Immunobiology of cytomegalovirus infection in patients with haematological malignancies undergoing treatment with small molecule inhibitors.

Molecular targeting agents comprising highly specific small molecules such as antiapoptotic B-cell lymphoma 2 protein inhibitors (venetoclax), multiple tyrosine kinase inhibitors (TKI) (i.e. dasatinib) and intracellular Janus kinase (JAK) signal transducer and activator of transcription (STAT) pathway inhibitors (i.e. ruxolitinib) are potent drugs currently used to treat a number of haematological malignancies.1 Infections of any grade due to a variety of microorganisms have been documented to occur at variable frequencies in patients treated with these drugs.2 In this context, cytomegalovirus (CMV) end-organ diseases, including retinitis, pneumonitis and colitis, has been anecdotally reported.3-5 Nevertheless, there is scarce information as to how these treatments impact on the immunobiology of CMV infection,6 a plausible line of research given their pleiotropic effects on B- and T-cell-mediated immunity.2 In this context, dasatinib-induced T-cell or NK-cell large granular lymphocyte (LGL) expansions, which appear to associate with long-lasting therapeutic responses in patients with advanced leukaemia, were reported to be linked with CMV reactivation and high frequencies of CMV-specific T cells targeting an immunodominant epitope on CMV pp65.7 To further address this issue, in this multicentre, prospective, observational study, we enrolled 36 adult (≥18 years old) CMV-seropositive patients (at baseline) with haematologic malignancies (Table 1) who were administered venetoclax (n = 8; group 1), BCR-ABL1 TKIs (n = 7; group 2) or ruxolitinib (n = 21; group 3) as first-line therapy or after conventional chemotherapy between January 2019 and January 2022 at Hospital Clínico Universitario of Valencia (n = 15), Hospital Politécnico Universitario La Fe, Valencia (n = 9), Hospital Ramón y Cajal, Madrid (n = 5), Hospital Vall d'Hebrón, Barcelona (n = 5) and Hospital Morales Meseguer, Murcia (n = 2), all in Spain. The study period comprised the first 180 days after treatment initiation. The current study was approved by the respective Clinical Research Ethics Committees (CEIC) of the participating centres. Informed consent was obtained from all participants. Monitoring of plasma CMV DNA load was conducted on a biweekly basis until day 90+ after treatment initiation and monthly from day +90 until day +180 by the CMV RealTime CMV polymerase. chain reaction (PCR) (Abbott Molecular). As per local guidelines, no patient testing positive for CMV DNA in plasma was started on pre-emptive antiviral therapy. A total of 20/36 patients (55%) developed one (n = 12) or two (recurrences; n = 8) episodes of CMV DNAemia within the study period, the initial one occurring at a median of 30 days after treatment initiation (range, 0–150) (Table 1). Three patients had CMV DNAemia at baseline. Of the 28 detected episodes of CMV DNAemia, 14, developing in 12 patients, were blips (single positive PCR result). All episodes cleared spontaneously. Patients undergoing treatment with BCR-ABL1 TKI (5/7) and ruxolitinib (12/21) had a higher incidence of CMV DNAemia than those receiving venetoclax (1/8) (p = 0.04; chi-squared test). Median CMV DNA peak load was comparable across groups (p = 0.29; Wilcoxon test) and in no case was higher than 182 iu/ml (Figure S1). No patient developed CMV end-organ disease. We next enumerated pp65/IE-1-reactive IFNγ-producing CD8+ and CD4+ T cells in whole blood by flow cytometry with intracellular cytokine staining (ICS), and quantified Tregs by days +30, +60, +90, +120, +150, and +180 after treatment initiation (Supplementary Material).8, 9 As shown in Supplementary Table S1, at baseline, all but one patient with specimen available (n = 32) had detectable CMV-specific CD8+, CD4+ T cells or both. Furthermore, most patients had detectable CMV-specific T-cell responses over the study period. From a quantitative standpoint (Table 2), although both CMV-specific CD4+ and CD8+ T cells tended to fluctuate regardless of the therapy the patients received and varied widely across individuals, two relatively distinctive patterns were identified. First, patients undergoing treatment with BCR-ABL1 TKI appeared to