Ebolaviruses Associated with Differential Pathogenicity Induce Distinct Host Responses in Human Macrophages

Abstract

Ebola virus (EBOV) and Reston virus (RESTV) are members of the Ebolavirus genus which greatly differ in their pathogenicity. While EBOV causes a severe disease in humans characterized by a dysregulated inflammatory response and elevated cytokine and chemokine production, there are no reported disease-associated human cases of RESTV infection, suggesting that RESTV is nonpathogenic for humans. The underlying mechanisms determining the pathogenicity of different ebolavirus species are not yet known. In this study, we dissected the host response to EBOV and RESTV infection in primary human monocyte-derived macrophages (MDMs). As expected, EBOV infection led to a profound proinflammatory response, including strong induction of type I and type III interferons (IFNs). In contrast, RESTV-infected macrophages remained surprisingly silent. Early activation of IFN regulatory factor 3 (IRF3) and NF-κB was observed in EBOV-infected, but not in RESTV-infected, MDMs. In concordance with previous results, MDMs treated with inactivated EBOV and Ebola virus-like particles (VLPs) induced NF-κB activation mediated by Toll-like receptor 4 (TLR4) in a glycoprotein (GP)-dependent manner. This was not the case in cells exposed to live RESTV, inactivated RESTV, or VLPs containing RESTV GP, indicating that RESTV GP does not trigger TLR4 signaling. Our results suggest that the lack of immune activation in RESTV-infected MDMs contributes to lower pathogenicity by preventing the cytokine storm observed in EBOV infection. We further demonstrate that inhibition of TLR4 signaling abolishes EBOV GP-mediated NF-κB activation. This finding indicates that limiting the excessive TLR4-mediated proinflammatory response in EBOV infection should be considered as a potential supportive treatment option for EBOV disease.IMPORTANCE Emerging infectious diseases are a major public health concern, as exemplified by the recent devastating Ebola virus (EBOV) outbreak. Different ebolavirus species are associated with widely varying pathogenicity in humans, ranging from asymptomatic infections for Reston virus (RESTV) to severe disease with fatal outcomes for EBOV. In this comparative study of EBOV- and RESTV-infected human macrophages, we identified key differences in host cell responses. Consistent with previous data, EBOV infection is associated with a proinflammatory signature triggered by the surface glycoprotein (GP), which can be inhibited by blocking TLR4 signaling. In contrast, infection with RESTV failed to stimulate a strong host response in infected macrophages due to the inability of RESTV GP to stimulate TLR4. We propose that disparate proinflammatory host signatures contribute to the differences in pathogenicity reported for ebolavirus species and suggest that proinflammatory pathways represent an intriguing target for the development of novel therapeutics.

Keywords: Ebola virus; Reston virus; Toll-like receptor 4; Toll-like receptors; chemokines; cytokines; filovirus; host response; interferons; macrophages.

FIG 1

MDMs are infected with EBOV and RESTV at comparable rates. (A) MDMs grown in chamber slides were infected with EBOV or RESTV or left uninfected (Mock) and examined by immunofluorescence analysis at 1 dpi using antibodies directed against EBOV NP or RESTV NP (red). Cell nuclei were stained with DAPI (blue). Infection was performed in five separate donors in independent experiment; representative images are shown. (B) Infection rate was determined by counting infected MDMs from five different donors as described for panel A. (C) RNA-Seq analysis was performed with samples generated from three different donors in independent experiments. Reads mapping to the VP35 sequence in the EBOV or RESTV genomes were counted using HTseq-count. Horizontal bars represent mean values; each symbol represents an individual donor. EBOV and RESTV VP35 expression levels were compared using two-tailed t test (GraphPad Prism 5 software), and no statistically significant differences were determined (P > 0.05 for all samples). (D) qRT-PCR analysis of EBOV and RESTV VP35 mRNA. Experiments were performed in duplicate using MDMs obtained from the same 3 individual donors as used for RNA-Seq in panel C. EBOV and RESTV VP35 expression levels were compared using two-tailed t test. *, P ≤ 0.05.

FIG 2

EBOV infection and LPS stimulation induce similar transcriptional changes in MDMs that are absent in RESTV infection. (A) Number of highly DE genes following infection or LPS stimulation relative to time-matched mock-treated samples of MDMs generated from three separate donors in independent experiments. Differential gene expression cutoff was set to |log2FC| >1, and a P value of ≤0.05 was calculated using a generalized linear model with subsequent Benjamini-Hochberg correction. (B) Distribution of DE genes following infection with EBOV or RESTV or LPS treatment. A Venn diagram shows the overlap of DE genes at 6 hpi and 1 and 2 dpi. (C) Hierarchical clustering of the union of 3,574 DE genes based on Euclidean distances reveals distinct modules of gene expression in infected and LPS-stimulated MDMs. The heat map represents the average expression intensities for each condition and time point relative to the average of time-matched mock-infected samples. Each module is identified by a unique color, represented on the left of the heat map. Bar graphs represent canonical pathways associated with genes identified in each of the modules.

