Effect of Cell Culture Conditions on the Anticytomegalovirus Activity of Maribavir

Abstract

The cytomegalovirus UL97 kinase inhibitor maribavir suppressed viral growth more effectively in lung fibroblasts than in skin fibroblasts, and some cellular kinase inhibitors enhanced its antiviral activity. These effects influence the phenotypic assay of drug susceptibility and suggest the possibility of therapeutically useful combinations of maribavir and cellular kinase inhibitors.

Maribavir (MBV) (1263W94) is an l-benzimidazole riboside that is undergoing clinical antiviral trials with human cytomegalovirus (CMV) infection (6). Unlike current CMV antivirals (ganciclovir, foscarnet, and cidofovir), MBV potently inhibits the viral UL97 kinase (phosphotransferase) and retains activity against CMV strains resistant to current drugs (1).

Standards for the phenotypic determination of MBV sensitivity have not been established. Widely differing MBV 50% inhibitory concentrations have been reported by traditional plaque reduction assay (1, 10). Because the observed variability could be related to cell culture factors and host cell kinase activity, we performed phenotypic assays with different fibroblast lines and studied the effects of several cellular kinase inhibitors on the antiviral activity of MBV.

We used a drug-sensitive baseline strain, T2233, derived from strain AD169, which contains a secreted alkaline phosphatase (SEAP) reporter gene (3). In this phenotyping system, SEAP activity in the culture supernatant is used for viral quantitation during multiple cycles of replication (3). For this study, an additional SEAP-expressing CMV strain, T2264, containing the L397R mutation of UL97, which confers high-level MBV resistance (1), was derived from strain T2211 as described previously (3). MBV was obtained from GlaxoSmithKline. Cellular kinase inhibitors roscovitine, rapamycin, and LY294002 were obtained from LC Laboratories (Woburn, MA). The leflunomide metabolite A77-1726 [N-(4-trifluoromethylphenyl)-2-cyano-3-hydroxycrotoamide] was obtained from EMD Biosciences (San Diego, CA).

Human embryonic lung (HEL) fibroblasts and human foreskin fibroblasts (HFF) were derived from anonymous tissue at Stanford University and used at passages 10 to 20 (HEL fibroblasts) or 20 to 30 (HFF). MRHF human foreskin fibroblasts (passage 26; Biowhittaker, Walkersville, MD) and MRC-5 human embryonic lung fibroblasts (passage 19; American Type Culture Collection) were commercially obtained. Cells were cultured at 37°C in 24-well tissue culture plates in minimal essential medium with Earle's salts supplemented with 2 mM l-glutamine and 10% fetal bovine serum and used when confluent 6 days later at a density of 1.8 × 10⁵ to 2.4 × 10⁵ cells per well. After viral inoculation, maintenance medium was the same except for reduction of fetal bovine serum to 3%.

For SEAP growth assays, wells of 24-well culture plates of fibroblasts were inoculated with cell-free virus stock at a multiplicity of infection of 0.01 to 0.03 (3). Some wells contained an added drug(s) at the specified concentration(s). At intervals of 1, 4, 5, 6, 7, and 8 days after incubation at 37°C, 80-μl aliquots of supernatant medium were collected, frozen, and later assayed together for SEAP activity, measured as relative light units, by using a chemiluminescent substrate (3). SEAP yield reduction assays were performed using serial drug dilutions as recently published (3), except that in some experiments a fixed concentration of a second drug was added to all six wells used to perform each assay. The drug concentration that reduced the SEAP activity to 50% of the control value (in the no-drug well) (50% effective concentration [EC₅₀]) was calculated by fitting an exponential curve to the measured SEAP activities.

Figure 1 shows the SEAP growth curves for strain T2233 under increasing concentrations of MBV in HEL cells and HFF. The main difference between HEL cells (Fig. 1A) and HFF (Fig. 1B) is that substantial viral growth continued in HFF under much higher concentrations of MBV, approaching cytotoxicity (32 μM). Figure 2 shows a cross-sectional view of the curves in Fig. 1 at a single postinoculation interval (5.5 days), during the exponential growth phase as new cycles of viral replication take place. In HEL and MRC-5 cells, which are both lung fibroblasts, low concentrations of MBV (less than 0.5 μM) reduce the SEAP yield to well under 50%, resulting in a sharply defined MBV yield reduction EC₅₀ for HEL cells of 0.14 μM (Table 1). In HFF and MRHF cells, which are also foreskin fibroblasts, there is an initial MBV dose-related yield reduction up to about 1 μM of drug, followed by a relatively flat part of the curve where the SEAP yield is reduced by 40 to 60% over a wide range of MBV concentrations. Under these circumstances, EC₅₀ values may fluctuate considerably depending on exact cell culture conditions and the formulas used to calculate the EC₅₀s. Based on 16 experiments, the MBV yield reduction EC₅₀ for HFF is 13.3 μM (Table 1), or almost 100 times the value observed for HEL cells. These observations may explain why published plaque reduction MBV EC₅₀ values for the same AD169 strain range from 0.5 μM (1) to 19 ± 19 μM (10). The lower value was obtained with MRC-5 lung fibroblasts and the higher value in a different laboratory using foreskin fibroblasts. It is not known if the variable levels of activity of MBV in cell culture affect the activity of the drug in vivo at different tissue sites. A preliminary clinical trial suggests good activity of MBV in reducing viral shedding (6), but there is no information yet regarding tissue-invasive CMV disease.

FIG. 1.

SEAP growth curves for CMV strain T2233 under different MBV concentrations in (A) HEL cells and (B) HFF. Data points are the means ± standard errors of the means from at least four assays. RLU, relative light units.

