A systems approach to prion disease

Abstract

Prions cause transmissible neurodegenerative diseases and replicate by conformational conversion of normal benign forms of prion protein (PrP(C)) to disease-causing PrP(Sc) isoforms. A systems approach to disease postulates that disease arises from perturbation of biological networks in the relevant organ. We tracked global gene expression in the brains of eight distinct mouse strain-prion strain combinations throughout the progression of the disease to capture the effects of prion strain, host genetics, and PrP concentration on disease incubation time. Subtractive analyses exploiting various aspects of prion biology and infection identified a core of 333 differentially expressed genes (DEGs) that appeared central to prion disease. DEGs were mapped into functional pathways and networks reflecting defined neuropathological events and PrP(Sc) replication and accumulation, enabling the identification of novel modules and modules that may be involved in genetic effects on incubation time and in prion strain specificity. Our systems analysis provides a comprehensive basis for developing models for prion replication and disease, and suggests some possible therapeutic approaches.

Figure 1

Strategies for identification of 333 core differentially expressed genes (DEGs) and their functional analysis in mouse prion diseases. Two prion strains (RML and 301V) were used for inoculating mice from six different genetic backgrounds (B6, B6.I, FVB, Tg4053, 0/+, and 0/0) to generate eight prion–mouse combinations. From the list of 7400 DEGs identified from at least one of the five combinations with normal levels of prion protein (1X), 333 DEGs shared by all five were selected through novel statistical methods to represent perturbed networks essential to prion pathophysiology. Venn diagram shows the overlap of the 333 DEGs with DEGs from Tg4053-RML (mice expressing eight times of normal prion protein levels) and from 0/+-RML (mice expressing one-half of normal prion protein levels). Among 333 DEGs, 161 genes were mapped to networks through protein–protein interaction network or metabolic pathways. Also, by comparison of 333 DEGs with previous prion microarray studies, we identified 178 DEGs that have not been reported in connection with prion disease.

Figure 2

Expression profiles of genes shared among comparison of five mouse–prion combinations or unique to Tg4053-RML. Dynamic gene expression profiles in all eight mouse strain–prion strain combinations are shown for three cases in the order of B6-RML, B6-301V, B6.I-RML, B6.I-301V, FVB-RML, Tg4053-RML, 0/+-RML, and 0/0-RML: (A) top 20 genes with highest statistical significance for involvement in prion pathology among 333 differentially expressed genes (DEGs) shared by five combinations, B6-RML, B6-301V, B6.I-RML, B6.I-301V, and FVB-RML; (B) bottom 20 genes with lowest statistical significance for involvement in prion pathology among 333 shared DEGs; and (C) top 20 DEGs with highest statistical significance for involvement in prion pathology among RML-infected Tg4053 mice.

Figure 3

Time course histoblots and profiles of selected DEGs in three mouse strains with varying PrP^C concentration. (A) Histoblots are shown for dynamic patterns of accumulation of proteinase K-resistant PrP^Sc in brains of RML-infected mice with wild-type PrP^C concentration (FVB), half of wild-type PrP^C concentration (0/+), and approximately eight times of wild-type PrP^C concentration (Tg4053). (B) Expression profiles of shared DEGs representing key network modules are shown for FVB, Tg4053, and 0/+ mice. No significant changes were detected for these genes in mice lacking the Prnp gene (0/0). Percentage of the incubation times (incubation time is defined in this paper as the time period from prion inoculation to end point) for the time points when gene expression changes start to occur are indicated with arrows in histoblot series (A) and gene expression profiles (B).

Figure 4

Dynamic PrP replication and accumulation network. Hypothetical networks of proteins and metabolites that are potentially involved in PrP replication and accumulation were constructed starting from the list of 333 shared DEGs and protein–protein interaction/metabolic pathway information from public databases. Relative changes of the transcripts for the corresponding proteins are represented in color changes: red—upregulation, green—downregulation, yellow—no change. Data for the transcriptional changes are from BL6 mice infected with RML prions, at 6 weeks (A), 10 weeks (B), 14 weeks (C), 18 weeks (D), 20 weeks (E), and 22 weeks (F) after inoculation. See main text for detailed description of modules 1 through 6. Ch: cholesterol; ECM: extracellular matrix; PM: plasma membrane. Large nodes indicate DEGs whose expression change patterns are shared by five prion–mouse combinations (B6-RML, B6-301V, B6.I-RML, B6.I-301V, and FVB-RML); asterisks indicate DEGs whose prion-related changes are unique in this study; genes in blue are also DEGs in RML-infected 0/+ mice; genes in bold are also DEGs in RML-infected Tg4053 mice; genes with solid underline were changed only in mice with short incubation times; genes with dotted underline were changed only in RML-infected mice. Source data is available for this figure at www.nature.com/msb

Figure 5

Microglial and astrocyte activation network. Hypothetical networks of proteins and metabolites that are potentially involved in activation of microglia and astrocytes were constructed starting from the list of 333 shared DEGs and protein–protein interaction information from public databases. Modules A–F denote distinct functional modules within the network: (A) Complement activation, (B) pattern recognition receptors, (C) cytokines and chemokines, (D) growth factors, (E) reactive astrogliosis, and (F) leukocyte extravasation. Relative changes of the transcripts for the corresponding proteins are represented in color changes: red—upregulation, green—downregulation, yellow—no change. Data for differential expression are from BL6 mice infected with RML prions. The temporal expression changes at six time points (10, 14, 16, 18, 20, and 22 weeks) are represented by the circular heatmaps of the network nodes: the center color of each node indicates the differential expression at 10 weeks, while the outer circle colors represent expression changes at increasing time points. Large nodes indicate DEGs whose expression change patterns are shared by five prion-mouse combinations; asterisks indicate DEGs whose prion-related changes are unique in this study; genes in blue are also DEGs in RML-infected 0/+ mice; genes in bold are also DEGs in RML-infected Tg4053 mice; genes with solid underline were changed only in mice with short incubation times. Source data is available for this figure at www.nature.com/msb

Figure 6

Novel functional modules with dynamic gene expression changes potentially linked to prion pathophysiology: (A) androgen metabolism, (B) iron metabolism, (C) arachidonate/phospholipid metabolism. Hypothetical networks of proteins and metabolites that are potentially involved in three novel functional modules were constructed based on genes selected from the list of 333 shared DEGs and metabolic pathway information from literature. The temporal expression changes at six time points (10, 14, 16, 18, 20, and 22 weeks after inoculating BL6 mice with RML prions) are represented by the circular heatmaps of the network nodes: the center color of each node indicates the differential expression at 10 weeks, whereas the outer circle colors represent expression changes at increasing time points. Large nodes indicate DEGs whose expression change patterns are shared by five mouse-prion combinations; asterisks indicate DEGs whose prion-related changes are unique in this study. See text for detailed description of the three pathways.

Figure 7

Expression profiles of DEGs changing predominantly in three short incubation time combinations or in two RML-infected mice. (A) Top 10 genes that show selective changes in short incubation time combinations, B6-RML, B6.I-301V, and FVB-RML as compared with long incubation time combinations (B6-301V and B6.I-RML). The changes of these genes in 0/+ is also shown as another example of mouse–prion combination with long incubation time; (B) top 10 genes that show selective changes in RML-infected combinations, B6-RML and B6.I-RML as compared with 301V-infected combinations, B6-301V and B6.I-301V.