Investigating N-arylpyrimidinamine (NAPA) compounds as early-stage inhibitors against human cytomegalovirus

Abstract

Human cytomegalovirus (CMV) is a ubiquitous β-herpesvirus that establishes latent asymptomatic infections in healthy individuals but can cause serious infections in immunocompromised people, resulting in increased risk of morbidity and mortality. The current FDA-approved CMV drugs target late stages of the CMV life-cycle. While these drugs are effective in most cases, they have serious drawbacks, including poor oral bioavailability, dose-limiting toxicity, and a low barrier to resistance. Given the clinical relevance of CMV-associated diseases, novel therapies are needed. Thus, a novel class of compounds that inhibits the early stages of the CMV life-cycle was identified and found to block infection of different strains in physiologically relevant cell types. This class of compounds, N-arylpyrimidinamine (NAPA), demonstrated potent anti-CMV activity against ganciclovir-sensitive and -resistant strains in in vitro replication assays, a selectivity index >30, and favorable in vitro ADME properties. Mechanism of action studies demonstrated that NAPA compounds inhibit an early step of virus infection. NAPA compounds are specific inhibitors of cytomegaloviruses and do not exhibit anti-viral activity against other herpesviruses. Collectively, we have identified a novel class of CMV inhibitor that effectively limits viral infection and proliferation.

1. Introduction

Human cytomegalovirus (CMV) is a ubiquitous pathogen with high seroprevalence (~83%) worldwide(Adane and Getawa, 2021). CMV can spread through bodily fluids, during organ transplants, and across the placenta from mother to fetus. Virus proliferation in the immunocompromised leads to increased morbidity and mortality in newborns, organ transplant recipients, AIDS patients, and the elderly. CMV is the leading infectious cause of birth defects, affecting ~0.5-2% of newborns worldwide, with up to 40,000 new cases/year of congenital CMV infection in the United States(Damato and Winnen, 2002; Goderis et al., 2014; Ssentongo et al., 2021). CMV can also exacerbate cardiovascular diseases associated with atherosclerosis (Ji et al., 2012; Lee et al., 2014) and cardiac allograft vasculopathy (Petrakopoulou et al., 2004; Simmonds et al., 2008). Given the patient populations at risk for CMV-associated diseases and the estimated cost-of-illness is $0.4-1.3billion/year for congenital CMV(Grosse et al., 2021)) and ~$4.4billion/year overall(Dove, 2006)), CMV is a significant health challenge requiring the development of a multi-faceted therapeutic strategies.

While several drugs have been approved by the FDA for treatment of CMV, including ganciclovir (GCV), foscarnet (PFA), cidofovir (CDV) and letermovir (LTV), these drugs exhibit high frequencies of drug resistance(Chou, 2008, 2015; Jabs et al., 1998; Weinberg et al., 2003) and can elicit severe side effects(Kenneson and Cannon, 2007). Recently, maribavir (LIVTENCITY^™) has been approved for post-transplant CMV disease in patients refractory to ganciclovir, valganciclovir, cidofovir, or foscarnet(Kang, 2022). Approved anti-CMV therapies target the DNA polymerase UL54 (GCV, PFA, and CDV) and the terminase protein UL56 (LTV), both of which are essential in later phases of the CMV life-cycle(Lurain and Chou, 2010). The CMV genome encodes many unexploited potential anti-viral targets for novel therapies. In particular, early stages of the CMV life-cycle mediate several essential functions, such as binding and entry into host cells, membrane fusion, transfer of the genome to the nucleus, and expression of immediate early genes, which are carried out by cellular and viral proteins.

To explore the discovery of novel drugs that act on unexploited targets, we developed a high-content screening (HCS) assay to identify inhibitors of the early stages of the CMV life-cycle, which is driven by proteins not targeted by existing drugs. The assay is comprised of a CMV reporter virus (AD169^IE2/YFP) that expresses an IE2:yellow-fluorescent protein (YFP) fusion protein(Gardner et al., 2015). We screened >112,000 compounds using AD169^IE2/YFP-infected fibroblasts(Gardner et al., 2015) and found several distinct compounds that effectively limit CMV infection. One of these, a N-arylpyrimidinamine (NAPA), exhibited potent anti-CMV activity with limited cytotoxicity. The current study describes NAPA compounds as effective inhibitors of in infection and dissemination.