FIG 3

EBOV and LPS, but not RESTV, activate similar immune pathways in MDMs. Ingenuity Pathway Analysis (IPA) was used to assess enrichment of canonical pathways in the data sets obtained from each infectious/stimulation condition across time as described in the legend to Fig. 2. The P values are represented by the diameters of circles with more significant pathway enrichment as determined by the Fisher exact test correlating with bigger diameters. “Ratio” represents the number of genes found to be DE relative to the total number of genes that represent the pathway, with darker red color correlating with a greater number of pathway-associated DE genes. Pathway enrichment is reported as −log₁₀ P value, where values greater than 1.3 are considered significant. Gray X's represent a lack of enrichment for a canonical pathway.

FIG 4

Infection with EBOV, but not RESTV, induces a type I and type III IFN response in MDMs. (A) MDMs were either infected with EBOV or RESTV, stimulated with LPS, or left untreated, and cellular RNA was isolated at the indicated time points. The heat map represents the average expression intensities for type I and type III IFN genes and selected ISGs in MDMs from three separate donors in independent experiments. Each column represents an experimental condition and time point, and each row represents an individual gene. Upregulated genes are represented in red, and downregulated genes are represented in blue. (B) Luminex analysis of supernatants from MDMs infected with EBOV or RESTV. As controls, cells were stimulated with LPS or left untreated (Mock). Time points analyzed by RNA-Seq (6 h, 1 day [1d], or 2d as for panel A) were examined using samples from four separate donors, including the corresponding supernatants from the three donors used to generate the RNA-Seq data for all conditions. Extended analysis of late time points postinfection (3d and 4d) were performed with samples from two donors (in gray). Each donor is represented by a different symbol. Horizontal bars represent mean values. Statistically significant differences from time-matched Mock samples as determined by one-way ANOVA are indicated by asterisks (***, P ≤ 0.0001; **, P ≤ 0.001; *, P ≤ 0.05). 3d and 4d time points were not included in the statistical analysis.

FIG 5

RESTV infection does not induce nuclear IRF3 translocation. (A) MDMs grown in chamber slides were infected with EBOV or RESTV, fixed at the indicated times postinfection, and examined by immunofluorescence analysis using antibodies directed against EBOV NP or RESTV NP (red) and IRF3 (green). Cell nuclei were stained with DAPI (blue). The experiment was performed twice with cells obtained from two separate donors; representative images are shown. A representative image of mock-infected cells for one time point and one antibody combination is shown. (B) The percentage of cells (+SEM) showing nuclear IRF3 localization was determined by counting at least 100 infected MDMs from two different donors used in two independent experiments. Gray bar, mock infected; black bars, EBOV infected; white bars, RESTV infected. Statistically significant differences compared to Mock 30 min as determined by one-way ANOVA are indicated by asterisks (***, P ≤ 0.0001).

FIG 6

Differences in NF-κB activation in EBOV- and RESTV-infected MDMs. (A) MDMs were either infected with EBOV or RESTV, stimulated with LPS, or left untreated, and cellular mRNA was isolated at the indicated time points and used for RNA-Seq analysis. The heat map represents the average expression intensities for 137 genes either involved in NF-κB signal transduction or being NF-κB target genes, as identified by the GeneCards database, relative to the time-matched mock-treated samples of MDMs generated from three separate donors in independent experiments. Each column represents an experimental condition and time point, and each row represents an individual gene. Upregulated genes are represented in red and downregulated genes are represented in blue. (B) Average gene expression intensity for proinflammatory cytokine genes, relative to time-matched mock-treated samples. (C) Luminex analysis of supernatants from MDMs infected with EBOV or RESTV. As controls, cells were stimulated with LPS or left untreated (Mock). Time points analyzed by RNA-Seq (6 h, 1d, and 2d as in panel A) were examined using samples from four separate donors, including the corresponding supernatants from the three donors used to generate the RNA-Seq data for all conditions. Later time points (3 and 4 dpi) were examined using samples from 2 individual donors (shown in gray). Each donor is represented by a different symbol. Horizontal bars represent mean values. Statistically significant differences to time-matched Mock samples were determined by one-way ANOVA (***, P ≤ 0.0001; **, P ≤ 0.001; *, P ≤ 0.05). 3d and 4d time points were not included in the statistical analysis.

FIG 7

EBOV, but not RESTV, infection induces p65 nuclear translocation in MDMs. (A) MDMs grown in chamber slides were infected with EBOV or RESTV, treated with LPS, or left noninfected (Mock). At the indicated time points, cells were examined by immunofluorescence analysis using antibodies targeting EBOV NP or RESTV NP (red) and p65 (green). Cell nuclei were stained with DAPI (blue). A representative image of mock-infected cells for one time point and one antibody combination is shown. The experiment was performed four times with cells from four separate donors; representative images are shown. (B) The percentage of cells (+SEM) showing nuclear p65 localization was determined by counting at least 100 infected MDMs from four different donors used in four independent experiments. Light gray bars, mock infected; dark gray bars, LPS treated; black bars, EBOV infected; white bars, RESTV infected. Statistically significant differences from Mock 30 min as determined by one-way ANOVA are indicated (***, P ≤ 0.0001).

