Raman microspectroscopy reveals unsaturation heterogeneity at the lipid droplet level and validates an in vitro model of bone marrow adipocyte subtypes

Bone marrow adipocytes (BMAds) constitute the most abundant stromal component of adult human bone marrow. Two subtypes of BMAds have been described, the more labile regulated adipocytes (rBMAds) and the more stable constitutive adipocytes (cBMAds), which develop earlier in life and are more resilient to environmental and metabolic disruptions. In vivo, rBMAds are enriched in saturated fatty acids, contain smaller lipid droplets (LDs) and more readily provide hematopoietic support than their cBMAd counterparts. Mouse models have been used for BMAds research, but isolation of primary BMAds presents many challenges, and thus in vitro models remain the current standard to study nuances of adipocyte differentiation. No in vitro model has yet been described for the study of rBMAds/cBMAds. Here, we present an in vitro model of BM adipogenesis with differential rBMAd and cBMAd-like characteristics. We used OP9 BM stromal cells derived from a (C57BL/6xC3H)F2-op/op mouse, which have been extensively characterized as feeder layer for hematopoiesis research. We observed similar canonical adipogenesis transcriptional signatures for spontaneously-differentiated (sOP9) and induced (iOP9) cultures, while fatty acid composition and desaturase expression of Scd1 and Fads2 differed at the population level. To resolve differences at the single adipocyte level we tested Raman microspectroscopy and show it constitutes a high-resolution method for studying adipogenesis in vitro in a label-free manner, with resolution to individual LDs. We found sOP9 adipocytes have lower unsaturation ratios, smaller LDs and higher hematopoietic support than iOP9 adipocytes, thus functionally resembling rBMAds, while iOP9 more closely resembled cBMAds. Validation in human primary samples confirmed a higher unsaturation ratio for lipids extracted from stable cBMAd-rich sites (femoral head upon hip-replacement surgery) versus labile rBMAds (iliac crest after chemotherapy). As a result, the 16:1/16:0 fatty acid unsaturation ratio, which was already shown to discriminate BMAd subtypes in rabbit and rat marrow, was validated to discriminate cBMAds from rBMAd in both the OP9 model in vitro system and in human samples. We expect our model will be useful for cBMAd and rBMAd studies, particularly where isolation of primary BMAds is a limiting step.