Impaired T cells and "memory-like" NK-cell reconstitution is linked to late-onset HCMV reactivation after letermovir cessation

Abstract

Allogeneic hematopoietic stem cell transplantation (alloSCT) is the only cure for many hematologic malignancies. However, alloSCT recipients are susceptible to opportunistic pathogens, such as human cytomegalovirus (HCMV). Letermovir prophylaxis has revolutionized HCMV management, but the challenge of late HCMV reactivations has emerged. Immunological surrogates of clinically significant HCMV infection (csCMVi) after discontinuation of letermovir remain to be defined. Therefore, we studied natural killer (NK)-cell reconstitution along with the global and HCMV pp65-specific T-cell repertoire of 24 alloSCT recipients at 7 time points before (day +90) and after (days +120-270) cessation of letermovir prophylaxis. Patients who experienced csCMVi had lower counts of IFN-γ+ HCMV-specific CD4+ and CD8+ T cells than HCMV controllers. Furthermore, patients with csCMVi displayed late impairment of NK-cell reconstitution, especially suppression of "memory-like" CD159c+CD56dim NK-cell counts that preceded csCMVi events in most patients. Moreover, several surrogates of immune reconstitution were associated with the severity of HCMV manifestation, with patients suffering from HCMV end-organ disease and/or refractory HCMV infection harboring least HCMV-specific T cells and "memory-like" NK cells. Altogether, our findings establish an association of delayed or insufficient proliferation of both HCMV-specific T cells and "memory-like" NK cells with csCMVi and the severity of HCMV manifestations after discontinuation of letermovir prophylaxis.

Graphical abstract

Figure 1.

Connection of HCMV load levels and clinical severity of HCMV manifestation. (A) The incidence of csCMVi onset among patients with csCMVi was monitored from day 0 to day +360. CsCMVi events were reported at the end of each month based on their diagnosis date. (B) Peak HCMV DNA copies per milliliter measured by polymerase chain reaction until day +365 in patients who underwent alloSCT, serving as control for HCMV (controllers, blue) or those with csCMVi (brown). Patients with csCMVi were further subdivided into patients without (gray) or with Rf/EOD (red). ∗P < .05, ∗∗∗P < .001 using Mann-Whitney U test.

Figure 2.

Poor HCMV–specific T-cell reconstitution in csCMVi patients after discontinuation of letermovir. Samples from alloSCT recipients receiving letermovir prophylaxis until day +100 who controlled HCMV (controllers, blue) and those with csCMVi (brown) were compared using flow cytometry. (A) Heat map comparing the MMRs (csCMVi/controllers) of selected T-cell populations between the 2 cohorts. (B-E) Background-corrected median numbers (B) and frequencies (C) of HCMV–specific Th cells (CD4⁺IFN-γ⁺) as well as numbers (D) and frequencies (E) of cytotoxic T cells (CD8⁺IFN-γ⁺CD107a⁺). (A-E) Mann-Whitney U test and Benjamini-Hochberg procedure to test for an FDR of <0.2. (F,G) Background-corrected individual numbers of HCMV–specific (F) Th cells (CD4⁺IFN-γ⁺) and (G) cytotoxic T cells (CD8⁺IFN-γ⁺ CD107a⁺) on day +90 and day +210 are shown. (H-I) Background-corrected individual numbers of HCMV–specific (H) Th cells (CD4⁺IFN-γ⁺) and (I) cytotoxic T cells (CD8⁺IFN-γ⁺CD107a⁺) before and after csCMVi are displayed. (F-I) ∗P < .05, ∗∗P < .01, ∗∗∗P < .001 using paired Wilcoxon test. CD, cluster of differentiation; d, day; MMR, median-to-median ratio.

Figure 3.

Higher numbers and frequencies of NK cells and “memory-like” NK cells are associated with HCMV control in alloSCT recipients after discontinuation of letermovir. NK-cell reconstitution was compared in alloSCT recipients who controlled HCMV (controllers, blue) and those with csCMVi (brown). (A) Heat map comparing the MMRs (csCMVi/controllers) of selected NK-cell populations between the 2 cohorts. (B-E) Median absolute numbers (B) and frequencies (C) of CD56^dimCD159c⁺(CD57⁺) cells as well as numbers (D) and frequencies (E) of CD56^dimCD159c⁺⁺ (CD57⁺) cells are shown. (A-E) Mann-Whitney U test and Benjamini-Hochberg procedure to test for a FDR of <0.2. Significant results with an FDR >0.2 are indicated by the gray stars. (F-I) Individual numbers of (F) CD56^dimCD159c⁺, (G) CD56^dimCD159c⁺CD57⁺, (H) CD56^dimCD159c⁺⁺, (I) CD56^dimCD159c⁺⁺CD57⁺ on day +90 and day +210 are shown. (J,K) Individual absolute numbers of (J) CD56^dimCD159c⁺ and CD56^dimCD159c⁺CD57⁺ as well as (K) CD56^dimCD159c⁺⁺ and CD56^dimCD159c⁺⁺CD57⁺ before and after csCMVi are displayed. (F-K) ∗P < .05, ∗∗P < .01 using paired Wilcoxon test. CD, cluster of differentiation; d, day; MMR, median-to-median ratio.

Figure 4.

Severity of HCMV infection is linked to numbers and frequencies of HCMV–specific T cells. T-cell reconstitution was compared in alloSCT recipients who controlled HCMV (controllers, blue), those who experienced csCMVi without Rf/EOD (gray), and those who had csCMVi with Rf/EOD (red). (A) Absolute numbers of total Th cells (CD4⁺) and cytotoxic T cells (CD8⁺). (B-C) Background-corrected numbers and percentages of HCMV–specific Th cells (CD4⁺IFN-γ⁺, B) and cytotoxic T cells (CD8⁺IFN-γ⁺CD107a⁺, C) after 16 to 18 h of stimulation with pp65 HCMV-peptide mix. (A-C) Symbols/lines and error bars indicate medians and interquartile ranges, respectively. Kruskal-Wallis test and Benjamini-Hochberg procedure to test for a FDR of <0.2. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001. CD, cluster of differentiation.

Figure 5.

Numbers and frequencies of “memory-like” NK cells are associated with the severity of HCMV infection. NK-cell reconstitution was compared in alloSCT recipients who controlled HCMV (controllers, blue), those who experienced csCMVi without Rf/EOD (gray), and those who faced csCMVi with Rf/EOD (red). (A) CD56⁺CD3⁻, CD56^brightCD3⁻ and CD56^dimCD3⁻ NK-cell numbers. (B-C) Absolute cell numbers and frequencies of (B) CD56^dimCD159c⁺(CD57⁺) and (C) CD56^dimCD159c⁺⁺(CD57⁺) “memory-like” NK cells. (A-C) Symbols/lines and error bars indicate medians and interquartile ranges, respectively. Kruskal-Wallis test and Benjamini-Hochberg procedure to test for a FDR of <0.2. The significant result with an FDR >0.2 is indicated by a gray star. ∗P < .05, ∗∗P < .01. CD, cluster of differentiation.

Figure 6.

Impaired reconstitution of “memory-like” NK cells and HCMV–specific T cells in patients with Rf/EOD CMVi. HCMV–specific CD4⁺ and CD8⁺ T cells as well as “memory-like” NK cells were quantified in patients with Rf/EOD-CMVi. Individual kinetics of all 7 patients with Rf/EOD-CMVi are shown. CD, cluster of differentiation; PCR, polymerase chain reaction.