Impact of letermovir on cytomegalovirus-specific T-cell reconstitution after allogeneic hematopoietic stem cell transplantation in the post-transplant cyclophosphamide era

The approval of letermovir (LTV) in prophylaxis after allogeneic-hematopoietic stem cell transplantation (allo-HSCT) has reduced the incidence of cytomegalovirus (CMV) clinically relevant reactivations (CRE) in the early post-HSCT period, postponing these events after LTV cessation in high-risk patients.^1-3 T-cell reconstitution is impaired in LTV recipients, and levels of both polyclonal and CMV-specific T cells are restored only some months after LTV treatment.^4-8 Since the presence of CMV-specific T cells correlates with protection against CRE,^9-13 this delay potentially exposes patients to viral reactivations after LTV discontinuation.

To identify the factors stimulating the emergence of protective CMV-specific immunity during or immediately after LTV treatment, we analyzed the dynamics of CMV-specific CD8⁺ T-cell reconstitution in patients receiving allo-HSCT with post-transplant cyclophosphamide (PT-Cy) and calcineurin inhibitor (CNI)-free graft-versus-host disease (GvHD) prophylaxis. Additionally, we explored the correlation between reconstitution of CMV-specific T cells and key clinical outcomes, such as the incidence and severity of CMV reactivation and of acute and chronic GvHD (aGvHD and cGvHD). Understanding whether CMV-specific T cells can serve as a predictive biomarker may have significant implications for risk stratification and personalized management of post-HSCT patients.

In a single-center retrospective observational study, we analyzed 42 CMV-seropositive adult patients with hematological malignancies undergoing allo-HSCT with sirolimus. All patients were treated according to institutional guidelines.¹⁴ GvHD prophylaxis was based on PT-Cy, except in patients receiving graft from cord blood. Anti-thymocite globulin was not used. The conditioning regimen was treosulfan-based; an intensified conditioning with the addition of a second alkylating agent was classified as myeloablative.¹⁴

Fifteen patients (36%) were treated with allo-HSCT from 2019 to 2022 and received 480 mg of LTV orally once daily from day (D)0 until D100 after transplant as CMV prophylaxis. The control group (no-LTV) was composed of 27 (64%) patients transplanted from 2016 to 2018, before LTV approval. Only patients with peripheral blood mononuclear cells (PBMC) available at 90 (D90) and 180 (D180) days after HSCT and whose donors expressed at least one HLA evaluable with Dextramer^® CMV-Kit were enrolled. All patients signed written informed consent according to the principles of the Declaration of Helsinki and the non-interventional ALMON study approved by the Milan San Raffaele Hospital Institutional Ethical Committee in 2007. Samples were processed and stored by the institutional Biological Resource Center (Num_ID in BBMRI-ERIC: bbmri-eric:ID:IT_1383758011993577).

Cytomegalovirus DNAemia was quantified weekly by real-time polymerase chain reaction (PCR) on whole blood samples until D100, twice weekly in case of positivity. Subsequently, patients were routinely monitored for CMV viremia and/or presence of CMV-related symptoms: in detail, patients with no CMV reactivation or CMV-related symptoms were monitored for CMV viremia every 2-3 weeks until D180, as per institutional guidelines.⁵ CRE was defined as onset of CMV disease or initiation of pre-emptive therapy, based on CMV DNAemia above 10,000 copies/mL and the CMV-infection risk of the patient, or repeated positive samples showing increased viral load during LTV prophylaxis.¹³ CMV blips were defined as the detection of a single positivity for CMV DNA with subsequent spontaneous clearance.¹⁵ Patients’, transplant and immune characteristics are detailed in Table 1. CMV-specific CD8⁺ T cells were quantified by flow cytometry using the In Vitro Diagnostic Regulation (IVDR) Dextramer® CMV-Kit (Immudex), as previously described,¹⁰ in PBMC frozen at D90 and at D180. Dextramer reagents are composed of multiple MHC-peptide complexes and fluorophores covalently bound to a dextran backbone. This tool allows the evaluation of CMV-specific immunity with a good correlation with functional assays and higher sensitivity.¹⁰ Data were considered evaluable if at least 15,000 events in the CD3⁺CD8⁺ region were acquired. Absolute counts of CMV-specific CD8⁺ T cells were calculated based on the percentage of CD3⁺CD8⁺Dextramer⁺ lymphocytes in frozen samples and on the absolute counts of CD3⁺ T cells in corresponding whole fresh blood. Protective anti-viral immunity was defined based on the threshold of 0.5 CMV-specific cells/μL, which we had previously identified as an independent protective factor from subsequent CRE.¹⁰ Absolute CD3⁺ counts and percentages of CD8⁺ T cells were similar in the two groups at both timepoints, with a trend for a slower CD3⁺ reconstitution in LTV patients, in agreement with previous observations.^4,5,8 Only a minority of patients (N=2 at D90 and N=4 at D180) were treated with a dose of steroids with a potential impact on T-cell amount and functionality. As expected, the number and incidence of both CRE and overall CMV reactivations within D180 was lower in LTV compared with no-LTV patients (CRE, P=0.007; CMV reactivations, P=0.008) (Table 1, Online Supplementary Figure 1A, B). In agreement with our previous data in a wider cohort,⁵ in patients receiving LTV prophylaxis, we also observed a reduction in moderate-to-severe cGvHD (P=0.009) and a trend towards a lower incidence of any grade aGvHD (P=0.06) and cGvHD (P=0.07) (Online Supplementary Figure 1C-E).

