Dynamics of human adipose lipid turnover in health and metabolic disease

Abstract

Adipose tissue mass is determined by the storage and removal of triglycerides in adipocytes. Little is known, however, about adipose lipid turnover in humans in health and pathology. To study this in vivo, lipid age was determined by measuring nuclear bomb test-derived ¹⁴C in adipocyte lipids. We report that during the average ten year life span of human adipocytes, triglycerides are renewed six times. Lipid age is independent of adipocyte size, is very stable across a wide range of adult ages and does not differ between genders. Adipocyte lipid turnover, however, is strongly related to conditions with disturbed lipid metabolism. In obesity, triglyceride removal rate from fat tissue is decreased and the amount of triglycerides stored each year is increased. In contrast, both lipid removal and storage rates are decreased in non-obese patients diagnosed with the most common hereditary form of dyslipidemia, familial combined hyperlipidemia. Lipid removal rate is positively correlated with the capacity of adipocytes to break down triglycerides, as assessed through lipolysis and is inversely related to insulin resistance. Our data support a mechanism in which adipocyte lipid storage and removal play different roles in health and pathology. High storage but low triglyceride removal promotes fat tissue accumulation and obesity. Reduction of both triglyceride storage and removal decreases lipid shunting through adipose tissue and thus promotes dyslipidemia. We identify adipocyte lipid turnover as a novel target for prevention and treatment of metabolic disease.

A major function of adipose tissue is to store and release fatty acids, which are incorporated into adipocyte triglycerides according to whole body energy demands. Body fat mass is determined by the balance between triglyceride storage and removal in adipocytes, by either enzymatic hydrolysis (lipolysis) and subsequent fatty acid oxidation and/or ectopic deposition in non-adipose tissues. Very little is known about the dynamics of these processes in humans. Although isotope tracer methods have been used to estimate lipid turnover in human adipose tissue, these studies have been limited to short-term experimental conditions(1-4). To study long-term adipose tissue lipid turnover in vivo and across the adult lifespan, we developed a method to retrospectively determine the age of adipocyte triglycerides in humans. Triglycerides are the major component of the adipocyte lipid droplet. Lipid age was assessed by measuring the ¹⁴C content in the lipid compartment of adipocytes from human subcutaneous adipose tissue, the major fat depot in humans. ¹⁴C levels in the atmosphere remained remarkably stable until above ground nuclear bomb tests between ca. 1955 and 1963 caused a significant increase in ¹⁴C relative to stable carbon isotope levels (5) (Fig. 1A). Following the Limited Nuclear Test Ban Treaty signed in 1963, ¹⁴C levels in the atmosphere have decreased exponentially. This is not due to radioactive decay (T_1/2 for ¹⁴C is 5,730 years), but to diffusion of ¹⁴C out of the atmosphere (6). ¹⁴C in the atmosphere oxidises to form CO₂, which is taken up in the biotope by photosynthesis. Since we eat plants, or animals that live off plants, the ¹⁴C content in the atmosphere is directly mirrored in the human body.

Fig. 1.

Atmospheric ¹⁴C over time. A: Above ground nuclear bomb testing during the period of the cold war caused an increase in atmospheric levels of ¹⁴C. These values decreased exponentially following implementation of a limited world-wide test ban treaty in 1963 (blue curve). Lipid age is determined by measuring ¹⁴C levels in lipids (1) and plotting this value against the bomb-curve (2) to determine the difference between the year corresponding to the atmospheric ¹⁴C concentration (3) and the biopsy collection date (dashed line). Atmospheric ¹⁴C levels are presented as ¹⁴C/¹²C ratios in units of Fraction Modern (for a definition of ‘Modern’ see Supplement 2). B: Lipid age and turnover do not change as a function of person age. Lipid age is shown for three individuals born in 1940.2, 1959.9 and 1967.9. Lipid age was shown to be the same for all individuals, despite markedly different subject ages. Fat biopsies were collected from all individuals on the same date (dashed vertical line). The solid vertical line indicates the date of birth (DOB). The small dashed lines show the ¹⁴C lipid value for each individual (only two 14C lipid values are shown, since two of them were equal).

Radiocarbon dating has previously been used to study the incorporation of atmospheric ¹⁴C into DNA in order to determine the age of different human cell types, including adipocytes (7-11). Here, we compared the incorporation of ¹⁴C into adipocyte triglycerides with the dynamic changes in atmospheric ¹⁴C described above. Triglyceride age was determined by using a linear lipid replacement model in which the age distribution of lipids within an individual was exponentially distributed corresponding to a constant turnover rate (per year) (12). The associated mean age, termed lipid age, is the inverse of the turnover rate and reflects the irreversible removal of lipids from adipose stores (Supplement 1).

