Human cytomegalovirus (CMV), Epstein–Barr virus (EBV), and human herpesvirus 6 (HHV‑6) remain frequent causes of morbidity after allogeneic hematopoietic cell transplantation (HCT). Current surveillance relies on quantitative PCR but biomarkers that reflect biological processes preceding reactivation are limited. This exploratory study evaluated whether longitudinal extracellular vesicle (EV) surface marker profiles associate with CMV, EBV, and HHV‑6 reactivation after haploidentical HCT (hHCT) with post‑transplant cyclophosphamide (PT‑Cy), with the goal of identifying candidate liquid‑biopsy indicators for prediction and monitoring.
Thirty‑two consecutive patients undergoing hHCT at Humanitas Research Hospital (Rozzano, Milan) were prospectively enrolled before local introduction of letermovir. Graft‑versus‑host disease (GvHD) prophylaxis included PT‑Cy (50 mg/kg on days +3 and +4), tacrolimus and/or cyclosporin A, and mycophenolate mofetil from day +5. Weekly whole‑blood real‑time PCR assays (commercial R‑GENE kits) were used to monitor CMV, EBV, and HHV‑6. For study analyses, viral reactivation was defined as DNA >1000 IU/mL. Local clinical thresholds for pre‑emptive CMV therapy were >4000 IU/mL; no routine pre‑emptive therapy was given for EBV or HHV‑6 unless clinically indicated.
Peripheral blood samples for EV analysis were collected before transplant (day −6), at transplant (day 0), and at predefined intervals across the first three months post‑HCT (including days +3, +7, +14, +21, +30, +45 and months 2, 2.5, and 3). Plasma was obtained after density‑gradient separation, stored at −80 °C, and processed to remove platelets. EV precipitation, resuspension in RPMI, and a single freeze‑thaw preceded analyses.
EV integrity and preparations were confirmed by transmission electron microscopy, nanoparticle tracking analysis (NTA) for size and concentration, and Western blot to evaluate EV enrichment (TSG101, CD63) and potential contaminants (GM130, ApoB). NTA dilution and bead‑based calibration data indicated the flow cytometry approach preferentially detected larger EVs, with a lower bead‑based threshold estimated around 100–150 nm.
The study reanalyzed an existing longitudinal EV flow cytometry dataset from this cohort. No new antibodies were added; the predefined panel included markers related to immune activation, endothelial biology, and epithelial injury. CD9 and CD81 tetraspanins were measured to identify tetraspanin‑positive EV events, though tetraspanin positivity was not equated with definitive exosomal origin. EV samples (approximately 3 × 10^9 particles) were incubated with FITC‑ or PE‑conjugated antibodies, acquired on a Guava easyCyte instrument, and calibrated with multiple size beads. For each marker, 5,000–10,000 events were recorded and mean fluorescence intensity (MFI), percent positive, and estimated concentration of marker‑positive EVs calculated from replicate measurements.
Associations between EV phenotypes and viral reactivation were evaluated using logistic regression and time‑varying Cox proportional hazards models. Multivariable models adjusted for potential confounders, including acute GvHD (aGvHD) and concomitant viral infections, and a bootstrap‑based assessment was used to evaluate marker‑selection stability across post‑HCT time windows. The pre‑specified reactivation threshold for analyses was >1000 IU/mL for each virus.
CMV reactivation was observed in 20 of 32 patients. Multivariable analyses identified an association between CMV reactivation and increased EV expression (concentration or MFI, as reported in the source dataset) of markers linked to endothelial and receptor biology, specifically CD44, CD86, CD140a (PDGFRα), and CD144. These markers align with known CMV biology — for example, PDGFRα (CD140a) has been implicated in CMV entry and endothelial cells are recognized viral reservoirs associated with CD44 upregulation.
EBV reactivation occurred in 11 patients. EBV reactivation correlated with reduced concentrations of CD30‑positive EVs and increased expression of KRT18 on EVs. The identified pattern suggests distinct epithelial or immune activation signatures on circulating EVs associated with EBV biology in this cohort.
HHV‑6 reactivation was detected in 17 patients. Altered EV profiles associated with HHV‑6 included changes in CD26, CD86, CD120a, and CD144. These markers implicate pathways of immune activation and endothelial/adhesion biology that may reflect HHV‑6 tropism for hematopoietic and endothelial compartments.
Reported associations between EV markers and viral reactivation remained statistically significant after adjusting for aGvHD and concurrent viral infections in multivariable models. The authors also applied bootstrap procedures to assess selection stability across post‑HCT time periods; details on individual marker stability metrics are presented in the source tables. These analyses support that EV marker–virus associations were not solely explained by GvHD or overlapping infections.
The study provides exploratory evidence that distinct circulating EV phenotypes are associated with reactivation of CMV, EBV, and HHV‑6 after hHCT with PT‑Cy. Because EVs may carry host‑derived molecules that reflect early virus‑induced perturbations, EV profiling could potentially serve as a complementary liquid‑biopsy approach to identify patients at higher risk of imminent viral reactivation or to monitor host responses to infection. The observed virus‑specific marker patterns are biologically plausible given known viral tropisms and entry receptors.
As an exploratory secondary analysis of a single‑center cohort (n=32) using a predefined EV panel, findings require external validation. The flow cytometry approach favored detection of larger EVs and may have missed smaller vesicles. The antibody panel was not specifically developed for viral reactivation and no new markers were introduced. Details such as quantitative effect sizes, confidence intervals, and full bootstrap stability metrics are reported in the source article tables and figures and should be reviewed for assessment of clinical performance.
Distinct circulating EV surface marker patterns were associated with CMV, EBV, and HHV‑6 reactivation in this cohort of hHCT recipients treated with PT‑Cy. These exploratory results support further validation of EV immunophenotypes as candidate biomarkers for predicting and monitoring herpesvirus reactivation after transplantation.