Table 1.Patients and transplant characteristics (total number of patients = 42).

In our cohort, absolute numbers of CMV-specific CD8⁺ T cells were enriched from D90 to D180 (P=0.02) (Online Supplementary Figure 2A). Analyzing separately the two different CMV prophylaxis groups, this increase was evident only in patients receiving LTV (P=0.03) (Figure 1A). Accordingly, the absolute numbers of CMV-specific T lymphocytes and the percentages of patients harboring protective levels of CMV-specific CD8⁺ T cells were lower in LTV compared to no-LTV patients at D90 (P=0.009 and P=0.04 respectively), but these values are comparable in the two groups at D180 as a consequence of CMV-specific immunity restoration in LTV patients (Figure 1B, C). Interestingly, the incidence of cGvHD (both overall and moderate-to-severe) was not affected by the presence of CMV-specific T cells (Figure 1D, Online Supplementary Figure 2B) but was increased in patients experiencing CMV reactivations (including not-CRE) before either D90 or D180, especially when considering only moderate-to-severe cGvHD (Figure 1E, Online Supplementary Figure 2C). These data suggest that, while CMV-specific lymphocytes do not seem to be involved in driving cGvHD onset and progression, its occurrence might be influenced by the increased inflammation associated with viral reactivation. Larger-scale investigations are needed to substantiate this hypothesis.

In bivariate analyses, the variables associated with the presence of protective CMV-specific immunity at D90 were: absence of LTV prophylaxis, the percentage of CD8⁺ T cells, and the number and occurrence of CMV reactivations (including those not-clinically relevant) before D90. (See Table 2 for further details.) For D180, statistically significant correlation was found only for viral reactivations before D180, with a trend towards positive correlation with absolute CD3⁺ T-cell counts and donor CMV serostatus (Table 2). Of note, CMV-specific T cells were not affected by either the type of transplant or the intensity of the conditioning regimen, which are the other two variables that are significantly different between LTV and no-LTV groups (Tables 1 and 2). Multivariate binomial logistic regression confirmed for both timepoints a positive correlation between the presence of CMV-specific T cells above the protective threshold and previous CMV reactivation, regardless of their number (Table 2 and data not shown). Interestingly, at D180 this correlation was significant only including the occurrence of CMV blips in the analysis: D90, no blips P=0.010 and with blips P=0.007; D180, no blips P=0.159 and with blips P=0.028. Furthermore, at D90, also the percentage of CD8⁺ T cells was significantly correlated with CMV-specific immunity (P=0.025), which is possibly dependent on the type of assay, while in this cohort with good immune-reconstitution, absolute T-cell counts had no impact (Table 2).

In this study we quantified CMV-specific CD8⁺ T cells by a flow cytometry IVDR kit. Compared to the detection of interferon(IFN)-γ producing CMV-specific T cells by flow cytometry, ELISpot or QuantiFERON, previously used to evaluate CMV-specific immunity in this setting,^6-8,16 the use of this kit strongly reduces the analysis time and is characterized by standardization, ease of use, and high reproducibility, in agreement with the recent European Regulation IVD-R 2017/746. Furthermore, we have recently demonstrated that the quantification of CMV-specific T cells by Dextramer staining correlates with functional assays such as IFN-γ ELISpot and QuantiFERON.¹⁰

The GvHD prophylaxis regimen in this cohort was CNI-free. Instead of the more traditional tacrolimus or cyclosporine backbones, it relied predominantly on PT-Cy (omitted only in patients receiving cord blood units as graft source), sirolimus and mycophenolate mofetil (MMF). The strategy of omitting CNI aims to reduce nephrotoxicity, neurotoxicity and metabolic complications, while still harnessing the potent in vivo T-cell modulation activity of PT-Cy and the synergistic immunoregulation of sirolimus plus MMF to prevent both acute and chronic GvHD.

