Caveolin-1 expression and cavin stability regulate caveolae dynamics in adipocyte lipid store fluctuation

Abstract

Adipocytes specialized in the storage of energy as fat are among the most caveolae-enriched cell types. Loss of caveolae produces lipodystrophic diabetes in humans, which cannot be reversed by endothelial rescue of caveolin expression in mice, indicating major importance of adipocyte caveolae. However, how caveolae participate in fat cell functions is poorly understood. We investigated dynamic conditions of lipid store fluctuations and demonstrate reciprocal regulation of caveolae density and fat cell lipid droplet storage. We identified caveolin-1 expression as a crucial step in adipose cell lines and in mice to raise the density of caveolae, to increase adipocyte ability to accommodate larger lipid droplets, and to promote cell expansion by increased glucose utilization. In human subjects enrolled in a trial of 8 weeks of overfeeding to promote fattening, adipocyte expansion response correlated with initial caveolin-1 expression. Conversely, lipid mobilization in cultured adipocytes to induce lipid droplet shrinkage led to biphasic response of cavin-1 with ultimate loss of expression of cavin-1 and -3 and EHD2 by protein degradation, coincident with caveolae disassembly. We have identified the key steps in cavin/caveolin interplay regulating adipocyte caveolae dynamics. Our data establish that caveolae participate in a unique cell response connected to lipid store fluctuation, suggesting lipid-induced mechanotension in adipocytes.

Figure 1

Caveolin-1 overexpression in 3T3-L1 increases adipocyte caveolae density. A: Cav1-RFP transgene expression in cells transduced with retroviral constructs containing empty vector (pBabe) or a caveolin-1 cDNA fused to RFP. Membranes were immunoblotted (IB) with an anti-RFP antibody or β-actin. B: Endogenous caveolin-1 and exogenous Cav1-RFP distribution into detergent-resistant membrane fractions. Cells stably expressing an empty pBabe vector or Cav1-RFP were lysed in the presence of cold Triton X-100. Detergent-resistant (fraction 1–5) or detergent-soluble (fraction 7–12) fractions were obtained after gradient centrifugation. C: Fluorescent imaging of Cav1-RFP cell lines by confocal microscopy. Bar scale is 20 μm. D and F: Relative mRNA expression in Cav1-RFP versus control cell lines. Indicated mRNA levels were measured by RT-QPCR and normalized to 18S or 36B4 mRNA. Values are means ± SEM obtained in three independent pools of antibiotic-selected clones. E: Relative protein expression in Cav1-RFP versus control cell lines. Indicated proteins were assessed by Western blotting and normalized to β-actin. According to cavin nomenclature, PTRF is cavin-1, SRBC is cavin-2, and SDPR is cavin-3. G: Electron microscopy images of 3T3-L1 adipocytes transduced with an empty vector (upper panel) stably expressing Cav1-RFP (middle panel) or adipocytes of subcutaneous adipose tissue of mice (lower panel). Bar scale: 300 nm. H: Quantification of caveolae density from electron microscopy images. A total of 40-μm membrane stretches were used for caveolae quantification in each group using ImageJ software. Caveolae density is expressed as the number of invaginated caveolae per micrometer membrane length, and values are means ± SEM of 6–10 image sections. Significant differences between groups by Student t test are indicated as follows: ***P < 0.001, *P < 0.05.

Figure 2

Caveolin-1 overexpression increases adipocyte basal metabolic activity. A: Expression of insulin receptor (Ins Rec) and PKB mRNA in control or Cav1-RFP differentiated adipocytes. 36B4 mRNA was used for normalization. B: Protein expression of IRS-1, PKB, and Erk1 in Cav1-RFP adipocytes. β-Actin normalization was used in Western blots. C–E: Insulin-dependent phosphorylation of PKB (phospho-Ser⁴⁷³; Cell Signaling) or IRS-1 (phospho-Tyr⁶¹²; Upstate) and Erk1/2 (phospho-Thr²⁰²/Tyr²⁰⁴; Cell Signaling) was evaluated by Western blotting with phospho-specific antibodies, and signals were normalized to total protein (Tot). Incubation with indicated insulin concentrations was for 30 min, and total cell lysates were immediately frozen. Values are means ± SEM of four experiments in independent transfectants pools, expressed relative to values of maximally insulin-stimulated control cells. A log scale is used for dose-response curves. F: 2-Deoxyglucose (2-DG) transport was measured after overnight insulin deprivation. 2-DG (0.2 mmol/L) was added for 8 min, and reaction was stopped by addition of ice-cold buffer. Values are means ± SEM obtained in three independent experiments. G: Lactate production by control or Cav1-RFP adipocytes. Cells were exposed to insulin (100 nmol/L) for 6 h in fresh DMEM. Extracellular lactate concentration was measured using a commercial kit. *Significant differences by paired t test. prot, protein.

