Alpha/Beta interferon and gamma interferon synergize to inhibit the replication of herpes simplex virus type 1

Abstract

In vivo evidence suggests that T-cell-derived gamma interferon (IFN-gamma) can directly inhibit the replication of herpes simplex virus type 1 (HSV-1). However, IFN-gamma is a weak inhibitor of HSV-1 replication in vitro. We have found that IFN-gamma synergizes with the innate IFNs (IFN-alpha and -beta) to potently inhibit HSV-1 replication in vitro and in vivo. Treatment of Vero cells with either IFN-beta or IFN-gamma inhibits HSV-1 replication by <20-fold, whereas treatment with both IFN-beta and IFN-gamma inhibits HSV-1 replication by approximately 1,000-fold. Treatment with IFN-beta and IFN-gamma does not prevent HSV-1 entry into Vero cells, and the inhibitory effect can be overcome by increasing the multiplicity of HSV-1 infection. The capacity of IFN-beta and IFN-gamma to synergistically inhibit HSV-1 replication is not virus strain specific and has been observed in three different cell types. For two of the three virus strains tested, IFN-beta and IFN-gamma inhibit HSV-1 replication with a potency that approaches that achieved by a high dose of acyclovir. Pretreatment of mouse eyes with IFN-beta and IFN-gamma reduces HSV-1 replication to nearly undetectable levels, prevents the development of disease, and reduces the latent HSV-1 genome load per trigeminal ganglion by approximately 200-fold. Thus, simultaneous activation of IFN-alpha/beta receptors and IFN-gamma receptors appears to render cells highly resistant to the replication of HSV-1. Because IFN-alpha or IFN-beta is produced by most cells as an innate response to virus infection, the results imply that IFN-gamma secreted by T cells may provide a critical second signal that potently inhibits HSV-1 replication in vivo.

FIG. 1.

IFN-β and IFN-γ synergize to inhibit HSV-1 plaque formation on Vero cells. (A) Efficiency of HSV-1 plaque formation on Vero cells treated with 100 U/ml of IFN-β and variable doses of IFN-γ (n = 3 per group). The dashed line indicates the number of plaques that formed on vehicle-treated cells. Significant reductions in plaque counts relative to cells treated with 100 U/ml of IFN-β alone are denoted by a single asterisk (P < 0.05, one-way ANOVA and Duncan's multiple-range test). (B) Efficiency of HSV-1 plaque formation on Vero cells treated with variable doses of IFN-β (n = 3 per group) and 100 U/ml of IFN-γ. The dashed line indicates the number of plaques that formed on vehicle-treated cells. Significant reductions in plaque counts relative to cells treated with 100 U/ml of IFN-γ alone are denoted by a single asterisk (P < 0.05, one-way ANOVA and Duncan's multiple-range test).

FIG. 2.

IFN-β and IFN-γ synergize to inhibit HSV-1 replication in Vero cells. Virus yield in Vero cells treated with vehicle, IFN-β (100, 316, or 1,000 U/ml), IFN-γ (100, 316, or 1,000 U/ml), IFN-β and IFN-γ (100 U/ml of each), or ACV (300 μM). Virus titers were determined 24 h after infection with 0.1 PFU of HSV-1 strain KOS/cell (n = 4 per group). Treatments that significantly reduced virus titer relative to vehicle-treated cultures are indicated with an asterisk (P < 0.05, one-way ANOVA and Tukey's post hoc t test). The dashed line represents the lower limit of detection of the plaque assay used to measure viral titers.

FIG. 3.

IFN-β and IFN-γ do not inhibit HSV-1 adsorption to Vero cells. (A) Ethidium bromide-stained RR PCR products amplified from HSV-1-infected Vero cells treated with either vehicle (upper panel) or IFN-β and IFN-γ (lower panel, 100 U/ml of each). From left to right, PCR tubes received no template (H₂O), 100 ng of uninfected (UI) Vero cell DNA, or 100 ng of HSV-1-infected Vero cell DNA harvested from cells inoculated with 0.1 to 20 PFU/cell. (B) RR PCR product yield plotted as a function of viral MOI in Vero cells treated with vehicle or both IFN-β and IFN-γ. The correlation coefficient between PCR product yield and viral MOI was r = 0.81 in vehicle-treated cells and was r = 0.84 in cells treated with IFN-β and IFN-γ.

FIG. 4.

