figure a

Introduction

Pleiotrophin (PTN) is a highly conserved cytokine that belongs to a family of heparin-binding growth factors [1, 2]. PTN is found in cells in the early stages of differentiation, in particular during embryonic development [3, 4]. PTN contributes to epithelial–mesenchymal interactions in organs undergoing branching morphogenesis [5,6,7] and participates in bone formation [8].

In adult rodents, residual Ptn expression is restricted to uterine cells and discrete populations in the nervous system [9]. An almost identical expression profile has been found in human adult tissues [10]. PTN levels are, however, upregulated in the uterus and placenta during pregnancy [9], and in the early stages of differentiation of cell types involved in repair [11, 12] and inflammatory processes [13].

Besides its role in tumour growth, the functions of PTN include regulation of cell growth, migration and survival [14]. Importantly, PTN signalling may have an inhibitory role in the differentiation of pre-adipocytes in vitro [15], which involves crosstalk between the PTN/phosphoinositide 3-kinase (PI3K)/Akt/glycogen synthase kinase (GSK)-3β/β-catenin and Wnt/β-catenin signalling pathways to repress adipogenesis [16]. Upon induction of white adipocyte differentiation, levels of endogenous PTN decrease [17], and the in vitro administration of recombinant PTN (rPTN) inhibits white adipocyte differentiation, leading to a decreased expression of white adipocyte markers such as Ppar-γ2 (also known as Pparg2) [16, 17].

Our hypothesis proposes a role for PTN as a modulator of lipid and glucose homeostasis that might be involved in fat accumulation and body composition due to specific actions on adipose tissue. To address this question, we phenotyped a whole-body constitutive Ptn knockout mouse model (Ptn−/−) at different stages of adult life.

Methods

Animals

PTN genetically deficient (Ptn−/−) mice on a C57BL/6 J background were kindly provided by T. F. Deuel (The Scripps Research Institute, La Jolla, CA, USA) [18, 19]. Female Ptn−/− and wild-type (Ptn+/+) mice were housed at 22–24°C with 12 h/12 h light/dark cycles of 08:00/20:00 hours and free access to water and a chow diet (Panlab, Barcelona, Spain). Animals were maintained in accordance with European Union Laboratory Animal Care Rules (86/609/ECC directive) and protocols approved by the Animal Research Committee of CEU San Pablo University. At 3, 6, 12 and 15 months of age, randomly selected mice from each genotype were fasted for 5 h, exposed to carbon dioxide and killed by decapitation. Researchers and animal caretakers were blinded to group assignment. Plasma and organs were collected and preserved at −80°C.

Determination of body fat

Total body fat was extracted from dried cadavers of 6-month-old mice in an automatic Soxhlet extraction system (Buchi extraction system B-811, Flawil, Switzerland) using chloroform–methanol (3:1, vol./vol.).

Plasma analysis, estimation of insulin resistance and GTT

Glucose (using the glucose oxidase–peroxidase, aminoantipyrine, phenol method [GOD-PAP]; Roche Diagnostics, Barcelona, Spain), triacylglycerols (using lipoprotein lipase [LPL]/glycerol phosphate oxydase [GPO]/-Trinder; Roche Diagnostics) and NEFA (Acyl-CoA synthase–Acyl-CoA oxidase [ACS-ACOD] method; Wako Chemicals, Neuss, Germany) were determined by enzymatic colorimetric tests. Plasma insulin measurement (Mercodia, Uppsala, Sweden) and free T4 levels (DRG, NJ, USA) were determined using immunoassay kits. Plasma leptin was measured with a Bio-Plex Pro mouse diabetes immunoassay kit (Bio-Rad, Hercules, CA, USA). The QUICKI insulin sensitivity index was calculated as the inverse log sum of fasting insulin (in pmol/l) and fasting glucose (in mmol/l), as previously described [20]. Glucose tolerance tests (GTTs) using intraperitoneal glucose administration (2 g/kg) were performed in mice that had been fasted for 6 h; AUCs for glucose were calculated.

Quantitative real-time PCR in white adipose tissue

RNA was isolated using RNeasy Mini Kits (Qiagen, Valencia, CA, USA). First-strand cDNAs were synthesised using the iScript cDNA Synthesis Kit (Bio-Rad) and subjected to quantitative real-time PCR (qPCR) analysis by SYBR green method (Bio-Rad) in a CFX96 real-time system (Bio-Rad); the primer sequences are shown in Table 1. The relative expression of each gene was normalised against Hprt and Rpl13, used as reference standards.

