Quantitative histological image analyses of reticulin fibers in a myelofibrotic mouse.

Abstract

Bone marrow (BM) reticulin fibrosis (RF), revealed by silver staining of tissue sections, is associated with myeloproliferative neoplasms, while tools for quantitative assessment of reticulin deposition throughout a femur BM are still in need. Here, we present such a method, allowing via analysis of hundreds of composite images to identify a patchy nature of RF throughout the BM during disease progression in a mouse model of myelofibrosis. To this end, initial conversion of silver stained BM color images into binary images identified two limitations: variable color, owing to polychromatic staining of reticulin fibers, and variable background in different sections of the same batch, limiting application of the color deconvolution method, and use of constant threshold, respectively. By blind coding image identities, to allow for threshold input (still within a narrow range), and using shape filtering to further eliminate background we were able to quantitate RF in myelofibrotic Gata-1low (experimental) and wild type (control) mice as a function of animal age. Color images spanning the whole femur BM were batch-analyzed using ImageJ software, aided by our two newly added macros. The results show heterogeneous RF density in different areas of the marrow of Gata-1low mice, with degrees of heterogeneity reduced upon aging. This method can be applied uniformly across laboratories in studies assessing RF remodeling induced by aging or other conditions in animal models.