Validation and characterization of five distinct novel inhibitors of human cytomegalovirus

Abstract

The critical consequences of human cytomegalovirus (HCMV) infection in the transplant population and in congenitally infected infants, the limited treatment options for HCMV, and the rise of resistant mutants towards existing therapies has fueled the search for new anti-HCMV agents. A pp28-luciferase recombinant HCMV was used as a reporter system for high-throughput screening of HCMV inhibitors. Approximately 400,000 compounds from existing libraries were screened. Subsequent validation assays using resynthesized compounds, several virus strains and detailed virology assays resulted in the identification of five structurally unique and selective HCMV inhibitors, active at sub to low μM concentrations. Further characterization revealed that each compound inhibited a specific stage of HCMV replication. One compound was also active against herpes simplex virus (HSV1 and HSV2) and another compound was active against Epstein Barr Virus (EBV). Drug combination studies revealed that all five compounds were additive with ganciclovir or letermovir. Future studies will focus on optimization of these new anti-HCMV compounds along with mechanistic studies.

Graphical Abstract

Time of maximal activity of the five new inhibitors and GCV during HCMV replication

Addon and removal assays were performed for each of the five compounds as described in the materials and methods section and shown for each compound in Figs 5–8 C&D. The timing of 75% reduction in luciferase activity was calculated for each compound both when added and when removed and is depicted in a unique color. The individual addon and removal assays are provided in Fig 5,6,7,8 C&D.

1. INTRODUCTION

Within the world population, human cytomegalovirus (HCMV) is one of the most common pathogens. Seroprevalence rates approach 90% in individuals over 80 years old¹. Although HCMV infection is usually asymptomatic, it gravely threatens immunocompromised hosts, including those with AIDS and transplant recipients. HCMV is the most common infection transmitted during pregnancy from mother to fetus, leading to hearing impairment and global disabilities, in children^2–6.

Ganciclovir (GCV), a nucleoside analog, and its oral formulation val-GCV have dramatically improved the outcome of transplantations by reducing sequelae from HCMV replication and disease^7,8. GCV was also shown to prevent hearing deterioration in congenitally infected children⁹. However, prolonged therapy with GCV or val-GCV often results in serious bone marrow toxicities^10–13. A phase III clinical trial of oral val-GCV in infants with congenital HCMV infection found that 6 months’ therapy resulted in better neurological outcomes compared to 6 weeks’ therapy; however GCV-resistant mutants developed in study participants¹⁴. Similarly, GCV-resistant mutants emerge in transplant recipients^15,16. When GCV resistance arises, the limited therapeutic alternatives, foscarnet and cidofovir, are highly nephrotoxic and can only be administered intravenously. Their use in cases of resistant or refractory HCMV is associated with high morbidity and mortality¹³. Several new drugs have advanced into FDA approval and clinical trials.

Letermovir, a terminase inhibitor, is approved for HCMV prophylaxis after hematopoietic bone marrow transplantation¹⁷. Rapid selection of UL56 mutations in cell culture¹⁶ and letermovir resistance in patients is being reported¹⁸. Maribavir, a viral UL97 kinase inhibitor, is in clinical trials¹⁹.

There is a clear need to improve the anti-HCMV drug armamentarium, either as monotherapy or combination therapy. Towards this goal, we completed a large screening campaign of ~400,000 compounds using our highly-sensitive pp28-luciferase HCMV reporter and several collections of chemical libraries²⁰. In-depth validation studies resulted in the identification and characterization of five structurally distinct and selective compounds active in vitro at sub to low μM concentrations.

2. RESULTS

2.1. High throughput screening for HCMV inhibitors:

The pp28 luciferase HCMV was used for screening of the Molecular Library Small Molecule Repository (MLSMR) collection of 370,000 compounds, the NIH Chemical Genomic Center (NCGC) diversity collection of 65,000 compounds and the NCGC Pharmaceutical Collection of investigational and approved drugs. The screening used two compound doses plated before infection and luciferase activity was quantified at 72 hours post-infection (hpi) in lysates. A total of 2215 inhibitors of luciferase activity in our engineered HCMV were selected using published compound dose-response curve algorithms²¹. A secondary screen was performed to remove compounds that showed toxicity in human foreskin fibroblasts (HFFs) resulting in 847 confirmed, non-toxic HCMV inhibitors (Fig. 1).

Figure 1.

Flow chart of HTS.

2.2. Validation of anti-HCMV activity:

Additional studies including luciferase activity, plaque reduction and toxicity assays were performed on 20 compounds which passed the primary and secondary screen. This iterative process resulted in the identification of five chemotypes (Fig. 2) that were selected for further characterization and validation in several antiviral assays, using compounds that were resynthesized or obtained from vendors. Each of the five compounds was tested in multiple virological and molecular assays and is represented in a unique color in each figure and the table of contents. The last four digits of each compound are used in all subsequent figures.

Figure 2.

Chemical structure and MW of the five compounds that resulted from HTS.

HCMV inhibition was measured by luciferase activity of pp28-recombinant Towne and pp28 GCV-resistant Towne strains of HCMV (72 hpi) and by plaque reduction using HCMV Towne or TB40 (at 8–10 days). Western blot was used to measure viral protein expression. A colorimetric MTT assay was performed in non-infected HFFs to determine cell cytotoxicity in parallel to the infectivity assay, i.e. at 72 h and at 10 days, and is represented as CC₅₀. All five compounds were active at sub to low μM concentrations against all the HCMV strains used. Based on plaque reduction, MLS8969 had the best activity against HCMV followed by MLS8091, MLS8554, NCGC2955 and NFU1827 (Table 1). The slope parameter (similar to the Hill coefficient) was calculated for each compound. A slope equal to 1 signifies either one ligand binding site or a lack of cooperativity between multiple sites of binding. Positive cooperativity indicates that binding of a ligand to one site strengthens affinity of that ligand at a secondary site (slope>1). In contrast, negative cooperativity (slope<1) indicates binding towards one site weakens ligand affinity at a secondary site. In HIV infection, modeling the slope parameter showed that this parameter could distinguish drug classes with distinct mechanisms of action. Moreover, the slope was reduced even in cases where the EC₅₀ value was the same in resistant virus mutants^22,23. Based on plaque reduction assays we found that all compounds except for MLS8969 had a slope near 1.