Effect of IFNs and/or ACV on the entry and spread of KOS-GFP in Vero cells. (A) Representative photomicrographs of Vero cells taken 36 h after infection with KOS-GFP (MOI = 0.03), as seen when illuminated with the 360- to 400-nm spectrum of light that excites GFP fluorescence. Magnification, ×10. Vero cells were treated with vehicle (VEH), IFN-β (100 U/ml), IFN-γ (100 U/ml), or IFN-β and IFN-γ (100 U/ml of each) and were treated with either no ACV (−ACV) or 300 μM ACV (+ACV) after infection with KOS-GFP. (B and C) Flow cytometric analysis of GFP fluorescence in Vero cells 1, 12, 24, and 36 h after infection with KOS-GFP (MOI = 0.03). IFN treatments were the same as described above, and cells were secondarily treated with either (B) no ACV (−ACV) or (C) 300 μM ACV (+ACV). The mean percentage ± the standard error (SE) of GFP-positive cells in each treatment was determined in three replicate cultures per time point. The results are presented as the mean percentage ± the SE of GFP-positive cells after subtracting out the background frequency of fluorescence observed in uninfected cultures of Vero cells (i.e., 48 ± 5 per 24,000 cells evaluated).

FIG. 5.

Inhibition by IFN-β and IFN-γ is overcome by increasing the MOI. Virus yield was plotted as a function of MOI. Vero cells were treated with vehicle, IFN-β (100 U/ml), IFN-γ (100 U/ml), or IFN-β plus IFN-γ (100 U/ml of each). Virus titers were determined 24 h after infection with 0.1 to 32 PFU/cell of HSV-1 strain KOS (n = 4 per group). Regression analysis demonstrated that the titer of virus recovered from cells treated with IFN-β and IFN-γ was linearly dependent on MOI (r² = 0.99, P = 10⁻⁶).

FIG. 6.

Pretreatment with IFN-β and IFN-γ inhibits HSV-1 replication in vivo. (A) Virus titers recovered from mouse eyes treated with vehicle, IFN-β, IFN-γ, or both IFN-β and IFN-γ (n = 8 mice per group). Mouse eyes were scarified and treated 0, 4, 8, and 12 h later with vehicle, 200 U of IFN-β, 200 U of IFN-γ, or IFN-β and IFN-γ (200 U of each). After the final treatment, mouse eyes were inoculated with 2 × 10⁵ PFU of HSV-1 strain KOS. The titer of virus recovered from mouse eyes was measured by plaque assay, and the dashed line indicates the lower limit of detection of this assay. Two-way ANOVA demonstrated that each IFN treatment significantly reduced the course of ocular HSV-1 shedding in mice relative to vehicle-treated mice (P < 0.001). (B) Photograph taken 8 days after ocular HSV-1 infection of mice treated with vehicle or both IFN-β and IFN-γ.

FIG. 7.

Pretreatment with IFN-β and IFN-γ reduces the establishment of latent HSV-1 in the TG. (A) RR and competitor PCR products amplified from viral DNA standards. These DNA standards consisted of (i) twofold dilutions of HSV-1 genomes, (ii) a constant amount of competitor (∼1,400 copies), and (iii) 100 ng of uninfected TG DNA. Duplicate blots of PCR products were hybridized to RR-specific or competitor-specific probes, and the ratio of RR to competitor PCR product was used to define the “normalized RR PCR product yield.” The relationship between the number of viral genomes per TG (x, input) and the normalized RR PCR product yield (y, output) was sigmoidal and was described by the polynomial equation: x = 0.328y³ + 0.551y² + 1.4851y + 3.4423 (r² = 0.99). The linear range of the standard curve was between 6.8 × 10³ and 1.7 × 10⁶ viral genomes/TG. Only ∼1/230th of the DNA from each TG was included in the PCR (i.e., 100 ng); thus, the lower limit of accurate quantitation was 30 HSV-1 genomes per PCR. (B) Dot blot of HSV-1 RR PCR products amplified from TG of 3 uninfected (UI) mice and 31 HSV-1 latently infected mice (sacrificed 40 days after inoculation) that were treated prior to HSV-1 infection with vehicle (n = 7), IFN-β (n = 8), IFN-γ (n = 8), or both IFN-β and IFN-γ (n = 8). (C) Effect of IFN-β, IFN-γ, or IFN-β plus IFN-γ on HSV-1 genome load in latently infected TG. The dashed line indicates the lower limit of the linear range of the assay. Significant differences in viral genome load relative to vehicle-treated mice are indicated by an asterisk (P < 0.05, one-way ANOVA and Tukey's post hoc t test).