The alphaherpesvirus gE/gI glycoprotein complex and proteases jointly orchestrate invasion across the host's upper respiratory epithelial barrier

Abstract

Alphaherpesviruses, including herpes simplex virus type 1 (HSV-1), pseudorabies virus (PRV), and bovine herpesvirus type 1 (BoHV-1), are significant pathogens affecting humans and animals. These viruses penetrate the upper respiratory tract mucosa, yet the mechanisms facilitating this invasion are not fully understood. This study investigates the role of the gE/gI glycoprotein complex and proteases in mucosal invasion by these viruses. Using species-specific respiratory mucosal explants, we observed that the removal of extracellular calcium disrupts epithelial junction integrity, enhancing viral infection across all viruses and suggesting a common mechanism of targeting a basolaterally located receptor. PRV exhibited significantly faster replication and deeper invasion compared to HSV-1 and BoHV-1. The gE glycoprotein was consistently polarized at the basement membrane across all viruses, indicating a critical role in the process of viral entry and subsequent spread through the epithelium. In this context, "infection" refers to the virus's attachment to its cell-surface receptor, entry into the cell, and completion of the viral life cycle, culminating in the production of progeny virions. Notably, in gE/gI null mutants of PRV and HSV-1, while the infection was not abortive and the viral life cycle was completed, the infection was delayed, and the invasion into the deeper layers of the epithelium and underlying mucosa was significantly reduced. In BoHV-1 mutants, this effect was even more pronounced, with infection restricted to the apical cells, failing to progress to the basal cells. In addition, PRV and HSV-1 invasion involved serine protease activity, unlike BoHV-1, which correlates with its slower invasion pace. Notably, the protease facilitating PRV invasion was identified as a urokinase plasminogen activator (uPA), while the specific protease for HSV-1 remains unidentified. These findings highlight the critical roles of the gE/gI complex and proteases in alphaherpesvirus pathogenesis, offering potential targets for therapeutic intervention.

Importance: Herpes simplex virus type 1 (HSV-1) infections are a worldwide issue. More than three billion people are infected with HSV-1 globally. Although most infections with HSV-1 occur subclinically, severe symptoms and complications are numerous and can be life-threatening. Complications include encephalitis and blindness. Recently, HSV-1 infections have been associated with the development of Alzheimer's Disease. To date, no effective vaccines against HSV-1 are on the market. Pseudorabies virus (PRV) and bovine herpesvirus type 1 (BoHV-1) are two alphaherpesviruses of major veterinary importance. Although efforts have been made to eradicate these viruses from livestock animals, clinical problems still occur, resulting in great economic losses for farmers. It is evident that new insights into the pathogenesis of alphaherpesviruses are needed, to develop effective treatments and novel preventive therapies.

Keywords: alphaherpesvirus; gE/gI complex; pathogenesis, mucosa invasion; proteases; upper respiratory tract; urokinase plasminogen activator.

Fig 1

EGTA treatment of the respiratory mucosa destroys epithelial junctions and thereby enhances alphaherpesvirus infection in swine, bovine, and human explants. (A) Representative HE stainings showing the respiratory mucosa under normal conditions (left), and EGTA treatment (right) for the three different species. Arrows point out intercellular spaces. (B) Bar plot indicating central tendencies (mean ± SD) for the percentage of intercellular spaces in the epithelium. (C) Representative transmission electron microscopy pictures of non-EGTA (top) and EGTA-treated (bottom) porcine respiratory epithelium. White arrows point to cellular junctions. Scale bars represent 2 µm. (D) Representative immunofluorescent images of WT PRV, WT BoHV-1, and WT HSV-1 at 48 hpi, after serum-free medium treatment (negative control) (left column), or EGTA treatment (right column). The basement membrane (collagen type VII) is depicted by the red fluorescent color. (E) Bar plot depicting the percentage of WT PRV, WT BoHV-1, and WT HSV-1 infection in the epithelium at 48 hpi, after non-EGTA or EGTA treatment. Significant differences between the means of three independent experiments are depicted on the bar plots by asterisks (*P-value <0.05, **P-value <0.01, ***P-value <0.001, ****P-value <0.000.