Table 1.

Anti-HCMV activity of the five compounds against different strains of HCMV, toxicity and slope.

Compound	EC50 - Towne (μM)	EC50 - GCVR Towne (μM)	EC50 - TB40 [μM]	CC50 (μM)	Slope
MLS8969	0.55±0.1*	0.25±0.1*	0.12±0.0*	>500	0.64±0.1*
NFU1827	0.85±0.1	0.84±0.1	1.21±0.1	70.9	1.27±0.1
MLS8554	0.64±0.1	0.17+0.0	0.34±0.0	86.0	1.02±0.1
MLS8091	0.39±0.0	0.26±0.0	0.33±0.0	>250	1.37±0.3
NCGC2955	4.61±1.4	1.4±0.2**	1.18±0.2	>200	0.80±0.1
GCV	2.1±0.3	29.7±1.2	0.72±0.0	>200	0.93±0.1

^* = pre treatment followed by plaque reduction assay

^** = plaque reduction assay

2.3. Inhibition of other herpesviruses:

Inhibition of herpes simplex 1 (HSV1) and herpes simplex 2 (HSV2), mouse CMV (MCMV), and Epstein Barr virus (EBV) was tested with the five compounds. Plaque reduction assays were used for HSV1, HSV2 and MCMV. Inhibition of lytic EBV replication was quantified in AKATA cells following IgG induction by real-time PCR. Only MLS8554 was active against HSV1, HSV2 and MCMV (Table 2). NCGC2955 was active against MCMV (EC₅₀ of 1±0.0 μM, Table 2), but toxicity to MEFs was observed at 10 μM and higher concentrations. NFU1827 showed activity against EBV (EC₅₀ − 1.92 ±0.1 μM). The other compounds, MLS8969 and MLS8091 had specific activity against HCMV and did not inhibit HSV1, HSV2, EBV or MCMV.

Table 2.

Activity of the five compounds against other human herpes viruses and mouse CMV.

Compound	EC50 - HSV1-CI (μM)	EC50 - HSV2-CI (μM)	EC50 - EBV [μM]	EC50 - MCMV (μM)
MLS8969	NA	NA	>10	NA
NFU1827	NA	NA	1.92±0.1	NA
MLS8554	1.78±0.2	1.3±0.1	NA	11.67±2.0
MLS8091	NA	NA	>10	NA
NCGC2955	NA	NA	NA	1.0±0.0
GCV	0.04±0.0	0.58±0.0	5.2±0.4	0.96±0.1

NA- Not active; CI- Clinical isolate

2.4. HCMV inhibition timing:

Add-on and removal assays were carried out at 0, 6, 12, 24, 36 and 48 hpi, and luciferase activity was quantified at 72 hpi. Based on these assays each compound displayed a specific time of maximal activity. The compounds are reported below based on the timing of maximal virus suppression.

2.5. Inhibition of HCMV entry:

Compound MLS8969 was first tested after infection at MOI of 1 PFU/cell using a luciferase assay at 72 hpi, but minimal reduction in luciferase activity was observed (Fig. 3A). However, a plaque assay showed good anti-HCMV activity of MLS8969 (Fig. 3B).

Figure 3.

MLS8969 inhibits HCMV replication. A, Following infection of HFFs with HCMV pp28-luciferase Towne (MOI= 1 PFU/cell), cells were treated with MLS8969 or GCV. Luciferase activity was measured at 72 hpi. B, Following infection of HFFs with HCMV TB40 (100 PFU/well), cells were treated with MLS8969 or GCV. Plaques were stained and enumerated after 10 days. C, HFFs were pretreated with MLS8969 or GCV for 24 h followed by infection with HCMV Towne at MOIs 1, 0.1, and 0.01. Luciferase activity was quantified after 3 days. D, E, HFFs were pretreated with the specified concentrations of MLS8969 or GCV for 24 h followed by infection with HCMV TB40 (D) or GCV-resistant HCMV Towne (E), at 100 PFU/well. After 10 days plaques were stained and enumerated. A-E Data are average ± SD of triplicate values of a representative experiment. F, HFFs were either pre-treated or infected followed by treatment. Infection (Towne) was performed at MOI 1, 0.1 and 0.01. The expression of viral proteins from cell lysates was quantified at 72 hpi. WB images are from a representative experiment.

Because in the high-throughput screens HFFs were treated with compounds before infection (pre-treatment), and our plaque assay, in which MLS8969 was added after infection, revealed an EC₅₀ of 0.12 ± 0.0 μM (Table 1), we suspected lack of inhibition of luciferase activity could represent a MOI-dependent phenomenon, and thus suggestive of entry inhibition²⁴. The assay was therefore performed using a range of MOIs (0.01, 0.1 and PFU/cell) (Fig. 3C) and MLS8969 was added before infection. MOI-dependent inhibition of luciferase activity was observed (Fig. 3C). TB40 and a GCV-resistant HCMV Towne were also inhibited with MLS8969 pretreatment in the plaque reduction assay (Fig. 3D&E). Western blot analysis corroborated these data, showing no inhibition of viral protein expression at post or pretreatment at MOI of 1 PFU/cell, while pretreatment followed by infection at low MOI (0.01) resulted in decreased IE1/2, pp65, UL84 and UL44 protein levels (Fig. 3F). Further evidence for MLS8969 being an entry inhibitor was obtained by an indirect immunofluorescence assay (IFA) for HCMV pp65 (Fig. 4). MLS8969 inhibited HCMV entry only at MOI of 0.1; irrespective of MOI, GCV did not affect pp65 nuclear staining (Fig. 4). Heparin (positive control) inhibited HCMV entry.

Figure 4.

MLS8969 is an entry inhibitor of HCMV. HFFs were pretreated (24 h) with MLS8969 (3 μM) followed by infection with HCMV Towne (MOI = 1, 0.1, 0.01). Heparin (30 μM) was used as a positive control for blocking viral entry into the cells and GCV pretreatment (5 μM) was used as a negative control. IFA for HCMV-encoded pp65 was performed 2 hpi to determine the amount of pp65 localized in the nucleus under the different conditions. The experiment was repeated independently three times, and images from a representative experiment are shown.

