Donor NKG2C Copy Number: An Independent Predictor for CMV Reactivation after Double Cord Blood Transplantation

Cytomegalovirus (CMV) is a member of the herpesviridae family and remains latent in the host cells after primary infection. The virus reactivates in the immunosuppressed host and is a leading cause of morbidity following allogeneic hematopoietic stem cell transplant (allo-HSCT). Natural killer (NK) cells provide the first-line defence against viruses and a subset of NK cells expressing NKG2C have been shown to play a role in the immune surveillance of human CMV. Deletion of the NKG2C gene has been reported as a risk factor for viral infections including CMV and HIV, but current data on the influence of NKG2C gene-copy number variations in the donor graft on the risk of CMV reactivation after allo-HSCT is limited. Following cord blood transplantation (CBT), where the T-cell compartment is functionally naïve, the NKG2C genotype may have an even more pronounced impact on the risk of CMV reactivation. Therefore, we studied NKG2C copy number in the donor graft and the risk of CMV reactivation after double umbilical cord blood transplantation (DUCBT) in 100 CMV seropositive DUCBT recipients and their corresponding cord blood grafts (n=200). The gene encoding NKG2C is present at different copy numbers in the genomes of different individuals, with possible genotypes including 2 copies (wt/wt), 1 copy due to heterozygous deletion (wt/del) or 0 copies due to homozygous deletion (del/del). Among the donor CB units, approximately 2/3 had both copies of the gene (wt/wt), 1/3 had only one copy (wt/del) and only a minority of units had 0 copies with both alleles deleted (del/del). In the setting of DUCBT, the combined graft may contain 0 to 4 functional copies of NKG2C gene. In our cohort, patients whose combined grafts contained only 1 or 2 NKG2C gene copies had a significantly higher probability of CMV reactivation than patients whose combined grafts had 3 or 4 NKG2C copies. The 6-month cumulative incidence of CMV reactivation for the 4 groups was 100%, 92.9%, 60.5% and 55.9% respectively (p=0.005). No patient received a graft with zero gene copies. For the rest of the analysis we divided the patients into two groups, namely the 84 patients who received a combined graft with 3 or 4 NKG2C copies and the 16 patients who received two units with 1 or 2 NKG2C copies. The 6-month cumulative incidence of CMV reactivation for the two groups was 58.4% and 93.7% respectively (p=0.0003). In a multivariate analysis, receiving a combined graft with low NKG2C copy number (1 or 2 copies) (HR=2.72, CI=1.59-4.64; p<0.0001) and reduced intensity conditioning (RIC) (HR=0.59, CI=0.35-0.99, p=0.046) were the only independent predictors for CMV reactivation. Interestingly, the NKG2C copy number of the dominant cord was not predictive for CMV reactivation, suggesting that both cord blood units contribute to the antiviral response early post-DUCBT. Our study points to an important role for donor NKG2C in protection against CMV reactivation after DUCBT. If confirmed in larger numbers of CBT recipients, NKG2C genotyping of the CB graft may be a useful biomarker for predicting the risk of CMV infection after CBT, thus guiding the intensity of CMV prophylaxis for individual patients. Moreover these novel findings may provide a compelling rationale for considering NKG2C genotype as a criterion in the algorithm of CB selection for DUCBT.