Infrared Mapping as a Predictive Tool in Wound Care

Introduction

Aberrant wound healing, as observed in the presence of co-morbidities such as diabetes, represents a major clinical problem. To improve wound care, a better understanding of the cellular and molecular details of healing processes is required. Our aim was to utilize Focal Plane Array (FPA) and synchrotron – Fourier transform infrared (FTIR) chemical mapping to further characterize the wound healing process. Human skin tissue samples were obtained from a chronic diabetic venous ulcer and a full-thickness burn with fully informed patient consent. Normal tissue was obtained from murine skin samples.

Results

Morphological differences in skin layers of diabetic foot ulcer and burn patients compared to healthy mouse skin

Figure 2. Overview of histological staining performed of control mouse ear skin, human burn (two sections of the same tissue; S1, S2) and diabetic foot ulcer patients at two time points (T1, T2). H&E, Massons’ trichrome and picrosirius red staining respectively highlights increased inflammation, collagen deposition, specifically Type I in the diabetic foot ulcer at T2 and in the burned skin in comparison to the mouse control skin and diabetic foot ulcer at T1. Silver impregnation shows greatest amounts of reticular fibres are present in the diabetic foot ulcer at T1 and burn samples. Evidence of inflammation and collagen is suggestive of blood flow and scaffold formation in the wound healing process.

Material and Methods

Histology

Control murine ear skin samples and human burn and diabetic foot ulcer patient samples were sectioned, fixed and stained using standard: haematoxylin & eosin (H&E), Massons’ trichrome, silver impregnation (reticular fibres) and picrosirius red protocols.

Focal Plane Array (FPA) microspectroscopy

Samples were fixed to CaF2 windows (Crystan) and imaged with a 15x objective in transmission mode using the Bruker Hyperion 2000 FTIR microscope (Bruker) at the Australian Synchrotron. Interferograms were recorded at a spectral resolution of 8cm-1 over the spectral range 4000-800cm-1, using a co-addition of 64 scans. The spectra were collected and processed through Opus v7.2 (Bruker) software using a Blackman-Harris 3-Term apodization, Power-Spectrum phase correction and a zero-filling factor of two and regions of interest were integrated as demonstrated in figure 1.

dth=”705″]

Figure 1. Overview of spectral processing of data collected using FPA-FTIR and synchrotron-FTIR. dividual spectra are extracted from chemical maps, and either integrated or tested using principal ponent analysis (PCA) from their second derivative.

Synchrotron – FTIR microspectroscopy and analysis

Samples were imaged in transmission mode with a 36x objective on the infrared microspectroscopy beamline (IRM) at the Australian Synchrotron. Interferograms were recorded at a spectral resolution of 4cm-1 over the spectral range 3900-750cm-1. Hypercubes were collected in rectangular grids of 140×20, 25×25 with a 5µm aperture for the mouse and burn skin respectively and a 50×50 grid map, with an aperture of ~6.9 µm for the diabetic foot ulcer. CytoSpec™ was used to separate individual spectra in the hypercube and group into epidermal and dermal regions, followed by spectral processing shown in figure 1.

Results

FPA-FTIR and synchrontrom-FTIR microspectroscopy of normal mouse ear, human burn skin and human diabetic foot ulcer

Figure 3. (A) FPA-FTIR chemical maps of mouse skin and human burn skin. The spectrum was integrated for regions; lipid composition (3000 – 2830 cm^-1), inflammatory cells (1755 – 1720 cm^-1), amide I (1695 – 1600 cm^-1), keratin (1660 – 1645 cm^-1), melanin (1410 – 1390 cm^-1) and collagen (1355 – 1325 cm^-1). Keratin was highly expressed in the epidermis of the mouse ear skin, compared to the burn sample. Collagen expression was high within the dermis of all samples. The localization of the keratin and collagen in the human skin samples aided to differentiate the epidermis and dermis of the samples. (B) Spectra collected from synchrotron-FTIR highlight the epidermal and dermal regions, as displayed in the (i) absorption spectrum, and (ii) second derivative. Significant peaks leading to the separation of the normal and burn samples are seen in the (iii) loadings plot and (iv) PCA plot. Strong absorption peaks in the lipids at 1734 and 1735 cm^-1 is present within the normal skin in both regions yet this band is absent within the burn skin. The amide III region (1350 – 1250 cm^-1) displays protein secondary structural changes within the samples. The peak for collagen is (1338cm^-1) at a lower absorption in the burn samples, yet show strong peaks at 1312 and 1280 cm^-1. This indicates that in the burn sample the collagen α-helical nature is reducing and changing to alternative secondary structures. The peak at 1312 cm-1 may be attributed to β-sheets or α-helices, whilst 1286 cm^-1 represents random coil secondary structure formation. (C) Infrared spectral analyses; (i) absorption spectrum, (ii) second derivative, (iii) loadings plot and (iv) PCA plot of a human diabetic foot ulcer patient wound healing over time and (D) corresponding synchrotron-FTIR chemical map. The PC-1 loadings plot shows separation (48%) is attributed to CH stretching in the lipid range (2916 and 2847 cm-1), asymmetric bending of methyl groups (1450 cm^-1), collagen deposition (1338 cm^-1), and PO2- asymmetric stretching (1238 cm^-1). Amides, type I collagen at 1338 cm^-1, and type III collagen (unordered) are both predominant.

Conclusions

Collectively, we show from histological staining and chemical mapping using FPA and synchrotron-FTIR, patient diabetic foot ulcer skin and burn skin are in the initial phases of the wound healing process. Inflammation and collagen deposition found in the later skin sections suggest scaffold formation and remodeling of the skin. Overall, tissue analysis using Synchrotron-FTIR and FPA-based FTIR, highlight the potential use of these technologies as diagnostic tools in wound care.