2.6. Immediate early-to early stage inhibitors of HCMV replication:

Two compounds showed anti-HCMV activity between 0–24 hpi (NFU1827 and MLS8554) NFU1827 was active against HCMV and GCV-R HCMV, EC₅₀ 0.85±0.1 and 0.84±0.1 μM, respectively (Table 1, Fig. 5 A&B), as well as EBV (EC₅₀ 1.21±0.1 μM), but had no activity against HSV1&2 or MCMV (Table 2). Add on and removal assays revealed an immediate early to early time of activity. When added to infected cells after 24 h it lost its ability to inhibit HCMV. Removal of NFU1827 after 6 h already achieved significant HCMV suppression (Fig. 5 C&D). NFU1827 reduced the level of both IE1 and IE2 (Fig. 5E). In accordance with its immediate early to early activity, viral DNA replication and DNA yield in supernatants were significantly reduced (Fig. 5 F&G).

Figure 5.

NFU1827 is an immediate early-to early inhibitor of HCMV replication. A, Following infection of HFFs with HCMV TB40 (100 PFU/cell), cells were treated with the specified concentrations of NFU1827. At 8 dpi plaques were stained and enumerated. The experiment was repeated three times and data shown are average ± SD of triplicate values of a demonstrative experiment. B, Following infection of HFFs with a GCV-resistant HCMV Towne (MOI = 0.1), cells were treated with the specified concentrations of NFU1827. Luciferase activity was assayed in cell lysates at 72 hpi. Data shown are average ± SD of triplicate values of a representative experiment. C, D, HFFs were infected with HCMV Towne (MOI = 1 PFU/cell) followed by addition or removal of compounds (0, 6, 12, 24, 48, 72 hpi). Luciferase activity was quantified in cell lysates after 3 days. Data shown are average ± SD of triplicate values of a demonstrative experiment. E, Following infection of HFFs with HCMV Towne (MOI = 1), cells were treated with the specified concentrations of compounds. The expression level of viral proteins and cellular β-actin was determined at 72 hpi. The experiment was repeated twice, and WB images are from a demonstrative experiment. F, G Following infection of HFFs with HCMV Towne (MOI = 0.1), cells were treated with NFU1827 (5 μM) or GCV (5 μM). Viral DNA replication in cells (F), and viral DNA load in supernatants (G) was quantified by US17 real-time PCR. The experiment was repeated three times and data shown are average ± SD of triplicate values of a demonstrative experiment.

MLS8554 was active in most antiviral assays inhibiting HCMV, GCV-R HCMV, HSV1&2 and MCMV (Fig. 6, Table 1&2). Its maximal activity was measured at immediate early to early time after infection. When added at 24 hpi or after, its activity was significantly reduced, and when removed at 6 h HCMV suppression was already near complete (Fig. 6 C&D). Western blot showed reduced IE2, UL44, UL84 and pp65 protein levels at 72 hpi (Fig. 6E). In contrast to NFU1827 which reduced both IE1 and IE2, MLS8554 mostly reduced IE2 expression. Both viral DNA replication and DNA yield in supernatants were significantly reduced (Fig. 6 F&G).

Figure 6.

MLS8554 is an immediate early to early inhibitor of HCMV replication. A, Following infection of HFFs with HCMV TB40 (100 PFU/well), cells were treated with the specified concentrations of MLS8554. Plaques were stained and enumerated at 8 dpi. The experiment was repeated three times and data from a representative experiment are shown. B, Following infection of HFFs with a GCV-resistant HCMV Towne (MOI = 0.1), cells were treated with the specified concentrations of MLS8554. Luciferase activity was measured in cell lysates at 72 hpi. Data shown are average ± SD of triplicate values from a single experiment. C, D Following infection of HFFs with HCMV Towne (MOI = 1) compounds were either added or removed at the specified time points. Luciferase activity was quantified in the cell lysates after 3 days. The experiment was repeated independently twice, and representative data from a single experiment are shown. E, Following infection of HFFs with HCMV Towne (MOI = 1), cells were treated with MLS8554 or GCV at the indicated concentrations. The expression level of viral proteins IE1/2, UL84, UL44, pp65 and cellular β-actin was determined after 3 days. WB data are from a representative experiment. F, G HFFs were infected with HCMV Towne (MOI = 0.1) followed by treatment with MLS8554 (2 μM) or GCV (5 μM). Supernatants were harvested and cells were lysed at the specified time points to isolate DNA. Viral DNA replication in cells (F), and viral DNA load in supernatants (G) were quantified by US17 real-time PCR. Data shown are average ± SD of quadruplicate values from two separate experiments.

2.7. Early to late stage inhibitors of HCMV replication:

MLS8091 showed a dose-response against HCMV and GCV-resistant HCMV with EC₅₀ values of 0.39 ± .0.5 and 0.26±0.0 μM, respectively (Table 1, Fig. 7), but was not active against HSV1&2, EBV, or MCMV (Table 2). The timing of MLS8091 activity overlapped with that of GCV in the add-on assay, but in the removal assay MLS8091 showed earlier effects (Fig. 7 C&D). Inhibition of all viral proteins was observed at 72 hpi (Fig. 7E). Inhibition of viral DNA replication was higher than that achieved with GCV, although the effects on DNA yield were not significantly different from GCV (Fig. 7 F&G).

Figure 7.

MLS8091 inhibits HCMV replication at early to late stage. A, Following infection of HFFs with HCMV TB40 (100 PFU/well), cells were treated with the specified concentrations of MLS8091. At 8 dpi plaques were stained and enumerated. The experiment was repeated independently three times, and data from a single representative experiment are shown. B, Following infection of HFFs with a GCV-resistant HCMV Towne (MOI = 0.1), cells were treated with the specified concentrations of MLS8091. Luciferase activity was quantified in cell lysates (72 hpi). Data shown are average ± SD of triplicate values from a single experiment. C, D HFFs were infected with HCMV Towne (MOI = 1) followed by addition or removal of compounds at different time points after infection (0, 6, 24, 48, 72 h). Luciferase activity was quantified in the cell lysates after 3 days. The experiment was repeated independently twice, and representative data from a single experiment are shown. E, Following infection of HFFs with HCMV Towne (MOI = 1), expression level of viral proteins and cellular β-actin was determined at 72 hpi. WB data are from a representative experiment. F, G HCMV-infected HFFs (MOI = 0.1) were treated with MLS8091 (1.5 μM) or GCV (5 μM). Supernatants were harvested and cells were lysed at specified time points to isolate DNA. Viral DNA replication (F), and viral DNA load in supernatants (G) were quantified by real-time PCR. Experiments were repeated independently twice. Data shown are average ± SD of triplicate values from a single experiment.

