Quantitative assessment of morphology and sub-cellular changes in macrophages and trophoblasts during inflammation

Biomed Opt Express. 2020 Jun 12;11(7):3733-3752. doi: 10.1364/BOE.389350. eCollection 2020 Jul 1.

Abstract

In pregnancy during an inflammatory condition, macrophages present at the feto-maternal junction release an increased amount of nitric oxide (NO) and pro-inflammatory cytokines such as TNF-α and INF-γ, which can disturb the trophoblast functions and pregnancy outcome. Measurement of the cellular and sub-cellular morphological modifications associated with inflammatory responses are important in order to quantify the extent of trophoblast dysfunction for clinical implication. With this motivation, we investigated morphological, cellular and sub-cellular changes in externally inflamed RAW264.7 (macrophage) and HTR-8/SVneo (trophoblast) using structured illumination microscopy (SIM) and quantitative phase microscopy (QPM). We monitored the production of NO, changes in cell membrane and mitochondrial structure of macrophages and trophoblasts when exposed to different concentrations of pro-inflammatory agents (LPS and TNF-α). In vitro NO production by LPS-induced macrophages increased 22-fold as compared to controls, whereas no significant NO production was seen after the TNF-α challenge. Under similar conditions as with macrophages, trophoblasts did not produce NO following either LPS or the TNF-α challenge. Super-resolution SIM imaging showed changes in the morphology of mitochondria and the plasma membrane in macrophages following the LPS challenge and in trophoblasts following the TNF-α challenge. Label-free QPM showed a decrease in the optical thickness of the LPS-challenged macrophages while TNF-α having no effect. The vice-versa is observed for the trophoblasts. We further exploited machine learning approaches on a QPM dataset to detect and to classify the inflammation with an accuracy of 99.9% for LPS-challenged macrophages and 98.3% for TNF-α-challenged trophoblasts. We believe that the multi-modal advanced microscopy methodologies coupled with machine learning approach could be a potential way for early detection of inflammation.