Table 1 Primer sets used for qPCR analysis
Full size table

Histological studies

Randomly selected samples of periovarian adipose tissue were stained with haematoxylin and eosin and analysed by optical microscopy.

Lipolysis analysis in isolated adipocytes

Adipocytes were isolated from the periovarian adipose tissue of 15-month-old mice as previously described [21]. To test catecholamine-stimulated lipolysis, isolated adipocytes were incubated with 1 U/ml of adenosine deaminase (Sigma-Aldrich, Madrid, Spain) in the absence or presence of isoprenaline (Sigma-Aldrich). Lipolysis was quantified as liberation of glycerol into the medium, determined by the GPO-Trinder method (Sigma-Aldrich). Basal lipolysis, half-maximal effective agonist concentration (EC50) and maximum effect (Emax) for lipolysis were estimated from the concentration vs effect curves of glycerol. To measure the effect of insulin on lipolysis, isolated adipocytes were incubated with 0, 1, 10 or 100 nmol/l insulin (Sigma-Aldrich) prior to the addition of isoprenaline (100 nmol/l) and incubated for 90 min. The IC50 and maximum inhibitory effect (Imax) values were estimated.

Indirect calorimetry

Six-month-old Ptn+/+ and Ptn−/− mice were randomly selected and individually housed in chambers at 22–24°C with free access to chow and water. After adaptation for 72 h, the mice were housed for 3 days at 24°C, and then 7 days at 30°C. Oxygen consumption (\( \dot{V}{\mathrm{O}}_2 \)) and carbon dioxide production (\( \dot{V}{\mathrm{CO}}_2 \)) of individual mice were measured every 30 min over 24 h in LabMaster metabolic cages (TSE Systems, Bad Homburg, Germany). Respiratory exchange ratios (RERs) were calculated as \( \dot{V}{\mathrm{CO}}_2 \)/\( \dot{V}{\mathrm{O}}_2 \). Data on total energy expenditure (EE) were used to calculate cold-induced thermogenesis as a fraction of total daily EE [22].

Analysis in brown adipose tissue

Total T3 and T4 concentrations were determined in brown adipose tissue (BAT) from 6-month-old mice by radioimmunoassay [23]. Deiodinase 2 (DIO2) activity was assayed in BAT homogenates as previously described [24]. Mouse Dio2 mRNA was quantified by qPCR using specific Taqman probes for murine Dio2 (Applied Biosystems, Foster City, CA, USA). The results were normalised to cyclophilin. Production of uncoupling protein 1 (UCP-1) was determined by Western blotting, using a 1/1000 dilution of a specific rabbit anti-UCP1 antibody validated for western Blot (ab1426, Chemicon, Temecula, CA, USA), the corresponding 1/20000 dilution of a rabbit secondary horse radish peroxidase antibody (A0545, Sigma-Aldrich) and visualisation by enhanced chemiluminescence (Amersham-GE Healthcare, Barcelona, Spain); values were normalised to β-actin using a 1/100 dilution of a specific rabbit anti-β-actin (A2066, Sigma-Aldrich).

Cell culture and differentiation of brown adipocytes

Immortalised mBAs were generated as previously described [25], grown in DMEM, supplemented with 10% (vol./vol.) fetal serum, 1% (vol./vol.) penicillin/streptomycin (Sigma-Aldrich) and 2 mmol/l HEPES, and differentiated until they exhibited numerous multilocular cytoplasmic lipid droplets [26]. Undifferentiated (day 0) pre-adipocytes were treated with 0.1 μg/ml PTN (Sigma-Aldrich) until day 6 of differentiation; at this time point, non-treated pre-adipocytes typically achieved a fully differentiated phenotype. Markers of mBA differentiation (Cidea, Pgc1-α [also known as Ppargc1a] and Prdm16) and Ptn mRNA levels were measured by qPCR (Table 1).

Statistical analysis

Results are expressed as mean ± SEM. When data were not normally distributed, log10-transformed values were used for statistical analysis, and a Grubbs’ test was run to detect outliers. Statistical comparisons between two groups were made using the unpaired Student’s t test; comparisons between three or more groups were made by one- or two-way ANOVA, followed by a Student–Newman–Keuls post hoc test, using GraphPad Prism v7 (San Diego, CA, USA).

