Improving Visualization of cAMP Gradients Using Algorithmic Modelling.

A ubiquitous second messenger molecule, cAMP is responsible for orchestrating many different cellular functions through a variety of pathways. Fӧrster resonance energy transfer (FRET) probes have been used to visualize cAMP spatial gradients in pulmonary microvascular endothelial cells (PMVECs). However, FRET probes have inherently low signal-to-noise ratios; multiple sources of noise can obscure accurate visualization of cAMP gradients using a hyperspectral imaging system. FRET probes have also been used to measure cAMP gradients in 3D; however, it can be difficult to differentiate between true FRET signals and noise. To further understand the effects of noise on experimental data, a model was developed to simulate cAMP gradients under experimental conditions. The model uses a theoretical cAMP heatmap generated using finite element analysis. This heatmap was converted to simulate the FRET probe signal that would be detected experimentally with a hyperspectral imaging system. The signal was mapped onto an image of unlabeled PMVECs. The result was a time lapse model of cAMP gradients obscured by autofluorescence, as visualized with FRET probes. Additionally, the model allowed the simulated expression level of FRET signal to be varied. This allowed accurate attribution of signal to FRET and autofluorescence. Comparing experimental data to the model results at different levels of FRET efficiency has allowed improved understanding of FRET signal specificity and how autofluorescence interferes with FRET signal detection. In conclusion, this model can more accurately determine cAMP gradients in PMVECs. This work was supported by NIH award P01HL066299, R01HL58506 and NSF award 1725937.