Cancer Technologies
Marlee Pincus
Student
Cornell University
Yorktown Heights, New York, United States
Obesity is linked to the development of thirteen different types of cancer, including breast cancer. In obese individuals, adipose tissue is characterized by hypertrophic expansion of adipocytes, indicated by increased cell size due to lipid accumulation. In normal physiology, adipose tissue acts as an energy storing organ which packages and releases lipids in response to endocrine signals. In breast cancer, tumor cells interact with adipocyte-rich breast tissue as tumors invade their surroundings. Cancer-associated adipocytes (CAA) are known to experience metabolic dysfunction and secrete greater amounts of chemokines and lipids which promote tumor cell invasion. For women with pre-invasive breast cancer, adipocyte hypertrophy correlates with progression to invasive disease. Yet, how hypertrophy alters adipocyte phenotypes and whether these changes contribute to this correlation are unknown.
While the biochemical interactions between adipocytes and breast cancer cells are well studied, the impact of adipocyte mechanics on tumor invasion remains unclear. Preliminary results from our lab reveal drastic remodeling of the actin cytoskeleton as cells transition from preadipocytes to mature adipocytes to hypertrophic adipocytes. Given that altered tissue mechanics drive tumor invasion, understanding the changes in actin organization and cell mechanics throughout differentiation will reveal if hypertrophy-associated changes impact invasion. To examine this, 3T3-L1 preadipocytes were differentiated under standard adipogenic conditions and in the presence of sodium palmitate to mimic elevated fatty acids in obesity. The degree of lipid accumulation and actin polymerization were then analyzed at different time points over the course of differentiation.
For differentiation experiments, 3T3-L1 preadipocytes were seeded at a density of 5,000 cells/cm2 on coverslips coated with 50 μg/mL of rat tail collagen. Adipogenic differentiation was induced two days after seeding by exchanging standard DMEM (5% FBS, 1% P/S) for media supplemented with 0.5 mM IBMX, 100 µm indomethacin, 1 µm dexamethasone, and 1 µm insulin. After two more days, differentiated coverslips were switched to either maintenance media (DMEM + 1 µm insulin) or maintenance media supplemented with 200 µm palmitic acid. Media changes were then performed every two days for twelve days. All control coverslips were kept in standard DMEM for the duration of the experiment. At days five, nine, and twelve, coverslips from each condition were fixed with 4% paraformaldehyde and stained with DAPI (DNA), phalloidin (F-actin), and Nile Red (neutral lipid). Confocal z-stacks were taken using a 63x objective on a Zeiss 880 laser scanning microscope. All images were analyzed using FIJI. For analysis of the actin cytoskeleton, the “Analyze Skeleton” plugin was used to quantify the branching features of binary images. For analysis of lipid accumulation, thresholding and the “Analyze Particles” feature were used to quantify average lipid droplet area per cell in each of the conditions. Similar image analysis was repeated on images of primary adipocytes of different sizes isolated from wildtype (WT) or genetically obese (ob/ob) mice for comparison.
Our results confirm that significant lipid accumulation occurs during the first two weeks of adipogenic differentiation of 3T3-L1 cells. In both standard adipogenic (induced) and fatty acid supplemented (FA) conditions, the average number of lipid droplets per cell, the size of individual droplets as well as total area covered by lipid per cell increased (Fig. 1A). Notably, the size of individual droplets and the total area covered were greater in FA compared to induced conditions. This indicates that in obesity-mimicking conditions, adipogenically induced 3T3-L1s experience greater lipid uptake similar to hypertrophic adipocytes in vivo.
Image analysis of actin fibers revealed that the average actin fiber length increases in induced conditions whereas the number of branches per cell decreases (Figure. 1B). On the other hand, control (non-induced) 3T3-L1s exhibit the opposite trend. This suggests that as adipogenic differentiation occurs, the actin cytoskeleton becomes less branched with fewer cross links, yet individual fibers become longer. No significant difference between the induced and FA conditions in terms of actin morphology were observed. However, for FA conditions, the average length and number of branches remained relatively constant between days 5 and 9 before increasing on day 12. Further studies with more replicates will be required to make definitive conclusions. Of note, these results contrast findings in primary adipocytes which suggest increased actin polymerization with lipid accumulation (Fig. 1D). This discrepancy may be due to the relatively early stage of adipogenesis achieved after 12 days of differentiation and longer term studies will be needed moving forward.
Overall, our data suggests that differentiation of 3T3-L1 preadipocytes induces lipid accumulation and remodeling of the actin cytoskeleton. While in vitro differentiation of 3T3-L1s reduces actin polymerization early during adipogenesis, hypertrophic expansion of primary adipocytes is associated with increased actin branching. How this initial decrease and subsequent increase in actin polymerization with lipid accumulation affect adipocyte mechanics and behavior is unknown. Ultimately, understanding how the biophysical properties of adipocytes evolve over differentiation and hypertrophic expansion will reveal if these changes contribute to the pathogenesis of obesity-associated cancers and metabolic disease.