Dysregulation of gene expression in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned mouse substantia nigra

Abstract

Parkinson's disease pathogenesis proceeds through several phases, culminating in the loss of dopaminergic neurons of the substantia nigra (SN). Although the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) model of oxidative SN injury is frequently used to study degeneration of dopaminergic neurons in mice and non-human primates, an understanding of the temporal sequence of molecular events from inhibition of mitochondrial complex 1 to neuronal cell death is limited. Here, microarray analysis and integrative data mining were used to uncover pathways implicated in the progression of changes in dopaminergic neurons after MPTP administration. This approach enabled the identification of small, yet consistently significant, changes in gene expression within the SN of MPTP-treated animals. Such an analysis disclosed dysregulation of genes in three main areas related to neuronal function: cytoskeletal stability and maintenance, synaptic integrity, and cell cycle and apoptosis. The discovery and validation of these alterations provide molecular evidence for an evolving cascade of injury, dysfunction, and cell death.

Figure 1.

Analysis of MPTP-induced injury. Unbiased stereology was used to count the number of TH-positive neurons after four doses of MPTP at either 7 or 21 d after the last dose. There was a significant reduction in the number of TH-positive cells in MPTP mice compared with saline-injected mice at both time points (^*p < 0.01). The 35% loss of cells appears to be stable, because additional death did not occur between 7 and 21 d (A). Data expressed as mean + SEM. MPTP also caused the selective loss of DAT protein in the striatum, an indication of the number of dopaminergic axons that innervate target cells there (B). Western blot illustrates striatum DAT levels from each of three mice for each group that when scanned densitometrically were significantly reduced compared with control (^*p < 0.01).

Figure 2.

Comparison of significant genes identified by two array platforms. We hypothesized that gene changes important for MPTP pathophysiology would be identified by both Affymetrix and CodeLink arrays. Only genes present on both array platforms were included in the analysis. The overlap was determined by comparing gene information associated with the probe of each platform, using only genes that were significantly differentially expressed (for MME, 425 of 729 genes; for MML, 431 of 763 genes overlapped). Although there were hundreds of gene expression changes that surpassed the significance cutoff, only 11 (A; MME vs MC) or 12 (B; MML vs MC) genes were consistently altered by ALL methods for both array platforms in our study. Of these, four were downregulated at both of the time points examined after MPTP treatment (C).

Figure 3.

ISH to validate selected gene expression changes in dopamine neurons. Quantitative grain density counts were obtained for three genes that were identified as significantly differentially expressed between MPTP- and saline-treated mice and validated by qRTPCR. These genes were MAP-2, DAT, and Pur-α. Significant reductions in grain density were observed over TH-immunopositive neurons for MAP-2, DAT, and Pur-α (^*p < 0.05). Data presented are mean + SEM.

Figure 4.

Effective sample clustering by informative genes. Affymetrix U74A probe sets were filtered by dChip to yield a 2102 member gene list for use in sample clustering. The expression patterns of these genes in the 12 arrays (saline and chronic MPTP treatments) segregated arrays into the proper treatment groups. Although all arrays are correctly partitioned by injury status, the analysis showed that the MML group was more closely related to the control group, possibly indicating some transcriptional recovery at this later time point. Numbers represent the four samples (arrays) in each of the three groups.

Figure 5.

Sample clustering by groups of functionally similar genes. Using Affymetrix U74A GeneChip data, 23 groups of genes having the same general cell function were tested for their ability to cluster the 12 SN samples from saline (MC; n = 4), MPTP early (MME; n = 4) and MPTP late (MML; n = 4). The dendrograms indicate the distance between groups. Below each dendrogram are the gene class names that created it, with the number of genes in that class in parentheses. The majority of gene groups that were able to correctly segregate the 12 samples into the three disease states did so indicating that the MML group was more closely related to the controls than the MME group. However, two gene groups were identified that not only correctly assigned each sample to the appropriate treatment group, but the expression of these genes was more similar between the two MPTP time points than either MPTP group was to controls. Nine of the functional gene groups failed to correctly segregate samples into three distinct treatment groups.

Figure 6.

PCA of MPTP-treated samples. PCA of the Affymetrix GeneChip data indicates that each treatment group's samples occupy nonoverlapping areas in the two-dimensional space defined by PC1 and PC2. The samples can also be seen to fall into two different pairings: one in which the controls and MPTP late samples (▪) share the bottom left area and the MPTP early samples (▧) are alone in the top right area, and another in which the controls (□) are alone in the far left area, whereas the two MPTP groups occupy the right side of the space. These are precisely the two different pairings seen in hierarchical clustering.