Detection of melanin-mediated false-positive for BaxΔ2 immunohistochemical staining in human skin tissues

Abstract

BaxΔ2 is a pro-apoptotic isoform of the Bax family. We have previously shown that BaxΔ2 protein expression is low in most healthy human tissues with the protein mainly scattered in the connective tissues. Surprisingly, in skin tissue, the BaxΔ2-positive staining was strikingly strong, especially in the basal layer of the epidermis. It has been documented that melanin in the basal cells displays a brown color that could give false-positive staining in the commonly used DAB-based (3-3-diaminobenzidine) immunostaining. To avoid the false-positive staining from DAB-melanin, we performed immunofluorescent staining on the same set of tissues which were positive for BaxΔ2 in the DAB immunostaining. The results from co-immunofluorescent staining of BaxΔ2 and a basal cell marker CK-14 showed that the previously detected DAB-stained BaxΔ2-positive epidermal cells were CK-14 positive but BaxΔ2 negative. Consistent with our previous finding, the BaxΔ2-positive cells in the dermis connective tissues are positive in both DAB-based and fluorescence-based immunostainings. These data indicate that the DAB-stained BaxΔ2 positive cells in the basal layer of skin tissue are false-positive due to a misinterpretation of the melanin color, but the dermal connective tissue still presents real BaxΔ2 positive staining.

Combined colour deconvolution and artificial intelligence approach for region‐selective immunohistochemical labelling quantification: The example of alpha smooth muscle actin in mouse kidney

“…With the advent of whole slide images (WSI) and the increase in computational power available to investigators, digital image analysis on a WSI is possible and within reach [10]. However, even when digital image analysis based on colour identification is applied to immunohistochemically stained slides, two main challenges remain: Firstly, the presence of false positive areas arising from possible background staining or similarly coloured structures (e.g., melanin, hemosiderin, bile pigments) necessitates manual removal to avoid overestimation of protein expression [11,12]. Secondly, in scenarios where the protein expression of interest is confined to a specific tissue compartment or region of interest (ROI), analysing the entire tissue as a whole, introduces bias, particularly when other regions express the protein at baseline levels under normal conditions.…”