Proteomics, the large-scale study of proteins, plays a pivotal role in decoding clinical samples and unraveling the intricate molecular underpinnings of various health conditions. By scrutinizing the protein composition within clinical samples, we can gain profound insights into the underlying mechanisms of diseases, identify potential biomarkers, and discover novel therapeutic targets.
In this context, proteomics leverages advanced techniques to extract, identify, and quantify proteins in biological samples. The methodology typically involves several key steps: sample preparation, mass spectrometry analysis, and subsequent data processing. It's a process that demands precision, sensitivity, and extensive data handling.
Let's delve into a real-life case to illustrate the power of proteomics in clinical sample interpretation.
Case. Proteomic Analysis of COVID-19 and Associated Conditions Reveals Tissue-Specific Dysregulation [1]
Background
The goal of the study is to comprehend the proteome alterations brought on by COVID-19 and how they affect different human organs. It aims to provide light on the molecular pathogenesis of SARS-CoV-2 infection as well as offer suggestions for new treatments.
Samples
For the study, formalin-fixed, paraffin-embedded (FFPE) tissue samples from COVID-19 patients and non-COVID-19 patients who underwent autopsies were used. The following organs' tissues—heart, lung, liver, kidney, spleen, testis, and thyroid—were examined alongside controls from COVID-19 patients.
Technical Methods
Sample Preparation: Approximately 1 to 1.5 mg of formalin-fixed, paraffin-embedded (FFPE) tissue samples were processed for peptide extraction. The tissue samples underwent a series of steps, including dewaxing, rehydration, acidic and basic hydrolysis, and lysis with a buffer containing 6 M urea and 2 M thiourea. Reduction and alkylation were carried out using Tris(2-carboxyethyl)phosphine (TCEP) and iodoacetamide (IAA). Enzymatic digestion was performed using a combination of Lys-C and trypsin, and the reaction was quenched with trifluoroacetic acid (TFA).
TMT Labeling: The resulting peptides from each of the 218 tissue samples, as well as 37 technical replicates and 34 pooling samples, were labeled with TMTpro 16plex reagents. Technical replicates were included to assess the reproducibility of the proteomic analysis.
Sample Pooling: Two types of pooling were done: common pooled samples and tissue-specific pooled samples. Common pooled samples contained equal amounts of peptides from six different human organs, while tissue-specific pooled samples combined peptides from all samples of a specific organ. Pooling was essential for batch alignment and quantitative accuracy in the subsequent analysis.
Peptide Fractionation: The labeled peptides were fractionated using liquid chromatography (LC) with a 60-minute gradient from 5% to 35% acetonitrile (ACN) in 10 mM ammonia (pH = 10.0). The fractionation resulted in 60 fractions that were later combined into 30 samples to reduce complexity.
Mass Spectrometry: The peptide samples were analyzed using a Q Exactive HF-X hybrid Quadrupole-Orbitrap mass spectrometer, or a Q Exactive HF hybrid Quadrupole-Orbitrap. The instrument settings and LC-MS/MS conditions were consistent with previous studies.
Proteomics Data Analysis: Raw mass spectrometry data were processed using Proteome Discoverer software (Version 2.4.1.15, Thermo Fisher Scientific). The FASTA database for analysis included 20,365 reviewed Homo sapiens proteins and 10 SARS-CoV-2 virus proteins. Data analysis parameters included precursor and product ion mass tolerances and other settings as per established protocols.
Quality Control: Rigorous quality control measures were implemented at multiple levels to ensure data accuracy and reliability. QC included instrument performance evaluation, blank injections, and the calculation of coefficients of variation (CV) for proteomic data. Unsupervised clustering techniques like heatmaps and t-distributed stochastic neighbor embedding (t-SNE) were used to assess data quality.
Statistical Analysis: Proteomic data were subjected to statistical analysis to identify differentially expressed proteins. Significantly dysregulated proteins were selected based on adjusted p-values (B-H correction) and fold-change criteria.