2. Materials and Methods

2.1. Cells and viruses

HFF (ATCC#SCRC-1041), MRC-5 (ATCC #CCL-171), NHDF (ATCC #PSC-201-010) fibroblast cells and ARPE-19 human retinal epithelial cells (ATCC #CRL-2302) were cultured as described(Parsons et al., 2022). CMV strains AD169 (ATCC:VR-538), AD169^IE2-YFP (gift from Dr. Leor Weinberger, Gladstone Inst/UCSF), AD169 (BADrUL131-C4) containing the UL131 open reading frame of the CMV strain TR that expresses EGFP (AD169R)(Wang and Shenk, 2005) (gift from Dr. Thomas Shenk, Princeton U.), and TB40/E and TB40/E^UL32-EGFP (Sampaio et al., 2005) (gifts from Dr. Christian Sinzger, Ulm U.) were propagated and titered as described (Gardner et al., 2013).

2.2. High-content screening

The optimized screening assay was performed using 384-well assay plates (Corning #COR-3712) seeded with 3000 HFF cells (40μL DMEM) for 24 hours. The next day, 0.2μL compound (stock: 2.5mM, final: 10μM) from screening libraries was added to each well in columns 3-22 using a Caliper robot with a pin tool array. DMSO (0.2μl) was added to control columns 1-2 and 23-24. After incubation at 37°C for 1 hour, 10μL diluted AD169^IE2-YFP (multiplicity of infection, MOI=1) was added to wells in columns 1-22 and spin-infected (2000 rpm for 30min at room temperature (RT)) to achieve 30-50% infection. Columns 23-24 were uninfected controls. After 17 hours post-infection (hpi) at 37°C, cells were fixed (7.4% formaldehyde final in PBS) for 30 min at RT, then washed with PBS using a BioTek Elisa plate washer prior to adding Hoechst 33342 (30μL, 25μg/mL). Cells were washed with PBS, and 30μL PBS was added to each well prior to imaging with a Cytation 5 Cell Imaging Multi-Mode plate reader (BioTek) to quantify YFP+ and H33342-stained nuclei in 5 separate fields of view. The numbers of YFP+ and total nuclei (H33342 stained) were normalized to average counts of DMSO-treated controls (columns 1-2). Compounds that reduced YFP+ nuclei by ≥70% but had a minimal effect (≤25% reduction) on the number of total nuclei (H33342 stained) were scored as “primary hits.” Primary hits were confirmed (four technical replicates) using the screening assay, and the average reduction of YFP+ and total nuclei as compared to negative controls was determined. We screened a total of 112,197 small molecules from commercially available compound collections, comprised of compounds purchased from MicroSource Discovery (Gaylordsville, CT), Chembridge (San Diego, CA), Biomol GmbH (Hamburg, Germany), Life Chemicals (Niagara-on-the-Lake, Canada), ChemDiv (San Diego, CA), and TimTec (Newark, DE).

2.3. Dose response assay

The half-maximal inhibitory concentration (IC₅₀) of primary hits and analogs was determined by measuring infection with a two-fold compound dilution series (0-100 μM) using the screening assay protocol. Dose-response data was normalized to untreated cells, and the IC₅₀ was determined using a four-parameter curve-fitting algorithm (GraphPad Prism). Each assay was performed over three replicates and repeated at least once. The reported value is the average of at least six technical replicates.

2.4. Compound evaluation/assay funnel

NAPA analogs were evaluated for solubility, cytotoxicity, and predicted ADME properties using the following assays. Each assay was performed with three replicates repeated at least once. The reported value is the average of at least six technical replicates. Solubility. The maximum aqueous solubility of analogs was determined by nephelometry using published methods(Bevan and Lloyd, 2000). Cytotoxicity. We determined the concentration of each compound that produces a half-maximal decrease in cytotoxicity (CC₅₀) of Hela cells exposed to inhibitor (0.4-300μM) using the vital stain MTS (Abcam, ab197010). The dose-response data was normalized to untreated cells, and the CC₅₀ was determined using a four-parameter curve-fitting algorithm (GraphPad Prism) (Butler et al., 2007). A CellTiter Glo Luminescent Assay (Promega Inc, Madison, WI) was performed on MBX-4992-treated NHDF cells based on manufacture’s instructions. Mouse liver microsome stability. To examine the potential for first-pass metabolism of analogs in the liver, the half-life (T_1/2) was assessed in the presence of mouse liver microsome preparations (XenoTech) and of NADPH (Kuhnz and Gieschen, 1998). Cytochrome 450 (CYP450 3A4) inhibition. Inhibition of CYP450 3A4 by NAPA analogs was measured using commercially available kits for CYP3A4 activity (BD Gentest Corp, Woburn, MA).