Results

Ptn −/− mice show reduced body weight and adiposity

The survival rate was similar in both genotypes (Fig. 1a). We first investigated whether Ptn deletion would affect body weight and fat distribution. The body weight of Ptn−/− mice was significantly lower over the whole time course of the study (Fig. 1b). Furthermore, we found that the weights (Fig. 1c, d) of periovarian and retroperitoneal adipose tissue were significantly higher in young Ptn−/− mice, whereas in older mice the weight of both fat depots was significantly lower in Ptn−/− mice. This pointed to a switch in fat distribution by deletion of Ptn, which evolved from a higher degree of adiposity in young animals to a slimness in older mice. Analysis of total body composition in a subset of 6-month-old mice corroborated these data. Ptn−/− mice showed a significant reduction in total body fat (20.2% in Ptn+/+ vs 13.9% in Ptn−/− mice), while water content was increased, compared with wild-type controls (Fig. 1e).

Fig. 1
figure 1

Ptn−/− mice show decreased body weight and reduced adiposity. (a) Survival curve, (b) body weight, (c) periovarian adipose tissue weight, (d) retroperitoneal adipose tissue weight, (e) body fat and water content analysis, and (f) circulating plasma leptin in Ptn+/+ (grey lines and bars) and Ptn−/− (blue lines and bars) female mice. Data are presented as mean ± SEM for n = 8 mice/group (a–d), n = 6 mice/group (e) and n = 5 mice/group (f). †p < 0.05, ††p < 0.01 for differences in the effect of ageing in Ptn+/+ female mice (vs 3 months); ‡p < 0.05, ‡‡p < 0.01, ‡‡‡p < 0.001 for differences in the effect of ageing in Ptn−/− female mice (vs 3 months). *p < 0.05; **p < 0.01 for differences between Ptn−/− and Ptn+/+ mice

Full size image

Increased adiposity observed in young Ptn−/− mice paralleled higher leptin levels when compared with age-matched controls (Fig. 1f). Moreover, Ptn+/+ mice exhibited an age-related hyperleptinaemia, while circulating leptin in Ptn−/− animals did not increase with age compared with their wild-type counterparts, which can be explained by their reduced adiposity.

Altered circulating lipid profile, impaired glucose tolerance and insulin resistance are age-related phenomena in Ptn −/− mice

Energy imbalance and altered adiposity are related to deteriorated lipid and glucose homeostasis. Accordingly, Ptn−/− mice showed lower plasma triacylglycerol levels from 6 to 15 months of age than Ptn+/+ controls (Fig. 2a), and circulating plasma NEFA were also significantly lower in 15-month-old Ptn−/− mice (Fig. 2b). Ptn−/− mice exhibited an altered lipid profile, with their plasma triacylglycerol and NEFA concentrations not increasing with age, as happens physiologically in Ptn+/+ mice.

Fig. 2
figure 2

Altered lipid profiles, impaired glucose tolerance and insulin resistance are age-related phenomena in Ptn−/− mice. Plasma biochemistry and insulin sensitivity indexes in Ptn+/+ female mice (grey lines) and Ptn−/− (blue lines) female mice at 3, 6, 12 and 15 months of age. (a) Triacylglycerol, (b) NEFA, (c) glucose, (d) insulin, (e) AUC for glucose (AUC-G) and (f) QUICKI. AUC-G represents the AUC for glucose during the intraperitoneal GTT in female mice fasted for 6 h. Data are mean ± SEM of n = 8 mice/group. †p < 0.05, ††p < 0.01 for differences in the effect of ageing in Ptn+/+ female mice (vs 3 months); ‡p < 0.05, ‡‡p < 0.01, ‡‡‡p < 0.001 for differences in the effect of ageing in Ptn−/− female mice (vs 3 months). *p < 0.05; **p < 0.01 for differences between Ptn−/− and Ptn+/+ mice

Full size image

Next, we explored the impact of Ptn deficiency on glucose homeostasis. Although fasting plasma glucose was similar in both groups (Fig. 2c), we found a switch of plasma insulin in Ptn−/− mice, departing from lower insulin levels in young mice to hyperinsulinaemia in 15-month-old animals (Fig. 2d).

GTTs and estimated insulin sensitivity did not reveal any differences in the AUC for glucose at 3 and 6 months of age (Fig. 2e). At 9, 12 and 15 months, however, the AUC for glucose was significantly higher in Ptn−/− than in Ptn+/+ mice, indicating a deterioration of glucose tolerance in Ptn-deficient animals with age. Furthermore, 15-month-old Ptn−/− mice showed a significantly lower QUICKI value than Ptn+/+ control mice (Fig. 2f), suggesting impaired insulin sensitivity in later life.

