Effects of Pioglitazone on Adipose Tissue Remodeling ... - Diabetes

Effects of Pioglitazone on Adipose Tissue Remodeling Within the Setting of Obesity and Insulin Resistance Christopher J. de Souza,1 Michele Eckhardt,1 Karen Gagen,1 Mei Dong,1 Wei Chen,1 Didier Laurent,2 and Bryan F. Burkey1

Obesity and dysfunctional energy partitioning can lead to the development of insulin resistance and type 2 diabetes. The antidiabetic thiazolidinediones shift the energy balance toward storage, leading to an increase in whole-body adiposity. These studies examine the effects of pioglitazone (Pio) on adipose tissue physiology, accumulation, and distribution in female Zucker (fa/fa) rats. Pio treatment (up to 28 days) decreased the insulin-resistant and hyperlipidemic states and increased food consumption and whole-body adiposity. Magnetic resonance imaging (MRI) analysis and weights of fat pads demonstrated that the increase in adiposity was not only limited to the major fat depots but also to fat deposition throughout the body. Adipocyte sizing profiles, fat pad histology, and DNA content show that Pio treatment increased the number of small adipocytes because of both the appearance of new adipocytes and the shrinkage and/or disappearance of existing mature adipocytes. The remodeling was time dependent, with new small adipocytes appearing in clusters throughout the fat pad, and accompanied by a three- to fourfold increase in citrate synthase and fatty acid synthase activity. The appearance of new fat cells and the increase in fat mass were depot specific, with a rank order of responsiveness of ovarian > retroperitoneal > subcutaneous. This differential depot effect resulted in a redistribution of the fat mass in the abdominal region such that there was an increase in the visceral:subcutaneous ratio, as confirmed by MRI analysis. Although the increased adiposity is paradoxical to an improvement in insulin sensitivity, the quantitative increase of adipose mass should be viewed in context of the qualitative changes in adipose tissue, including the remodeling of adipocytes to a smaller size with higher lipid storage potential. This shift in energy balance is likely to result in lower circulating free fatty acid levels, ultimately improving insulin sensitivity and the metabolic state. Diabetes 50:1863–1871, 2001

From the 1Department of Metabolic and Cardiovascular Diseases and 2Core Technologies, Novartis Institute of Biomedical Research, Novartis Pharmaceutical, Summit, New Jersey. Address correspondence and reprint requests to Christopher J. de Souza, Novartis Institute of Biomedical Research, 556 Morris Ave., Summit, NJ 07901-1398. E-mail: [email protected]. Received for publication 10 October 2000 and accepted in revised form 2 May 2001. ANOVA, analysis of variance; aSC, abdominal region of the subcutaneous fat; FFA, free fatty acid; IBAT, interscalpular brown adipose tissue; KRPH, Krebs-Ringer phosphate-HEPES; MRI, magnetic resonance imaging; OV, ovarian; Pio, pioglitazone; PPAR, peroxisome proliferator–activated receptor; RP, retroperitoneal; SC, subcutaneous; SNK, Student-Newman-Keuls; TZD, thiazolidinedione; V, visceral. DIABETES, VOL. 50, AUGUST 2001