NCGC2955 inhibited HCMV and GCV-R HCMV (Fig. 8 A&B). For this compound plaque assays were preferred, since it did not show a full dose response with the pp28-luciferase assay. The timing of NCGC2955 activity overlapped with that of GCV in the add-on assay, but in the removal assay NCGC2955 appeared to have longer activity than GCV (Fig. 8 C&D). At 72 hpi there was a strong reduction in viral protein expression including IE1 and IE2 (Fig. 8E). Despite inhibition of viral progeny by plaque assays and the observed decrease in viral protein levels, the influence of NCGC2955 on viral DNA replication and viral DNA yield was modest (Fig. 8 F&G), suggesting inhibition via mechanisms that do not involve the DNA replication machinery.

Figure 8.

NCGC2955 is an early to-late inhibitor of HCMV replication. A, Following infection of HFFs with HCMV TB40 (100 PFU/well), cells were treated with specified concentrations of NCGC2955. Plaques were stained and enumerated at 8 dpi. The experiment was repeated independently three times, and data from a single representative experiment are shown. B, Following infection of HFFs with a GCV-resistant HCMV Towne (100 PFU/well), cells were treated with the specified concentrations of NCGC2955. Plaques were stained and enumerated at 8 dpi Data shown are average ± SD of triplicate values from a single experiment. C, D HFFs were infected with HCMV Towne (MOI = 1) followed by addition or removal of compounds at different times after infection (0, 6, 24, 48, 72 h). Luciferase activity was quantified in the cell lysates after 3 days. Data shown are the average of three independent experiments (average ± SD). E, Following infection of HFFs with HCMV Towne (MOI = 1), expression levels of viral proteins was determined at 72 hpi. The experiment was repeated independently three times, and data from a single representative experiment are shown. F, G Following infection of HFFs with HCMV Towne (MOI = 0.1), cells were treated with NCGC2955 (3 μM) or GCV (5 μM). Supernatants were harvested and cells were lysed at specified time points to isolate DNA. Cellular viral DNA replication (F), and viral DNA load in supernatants (G) were quantified by real-time PCR. Data shown are the average ± SD of quadruplicate values from two independent experiments.

2.8. Patterns of drug combination with GCV and letermovir:

The effect of each compound by itself or combined with GCV or letermovir was tested in HFFs infected with HCMV. An additive or mild synergistic effect was observed with GCV (Fig. 9) or letermovir (Fig. 10) for all five agents. A ratio of observed over expected fold inhibition of > 1 was considered drug synergy, < 1 as drug antagonism, and = 1 as additive. The Bliss coefficient of drug combination is provided in Table 3. For letermovir and its combination with each of the five compounds the EC₅₀ was determined based on plaque assay. The combination of GCV and 2955 was also tested by plaque assay. All other combinations and Bliss coefficients were measured by pp28-luciferase. All Bliss coefficients were close to 1, suggesting an additive effect between the compounds.

Figure 9.

Combination of newly identified HCMV inhibitors and GCV is additive. A-E HFFs were infected with HCMV Towne (MOI = 1) and first treated with each individual compound, followed by combination of each compound with GCV at different doses. In the drug combination experiments HCMV-infected HFFs were treated with an initial drug concentration of twice the EC₅₀ value of individual compound and two-fold serial dilutions. Luciferase activity was quantified in cell lysates after 3 days and the Bliss Model was used to calculate the antiviral activity of compounds in combination. Solid lines indicate observed HCMV inhibition (dose- response) and dotted lines indicate expected HCMV inhibition at each dose of drug combination. A-D Experiments were repeated three times independently. Data from a single representative experiment are shown. E, Data shown represent average ± SD of triplicate values from a single experiment.

Figure 10.

Combination of newly identified HCMV inhibitors with letermovir is additive. A-E HFFs were infected with TB40 (100 PFUs/well) and treated with each compound alone, and combination of each compound with letermovir at different doses. For drug combination HCMV-infected HFFs were treated with starting concentration of twice the EC₅₀ value of individual compound followed by two-fold serial dilutions. The number of plaques in each condition was enumerated at 10 dpi and antiviral activity of compounds was calculated using the Bliss Model. Solid lines indicate observed HCMV inhibition (dose-response) and dotted lines indicate expected HCMV inhibition at each dose of drug combination. Data shown are average ± SD of triplicate values from a single experiment.

Table 3.

Bliss coefficient for drug combination. The EC₅₀ was determined for each compound in the combination experiment. For the letermovir combinations, a plaque reduction assay was performed with HCMV Towne. The GCV combinations used the pp28-luciferase HCMV Towne except for GCV and NCGC2995 which was performed with HCMV TB40. Letermovir combination assays were performed in a single experiment, whereas GCV combination with the specified compounds was performed in separate experiments for which the EC₅₀ was determined.

Compound 1	EC50	Compound 2	EC50 (μM)	Bliss Coefficient
Letermovir	1.1 ± 0.1 (nM)	MLS8969	0.32 ±0.0	1.2
		NFU1827	0.61 ±0.1	1.2
		MLS8554	0.14 ±0.0	1.0
		MLS8091	0.13 ±0.1	1.3
		NCGC2955	3.19 ±0.2	1.1
GCV	1.33 ± 0.2 (μM)	MLS8969	0.72 ±0.1	0.9
GCV	1.63 ±0.2	NFU1827	0.69 ± 0.1	1.2
GCV	1.91 ±0.6	MLS8554	0.55 ± 0.0	1.1
GCV	1.25 ±0.6	MLS8091	0.57 ±0.1	1.3
GCV	0.77 ±0.1	NCGC2955	1.51 ±0.0	0.9

3. DISCUSSION AND CONCLUSIONS

We report the results of the largest HTS to date for HCMV inhibitors using our pp28-luciferase HCMV. This reporter virus has shown high sensitivity and reproducibility for such screening²⁰. Its advantage is that HCMV-encoded pp28 is a true late gene, thus its activation represents a near complete cycle of HCMV replication.

