Method development and characterisation of the low-molecular-weight peptidome of human wound fluids

Abstract

The normal wound healing process is characterised by proteolytic events, whereas infection results in dysfunctional activations by endogenous and bacterial proteases. Peptides, downstream reporters of these proteolytic actions, could therefore serve as a promising tool for diagnosis of wounds. Using mass-spectrometry analyses, we here for the first time characterise the peptidome of human wound fluids. Sterile post-surgical wound fluids were found to contain a high degree of peptides in comparison to human plasma. Analyses of the peptidome from uninfected healing wounds and Staphylococcus aureus -infected wounds identify unique peptide patterns of various proteins, including coagulation and complement factors, proteases, and antiproteinases. Together, the work defines a workflow for analysis of peptides derived from wound fluids and demonstrates a proof-of-concept that such fluids can be used for analysis of qualitative differences of peptide patterns from larger patient cohorts, providing potential biomarkers for wound healing and infection.

Keywords: bacterial infection; biochemistry; biomarkers; chemical biology; human; immunology; inflammation; peptide profiles; peptidomics; plasma; wound fluid.

Figure 1.. Comparison of sample preparation methods.

(A) Schematic overview of the workflow. Peptides were extracted from 25 µL, 50 µL, or 100 µL of wound fluid in 6 M urea (U), 6 M urea +0.05% RapiGest (U + R), or 6M urea +0.1% TFA (U + T), using 30 kDa cut-off filters. Stored filtrates were defrosted, followed by peptide concentration using StageTips and finally 1.6 µL of the original wound fluids were analysed by nano-LC–MS/MS. (B) Representative example of a 10–20% Tricine gel run with 1 µL of sample before filtration (BF) or 22.5 µL of sample after filtration, extracted from 100 µL of wound fluid, ran under non-reducing conditions, and stained with SilverQuest stain. (C) Total numbers of identified peptides and corresponding proteins for the different buffers and amounts of wound fluid as analysed by MS. Results are shown as combined data of two injections per sample.

Figure 2.. Robustness of sample preparation method.

Peptides were extracted from 100 µL wound fluid in 6 M urea supplemented with 0.05% RapiGest using the workflow shown in Figure 1. (A) To investigate possible influence of the used filters on the obtained results, one wound fluid preparation was divided over two 30 kDa cut-off filters, centrifuged and filtrates were analysed using SDS–PAGE (U + R) and nano-LC–MS/MS. Combined results from two injections per sample are shown in Venn diagrams depicting proteins and unique peptides. For SDS–PAGE, 1 µL of sample before filtration (BF) or 22.5 µL of sample after filtration were run on a 10–20% Tricine gel under non-reducing conditions and stained with SilverQuest stain; extractions using 6M urea (U) without RapiGest are shown for comparison. (B) Reproducibility of sample injection using one preparation injected three times on the same day or (C) injected twice on two different days. (D) Reproducibility of sample preparation using two independently generated samples of the same wound fluid on different days. Combined results of two injections per sample are expressed as the protein score, the number of unique peptides per protein, and the percentage of total coverage of each protein by the identified peptides; r² values are indicated in each graph.

Figure 3.. Comparison of plasma and wound fluids.

Peptides were extracted from 100 µL acute wound fluid (aWF) or citrated plasma in 6 M urea supplemented with 0.05% RapiGest using the workflow shown in Figure 1. (A) Comparison of three wound fluids and three plasma samples as analysed using a 10–20% Tricine gel ran under non-reducing conditions and stained with SilverQuest stain. (B) Representative LC–MS/MS chromatograms of wound fluid (top) and plasma (bottom) preparations. (C) Comparison of the pooled results of three wound fluids with three plasma preparations using Venn diagrams depicting total number of identified proteins and unique peptides. (D) Distribution of peptides from representative wound fluid and plasma preparations based on molecular weight. The results are shown as total numbers (left) and normalised values (right). (E) Heatmaps comparing wound fluids and plasma depicting the protein score, the number of unique peptides per protein, and the percentage of total coverage of each protein by the identified peptides. Results are shown as combined data of two injections per sample.