E

xcessive adipose accumulation is a key pathological contributor to insulin resistance, which, in the presence of dysfunctional ␤-cell insulin secretion, leads to type 2 diabetes. Adipose tissue serves not only as a depot for triglyceride but also as a dynamic endocrine organ involved in the control of energy balance. Evidence from human and animal studies substantiate the importance of adipose tissue, where too little, as in lipodystrophy (1), or too much, as in obesity (2,3), leads to insulin resistance and diabetes. In such cases, restoration of normal adipose levels results in an alleviation of the insulin-resistant state (4,5). The recent development of thiazolidinediones (TZDs) as insulin sensitizers presents a new line of therapy for the treatment of type 2 diabetes. At the molecular level, these agents are ligands for peroxisome proliferator–activated receptor (PPAR)-␥, a nuclear hormone receptor expressed predominantly in adipose tissue. The TZDs have been shown to enhance insulin action and improve glycemic control by increasing peripheral glucose disposal and reducing hepatic glucose output. Although the exact mechanism(s) by which these agents exert their antidiabetic effects is still not fully understood, modulations of adipose tissue energy balanced via PPAR-␥ activation is believed to be a key contributor. Evidence that TZDs directly affect adipose tissue includes enhanced differentiation of preadipocytes into mature adipocytes and the regulation of gene expression in adipose tissue leading to the coordinated regulation of lipid metabolism (6,7). The antidiabetic efficacy of the TZDs correlates well with their rank order of binding affinity to PPAR-␥ (8). Hence, it is reasonable to infer that most of the antidiabetic effects of TZDs result from PPAR-␥–mediated regulation of adipocyte gene expression and the subsequent improvement in adipose physiology. In humans, insulin resistance and dyslipidemia are most closely associated with increased visceral adipose tissue mass (3,9 –11). The factors that underlie the pathophysiology associated with android body fat distribution are complex and controversial. Regardless, there are numerous reports demonstrating regional adipose depot differences in gene regulation (12–15), lipolysis (16,17), response to weight loss (18,19), and even metabolic response to feeding (20). In each case, these differences in regional adipose physiology are implicated as potential factors involved in the pathogenesis of insulin resistance. In animal studies, TZDs increase body weights due largely to increases in fat 1863

PIOGLITAZONE AND ADIPOSE REMODELING

pad mass (21,22) and alteration of adipocyte size and number (23,24). Similar increases in body weight have been reported in humans after 26 weeks of monotherapy with rosiglitazone or pioglitazone (Pio) (25,26). As in animals, the body weight increase in humans is due in part to an increase in adipose tissue mass (27,28). Equally important are the observations in humans that troglitazone and Pio treatment results in a redistribution of fat from the metabolically deleterious visceral adipose tissue to the more inert subcutaneous (SC) fat depots (27–30). Therefore, in light of the antidiabetic effects of TZDs, which occur in the presence of both an increase and redistribution of adipose mass, these studies were performed to better understand the nature of the TZD-induced adipose remodeling. To do so, female Zucker fa/fa rats were treated with Pio for up to 28 days to assess the quantitative and qualitative effects on white adipose tissue and their relationship to the known antidiabetic effects of this compound. RESEARCH DESIGN AND METHODS Animal studies. Female fatty Zucker (fa/fa) rats and their lean (FA/?) littermates were purchased from Charles River Labs (Wilmington, MA) at 10 weeks of age. All procedures in this study are in compliance with the Animal Welfare Act Regulations 9, Code of Federal Regulation parts 1, 2, and 3, and the Guide for the Care and Use of Laboratory Animals (31). The animals had free access to powdered rodent food (Harlan Teklad Premier Chow, #8604) and water and were maintained at 21 ⫾ 1°C on a 12:12-h light:dark reversephase photo period (lights on from 6:00 P.M. to 6:00 A.M.). Pio was administered as an admixture in rodent food and was set to deliver a dosage of at least 20 mg 䡠 kg–1 䡠 day–1. Obese controls and lean animals received normal powdered rodent food. Body weights and food consumption, corrected for spillage, were measured periodically over the course of the treatment period. Oral glucose tolerance test and tissue analysis. After a 12-h fast, animals were administered a 1.35 g/kg glucose bolus orally, and blood samples were obtained via a tail nick at ⫺10, 0, 15, 30, 45, 60, and 120 min after glucose administration. These experiments commenced at 7:00 A.M., 1 h after light offset, and were performed under the illumination of a red light. After a 2-h recovery period, the animals were anesthetized (50 mg/kg sodium pentobarbital i.p.), and the inguinal SC, retroperitoneal (RP), ovarian (OV) white adipose tissue, and interscalpular brown adipose tissue (IBAT) depots were rapidly removed. These tissues were weighed, and pieces were taken fresh for adipocyte isolation, fixed for histology, or snap-frozen in liquid nitrogen for biochemical analysis. Plasma glucose, insulin, free fatty acid (FFA), and triglyceride levels were assayed as previously described (21). Tissue protein and DNA concentrations, in addition to citrate synthase and fatty acid synthase activity measurements, were performed as previously described (21). Adipocyte number and sizing. Freshly isolated rat adipocytes were prepared by collagenase digestion in Krebs-Ringer phosphate-HEPES (KRPH) buffer containing 5% bovine albumin (Bovuminar Cohn Fraction V; Intergen, Purchase, NY) and 0.6 mmol/l D-glucose, pH 7.4, as previously described (21). The final adipose cell suspension (⬃20% vol/vol) was maintained in KRPH buffer with constant shaking at 40 cycles/min at 37°C. Adipocyte number and size distributions were performed on the freshly isolated adipocytes using a Coulter Counter equipped with a 400-␮m orifice tube, a stirred sample chamber, and a multichannel particle analyzer (Multisizer II; Coulter Electronics, Miami, FL). Adipocytes were diluted into 25 ml electrolyte solution (Isoton II containing 10% glycerol) and stirred to maintain an even suspension. Cell number and size profiles were generated using the siphon option with the manometer set for 2 ml, and adipocyte diameters were distributed across 256 channels. Using the Accucomp software supplied with the Multisizer II, analysis of adipocyte size distributions were performed on cells ranging from 20 to 200 ␮m in diameter, where each channel was expressed as a percentage of the total number of cells within a given size range. Histology and microscopy. White adipose tissue samples were fixed in 10% buffered formalin and embedded in paraffin. Tissue was sectioned (10 ␮m) and stained with hematoxylin and eosin. A total of two to three tissue blocks were processed for each treatment group and surveyed using an unbiased sampling method. Micrographs were taken at 125⫻ magnification, and one representative micrograph per treatment group was selected. 1864

