Cell-cell-neighborhood relations in tissue sections--a quantitative model for tissue cytometry

Abstract

Physical interactions between different cell types are a requirement for the initiation and maintenance of immune responses. The distribution pattern of cells within a tissue may result from specific cell-cell-interactions or random distribution. Tissue architecture, degree of inflammation, frequencies of cells, number of contact partners, cell type, and size as well as cell movement and contact time determine the distribution of cells within tissues. We developed a matrix model to determine the degree of expected random distribution of two cell types (A and B) and cell-cell-contacts within tissue sections. The model predictions were compared with experimental data derived from immunofluorescence microscopy. We implemented a computer algorithm for automatic image analysis to visualize and quantify cell-cell-neighborhood relations. Using the number of cells type A (a), the total cell number (t) and the mean number of cells that are in contact with cells type B (c(B)), the ratio of cells type B in contact with cells type A can be described by b(A)/b = 1- (1- (a/t))[symbol: see text]c(B). We applied the model system to investigate the distribution of Foxp3(+) regulatory T cells with Ki-67(+) proliferating cells within mouse tissue sections. The matrix model provides a tool to describe the expected distribution of two different cell types and their cell-cell-contacts within tissues. Comparing the degree of expected random distribution with experimental data might help to propose functional cell-cell-interactions in tissue sections.