Cytomegalovirus-specific T-cell reconstitution following letermovir prophylaxis after hematopoietic cell transplantation

Abstract

Decreased cytomegalovirus (CMV)-specific immunity after hematopoietic cell transplantation (HCT) is associated with late CMV reactivation and increased mortality. Whether letermovir prophylaxis-associated reduction in viral exposure influences CMV-specific immune reconstitution is unknown. In a prospective cohort of allogeneic HCT recipients who received letermovir, we compared polyfunctional CMV-specific T-cell responses to those of controls who received PCR-guided preemptive therapy before the introduction of letermovir. Thirteen-color flow cytometry was used to assess T-cell responses at 3 months after HCT following stimulation with CMV immediate early-1 (IE-1) antigen and phosphoprotein 65 (pp65) antigens. Polyfunctionality was characterized by combinatorial polyfunctionality analysis of antigen-specific T-cell subsets. Use of letermovir and reduction of viral exposure were assessed for their association with CMV-specific T-cell immunity. Polyfunctional T-cell responses to IE-1 and pp65 were decreased in letermovir recipients and remained diminished after adjustment for donor CMV serostatus, absolute lymphocyte count, and steroid use. Among letermovir recipients, greater peak CMV DNAemia and increased viral shedding were associated with stronger CD8+ responses to pp65, whereas the CMV shedding rate was associated with greater CD4+ responses to IE-1. In summary, our study provided initial evidence that letermovir may delay CMV-specific cellular reconstitution, possibly related to decreased CMV antigen exposure. Evaluating T-cell polyfunctionality may identify patients at risk for late CMV infection after HCT.

Graphical abstract

Figure 1.

Cumulative incidence of CMV infection in the first 270 days after HCT. The cumulative incidence of CMV infection from days 0 through 270 after HCT in letermovir recipients and controls with preemptive treatment. A positive viral event was defined as a composite of CMV end-organ disease or CMV PCR DNAemia: at any level, ≥150 IU/mL, or ≥500 IU/mL for the assay used. Death was treated as a competing risk. Note that only a proportion of patients in both groups had late (after post-HCT day 100) CMV DNA PCR results available for analysis (36 letermovir recipients, 75 controls). Patients without late CMV DNA PCR surveillance were censored only if they did not have late CMV disease data available via review of the electronic medical record or our long-term follow-up database.

Figure 2.

Relative proportion of functional CMV-specific T-cell phenotypes. The proportion of CMV-specific CD4⁺ (A) and CD8⁺ (B) T-cell subset phenotypes based on the total number of functional markers expressed at approximately day 90 after HCT in letermovir recipients and preemptive therapy recipients. Responses were measured after stimulation with CMV IE-1, pp65, or SEB (positive control). Granzyme B monofunctional cytokine responses were excluded from these charts, as some cell subsets have been noted to express this cytokine constitutively (see "Letermovir is associated with decreased CMV-specific polyfunctional responses").

Figure 3.

Proportion of patients with positive polyfunctional CMV-specific T-cell responses. The proportion of patients with positive polyfunctional CMV-specific CD4⁺ (A) and CD8⁺ (B) T-cell responses (defined by the phenotype IFN-γ plus ≥1 functional marker) at approximately day 90 after HCT in letermovir recipients and preemptive therapy recipients. Positive T-cell responses were further defined as those >.05% above background and at least threefold greater than the DMSO (negative control) response in the same cell subset.

Figure 4.

Absolute polyfunctional CMV-specific T-cell counts. Absolute polyfunctional CMV-specific CD4⁺ (top) and CD8⁺ (bottom) T-cell counts (defined by the phenotype IFN-γ plus ≥1 functional marker) at approximately day 90 after HCT in letermovir recipients and preemptive therapy recipients. Cytokine responses were generated after stimulation with CMV IE-1, pp65, and SEB (positive control) antigens. Violin plots show frequency distribution of data and width of sections represent the proportion of a specific value (ie, wider sections represent a greater proportion). Dotted lines represent the interquartile range and solid lines represent the median. Comparisons were made by Wilcoxon rank-sum tests.

Figure 5.

COMPASS-generated PFSs. COMPASS-generated PFSs at approximately day 90 after HCT in letermovir recipients and preemptive therapy recipients. Cytokine responses were generated after stimulation with CMV IE-1, pp65, and SEB (positive control) antigens. Violin plots show frequency distribution of data and width of sections represent the proportion of a specific value (ie, wider sections represent a greater proportion). Dotted lines represent the interquartile range, and solid lines represent the median. Comparisons were made by Wilcoxon rank-sum tests.

Figure 6.

Multivariable linear regression of CMV kinetics with polyfunctional CMV-specific T-cell immunity in letermovir recipients only. Multivariable linear regression examining association of viral kinetics in the first 100 days after HCT with polyfunctional CMV-specific T-cell responses at approximately day 90 after HCT among letermovir recipients only. Comparisons were adjusted for either peak CMV DNAemia (in log₁₀ scale as a continuous variable; A) or CMV shedding (percentage of positive CMV PCR tests collected per 10% increase; B). CMV-specific polyfunctionality was measured as absolute polyfunctional T-cell counts (as defined by the phenotype IFN-γ plus ≥1 functional marker; top half) or COMPASS PFSs (bottom half). Adjustments were also made for donor CMV serostatus, ALCs (≤300 vs >300 cells per mm³), and AUC of daily weight-based prednisone equivalent steroid dose in the first 100 days after HCT. CI, confidence interval; MD, mean difference.

Figure 7.

Functional CMV IE-1 CD8⁺ T-cell subsets at the patient level grouped by late CMV reactivation. Heat map of functional CD8⁺ T-cell subsets to CMV IE-1 at the subject level quantified by posterior probabilities and grouped according to late CMV reactivation (lower violet) vs no reactivation (upper green). Rows correspond to entire antigen-specific cell subsets represented in a patient and are arranged in columns by ascending degree of functionality, from 1 (light blue) to 5 (pink) functions. Individual heat map cells show posterior probability that a functional antigen-specific cell subset is present in a patient and are color coded from white (0) to dark purple (1). For example, patient 1 has a probability of 1.0 of possessing CD8⁺ T cells that express all 5 functional markers but 0 probability of possessing CD8⁺ T cells expressing only TNF-α and granzyme B in response to stimulation with IE-1. Note the increased proportion of subjects with 4 and 5 functional type CD8⁺ T-cell subsets (red box) among patients without CMV reactivation vs those with CMV reactivation.