TABLE 1 Plasma hormone and metabolite parameters Parameter Plasma glucose (mg/dl) Plasma insulin (␮U/ml) Plasma FFA (mEq/l)

Plasma triglycerides (mg/dl) ⌬AUC glucose (g 䡠 min⫺1 䡠 dl⫺1) ⌬AUC insulin (mU 䡠 min⫺1 䡠 dl⫺1)

Days† 0 12 19 0 12 19 0 12 19 0 12 19 12 25 12 25

Lean

Control

Pio

93 ⫾ 10 132 ⫾ 6 130 ⫾ 5 112 ⫾ 11 140 ⫾ 6 128 ⫾ 7 ND 147 ⫾ 7 110 ⫾ 8* 22 ⫾ 4* 686 ⫾ 120 735 ⫾ 68 12 ⫾ 7* 764 ⫾ 107 274 ⫾ 50* ND 864 ⫾ 106 196 ⫾ 23* 1.03 ⫾ 0.15* 2.49 ⫾ 0.30 2.72 ⫾ 0.22 1.80 ⫾ 0.23 3.35 ⫾ 0.42 0.57 ⫾ 0.08* ND 3.56 ⫾ 0.40 0.60 ⫾ 0.08* 122 ⫾ 21* 587 ⫾ 32 656 ⫾ 20 188 ⫾ 41* 649 ⫾ 34 176 ⫾ 35* ND 648 ⫾ 31 128 ⫾ 20* 4.90 ⫾ 0.98* 11.28 ⫾ 0.74 6.03 ⫾ 1.33* 2.64 ⫾ .83* 8.98 ⫾ 1.09 0.98 ⫾ 1.31* 1.45 ⫾ 0.47 4.21 ⫾ 1.64 5.42 ⫾ .80 2.64 ⫾ 0.83* 8.98 ⫾ 1.09 0.39 ⫾ 1.50*

