Adipocyte triglyceride turnover and lipolysis in lean and overweight subjects

Abstract

Human obesity is associated with decreased triglyceride turnover and impaired lipolysis in adipocytes. We determined whether such defects also occur in subjects with only moderate increase in fat mass. Human abdominal subcutaneous adipose tissue was investigated in healthy, nonobese subjects [body mass index (BMI) > 17 kg/m² and BMI < 30 kg/m²]. Triglyceride age, reflecting lipid turnover, was examined in 41 subjects by assessing the incorporation of atmospheric ¹⁴C into adipose lipids. Adipocyte lipolysis was examined as the ability of lipolytic agents to stimulate glycerol release in 333 subjects. Adipocyte triglyceride age was markedly increased in overweight (BMI ≥ 25 kg/m²) compared with lean subjects (P = 0.017) with triglyceride T_1/2 of 14 and 9 months, respectively (P = 0.04). Triglyceride age correlated positively with BMI (P = 0.002) but not with adipocyte volume (P = 0.2). Noradrenaline-, isoprenaline- or dibutyryl cyclic AMP-induced lipolysis was inversely correlated with triglyceride age (P < 0.01) and BMI (P < 0.0001) independently of basal lipolysis, gender, and nicotine use. Current, but not the highest or lowest BMI in adult life, correlated significantly (inversely) with lipolysis. In conclusion, adipocyte triglyceride turnover and lipolytic activity are decreased in overweight subjects and reflect the current BMI status. These changes may confer an increased risk for early development and/or maintenance of excess body fat.

Excess body fat is a well-established risk factor for numerous complications in the overweight and obese states (1). Adipose tissue mass is determined by its lipid turnover; i.e., the balance between incorporation and removal of triglycerides (TG) into adipocytes where adipocyte lipolysis (hydrolysis of intracellular TGs) is the most important factor for lipid removal. The factors regulating the expansion of human adipose are less well known. Recent studies have shown that adult human adipose tissue is in a highly dynamic state (2). Fat cells turn over at a high rate in which ∼10% of the cells are renewed each year (2), and the number of new fat cells is markedly increased in obesity. Furthermore, the lipid content in fat cells (primarily TGs) also undergoes a high turnover (3). These lipids constitute a metabolically active pool which is replaced about six times during the ten-year life span of an adipocyte. By determining the lipid age in fat cells, it is possible to estimate lipid turnover. A higher age is indicative of lower turnover and vice versa. We recently developed a method to determine adipocyte lipid age in humans by measuring the incorporation of atmospheric ¹⁴C into adipocyte TGs (3). In subcutaneous adipose tissue, the body's largest adipose region, lipid age was markedly increased in obese compared with nonobese subjects, which is indicative of low TG turnover in fat cells of the obese. This may be secondary to defects in lipolysis since many studies have shown that the ability of catecholamines to stimulate adipose lipolysis, both in vivo and in vitro, is decreased in obesity (reviewed in Refs. 4, 5).

Obesity is preceded by an overweight state that is associated with increased risk for type 2 diabetes (6, 7). There is no information about adipose lipid turnover and lipolysis in overweight subjects. We therefore investigated these factors in subcutaneous white adipose tissue of apparently healthy lean [body mass index (BMI) > 17 kg/m² and BMI < 25 kg/m²] or overweight (BMI ≥ 25 kg/m² and BMI < 30 kg/m²) subjects. In one cohort (n = 41), we determined adipocyte lipid age by assessing the incorporation of atmospheric ¹⁴C into TGs. In a second cohort (n = 333), we investigated the ability of lipolytic agents to stimulate glycerol release (lipolysis index) from adipocytes.

