Adipogenesis, lipogenesis and lipolysis is stimulated by mild but not severe hypoxia in 3T3-L1 cells

Abstract

In-vitro investigation of the effects of hypoxia is limited by physical laws of gas diffusion and cellular O2 consumption, making prolonged exposures to stable O2 concentrations impossible. Using a gas- permeable cultureware, chronic effects of mild and severe hypoxia on triglyceride accumulation, lipid droplet size distribution, spontaneous lipolysis and gene expression of adipocyte-specific markers were assessed. 3T3-L1 cells were differentiated under 20%, 4% or 1% O2 using a gas-permeable cultureware. Triglyceride accumulation, expression of genes characteristic for advanced adipocyte differentiation and involvement of key lipogenesis enzymes were assessed after exposures. Lipogenesis increased by 375% under mild hypoxia, but dropped by 43% in severe hypoxia. Mild, but not severe, hypoxia increased formation of large lipid droplets 6.4 fold and strongly induced gene expression of adipocyte-specific markers. Spontaneous lipolysis increased by 488% in mild, but only by 135% in severe hypoxia. Inhibi- tion of ATP-dependent citrate lyase suppressed hypoxia-induced lipogenesis by 81% and 85%. Activation of HIF inhibited lipogenesis by 59%. Mild, but not severe, hypoxia stimulates lipolysis and promotes adipocyte differentiation, probably through excess of acetyl-CoA originating from tricarboxylic acid cycle independently of HIF activation.

1. Introduction

Excessive accumulation of adipose tissue is causally linked to the development of insulin resistance, Type 2 diabetes mellitus (T2DM) and increased cardiovascular as well as all cause mortality [1]. Although epidemiological evidence is convincing, mechanisms mediating the adverse metabolic effects of obesity remain only partially elucidated. Multiple factors have been suggested, including elevated release of free fatty acids from adipocytes, modified spectrum of adipose tissue secreted endocrine factors (adipokines) or induction of low-grade pro-inflammatory state.

However, substantial proportion of obese individuals do not develop metabolic impairments, suggesting that other factors probably also play an important role. It has been established recently, that adipose tissue hypoxia represents a significant factor influencing multiple adipocyte functions and might thus contribute to the development of meta- bolic impairments [2]. In vitro studies demonstrated the that hypoxia could modify gene expression, adipokine secretion and induce insulin resistance in adipocytes and regulate cell growth, viability, differentiation and substrate metabolism [2,3]. Adipose tissue O2 levels are reduced in obese rodents and humans by ~20e30 mmHg, probably as a consequence of increased adipocyte size and reduced tissue capillarity [4e6], and that such tissue hypoxia is associated with increased lipolysis and modified spec- trum of secreted adipokines [2]. Furthermore, it was observed that diseases characterized by decreased oxygen hemoglobin saturation and whole-body hypoxia, such as obstructive sleep apnea or chronic obstructive pulmonary disease, are associated with insulin resistance, glucose intolerance and represent a risk factor for T2DM [7].

Physical limitations of gas diffusion through culture media compromise current in vitro paradigms as pericellular O2 levels were shown to be dramatically different from the O2 concen- tration in the gas atmosphere inside the CO2 incubator [8e10]. Additionally, oxygen consumption of adherent cells leads to progressively decreasing pericellular O2 levels, ultimately leading to severe cellular hypoxia or anoxia even when cells are cultured in a CO2 incubator with 20% O2 in the gas phase [10,11]. As a result, published studies typically utilize acute, short-term hyp- oxic exposures ranging from several hours to 48 h [2,4,6], which does not adequately reflect the pathophysiological context of the development of obesity and T2DM, requiring prolonged, chronic exposures to precisely defined and constant pericellular O2 levels.

The aim of this study was to investigate the effects of mild (4% O2) and severe (1% O2) cellular hypoxia on adipocyte differentia- tion, triglyceride accumulation and lipolysis, representing key adipocyte functions. We also aimed address potential mechanisms mediating hypoxia-induced changes in metabolic pathways. To achieve these goals, we used an innovative approach based on a membrane-bottom cultureware enabling rapid exchange of gases and thus prolonged exposure of adherent cells to predictable and reproducible levels of pericellular O2 throughout the whole experiment [12,13].

