Dysregulated Gene Expression in Lymphoblasts from Parkinson's Disease

Abstract

Parkinson's disease is the second largest neurodegenerative disease worldwide and is caused by a combination of genetics and environment. It is characterized by the death of neurons in the substantia nigra of the brain but is not solely a disease of the brain, as it affects multiple tissues and organs. Studying Parkinson's disease in accessible tissues such as skin and blood has increased our understanding of the disease's pathogenesis. Here, we used lymphoblast cell lines generated from Parkinson's disease patient and healthy age- and sex-matched control groups and obtained their whole-cell transcriptomes and proteomes. Our analysis revealed, in both the transcriptomes and the proteomes of PD cells, a global downregulation of genes involved in protein synthesis, as well as the upregulation of immune processes and sphingolipid metabolism. In contrast, we discovered an uncoupling of mRNA and protein expression in processes associated with mitochondrial respiration in the form of a general downregulation in associated transcripts and an upregulation in proteins. Complex V was different to the other oxidative phosphorylation complexes in that the levels of its associated transcripts were also lower, but the levels of their encoded polypeptides were not elevated. This may suggest that further layers of regulation specific to Complex V are in play.

Keywords: Parkinson’s disease; cell models; mitochondria; oxidative phosphorylation; protein synthesis; proteome; transcriptome.

Figure 1

Global changes in protein and RNA expression.

Figure 2

Most significantly enriched cellular components containing downregulated proteins. The Panther analysis of the annotated GO cellular components yielded a total of 30 cellular components that were significantly enriched in the downregulated proteins (Supplementary Information Table S1). Many components were represented by multiple GO terms—e.g., proton-transporting ATP synthase complex and mitochondrial proton-transporting ATP synthase complex—and were merged together. The merged GO cellular components were plotted against the mean fold enrichment, with lower and higher values represented by the error bars. The mean false discovery rate (FDR) for each merged GO cellular component is shown.

Figure 3

Most significantly enriched cellular components containing upregulated proteins. PanTable 26. cellular components that were significantly enriched in the upregulated proteins (Supplementary Information Table S2). Many components were represented by multiple GO terms and were merged together. The merged GO cellular components were plotted against the mean fold enrichment, with lower and higher values represented by the error bars. The mean p value for each merged cellular component, as calculated by t tests, are shown.

Figure 4

Most significantly enriched cellular components in downregulated transcripts. Panther analysis of annotated GO cellular components yielded a total of 101 cellular components that were significantly enriched in the downregulated transcripts (Supplementary Information Table S3). Many components were represented by multiple GO terms and were merged together. The merged GO cellular components were plotted against the mean fold enrichment, with lower and higher values represented by the error bars. The mean False Discovery Rates (FDR) are shown.

Figure 5

Most significantly enriched cellular components in upregulated transcripts. Panther analysis identified a total of 20 cellular components that were significantly enriched in the upregulated transcripts (Listed in Supplementary Information Table S4). GO cellular components represented by more than one GO term were merged together, reducing the list to eight. The mean fold enrichment of each cellular component is shown, with error bars representing higher and lower fold enrichments for the merged terms. False Discovery Rates (FDR) are shown for each cellular component.

Figure 6

Enrichment of GO biological processes in significantly downregulated proteins. Panther pathway analysis of GO biological processes enriched in the downregulated proteins identified 13 processes all associated with protein biosynthesis. The GO biological process was plotted against the fold enrichment and the significance is shown for each as a False Discovery Rate (FDR) value.

Figure 7

Most significantly enriched GO biological processes in upregulated proteins. The top 10 most enriched GO biological processes in the upregulated proteins were plotted against fold enrichment. The highest fold-change was 8.78, and this was the case for all of the top ten processes. In each case, all of the proteins identified for a biological process in the full reference list were also present in the upregulated proteins. The statistical significance (p value) as determined by t-tests varied depending on how many proteins were identified in the biological process. Redundant terms and processes not directly relevant to the cell type were removed from this list. The full list of biological processes can be found in Supplementary Material Table S5.

Figure 8

Most significantly enriched GO biological processes in downregulated transcripts. Panther pathway analysis was used to detect the enriched biological processes represented in the downregulated transcripts. Redundant overlapping terms were merged together and processes not directly relevant to the cell type were removed. The top 10 most enriched processes are depicted with their mean fold enrichment. Error bars represent the upper- and lower-fold enrichment values for the merged terms. Statistical significance is indicated by the false discovery rate (FDR)—for the merged terms, this is the mean FDR.

Figure 9

Most significantly enriched GO biological processes in upregulated transcripts. Using Panther pathway analysis in upregulated transcripts and with correction for false discovery rate, no GO biological processes were enriched. In case the FDR was too conservative and the biological processes were spread across many processes, the FDR correction was removed and the analysis repeated. This identified 606 GO biological processes—the full list is shown in Supplementary Material Table S7. Redundant terms were merged together and processes not directly relevant to the cell type were removed. The top 10 enriched biological processes are shown with their mean fold enrichment, and the significance determined by t-tests is indicated by the p value. Many enriched pathways had a fold-change of >100, and in these cases, had only one transcript identified as belonging to that pathway in both the full list of transcripts and the upregulated transcripts. Other enriched biological processes had a fold-change enrichment of 59.38 and had one of two identified transcripts in the full dataset that were present in the upregulated transcripts.

Figure 10

Changes in the abundance of mitochondrial transcripts and proteins. Venn diagrams depict the number of differentially expressed gene products associated with mitochondria in lymphoblast cell lines generated from Parkinson’s disease patients compared to healthy control samples in whole-cell proteomes and whole-cell transcriptomes. A significantly greater number of mitochondrial proteins were upregulated than downregulated in the proteome (2.2 × 10⁻¹⁶), whereas a significantly higher number of mitochondrial transcripts were downregulated in the transcriptome than were upregulated (p = 9.29 × 10⁻¹²). The p values were calculated using a proportion test in R.