Supplementary Figure 1. Fold-increase in transcripts for indicated host and viral genes from a pooled sample is representative of data from individual mice. To ensure that pooling was an appropriate strategy for generating experimental samples for microarray analysis, samples from individual infected aged BALB/c mice obtained at day 2 were checked for similarity of cellular and viral transcripts using a panel of Taqman primer/probes (e.g., Irf1, Irf7, Ccl4, Stat1, Gapdh, SARS-CoV M, SARS-CoV NP). The genes are indicated on the x-axis and the fold increase of transcripts indicated on the y-axis. The first four bars are individual mouse samples. The final bar is an average with standard deviation.
Supplementary Figure 2 Didactic Figure to Show Experimental Design. The left side shows how samples A, B, C, and D are physically co-hybridized to microarrays to generate gene expression data. With the physical hybridization, young mock-infected mice were used as the “reference sample” so that data could be compared with data from our historical microarray database that study young mice. Computationally, the physical hybridization experiments can be manipulated to create in silico hybridization experiments. Using the “re-ratio” feature of Resolver, we compared each aged infected mouse hybridization (A/D) against that of the mock-infected mouse hybridization (C/D) virtually (in silico) and in this manner obtained a time course where each aged infected hybridization was compared to an aged mock-infected hybridization. The re-ratio design creates new data sets based on comparisons of channels from existing ratio hybridizations without having to prepare new hybridizations (32).
Supplementary Figure 3 Heat maps show early, intermediate, late, early/intermediate, and intermediate/late differential cellular gene expression in lung samples of young SARS-CoV infected mice. Data are presented for young (Y) mice on indicated days (D) 1, 2, 5 and 7 post-infection. Heat maps highlight the different areas of the Venn diagram in Figure 2. Early, intermediate and late genes were defined in the text. Briefly, genes shown in the Venn diagram and subsequent heat map met the criteria of an absolute fold-change ³ 2 and p £ 0.01 in at least two time points. Early genes met the criteria on both days 1 and 2, intermediate genes met the criteria on both days 2 and 5, and late genes met the criteria on both days 5 and 7 post-infection. Therefore this representation is a subset of the global gene expression present in Figure 1C. Gene expression profiles are the result of comparing gene expression in lungs of experimentally infected mice (pooled, n=4) with gene expression in lungs of mock infected mice (pooled, n=5)
Supplementary Figure 4 Heat maps show early, intermediate, late, early/intermediate, and intermediate/late differential cellular gene expression in lung samples of aged SARS-CoV infected mice. Data are presented for aged (A) mice on indicated days (D) 1, 2, 5 and 7 post-infection. Heat maps highlight the different areas of the Venn diagram in Figure 3. Early, intermediate and late genes were defined in the text. Briefly, genes shown in the Venn diagram and subsequent heat map met the criteria of an absolute fold-change ³ 2 and p £ 0.01 in at least two time points. Early genes met the criteria on both days 1 and 2, intermediate genes met the criteria on both days 2 and 5, and late genes met the criteria on both days 5 and 7 post-infection. Therefore this representation is a subset of the global gene expression present in Figure 1C. Gene expression profiles are the result of comparing gene expression in lungs of experimentally infected mice (pooled, n=4) with gene expression in lungs of mock infected mice (pooled, n=5)
Supplementary Figure 5 Aged mice showed differences compared to young mice even before viral infection. Many genes involved in cell cycle and T-cell receptor signaling were down regulated. Immunoglobulin related genes were also affected. Fold changes in gene expression are the result of comparing gene expression in lungs of aged mock-infected mice (pooled, n=4) with gene expression in lungs of young mock-infected mice (pooled, n=4).
Supplementary Figure 6 Network interactions show that a number of genes with lower levels of gene expression in mock-infected aged mice when compared with mock-infected young mice. These changes are shown mapped onto the merged network from Figures 6A and 6B. These diagrams, which comprise the two top networks, represent the interactions, both direct (solid lines) and indirect (dashed lines), between genes and gene products identified during peak viral replication. Network 1 was the most prominent network detected at day 2 in young SARS-CoV infected mice and was composed of cell cycle, cellular development and hematological system development functions. Network 2 was the most prominent network detected at day 2 in aged SARS CoV infected mice and was composed of immune response, inflammatory response, and cell-to-cell signaling functions. Pink indicates a fold change ³ 2, and grey indicates a fold change < 2. Dark green indicates a fold change ³ 5, light green indicates a fold change ³ 2, and grey indicates an absolute fold change < 2. Gene expression profiles are the result of comparing gene expression in lungs of experimentally 1 infected mice (pooled, n=4) with gene expression in lungs of mock-infected mice (pooled, n=5).
Supplementary Figure 7 Pathway Analyses of Jak/Stat/Ifn in (A) aged and (B) young SARS-CoV infected mice. The top bar of the individual gene heat maps shows expression at day 2, the middle bar shows expression at day 5, and the lower bar shows expression at day 7. A gene is shown in the pathway if it meets the criteria of an absolute fold-change ³ 2 and p £ 0.01 in at least one experiment. We selected probes with the strongest signals for any one gene for the purpose of these diagrams. This figure was produced using the Resolver® 4.0 (Rosetta Biosoftware) and Pathway Builder Tool (Protein Lounge, San Diego, CA). A small subset of genes, Cxcl10, Ddx58, Ifnb, Irf7, Isgf3g, Mx2, Stat1, and Stat2 indicated on the x-axis were evaluated using Taqman in aged (top right) and young (bottom right) mice. The log (ratio) is indicated on the y-axis. Gene expression profiles are the result of comparing gene expression in lungs of experimentally infected mice (pooled, n=4) with gene expression in lungs of mock infected mice (pooled, n=5).
Supplementary Figure 8 Global gene expression shows that aged mice mount an inappropriate response to virus. A union of the genes was created from young mouse response (196 from Figure 2) and aged mouse response (237 from Figure 3). Data are presented for aged (A) or young (Y) mice on indicated days (D) 1, 2, 3, 5 and 7 postinfection. Aged mice infected with a non-lethal SARS coronavirus show more similarity to young mice infected with a highly virulent and lethal influenza virus than young mice infected with a non-lethal influenza virus. The mice in these experiments were all BALB/c mice. Gene expression profiles for SARS-CoV experiments are the result of comparing gene expression in lungs of experimentally infected mice (pooled, n=4) with gene expression in lungs of mock-infected mice (pooled, n=5). Gene expression profiles for influenza experiments are the result of comparing gene expression in lungs of experimentally infected mice (pooled, n=3) with gene expression in lungs of mock infected mice (pooled, n=3) (18).