Deletion of Ptn is associated with differential expression of genes involved in lipid and glucose metabolism in visceral adipose tissue

For gene expression analysis, we focused on periovarian adipose tissue. Compared with wild-type controls, 3-month-old Ptn−/− mice showed a downregulation of mRNA of the Ppar-γ isoforms 1 and 2 (Fig. 3a, b) and their cofactor, Pgc1-α (Fig. 3c). Although the expression of these genes decreased with age in both genotypes (at 12 vs 3 months), the age-related decrease in expression was less pronounced in Ptn−/− mice. These results suggest that the observed Ptn−/− phenotype may be a consequence of defective Ppar-γ activation, causing impaired lipid and glucose homeostasis.

Fig. 3
figure 3

Deletion of Ptn is associated with differential expression of genes involved in lipid and glucose metabolism in periovarian adipose tissue. qPCR analyses of (a) Ppar-γ1, (b) Ppar-γ2, (c) Pgc1-α, (d) Lpl, (e) Glut-4, (f) Cpt1α, (g) Ucp-2, (h) Adr-β3 and (i) Tnf-α in periovarian adipose tissue of 3- and 12-month-old Ptn+/+ (grey bars) and Ptn−/− (blue bars) female mice. Data are mean ± SEM of n = 5 mice/group, except for 12-month-old Ptn−/− mice (n = 4). *p < 0.05, **p < 0.01 vs Ptn+/+ mice of the same age. †p < 0.05, ††p < 0.01, †††p < 0.001 for differences in the effect of ageing in Ptn+/+ (12 vs 3 months)

Full size image

To corroborate this notion, we analysed two pivotal peroxisome proliferator-activated receptor (PPAR)-regulated genes, Lpl and Glut-4 (also known as Slc2a4), that control lipid and glucose metabolism in white adipocytes. In 3-month-old mice, expression of Lpl was significantly lower in Ptn−/− animals than in age-matched controls (Fig. 3d, e). Although both genotypes exhibited an age-related downregulation of Lpl and Glut-4, the reduced expression was more pronounced in Ptn−/− mice. Thus, both Ptn deletion and ageing can cause downregulation of these genes to a similar extent, which is paralleled by changes in Ppar-γ1 and Ppar-γ2 mRNA. In fact, we found a significant positive correlation of the expression of Lpl and Glut-4 with that of Ppar-γ1 and Ppar-γ2 (p < 0.001 for all Pearson coefficients).

As these changes may be related to an altered cellular energy balance, we investigated the mRNA levels of Cpt1α (also known as Cpt1a) and Ucp-2 in periovarian adipose tissue. We found that Cpt1α expression was downregulated by Ptn deletion in 3-month-old mice, whereas Ucp-2 was upregulated (Fig. 3f, g). Although ageing was accompanied by a decrease in Cpt1α and an increase in Ucp-2 expression in Ptn+/+ animals, we found no change in Ptn−/− mice (Fig. 3f, g). Analysis of mRNA of genes related to lipid mobilisation, such as β3-adrenoceptor (Adr-β3, also known as Adrb3), revealed a decreased expression of this receptor in the periovarian adipose tissue of young Ptn−/− animals compared with wild-type controls (Fig. 3h). Interestingly, in Ptn−/− mice, ageing was accompanied by a significant increase in β3-adrenoceptor mRNA, whereas in wild-type controls expression of this receptor decreased with age.

All these changes may be related to a state of low-grade inflammation affecting visceral adipocytes, which is caused by the absence of Ptn. Indeed, mRNA expression of Tnf-α (also known as Tnf) was increased in adipocytes from 3-month-old Ptn−/− mice (Fig. 3i). In wild-type animals, ageing was accompanied by increased Tnf-α expression, but no changes were observed in Ptn−/− mice.

Genetic inactivation of Ptn affects adipocyte size and β-adrenergic-stimulated lipolysis

We found a significant decrease of adipocyte size in periovarian adipose tissue of 15-month-old Ptn−/− mice compared with Ptn+/+ controls (Fig. 4a, b). To further characterise the alteration of adipose tissue by Ptn deletion, catecholamine-stimulated lipolysis was analysed in isolated periovarian adipocytes. Basal lipolysis was significantly lower, and the lipolytic response to isoprenaline significantly higher, in isolated Ptn−/− adipocytes (Fig. 4c). Furthermore, maximum inhibitory actions of insulin on catecholamine-induced lipolysis were significantly lower in Ptn−/− adipocytes compared with wild-type cells (Fig. 4d).

