Expression microarray analysis implicates apoptosis and interferon-responsive mechanisms in susceptibility to experimental cerebral malaria

Abstract

Specific local brain responses, influenced by parasite sequestration and host immune system activation, have been implicated in the development of cerebral malaria. This study assessed whole-brain transcriptional responses over the course of experimental cerebral malaria by comparing genetically resistant and susceptible inbred mouse strains infected with Plasmodium berghei ANKA. Computational methods were used to identify differential patterns of gene expression. Overall, genes that showed the most transcriptional activity were differentially expressed in susceptible mice 1 to 2 days before the onset of characteristic symptoms of cerebral malaria. Most of the differentially expressed genes identified were associated with immune-related gene ontology categories. Further analysis to identify interaction networks and to examine patterns of transcriptional regulation within the set of identified genes implicated a central role for both interferon-regulated processes and apoptosis in the pathogenesis of cerebral malaria. Biological relevance of these genes and pathways was confirmed using quantitative RT-PCR and histopathological examination of the brain for apoptosis. The application of computational biology tools to examine systematically the disease progression in cerebral malaria can identify important transcriptional programs activated during its pathogenesis and may serve as a methodological approach to identify novel targets for therapeutic intervention.

Figure 1

Overview of differentially expressed genes in CM-susceptible and -resistant mouse brains. Application of the extraction and analysis of differential gene expression time course algorithm to the microarray data set identified more than 1100 genes as differentially expressed between CM-susceptible (C57BL/6) and -resistant (BALB/c) mice. A: Hierarchical clustering of differentially expressed genes revealed that in CM-susceptible animals, the majority underwent intense transcriptional activity late in infection. In the heatmap, red indicates up-regulation and green indicates down-regulation. B: Principal components analysis of the ∼1100 genes identified three prominent clusters of animals: CM-susceptible mice at day 6, CM-susceptible mice (days 0 to 3), and CM-resistant animals (all days). The first principal component, which primarily accounts for CM-susceptible mice on day 6, encompassed more than 50% of the total gene expression variability, reinforcing what was initially observed by clustering the temporal gene-expression profile.

Figure 2

Functional categories significantly associated with gene expression differences between C57BL/6 and BALB/c mouse brains. Functional analysis of genes differentially expressed between CM-susceptible and -resistant mice was performed using expression analysis systematic explorer software. Enriched GO categories, which are shown for each time point (columns 1 to 4) and for all differentially expressed genes (column 5), included various immunological, inflammatory, and pro-apoptotic programs. The majority of functional categories were enriched predominantly at day 6; however, some antigen presentation-related GO categories were enriched across infection.

Figure 3

Interaction network of differentially expressed genes. A gene network based on known gene product interactions was generated using IPA software and displayed using Pajek’s visualization tool (left). Genes that are hubs or nodes of high interconnectivity (≥10) are shown on the outer rim of the network. A complete list of hubs is provided in Supplemental Table 3 (see http://ajp.amjpathol.org). A module within the interactome corresponding to apoptosis-related pathways was amplified and displayed according to subcellular localization of its gene products (right). In the apoptosis module, red ovals are differentially up-regulated and green ovals are differentially down-regulated genes in CM-susceptible (C57BL/6) relative to CM-resistant (BALB/c) mice. The majority of these genes were differentially up-regulated in C57BL/6 mice, demonstrating that there is widespread activation of apoptotic pathways in the brains of CM-susceptible mice.

Figure 4

Transcription factor binding sites over-represented among differentially expressed genes. Analysis of promoter elements in differentially expressed genes using the promoter integration in microarray analysis algorithm identified transcription factors that may regulate transcriptional programs activated in the CM-susceptible versus -resistant brain. The highly enriched promoter sites IRF, IRF-1, IRF-7, and ISRE are all associated with various aspects of interferon-mediated regulation of gene expression. The number of differentially expressed genes recognized by these regulatory factors is show above each column.

Figure 5

Real-time RT-PCR validation of expression levels of immune-related transcripts. Several immune-related genes, which were highly up-regulated at day 6 in CM-susceptible compared with resistant mice, were analyzed using quantitative real-time RT-PCR to confirm microarray expression levels. To compare the microarray data (□) with the qRT-PCR data (▪), log₂ values of the expression ratio are shown indicating twofold change over baseline. The following genes that were analyzed showed a close correlation between microarray and PCR expression values: complement component 4B (C4b; A), chemokine (C-C motif) ligand 12 (Ccl12; B), chemokine (C-X-C motif) ligand 10 (Cxcl10 or IP10; C), granzyme A (D), intercellular adhesion molecule 1 (Icam1; E), signal transducer and activator of transcription 1 (Stat1; F) interferon regulatory factor 1 (Irf1; G), and interferon regulatory factor 7 (Irf7; H).

Figure 6

Histopathological analysis of mouse brains during PbA infection. Pathological changes in the brains of CM-susceptible (C57BL/6) and -resistant (BALB/c) mice were examined during PbA infections. Brains were formaldehyde fixed and examined using histopathology by observers blinded to the genetic background of the mice. Quantification of TUNEL-positive cells confirmed that significantly higher numbers of apoptotic cells were found in C57BL/6 mice compared with BALB/c. A: Average TUNEL-positive cells per slide were quantitated from four animals per strain (n = 20 slides per strain) and were significantly higher in C57BL/6 mice (Mann-Whitney test, P < 0.001). B: TUNEL-positive cells per slide are shown for each mouse examined. C: Representative images of TUNEL-positive neuronal cells in the cerebral cortex of an infected C57BL/6 mouse (left) and an infected BALB/c mouse (right). Scale bars represent 20 μm. TUNEL-labeled brain sections showing positive nuclear staining (brown color; inset at magnification, ×1000) suggested that apoptosis occurred in infected CM-susceptible C57BL/6 mice at day 6 after infection when compared with the equally parasitemic but CM-resistant BALB/c.