MBX-4992 compound synthesis.

To a solution of furan-2-carboxylic acid (1.0eq), N1-(2-methyl-6-(pyrrolidin-1-yl)pyrimidin-4-yl)benzene-1,4-diamine (1.2eq), and HBTU (1.1eq) in DMF (0.3mM) was added diisopropylethylamine (1.3eq) and stirred at 25°C for 72hr; this mixture was purified on prep-HPLC and freeze-dried to light yellow solid. Compounds were resuspended in DMSO (final: 50mM).

2.5. Analysis of inhibitory properties of the NAPA compounds

Cells (10⁴/well, 96-well plates (Greiner) plated overnight were pretreated for 1 hour at 37°C with the NAPA compounds or DMSO in triplicate prior to virus infection (MOI that yields ~20-30% infection). Virus infection was evaluated at 24hpi using a rabbit anti-IE1 antibody (Tortorella lab) as described (Stein et al., 2019). The relative percent infection of NAPA-treated cells was determined using DMSO-treated cells as 100% infection. The half-maximal inhibitory concentration (IC₅₀) for each compound was calculated using a nonlinear four-parameter curve fitting algorithm (GraphPad Prism). The immunoblot assay was performed as described(Stein et al., 2019). The mechanistic studies and herpes virus inhibition (Keith et al., 2018; Hartline et al., 2018) protocols are described in Supplemental Information.

3. Results

3.1. High-content screen (HCS) for early-stage CMV replication inhibitors

We utilized the reporter virus AD169^IE2-YFP(Gardner et al., 2015) to screen a library of discreet synthetic small molecules for compounds that reduce expression of the IE2/yellow fluorescent protein (YFP) (IE2-YFP) chimera. The IE1 and IE2 CMV gene products are translated within 3-6hpi and are key markers of the early stages of CMV replication(Cherrington and Mocarski, 1989; Meier and Stinski, 1997). The IE2-YFP chimera exhibits expression kinetics and nuclear localization like that of IE2 (Teng et al., 2012). Expression of the nucleus-localized IE2-YFP can be readily quantified using a confocal fluorescence microscopic plate reader. The AD169^IE2-YFP reporter virus thus provides the basis for the HCS assay for identification of inhibitors to the early stages of CMV’s life cycle.

After optimization of the AD169^IE2-YFP reporter assay (average Z’ factor ~0.5), we screened a library of 112,197 small molecules from commercially-available compound collections and quantified the percentage of nuclei containing YFP (YFP+) as compared to total nuclei (H33342-stained) using a confocal microscopic plate reader. Compounds that reduced YFP+ nuclei by ≥75% but had a minimal effect (≤25% reduction) on the number of total nuclei (H33342 stained) were scored as “primary hits.” The workflow for screening and evaluation of primary hits is summarized in Fig. 1A. We identified 300 primary hits, which were further evaluated using an assay funnel designed to prioritize compounds based on antiviral activity, specificity, and drug-like properties, ultimately uncovering two potent, selective, drug-like candidates for hit-to-lead optimization that were not Pan Assay Interference compounds (PAINS)(Baell and Holloway, 2010; Capuzzi et al., 2017; Dahlin and Walters, 2016). The primary hit-to-lead series selected was an N-arylpyrimidinamine (NAPA) compound referred to as MBXC-4302 (Fig. 1B), in which secondary validation assays demonstrated that this compound exhibited potent anti-CMV activity (IC₅₀=3.2μM) with limited cytotoxicity (CC₅₀ > 100μM).

Figure 1: Screening strategy to identify early stage CMV inhibitors.

(A) The schematic and criteria of the high-content screen to identify inhibitors of CMV infection using AD169^IE2-YFP/fibroblast infection. (B) Structure of MBXC-4302.