Screening campaigns may result in false positive hits and require intense validation. Our combined approach using multiple HCMV strains and several detailed anti-viral assays resulted in the identification and characterization of five distinct hits, each with its unique characteristics. Of the five hits MLS8969 inhibited HCMV entry, MLS8554 and NFU1827 inhibited an immediate early to early stage of HCMV replication (but were different from one another in their effects on other herpesviruses), MLS8091 and NCGC2955 were active at an early to late stage, but NCGC2955 did not have strong suppression of viral DNA replication. All five compounds demonstrated additive effects or mild synergy when combined with GCV or letermovir. No antagonism was found in any of the drug combination experiments. In addition, the five compounds were active against GCV-resistant HCMV, indicating a mechanism separate from the viral DNA polymerase. Future studies should provide insights into the mechanism of action of each of five compounds. Based on data presented here, we expect each compound to have a unique mechanism for HCMV suppression.

Screens with libraries containing a limited number of small molecules, different HCMV strains, cell types and timing of treatment have been performed. Three recombinant HCMV strains carrying enhanced yellow fluorescent protein (EYFP) fusions of the immediate early 2 (IE-2), ppUL32 (pp150), and ppUL83 (pp65) viral proteins were tested against a cellular kinase library (80 compounds) as a pilot screen. Several kinase inhibitors blocked HCMV replication as shown by decreased fluorescent intensity²⁵. This screen used a fixed drug concentration (10 μM) at time of infection and a limited number of hits had no apparent cell toxicity. Reporter lines have previously been created to screen for anti-HCMV compounds^{26, 27}. One approach used a promoter in front of a luciferase reporter that was activated by IE proteins. Consequently, inhibitors of later HCMV infection stages could not be discovered using this reporter cell line²⁷. The Gray Kinase Inhibitor library (187 compounds) was screened against the HCMV strain AD169, and cells were immunostained for the HCMV antigen pp28. Three kinase inhibitors were identified as inhibitors of IE2 production²⁸. Recent drug repurposing screening used the Spectrum Collection of 2560 compounds from Micro Discovery System, Inc^{29, 30}. Mercorelli used a mechanism-based screen that identified compounds that interfered with the transactivating activity of IE2. Phenotypic assays measured the expression of EGFP fused to IE2³¹ or UL9 9³². A major success of HTS was the identification of letermovir (AIC246) from a screen of a compound library³³, which in subsequent studies was shown to harbor a novel mode of HCMV inhibition³⁴.

In conclusion, our study provides several novel compounds that inhibit HCMV replication at sub to low μM concentrations. Each of these compounds represents a good candidate for future mechanistic studies and drug development.

4. EXPERIMENTAL SECTION:

4.1. Compounds:

The compounds were identified from a high throughput screen (HTS) and then validated using several HCMV strains, multiple antiviral assays and new compound preparations. In addition to compounds obtained from the libraries used in the screening, NFU1827 was synthesized in the Department of Biochemistry, Program in Chemical Biology at the Medical College of Wisconsin, MLS8091 was purchased from Princeton Biomolecular Research, Inc., MLS8969 was purchased from ChemBridge Corporation, San Diego, CA, MLS8554 was purchased from InterBioScreen STOCK1N-59867, and NCGC2955 was synthesized in the Chemistry Core, Johns Hopkins University School of Medicine. The purity of all synthesized and purchased compounds used in bioassays was determined by HPLC using either a Phenomenex Luna C18 3.0 × 75 mm column with a 7 min gradient of 4–100% ACN in H₂O with 0.05% v/v TFA or a Higgins Analytical, Inc. Targa C18 5 μm 4.6 × 150 mm column with a 30 min gradient of 0–100% ACN in H₂O with 0.01% TFA and either absorbance detection at 254 nm or evaporative light scattering detection. All compounds used in bioassays exhibited NMR and MS data consistent with their structures and purities of >95% as determined by HPLC.

Dimethyl sulfoxide (DMSO) was used to dissolve compounds, and stock solutions of 50 and 10 mM were stored at −80 °C. Ganciclovir was purchased from Sigma Aldrich (St. Louis, MO), a 10 mM stock solution was prepared in ddH₂O.

Synthesis of NCGC2955:

tert-butyl 4-(isopropylcarbamoyl) piperidine-1-carboxylate (1). To a solution of 1-(tert-butoxycarbonyl) piperidine-4-carboxylic acid³⁵ (3.94 g, 17.2 mmol, 1.0 equiv.) in CH₂Cl₂ (39 mL) was added N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (3.46 g, 18.0 mmol, 1.05 equiv.), 1-hydroxybenzotriazole monohydrate (2.83 g, 18.5 mmol, 1.075 equiv.) and isopropylamine (5.1 mL, 60.15 mmol, 3.5 equiv.). After stirring at room temperature for 21 h, the reaction was diluted with CH₂Cl₂, and washed with 5% aq. HCl. The aqueous layer was extracted with CH₂Cl₂ (3 × 20 mL). The organic layers were combined, dried with anhydrous MgSO₄, and concentrated in vacuo. Purification by flash chromatography (0 to 100% EtOAc/hexanes) provided 2.95 g compound 1 as a solid in 63% yield. ¹H NMR (500 MHz, CDCl₃) δ 5.28 (d, J = 6.45 Hz, 1H), 4.01 – 4.25 (m, 3H), 2.73 (br. s., 2H), 2.16 (tt, J = 3.69, 11.63 Hz, 1H), 1.79 (d, J = 11.32 Hz, 2H), 1.61 (dq, J = 4.24, 12.42 Hz, 2H), 1.45 (s, 9H), 1.14 (d, J = 6.60 Hz, 6H).