2. Material and methods

2.1. 3T3L1 cells culture, differentiation and hypoxic exposure

Murine 3T3-L1 fibroblasts (Zen-Bio Inc., NC, USA) were cultured in a CO2 incubator at 37 ◦C and 5% CO2 on fluorocarbon-bottom culture plates (Prod. #94.6077.410, Sarstedt AG & Co, Nümbrecht, Germany). Cells were expanded at cell passage number 11, for 4e5 days, in T75 flasks, harvested and plated (5000 cells/cm2) in fluorocarbon-bottom plates. After reaching confluence, the culture medium was changed to differentiation medium (DM2-L1, Zen-Bio Inc., NC, USA) and dishes placed in modular incubators (Billups- Rothenberg Inc., Del Mar, CA, USA) flushed with calibration quality gas mixtures of 4% O2 + 5% CO2 or 1% O2 + 5% CO2 (Linde Gas a.s., Prague, Czech Republic) to achieve mild and severe hypoxic environments, respectively. Control exposures were performed in a standard CO2 incubator (20% O2 + 5% CO2). Cells were differentiated in modular chamber incubators for 14 days.

2.2. Spontaneous lipolysis determination

Differentiated cells were starved for 24 h in a serum-free me- dium (Zen-Bio Inc., NC, USA) and incubated in KRHBA buffer (Krebs Ringer Bicarbonate buffer containing 10 mmol/L HEPES, 2% fatty acid free bovine serum albumin and 6 mmol/L glucose at pH 7.4) for 1 h to recover and subsequently for 3 h to assess spontaneous lipolytic rate. Glycerol released into media was measured colori- metrically (Prod. No.: F6428, Sigma-Aldrich, St. Louis, MO, USA), and normalized to total lipid content as previously described [13].

2.3. Hypoxia inducible factor (HIF) activity assay and lipogenesis activators/inhibitors

HIF Luciferase Reporter NIH3T3 Stable Cell Line (Signosis Inc., Santa Clara, CA, USA) was used to evaluate changes in hypoxia- inducible factor (HIF) transcriptional activity after dimethylox- aloylglycine administration (DMOG, Prod.No.: D3695, Sigma-Aldrich, St. Louis, MO, USA) following manufacturer’s instructions. Transcriptional activity of HIF was normalized to protein amount measured with bicinchoninic acid assay (Thermo Fischer Scientific, Waltham, MA, USA).

To assess the influence of selected pathways regulating lipo- genesis, cells were incubated with the following chemicals added to the culture media for the whole duration of differen- tiation (14 days): a) pyruvate dehydrogenase activator – 5 mM dichloroacetate (Prod.No.: D54702, Sigma-Aldrich, St. Louis, MO, USA), b) ATP-dependent citrate lyase inhibitor: SB204990 (Prod.No.: 4962, Tocris Bioscience, Bristol, UK) and c) 1 mM and 2 mM dimethyloxalylglycine – stabilizing HIF transcriptional factor (DMOG, Prod.No.: D3695, Sigma-Aldrich, St. Louis, MO, USA).

2.4. Triglyceride accumulation and droplet size distribution

Cells were fixed 1 h in 10% formalin, washed twice with 1 mL phosphate-buffered saline and stained with a working solution of Oil Red O (Prod. No. O0625, Sigma Aldrich, St. Louis, MO, USA) for 1.5 h. Subsequently, cells were washed, air dried and Oil Red O extracted with 0.5 mL isopropanol. Total lipid content was deter- mined quantitatively [14]. Droplet size distribution was determined from digital images (20×, Leica DMLB 100T, Leica Microsystems, Wetzlar, Germany). For each experimental condition, three independent sets of cells were used. In each set, three microscopic images were taken in randomly selected fields (cultureware dish bottom was divided into 100 numbered squares and 3 square numbers were selected for imaging using random number generator).

