Comparative analysis of colorimetric staining in skin using open-source software

Abstract

Colorimetric staining techniques such as immunohistochemistry (IHC), immunofluorescence (IF) and histochemistry (HC) provide useful information regarding the localization and relative amount of a molecule/substance in skin. We have developed a novel, straightforward method to assess colorimetric staining by combining features from two open-source software programs. As a proof of principle, we demonstrate the utility of this approach by analyzing changes in skin melanin deposition during the radiation-induced tanning response of Yucatan mini-pigs. This method includes a visualization step to validate the accuracy of color selection before quantitation to ensure accuracy. The data show that this method is robust and will provide a means to obtain accurate comparative analyses of staining in IHC/IF/HC samples.

Background

Immunohistochemical (IHC), immunofluorescence (IF) and histochemical (HC) staining techniques are commonly used to determine the location and relative levels of a molecule or substance in skin. This information is critically important for understanding development and the mechanisms underlying pathologic conditions. Current techniques for assessing colorimetric staining can be subjective, using arbitrary visual grading scales applied by an experienced histopathologist, while other methods require complicated algorithms and are cumbersome ^1–3.

This report describes a new methodology for comparatively analyzing Fontana-Masson stained skin sections using open source software. This technique was developed to study the impact of ionizing radiation on skin in doses that could be encountered by astronauts during extended space travel ⁴. As a model system, Yucatan mini-pigs were exposed to varying doses of 6 MeV electrons to mimick the dose distribution of solar particle event (SPE) proton radiation ⁵. This ionizing radiation exposure induced prominent epidermal hyperpigmentation associated with dermal pigment incontinence. The pigmentary changes were analyzed by using GIMP software to specifically extract the pixels associated with pigment staining from photomicrographs of Fontana-Masson stained skin sections; and quantitating the number of extracted pixels using Image J. Our results demonstrate this technique can precisely determine changes in skin pigmentation, secondary to ionizing radiation. This approach should have wide applicability and be readily adaptable to any project comparing the magnitude of IHC/IF/HC signals in different cell populations (Fig. S1).

Experimental Design

Animals

Yucatan mini pigs were obtained from Sinclair BioResources (Columbia, MO) and were irradiated with 6 MeV electrons (see supplemental information) ^5–9.

Tissue Biopsies, Processing and Analysis

Skin punch biopsy samples were obtained prior to irradiation and at 7-, 14- and 30-days following radiation exposure. Punch biopsies (4 mm) were fixed in 10% buffered formalin and embedded in paraffin. Paraffin sections were stained with Hematoxylin/Eosin (H+E) and Fontana-Masson for visualization of melanin.

Melanin quantitation

Melanin was quantitated using two open source optical imaging software programs. Photomicrographs (jpeg format) of Fontana-Masson and H+E stained tissue sections were opened in GIMP- GNU Image Manipulation Program (www.gimp.org). The brown-black color signals corresponding to melanin grains on stained sections were selected, copied and pasted into a new image and saved as a jpeg file; this jpeg file consists solely of black/brown (melanin grains) on a white background. This image was subsequently opened using the ImageJ program (Link: rsbweb.nih.gov/ij). A histogram of the image was created which separates the total number of pixels in the image into 255 color categories spanning the visible spectrum. The peak corresponding to the brown-black color (melanin) was determined by cutting and summing the appropriate counts from each channel of the melanin peak. Alternatively, the numbers corresponding to the pigment peak could be pasted into an Excel spreadsheet and summed.

The pigmentation index was obtained by taking the absolute number of black pixels in a representative 20X field as determined by Image J and multiplying it by 10⁻³.

Results and Discussion

The porcine back epidermis and dermis manifest a thickness and structure similar to human skin, and is an accepted in vivo model for radiation studies. The baseline skin color of normal Yucatan mini pigs is whitish-grey (Fig S2a), and demonstrates mild pigmentation in basilar keratinocytes (Fig 1 and Fig S2C). The un-irradiated dermis exhibits spindled cells and a sparse mononuclear inflammatory infiltrate without pigment deposition (Fig S2C); keratinocyte necrosis was not identified (Sup Fig 2C).

