Refeeding-induced brown adipose tissue glycogen hyper-accumulation in mice is mediated by insulin and catecholamines

Abstract

Brown adipose tissue (BAT) generates heat during adaptive thermogenesis through a combination of oxidative metabolism and uncoupling protein 1-mediated electron transport chain uncoupling, using both free-fatty acids and glucose as substrate. Previous rat-based work in 1942 showed that prolonged partial fasting followed by refeeding led to a dramatic, transient increase in glycogen stores in multiple fat depots. In the present study, the protocol was replicated in male CD1 mice, resulting in a 2000-fold increase in interscapular BAT (IBAT) glycogen levels within 4-12 hours (hr) of refeeding, with IBAT glycogen stores reaching levels comparable to fed liver glycogen. Lesser effects occurred in white adipose tissues (WAT). Over the next 36 hr, glycogen levels dissipated and histological analysis revealed an over-accumulation of lipid droplets, suggesting a potential metabolic connection between glycogenolysis and lipid synthesis. 24 hr of total starvation followed by refeeding induced a robust and consistent glycogen over-accumulation similar in magnitude and time course to the prolonged partial fast. Experimentation demonstrated that hyperglycemia was not sufficient to drive glycogen accumulation in IBAT, but that elevated circulating insulin was sufficient. Additionally, pharmacological inhibition of catecholamine production reduced refeeding-induced IBAT glycogen storage, providing evidence of a contribution from the central nervous system. These findings highlight IBAT as a tissue that integrates both canonically-anabolic and catabolic stimulation for the promotion of glycogen storage during recovery from caloric deficit. The preservation of this robust response through many generations of animals not subjected to food deprivation suggests that the over-accumulation phenomenon plays a critical role in IBAT physiology.

Figure 1. Fed-mouse tissue glycogen concentrations.

Group-housed, fed, male, wild-type, CD1 mice were sacrificed at 19∶30, 23∶30, or 07∶30 and tissues were harvested and snap-frozen in liquid nitrogen. Tissue glycogen concentrations were measured in epididymal adipose tissue (EPI) (A), interscapular brown adipose tissue (IBAT) (B), and liver (C). Error bars are ±SEM. Statistical comparisons for each tissue were made between time points using a 2-tailed Student’s t-test. ^#, p<0.001. 10–30 mice were used for each time point.

Figure 2. Tissue glycogen during 5 days of partial fasting and refeeding.

Male, wild-type, CD1 mice were fed 60% of their normal daily intake of chow once daily for 5 days and then sacrificed (0) or allowed to refeed ad libitum 4–48 hr before sacrifice (4–48). Upon sacrifice, tissues were harvested and snap-frozen in liquid nitrogen. Tissue glycogen was measured in EPI (A), mesenteric adipose tissue (B), perirenal adipose tissue (C), subcutaneous adipose tissue (D), IBAT (E), and the liver (F). Error bars are ±SEM. Statistical comparisons were made between the partial-fasting time point (0) and each refeeding time point (4–48) for each tissue using a 2-tailed Student’s t-test. *, p<0.05; **, p<0.01; ^#, p<0.001; ^&, p<0.0001. Fasted time points utilized 2 mice. All other time points were obtained from 4–8 mice.

Figure 3. Histological visualization of glycogen storage in IBAT (top row) and liver (bottom row) using Periodic Acid Schiff staining.

Male, wild-type, CD1 mice were either fed ad libitum (A, E), fed 60% of their normal daily intake of chow once daily for 5 days and then sacrificed (B, F), or fed 60% of their normal daily intake of chow once daily for 5 days and then allowed to refeed ad libitum 4 hr (C, G) or 48 hr (D, H) before sacrifice. IBAT (A–D) and liver (E–H) were harvested and immediately placed in formalin fixative solution. Fixed samples were mounted in paraffin, sectioned, and stained using Periodic Acid Schiff staining with Hematoxylin and Eosin counter-staining by the University of Chicago Human Tissue Resources Center.

Figure 4. Tissue glycogen during refeeding after total starvation.

Male, wild-type, CD1 mice were starved for either 72 hr (A–C) or 24 hr (D,E) and then sacrificed (0) or allowed to refeed ad libitum 1–48 hr before sacrifice (1–48). Tissues were harvested and snap-frozen in liquid nitrogen. Tissue glycogen was measured in EPI (A), IBAT (B, D), and liver (C, E). Error bars are ±SEM. Statistical comparisons were made between the starved time point (0) and each refeeding time point (1–48) for each tissue using a 2-tailed Student’s t-test. *, p<0.05; **, p<0.01; ^#, p<0.001; ^&, p<0.0001. Fasting time points utilized 3–5 mice, except 24 hr fasting BAT glycogen which utilized 19 mice. All refed measurements were taken from 3–9 mice except 24 hr fasted/refed 2 hr BAT time point, which was measured in 15 mice.

Figure 5. IBAT glycogen and serum parameters after 24 hr starvation and a hyperinsulinemic, hyperglycemic clamp.

Male, wild-type, CD1 mice were starved for 24 hr and then subjected to a 2 hr hyperinsulinemic, hyperglycemic clamp (24H-SC) in which glucose was clamped at 300 mg/dL and insulin was infused at a rate of 3 mU/kg/min. At the end of the clamp, mice were sacrificed, serum was collected, and tissues were collected and snap-frozen in liquid nitrogen. Final serum glucose (A), serum insulin (B), and IBAT glycogen (C) were compared to those of fed mice sacrificed at 19∶30 (Fed) and 24 hr starved/refed mice sacrificed 2 hr after the onset of refeeding (24H-SR). Error bars are ±SEM. Statistical comparisons were made between experimental conditions using a 2-tailed Student’s t-test. **, p<0.01; ^#, p<0.001; ^&, p<0.0001. All measurements represent 8–9 mice, except Fed 19∶30 and 24H-SR IBAT glycogen, obtained from 30 and 15 mice, respectively.

Figure 6. Effects of catecholamine-synthesis inhibition on tissue glycogen and serum parameters during starvation and refeeding.

Male, wild-type CD1 mice were used for these experiments. 24H-SR mice were starved for 24 hr and either sacrificed (0) or allowed to refeed ad libitum 2–4 hours and then sacrificed (2, 4). 24H-SAR mice were starved for 24 hr, but also received an IP injection of 300 mg/kg AMPT in 0.9% NaCl 2 hr prior to refeeding. 24H-SAR mice were either sacrificed following the full 24 hr starvation (0), or allowed to refeed 2 or 4 hr prior to sacrifice (2, 4 respectively). Upon sacrifice, tissues were snap-frozen in liquid nitrogen and serum was collected. Glycogen was measured in IBAT (A) and liver (B), and circulating glucose (C) and insulin (D) were also measured. Error bars are ±SEM. Statistical comparisons were made between experimental conditions using a 2-tailed Student’s t-test. ^#, p<0.001. Measurements for each time point from AMPT-injected mice were obtained for 7–12 mice.

Figure 7. Mechanism by which glycogen stores may enhance free fatty acid esterification in adipose tissues during refeeding.

During the first several hours of refeeding, hyperaccumulation of intracellular BAT glycogen occurs. The eventual glycogenolytic and subsequent glycolytic processing of glycogen stores may generate elevated levels of glycerol-3-phosphate, a precursor to monoglyceride formation. An increase in this precursor may enhance the rate of esterification of free fatty acids (FFA), both from de novo synthesized sources as well as FFA recently liberated from stored neutral lipids. This process may greatly enhance lipid repletion during refeeding in order to maximize recovery from the starvation state.