2.5. Image analysis

Lipid droplets were counted using Matlab 7.12.0 (MathWorks Inc., Natick, MA, USA) algorithm using a disk structural element (SE) to detect lipid droplets and determine assignment into indi- vidual size categories. DAPI (4′,6-diamidino-2-phenylindole) counterstaining was used to count nuclei.

2.6. Gene expression analysis

Isolated RNA was treated with DNAse (Roche Diagnostics, Mannheim, Germany) and gene expression of DGAT (diacylglycerol O-acyltransferase 2), FABP4 (fatty acid binding protein 4), FASN (adipocyte fatty acid synthase), HSL (hormone sensitive lipase), PLIN1 (perilipin 1), ACLY (ATP-dependent citrate lyase) and TBP (TATA box binding protein) was assessed by qPCR (Applied Bio- systems, Carlsbad, CA) using TaqMan probes (Product ID: Mm0049- 9536_m1, Mm00445878_m1, Mm00662319_m1, Mm01197698- _m1, Mm00495359_m1, Mm00558672_m1, Mm01302282_m1,Mm00446971_m1). TBP was used as an endogenous control and results were expressed as DDCt (threshold cycle) values.

2.7. Statistical analysis

The effect of mild and severe hypoxic exposure on outcome variables was analyzed using ANOVA test with Tukey’s post-hoc analysis using the GraphPad (GraphPad Software, Inc., La Jolla, CA, USA). The effect of pharmacological substances on outcome vari- ables under various O2 levels was analyzed using 2-way ANOVA and interaction between the pharmacological treatment and O2 exposure was determined. Data represent mean ± SEM. A value of p < 0.05 was considered significant.

3. Results

3.1. Lipogenesis and lipid droplet size distribution

Exposure to mild hypoxia markedly stimulated lipogenesis by 375% as assessed by total triglyceride accumulation, while severe hypoxia had opposite effect, it inhibited lipogenesis by 43% (0.07 ± 0.004 vs. 0.19 ± 0.01 vs. 0.04 ± 0.002 mg lipids/mg DNA for control, mild and severe hypoxia, respectively, all p < 0.05, ANOVA), Fig. 1A. More detailed analysis using droplet size distribution revealed that mild hypoxia was associated with increased number of the largest lipid droplets in adipocytes (>30 mm diameter) by 540% compared to control conditions (25.6 ± 1.12 vs. 4.0 ± 0.4 droplets/100 cells, p < 0.05). Importantly, exposure to severe hyp- oxia stimulated formation of predominantly smaller lipid droplets, while large droplets were not formed, leading to lower total tri- glyceride content than in mild hypoxia (Fig. 1B).

3.2. Adipocyte differentiation markers

Gene expression of key adipocyte differentiation markers under normoxic, mild and severe hypoxic exposures is summarized in Fig. 1C. Data suggest that mild hypoxia (4% O2) promoted differ- entiation/adipogenesis and increased gene expression of FABP4, DGAT, HSL and perilipin 2.3 ± 0.1, 12.1 ± 3.3, 3.8 ± 0.3 and 5.3 ± 0.2 fold (all p < 0.05). In contrast, severe hypoxia (1% O2) led to significantly smaller induction of DGAT, HSL and perilipin gene expression (8.5 ± 2.3, 1.9 ± 0.4 and 2.4 ± 0.3 fold increase, all p < 0.05), while no change in FABP4 expression was detected (0.9 ± 0.1 fold change, p > 0.05). Among the analyzed genes, fatty acid synthase (FASN) was the only gene exhibiting a dose-response relationship with hypoxic exposure (4.9 ± 1.0 and 6.7 ± 1.6 fold change, respectively, p < 0.05), while ACLY was not affected by neither mild nor severe hypoxic exposure (1.4 ± 0.1 and 1.7 ± 0.5 fold change, respectively, p > 0.05).