To evaluate the potency of MBXC-4302 in blocking CMV infection of physiologically relevant cell types, infectivity assays were performed in human fibroblasts (NHDF) and epithelial (ARPE-19) cells. AD169^R and TB40/E (MOI=0.5)-infected ARPE-19 and NHDF cells treated with MBXC-4302 were analyzed for virus infection at 24hpi (Materials and Methods). MBXC-4302 inhibited AD169^R-infection of NHDF and ARPE-19 cells with IC₅₀ values of 3 μM and 5.8 μM respectively (Fig. 2A and C). The MBXC-4302 IC₅₀ values for TB40/E-infected NHDF and ARPE-19 cells were ~4.5 μM for both cell types (Fig. 2B and D). These data demonstrate that MBXC-4302 broadly inhibits CMV infection of physiologically relevant cell types.

Figure 2. MBXC-4302 inhibits CMV infection.

CMV strains AD169^R (A and C) and TB40/E (B and D) infection (MOI=0.2) of fibroblasts (A-B) and ARPE-19 cells (C-D) pretreated with 0, 1, 5, and 20 μM MBXC-4302 were assessed 24hpi based on GFP fluorescence (A,C) or IE1 (B,D) expression by an anti-IE1 immunostain. Relative % infection was determined using DMSO-treated cells as 100% infection. The data points are the averages of three replicates and the error bars are the standard deviation. Statistical tests were performed using one-way ANOVA with multiple comparisons to DMSO-treated cells as a control and a Dunnett’s post-test; ****, p<0.0001.

3.2. Preliminary structure activity relationships of the NAPA series

Preliminary structure activity relationship (SAR) studies of the NAPA series demonstrated responsiveness in various biological activities to structural changes within the five zones of MBXC-4302 defined in Table 1. Modulations in Zone 1 were found to alter compounds’ potency and selectivity. Switching from an heteroaromatic pyrazole in Zone 1 of MBXC-4302 (entry 7) to an aromatic secondary amine (entries 8-10) led to a 10fold increase in potency, highlighting a potential hydrophobic binding pocket. This is corroborated by the decrease in potency when the aryl ring was replaced by a polar acyclic ethylamine (entry 1), resulting in potency similar to MBXC-4302. Increasing lipophilicity (entries 2-5) also improved potency of inhibition, which was restored to sub-micromolar levels with the piperidine substitution (entry 5). Zone 1 analogs varied widely in CC₅₀ values, with a selectivity index ranging from 7 to >100, consistent with a specific antiviral mechanism independent of any cytotoxic pathway. Similarly, tertiary amines (entries 2-6) were highly selective relative to the phenyl analogs. With regards to analogs with modifications in Zone 2 (entries 8-10), MBXC-4297 had a selectivity index of >50, while the other two isomeric pyrimidine ligands had index values of <20, suggesting the electronics of the pyrimidine moiety plays a role in selectivity. In Zone 4, addition of a sulfone linker (entry 11) was not tolerated, but a urea linker (entry 12) improved potency, suggesting a potential hydrogen-bonding site in this binding pocket. Finally, alterations in Zone 5 from a furan to a thiophene or a phenyl ring (entries 13, 14) resulted in complete loss of activity, suggesting the need for further optimization with regards to this Zone.

Table 1.

Structure Activity Relationships of the NAPA Series

^aHeLa cell lines;

^bMaximum solubility in water determined by nephlometry;

^cMurine liver microsomal stability;

^d% remaining after 60 min at 37 °C in the presence of NADPH.

Analogs were further evaluated for solubility and in vitro ADME properties. The basic amines (entries 1-6) are highly soluble compared to other analogs, and the majority of compounds were stable in the presence of mouse liver microsomes, supporting the analysis of effective analogs in future in vivo studies. While some compounds exhibited CYP3A4 inhibition, modulation of Zone 1 provides a potential solution to overcome or avoid drug-drug interactions. Based on the data gathered from in vitro assays and a cytotoxicity study in NHDFs (Fig. S1A), MBX-4992 (entry 4) was found to exhibit favorable parameters, and was therefore selected for further characterization as a novel CMV inhibitor.