N-isopropylpiperidine-4-carboxamide trifluoroacetic acid salt (2). To a solution of compound 1 (2.95 g, 10.9 mmol) in CH2Cl2 (5 mL) was added triisopropylsilane (250 μL) and trifluoroacetic acid (3 mL) at room temperature. After stirring for 24 h at room temperature, the volatiles were removed in vacuo providing a viscous syrup used without further purification. 1H NMR (500 MHz, CDCl3) δ 4.06 (qd, J = 6.86, 13.38 Hz, 1H), 3.57 (d, J = 12.89 Hz, 2H), 3.10 (br. s., 2H), 2.45 – 2.63 (m, 1H), 2.05 – 2.16 (m, 4H), 1.18 (d, J = 6.60 Hz, 6H).

methyl 4-(4-chlorobenzyl)-4H-thieno[3,2-b] pyrrole-5-carboxylate (3). To a suspension of methyl 4H-thieno[3,2-b] pyrrole-5- carboxylate36 (1.44 g, 7.99 mmol, 1.0 equiv.) and cesium carbonate (3.91 g, 12.0 mmol, 1.5 equiv.) in anhydrous DMF (10 mL) was added 4-chlorobenzyl bromide (1.97 g, 9.59 mmol, 1.2 equiv.) in one portion at rt. After stirring at rt until complete (ca. 4 h), the reaction was diluted with water and extracted with EtOAc (3 × 20 mL). The organic layers were combined, washed with brine, dried with anhydrous MgSO4, and concentrated in vacuo. Purification by flash chromatography (3m 0%, 6m gradient 0 to 50%, 2m 50% EtOAc/hexanes) provided 2.40 g of methyl ester 3 as a pale yellow solid in 98% yield. 1H NMR (500 MHz, CDCl3) δ 7.33 (d, J = 5.34 Hz, 1H), 7.21 – 7.29 (m, 3H), 7.05 (d, J = 8.17 Hz, 2H), 6.85 (d, J = 5.50 Hz, 1H), 5.71 (s, 2H), 3.83 (s, 3H).

4-(4-chlorobenzyl)-4H-thieno[3,2-b] pyrrole-5-carboxylic acid (4). To a mixture of methyl ester 3 (3.58 g, 11.7 mmol, 1.0 equiv.) in THF (12 mL) and MeOH (12 mL) was added a solution of lithium hydroxide monohydrate in water (12 mL) in one portion. After 2 d of rapid magnetic stirring at rt, the starting material was consumed, and the organic solvents were removed in vacuo. The remaining aqueous solution was diluted with water (ca. 20 mL) and under vigorous magnetic stirring the pH was adjusted to pH = 1 with conc. aq. HCl. A precipitate formed during acidification, and was collected via vacuum filtration, washed with water and dried under vacuum to provide 3.34 g of compound 4 in 98% yield. 1H NMR (500 MHz, DMSO-d6) δ 12.57 (br. s., 1H), 7.56 (d, J = 5.19 Hz, 1H), 7.29 – 7.43 (m, J = 8.49 Hz, 2H), 7.17 – 7.29 (m, 2H), 7.05 – 7.17 (m, J = 8.49 Hz, 2H), 5.76 (s, 2H).

1-(4-(4-chIorobenzyl)-4H-thieno[3,2-b] pyrrole-5-carbonyl)-N-isopropyIpiperidine-4-carboxamide (NCGC2955). To a mixture of compound 4 (51.9 mg, 0.177 mmol, 1.0 equiv.) in CH₂Cl₂ (1.7 mL) was added N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (37.5 mg, 0.19 mmol, 1.1 equiv.), 1-hydroxybenzotriazole monohydrate (31.0 mg, 0.204 mmol, 1.15 equiv.) and diisopropylethylamine (123 μL, 0.71 mmol, 4 equiv.). After stirring at room temperature for 5 min, TFA salt 2 (60.7 mg, 0.21 mmol, 1.2 equiv.) was added as a solution in CH₂Cl₂ (0.2 mL, plus 0.2 mL rinse) and the reaction stirred at rt. After 24 h, the reaction was diluted with CH₂Cl₂, and washed with 5% aq. HCl. The aqueous layer was extracted with EtOAc (3 × 10 mL). The organic layers were combined, dried with anhydrous MgSO₄, and concentrated in vacuo. Purification by flash chromatography (0 to 100% EtOAc/hexanes) provided 48.9 mg of NCGC2955 as a white solid in 62% yield. ¹H NMR (500 MHz, CDCl₃) δ 7.24 (d, J = 8.33 Hz, 2H), 7.18 (d, J = 5.19 Hz, 1H), 7.08 (d, J = 8.33 Hz, 2H), 6.83 (d, J = 5.19 Hz, 1H), 6.58 (s, 1H), 5.43 (s, 2H), 5.19 (d, J = 8.02 Hz, 1H), 4.44 (d, J = 12.89 Hz, 2H), 4.08 (qd, J = 6.77, 13.34 Hz, 1H), 2.92 (t, J = 12.34 Hz, 2H), 2.24 (tt, J = 3.62, 11.08 Hz, 1H), 1.80 (d, J = 14.93 Hz, 2H), 1.43 – 1.63 (m, 6H), 1.15 (d, J = 6.60 Hz, 6H).

Synthesis of NFU1827:

4-(4-methoxyphenyl)-7-thia-2,5-diazatricyclo [6.4.0.02,6] dodeca- 1(8),3,5,9,11-pentaene-10-carboxylic acid (5). 2-Aminobenzothiazole-6-carboxylic acid (2.0 g, 10.30 mmol, 1.0 equiv.) and 2-bromo-4’- methoxyacetophenone (2.60 g, 11.33 mmol, 1.1 equiv.) were charged to a pressure bottle and suspended in 2-methoxyethanol (50 mL). The suspension was heated to 40 °C for 24 h and at 140 °C for 19 h. The reaction cooled to rt and concentrated in vacuo. The crude compound was purified by flash column chromatography (5 to 75% EtOAc/hexanes) to afford 1.83 g of compound 6 in 55% yield. ¹H NMR (500 MHz, DMSO-d₆) δ 13.17 (broad s, 1H), 8.69 (s, 1H), 8.63 (d, J = 1.2 Hz, 1H), 8.10 (dd, J = 1.6, 8.3 Hz, 1H), 8.01 (d, J = 8.4 Hz, 1H), 7.78 (d, J = 8.7 Hz, 1H), 6.99 (d, J = 8.8 Hz, 1H), 3.77 (s, 3H).