3.3. Spontaneous lipolysis

Exposure to mild hypoxia (4% O2) augmented spontaneous lipolytic rate by 488% (glycerol in media: 0.26 ± 0.01 versus 1.53 ± 0.07 mmol/mg lipids/180min, p < 0.05), while severe hypoxia (1% O2) exhibited smaller effect by stimulating spontaneous lipol- ysis by 135% (0.26 ± 0.01 versus 0.61 ± 0.30 mmol/mg lipids/180min, p < 0.05), Fig. 1D.

3.4. Mechanisms mediating the effects of hypoxia

We further investigated possible mechanisms mediating lipogenesis-stimulating effect of mild hypoxia. First, we aimed to elucidate the metabolic source of the key substrate for fatty acid synthesis, acetyl-CoA. Citrate produced in tricarboxylic acid cycle can be converted to acetyl-CoA by ACLY (ATP-dependent citrase lyase) and used for triglyceride synthesis. Inhibition of ACLY with a cell permeable inhibitor SB204990 suppressed lipogenesis by 81% and 85% under mild and severe hypoxia, respectively (p < 0.05). Subsequently, we assessed the contribution of glycolysis to the acetyl-CoA pool by activating PDH (pyruvate dehydrogenase), providing acetyl-CoA from glycolytic pathway, with dichloroacetate (DCA). Under normoxic conditions, DCA treatment increased tri- glyceride cell content by 61% (p < 0.05), however, it had no effect on lipogenesis at 4% O2 and suppressed lipogenesis by 52% under 1% O2 (p < 0.05).

Finally, we assessed whether mild stimulation of triglyceride accumulation observed in mild hypoxia could be mediated by activation of HIF-1 (hypoxia inducible factor-1) transcriptional regulator. As shown in Fig. 2., 1 mM and 2 mM DMOG treatment effectively activated HIF-1, but it suppressed lipogenesis by 59.1 ± 9.6% and 32.8 ± 3.1%, respectively, (p < 0.05) and was thus not following the pattern of hypoxia-stimulated triglyceride accumulation.

4. Discussion

The unique cultureware technology used in this study enabled us to demonstrate that adipocyte morphology and metabolic functions are regulated differently under mild and severe hypoxia. While exposure of 3T3-L1 cells to mild hypoxia (4% O2) increased gene expression of adipocyte-specific markers, augmented tri- glyceride accumulation and promoted formation of large intracel- lular lipid droplets, more severe hypoxic exposure (1% O2) induced smaller increase in adipogenesis markers expression and have not stimulated formation of large lipid droplets. Similarly, spontaneous lipolysis was markedly stimulated after exposure to 4% O2, but less induced under 1% O2.

Hypoxia was associated with impaired adipocyte function including increased expression and production of pro- inflammatory adipokines [15e19] attributed to HIF-1 and/or NFkB transcriptional activation under hypoxia [3,20]. Additionally, in- duction of adipocyte insulin resistance was observed after acute exposure to hypoxia [19,21], while genes involved in insulin- independent glucose uptake pathways were enhanced under hypoxia [22,23]. The present study utilized a paradigm based on a prolonged exposure of cultured cells to defined and stable peri- cellular O2 levels for 14 days [12] expanding thus the available knowledge beyond the 48 h timepoint [2,4,6]. Prolonged exposures more closely model the effects of tissue hypoxia observed in obesity or OSA as acute adaptations are often different from the effects of chronic exposure to hypoxia. For example, prolonged, but not acute, hypoxic exposure was shown to stimulate spontaneous lipolysis in 3T3-L1 adipocytes [24]. Similarly, the expression of GLUT 4 receptor gene was found to be initially up-regulated, but subsequently reduced after two days of hypoxic stimulus [25]. Additionally, transcriptional response mediated by HIF-1, a key mediator for adipocyte adaptation to hypoxia, is transient and diminishes significantly over the period of several hours [3,19,26]. Similarly,
gene expression profiles in vivo show distinct patterns after acute versus chronic hypoxic exposures [27].