Data are means ⫾ SE. Plasma levels are from fed animals while the oral glucose tolerance test was performed after a 12-h fast. *P ⬍ 0.05 vs. same-day control, as determined by ANOVA and the SNK test; †days of treatment; ‡incremental area under the curve (AUC) from 0 to 120 min after an oral glucose load (1.35 g/kg) and calculated using the trapezoid method. ND, no data. Magnetic resonance imaging. Assessments of body fat were performed by magnetic resonance imaging (MRI) on a Bruker Avance 3.0 T/60 cm wide-bore instrument (Bruker Medical, Billerica, MA) equipped with a 12-cm i.d. actively shielded gradient insert. Radio frequency pulses were transmitted, and the nuclear magnetic resonance signal was acquired with either a homogeneous birdcage Bruker resonator (i.d. 72-mm) or a custom-built Alderman-Grant resonator (i.d. 94-mm). Contiguous transversal slices were obtained along the rat length using a turbo-spin echo technique with 32 echoes per excitation and 128 phase-encoding steps. The imaging sequence was optimized for short echo time (echo time 6 ms) and short repetition time (repetition time 250 ms) to allow for signal suppression from other tissues than fat. The spatial resolution in plane varied with the rat size from 625 to 859 ␮m2 (2). Usually, 2-mm thick slices were obtained in three to four blocks (30 slices/block) to cover the whole body in ⬃30 min of total acquisition time. Two-dimensional image series were then imported into IDL software (IDL, Boulder, CO) for pixel counting– based determination of fat volumes. Fat distribution was determined by manually outlining the visceral fat along the well-defined mesenteric line in each slice within the abdominal region. To assess regional changes in fat mass, the whole-body fat mass was segmented into total, intra-abdominal, or visceral; total SC; SC fat in the abdominal region; and thoracic fat in the region anterior to the diaphragm, which was mainly SC. A signal threshold was used after applying a Gauss filter, a quantization classifier, and a class select interaction to exclude all nonfat tissues in each slice. A density factor of 0.9 g/ml was used to convert fat volume (milliliters) into fat mass (grams). The MRI technique for determining total body fat mass was validated by comparing data in lean and obese fa/fa animals (linear regression, slope ⫽ 0.9 and r2 ⫽ 0.98) using both MRI and the tritiated water dilution technique, as previously described (32). Statistical analysis. Unless otherwise stated, data are presented as means ⫾ SE. Statistical analysis were performed using either a Student’s t test or a one-way analysis of variance (ANOVA) with a Student-Newman-Keuls (SNK) test for post-ANOVA comparisons using SigmaStat v2.0 (Jandel Scientific, San Rafael, CA).

RESULTS

The antidiabetic effects of Pio were assessed in obese female fa/fa Zucker rats (Table 1). When compared with their lean littermates (lean), the obese control rats (control) were extremely insulin-resistant, mildly hyperglycemic, and hyperlipidemic. Treatment of obese rats with Pio resulted in a marked reduction in insulin resistance and greatly decreased plasma insulin, FFA, and triglyceride levels. This reduction was sustained or enhanced over the treatment period. Similarly, glucose tolerance, following DIABETES, VOL. 50, AUGUST 2001

C.J. DE SOUZA AND ASSOCIATES

Zucker rats, the weight of the IBAT, OV, SC, and RP fat pads were significantly heavier than those of the lean animals. Administration of Pio for 14 days resulted in a 130% increase in the mass of the IBAT, a 20% increase in the OV fat pad, and no change in either SC or RP white fat pad mass (Fig. 2A). After 27 days of Pio treatment, there was a further increase in the weights of the IBAT (240%) and OV fat pads (34%), along with a 50% increase in the weight of the RP. In contrast, the weight of the SC white fat pads remained unchanged (Fig. 2B). The increased adiposity was confirmed by the MRI measurement where the fa/fa rats treated with Pio for 28 days had a greater total body fat mass, along with a decreased lean body mass (Table 2). On a regional basis (Fig. 3), the fat mass increase in the thoracic region of the SC fat mass was greatest at 33% (125 ⫾ 9 vs. 166 ⫾ 7 g for control and Pio, respectively), followed by a 31% increase in the visceral fat mass. The smallest increase occurred in the abdominal region of the SC fat (aSC) at 13%. Consequently, there was a redistribution of fat from the SC mass in the aSC to the intra-abdominal visceral mass (V), as indicated by a greater V:aSC value (Table 2). Markers of adipose tissue cellularity and metabolic activity were measured as a way to assess adipose tissue composition. In the OV fat depots of obese control ani-