Figure 4.. Comparison of five acute wound fluids.

Peptides were extracted from 100 µL wound fluid in 6 M urea supplemented with 0.05% RapiGest using the workflow shown in Figure 1. (A) Comparison of five wound fluids using Venn diagrams depicting proteins, unique peptides of the 74 proteins common for all five wound fluids and unique peptides of all identified proteins. (B) Heatmaps comparing the five wound fluids depicting the protein score, the number of unique peptides per protein, and the percentage of total coverage of each protein by the identified peptides. Results are shown as combined data of four injections per sample.

Figure 5.. Comparison of healing and non-healing infected wounds.

(A) Photos of wounds, 7 days after surgery, of six patients who had undergone facial full-thickness skin grafting. On the left side are the three that healed well and showed low inflammation and no infection, while on the right side wounds are depicted that were highly inflamed and infected with i.a. Staphylococcus aureus. Dressing extracts were made of the seven day old dressings derived from each wound and analysed for cytokine content (B), protein and peptide composition using SDS–PAGE (C), and enzymatic activity using zymograms (D). Peptides were extracted from 280 µg of wound dressing extract in 6M urea supplemented with 0.05% RapiGest using the workflow shown in Figure 1. (E) Comparison of the pooled results of the three low inflammation samples with the three high inflammation samples using Venn diagrams depicting total number of identified proteins and their unique peptides. (F) Heatmaps comparing the six samples depicting the protein score, the number of unique peptides per protein, and the percentage of total coverage of each protein by the identified peptides. Results are shown as combined data of two independent sample preparations with two injections per sample. Notably, (A) and (B) are derived from previously published results. Figure 5A is reproduced from Figure 2 of Saleh et al., 2019, and Figure 5B has been adapted from Figure 4A of Saleh et al., 2019.

Figure 5—figure supplement 1.. Venn diagrams comparing the numbers of identified proteins and their peptides of each sample in the low and high inflammation groups.

Figure 6.. Peptide profiles of fibrinogen and selected proteases.

Peptide profiles and peptide alignment maps of three fibrinogen chains and the proteases prothrombin, cathepsin G, and neutrophil elastase, were generated from the UniProt IDs, peptide sequences, start and end, and intensities for each protein using the web-based application Peptigram. The height of the green bars is proportional to the number of peptides overlapping the amino acid residue, while the intensity of the colour (green) is proportional to the sum of the intensities overlapping this position. Interesting peptide regions are highlighted by blue boxes, and of the corresponding peptides, one sequence of each identified N-terminal is shown for illustration purposes: 1–3, low inflammation samples and 4–6, high inflammation samples.

Figure 7.. Peptide profiles of selected protease inhibitors.

Peptide profiles and peptide alignment maps were generated for the protease inhibitors kininogen, alpha-1-antitrypsin, alpha-2-antiplasmin, and inter-alpha-trypsin inhibitor heavy chain H4. Interesting peptide regions are highlighted by blue boxes, and a selection of the corresponding peptides is shown for illustration purposes: 1–3, low inflammation samples and 4–6, high inflammation samples. The orange arrow indicates non-highlighted peptide sequences unique to the three low inflammation samples.

Figure 8.. Peptide profiles of selected complement factors and additional proteins.

Peptide profiles and peptide alignment maps were generated for the complement factors B, C3, and C9, as well as the proteins serum amyloid A-1, dermcidin, fetuin-A, and albumin. Interesting peptide regions are highlighted by blue boxes, and a selection of the corresponding peptides is shown for illustration purposes: 1–3, low inflammation samples and 4–6, high inflammation samples. The orange arrows indicate non-highlighted peptide sequences unique to the three low inflammation samples, while the purple arrow indicates those to the high inflammation group.