Fig. 4
figure 4

Genetic inactivation of Ptn affects adipocyte size and lipolytic activity after β-adrenergic agonist stimulation. (a) Haematoxylin and eosin staining of formalin-fixed periovarian adipose tissue cryosections; ×10 for top panels (scale bar, 200 μm); ×20 for enlargements below (scale bar 100 μm). (b) Quantification of periovarian adipocyte cell surface area (n = 850 cells for Ptn+/+ and 1160 for Ptn−/− mice). (c) Concentration–response curves of isoprenaline-stimulated lipolysis in isolated adipocytes from 15-month-old Ptn+/+ and Ptn−/− female mice. Data represent lipolytic activities normalised in each experiment to the lipolytic activity in the absence of isoprenaline (basal). Basal lipolysis was 30.7 ± 5.0 and 16.6 ± 2.0 nmol glycerol/mg lipid for Ptn+/+ and Ptn−/−, respectively (p < 0.05). Emax (%) was 189.1 ± 16.8 and 273.2 ± 43.3 for Ptn+/+ and Ptn−/−, respectively (p < 0.05). EC50 was 81.1 ± 28.3 and 407.8 ± 145.8 nmol/l for Ptn+/+ and Ptn−/−, respectively (p < 0.05). (d) Concentration-dependent inhibition of isoprenaline-induced lipolysis by insulin in isolated adipocytes from 15-month-old Ptn+/+ and Ptn−/− female mice. Data represent lipolytic activities normalised in each experiment to the lipolytic value in the absence of insulin (control). Imax (%) was 26.7 ± 2.1 and 16.3 ± 3.1 for Ptn+/+ and Ptn−/−, respectively (p < 0.05). IC50 was 1.47 ± 0.5 and 0.97 ± 0.4 nmol/l for Ptn+/+ and Ptn−/−. Grey lines and bars, Ptn+/+ mice; blue lines and bars, Ptn−/− mice, throughout. (c, d) Data are mean ± SEM of n = 7 mice/group. *p < 0.05, ***p < 0.001 for differences between Ptn−/− and Ptn+/+ mice

Full size image

Metabolic activity under standard and thermoneutral conditions, and BAT activity in Ptn −/− mice

The effect of Ptn deletion on whole-body metabolic activity at 24°C was monitored in metabolic cages in 6-month-old female mice for 3 days, and subsequently for 7 days under thermoneutrality (30°C). Ptn+/+ mice were more active than Ptn−/− mice during the light and dark phases (Fig. 5a), while no differences were observed in total EE in any housing conditions (Fig. 5b). When we calculated the contribution of cold-induced thermogenesis to EE, we found that, at 24°C, the fractions of EE corresponding to cold-induced thermogenesis were higher in Ptn−/− mice than in control animals, both in the light (50.8 ± 3.5% in Ptn−/− vs 39.8 ± 2.1% in Ptn+/+ mice, p < 0.05) and in the dark phase (35.9 ± 1.9% in Ptn−/− vs 27.1 ± 2.3% in Ptn+/+ mice, p < 0.05), with a mean daily contribution of thermogenesis to EE of 42.6% and 33.6% for Ptn−/− and Ptn+/+ mice, respectively. These differences disappeared when the animals were housed at thermoneutrality.

Fig. 5
figure 5

Altered metabolic activity under standard and thermoneutral conditions with Ptn deletion. (a) Physical activity in the X and Y axes, (activity XY) and (b) EE during the light and dark cycle in 6-month old Ptn+/+ and Ptn−/− female mice housed for 3 days at 24°C and 7 days at 30°C. (c) Food intake during the dark and light periods and (d) RER during the dark and light cycles. Grey lines and bars, Ptn+/+ mice; blue lines and bars, Ptn−/− mice. Data are mean ± SEM of the daily intake (3 days at 24°C and 7 days at 30°C) of four mice/group. *p < 0.05, **p < 0.01 vs Ptn+/+ mice at the same temperature

Full size image

Analysis of food intake at 24°C and 30°C did not reveal any statistically significant differences between phenotypes (Fig. 5c). We found, however, that the RER at 24°C was lower in Ptn−/− than in Ptn+/+ mice (Fig. 5d). The analysis of the maximal and minimal RER shows that Ptn+/+ mice consumed carbohydrates as the main source of fuel during the dark period, and a combination of glucose and fatty acids during the light period. Ptn−/− mice were already consuming fatty acids during the dark period, and consumption increased during the light period, indicating that deletion of Ptn increases the use of fatty acids for energy production. In thermoneutral conditions, no differences were found in RER between genotypes during either the dark or the light period (Fig. 5d).