3.3. MBX-4992 broadly inhibits CMV infection

To investigate the broad inhibitory potential of the NAPA analog MBX-4992, NHDF and ARPE-19 cells were treated with MBX-4992 (1, 5, and 10μM) followed by infection with AD169^R and TB40/E strains due to their wide tropism. Cells were then examined for infection at 24hpi, revealing IC₅₀ values ranging from 1.9-4.8 μM (Figs. 3A–D). To confirm that MBX-4992 reduces expression of the CMV IE1 protein during infection, we examined IE1 protein levels in AD169^R-infected NHDFs treated with MBX-4992 (Fig. 3E). A decrease in IE1 expression was observed upon treatment with MBX-4992 (Fig. 3E, lanes 1-4). As expected, treatment with ganciclovir (GCV) did not impact IE1 expression, while convallatoxin (CVT), a compound targeting the Na⁺/K⁺ ATPase(Cohen et al., 2016) inhibited IE1 expression (Fig. 3E, lanes 5-6). Analysis of IE1 expression at 2 days post-infection further validated that MBX-4992 prevents virus infection (Fig. S1B). These findings demonstrate that MBX-4992 inhibits CMV infection.

Figure 3. NAPA compound MBX-4992 significantly limits CMV infection in vitro.

CMV strains AD169^R (A,C) and TB40/E (B,D) infection (MOI=0.2 or 0.4, respectively) of fibroblasts (A-B) and ARPE-19 cells (C-D) pre-treated with 0, 1, 5, and 20 μM MBX-4992 were assessed 24hpi based on GFP fluorescence (A,C) or anti-IE1 immunostain (B,D). Relative % infection was determined using DMSO-treated cells as 100% infection. The data points are the averages of three replicates with error bars indicating standard deviation. Statistical tests were performed using one-way ANOVA with multiple comparisons to DMSO-treated cells as a control and a Dunnett’s post-test; ***, p<0.001; ****, p<0.0001. (E) Total cell lysates from AD169^R-infected (MOI=0.2) NHDF cells pretreated with DMSO, MBX-4992 (1-10 μM), ganciclovir (5 μM) or convallotoxin (0.01 μM) collected at 24hpi were subjected to Western blot analysis using a rabbit anti-IE1 antibody (Lanes 1-6) and an anti-GAPDH antibody (Lanes 7-12) as a loading control. The molecular weight markers and respective proteins are indicated.

3.4. MBX-4992 inhibits several herpes viruses

To determine the spectrum of antiviral activity of MBX-4992, MBX-4992 was evaluated for inhibition of a panel of herpes viruses (Table 2). MBX-4992 exhibited potent antiviral activity against human CMV (IC₅₀=1.7μM) that was comparable to that of GCV (IC₅₀=0.9μM). MBX-4992 was also effective against a GCV-resistant strain of human CMV (IC₅₀=1.8μM). Interestingly, the compound exhibited modest activity against murine CMV and was highly active against HHV-8. However, MBX-4992 did not exhibit activity against any additional herpes viruses or other viral pathogens, such as hepatitis B, dengue, influenza A H1N1, respiratory syncytial, yellow fever, or Zika virus (data not shown). Thus, antiviral activity of MBX-4992 is the most sensitive against cytomegaloviruses.

Table 2.

Spectrum of anti-viral activity of MBX-4992 against herpes viruses.

				MBX-4992 (μM)			GCV (μM)
Virus	Strain	Cell line	Assay	IC50	CC50	SI	IC50*
EBV	Akata	Akata	qPCR	30.9	>100	>3	ND
GpCMV	22122	GP Lung	qPCR	>30	118.8	<4	ND
HCMV	AD169	HFF	CPE	1.7	>150	>88	0.9
HCMV	GDGr K17 (GCVR)	HFF	CPE	1.8	>151	>83	10.4
HHV-6B	Z29	MOLT-3	qPCR	43.7	>100	>2	ND
HHV-8	BCBL-1	BCBL-1	qPCR	10	>100	>10	ND
HSV-1	E-377	HFF	CPE	>30	60.6	<2	ND
HSV-2	G	HFF	CPE	>150	>150	1	ND
MCMV	Smith	MEF	qPCR	4.5	>150	>33	0.2
VZV	Ellen	HFF	CPE	>30	139.2	<5	ND

Abbreviations: EBV, Epstein-Barr virus; GpCMV, Guinea pig cytomegalovirus; HCMV, Human cytomegalovirus; HHV-6, Human herpes virus 6B; HHV-8, Human herpes virus 8; HSV-1, Herpes simplex virus 1; HSV-2, Herpes simplex virus 2; MCMV, Murine cytomegalovirus; VZV, Varicella-Zoster virus; CPE, cytopathic effects; qPCR, quantitative PCR; HFF, Human foreskin fibroblasts; MEF, mouse embryonic fibroblasts.

