Proteomic profiling of striatal tissue of a rat model of Parkinson's disease after implantation of collagen-encapsulated human umbilical cord mesenchymal stem cells

Abstract

Parkinson's disease (PD) is the most common neurodegenerative disorder of movement worldwide. To date, only symptomatic treatments are available. Implantation of collagen-encapsulated human umbilical cord mesenchymal stem cells (hUC-MSCs) is being developed as a novel therapeutic approach to potentially modify PD progression. However, implanted collagen scaffolds may induce a host tissue response. To gain insight into such response, hUC-MSCs were encapsulated into collagen hydrogels and implanted into the striatum of hemi-Parkinsonian male Sprague-Dawley rats. One or 14 days after implantation, the area of interest was dissected using a cryostat. Total protein extracts were subjected to tryptic digestion and subsequent LC-MS/MS analyses for protein expression profiling. Univariate and multivariate analyses were performed to identify differentially expressed protein profiles with subsequent gene ontology and pathway analysis for biological interpretation of the data; 2,219 proteins were identified by MaxQuant at 1% false discovery rate. A high correlation of label-free quantification (LFQ) protein values between biological replicates (r = .95) was observed. No significant differences were observed between brains treated with encapsulated hUC-MSCs compared to appropriate controls. Proteomic data were highly robust and reproducible, indicating the suitability of this approach to map differential protein expression caused by the implants. The lack of differences between conditions suggests that the effects of implantation may be minimal. Alternatively, effects may only have been focal and/or could have been masked by nonrelevant high-abundant proteins. For follow-up assessment of local changes, a more accurate dissection technique, such as laser micro dissection, and analysis method are recommended.

Keywords: Parkinson's disease; collagen hydrogels; mesenchymal stem cells; proteomics; rat model; regenerative medicine.

FIGURE 1

Injection area. (a) Stereotaxic coordinates. In black, lesion in the medial forebrain bundle using 6‐hydroxydopamine (AP: −4.0‐mm bregma). In grey, transplantation plane of collagen‐encapsulated cells in the striatum (AP: 0.0‐mm bregma). Both lesion and transplantation were performed unilaterally at the right hemisphere. (b) Coronal view of the transplantation side. The dot represents the implant and the square the dissected area. AP, anterior–posterior [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

Pearson's correlations. (a) Matrix of Pearson's correlations across all samples in all treatment groups. All samples presented similar expression profiles (r > .88). (b) Example of the correlation of all 2,219 identified proteins individually for the highest 1‐to‐1 correlation that was observed (r = .99). (c) Example of the correlation of all 2,219 identified proteins individually for the lowest 1‐to‐1 correlation that was observed (r = .88). Blood‐derived proteins were observed in one sample only, a tissue that presented a red bloody area at dissection time and resulted in contamination with blood‐derived proteins in the lysate. LFQ, label free quantification [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 3

Principal component analysis of protein expression profiles of rat striatum indicating variations between intact brain and implanted brains with collagen scaffold, human umbilical cord mesenchymal stem cells (hUC‐MSCs), or collagen‐encapsulated hUC‐MSCs at either 1 or 14 days after implantation. Similar characteristics between groups can be observed. PC, principal component [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 4

Treatment group comparisons. (a) One‐way ANOVA with post hoc test. The log 10 of the non‐corrected p value for all quantified proteins was plotted on the y‐axis and each protein at the x‐axis. No significant different protein levels between groups were found. The post hoc Fisher's least significant difference adjusted p value cutoff was .05. (b) Example of a volcano plot (intact brain Day 1 vs. implanted collagen‐encapsulated hUC‐MSC Day 1). Volcano plots connect mathematical and biological differences. At the y‐axis, the negative logarithm base 10 of the non‐corrected p value, −log 10(non‐corrected p value), was plotted. Threshold was set at p value = .05, which corresponds to a value of 1.3 after log transformation. At the x‐axis, the logarithm base 2 of the fold change, log2(FC), was plotted and the threshold was set at 2, which corresponds to a log2(FC) of 1. Note that the graph shows the non‐corrected p value and not the adjusted p value after post hoc test. After post hoc, no significant differences on expression profiles were observed

FIGURE 5

Pearson's correlation matrix between lesioned and intact area (I). High correlations between samples were observed (r > .89) [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 6

Group comparison. (a) T‐test with Bonferroni post hoc test. The logarithm base 10 of the non‐corrected p value for all identified proteins was plotted on the y‐axis. No significant different protein levels were found. Note that the adjusted p values after post hoc test are not plotted. (b) Volcano plot. On the y‐axis, the logarithm base 10 of the non‐corrected p values are plotted at a threshold of .5, corresponding to a value of 1.3 after log transformation (horizontal line). On the x‐axis, the logarithm base 2 of the fold change (FC) is plotted with a threshold set at 2.0, which corresponds to 1 after log transformation (vertical lines). Note that the graph shows the non‐corrected p value and not the adjusted p value after post hoc test. After post hoc, no significant differences on expression profile were observed