Fig 2

Infection and invasion of WT PRV, BoHV-1, and HSV-1 in porcine, bovine, and human respiratory mucosal explants. (A) Representative immunofluorescent pictures of uninfected URT (“mock”) and infected URT at 12, 24, 48, and 72 hpi. Note: The images used in Fig. 2A (48 hpi) are the same as those used in Fig. 1D (right column), as they represent identical experimental conditions (EGTA-treated explants inoculated with WT virus at 48 hpi) across both figures. This duplication reflects consistent observations from replicate experiments. The glycoprotein B (gB) of each alphaherpesvirus was stained in green (AlexaFluor 488) and the BM (collagen type VII) was stained in red (AlexaFluor 594). Nuclei were counterstained in blue (Hoechst 33342) (scale bar represents 100 µm). (B) Central tendencies for the amount of infection in the respiratory epithelium are indicated by the upper three bar plots [left graph: percentage of infection in the epithelium; middle graph: number of infectious centers (i.e., viral plaques, SIC and groups of 2–3 infected cells)/2 mm² epithelium; right graph: average plaque diameter (µm)]. (C) Viral invasion parameters are presented in the three lower bar plots [left graph: percentage of plaques infecting basal cells; middle graph: percentage of plaques crossing the BM and invading the lamina propria; right graph: maximum invasion depth of the plaques underneath the BM (µm)]. Values in all graphs are presented as means ± standard deviations (SD). Significant differences between the means of three independent experiments are depicted on the bar plots by asterisks (*P-value <0.05, **P-value <0.01, ***P-value <0.001, ****P-value <0.0001).

Fig 3

Polarized expression of glycoprotein B, C, D, and E of WT PRV, HSV-1, and BoHV-1 in the URT epithelium. (A, B, and C) Representative immunofluorescent pictures of WT PRV, WT HSV-1, and WT BoHV-1 infection and invasion at 48 hpi. The glycoprotein E (gE) of each alphaherpesvirus was stained in green (AlexaFluor 488) and the glycoproteins B, C, and D (gB, gC, and gD) were stained in red (AlexaFluor 594). Nuclei were counterstained in blue (Hoechst 33342). The BM is depicted by a white dashed line. The scale bar represents 100 µm. An ROI around the BM was enlarged to show the polarization. White arrows point out glycoprotein expression at the BM. (D) Central tendencies for the percentage of glycoprotein expression at the level of the BM are indicated by the bar plot. Values in the graph are presented as means ± standard deviations (SD). Significant differences between the means of three independent experiments are depicted on the bar plots by asterisks (*P-value <0.05).

Fig 4

Inhibition of serine proteases in WT PRV and WT HSV-1 reduces BM invasion, contrary to WT BoHV-1. (A) Representative immunofluorescent pictures at 36 hpi. The glycoprotein B (gB) of each alphaherpesvirus was stained in green (AlexaFluor 488) and the BM (collagen type VII) was stained in red (AlexaFluor 594). Nuclei were counterstained in blue (Hoechst 33342) (pictures taken with a 20x objective, scale bar represents 100 µm) (B) The bar plot depicts the maximum invasion depth underneath the BM (µm) at 36 hpi, after protease inhibitor or control treatment. Each black dot represents one replicate value (=1 animal experiment). The colored bars represent the mean values. Significant differences between the means of three independent experiments are depicted on the bar plots by asterisks (*P-value <0.05, **P-value <0.01, ***P-value <0.001).

Fig 5

Comparative replication kinetics of PRV, BoHV-1, and HSV-1 gE/gI deletion mutants (gE/gI null). (A) Representative immunofluorescent pictures at 12, 24, 48, and 72 hpi. The glycoprotein C (gC) of each alphaherpesvirus was stained in green (AlexaFluor 488) and the BM (collagen type VII) was stained in red (AlexaFluor 594). Nuclei were counterstained in blue (Hoechst 33342) (pictures taken with a 20x objective, scale bar represents 100 µm). (B) Central tendencies for the amount of infection in the respiratory epithelium are indicated by the upper three bar plots [left graph: percentage of infection in the epithelium; middle graph: number of infectious centers (i.e. viral plaques, SIC and groups of 2–3 infected cells)/2 mm² epithelium; right graph: average plaque diameter (µm)]. (C) Viral invasion parameters are presented in the three lower bar plots [left graph: percentage of plaques infecting basal cells; middle graph: percentage of plaques crossing the BM and invading the lamina propria; right graph: maximum invasion depth of the plaques underneath the BM (µm)]. Values in all graphs are presented as means ± standard deviations (SD). Significant differences between the means of three independent experiments are depicted on the bar plots by asterisks (*P-value <0.05, **P-value <0.01, ***P-value <0.001, ****P-value <0.0001).

