In vitro assessment of the combined effect of letermovir and sirolimus on cytomegalovirus replication

Abstract

Introduction

Letermovir (LMV) is used for prophylaxis of cytomegalovirus (CMV) reactivation and end-organ disease in adult CMV-seropositive allogeneic hematopoietic stem cell transplant recipients (allo-HSCT). In turn, sirolimus (SLM) which displays in vitro anti-CMV activity, is frequently employed for prophylaxis of Graft vs. Host disease in allo-HSCT. Here, we aimed at assessing whether LMV and SLM used in combination may act synergistically in vitro on inhibiting CMV replication.

Material and methods

The antiviral activity of LMV and SLM alone or in combination was evaluated by a checkerboard assay, using ARPE-19 cells infected with CMV strain BADrUL131-Y. LMV and SLM were used at concentrations ranging from 24 nM to 0.38 nM and 16 nM to 0.06 nM, respectively.

Results

The mean EC₅₀ for LMV and SLM was 2.44 nM (95% CI, 1.66-3.60) and 1.40 nM (95% CI, 0.41-4.74), respective. LMV and SLM interaction yielded mainly additive effects over the range of concentrations tested.

Conclusion

The additive nature of the combination of LMV and SLM against CMV may have relevant clinical implications in management of CMV infection in allo-HSCT recipients undergoing prophylaxis with LMV.

INTRODUCTION

Cytomegalovirus (CMV) infection remains a significant cause of morbidity and mortality in the allogeneic hematopoietic stem cell transplantation (allo-HSCT) setting [1]. Although preemptive antiviral therapy (PET) has dramatically decreased the incidence of CMV end-organ disease [1], there is growing evidence that CMV DNAemia, especially if requiring PET, may be detrimental by increasing the risk of overall and non-re-lapse mortality [2-5]. Letermovir (LMV), an antiviral compound that blocks virion maturation by inhibiting the CMV DNA terminase complex [6], has been approved for prophylaxis of CMV reactivation and end-organ disease in adult CMV-sero-positive allo-HSCT recipients. LMV efficiently suppresses CMV replication and seemingly increases survival by week 24 after allo-HSCT [7]. Nowadays, sirolimus (SLM), an mTOR inhibitor that inhibits T cell activation specifically by binding the cytosolic protein FKPB-12, is currently in wide use combined with other immunosuppressive agents for prophylaxis of acute graft-versus-host disease (aGvHD) in allo-HSCT. SLM exposure has been shown to dose-dependently decrease the risk of CMV DNAemia in a cohort of allo-HSCT recipients at high risk of CMV end-organ disease [8]. Furthermore, risk of CMV DNAemia requiring PET was shown to fall by 6% for each 1 ng/mL increase in SLM trough concentration [9]. The protective effect of SLM against CMV infection may be mechanistically related to the key role of the mTOR pathway in CMV protein synthesis and replication [10-14] and improvement of CMV-specific T-cell function via modulation of the environmental milieu [15]. The use of LMV and SLM in combination may improve the management of CMV infection in allo-HSCT by increasing clinical efficacy, reducing adverse effects and minimizing the likelihood of emergence of LMV-resistant variants. Evidence partly supporting the assumption that the association of LMV and SLM may reduce the rate of clinically significant CMV infection (CMV DNAemia that requires preemptive antiviral therapy/ CMV end-organ disease) was recently provided [16]. A synergistic effect has been shown in vitro for maribavir, another anti-CMV drug inhibiting virus DNA replication, and SLM [17]. Here, we evaluated in vitro the potential interaction between SLM and LMV regarding its anti-CMV activity.

METHODS

Cells and virus. Human ARPE-19 retinal pigment epithelial cells (ATCC CRL-2302) were cultured in high-glucose Dulbecco’s Modified Eagle Medium Nutrient Mixture F-12 (DMEM:F12K) (Gibco, ThermoFisher Scientific, Waltham, Massachusetts, USA), supplemented with 10 % FCS (HyClone Laboratories Inc, Cytiva, USA), 10,000 IU penicillin (Gibco, ThermoFisher Scientific, Waltham, Massachusetts, USA), and 10 mg streptomycin (Gibco, ThermoFisher Scientific, Waltham, Massachusetts, USA). The CMV strain BADrUL131-Y4 [18] was kindly provided by Dr Pilar Pérez (ISCIII, Madrid, Spain). This strain is derived from a bacterial artificial chromosome clone of the CMV strain AD169 genome that was modified in Escherichia coli to express a functional UL131 protein, which permits replication in ARPE-19 and MRC-5 cells. Viral titers of CMV BADrUL131-Y4 were determined by limiting dilution in 96-well plates using MRC-5 cells.