3.5. MBX-4992 targets an early stage of infection

A time-of-addition assay was performed to elucidate the step of CMV infection targeted by MBX-4992 (Fig. 4A). MBX-4992 (10 μM) was added at times relative to infection (−60, 0, 15, 30, 60, and 90 mpi). MBX-4992 was most effective at preventing infection when introduced between −60 min prior to infection to 30 mpi. The inhibitory activity of MBX-4992 began to diminish at 60 mpi and further at 90 mpi. A similar study in ARPE-19 cells demonstrated similar kinetics to loss of inhibition at 60 and 90 mpi (Fig. S2A). MBX-4992 likely inhibits an early-step of infection based on CMV expression of IE1/2 and localization of pp65 within 2hpi(Bodaghi et al., 1999; Stenberg, 1996)}. Further, virus infection was evaluated using a TB40/E variant expresses the chimeric capsid protein UL32 with EGFP (TB40/E^UL32-EGFP)(Sampaio et al., 2005) (Fig. S2B). We observed a dramatic decrease in TB40/E^UL32-EGFP particles after 2hpi in MBX-4992 treated cells compared to no treatment. Also, the GFP signal from the virus particles appears to localize at the cell membrane. The controls anti-gH neutralizing mAb 4E7(Parsons et al., 2022) and heparin(Kapoor et al., 2020) that block virus entry also displayed limited the virus particles in the cells. These studies are consistent with the paradigm that MBX-4992 is targeting an early-step of virus infection.

Figure 4. NAPA compounds target an early, post-attachment step.

(A) MBX-4992 (10 μM) was added along a time course following infection with AD169 virus (MOI=0.2) from −60 mins prior to infection to 90 mpi on NHDF cells. Infection at 24 hours was quantified with anti-IE1 immunostain and untreated virus-infected cells was normalized to 100%. (B) NHDFs were preincubated with DMSO, MBX-4992, or heparin for 30 mins at 37°C followed by the addition of AD169^R (MOI=0.2) for 30 minutes at 4°C. Cell media was either unchanged (No Wash) or removed (Wash) and replaced with media containing DMSO, MBX-4992, or heparin. (C) AD169^R (MOI=0.2) pre-incubated with DMSO, MBX-4992 (1 μM), or heparin (50μg/ml) for 30 mins at 37°C was added to fibroblasts for 30 min at 4°C. Cells were not washed (No Wash) or media was removed (Wash) and replaced with media containing either DMSO, MBX-4992, or heparin shifting cells to 37°C. Virus infection was quantified at 24hpi by GFP fluorescence with DMSO-treated virus normalized to 100%. The data points are the averages of three replicates with error bars. Statistical tests were performed using ordinary one-way ANOVA with multiple comparisons to DMSO-treated cells as a control and a Dunnett’s post-test: *, p<0.05; ****, p<0.0001.

Does pretreating cells with MBX-4922 limit infection? To address this question, cells were pre-treated with MBX-4992 followed by the addition of AD169^R 30 mins at 4°C allowing the virus to bind. The cells were not washed or washed/replaced with media containing DMSO or MBX-4992 (10μM) followed by analysis of virus infection at 24hpi in which DMSO-treated cells was defined as 100% (Fig. 4B). Heparin treatment (50μg/ml) was used an inhibition control. MBX-4992 and heparin significantly inhibited virus during infection. Yet, pre-incubation of cells with MBX-4992 and its subsequent removal did not significantly virus infection suggesting MBX-4992 requires to be present at time of infection. We next addressed whether MBX-4992 irreversibly targets the virion (Fig. 4C). MBX-4992 (10μM) pre-incubated with CMV at 37°C for 30 min and then added to cells for 30 min at 4°C. The cells were not washed or washed/replaced with media containing DMSO or MBX-4992 (10μM). Virus infection was evaluated at 24hpi in which DMSO treated cells was normalized to 100% (Fig. 4C). Heparin treatment (50μg/ml) was used an inhibition control. Consistent with previous results, MBX-4992 or heparin treatment significantly decreased infection by >75% under all conditions. Interestingly, virus pre-incubated with MBX-4992 and then washed decreased virus infection by ~25% suggesting MBX-4992 has a slight impact on the virus binding. Together, the data connote that MBX-4992 is the most effective when incubated with virus and cells during infection.