Fig 6

Proliferating basal cells are resistant to BoHV-1 gE/gI null infection. (A) Representative immunofluorescent pictures. Top row: the glycoprotein C (gC) of BoHV-1 gE/gI null is stained in green (AlexaFluor 488), the basal cells (CK15) are stained in red (AlexaFluor 594). The dashed line represents the BM. Middle row: the glycoprotein C (gC) is stained in red and the integrin α6 is stained in green. Bottom row: The basal cells are shown in red and the integrin α6 is shown in green. Nuclei were counterstained with Hoechst. The scale bar of the top and middle row represents 100 µm. The scale bar of the bottom row represents 50 µm. (B) Top graph: the percentage of basal cells in the epithelium. Central tendencies for the percentage of CK15-positive cells are depicted in different ROIs. (ROI₁ = area underneath a viral plaque in BoHV-1 gE/gI null infected epithelium; ROI_1’ = an area in the infected epithelium, adjacent to a viral plaque; ROI_1’’ = an area of similar size in mock-inoculated epithelium). Bottom graph: The expression of integrin α6 in the epithelium. Central tendencies for the percentage of integrin α6-positive cells are depicted in the same ROI’s as mentioned above. Values in all graphs are presented as means + standard deviations (SD). Significant differences between the means of three independent experiments are depicted on the bar plots by asterisks (*P-value <0.05, **P-value <0.01, ***P-value <0.001, ****P-value <0.0001).

Fig 7

Comparative hypothetical model of alphaherpesvirus infection and invasion of the URT between species. (A) Situation at 72 hpi in compromised respiratory epithelium (caused by EGTA treatment or the action of environmental hazards). Top: PRV, middle: HSV-1, bottom: BoHV-1. Left of the dashed line: WT virus. Right of the dashed line: gE/gI null mutant virus. WT PRV infects faster and invades deeper the lamina propria than in HSV-1 and BoHV-1. The percentage of infection and plaque diameters are higher in PRV than HSV-1 and BoHV-1. PRV invasion goes up to 218 µm in 100% of the plaques. On the contrary, HSV-1 and BoHV-1 show less plaques with smaller plaque diameters. The percentage of plaques invading the BM and maximum invasion depths are lower as well (95 µm in 75% of plaques for HSV-1 and 76 µm in 45% of plaques for BoHV-1). Viral glycoproteins E, B, C, and D are more expressed at the BM by WT HSV-1 and WT BoHV-1 than by WT PRV. WT PRV invasion is facilitated by the urokinase plasminogen activator (uPA). HSV-1 invasion is also reliant upon a yet unidentified serine protease. No proteases are involved in BoHV-1 invasion. The glycoprotein E is crucial for BM invasion in all three viruses. Infection with gE/gI null mutants is both delayed and reduced in the three distinct epithelia compared to WT strains. Invasion of the BM is similarly reduced and delayed upon infection with gE/gI null viruses. Interestingly, BoHV-1 gE/gI null is additionally hampered in its invasion, as basal cells remain uninfected up to 72 hpi, suggesting that the gE/gI complex is necessary for basal cell infection in BoHV-1. Moreover, basal cells reorganize integrin α6 expression to detach from the BM. Next, they migrate to and proliferate underneath viral plaques to form a wedge-like shape. They push the infected cells away from the BM (sequestration process). (B) Hypothetical model explaining the different observed phenotypes. In all three WT viruses, the tail of the gE/gI complex binds to the US9 protein, which, in turn, binds to a kinesin motor protein. This kinesin specifically targets the basal side of the epithelial cell by moving along microtubuli. The extracellular domain of the gE/gI binds target proteins and traffics them to the basal side of the cell. In WT PRV, the urokinase plasminogen activator (uPA) is bound. In HSV-1, a yet unidentified serine protease is bound, as well as glycoproteins B, C, and D. In BoHV-1, no proteases are bound by the gE/gI complex. Only glycoproteins B, C, and D are bound and are transported to the basal side of the epithelial cells. Here, proteases facilitate invasion (WT PRV, WT HSV-1).