FIG 8

EBOV activates canonical NF-κB signaling in MDMs. (A) MDMs were infected with EBOV, stimulated with LPS or TNF-α, or left untreated (Mock), and whole-cell lysates were prepared at the indicated time points. Western blot analysis was performed using antibodies directed against β-actin and IκBα. The experiment was performed in triplicate using MDMs from three different donors. A representative result is shown. (B) MDMs were infected with EBOV, stimulated with LPS, or left untreated (Mock). Nuclear lysates were prepared at the indicated time points and analyzed using an NF-κB subunit DNA binding assay. Raji cell extract was used as a positive control for DNA binding. Light gray bars, mock infected; dark gray bars, LPS treated; black bars, EBOV infected; checkered bars, Raji control. The experiment was performed in triplicate using MDMs from three different donors. Bars represent mean values (+SEM). Statistically significant differences to corresponding time-matched Mock samples as determined by one-way ANOVA are indicated (*, P ≤ 0.05).

FIG 9

Neither EBOV nor RESTV interferes with LPS-mediated NF-κB activation. (A) LPS treatment induces significant gene expression in RESTV-infected cells. MDMs generated from three different donors were infected with EBOV or RESTV or left uninfected and at 1 dpi treated with LPS for 6 h. Cellular mRNA was isolated and subjected to RNA-Seq analysis. The heat map represents the average expression intensities of the union of 4311 DE genes for each of the specified treatment conditions (left, EBOV or RESTV infection or LPS treatment compared to noninfected, untreated cells [Mock]; middle: EBOV or RESTV infection + LPS stimulation compared to infected cells only; right, EBOV or RESTV infection + LPS compared to Mock. Each column represents an experimental condition, and each row represents individual genes. (B) MDMs grown in chamber slides were infected with EBOV or RESTV or left uninfected and at 1 dpi were stimulated with LPS for 30 min where indicated. Immunofluorescence analysis was performed using antibodies directed against EBOV NP or RESTV NP (red) and p65 (green). Cell nuclei were stained with DAPI (blue). The experiment was performed three times with cells from three separate donors. Representative images are shown. EBOV NP staining is shown for Mock samples; no difference was observed for RESTV NP staining. (C) The percentage of cells (+SEM) showing nuclear p65 localization was determined by counting at least 100 infected MDMs from 3 different donors. Statistical analysis was performed using one-way ANOVA (***, P ≤ 0.0001). (D) MDMs were infected with EBOV or RESTV and at 1 dpi were stimulated with LPS for 6 h or 1 day. Cell supernatants were analyzed by Luminex assay. As controls, cells were infected with EBOV or RESTV for the indicated times without subsequent LPS stimulation, stimulated with LPS, or left untreated (Mock) for the indicated times. No statistically significant differences were observed between LPS-treated cells and cells infected with EBOV or RESTV prior to LPS treatment as determined with one-way ANOVA (P > 0.05). The experiment was performed three times using MDMs from 3 individual donors; each donor is represented by a different symbol. Horizontal bars represent mean values.

FIG 10

EBOV GP, but not RESTV GP, induces TLR4-mediated p65 nuclear translocation. (A, D, and E) MDMs grown on coverslips were treated as indicated and examined for p65 nuclear translocation by immunofluorescence analysis using a p65 antibody (green). Cell nuclei were stained with DAPI (blue). Experiments were performed three times using MDMs from three separate donors. Shown are representative images. (A) MDMs were treated with gamma-irradiated EBOV or RESTV or were mock treated for 1 h. (B) The percentage of cells (+SEM) showing nuclear p65 localization was determined by counting at least 100 cells from three different donors. Statistically significant differences from mock-treated samples as determined by one-way ANOVA are indicated (***, P ≤ 0.0001). (C) Western blot analysis of eVLPs and rVLPs using antibodies against EBOV GP1, RESTV GP1, and EBOV VP40. The EBOV VP40 antibody cross-reacts with RESTV VP40 at a lower affinity and was used to visualize both EBOV and RESTV VP40 proteins. Protein bands were visualized using the Li-Cor Odyssey system. E, EBOV proteins; R, RESTV proteins. (D) MDMs were treated with eVLPs or rVLPs for 1 h. Top row, GP-containing VLPs; middle row, GP-free VLPs; bottom row, VLPs containing the heterologous GP. (E) MDMs were treated for 3 h with LPS-RS or left untreated before stimulation with LPS or VLPs for 1 h. (F) The percentage of cells (+SEM) showing nuclear p65 localization was determined by counting at least 100 cells from three different donors. Light gray bars, untreated; dark gray bars, LPS treated; black bars, treated with VLPs containing EBOV GP; hatched bars, treated with VLPs without GP; white bars, treated with VLPs containing RESTV GP. Statistically significant differences of LPS-RS-treated samples from corresponding untreated samples as determined by one-way ANOVA are indicated (***, P ≤ 0.0001).