Antiviral compounds. LMV was kindly provided by Merck, Sharp & Dohme (MSD) and stored as a stock solution of 50 milimolar (mM) in dimethyl sulfoxide (DMSO). A fresh working solution of 50 nanomolar (nM) was prepared in DMEM/ F12K medium for each experiment and for medium renewal. SLM was purchased from Sigma Aldrich Merck (Merck KGaA, Darmstadt, Germany) and stored in a stock solution of 11 micromolar (µM) in DMSO. A working solution of 110 nM in in DMEM/F12K medium was prepared for each experiment.

Antiviral assay. The antiviral activity of LMV and SLM alone or in combination was evaluated by a checkerboard assay, as previously described [19]. ARPE-19 cells were seeded in 96-well microtiter plates (4x10⁴ cells/well) for 24 h at 37ºC and 5% CO₂ and infected with 0.1 MOI of CMV strain BADrUL131-Y for 2 h in DMEM/F-12K medium containing either LMV (two-fold dilutions from 24 nM-14 ng/ml- to 0.38 nM- 0.22 ng/ml-), SLM (two-fold dilutions from 16 nM-15 ng/ml- to 0.06 nM-0.05 ng/ml-), or both drugs. The resulting two-dimensional matrix included all possible combinations of LMV and SLM and individual concentrations of each drug. The range of LMV and SLM concentrations tested in the current study was derived from that used in previous studies addressing the in vitro anti-CMV activity of LMV [6,17,20] or the potential synergistic effect between SLM and maribavir [17], respectively. For both drugs, these ranges were centered on the reported EC₅₀ against CMV [6,17,20]. Four positive (CMV-infected cells in cell culture medium) and negative (CMV-uninfected cells in cell culture medium) controls were included in each plate. A total of five plates were prepared in three separate experiments. The plates were incubated for 6 days, and the medium (with or without the drugs) was replaced (with and without drugs, as appropriate for each well) every 2-3 days, as half-life for LMV and SLM are approximately of 12 and 60 hours, respectively, according to the corresponding manufacturer. After incubation, cells were washed, fixed with acetone/methanol (1:1), permeabilized with 0.2% Triton-X100 and stained for 1 h with 100 µl/well of a mouse IgG2a monoclonal antibody anti-Cytomegalo-virus immediate early 1 (IE1) protein (1:150) (ThermoFisher, Waltham, Massachusetts, USA). A secondary FITC-labelled goat anti-mouse IgG2a cross-adsorbed secondary antibody (ThermoFisher, Waltham, Massachusetts, USA) was added at a dilution of 1:150 and incubated in dark for 1 h. Green-fluorescent nuclei were counted in a fluorescence microscope. Potential drug cytotoxicity was assessed for each experiment using mock-infected cells and matched drug exposure. Cell viability was verified after 6 days of culture for each condition using the alamarBlue cell viability assay (ThermoFisher, Waltham, Massachusetts, USA), following the manufacturer’s recommendations.

Statistical analysis. EC₅₀ was calculated by non-linear regression and using the dose-response curve obtained from the average percentage inhibition for each compound alone (GraphPad Prism software). Drug interaction was evaluated using MacSynergy II software, and interpreted as detailed elsewhere [20]. Differences between frequencies were analyzed using the Chi Square test. A P value <0.05 was considered statistically significant.

RESULTS

Both LMV and SLM inhibited CMV replication in ARPE-19 cells in a dose-dependent manner. The mean EC₅₀ was 2.44 nM [1.40 ng/ml] (95% CI, 1.66-3.60) and 1.40 nM [1.28 ng/ ml] (95% CI, 0.41-4.74), respectively (Table 1 and Figure 1 for dose-response curves). Overall synergy/antagonism volumes were 0% for synergy and -117% for antagonism at 99% confidence interval (Table 1). Interaction between the two drugs was concentration-dependent and mainly yielded additive effects (linear isobologram) over the range of concentrations tested (Figure 2), although low-level antagonistic interactions could be found, particularly at low concentrations of LMV (0.75 nM [0.43 ng/ml] and relatively high concentrations of SLM (8 nM [7.31 ng/ml]. Both LMV and SLM induced no cytotoxicity at the concentration ranges used during the incubation time, as revealed by the alamarBlue cell viability assay (not shown).

Table 1.

