Abstract
Background
Cytomegalovirus (CMV) is a common infectious complication after allogeneic hematopoietic cell transplantation (alloHCT). Efforts to enhance immune reconstitution post-transplant have been pursued to help facilitate clearance of such infections. Assessment of natural killer (NK) cell allo-reactivity with investigation of killer cell immunoglobulin-like receptors has been reported to be associated with protection from CMV infection after alloHCT (Davis ZB et al, BBMT 2015). In addition, the activating NKG2D receptor controls immune responses by regulating NK cells, NKT cells and γδ-T cells. MHC class I chain-related gene A (MICA) is a polymorphic ligand of the NKG2D receptor on these immune effector cells. Given the potential benefit of NK cell allo-reactivity for protection from CMV infection after alloHCT we hypothesized that MICA polymorphisms may influence CMV infection rates after such transplants.
Methods
We conducted a single center, retrospective analysis of allogeneic HCTs for adults with hematologic malignancies in which MICA data were available for donors and recipients. Analysis was restricted to patients with T-cell replete HLA-8/8 matched related or unrelated donor. Fine and Gray regression was used to identify risk factors for CMV infection. The first episode of graft-versus-host-disease (GVHD) was analyzed relative to CMV as a time-dependent covariate. An analysis was performed examining dimorphisms at the MICA-129 position, which previously has been categorized as weaker (valine/valine: V/V), heterozygous (methionine/valine: M/V), or stronger (methionine/methionine: M/M) receptor binding affinity.
Results
From 2000-2016, 423 alloHCT patients were identified who had MICA data. Diagnoses included 197 AML, 82 MDS, 34 ALL, 34 NHL, 22 CML, 12 CMML, 9 CLL, 9 myelofibrosis, 9 plasma cell myeloma, 9 other leukemias, and 6 Hodgkin lymphoma. High, intermediate, and low co-morbidity index was seen in 42%, 33%, and 25% of patients, respectively. Median age at transplant was 52 years (range, 18-76), with 95% Caucasian. A myeloablative transplant was performed in 80% of patients and 52% had a bone marrow graft source. CMV infection occurred in 141 (33%) of patients at a median time of 46 days post-transplant (range, 0-609 days) with 29 (21%) occurring within 30 days, 108 (77%) within 100 days and 33 (23%) after day 100. Thirty-three (8%) patients were MICA mismatched with their donor. Donor MICA-129 dimorphisms included 203 (48%) V/V, 190 (45%) M/V and 30 (7%) M/M. Baseline donor(d)/recipient(r) CMV serostatus for V/V vs. M/V + M/M cohorts were 25% vs. 28% for d+/r+, 11% vs 9% for d+/r-, 39% vs. 37% for d-/r+, and 25% vs. 26% for d-/r- (P=0.75). In univariate analysis, MICA mismatch was associated with a higher risk of CMV (HR 1.64, CI 1.00-2.69, P=0.049), and V/V donor MICA-129 dimorphism with a marginally higher risk of CMV (HR 1.32, CI 0.85-1.83, P=0.10). In multivariable analysis, MICA mismatch was not associated with CMV infection (HR 1.38, CI 0.83-2.29, P=0.22) while V/V donor MICA-129 dimorphism was associated (HR 1.40, CI 1.00-1.96, P=0.05) (Figure 1). Other significant variables in multivariable analysis were year of transplant (HR 0.95, CI 0.92-0.99 P=0.01), non-Caucasian race (HR 2.15, CI 1.18-3.91, P=0.01), high-risk disease (HR 1.62, CI 1.13-2.32 P=0.008), baseline CMV serostatus (HR 7.51, CI 3.76-15.0, P<0.001 for d+/r+, HR 7.73, CI 3.90-15.3, P<0.001 for d-/r+, both relative to d-/r-), and development of GVHD prior to CMV (HR 2.02, CI 1.37-2.96, P<0.001). There was no association of MICA mismatch with acute (HR 1.05, CI 0.66-1.68, P=0.83) or chronic (HR 0.94, CI 0.51-1.76, P=0.85) GVHD. As compared to the V/M+M/M cohort, patients with V/V donors had a longer interval from diagnosis to transplant (median 8.2 vs 6.2 months, P=0.036) and more often received myeloablative conditioning (P=0.030).
Conclusion
We conclude that the donor MICA-129 V/V dimorphism with weak NKG2D receptor binding affinity is associated with increased risk of CMV infection after alloHCT. The presence of at least one M residue encoding allele may confer enhanced NK cell anti-viral reactivity. These observations potentially may have implications in optimizing donor selection. Further investigation of MICA may also help better predict which patients are at higher risk of CMV infection in order to consider intervening sooner with CMV specific therapy or more rapid tapering of immunosuppression.
Figure 1 Figure 1.
Disclosures
Gerds: CTI BioPharma: Consultancy; Incyte: Consultancy. Majhail: Sanofi: Honoraria; Anthem, Inc.: Consultancy.
Human NK Cells in Autologous Hematopoietic Stem Cell Transplantation for Cancer Treatment