FIG. 1. Effects of 19 days of treatment on body weight gains (A) and daily food consumption (B). Obese animals were administered Pio (f) as a food admixture at a daily dose of ⬃20 mg/kg, whereas the obese controls (Œ) and lean animals (E) received regular rodent food. Food efficiency (inset) was calculated as the total weight gain over the first 10-day treatment period divided by the total amount of food consumed during that period. Data are expressed as means ⴞ SE, with n ⴝ 6 lean and 17 obese animals. P < 0.05 vs. control (Cont) (ⴱ) and lean (#) as determined by ANOVA and the SNK test.

an oral glucose load, was markedly improved after 12 and 25 days of treatment. At the start of the experiment, average body weights in the control and the Pio groups were comparable (358 ⫾ 12 vs. 369 ⫾ 7 g, respectively) and greater than those in the lean group (181 ⫾ 3 g). Over the course of a 19-day treatment period, body weights increased in all three groups. However, body weight gain in the control group was further enhanced with the administration of Pio (Fig. 1A). At the end of 19 days of treatment, the lean group had gained 20 ⫾ 2 g, the control group 78 ⫾ 4 g, and the Pio group 136 ⫾ 5 g. In keeping with the Pio-mediated increase on body weight, food consumption was also increased in the Pio group (Fig. 1B)— up by 52% at day 19. When the higher body weights of the Pio group were factored in, food consumption per kilogram per day was still 30% higher relative to the control group (75 ⫾ 4 vs. 58 ⫾ 2 g 䡠 kg–1 䡠 day–1 for Pio and control; P ⬍ 0.05, as determined by the Student’s t test). The calculated feed efficiency (Fig. 1B inset) was 70% higher in the Pio group—an effect indicative of a greater body weight gain for the amount of food consumed (i.e., greater metabolic efficiency). In keeping with the known obese phenotype of fa/fa DIABETES, VOL. 50, AUGUST 2001

FIG. 2. Fat pad weights after 14 (A) and 27 (B) days of treatment with Pio delivered as a food admixture at a daily dose of ⬃20 mg/kg. The controls received normal rodent food. Data are expressed as means ⴞ SE, with n ⴝ 6 –7 animals. *P < 0.05 vs. control and lean groups, respectively, as determined by ANOVA and the SNK test. 1865

PIOGLITAZONE AND ADIPOSE REMODELING

TABLE 2 Assessment of whole-body adiposity and distribution of regional adiposity using MRI Parameter Body weight (g) Total fat mass (g) Non–fat mass (g) % Body weight Fat mass Non–fat mass Regional fat mass (g) V Total SC aSC tSC V/aSC V/SC V/tSC

Control

Pio

P

553 ⫾ 17 356 ⫾ 18 197 ⫾ 5

615 ⫾ 17 449 ⫾ 19 166 ⫾ 8

0.024 0.003 0.005

64 ⫾ 1 36 ⫾ 1

73 ⫾ 1 27 ⫾ 1

0.001 0.001

121 ⫾ 6 235 ⫾ 13 91 ⫾ 3 125 ⫾ 9 1.33 ⫾ 0.06 0.52 ⫾ 0.02 0.98 ⫾ 0.03

159 ⫾ 7 289 ⫾ 14 103 ⫾ 5 166 ⫾ 7 1.57 ⫾ 0.08 0.55 ⫾ 0.02 0.97 ⫾ 0.04

0.001 0.012 0.066 0.005 0.030 0.187 0.807

Data are means ⫾ SEM, with n ⫽ 9 controls and 7 Pio. Reported P values are based on the Student’s t test. tSC, thoracic SC.

