Targeted antiviral chemotherapy has greatly reduced the incidence of CMV disease post-HSCT, but adverse side effects still make CMV seropositivity a negative prognostic factor for survival. Immune protection from CMV disease depends on the adaptive immune response, and various strategies for adoptive cellular therapy for CMV have been evaluated in clinical trials to reduce the dependency on antiviral drugs. Characterizing the development of post-transplantation CMV-specific T cell responses may help identify minimum requirements for successful immunologic control of CMV reactivation. We investigated CMV-specific T cell immune reconstitution in 15 patients who all received ACT for CMV 4 weeks post-HSCT. The CMV seropositive donor-derived ACT product consisted of 1x105 CD4+ and CD8+ T cells/kg, generated through 2-week in vitro co-culture with CMV antigen-pulsed autologous dendritic cells. No graft versus host disease >grade II occurred in the study group. CMV-specific T cells were characterized at various time points before, during, and after CMV reactivation events by IFN-γ secretion assay, and/or by phenotyping with HLA CMV pp65 and/or IE-1 peptide pentamers. Three patients who remained CMV PCR negative throughout had no detectable levels of CMV-specific T cells, whereas high levels of CMV-specific T cells developed in the other 12 patients who all experienced episodes of viral reactivation. 11/12 of the patients with CMV reactivation had pre-emptive antiviral chemotherapy, 1/12 controlled CMV reactivation unaided. 3/12 patients experienced a second episode of viral reactivation, which they controlled unaided. At the time of first CMV DNA detection, 1 week before the start of pre-emptive chemotherapy, pentamer+ cells were already detectable in 3/5 patients tested, with absolute numbers ranging from 1150-4900/ml blood. Only after a further 1 to 4 more weeks did T cells secreting IFN-γ in response to in vitro restimulation with CMV peptides first appear in 5 patients tested. By the time viral DNA had become undetectable post-antiviral chemotherapy or post-immunological control of viral replication, pentamer+ cells were present in 8/8 patients, with absolute numbers of pentamer+ T cells >10000/ml in 7/8 patients, a level often quoted as ‘protective’ (median: 37530 cells/ml, range: 3200–139300/ml). CMV-specific CD8 T cells were maintained at high levels long after resolution of CMV viraemia (median: 24200/ml, range: 5100–75900). Absolute numbers of IFN-γ + T cells also reached high levels shortly before to just after the end of a CMV reactivation episode in 5/5 patients tested (CD8+IFN-γ + median: 48490/ml, range: 2150–138290), but subsequently temporarily dropped again to undetectable levels in some patients. Four patients had CMV reactivation episodes that were immunologically controlled unaided. Just prior these four reactivation events there was a total lack of CD8+ T cells functionally responsive to CMV. This was in spite of the presence CMV-specific CD8+ cells by phenotype in 3/3 patients tested, at >10000/ml in one patient. 0% of these pentamer+ cells secreted IFN-γ in response to CMV peptide in vitro. Soon after, by the time of the first CMV PCR+ result CD8+ IFN-γ + T cells were already detected in all patients tested (median: 2180 cells/ml, range: 550–48480, n=3), with up to 64% of pent+ cells secreting IFN-γ. Pent+ cells were above the 10000 threshold in 2/3 patients tested at this time point. These data show that characterizing T cells by phenotype alone does not permit conclusions about their functionality, because some patients experienced CMV reactivation events in the presence of high levels of pentamer+ cells. Viral reactivation episodes however appeared to be associated with a prior lack of CMV-specific CD8+ cells capable of IFN-γ secretion. Immunological control of reactivation, on the other hand, may be associated with the ability to make rapid CMV-induced IFN-γ responses, because only patients with self-resolving CMV infection developed CD8+IFN-γ + responses immediately at the onset of viraemia. However it is not possible to know how many of the patients who received antiviral chemotherapy would have been able to control the infection unaided.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author