N-[3-(4-ethylpiperazin-1-yl)propyl]-4-(4-methoxyphenyl)-7-thia-2,5-diazatricyclo [6.4.0.02,6] dodeca-1(8),3,5,9,11-pentaene-10-carboxamide (NFU1827). Compound 5 (100 mg, 0.308 mmol), N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (118.2 mg, 0.617 mmol), and 1-hydroxybenzotriazole hydrate (83.3 mg, 0.617 mmol) were dissolved in DMF and stirred at room temperature for 1 h. The reaction was then treated with 3-(4-ethylpiperazin-1-yl) propan-1-amine (0.2269 mL, 211.2 mg, 1.23 mmol) and stirred at room temperature for 18 h. The reaction was then poured into H₂O (20 mL). This was extracted with EtOAc (3 × 20 mL). The combined extracts were washed with H₂O (4 × 20 mL) followed by brine (1 × 20 mL). The organic layer was dried (Na₂SO₄), filtered, and the solvent removed in vacuo. The crude compound was purified by flash column chromatography using hexanes/EtOAc (gradient elution from 90:10 to 0:100) to afford 60 mg of NFU1827 in 41% yield. 1H NMR (500 MHz, CDCl3) δ 8.71 (s, 1H), 8.21 (s, 1H), 7.99 (d, J = 8.4 Hz, 1H), 7.91 (s, 1H), 7.80 (d, J = 8.7 Hz, 2H), 7.62 (d, J = 8.4 Hz, 1H), 6.96 (d, J = 8.7 Hz, 2H), 3.85 (s, 3H), 3.60 (q, J= 5.2 Hz, 2H), 2.63 (t, J = 5.5 Hz, 2H), 2.46 (q, J = 7.3 Hz, 2H), 1.82 (m, 2H), 1.78 (broad s, 8 H), 1.10 (t, J = 7.2 Hz, 3H). 13C NMR (126 MHz, CDCl3) δ 162.72, 156.34, 145.14, 130.89, 128.49, 127.24, 123.49, 123.25, 122.78, 120.26, 111.13, 109.09, 102.92, 57.32, 55.69, 52.27, 50.25, 49.91, 49.39, 38.42, 26.62, 20.55, 17.98, 11.12. HRMS (ESI⁺) m/z calcd for C₂₆H₃₁N₅O₂S [M + H]⁺ 478.2272, found 478.2271.

4.2. Cells:

Human Foreskin fibroblasts (HFFs), passage 12 to 16 (ATCC, CRL-2088) were grown in Dulbecco’s modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS) (Gibco, Carlsbad, CA) in a 5% CO₂ incubator at 37 °C. Mouse embryonic fibroblasts (MEFs; ATCC, CRL-1658) were used for infection with mouse CMV. Vero cells were obtained from the laboratory of Dr. Gary Hayward, Johns Hopkins University School of Medicine.

4.3. Cytotoxicity assay:

MTT assay was performed as described by manufacturer’s instructions (Sigma-Aldrich). Non-infected cells were treated with compounds for 72 h or 8–10 days (same time points as the antiviral assay), and 20 μL/well of MTT [3-(4, 5-Dimethyl-2-thiazolyl)-2, 5-diphenyl-2 H-tetrazolium bromide], 5 mg/mL in phosphate buffered saline (PBS)] was added to each well. After shaking at 150 rpm for 5 min the plates were incubated at 37 °C for 2–3 h. Conversion of yellow solution to dark blue formazan by mitochondrial dehydrogenases of living cells was quantified by measuring absorbance at 560 nm.

4.4. Viruses and antiviral assays:

A pp28-luciferase recombinant Towne HCMV strain expressing luciferase under the control of the UL99 (pp28) late promoter was reported to provide a sensitive and reproducible reporter for drug screening²⁰. A GCV-resistant HCMV pp28-luciferase that contains a C607Y mutation in UL97 was reported³⁷. The HCMV TB40 strain was obtained from ATCC (VR-1578). Clinical isolates of Human Herpesvirus 1 and 2 (HSV1, HSV2) were collected from the Johns Hopkins Microbiology laboratory without identifiers that could link them to a specific patient. Mouse CMV (MCMV) Smith strain (ATCC VR-1399) was used for infection of MEFs.

4.5. Luciferase activity:

Cell lysates were collected at 72 hpi and luciferase activity was determined using the Glomax-Multi + Detection System (Promega, Madison, WI) as described previously²⁰.

4.6. Plaque reduction assay:

HFFs were seeded into 12-well plates (2 × 10⁵ cells/well) and infected with HCMV TB40 at approximately 100 plaques/well. After 90 min, media were aspirated, and DMEM containing 0.5% carboxymethyl-cellulose (CMC), 4% fetal bovine serum (FBS), and compounds were added into duplicate wells. Following incubation at 37 °C for 8–10 days the overlay was removed, and plaques were counted after crystal violet staining. For HSV1 and HSV2 replication in Vero cells, the adsorption time was 60 min and plaques were counted after 36 h. Plaque assays in MEFs were completed after 3 days.

4.7. Inhibition of lytic Epstein Barr Virus (EBV) replication:

AKATA cells (0.5 × 10⁶/ well of a 24-well plate) latently infected with EBV were induced with 100 μg/mL of goat antihuman IgG (Sigma). Compounds and GCV control were added to triplicate wells right after addition of human IgG. Cells were collected 48 h after lytic induction for DNA quantification. Cellular DNA was purified using the Wizard SV Genomic Kit (Promega, Madison, WI). The effect of compounds on lytic induction of EBV was analyzed by real-time PCR of the EBV DNA polymerase gene, BALF5³⁸.

4.8. Add-on and removal assays:

These assays were performed to identify at which stage during HCMV replication the compounds had maximal activity. In the add-on group, compounds were added to infected HFFs at 0, 6, 24, and 48 hpi, and luciferase activity was measured at 72 hpi. In the removal group, compounds were added immediately after virus infection and were subsequently removed after 0, 6, 24, and 48 h; luciferase was measured at 72 hpi.