METHODS

Subjects

All subjects lived in the Stockholm (Sweden) area. Cohort one included 10 men and 31 women; all were healthy and free from any medication. The second cohort consisted of 121 men and 212 women. They were recruited as controls in an ongoing case-control study of obesity genetics. Exclusion criteria were diseases that influence body weight and continuous oral or injected medications influencing body weight and/or adipose lipolysis. All subjects were weight stable for at least 6 months prior to examination according to self-report. All subjects had body mass index (BMI) > 17 kg/m² and BMI < 30 kg/m². Overweight was defined as BMI ≥ 25 kg/m² and BMI < 30 kg/m². In cohort one, 13 subjects were lean and 28 were overweight. In cohort two, 54 subjects were overweight and 279 were lean. In 115 subjects of cohort two, we had information on weight variation; i.e., highest and lowest body weight in adult life. These weights differed by −14 ± 8% to +7 ± 8% of the weight at examination. All subjects were weight stable for at least six months prior to examination according to self-report. In cohort two, we had information regarding regular use of nicotine (tobacco, snuff, nicotine tablets, or chewing gum) from 321 subjects. After an overnight fast, an abdominal subcutaneous fat biopsy was obtained as a needle biopsy specimen under local anesthesia. Adipocyte lipolysis and fat cell volume were measured in freshly isolated adipocytes as described previously (8). In brief, fat cells were isolated and incubated in vitro in the absence or presence of increasing concentrations of noradrenaline (a natural catecholamine), isoprenaline (a synthetic catecholamine acting on β-adrenergic receptors), and dibutyryl AMP (a synthetic cyclic AMP analog activating protein kinase). Glycerol in the medium (lipolysis index) was determined, and the release in the absence of a lipolytic agent (basal) or in its presence at the maximum effective concentration was assessed. There is no consensus on how to express lipolysis data, such as per weight unit of adipocytes, per number of fat cells, or as a ratio of spontaneous (basal) lipolysis. We used the latter mode of expression since it fitted the data best. The study was approved by the Regional Ethics Board and was performed in accordance with the guidelines in the Declaration of Helsinki. Informed consent was obtained from all subjects.

Determination of triglyceride age

Since the Limited Nuclear Test Bomb Treaty in 1963, ¹⁴C levels in the atmosphere are decreasing exponentially. Due to the long isotope half-life, this is due to diffusion of ¹⁴CO₂ out of the atmosphere (9) which is taken up into the biotope. Because we eat plants and animals, the atmosphere ¹⁴C content is directly mirrored in the human body. In cohort one, samples of adipose tissue were subjected to lipid extraction, and the resulting sample was processed for measurements of ¹²C, ¹³C, and ¹⁴C as described (3). Data were presented as the ratio of ¹⁴C to ¹²C for each sample divided by that of a well-defined standard sample measured in the same year and incorporating ¹³C fractionation correction as described (3). These data were used to calculate TG age and half-life as described (3) using a linear replacement model in which age distribution of lipids within an individual was exponentially distributed corresponding to a constant turnover rate per year (10). This model is different than that for determining fat cell age (2). The latter model takes into account that DNA is “static” when a cell does not divide so ¹⁴C-DNA reflects the original ¹⁴C in the atmosphere the last time the fat cell divided. However, the lipids in fat cells are not static. If they were, their ¹⁴C content would be dependent on the time of collection (no or low levels before nuclear bomb testing and a gradually lower after stopping of bomb testing). However, no such relationship is observed (3). In theory, the incorporation of ¹⁴C from the atmosphere into the lipids could be influenced by seasonal factors (e.g., crops harvested and consumed by humans and animals at different times of the year). To address this methodological issue, we reanalyzed TG age data from subcutaneous adipose tissue of 91 subjects obtained from a previous study (3). The samples were divided into two seasonal categories: the warm period (n = 24) occurring between mid-April and mid-October and the cold period (n = 67) occurring the rest of the year. Adipose TG age (mean ± SD) was 1.55 ± 0.96 and 1.58 ± 0.77 years in the two groups (P = 0.86 by t-test). Thus, seasonal variations in ¹⁴C intake had no bearing on the data.

Statistics

Values are mean ± SD. They were compared by t-test, ANCOVA (ANCOVA) and single or multiple regression. Lipolysis data were log transformed to improve normalization.

RESULTS

In cohort one, TG age was determined by the ¹⁴C method (Fig. 1). Adipocyte TG age (Fig. 1A) and half-life (Fig. 1B) were markedly increased in overweight compared with lean subjects. BMI correlated positively and strongly with adipocyte TG age (Fig. 1C), but there was no relationship between mean fat cell volume and adipocyte TG age (Fig. 1D). Gender did not influence the relationship between BMI and TG age according to ANCOVA (F = 0.02; P = 0.89).