Earlier studies suggest that triglycerides in adipose tissue form two distinct pools with high or low turnover rates, respectively (13, 14). However, our data, obtained from individuals born before, during and after bomb testing, do not support the hypothesis of dual large lipid pools with different half lives (Fig. 1B). ¹⁴C data were modelled according to one or more pools of lipids with different lipid removal rates (details in Supplement 1). The existence of a very small pool of younger lipids cannot be excluded based on data modelling. According to a two-pool model (Supplement 1), the influence on the turnover rate is proportional to the fraction of lipid in the small pool. Triglyceride exchange between adipocytes and other small storage pools can affect turnover estimates. The two-pools model shows, however, that the non-adipose pool can be neglected when it makes up less than 20% of the lipids. Small pools with high turnover are more important for short term (days or weeks) rather than long term (years) triglyceride turnover.

Mean lipid age was 1.6 years (Fig. 2A) which is in the same range as in short term turnover studies (4). The distribution of lipid age was compared with that of adipocyte age reported previously in a comparable cohort (9). The mean age of adipocytes was 9.5 years (Fig. 2B). This implies that triglycerides, on average, are replaced six times during the life span of the adipocyte, enabling a dynamic regulation of lipid storage and mobilization over time.

Fig.2.

Adipocyte age and lipid turnover in subcutaneous fat. A: Distribution of values for lipid age in healthy non-obese or obese individuals. B: The distribution of values for human adipocyte age. Data are obtained from a previous publication (see main text). C: Lipid age in 48 non-obese, 30 obese and 13 non-obese FCHL subjects. D: Lipid storage in the same cohort as in C. T-bars indicate standard error. Overall effect is p<0.0001 by ANOVA in C and D. Results in graphs are from post-hoc test. Data are from Cohort 1 (See Suppl. 2).

There is a large variation in adipocyte size within and between individuals (Supplement 2, Table 1; ref 15). However, it is unlikely that the rate of triglyceride removal from adipocytes is important for these variations since lipid age was not related to adipocyte size when set in relation to the body fat mass (Supplementary Fig. 1 A,B), nor was there a difference in lipid age between large and small adipocytes isolated from the same adipose tissue sample (Supplementary Fig. 1 C,D). These data suggest that there is a continuous exchange of lipids between adipocytes within the adipose tissue which is not dependent on adipocyte size. Fatty acids produced by lipolysis in one adipocyte could, for example, be taken up by adjacent adipocytes and incorporated into their triglycerides.

Lipid age and total fat mass data were used to determine the net triglyceride storage in adipose tissue (kg/year) (see Supplement 1, Box 1). The net amount of lipid stored in adipose tissue each year is the sum of exogenous fat incorporation and endogenous synthesis, minus lipid removal. The removal rate is determined by the hydrolysis of triglycerides (lipolysis) and the irreversible removal of lipids from adipose stores due to fat oxidation. This implies that a high lipid age mirrors low removal rates. No relationship between lipid storage or removal and person age or gender was seen (Supplement 2, Fig. 2 A-D).

Two clinical conditions where altered lipid metabolism is observed were investigated – obesity and familial combined hyperlipidemia (FCHL). Obesity is a world-wide epidemic considered to be a disease burden comparable to smoking. FCHL constitutes the most common hereditary lipid disorder (reviewed in (16)). It has an unknown etiology and is a common hereditary cause of premature coronary heart disease. Adipocyte lipolysis is impaired in both conditions due to decreased cyclic AMP-dependent signaling, the major lipolytic pathway in adipocytes (17-19). In addition, as reviewed (20), both conditions display a similar metabolic phenotype (mixed dyslipidemia, elevated apolipoprotein B and insulin resistance). These clinical characteristics are confirmed in our study cohort (Supplement 2, Table 1). Importantly, FCHL affected individuals may present with a range of body fat levels, however for our analyses only non-obese FCHL patients were selected so as to remove the confound of obesity from the study.