FIG. 2.

SEAP yield reduction curves of CMV strain T2233 at 5.5 days under MBV. The effects of various concentrations of MBV were tested with four different cell types (HFF, MRHF, MRC-5, and HEL cells) and with HFF in the presence of rapamycin (Rapa), roscovitine (Rosco), or the leflunomide metabolite A77-1726 (A77). One hundred percent SEAP activity is defined as the level of activity in the wells without added drug. Data points are means ± standard errors of the means from at least four assays.

TABLE 1.

Antiviral EC₅₀ values under different cell culture conditions

Cell type and CMV strain (UL97 genotype)a	Drug A	Drug B, concn	EC50 for drug A (no. of replicates)b
HEL cells
T2233 (wt)	MBV	None	0.14 ± 0.04 (13)
	MBV	Roscovitine, 1.5 μM	0.06 ± 0.01 (9)
	MBV	Rapamycin, 0.1 nM	0.07 ± 0.02 (6)
	MBV	LY294002, 2 μM	0.11 ± 0.02 (9)
	MBV	A77-1726, 5 μM	0.25 ± 0.14 (9)
	Foscarnet	None	39 ± 10 (7)
	Foscarnet	Roscovitine, 1.5 μM	33 ± 7 (12)
	Foscarnet	Rapamycin, 0.1 nM	38 ± 14 (16)
	Foscarnet	LY294002, 2 μM	37 ± 3 (8)
	Roscovitine	None	1.2 ± 0.3 (6)
	Rapamycin	None	0.05 ± 0.01 (6)
	LY294002	None	1.6 ± 0.3 (6)
	A77-1726	None	7.5 ± 2.4 (6)
T2263 (L397R mutation)	MBV	None	24 ± 9.1 (7)
	MBV	Roscovitine, 1.5 μM	25 ± 11 (10)
	MBV	Rapamycin, 0.1 nM	15 ± 2.7 (9)
	MBV	LY294002, 2 μM	23 ± 5.0 (6)
	MBV	A77-1726, 5 μM	>40 (10)
	Roscovitine	None	0.93 ± 0.5 (6)
	Rapamycin	None	0.04 ± 0.01 (6)
	LY294002	None	2.4 ± 0.3 (8)
	A77-1726	None	13.7 ± 0.9 (6)
HFF
T2233 (wt)	MBV	None	13.3 ± 3.6 (16)
	MBV	Roscovitine, 1.5 μM	0.20 ± 0.05 (14)
	MBV	Roscovitine, 4 μM	0.17 ± 0.05 (6)
	MBV	Rapamycin, 0.1 nM	0.22 ± 0.0.06 (7)
	MBV	LY294002, 6 μM	0.22 ± 0.03 (7)
	MBV	A77-1726, 5 μM	>32 (6)
	Foscarnet	None	45 ± 8 (15)
	Foscarnet	Roscovitine, 1.5 μM	46 ± 12 (11)
	Foscarnet	Rapamycin, 0.1 nM	47 ± 11 (13)
	Foscarnet	LY294002, 2 μM	48 ± 12 (8)
	Roscovitine	None	3.9 ± 0.6 (7)
	Rapamycin	None	0.14 ± 0.06 (6)
	LY294002	None	5.3 ± 0.9 (6)
	A77-1726	None	5.0 ± 1.9 (9)

^awt, wild type.

^bEC₅₀s are expressed as μM (except for rapamycin EC₅₀s, which are expressed as nM) and are shown as means ± standard deviations. EC₅₀s for drug A were determined in the presence of a fixed concentration of drug B.

To test the possibility that cellular kinases may compensate for inhibition of the viral UL97 kinase by MBV and permit a moderate level of viral growth especially in HFF, MBV EC₅₀ values were determined in the presence of several cellular kinase inhibitors known to inhibit CMV in cell culture, including roscovitine (2, 8), rapamycin (5), LY294002 (4), and a leflunomide metabolite, A77-1726 (9). The EC₅₀ values for these substances by themselves for control strain T2233 are shown in Table 1, for both HEL cells and HFF. Roscovitine, rapamycin, and LY294002, when added to cultures at concentrations close to their EC₅₀, reduced the MBV EC₅₀ by over 50-fold in HFF (Table 1). This caused a shift in the HFF MBV dose-response curve closer to that observed with HEL cells (Fig. 2). The leflunomide metabolite A77-1726 had the opposite effect of increasing the MBV EC₅₀ (Table 1 and Fig. 2). As a control, the DNA polymerase inhibitor foscarnet was tested with and without concomitant cellular kinase inhibitors, and the EC₅₀ was unchanged (Table 1), suggesting that the effects described above in combination with MBV are not due to nonspecific cytotoxicity of the cellular kinase inhibitors. The high-level MBV resistance of CMV strain T2264, containing the L397R mutation of UL97 (1), persisted in the presence of cellular kinase inhibitors, even though the latter substances by themselves were about as active against strain T2264 as the control strain T2233.

The potentiation of MBV antiviral activity by some cellular kinase inhibitors may provide a clinically meaningful therapeutic strategy. In particular, the drug rapamycin (sirolimus) is in common use in the transplant setting, where CMV disease is prevalent, at therapeutic drug levels higher than those used here in combination with MBV. Interestingly, preliminary observations suggest that use of rapamycin itself is associated with less CMV disease in transplant recipients than some other standard immunosuppressive drugs (7). It is possible that a combination of MBV and rapamycin (or similar drugs) could provide an enhanced therapeutic option for CMV disease in this population.