4.9. DNA isolation and quantitative real-time (qPCR):

Total DNA was isolated from non-infected control and HCMV-infected HFFs using the Wizard SV genomic DNA isolation kit (Promega, Madison, WI). To determine viral load in the supernatant, total DNA was isolated from supernatants using automated DNA extraction on a BioRobot M48 instrument (Qiagen, Valencia, CA). A US17 real-time PCR assay that targets 151-bp from the highly conserved US17 region of the CMV genome was used³⁹. The primers and probe used for US17 were: forward 5′-GCGTGCTTTTTAGCCTCTGCA-3′, reverse 5′-AAAAGTTTGTGCCCCAACGGTA-3′ and US17 probe FAM-5′ TGATCGGGCGTTATCGCGTTCT-3′.

4.10. HCMV entry and indirect immunofluorescence assay:

MLS8969, GCV and heparin control were used to determine inhibition of HCMV entry. Compounds were diluted in serum-free media and combined with HFFs seeded on chamber slides 24 h prior to infection. After infection (MOI 1 or 0.1 PFU/cell) for 2 h, cells were fixed with 100% chilled methanol, blocked for 1 h with phosphate-buffered saline (PBS), 5% serum, 0.3% Triton X-100. Cells were then incubated with mouse monoclonal anti-pp65 antibody, 1:50 dilution (Vector Laboratories, Burlingame, CA) at 37°C in humidified chambers for 1 h, washed three times with TBST (0.1%), incubated with rhodamine-conjugated anti-mouse IgG, 1:500 dilution (Sigma) at 37 °C in humidified chambers for 1 h and washed with TBST (0.1%). A drop of mount oil containing DAPI (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the slides before visualization with a Zeiss ZI fluorescence microscope. Images were captured at 40x magnification.

4.11. SDS-polyacrylamide gel electrophoresis and immunoblot analysis:

Cell lysates containing an equal amount of protein were mixed with an equal volume of sample buffer (125 mM Tris-HCl, pH 6.8, 4% SDS, 20% glycerol and 5% β-mercaptoethanol) and heated for 10 min at 100 °C. Denatured proteins were resolved by Tris-glycine polyacrylamide gels (8–10%) and transferred to polyvinylidine difluoride (PVDF) membranes (Bio-Rad Laboratories, Hercules, CA) by electroblotting. Membranes were incubated in blocking buffer [5% w/v non-fat dry milk and 0.1% Tween-20 in PBS (PBST)] for 1 h, washed with PBST, and incubated with primary antibodies at 4 °C overnight. Membranes were washed with PBST and incubated with horseradish peroxidase-conjugated secondary antibodies in PBST for 1 h at room temperature. Following washing with PBST, protein bands were visualized by chemiluminescence using SuperSignal West Dura and Pico reagents (Pierce Chemical, Rockford, IL).

4.12. Antibodies:

Mouse monoclonal anti-HCMV IE1 and IE2 (MAB810) was from Millipore (Billerica, MA). Mouse monoclonal anti-HCMV UL83 (pp65) was from Vector Laboratories Inc. (Burlingame, CA). Mouse monoclonal anti-HCMV UL44, mouse monoclonal anti-HCMV UL84, and mouse anti-β-actin anti-mouse IgG was from Santa Cruz Biotechnology (Santa Cruz, CA). Horseradish peroxidase (HRP)-conjugated anti-mouse IgG was from GE Healthcare (Waukesha, WI). Horseradish peroxidase (HRP)-conjugated anti-rabbit IgG was from Cell Signaling (Beverly, MA).

4.13. Drug combination and analysis:

These experiments were performed as previously reported⁴⁰. The combination of GCV and each compound was tested using the pp28-luciferase. Briefly, 2×10⁶ HFFs/plate were seeded in 96-well plates and infected with the pp28-luciferase Towne HCMV strain (MOI=1). First, a dose-response curve was generated for each compound individually to determine its EC₅₀ value. Then, compounds were combined at twice their EC₅₀ values, diluted in DMEM with 4% FBS, followed by serial dilution and added in combination after infection. Luciferase activity of each compound individually and in combination was quantified at 72 hpi. The combinations of letermovir and each of the new compounds were tested by the plaque assay using TB40. The same principle applied in this assay for the drug combination effect, and plaques were counted at 8 dpi. The Bliss model was used to calculate the effect of each drug combination on pp28-luciferase activity and plaque reduction. In this model, the drug combination is represented in the following 4 probabilistically independent events⁴¹:

$$F_{\text{U}1+2}=F_{\text{U}1}\cdot F_{\text{U}2}=\frac{1}{1+{\left(\frac{D{}_1}{{\text{EC}}_{\text{50}\left(1\right)}}\right)}^{m_1}}\cdot \frac{1}{1+{\left(\frac{D{}_2}{{\text{EC}}_{\text{50}\left(2\right)}}\right)}^{m_2}}$$

Where D is the compound concentration, m is the slope parameter, and EC₅₀ is the effective concentration resulting in 50% virus inhibition. The combined effect of two compounds (F_U1+2) is calculated from the product of the individual effects of the two inhibitors (F_U1 and F_U2). In this model, synergistic compounds will yield a ratio > 1 of the observed fold inhibition divided by the expected fold inhibition. Antagonistic compounds will yield a ratio < 1. Additive compounds will yield a ratio = 1.

4.14. Statistical analysis:

Student t-tests were performed using Sigmaplot (Systat Software, San Jose, CA) and GraphPad Prism (GraphPad Software, La Jolla, CA). All sample groups were compared to the control group using a One-way ANOVA. P-values were adjusted for multiple comparisons. In all figures * indicates p value < 0.05, ** indicates p<0.01, and *** indicates p < 0.001. The curve fitting toolbox, MATLAB software (v7.10), MathWorks (Natick, MA) was used to determine EC₅₀ and CC₅₀ values using a four- parameter logistic regression.

ABBREVIATIONS

CMV

cytomegalovirus

MCMV

mouse cytomegalovirus

HSV

herpes simplex virus

EBV

Epstein Barr Virus

HTS

high throughput screening

HPLC

high performance liquid chromatography