Falsifying computational models of angiogenesis through quantitative comparison with in vitro models

During angiogenesis, endothelial cells migrate from existing vessels, proliferate and collectively organize into new capillaries. In vitro and in vivo experimentation is instrumental for identifying the molecular players and cell behavior that regulate angiogenesis. Alongside experimental work, computational and mathematical models of angiogenesis have helped to show if the current molecular and cellular understanding of cell behavior is sufficient. As input, the model takes (a subset of) the current knowledge or hypotheses of single cell behavior and captures it into a dynamical, mathematical description. As output, it predicts the multicellular behavior following from the actions of many individual cells, e.g., the formation of a sprout or the formation of a vascular network. Paradoxically, computational modeling based on different assumptions, i.e., completely different, sometimes non-intersecting sets of observed single cell behavior, can reproduce the same angiogenesis-like multicellular behavior, making it practically impossible to decide which, if any, of these models is correct. Here we present dynamic analyses of time-lapses of in vitro angiogenesis experiments and compare these with dynamic analyses of mathematical models of angiogenesis. We extract a variety of dynamical characteristics of endothelial cell network formation using a custom time-lapse video analysis pipeline in ImageJ. We compare the dynamical network characteristics of the in vitro experiments to those of the cellular networks produced by computational models. We test the response of the in silico dynamic cell network characteristics to key model parameters and make related changes in the composition of the in vitro environment. We present comparisons with computational model outcomes and argue how models that fail to reproduce these trends can be rejected. All in all, we show how our dynamic approach helps to clarify key endothelial cell interactions required for angiogenesis, and how the approach helps analyze what key changes in network properties can be traced back to changes in individual cell behavior.