Figure 1. Post-radiation Melanin Quantitation.

Photomicrographs of Fontana-Masson stained biopsies from a representative pig pre-irradiation (A) and at 7, 14 and 30 days following 25Gy of 6 MeV electrons radiation (C, E, G). Corresponding melanin pixel image derived from Fontana-Masson sections that will be integrated to determine melanin levels (B, D, F, H).

Animals were irradiated with 0–25 Gy of 6 MeV electrons ⁵. The animals manifested a tanning response detectable by 14 days post-irradiation (supplemental Fig 2B). Skin biopsy samples were obtained from animals pre-irradiation and at 7, 14 and 30 days following radiation exposure. Visual assessment of H+E and Fontana- Masson slides demonstrated increased pigmentation at 7 days, a peak at 14 days and decreased pigmentation at 30 days post-irradiation (Fig 1 and S3). A similar time course of pigmentation was observed using different doses (0–25 Gy) of radiation (data not shown).

The melanin content of tissue sections was assessed using the GIMP and ImageJ software programs and (see Experimental Design and supplemental information). The time course of melanin accumulation in pig skin following radiation treatment was determined (Fig 2A).

Figure 2. A. Comparative melanin levels as a function of time post radiation exposure.

The amount of melanin present in each section (See Fig 1; A, C, E, G) was determined using the reported method. Note maximal melanin deposition occurs 14 days following radiation exposure; results presented are mean ± SD. N= 3

Statistical analysis: Student’s t test: Pre-irradiation vs 7 days, p< 0.001; Pre-irradiation vs 14 days, p< 0.001; Pre-irradiation vs 30 days, p< 0.001.

B. Melanin deposition is proportional to radiation dose. Animals were irradiated with 0–25 Gy of 6 MeV electrons (3 animals/treatment group) and 14 days later skin biopsies were obtained and processed for melanin quantitation. Note that radiation increases melanin deposition in a dose-dependent manner; results presented are mean ± SD. Statistical analysis: Student’s t test:: Pre-irradiation vs 5 Gy, p= 0.0035; Pre-irradiation vs 10 Gy, p< 0.01; Pre-irradiation vs 15 Gy, p< 0.0001; Pre-irradiation vs 25 Gy, p< 0.001.

To determine the effect of radiation dose on pigmentation, animals received total body electron radiation (0–25 Gy) and skin biopsies were obtained pre- and at 7, 14 and 30 days following radiation exposure. Sections from each biopsy were stained with Fontana-Masson and the amount of melanin present in 14 day sections (peak melanin deposition) was determined. Increased melanin deposition correlated with radiation dose, yielding a linear dose-response curve (see Fig 2B). We also performed parallel analysis on sections stained with H+E from the same tissue blocks used above with similar results (Figs S3 and S4).

We have described a new method to compare colorimetric signals in sections of skin subjected to Fontana-Masson staining. The advantage of using Fontana-Masson is that it is easier to extract the black grains corresponding to melanin pigment and visually confirm if the color channel(s) selected represent all the melanin present in the tissue section. The described methodology can readily determine radiation-induced changes in skin pigmentation as a function of radiation dose and time post-radiation. The changes in skin pigmentation induced by ionizing radiation were similar regardless of whether the assessment was made on Fontana-Masson or H+E stained sections This technique does not require sophisticated equipment, and it can be used to determine relative changes in skin pigmentation secondary to wounding and human pigmentary disorders such as vitiligo or melasma, where non-involved skin would serve as the control ^{10, 11}. By using the color extraction capabilities of GIMP to create a separate JPEG representing the colorimetric signals to be analyzed, this method has an inherent visual quality control step that promotes accurate quantitation. This method should be applicable to analyzing any colorimetric signal present in photomicrographs obtained from IHC/IF/HC stained tissue sections.