.67, 1.00). Older age at diagnosis was associated with obesity (OR 1.14, 95% CI 1.02, 1.28) and depression (OR 1.16, 95% CI 1.02, 1.33). Hispanics were more likely than NHW to have communication disorder (OR 2.92, 95% CI 1.03, 8.27) and learning disability (OR 2.48, 95% CI 1.03, 5.98). EAA was estimated in 171 samples obtained from 96 Hispanics and 75 NHW that were an average of 9.0 years from diagnosis (range: 1.3-21.9y) and stratified into time periods in survivorship: 3-7 years from diagnosis (n=53), 7-13 years from diagnosis (n=73), and >13 years from diagnosis (n=32). CRT was the only factor associated with EAA, including at both the early (beta 3.30, SE 1.31) and late (beta 5.13, SE 1.71) time points. Although EAA did not differ significantly between Hispanics and NHW, Hispanics exposed to CRT demonstrated a more than 4 year increase in EAA (beta=4.47 EAA years, p-value=0.0002), with a null effect on EAA in NHW (Table 1). The relationship between EAA and CHC differed in early vs. late survivorship. In adjusted analyses, EAA determined from samples obtained 3-7 years from diagnosis was associated with any neuropsychiatric CHC, and specifically with learning disability (HR: 1.44, 95% CI: 1.10, 1.87). In contrast, EAA determined from samples obtained >7 years from diagnosis was associated with a greater risk for obesity (HR: 1.21, 95% CI: 1.04, 1.42) and a lower risk for osteopenia (HR: 0.83, 95% CI: 0.70, 0.98).
Conclusions: Results from this study confirm prior work indicating an association between EAA and risk for CHC in survivors of childhood cancer. In addition, our findings indicate that the relationship between EAA and specific CHC changes in early vs. late survivorship. Moreover, EAA results in Hispanics vs. NHW suggest a differential response to treatment exposures by survivor ethnicity, specifically with respect to CRT. Future directions of this work include validation of these findings in a diverse, multi-institutional survivor cohort.
Table 1: Linear regression model of EAA in Hispanic and Non-Hispanic White survivors of ALL
Disclosures
No relevant conflicts of interest to declare.
