Immunobiology of Cytomegalovirus Infections in Haploidentical Allogeneic Stem Cell Transplantation with Post-transplantation Cyclophosphamide

To clarify the discrepancies in scientific evidence regarding the incidence of CMV DNAemia and CMV disease in haploidentical hematopoietic stem cell transplantation associated with post-transplant cyclophosphamide (haplo-HSCT), we conducted several analyses. Initially, we determined the incidence of CMV DNAemia and studied its kinetics in haplo-HSCT, comparing it with data obtained from patients undergoing other allo-HSCT modalities. Overall, the cumulative incidence of CMV DNAemia, recurrent CMV DNAemia, and CMV DNAemia requiring preemptive antiviral treatment in patients undergoing haplo-HSCT was comparable to other transplant modalities (P = 0.22, P = 0.13, and P = 0.72, respectively). There was a trend toward longer episode durations and shorter CMV DNA doubling times in patients undergoing haplo-HSCT compared to other transplant modalities. Additionally, the median peak CMV DNA load was significantly higher in haplo-HSCT (P = 0.008), although overall mortality at days 180 and 365 was the same across all comparison groups. Five cases of CMV disease occurred, all in patients undergoing haplo-HSCT. Furthermore, we determined the incidence and characteristics of CMV DNAemia blips in this context. Eighty-three of the 187 patients who presented with DNAemia had one or more blips (n = 104). Blips occurred as a first episode of CMV DNAemia as opposed to prolonged CMV DNAemia (two consecutive positive PCR results) in 47 patients. Blips that presented as initial episodes occurred more frequently (P < 0.001) in patients who received an allograft from a CMV seropositive donor. The cumulative incidence of recurrent CMV DNAemia following initial blips, self-limited prolonged episodes of CMV DNAemia, or episodes treated preemptively with antivirals was not significantly different (P = 0.34). ROC curve analysis indicated that a CMV DNA load cutoff of 48 IU/mL provided the highest combined sensitivity (66%) and specificity (70.2%) for predicting a prolonged episode of CMV DNAemia. We also compared the reconstitution of functional CMV-specific T cells across different allo-HSCT modalities. We monitored CMV DNA load using real-time PCR and determined CD8+ and CD4+ IFN-γ-producing T cells specific for CMV by flow cytometry on days +30, +60, +90, and +180 post-transplantation. CMV DNAemia developed in 62 patients, occurring with comparable frequency in PT/Cy-haplo recipients and other modalities (P = 0.14). There were no significant differences between groups in the number of patients showing detectable CD8+ and CD4+ T cell responses specific for CMV or acquiring levels of CMV-specific T cells conferring protection against subsequent infections. The counts of CMV-specific T cells were comparable between groups at most time points analyzed, regardless of prior development of CMV DNAemia before monitoring. Additionally, we characterized the dynamics of regulatory T lymphocytes in peripheral blood (PB Treg) in patients undergoing haplo-HSCT to investigate if they correlate with the kinetics of reconstitution of CMV-specific T cells and evaluate the potential impact of CMV DNAemia on PB Treg reconstitution. We found that PB Treg counts measured on day +30 post-transplant were comparable (P = 0.63) between patients with (n = 20; median: 3.54 cells/μL; range: 0–24.37 cells/μL) or without (n = 19; median: 2.4 cells/μL; range: 0–29.46 cells/μL) subsequent CMV DNAemia and that PB Treg levels were also comparable (P = 0.34) between patients who developed (n = 17; median: 5.1 cells/μL; range: 0.5–52.7 cells/μL) or did not develop (n = 28; median: 4.6 cells/μL; range: 0–29.46 cells/μL) acute GvHD (grades II–IV). Similarly, no differences were found (P = 0.10) on day +60 between groups. Moreover, we investigated the impact of shared immunodominant HLA alleles between donor and recipient on T cell reconstitution and the incidence of CMV DNAemia in patients undergoing haploidentical transplantation. The cumulative incidence of CMV DNAemia, clinically significant episodes of CMV DNAemia (cs-CMVi), and recurrent CMV DNAemia was comparable between patients with HLA-I compatibility for ID-CMV and those without compatibility (71.8% vs. 80.9%, P = 0.95; 40.7% vs. 44.2%, P = 0.85; 16.4% vs. 28.1%; P = 0.43). The percentage of patients showing detectable IFN-γ-producing T cell responses specific for either CD8+ or CD4+ was similar between groups; however, significantly higher counts of CD8+ T cells specific for ID-CMV were observed in HLA-I compatible patients compared to those without compatibility on days +60 (P = 0.04) and +180 (P = 0.016) post-transplantation. Conclusions: - The incidence of initial and recurrent episodes of CMV DNAemia requiring or not requiring antiviral therapy in unmanipulated haplo-HSCT was comparable to other transplanted modalities within our cohort. - Initiating PET at very low thresholds for CMV DNA should not be common practice since blips resolving spontaneously showing low AUCs occur frequently while the virological course of a CMV DNAemia episode cannot be predicted clinically nor by initial CMV DNA load magnitude. - Data suggest PT/Cy-haplo recipients may reconstitute specific T-cell immunity to an extent similar to patients under HLA-matched allo-HSCT. - The magnitude of reconstituted specific immunity-CMV does not appear related to PB-Tregs enumeration nor does count provide useful information for predicting CMV DNAemia or GvHD. - ID-CMV HLA-I compatibility might impact the magnitude of specific CD8+ pp65/IE-1-CMV T-cell reconstitution but effect seems not to impact initial recurrent incidence nor cs-CMVi.