Pathway/Network Analysis: Pathway enrichment analysis was conducted using tools such as IPA, Metascape, and String. Immunological proteins were mapped against GSEA-immunologic gene sets in Metascape for further analysis.
Results
The study identified 5,336 significantly dysregulated proteins in COVID-19 patients across different organs, revealing changes related to immune response, protein translation, coagulation disorders, angiogenesis, and profibrotic processes. Crosstalk among organs indicated a hyperinflammatory environment and tissue hypoxia following SARS-CoV-2 infection. The research highlighted the potential roles of various proteins in different organs, shedding light on organ-specific impacts of COVID-19.
Experimental Design
Proteogenomic analysis was performed on 218 samples
Whole genome sequencing (WGS), RNA sequencing, and quantitative proteomics and phosphoproteomics analyses were performed on 218 samples. In total, 8,802 proteins and 18,235 phospho-sites were identified and quantified.
Disease Subtyping
Consensus clustering analysis of the proteomics data resulted in the classification of 8 subtypes: Ependy, Medullo, Aggressive, Cranio/LGG-BRAFV600E (C4), HGG-rich, Ganglio-rich, LGG BRAFWT-rich, and LGG BRAFFusion-rich (C8).
In the realm of proteomics, craniopharyngiomas (CP) exhibited a distinct categorization at the proteomic level, aligning with the C4 and C8 subtypes. This classification was consistent with the data derived from phosphorylated proteomics. However, at the RNA level, no such subtyping patterns were discerned, and the correlation with proteomic data was notably limited. In essence, while the Medullo subtype correlated with histological diagnoses, the remaining subtypes comprised a heterogeneous mixture of different diagnostic profiles.
To further validate the findings from proteomic and phosphorylated proteomic analyses, a targeted mass spectrometry (MRM) investigation was conducted on the same set of craniopharyngioma samples. The MRM measurements of key molecules within the MEK/ERK pathway convincingly revealed substantial disparities between craniopharyngioma C4 and C8, effectively classifying these two CP subtypes. This underscores the potential feasibility of employing MRM technology for subtype classification in a clinical context.
Immune Infiltration Characterization Analysis
Utilizing single-cell RNA-seq data from glioblastomas, five distinct tumor subtypes were delineated, each characterized by unique immune and stromal attributes: Cold-medullo, Cold-mixed, Neuronal, Epithelial, and Hot. A comparative analysis with the proteomic subtypes unveiled intriguing patterns: the more aggressive subtypes, namely aggressive, Medullo, and Ependy, exhibited diminished immune infiltration. In contrast, the Cranio/LGG-BRAFV600E, LGG BRAFWT-rich, and LGG BRAFFusion-rich subtypes displayed heightened levels of immune infiltration.
Assessing the Impact of Gene Mutations and Copy Number Variations on the RNA/Proteome Level
Compared to LGG BRAFWT tumors, LGG-BRAFV600E mutant tumors exhibited a significant decrease in BRAF protein abundance, while no significant decrease was observed at the transcript level. CTNNB1 mutations led to an increase in protein/RNA levels in chordoma samples, while NF1 mutations resulted in a downregulation of homologous proteins and transcripts in HGG. A total of 1541 genes were identified, where their transcript and protein abundances were concurrently influenced by alterations in their own gene copy numbers in one or more diagnostic categories.
Analysis of Potential Drug Targets
Further exploration of the proteogenomic data from pediatric brain cancer tissues, both LGG and HGG, suggests that targeting the MEK and mTOR pathways may be effective in inhibiting RAF/MAPK signaling in LGG tumors, while CDK inhibitors and MEK inhibitors may be effective in HGG tumors. This provides a theoretical basis for relevant therapeutic strategies.
Comparison of Primary and Recurrent Tumor Characteristics
Proteogenomic data analysis of 18 pairs of clinical samples reveals differences in activated pathways between most primary and recurrent genomes, indicating that tumor recurrence (or progression) may involve distinct tumorigenic mechanisms.
Reference
- Nie, Xiu, et al. "Multi-organ proteomic landscape of COVID-19 autopsies." Cell 184.3 (2021): 775-791.
You may be interested in:
