Immunovirology of cytomegalovirus infection in allogeneic stem cell transplant recipients undergoing prophylaxis with letermovir: A narrative review

On November 7, 2017, the US Food and Drug Administration approved the use of letermovir (LMV) for prophylaxis of cytomegalovirus (CMV) infection in adult CMV‐seropositive allogeneic stem cell transplant recipients. After 6 years of use, a large body of real‐world experience has been accumulated that validates the Phase III clinical trial results, in which LMV was shown to significantly reduce the risk of clinically significant CMV infection—defined as CMV end‐organ disease or CMV DNAemia requiring pre‐emptive antiviral therapy (PET)—and increase survival up to Week 24 after treatment inception. Notwithstanding, several issues still need to be settled, thus further investigation is required. First, since viral DNA may accumulate as a result of LMV‐driven abortive CMV infection, what is the optimal viral load threshold in the blood that would prompt LMV prophylaxis interruption and PET inception? Should this be adapted to the patient's risk? Second, what is the impact of LMV prophylaxis on the reconstitution of functional CMV‐specific T‐cell responses? Would it be a wise approach to individually tailor the duration of LMV treatment according to the number of peripheral blood CMV‐specific T cells at the end of regular prophylaxis? Third, how frequently do LMV‐resistant strains arise while patients are on LMV prophylaxis and how could this be minimized? Here, we discuss the literature addressing these topics.

CMV end-organ disease, either following primary infection or CMV reactivation, often occurs in the setting of high or increasing viral loads (viremia/pp65 antigenemia/virus DNAemia) in the blood compartment. 3,4 Thus, frequent monitoring of the viral burden in the blood may anticipate the occurrence of this clinical event, whose incidence can be drastically limited by the timely administration of antiviral drugs with intrinsic activity against CMV (pre-emptive antiviral therapy [PET]). 4,5 The implementation of PET strategies based upon monitoring the CMV DNA load in the blood (plasma or whole blood) by quantitative nucleic acid amplification methods, most commonly real-time polymerase chain reaction (PCR), has dramatically decreased the incidence of end-organ disease in the allo-SCT setting in the last decade (overall ≤5%). This is despite the great variability of nucleic acid amplification platforms for viral load measurements, which display noninterchangeable analytical performances and the different CMV DNA load cut-offs established for triggering the inception of antivirals across centers. 6 In turn, a large number of nonessential CMV genes (around one-third of the CMV genome) encode proteins with proinflammatory (human homologs of cytokines, chemokines, or chemokine receptors) or immunosuppressive (i.e., inhibitors of natural killer [NK]-cell activation or antigen presentation to CD4 + and CD8 + T cells) properties. 4,5 The proinflammatory and immunosuppressive nature of CMV mechanistically supports the association reported in some series between CMV infection and the occurrence of acute graft versus host disease (aGvHD) or virus superinfections (i.e., due to Epstein-Barr virus), 7,8 both among the so-called "indirect effects." Moreover, CMV DNAemia was suggested to be associated in a positive dose-response manner with an increased risk of overall mortality (OM) and nonrelapse mortality (NRM) in the first year after allo-SCT, independent of the use of PET. 9 The link between CMV DNAemia, most notably that requiring PET, and OM/NRM has also been observed in other studies, 10,11 including a meta-analysis and metaregression analysis. 12 Notwithstanding, a causal implication for CMV has not been proven.
Based upon the results of a randomized, placebo-controlled clinical trial published by Marty et al., 13 which showed that letermovir (LMV) significantly reduced the risk of clinically significant CMV infection (csCMV-I) (defined as CMV end-organ disease or CMV DNAemia requiring PET), and increased survival up to Week 24 after treatment inception, the US Food and Drug Administration approved this drug for prophylaxis of CMV infection and disease in adult CMVseropositive allo-SCT patients (November 7, 2017). Since then, a large body of real-life experience has been gathered, which, overall, confirms these beneficial effects 14 ; furthermore, LMV prophylaxis has been shown to be a cost-effective therapeutic option in adult allo-SCT compared with standard approaches based on PET. 15,16 Notwithstanding, several relevant aspects of the immunovirology of CMV infection in allo-SCT recipients undergoing LMV prophylaxis remain to be elucidated or fully settled, including how to optimally manage CMV DNAemia breakthrough episodes, how long LMV prophylaxis should be maintained, the potential impact of LMV use on the reconstitution of CMV-specific T-cell responses, or how frequently LMV-resistant strains arise while patients are on prophylaxis. Here, we review and discuss the available literature regarding these issues.

| MECHANISM OF ACTION OF LMV
A novel class of drug derivatives targeting the CMV terminase complex has been developed during the last 20 years, of which only quinazoline AIC246 (3,4-dihydro-quinazoline-4-yl-acetic acid)-LMV-has been approved by international drug agencies for clinical use. 17,18 CMV terminase is a heterotrimeric complex formed by the proteins pUL56, pUL89, and pUL51, which intervenes in the cleavage of concatemeric viral DNA, a catalytic activity required for packaging functional unit length linear genomes into preformed capsids. 19 The pUL56 subunit, the main target of LMV, binds the portal protein pUL104, a macromolecule that assembles into ring-like structures to form a channel in assembled capsids, and the packaging signal (pac) on concatemeric CMV DNA and mediates the translocation of the viral genome into viral capsids by virtue of its ATPase activity, which is finally cleaved by pUL89 (DNA duplex nicking). 20 [22][23][24] LMV is approximately 1000-fold more potent than ganciclovir and retains activity against CMV virus strains resistant to currently approved antivirals. [22][23][24][25] Importantly, LMV neither inhibits CMV DNA replication nor expression of late viral proteins in infected cells, rather it dramatically impairs nuclear egress and eventual release of infectious virions from cells, thus severely limiting subsequent rounds of CMV replication in adjacent or distant uninfected cells. 22,23 Remarkably, LMV activity associates with the accumulation of dense bodies (subviral infectious particles) in the cytoplasm of infected cells, which may drive the expansion of CMV-specific T cells. 23 Of potential clinical interest, by using a sensitive fluorescence reduction assay that closely correlates with plaque reduction assays, combinations of LMV with ganciclovir and cidofovir at different ratios were shown to yield consistently additive responses, with weighted average combination index (CI wt ) of 0.88-1.2. 26,27

| IMPACT OF LMV ON CMV-SPECIFIC T-CELL RECONSTITUTION
Definitive control of CMV replication in the allo-SCT setting critically depends upon the adequate activation and expansion of functional virus-specific CD4 + and CD8 + T cells. [3][4][5]28,29 A large fraction of the virus proteome elicits functional T-cell responses, 30 however, the tegument protein pp65 and immediate early protein 1 (IE-1) are immunodominant in most individuals, both triggering robust T-cell responses that may independently protect against the occurrence of active CMV infection and effectively contribute to virus clearance in this scenario. 28,29 The number of pp65/IE-1 monofunctional (interferon-gamma [IFN-γ]-producing) CD8 + or CD4 + T cells in whole or peripheral blood mononuclear cells can be easily assessed via flow cytometry for intracellular cytokine staining assays and interferongamma release assays (IGRA), such as the Quantiferon ® CMV assay or ELISpot assays. Several studies have consistently shown an inverse association between peripheral blood levels of CMV-specific IFN-γproducing CD8 + or CD4 + T cells at different time points after allo-SCT and the risk of subsequent CMV DNAemia (both requiring PET or not). Moreover, cut-off cell or IFN-γ levels affording protection were tentatively proposed, which overall displayed reasonable positive predictive values. [31][32][33][34][35][36][37][38][39] However, whether the quantification of CMV-specific polyfunctional T cells outperforms that of IFN-γproducing T cells in predicting the occurrence of CMV DNAemia remains unsettled. 40,41 The increased rate of first CMV DNAemia episodes after discontinuation of LMV prophylaxis in the original randomized controlled trial, 13  CMV-specific CD4 + T-cell counts, respectively. 44 Gabanti et al.,45 also using a flow cytometry assay, showed that CMV-specific T-cell recovery was delayed by approximately 100 days in LMV-treated patients compared with controls. Specifically, CMV-specific CD4 + and CD8 + T-cell counts were significantly lower in the LMV group at Days 120-360 and 90-120, respectively; this, however, appeared to have no significant impact on clinical outcomes. Via a flow cytometry assay, Giménez et al., 46 showed that the percentage of LMV and non-LMV patients exhibiting detectable CMV-specific IFN-γ-producing Tcell responses was comparable, nevertheless, CMV-specific CD4 + and CD8 + T-cell counts were lower in LMV patients at Days +60 and +90 after allo-SCT. Finally, Lauruschkat et al., 47 reported a trend toward lower CMV-specific IFN-γ-producing CD8 + and CD4 + T-cell counts from Day +30 until +120, as enumerated by flow cytometry, in patients managed via a PET approach compared with those under LMV prophylaxis. Nevertheless, interestingly, NK-cell counts were higher in the former, particularly at Day +120. Taken collectively, these data strongly suggest that LMV prophylaxis may inhibit to some extent CMV-specific T-cell reconstitution after allo-SCT, likely as a result of suboptimal antigen exposure. In this sense, it remains to be determined whether monitoring CMV-specific T-cell responses may be instrumental in individually tailoring the duration of LMV prophylaxis.

| EFFICACY OF LMV IN PREVENTING ACTIVE CMV INFECTION AND END-ORGAN DISEASE
There is solid evidence for the clinical efficacy of primary prophylaxis  Tables 1 and 2, respectively. The rate of reported breakthrough csCMV-I ranged from 0% to 32%, whereas that of active CMV infection was higher, between 5% and 47%. Consistent with the Phase III clinical trial, 13 most studies showed an increased rate of CMV DNAemia/pp65 antigenemia, including csCMV-I, after interrupting LMV prophylaxis.