In obese subjects, the rate of triglyceride storage (Fig. 2D) and mean lipid age (Fig. 2C) were markedly increased compared to non-obese individuals. Both lipid age (r=0.38, p=0.0005) and triglyceride storage (r=0.60, p<0.0001) correlated with body mass index when non obese and obese individual were pooled together. Similarly, in non-obese FCHL individuals lipid age was increased to values observed in obesity (Fig. 2C). In contrast to obesity however, the rate of triglyceride storage was markedly decreased compared to non-obese individuals (Fig. 2D). Thus, adipocyte triglyceride turnover is not just a mere reflection of the fat mass. Our data suggests a model where a combination of high storage and low lipid removal rates, as in obesity, facilitates triglyceride accumulation within adipose tissue, thereby promoting the development and/or maintenance of excess body fat mass. Conversely, a low rate of both triglyceride storage and removal, as in FCHL, leads to reduced triglyceride turnover and thereby a decreased ability of adipocytes to store and release fatty acids, despite a normal body fat mass. As discussed in detail elsewhere (21, 22), low lipid turnover in adipose tissue may result in fatty acids being shunted to the liver, which drives the synthesis of apolipoprotein B and increases the circulating levels of triglycerides. Adipocyte triglyceride turnover may also be involved in determining overall insulin effects. Insulin resistance (indirectly measured by the HOMA-IR index, see Supplement 2) and lipid turnover were assessed in 82 individuals. Triglyceride age was strongly related to levels of insulin resistance, though there was no relationship between triglyceride storage and insulin resistance (Fig. 3). There was no significant interaction between groups (lean, obese and non-obese FCHL) as determined by analysis of co-variance, indicating that the rate of triglyceride removal from adipocytes has an impact on whole body insulin sensitivity independent of underlying disorder.

Fig. 3.

Correlation between adipocyte triglyceride turnover and insulin resistance (HOMA-IR index). A linear regression analysis was performed on all individuals from cohort 1 having insulin resistance measures (n=82). HOMA-IR was correlated with lipid storage (A) and lipid age (B). The relationship between lipid age and HOMA-IR remained significant when body mass index (BMI), gender or group (non-obese, obese, FCHL) were included in the analysis (partial r=0.41, p=0.006 with BMI using multiple regression analysis and F=16.6, p=0.0001 F=4.8, p=0.03 for gender or group, respectively, using ANCOVA).

Since lipolysis is a major factor for the removal of lipids from adipose tissue we investigated the ability of the cyclic AMP system to activate lipolysis in vitro in adipocytes isolated from lean and obese individuals and compared this with in vivo measurements of lipid storage and removal (Fig. 4). Lipolysis activation was negatively correlated with triglyceride removal (lipid age) but was not related to the rate of triglyceride uptake (lipid storage). This was irrespective of whether lipolysis was induced using a cyclic AMP analogue (Fig 4 A,B), by activating endogenous adenylate cyclase (using forskolin; Fig. 4 C,D) or by administration of a synthetic beta adrenoceptor-selective catecholamine (isoprenaline; Fig 4. E,F). These data suggest that lipolysis determines lipid turnover in adipocytes by regulating the rate of triglyceride removal. The impact of subsequent fatty acid oxidation could not be examined in this study, however decreased lipid oxidations is frequently observed in obesity (23, 24).

Fig. 4.

Correlation between lipid turnover and adipocyte lipolysis. A and B: Lipid age and lipid removal rates with dibutyryl cyclic AMP (A,B), which is a phosphodiesterase resistant and stabile cyclic AMP analogue stimulating the protein kinase A complex; forskolin (C,D), a stimulator of adenylate cyclase which stimulates cyclic AMP production and isoprenaline (E,F), a synthetic beta adrenergic receptor-selective catecholamine acting as a lipolytic agents. Linear regression analysis was used. Data are from non-obese and obese individuals from cohort 1. Data with lipolytic agents versus lipid age were significant when analysed using body mass index as covariate in multiple regression analysis (partial r=−0.35; p≤0.004).

We are in the midst of a global epidemic of obesity with negative health and socio-economic consequences. We propose adipose triglyceride turnover as a novel target for the prevention and treatment of excess body fat and possibly its consequences for insulin resistance. The novel insights into abnormal triglyceride turnover in FCHL patients may also suggest novel treatment strategies for this complex disease targeting adipocytes.

METHODS SUMMARY

Subjects

Subcutaneous adipose tissue was obtained from two patient cohorts. Patient selection and collection of clinical data are described in Supplement 2.

Preparation of lipids

Triglycerides were extracted from pieces of adipose tissue or isolated adipocytes. Details of lipid extraction and adipocyte isolation are given in Supplement 2. Extracted lipids were subjected to accelerator mass spectrometry analysis, as described in Supplement 2.

Data analysis

Calculations between lipid turnover and clinical or adipocyte phenotypes are described in Supplement 2. Calculations of lipid age and net lipid uptake by adipose tissue are described in detail in Supplement 1. Conventional statistical methods were used to summarise and compare data.