Using a gas-permeable membrane bottom, we exposed cells to 4% O2, approximating the lower side of tissue O2 levels observed in vivo in obese rodents and humans or during intermittent oxygen desaturations characteristic for OSA [4,6,28]. This level of mild hypoxia promoted adipocyte differentiation into mature adipocytes with large intracellular lipid stores and big intracellular lipid droplets. These findings suggest that adipose tissue hypoxia thus might precede (such as in OSA) or perpetuate the development of predominantly hypertrophic adipocytes with secretory profile promoting low-grade inflammation and insulin resistance [29]. Our study extends previous observations of increased lipolytic rate after acute hypoxic exposure [19] by demonstration that chronic mild hypoxia increased spontaneous lipolytic rate. This finding is particularly important, as high rates of adipose tissue lipolysis contribute to elevated plasma levels of circulating FFA representing a causal factor in the development of insulin resistance, pancreatic dysfunction and eventually T2DM [30]. Even though exposure to mild hypoxia increased lipid droplet size, paralleling larger adipo- cytes in vivo, we showed increased lipolysis after normalization to lipid content, suggesting additional, size-independent effects of chronic mild hypoxia [31]. Such mechanisms might include modi- fied secretion of lipolysis stimulating adipokines such as tumor necrosis factor-a and interleukin-6 [2] as well as decreased expression of enzymes involved in lipid oxidation, rendering thus more FFA for extracellular transport and release [23].

In contrast to mild hypoxia, severe hypoxia of 1% O2 exhibited different effects on all investigated outcome variables, providing a rare documentation that responses to hypoxia do not follow a linear dose-response pattern. Although initially contra-intuitive, this finding can be readily explained by the fact that the key cellular regulator of hypoxic adaptations, HIF-1, is precisely up-regulated at O2 levels dropping below 4-5%, with the strongest up-regulation at 0.5% O2 [32,33]. At such low O2 levels, oxidative metabolic functions of cultured adipocytes are reduced, including adipogenesis, differ- entiation, lipolysis together with down-regulation key adipogenic regulators [23,34] Furthermore, insulin resistance induced by se- vere hypoxia might also decrease triglyceride accumulation as a consequence of limited effect of insulin present in the culture media [21]. Supporting the reduced oxidative capacity, we observed increased formation of predominantly small lipid drop- lets in severe hypoxia, which was previously associated specifically with states FFA overload [35]. Even though HIF-1 mediates changes in gene transcription regulating principal cell responses to acute hypoxia, our study showed that chemical stabilization of HIF-1 did not stimulate lipid synthesis, in fact, triglyceride accumulation was reduced in a dose-response manner. This finding supports the notion that mild hypoxia stimulates lipid synthesis by HIF- independent mechanisms, while severe exposure activating HIF-1 inhibits adipogenesis as suggested previously [36].

To identify potential drug targets, we sought to identify mechanisms providing substrates for increased triglyceride synthesis under mild hypoxia. We focused on the intracellular source of acetyl-CoA representing a key source of de novo fatty acid and triglyceride synthesis [37]. We found that ATP-dependent citrate lyase activity (ACLY) is essential for hypoxia-induced lipogenesis, suggesting that the major source of acetyl-CoA in hypoxic adipo- cytes is tricarboxylic acid cycle, as has been previously shown in other cell types and in cancer cells [38,39]. However, even under conditions of fully inhibited ACLY, lipogenesis was still higher in both hypoxic conditions than in the control exposure, pointing to the other intracellular sources of acetyl-CoA. Glycolytic source of acetyl-CoA seems to play a minor role as pharmacological activa- tion of pyruvate-dehydrogenase inhibited lipogenesis under severe hypoxia.

In summary, adipocytes respond to reduced O2 availability depending on the severity of hypoxia. Mild hypoxia stimulated adipogenesis, lipogenesis and increased spontaneous lipolytic rate, supporting the role of hypoxia in the development of adipocyte as well as whole body metabolic impairments. In contrast, more se- vere hypoxic exposure is characterized by lower stimulation of adipocyte differentiation and reduced stimulation of lipolysis. Be- sides obesity, we describe a possible mechanism for LF3 adipocyte impairments imposed by other hypoxic states, particularly obstructive sleep apnea.