Purpose. Cytomegalovirus (CMV) reactivation after transplant occurs frequently and is generally thought to be associated with an increased transplant-associated mortality. We aimed first to evaluate the cytotoxic effects of CMV on AML in vitro, and further, the clinical outcome of patients with AML with a documented CMV reactivation after transplant. We hypothesized that CMV directly affects AML cells thereby reducing the risk for leukemic relapse in AML-patients after transplant.

Patients and Methods. We retrospectively evaluated 236 patients with AML in two cohorts (108 patients in the 1st cohort transplanted from an HLA-identical sibling donor, and 127 patients in a 2nd cohort transplanted from an HLA-identical unrelated donor). All patients were transplanted in Essen. Endpoints of the study were probabilities of relapse and overall survival (OS). For the definition of a CMV reactivation treatment with ganciclovir and the detection of pp65 positive cells in peripheral mononuclear cells at minimum at two different occasions were required. For in vitro studies we used the cell lines Kasumi-1 (AML-M2), SD-1 (ALL) and K562 (CML in blast crises) and infected them with the AD169 CMV strain. 14 days after infection we evaluated the apoptosis rate by measuring annexin V by FACS, the proliferation rate by BRDU assay, the expression of disease-specific marker (AML1/ETO, BCR-ABL), pro-and antiapoptosis marker p21, c-myc by real-time PCR in infected cells and controls.

Results. Infection of CMV in Kasumi-1 cells (AML) induced in 99.8% of cells apoptosis. Apoptosis was induced also in SD-1 cells (ALL) in 31.3% of cells after CMV infection, whereas no difference in the apoptosis rate was seen for K562 cells (CML) after CMV infection (6.4%) compared to uninfected controls (9.0%). These results were in concordance with the clinical observation of a CMV reactivation after transplant in AML, ALL, and CML. In 25.9% of AML- patients of cohort 1 and in 29.9% of AML-patients of cohort 2 we detected a CMV reactivation. AML-Patients with a documented CMV reactivation had in both cohorts a markedly reduced risk for leukemic relapse (probability of risk for relapse at 5-year 9.2% versus 52.6% in cohort 1 (p<0.0005) and 12.2% versus 47.2% in cohort 2 (p<0.013), which resulted also in a significant improved outcome after transplant (cohort 1 OS 73.6% versus 42.5% (p<0.04) and 48.0% versus 33.5% in cohort 2 (p<0.04). There were no significant differences in the characteristics of patients with or without CMV reactivation as age, disease stage, donor type or conditioning regimens. Patients with CMV reactivation had a higher incidence of acute GVHD grade 2 – 4 (82% versus 38%, p <0.0001), but this was also seen in patients with ALL and CML who did not have a reduced risk for leukemic relapse after a CMV reactivation. When stratified according to the occurrence of acute GVHD grade 2–4, the probability for risk of relapse rate remained significant lower in patients with a CMV reactivation compared to their counterparts (p<0.003). In multivariate analyses, age of patients, acute GVHD grade2–4, chronic GVHD, disease stage, HLA-constellation between donor and recipient, donor type (unrelated versus sibling), sex-constellation between donor and recipient were tested in each cohort as well as combined in both cohorts together. The only significant factors impacting the outcome and relapse rate were disease stage, chronic GVHD and CMV reactivation.

Conclusions. We provide strong evidence that the occurrence of a CMV reactivation after HSCT is associated with a reduced risk for leukemic relapse in patients with AML. Our data indicate an unrecognized role of CMV reactivation after HSCT for AML, which is probably mediated by CMV induced apoptosis.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author