display lower CMV-specific CD4+ T-cell counts at all time points compared to those in patients treated with venetoclax or ruxolitinib, the differences reaching statistical significance over the first two months of treatment. Second, a trend towards decreasing CMV-specific CD4+ and CD8+ T-cell counts over time was evident in patients treated with venetoclax or ruxolitinib, but not with BCR-ABL1 TKI. Importantly, median CMV-specific CD4+ and CD8+ T cells measured by days +30, +60, +90, +120, +150, and +180 did not differ significantly from those quantified at baseline in any study group (Supplementary Table S2). We next assessed whether the occurrence of CMV DNAemia could be predicted on the basis of CMV-specific T-cell counts; this could only be reasonably investigated in patients receiving ruxolitinib. Data are shown in Supplementary Table S3 and indicated that CMV-specific CD4+ and CD8+ T-cell counts at baseline and day +30 were not significantly different in patients with or without subsequent CMV DNAemia. Treg counts were available from 185 blood samples from 32 patients, showing detectable levels in all but seven. Regardless of the treatment group to which the patients belonged, Treg peripheral blood levels fluctuated over time, with no apparent association with time elapsed since treatment initiation (Table 2). Moreover, Treg levels were comparable across groups at all time points explored. This is consistent with data reported in previous studies.10, 11 Overall, Treg levels did not correlate with either CMV-specific CD4+ (Spearman's rank test, 𝜌 = −0.049; p = 0.51), or CD8+ (𝜌 = −0.020; p = 0.79) T-cell counts. In summary, CMV DNAemia occurred frequently in patients treated with BCR-ABL1 TKI and ruxolitinib mostly within the first two months after treatment initiation, and less commonly in patients treated with venetoclax, perhaps reflecting a more pronounced effect of the former two drugs on T-cell immune responses compared to the latter; thus, although BCR-ABL1 TKI and ruxolitinib may have a detrimental effect on CMV-specific T-cell responses, the magnitude of this effect is insufficient to prevent ultimate control of CMV replication. In fact, no patient in our series developed CMV end-organ disease. Our data on CMV-specific T-cell immunity seem to support this assumption. In line with our data, a previous study found that almost 50% and around 90% of patients treated with ruxolitinib for less or more than five years, respectively, displayed CMV-specific T-cell responses (measured by IFN-γ ELISPOT assay) below the normal range, yet less than 10% of patients had detectable CMV DNAemia.6 The current study has a major limitation: its small sample size, which makes it merely exploratory. Furthermore, CMV-specific functional T-cell responses other than those targeting IE1 and pp65 and resulting in the secretion of IFN-γ were not assessed. Lastly, given the CMV DNA monitoring schedule and the lack of available samples at several time points, some episodes of CMV DNAemia may have been overlooked. From a practical clinical view, our data do not support systematic CMV DNA monitoring in blood in these patients. Whether CMV DNAemia in these patients may mediate ‘indirect effects’ as seen in the allo-HSCT setting12, 13 should be thoroughly investigated. Estela Giménez and Eliseo Albert hold a Juan Rodés research contract from the Carlos III Health Institute. Juan Carlos Hernández-Boluda, María José Terol, Carlos Solano and David Navarro designed the research and interpreted the data. Solano de la Asunción Carlos, Estela Giménez and Eliseo Albert, María José Remigia and Paula Amat performed analytical determinations. Solano de la Asunción Carlos and Eliseo Albert performed statistical analyses. Javier López-Jiménez, Rafael Andreu, Dolores García, and Laura Fox contributed to data collection and analysis. David Navarro wrote the manuscript. All authors reviewed and approved the final version of the manuscript. This research was supported by a grant from FIS PI18/00127 (Fondo de Investigaciones Sanitarias), Ministerio de Sanidad y Consumo, Spain. The authors report no conflicts of interest. Informed consent was obtained from all participants. Data S1. Table S1. Table S2. Table S3. Figure S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.