Fig. 1.

Adipose triglyceride turnover. Triglyceride age (A) and half-life (B) in lean and overweight subjects. Correlation between TG age and BMI (C) or fat cell volume (D).

The individual relationship between lipolysis stimulation and adipocyte TG age was investigated (Fig. 2). A strong negative correlation (r = 0.5; P < 0.01) with TG age was observed for all three pro-lipolytic agents (noradrenaline, isoprenaline, dibutyryl, cyclic AMP). Again, these relationships were not influenced by gender according to ANCOVA (values not shown).

Fig. 2.

Relationship between adipocyte TG age and adipocyte lipolysis stimulated with different lipolytic agents.

For lipolysis in cohort two (Fig. 3), similar results were obtained with the three lipolytic agents; i.e., a negative correlation between lipolysis and BMI (Fig. 3A–C; r = 0.41–0.48; P < 0.0001). These relationships were independent of gender or nicotine use according to ANCOVA (values not shown). In 115 subjects, we had information about weight changes in adult life. In a multiple regression analysis of isoprenaline-stimulated lipolysis versus present, highest, or lowest BMI in adult life, only the current BMI correlated (negatively) with lipolysis (Fig. 3D). Similar results were obtained with noradrenaline and dibutyryl cyclic AMP (data not shown).

Fig. 3.

Relationship between BMI and lipolysis. (A–C) Single regression of current BMI and lipolysis stimulated with different lipolytic agents. (D) Multiple regression of highest, lowest, and current BMI in adult life versus isoprenaline-stimulated lipolysis.

Since lipolysis data were expressed as a ratio between stimulated and basal lipolysis, the results might be influenced by variations in the rate of the latter. We therefore analyzed basal lipolysis (expressed as micromoles of glycerol per 2 h of incubation per gram of extracted lipids from the incubated sample) separately in cohort one and two. Basal lipolysis did not correlate with adipose TG age (r = 0.23; P = 0.18, cohort one) but did slightly so with BMI (r = 0.17; P < 0.001, cohort two). However, according to adjusted r², the latter could only explain 2.7% of the BMI-related variation in basal lipolysis. We also examined the relationship between BMI and ¹⁰log noradrenaline/basal, ¹⁰log isoprenaline/basal, or ¹⁰log dibutyryl cyclic AMP/basal lipolysis and corrected for the basal rate of lipolysis in multiple regression. The partial r values for the three measures of stimulated lipolysis were −0.41, −0.51, and −0.43, respectively, and P values were <0.0001 for all. Finally, correlation analyses in the two cohorts were performed using absolute values for isoprenaline and dibutyryl cyclic AMP-induced lipolysis (the two most pro-lipolytic agents tested) expressed as log micromoles of glycerol per 2 h of incubation per gram of extracted lipids. In cohort one, the correlation with adipose TG age was r = −0.43 and P = 0.009 for isoprenaline, and r = −0.56 and P = 0.0002 for dibuturyl cyclic AMP. In cohort two, the correlation between absolute values of stimulated lipolysis and BMI was the same for both lipolytic agents (r = −0.34 and P < 0.0001). Taken together, these findings suggest that the observed relationships are not dependent on variations in basal rates of lipolysis.

DISCUSSION

Although adipose tissue mass is determined by the storage and removal of adipocyte TGs (11), very little is known regarding how these factors contribute to body weight variations within the nonobese range. We investigated the role of TG turnover and lipolysis in lean and overweight subjects. To this end, we used our recently developed method to determine in vivo incorporation of atmospheric ¹⁴C into human fat cell TGs in order to estimate TG age. This reflects the ability of fat cells to remove and subsequently oxidize the lipids, and consequently, it is a valid measure of lipid turnover as previously discussed (3). We also investigated adipocyte lipolysis since hydrolysis of intracellular TGs is essential for lipid removal and an important factor determining adipocyte TG turnover.

A prominent finding was increased adipose TG age and half-life among overweight subjects. There was a continuous, positive relationship between TG age and BMI over the examined body weight spectrum (18–29 kg/m²). This suggests that adipose tissue turnover of TGs is highly sensitive to variations in BMI and that moderate weight increases, even within the normal weight range, affect TG turnover. The data are in line with previous observations comparing TG turnover in the nonobese versus obese (3) suggesting that there is a linear relationship between body weight and TG turnover over the entire BMI range.