4. Discussion

The complex life-cycle of CMV is reflected in the size of its genome (~235 kb), containing at least ~180 ORFs, 45 of which are essential for replication in fibroblasts (Dunn et al., 2003). Accordingly, there are many unexploited essential viral targets for anti-CMV drug discovery. We identified a novel series of N-arylpyrimidinamine (NAPA) compounds that broadly inhibit infection of diverse strains (Fig. 2). Further, NAPA compounds limited virus proliferation in vitro indicating that the compounds are effective inhibitors of virus dissemination (Table 2). Importantly, the NAPA compounds target early steps of the CMV life-cycle, likely impacting a post-binding step (Fig. 5). These inhibitors represent novel CMV inhibitors that target a step of the CMV life-cycle that differs from approved late-stage drug inhibitors of CMV.

The NAPA inhibitor MBX-4992 is specific for CMV replication based on its IC₅₀ values for HCMV (~1.8 μM) and MCMV (4.5 μM). While also demonstrating activity against HHV-8, MBX-4992 showed weak or no inhibition for EBV, GpCMV, HHV-6B, HSV-1, HSV-2, and VZV replication, supporting its relative specificity for cytomegaloviruses. Interestingly, the NAPA compounds have some structural similarities to previously identified benzyl-N’-arylthiourea inhibitors of CMV(Bloom et al., 2003; Bloom et al., 2004; Jones et al., 2004) with regards to Zones 3 and 4 of the NAPA compounds. The arylthiourea compounds target the HCMV envelope protein gB based on the generation of drug-resistant mutants of gB(Jones et al., 2004) and a structure of gB complexed with the arylthiourea analog WAY-174865(Liu et al., 2021). However, several critical gB residues of human CMV (Y155 and D714) that interact with the arylthiourea compound are not conserved in mouse CMV, a virus sensitive to the NAPA compound (Table 2). Despite these differences, the structural similarities of the arylthiourea compounds and the NAPA compounds are quite intriguing, indicating that NAPA compounds may indeed target an early step of infection.

The development of alternative CMV therapies is being pursued by modification of current drugs, drug repurposing, and screening of compound libraries for inhibitors that target different steps of the virus life-cycle(Bogner et al., 2021; Mercorelli et al., 2011). Potential CMV inhibitors include the antimalarial drugs artemisinins(Arav-Boger et al., 2010; Kaptein et al., 2006), which bind to the filament protein vimentin to prevent HCMV replication(Roy et al., 2020). In fact, an artemisinin derivative, artemisone, demonstrated antiviral activity against the early phase of virus replication(Oiknine-Djian et al., 2018). Other compounds that impact early steps of the life-cycle are valspodar(Parsons et al., 2021), dispirotripiperazines, tricyclic molecules with two quaternary positively-charged nitrogen atoms (or “spiro atoms”), modified lectins that block entry of herpes viruses(Egorova et al., 2021; Zoepfl et al., 2021;Lloyd et al., 2022)), and a furan-based molecule(Kapoor et al., 2020). In addition, marine sulfated glycans limited CMV and adenovirus infection by targeting viral attachment and entry(Zoepfl et al., 2021). These studies exemplify the diversity of inhibitors that target different cellular and viral entry factors and show promise as novel CMV therapeutics.

5. Conclusions

Human cytomegalovirus (CMV) has a high seroprevalence rates of 60-90% within the population. Virus proliferation significantly increases the morbidity and mortality of immunocompromised individuals, such as newborns and organ transplant recipients. The development of novel therapeutics that target different steps of the viral life cycle to limit virus propagation and dissemination would provide therapeutic for treating CMV-related diseases. A high-content screening assay using a CMV reporter assay identified an N-arylpyrimidinamine (NAPA) compound with anti-CMV infection activity and limited cytotoxicity. Thus, NAPA compounds represent novel CMV early-stage infection inhibitors that can potentially be developed into effective CMV therapeutics.

Highlights.

• NAPA compounds target an early step of human cytomegalovirus infection

• NAPA compounds inhibit cytomegalovirus infection and proliferation

• NAPA compounds represent a novel class of cytomegalovirus inhibitors