Single Nucleotide Polymorphisms (SNPs) Associated with Cognitive Impairment in Patients Treated with Hematopoietic Cell Transplantation (HCT): A Longitudinal Study

Purpose: Pre-HCT, HCT-related, and post-HCT therapeutic exposures place HCT recipients at risk for cognitive impairment, albeit with significant inter-individual variability in the risk, suggesting a role for genetic susceptibility. We hypothesized that the following mechanisms could impact cognitive functioning: impaired capacity to effectively pump genotoxic agents out of the cell, maintain telomere homeostasis, repair damaged DNA, in addition to reduced neural repair capacity and neurotransmitter activity. Methods: We established a prospective longitudinal cohort of patients with hematologic malignancies receiving HCT at a single institution. Comprehensive assessment of 8 domains of cognitive functioning (executive function; processing speed; working memory; auditory memory; visual memory; verbal speed; verbal fluency; and fine motor dexterity) using a battery of 14 standardized tests was performed at predefined time-points: pre-HCT, 6m, 1y, 2y, and 3y post-HCT. Intelligence quotient (IQ) was assessed pre-HCT as a measure of cognitive reserve. Demographic variables (age, sex, race, education, and income) were self-reported and clinical variables (primary diagnosis, conditioning regimen, type of HCT, risk of relapse at HCT, and remission status post-HCT) were abstracted from patient medical records. Blood or saliva for DNA were obtained pre-HCT. We used the Global Deficit Score (GDS), a widely accepted summary score of cognitive impairment, to represent the overall neuropsychological performance. (Carey et al. 2004; Blackstone et al. 2012) GDS ≥0.50 was used to indicate cognitive impairment - a cutoff previously shown to yield an optimal balance between sensitivity and specificity in classifying impairment. Generalized estimating equation models were fitted to test the association of each individual SNP with GDS, adjusting for time of neurocognitive testing, age at HCT, sex, race, pre-HCT IQ, and HCT type. Genetic association analysis involved 985 carefully-curated SNPs in 68 candidate genes. The number of independent tests was 326 (excluding SNPs in Linkage Disequilibrium) yielding an overall p-value threshold of 1.5 x 10-4using Bonferroni correction for multiple testing. Results: The study cohort included 277 patients (58.5% males; 68.6% non-Hispanic whites; median age at HCT: 51.6y, range: 19-73); 148 (53.4%) received an autologous HCT and 129 (46.6%) allogeneic HCT (41.9% matched related and 58.1% matched unrelated). Most common primary diagnoses were non-Hodgkin lymphoma (42.6%) and multiple myeloma (36.5%) in autologous HCT recipients, and acute myeloid leukemia (54.3%) in allogeneic HCT recipients. Factors significantly associated with higher risk of cognitive impairment included: age at HCT ≥50 years (odds ratio (OR)=2.85, 95% CI, 1.3-6.1, p=0.007), male gender (OR=2.62, 95%CI, 1.3-5.1, p=0.005), race other than non-Hispanic white (OR=2.83, 95%CI, 1.4-5.9, p=0.005), and pre-HCT IQ ≤ median (OR=6.58, 95%CI, 3.2-13.3, p<0.0001). Four SNPs were significantly associated with cognitive impairment: rs11837182(OR=4.2, 95%CI, 2.1-8.5, p=4.2 x 10-5) on SLCO1A2 (blood brain barrier);rs330792 (OR=9.2, 95%CI, 4.0-21.1, p=1.2 x 10-7) on MSH6 (DNA repair), rs4725015 (OR=10.7, 95%CI, 4.9-23.5, p=3.6 x 10-9) on RPA3 (DNA repair), and rs16900343 (OR=4.2, 95%CI, 2.0-8.9, p=1.3 x 10-4) on XRCC4 (DNA repair). We also conducted mechanism-specific analyses using the Bonferroni-corrected p-value threshold for each mechanism of cognitive impairment, based on the number of independent SNPs associated with each mechanism. Four additional associations were found to be statistically significant: rs4148734 (OR=6.1, 95%CI, 2.3-16.5, p=3.1 x 10-4) and rs10259849 (OR=6.1, 95%CI, 2.3-16.4, p=3.1 x 10-4) on ABCB1 (blood brain barrier, p-value threshold=1.04 x 10-3), and rs718742 (OR=3.7, 95%CI,1.8-7.7, p=4.3 x 10-4) and rs17152302 (OR=4.0, 95%CI, 1.8-8.6, p=4.1 x 10-4) on PINX1 (telomere homeostasis, p-value threshold=8.7 x 10-4). Conclusion: Our findings provide preliminary evidence for the role of blood brain barrier transporters, telomere homeostasis and DNA repair in the pathogenesis of cognitive impairment after HCT. These findings may help identify HCT recipients at highest risk for cognitive impairment, thus facilitating targeted interventions. Disclosures No relevant conflicts of interest to declare.