A greater estimated TG age in fat cells of overweight subjects could be due to increased volume of distribution of ¹⁴C-labeled lipids in overweight subjects. If important, such effect should also be observed at the fat cell level; i.e., large fat cells should have a greater lipid age than small ones. However, this is clearly not the case since there was no correlation between TG age and fat cell volume (Fig. 1D). Furthermore, in a previous study (3) TG age was the same in small and large fat cells isolated from the same fat depot in the same subject. Thus, differences in transit-time for ¹⁴C lipid cannot explain our findings in lean versus overweight subjects.

Lipolysis is a major factor determining adipocyte TG turnover. In a previous study of subjects not selected for BMI, we observed an inverse relationship between adipocyte TG age and catecholamine-stimulated lipolysis (3). Herein, we found similar associations in nonobese subjects, where negative correlations were observed between TG age and lipolysis induced by various pro-lipolytic agents. Importantly, these correlations were not influenced by differences in basal lipolysis rates. This suggests that decreased adipocyte TG turnover in the overweight state involves a (selective) reduction in the ability of fat cells to activate lipolysis. Furthermore, there was a strong inverse relationship between activation of lipolysis and BMI in the lean-to-overweight range of BMI irrespective of whether lipolysis was activated by a natural catecholamine (i.e., acting on both pro-lipolytic β-adrenoceptors and anti-lipolytic α adrenoceptors), by a synthetic agent acting only on β-adrenoceptors (isoprenaline), or by an agent acting more distally in the lipolysis cascade (i.e., at the postreceptor level; dibutyryl cyclic AMP). Again, basal lipolysis levels did not affect these associations. We therefore propose that TG turnover and removal of TGs through hormone-stimulated lipolysis are highly correlated in nonobese subjects so that small variations in BMI alter both parameters in a parallel fashion. Similar in vitro impairments of lipolysis are observed when nonobese and obese subjects are compared (as reviewed in Ref. 4). Thus, defects in lipolysis activation appear to be early alterations in the development of excess body fat mass and are localized at distal step(s) in the lipolytic cascade; i.e., near the lipases responsible for hydrolysis of TGs. At present we do not know whether these disturbances are a cause or consequence of increased fat mass/high BMI.

In theory, our findings could be influenced by BMI-dependent differences in adipocyte age. However, such differences cannot be of any importance since we previously (2) demonstrated that fat cell age is independent of BMI. It is also not likely that lipid recycling, leading to incorporation of older carbon atoms, could affect our estimates. This notion is based on the observation that ¹⁴C data obtained from individuals born before, during, and after nuclear bomb testing display similar TG age half-lives (3). In fact, as long as the major triglyceride pool is greater than 80% of the lipids, recycling lipids (if they exist) can be neglected in the calculation of TG age (3).

Although body weight and lipid metabolism are influenced by gender and nicotine use (as reviewed in Refs. 10, 11), these factors did not influence our findings in an important way. Nevertheless, there are a few important caveats with the present study. First, we only investigated abdominal subcutaneous adipose tissue. Since it is well established that adipose lipolysis is subjected to regional variations (reviewed in Refs. 4, 5), it is possible that results in other adipose depots could be different. Second, our study was cross-sectional. As mentioned above, we do not know whether low TG turnover in fat cells precedes the development of moderate fat mass gain. Since the methods to examine the lipid turnover in human fat cells have been developed quite recently, it will take some time before long-term prospective data can be generated. However, the relationship between adipose lipolysis and current (but not previous) body weight (Fig. 3D) indicates that altered lipolysis is probably a consequence of increased fat mass rather than a factor preceding the development of overweight/obesity. Finally, we do not know how lipolysis and adipocyte TG turnover are influenced during periods of increase or decrease in body weight within the nonobese range. Our subjects reported weight stability at the time of the investigation.

In conclusion, decreased TG turnover and lipolysis activation are adipose defects that are observed already in the overweight state. Whether these disturbances predispose for future weight gain remains to be determined in long-term prospective studies.