After 7 days housed under thermoneutrality, no significant differences were found in the weight of either the retroperitoneal adipose tissue (0.137 ± 0.013 g for Ptn+/+ vs 0.189 ± 0.018 g for Ptn−/− mice) or the periovarian adipose tissue (0.279 ± 0.028 g in Ptn+/+ vs 0.429 ± 0.134 g in Ptn−/− mice). Moreover, thermoneutrality also reverted the altered lipid profile in Ptn−/− 6-month-old mice. In this condition, no differences were found in circulating triacylglycerols (1.48 ± 0.23 mmol/l in Ptn+/+ vs 1.24 ± 0.21 mmol/l in Ptn−/− mice), and the levels of circulating NEFA were even higher in Ptn−/− mice (1.70 ± 0.05 μmol/l) than in Ptn+/+ animals (1.48 ± 0.16 μmol/l, p < 0.05).

To further assess altered thermogenic function in Ptn−/− mice, thyroid hormones and BAT activity were analysed in 6-month-old female mice. Body temperature was significantly higher (Fig. 6a) and plasma free T4 lower (Fig. 6b) in Ptn−/− than Ptn+/+ mice. An analysis of thyroid hormones in BAT revealed similar T4 levels in the two groups (Fig. 6c), whereas T3 concentration (Fig. 6d) was significantly higher in Ptn−/− mice; this was paralleled by an increased activity and expression in BAT of DIO2, the enzyme responsible for tissue T3 synthesis (Fig. 6e, f). Moreover, expression of mitochondrial UCP-1, responsible for facultative thermogenesis, was increased in the BAT of Ptn−/− mice (Fig. 6g).

Fig. 6
figure 6

Genetic inactivation of Ptn is associated with impaired thyroid hormones and BAT activity in 6-month old mice. (a) Body temperature. (b) Plasma concentration of free T4. Concentration of (c) T4 and (d) T3 in BAT. DIO2 (e) activity and (f) mRNA expression in BAT. (g) UCP-1 in BAT; AU, arbitrary units. Grey bars, Ptn+/+ female mice; blue bars, Ptn−/− female mice. Data are mean ± SEM of n = 14 and n = 12 mice/group (a), n = 7 mice/group (c–f) and n = 6 mice/group (b, g). *p < 0.05, **p < 0.01 for differences between Ptn−/− and Ptn+/+ mice

Full size image

Effect of PTN in brown adipocyte differentiation

We further analysed Ptn mRNA during the differentiation of mBAs in vitro. At day 3, we found a tenfold reduction of Ptn expression compared with undifferentiated cells; this persisted until day 6, when the cells were fully differentiated (Fig. 7a). Treatment of mBAs with rPTN during differentiation had no effect on the expression of endogenous Ptn (Fig. 7b). PTN supplementation, however, significantly decreased the expression of the mBA markers Cidea (20% reduction), Prdm16 (21% reduction) and Pgc1-α (11% reduction) (Fig. 7c–e), reflecting the inhibitory role of PTN in the differentiation of mBAs.

Fig. 7
figure 7

Effect of PTN in mBA differentiation. (a) Ptn expression during differentiation of brown pre-adipocyte tissue in mice. qPCR analyses of (b) Ptn (c) Cidea (d) Prdm16 and (e) Pgc1-α expression in undifferentiated pre-adipocytes (day 0, white bars) and differentiated adipocytes (day 6) in either the absence (blue bars) or presence (purple bars) of 0.1 μg/ml PTN. Data are mean ± SEM of n = 3 experiments performed in duplicate. ***p < 0.001 vs undifferentiated (day 0) adipocytes. ††p < 0.01 for differences in the effect of PTN on differentiation

Full size image

Discussion

Major causes of the high prevalence of metabolic disorders are imbalanced energy metabolism, deranged hormone biology and impaired adipocyte turnover. The characterisation of new targets for the modulation of energy metabolism is of particular interest for the treatment and prevention of metabolic disorders. Here, we provide novel insights into the adipose tissue-specific actions of PTN as a key player in the regulation of energy homeostasis and insulin sensitivity.