Figure 3

Caveolin-1 overexpression increases adipocyte lipid storage in cell culture and mice adipose tissue. A: Images of Oil Red O–stained cultured adipocytes transduced with an empty vector (pBabe) of stably expressing Cav1-RFP. B: Lipid droplet (LD) size distribution in control or Cav1-RFP stable 3T3-L1 clones. Lipid droplet size was measured on microscope images using ImageJ software. Mean distributions were calculated from values collected in four independent pools of stable 3T3-L1 clones and represent ~500 individual lipid droplets in each group. C: Untagged caveolin-1 overexpression in 3T3-L1 adipocytes by adenoviral vector encoding caveolin-1 (Ad Cav1) or GFP (Ad Null GFP). Two independent experiments are shown. D: Lipid droplet size distribution in adipocytes infected with Ad Cav1 or Ad Null GFP. E: 3T3-F442A preadipocytes (30 × 10⁶ cells) were injected subcutaneously in the dorsal (upper panel) or the ventral (lower panel) region of 6-week-old nude mice. After 2 months with ad libitum feeding, mice were given access to 30% sucrose in drinking water for 2 weeks and killed. Newly formed fat pads are shown. F and G: Histological sections were prepared from newly formed fat pads, and mean adipocyte area was determined with ImageJ software. Values are mean of fat cell surface of individual mice receiving matched injections of control preadipocytes in the dorsal and ventral regions. H: Mean adipocyte area in newly formed fat pads from mice injected twice with control (pBabe) and Cav1-RFP–expressing preadipocytes. Data collected on eight individual mice are shown. I: Mean values for relative adipocyte area in control (pBabe) versus Cav1-RFP–derived fat pads. Paired Student t test indicated significant changes (***P < 0.001). AU, arbitrary units.

Figure 4

Caveolin expression and adipose tissue expansion during overfeeding in healthy subjects. A: Clinical parameters of subjects participating in the 8-week overfeeding protocol are described in Research Design and Methods. The rational for group assignment is from the analysis of fat cell size change from preoverfeeding (baseline) to postoverfeeding, which defines hypertrophic (B) or hyperplastic (C) responses in individual patients. D: Spearman correlation of caveolin adipose tissue content to change in fat cell size in patients with hypertrophic response to overfeeding. Caveolin-1 content was assessed by Western blotting from total protein lysates (15 µg). β-Actin was used for normalization. AU, arbitrary units.

Figure 5

Lipid droplet shrinkage in 3T3-L1 adipocytes coincides with caveolae disassembly. A: Terminally differentiated 3T3-L1 adipocytes were cultured in DMEM supplemented with 10% FBS and treated with 1 mmol/L 8Br-cAMP or a combination of 10 µmol/L FSK and 500 µmol/L MIX. Rates of glycerol release to the extracellular medium (expressed per hour and per milligram cell protein) were measured for 2 h at indicated times over a 30-h period. Values are means ± SEM from four independent experiments. B: LD540 staining of neutral lipids in 3T3-L1 control adipocytes or cells treated for 30 h with 8Br-cAMP. Smaller lipid droplets after 8Br-cAMP denote effective lipid mobilization. Bar scale: 20 μm. C: Electron microscopy images of control or 8Br-cAMP–treated 3T3-L1 adipocytes. Note the presence of abundant caveolae on the surface of control cells only. Bar scale: 200 nm. D: Quantification of caveolae density from electron microscopy images. Caveolae invaginations were counted on linear membrane stretches representing a total of 30 μm membrane length in 6–10 images from different sections using ImageJ. Data represent means ± SEM (***P < 0.001). E: Caveolae quantification on electron microscopy adipose tissue sections of mice fasted for 48 h or fed (three mice per group). Caveolae number was compared on a per-cell basis (expressed as 100 caveolae/length of membrane perimeter). Data are means ± SEM. *P < 0.05.

Figure 6

Effect of prolonged lipid mobilization by 8Br-cAMP on caveolin-1. A: Time course of 8Br-cAMP exposure on caveolin-1 protein contents in 3T3-L1 adipocytes. β-Actin is used as a loading control for Western blot analysis. B and C: Quantitative analysis of caveolin-1 protein expression in control adipocytes or cells treated with 1 mmol/L 8Br-cAMP for 30 h (B) or at end point (C). Values are obtained by densitometric scanning of immunoblots and expressed relative to initial protein signals. Means ± SEM from 3–5 independent experiments are shown. *Significant difference by Student t test (P < 0.05). D: Adipocyte lysates were used to determine free cholesterol content. Mean values ± SEM were obtained from three independent experiments. E: Immunofluorescence labeling of endogenous caveolin-1 in differentiated 3T3-L1 adipocytes untreated (upper panel) or treated with 8Br-cAMP for 30 h (lower panel). Bar scale is 5 μm.

Figure 7

Prolonged lipid mobilization by 8Br-cAMP induces loss of cavins by protein degradation. A: Time course of 8Br-cAMP exposure on PTRF/cavin-1 and SDPR/cavin-2 protein contents in 3T3-L1 adipocytes. Fully differentiated adipocytes were treated or not with 1 mmol/L 8Br-cAMP as described above. β-Actin is used as a loading control. B–E: Quantitative changes in PTRF/cavin-1 (B), SDPR/cavin-2 (C), and EHD2 (D) protein expression with 8Br-cAMP treatment or at end point (30 h) with no effector (control), 1 mmol/L 8Br-cAMP, a combination of 10 µmol/L FSK and 500 µmol/L MIX, or 10 µmol/L FSK alone (E). Values are obtained by densitometric scanning of immunoblots and expressed relative to initial protein signals. Means ± SEM from at least three to five independent experiments are shown. *Significant difference by t test (P < 0.05). F–I: Fully differentiated 3T3-L1 adipocytes stably expressing PTRF-GFP (F), SRBC-GFP (G), or Cav1-RFP (H) were treated or not with 1 mmol/L 8Br-cAMP for 30 h. Lipids were stained using the neutral lipid probe LD540, and cells were analyzed by confocal microscopy (bar scale: 20 μm). Fluorescence intensity of confocal images (I) was evaluated using ImageJ software. Ten to 12 confocal images from different microscopic fields were acquired in the different cell lines. The sum of the fluorescence intensity was divided by microscopic field area. Data represent means ± SEM (*P < 0.05). AU, arbitrary units.