Adenovirus infection triggers a rapid, MyD88-regulated transcriptome response critical to acute-phase and adaptive immune responses in vivo

Abstract

Nearly 50 years ago, the discovery of interferon prompted the notion that host cells innately respond to viral invasion. Since that time, technological advances have allowed this response to be extensively characterized and dissected in vitro. However, these advances have only recently been applied to highly complex, in vivo biological systems. To this end, we exploited high-titer adenovirus (Ad) vectors to globally investigate the innate immune response to nonenveloped viral infection in vivo. Our results indicated a potent cellular transcriptome response shortly after infection, with global assessments revealing significant dysregulation in approximately 15% of the measured transcripts derived from Ad vector-transduced tissue. Bioinformatics-based transcriptome analysis revealed a complex innate response to Ad infection, with induction of proinflammatory responses (and suppression of metabolism and mitochondrial genes) akin to those observed when mice are challenged with lipopolysaccharide. Despite this commonality, there were many unique aspects of the Ad-dependent transcriptome response, including the upregulation of several RNA regulatory mechanisms and apoptosis-related pathways, accompanied by the suppression of lysosomal and endocytic genes. Our results also implicated the Toll-like receptors (TLRs) in these responses, prompting specific investigations into this pathway. By using MyD88KO mice, our results confirmed that Ad-induced dysregulation of five functionally related gene clusters are significantly dependent on this TLR adaptor gene. MyD88 deficiency also resulted in significantly diminished, although not abolished, adaptive and acute-phase immune responses to Ad, confirming the transcriptome data, as well as specifically identifying MyD88 as a significant Ad immunity amplifier and regulator in vivo.

FIG. 1.

Expression pattern of genes significantly dysregulated by Ad. (A) One-way ANOVA after hierarchical clustering of expression patterns of Ad significantly affected genes (P < 0.01) using MO20k arrays across the full time course of infection. n = 3 for all groups except the [E1-E2b-E3-] 6 hpi group, for which n = 2. (B) Expression patterns of Ad significantly affected genes (P < 0.05 [BH FDR correction]) using MO30k arrays at 6 hpi. In both panels, columns represent independent infections of mice arranged by the infectious parameters listed above. Horizontal rows represent particular genes (colored according to expression level) clustered by using a Pearson correlation, as shown to the left of the columns. Transcriptionally high-expression genes are indicated in red, intermediately expressed genes are yellow, and minimally expressed genes are blue. Intermediate colors are indicated in the legend. Nonhybridized probes are gray.

FIG. 2.

Gene groups dysregulated by Ad at 6 hpi. A proportional representation of each functional category among significantly Ad-affected GO gene groups whose expression significantly increased (top left), significantly increased and >3-fold (bottom left), significantly decreased (top right), or significantly decreased and >3-fold (bottom right) in response to Ad infection is shown. Significantly affected gene groups were determined by Ad infection at 6 hpi in an MO30k array experiment (one-way ANOVA, P = 0.05 [BH FDR correction]) and tested against all present genes on array for over-representation of GO functions by using EASE (P < 0.05) as described in Materials and Methods. Groups are listed proportionally as determined by gene number, with highly homologous groups of genes being merged together to help eliminate redundancy.

FIG. 3.

Identification of MyD88-dependent Ad-dysregulated genes. Heat maps of genes whose expression was found to be significantly different between Ad-infected MyD88^+/+, MyD88^+/−, or MyD88KO mice (one-way ANOVA, P = 0.05 [BH FDR correction]) were hierarchically clustered by using a Pearson correlation and subjected to QT clustering (correlation = 0.9, minimum group size = 50), forming five distinct clusters, whose functions are noted to the right. The MyD88 genotype is indicated at the top (with Mock data containing n = 4 of each MyD88^+/+, MyD88^+/−, and MyD88^−/−genotype), with each column representing the microarray results from an individual mouse.

FIG. 4.

Significantly elevated plasma cytokine and chemokine concentrations in Ad-infected MyD88+ (WT) mice compared to Ad-infected MyD88KO mice. n = 28 for mock injected mice (all time points); n = 10 for MyD88KO mice (all time points). n = 12, 12, and 10 at 1, 6, and 24 hpi, respectively, for MyD88⁺ mice; n = 12, 12, and 6 at the same respective time points for MyD88KO mice. Error bars indicate the standard deviations. These results were obtained over the course of five independent experiments. “*” or “**” indicate time points when the levels of the respective cytokine or chemokine were significantly different (P < 0.05 or P < 0.01, respectively) between Ad-injected and mock-injected mice; “#” or “##” indicate time points when the levels of cytokine or chemokine were significantly different (P < 0.05 or P < 0.01, respectively) between Ad-injected MyD88⁺ and MyD88KO mice.

FIG. 5.

Significantly diminished acute-phase and adaptive immune responses in Ad-infected MyD88KO mice compared to MyD88 heterozygous mice. “*” or “**” indicate time points when the levels of the respective cytokine or chemokine were significantly different (P < 0.05 or P < 0.01, respectively) between Ad-injected and mock-injected mice; “#” or “##” indicate time points when the levels of cytokine or chemokine were significantly different (P < 0.05 or P < 0.01, respectively) between Ad-injected MyD88Het (MyD88^+/−) and MyD88KO mice. (A) SAA concentration after Ad infection. n = 6 and 11 at the 24- and 48-hpi time points, respectively, for mock-injected mice; n = 7 and 14 for the respective time points in Ad-infected wild-type mice. n = 6 and 10 for these time points in Ad-injected MyD88KO mice. Error bars indicate the standard deviations. (B) Anti-Ad antibody ELISA. Serum from immunized mice (14 days after Ad injection) was quantified by using an Ad-specific ELISA as described in Materials and Methods. The data are shown as absorbance values at appropriate dilutions and are representative of multiple experiments using triplicate groups. (C) Day 14 CEA-specific ELISPOT IFN-γ responses of MyD88Het (MyD88^+/−) and MyD88KO mice after exposure to a CEA-transducing Ad. Splenocytes derived from the respectively treated mice (14 days after Ad injection) were cultured and exposed to the respective peptides (antigen-specific CEA or nonspecific pp65) as described in Materials and Methods. The results represent the triplicate average of three mice per group ± the standard deviation.