We show that Ptn deletion modulates adiposity and fat distribution through adipose tissue lipolytic activity and gene expression. Ptn−/−-deficient mice had altered body weight and metabolic activity, and their age-related increase in body weight was significantly reduced compared with wild-type animals. Although Ptn deletion in 3-month-old mice was associated with a slight increase in visceral adipose tissue depots, the age-related increase in adiposity was significantly reduced in Ptn−/− mice, suggesting a deterioration in adipose tissue expandability, which was further supported by analysis of the total fat mass.

Furthermore, we found alterations in adipocyte cell size and turnover, which are important factors for the development of obesity and metabolic disorders. In fact, in 15-month-old mice, adipocytes from periovarian adipose tissue were smaller in Ptn−/− than in wild-type mice, and this was accompanied by an increase in catecholamine-induced lipolysis. Interestingly, in 12-month-old mice, β3-adrenoceptor expression paralleled the catecholamine-activated lipolytic activity of the tissue, being significantly higher in the periovarian adipose tissue of Ptn−/− mice than in wild-type controls. Thus, the increased flux into the circulation of fatty acids and glycerol, released from adipose tissue in response to increased adrenergic stimulation of lipolysis, may account for the increased de novo synthesis of glucose and impaired glucose tolerance in Ptn−/− mice.

Importantly, the inhibitory response to insulin of catecholamine-stimulated lipolysis in Ptn-deficient periovarian adipocytes was significantly lower than in wild-type cells. These results suggest that Ptn−/− mice develop insulin resistance and/or an amelioration of insulin sensitivity in later life, which is accompanied by higher lipolytic activity and an inability to accumulate fat in white adipose tissue.

As evidenced by increased Tnf-α expression, the periovarian adipose tissue of Ptn−/− mice already exhibits low-grade inflammation at 3 months of age. Tnf-α is a well-known inducer of insulin resistance [27], and locally produced Tnf-α may act within adipose tissue as a potent autocrine and paracrine regulator of diverse metabolic processes [28]. Indeed, Ptn deficiency was associated with decreased Ppar-γ1 and Ppar-γ2 expression in the adipose tissue of young mice. Although rPTN was found to decrease Ppar-γ2 expression in vitro, the same study shows that in vivo injection of a PTN-neutralising antibody slightly decreased Ppar-γ2 levels in adipose tissue [17].

PPAR-γ is essential for adipocyte function, regulating target genes involved in lipid and glucose homeostasis [29]. In particular, PPAR-γ2 prevents peripheral lipotoxicity, promoting adipose tissue expansion and increasing the lipid-buffering capacity of the peripheral organs [30]. Deletion of Ppar-γ2 in obese mouse models reduces adipose tissue expandability, which is associated with severe insulin resistance [31]. Moreover, Smad3 knockout mice that have reduced expression of Ppar-γ2 also show impaired adipogenesis and altered lipid accumulation, caused by a decrease in adipocyte number and size [32]. Here, we show that decreased expression of Ppar-γ and of its coactivator, Pgc1-α, in periovarian adipose tissue is associated with PTN deficiency and may thus play a crucial role in the reduced visceral adiposity, the phenotypic hallmark of the Ptn knockout mouse model (Fig. 8).

Fig. 8
figure 8

Summary of the molecular mechanisms by which PTN deficiency alters the dynamics of adipose lipid turnover and impairs energy metabolism. PGC1-α, peroxisome proliferator-activated receptor γ coactivator 1-α

Full size image

Remarkably, the expression of Lpl, regulated by PPAR-γ, was also downregulated in the periovarian adipose tissue of young Ptn−/− mice; in these animals, Cpt1α mRNA was reduced, whereas Ucp-2 mRNA was increased to levels similar to those found in 12-month-old Ptn+/+ mice. Since Tnf-α has been proposed to play a role in the upregulation of Ucp-2 in broadly distributed tissues including BAT, white adipose tissue and skeletal muscle [33], it is tempting to speculate that deletion of Ptn may be associated with increased inflammation in periovarian adipose tissue, accelerating the age-related alterations in lipid and glucose metabolism, including insulin resistance and a decreased capacity to store lipids (Fig. 8).