Analysis of the combination of letermovir and sirolimus by checkerboard matrix

Drug A	Maxa(nM)	EC50b (95%CI)	Drug B	Maxc(nM)	EC50b(95%CI)	Mean vol µM2%d (95%)		Mean vol µM2%d (99%)		Mean vol µM2%d (99.9%)
						Synergy	Antagonism	Synergy	Antagonism	Synergy	Antagonism
Letermovir	24	2.44 (1.66-3.60)	Sirolimus	16	1.40 (0.41-4.74)	0	-172	0	-117	0	-76

95%CI, 95% Confidence Interval; EC50, 50% Effective Concentration; nM, nanoMolar; vol, volume.

^aA total number of n=7 2-fold dilutions were tested starting from 24nM.

^bThe mean concentration capable of inhibiting Cytomegalovirus infection by 50% (EC50) was calculated by using the individual dose-response curve for each compound (Graphpad Prism software, v.6).

^cA total number of n=9 2-fold dilutions were tested starting from 16nM.

^dVolumes (µM²%) of synergy and antagonism were calculated using the MacSynergyII software. Values >50 µM2% were considered moderate and significant in vivo. Volumes at statistical level of confidence of 95%, 99% and 99.9% are shown.

Figure 1.

Dose-response curves of letermovir (LMV) and sirolimus (SLM) on CMV strain BADrUL131-Y infection of ARPE-19 cells. The percentage of inhibition represents the mean vaule of 5 experiments

Figure 2.

Dimensional synergy plots for antiviral drug combinations with letermovir and sirolimus. The plots were generated from the results of 5 experiments of selected letermovir and sirolimus combinations against using MacSynergy II software. The difference between the observed viral growth inhibition and the amount expected from an additive effect of the drug concentrations was calculated at 99.9% confidence level.

DISCUSSION

Prophylaxis with combined LMV and SLM is currently common practice in the allo-HSCT setting for prevention of CMV end-organ disease and aGvHD. Since both compounds display anti-CMV activity [6,10-12], we were interested in determining whether their interaction could result in a synergistic anti-CMV effect in epithelial ARPE-19 cells. To address this issue, we employed a conventional checkerboard assay using LMV and SLM concentrations centered around their EC₅₀ against CMV [6,17,20] and SLM concentrations across the therapeutic range (4-12 ng/ml) [21]. Yet, due to the mechanism of action of both compounds and the choice of IE-1 as the assay read out full inhibition could not be achieved, as expected; in this sense, in the presence of LMV, CMV IE-1 can only be expressed in primarily infected cells (roughly 1/10 cells as an MOI of 0.1 was used); no further rounds of CMV replication ensue, as LMV blocks the generation of virus infectious particles. We obtained inhibition rates for LMV and SLM similar to previous studies [6,17,20], although our data indicated that the combination of LMV and SLM was additive at best. Notably, a low-level antagonistic effect could be observed in certain drug concentration combinations, in particular when we used high concentrations of SLM and low concentrations of LMV. Although speculative, the lack of synergy between the two drugs and the antagonistic interactions documented could be mechanistically related to the ability of SLM to interfere with CAP-dependent translation by inhibiting mTOR kinase [4], which may impair expression of the LMV viral target, the CMV terminase complex. Whether our in vitro observations mirror in vivo interaction between these drugs is uncertain; in this context, two facts need to be considered; first, peak LMV levels reached during prophylaxis could be 100-fold higher (approximately 5µM) than those used in our experiments [22], and thus CMV replication at mucosal and tissue sites could be drastically reduced in vivo, likely to a greater extent than in our in vitro model (around 50%); second, coadministration of LMV and SLM resulted in a 3.4-fold increases in area under the plasma concentration–time curve and 2.8-fold increases in maximum plasma concentration, respectively, of SLM [23].

The current study has several limitations. First, clinical CMV isolates were not used in the experiments; yet, given the nature of the recombinant CMV strain used herein (BAD-rUL131-Y4) [18], we do not foresee a straightforward biological reason that would account for LMV and SLM acting differently when employing clinical isolates. This assumption needs confirmation, though. Second, due to logistic reasons infectious virus yields could not be quantitated.

In summary, we have shown the additive nature of LMV and SLM in vitro interaction (over most drug concentrations assayed) in terms of anti-CMV activity in epithelial cells. Mechanistically the combination of LMV and SLM in vivo may thus shut off CMV replication in vivo at higher level than either drug used individually. Nevertheless, real-life data of the combined use of these drugs is required to gauge the potential clinical relevance of our in vitro observations.

FUNDING

The present study was (partially) funded by Merck Investigator Study Program #58812. The funder had no role in the design of the study, in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.