Figure 1.Dynamics of cytomegalovirus-specific T cells differ according to letermovir treatment without impact on moderate-to-severe chronic graft-versus-host disease. (A) Longitudinal variation in the absolute values of cytomegalovirus (CMV)-specific CD8⁺ T cells between Day (D)90 and D180 in patients receiving (right) or not receiving (left) letermovir (LTV) as CMV prophylaxis. Only patients with evaluable samples (≥15,000 CD8⁺ T cells acquired) at both D90 and D180 were included. Means of CMV-specific T cells: no-LTV, 7.8 cells/μL at D90 and 26.0 cells/μL at D180; LTV, 0.5 cells/μL at D90 and 9.6 cells/μL at D180. Statistical analyses by Wilcoxon test. (B and C) Comparison of the absolute counts of CMV-specific lymphocytes (B) and of the percentages of patients with levels of CMV-specific T cells above the protective threshold of 0.5/μL (C) at D90 and D180 after allogeneic hematopoietic stem cell transplantation (allo-HSCT) between patients treated or not treated with LTV. Only patients with evaluable samples (≥15,000 CD8⁺ T cells acquired) at either D90 or D180 were included. Means of CMV-specific T cells in (B): D90, 7.8 cells/μL in no-LTV and 0.5 cells/μL in LTV; D180, 25.0 cells/μL in no-LTV and 11.3 cells/μL in LTV. Statistical analyses by Mann-Whitney test in (B) and X² test in (C). (D) Incidence of moderate-to-severe chronic graft-versus-host disease (cGvHD) according to the presence of protective levels of CMV-specific T cells (DEX⁺ ≥0.5/μL) at D90 (left) or D180 (right). (E) Incidence of moderate-to-severe cGvHD according to the occurrence of CMV reactivations before D90 (left) or D180 (right). (D and E) Statistical analyses by Log-rank test. Lines indicate censored data.

Table 2.Correlation between the presence of protective cytomegalovirus-specific T cells and clinical/biological variables.

Our findings reinforce the protective role of LTV against clinically relevant CMV-CRE and confirm that LTV prophylaxis is associated with a delayed reconstitution of CMV-specific CD8⁺ T cells compared to no-LTV patients. Importantly, our data underscore the pivotal role of antigen exposure (even transient and at low levels, such as during CMV blips) in promoting the expansion of protective levels of CMV-specific T lymphocytes. This suggests that minimal antigenic stimulation is sufficient to boost protective CMV-specific immune responses in the context of ongoing reconstitution. Moreover, in this cohort, 4 patients (N=3 LTV; N=1 no-LTV) achieved protective levels of CMV-specific T cells by D180 despite the absence of previous CMV positivity (data not shown). Although we cannot exclude transient CMV reactivation escaping our monitoring schedule, this observation suggests that, while antigen exposure facilitates expansion, the recovery of CMV-specific immunity is also dependent on the broader trajectory of immune reconstitution, which improves at late post-transplant timepoints. A limitation of the method used in this study is that only patients with good immune reconstitution can be evaluated, due to the requirement of at least 15,000 CD3⁺CD8⁺ T cells. These findings have several implications. Firstly, a clearer definition of the stimuli that promote CMV-specific T-cell emergence (whether through subclinical antigen exposure or endogenous immune recovery) could help refine patients’ stratification strategies. Our work supports the potential utility of combining longitudinal immune monitoring with a CMV-specific Dextramer-based assay to identify patients with suboptimal CMV-specific immune reconstitution. Moreover, in the setting of extended prophylaxis that is currently being explored in clinical trials, Dextramer-based assays could distinguish those patients that could safely shorten the duration of therapy or who, on the other hand, would need even longer prophylaxis, allowing for better protection during the window of immune vulnerability.

Additionally, consistent with prior observations,⁵ our data suggest that the incidence of GvHD is lower in patients receiving LTV, but increases after CMV reactivation. This supports the hypothesis that reduced CMV reactivation and associated inflammation may contribute to a more favorable GvHD profile. Furthermore, our data provide indirect evidence that CMV-specific T cells are not major contributors to the pathogenesis of cGvHD, since we observed no increased risk of cGvHD in patients with higher levels of CMV-specific T cells.

Taken together, our study provides new insights into the dynamics of CMV-specific immunity under LTV prophylaxis and highlights actionable items for personalized antiviral strategies after allo-HSCT.

Footnotes

• Received May 26, 2025
• Accepted August 26, 2025

Correspondence

*ET and GO contributed equally as first authors.

^#CB and RG contributed equally as senior authors

R. GRECO - greco.raffaella@hsr.it

E. TASSI - tassi.elena@hsr.it

Disclosures

CB has been member of Advisory Board and Consultant for Intellia, TxCell, Novartis, GSK, Allogene, Kite/Gilead, Kiadis, Evir, Janssen, Genyo, Epsilen, and Alia, and received research support from Intellia Therapeutics. She is inventor of different patents on cancer immunotherapy and genetic engineering. RG declares speaker honoraria for educational events from Biotest, Pfizer, Medac, Quiagen, Kyverna MDS and Magenta, and advisory board from Pfizer and BMS. All the other authors have no conflicts of interest to disclose.

Contributions

Funding

CB is supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC 5 per Mille 22737 and AIRC-IG 18458), Italian Ministry of Research and University (PRIN 2017WC8499, 2022SLL3YZ), Italian Ministry of Health (Research project on CAR-T cells for hematological malignancies and solid tumors and RF-2019– 12370243; “Uncover and overcome senescence and dysfunction of genetically engineered T lymphocytes for cancer immunotherapy” RF-2021-12373598), EU (T2Evolve, Join4ATMP).

Acknowledgments

The authors would like to thank the Ospedale San Raffaele Hematology Trial Office, the participating patients and their families, and all nurses and data managers who contributed to this study.