We next explored whether Ptn deficiency was associated with the development of whole-body impaired glucose tolerance and insulin resistance. Plasma glucose was similar between groups, suggesting that Ptn deletion does not have any impact on glucose homeostasis. There is evidence, however, that impaired glucose tolerance and insulin resistance may develop under normoglycaemia as a consequence of altered lipid metabolism [34, 35]. In fact, although glucose tolerance and insulin responsiveness were higher in the young 3-month-old Ptn−/− mice, both variables were clearly impaired with ageing, suggesting that Ptn deletion could favour a prediabetic state rendering these animals more prone to developing diabetes mellitus later in life.

We further explored energy homeostasis in the Ptn−/− model and found that Ptn−/− mice consumed a combination of glucose and fatty acids during the dark period and further increased this consumption during the light period, pointing to an augmented energy production from fatty acids. The main oxidative tissue in which fatty acids are used as energy substrates by β-oxidation is skeletal muscle [36]. However, Ptn−/− mice exhibited reduced physical activity, suggesting that fatty acids may be redirected for oxidation in other tissues. In fact, the fraction of EE that accounts for cold-induced thermogenesis was significantly higher in Ptn−/− than Ptn+/+ mice. These results are in good agreement with previous studies showing that, at 22–24°C, mice require over one third of their EE to maintain core body temperature [22]. Since, in Ptn−/− mice, cold-induced thermogenesis represents more than 40% of EE, we looked at alterations in BAT, the tissue accounting for the production of heat by facultative thermogenesis. Brown fat thermogenesis depends on fatty acid utilisation provided by endogenous lipolysis, the mitochondrial machinery involved in fatty translocation and oxidation, and the uncoupling of ATP synthesis by UCP-1. In this regard, we found higher UCP-1 levels in the BAT of Ptn−/− than wild-type mice. Expression and activation of UCP-1 is regulated by fatty acids, cold exposure or noradrenaline (norepinephrine), and T3 is able to increase this stimulation by noradrenaline [37]. Subsequently, the conversion of T4 to T3 by DIO2 is required for the thermogenic function of BAT [38]. Analysis of BAT revealed increased DIO2 activity and expression, together with higher concentrations of T3 in BAT of Ptn−/− mice. Interestingly, these mice also had lower levels of plasma free T4. Thus, in BAT, Ptn deletion may be associated with an increased conversion of T4 to T3 by DIO2, accounting for the increased UCP-1 production and the ensuing elevation of body temperature. This process can be maintained by an increased oxidation of fatty acids as energy substrates, made available from an increased lipolytic activity of adipose tissue despite sufficient glucose availability. Such a preferential oxidation of fatty acids may contribute to the diminished plasma concentration of fatty acids and triacylglycerols, as well as to the reduced adiposity observed in the periovarian and retroperitoneal adipose tissue of Ptn−/− mice.

Further exploring the mechanism of altered thermogenesis in Ptn−/− mice, we found that the differentiation of brown pre-adipocytes (mBAs) blunted Ptn expression, and that treatment of mBAs with rPTN diminished the expression of brown fat markers (Cidea, Prdm16 and Pgc1-α). Thus, although it is known that Ptn expression is suppressed during the differentiation of white adipocytes [16], our results reveal for the first time that BAT differentiation is regulated by changes in Ptn expression and suggest an inhibitory role of this cytokine in brown fat differentiation and thermogenesis.

In conclusion, we propose that the lipodystrophic phenotype of Ptn−/− mice is related to enhanced thermogenesis in BAT that is maintained by active lipid mobilisation from white fat. In support of this hypothesis, we found that when housing animals at thermoneutrality, a condition that blunts thermogenesis, the differences in adiposity and circulating triacylglycerols observed at 24°C disappeared, and NEFA concentrations were even higher in Ptn−/− than control mice. These results, together with the increased RER observed at 30°C in Ptn−/− mice, point to a pronounced inhibition of unrestrained lipolysis at thermoneutrality in Ptn−/− mice.

Conclusion

This is the first study demonstrating that PTN expression is essential to preserve the dynamics of adipose lipid turnover and plasticity. In particular, in visceral and brown fat pads, ablation of Ptn renders mice more prone to developing insulin resistance and/or an amelioration of insulin sensitivity in later life. These conditions are concomitant with altered expansibility of periovarian adipose tissue and impaired thermogenesis of brown adipose cells. The finding that Ptn deficiency and ageing alter the expression profile of a set of genes regulating lipid uptake and utilisation supports the notion that defective PPAR-γ activation accounts for the phenotype of our mouse model, by inducing impairments in lipid and glucose homeostasis (Fig. 8).