mals, protein and DNA content per milligram tissue was lower than that in the lean group, consistent with adipose tissue from obese animals having fewer and larger lipidladen adipocytes per milligram tissue (Table 3). In comparing markers of adipose tissue metabolic activity per milligram wet weight, there were no differences in citrate synthase (an indicator of carbon entry into the Krebs’ cycle) and fatty acid synthase (a rate-limiting enzyme in lipogenesis) activity between the obese control and lean groups (Table 3). Pio treatment for 14 days increased protein and DNA levels per milligram tissue to levels comparable to those of the lean group, indicative of a greater number of adipocytes per milligram tissue mass. There was also an increase in the total amount of DNA and protein in the whole OV fat pad. Furthermore, citrate synthase and fatty acid synthase activity per milligram tissue, as well as total activity for the whole depot, were elevated two- to fourfold above levels in the control group (Table 3), reflective of greater lipid metabolism. Significant treatment-induced differences in the size distribution of adipocytes were also observed, and the magnitude and nature of these differences varied between the three white fat depots studied (Table 4 and Figs. 4 and 5). In the control group, size distribution of the adipocytes from all three depots distributed as cell populations with a single peak. The peak diameter of adipocytes from SC tissue (Fig. 4A) was 131 ␮m, smaller than the peaks for RP (Fig. 4B) and OV (Fig. 5A), which were both 147 ␮m in diameter. In the SC depot, after 14 days of treatment with Pio (Fig. 4A), the percentage of adipocytes in the mediumsized 90- to 150-␮m range decreased by 23%, whereas the percentage of small adipocytes in the 40- to 90-␮m range increased by 78%, while the distribution of the largest adipocytes (⬎150 ␮m) remained unchanged (Table 4). In the RP depot (Fig. 4B), treatment with Pio for 14 days resulted in a 38% decrease in the percentage of the largest adipocytes (⬎150 ␮m), an 85% increase in the percentage of the small adipocytes (40 –90 ␮m), with no change in the size distribution of the 90 –150 ␮mol/l medium-sized adipocytes (Table 4). The most dramatic changes occurred in the OV depot (Fig. 5A) in which, after 14 days of treatment with Pio, there was a 71% decrease in the percentage of 1866

adipocytes in the large range and a 35% decrease in the medium-size range, whereas the populations of small adipocytes in the 40- to 90-␮m range expanded by 311% (Table 4). Prolongation of Pio treatment to 28 days resulted in what appears to be a merging of the two peaks back into one population of smaller adipocytes, with a peak cell diameter of 110 ␮m (Fig. 5B). The appearance of the small fat cells is further illustrated in the micrographs of the OV fat pad taken after 14 days of treatment with Pio (Fig. 6). In the controls, there was an even distribution of equal-sized adipocytes throughout the depot (Fig. 6A), whereas in the Pio group, there were large numbers of small fat cells (Fig. 6B). However, these new small fat cells were not evenly distributed throughout the tissue but appeared as discrete clusters or patches of cells. DISCUSSION

In light of the well-documented relationship between obesity and insulin resistance (33,34), the treatment effects of TZDs appear to be paradoxical in that their insulinsensitizing effects occur in the presence of an increase in body weight and whole-body adiposity. However, these conflicting quantitative changes should be assessed in context of the quality of the increased fat mass. Results of our present studies help to explain these seemingly paradoxical effects. While Pio increases fat mass, it does so in part by 1) increasing the number of small adipocytes in fat depots, 2) dramatically increasing the lipogenic activity in white adipose tissue, and 3) causing depot-specific redistribution resulting in a remodeling of the adipose depots— especially the visceral adipose depots. The higher lipid storage capacity of the new, small, insulin-sensitive adipocytes likely serves to lower circulating plasma FFA levels and skeletal muscle triglyceride levels, which, in part, may account for the decrease in skeletal muscle insulin resistance (35). This is especially significant given the known relationship between hyperlipidemia, skeletal muscle lipid content, and insulin resistance (36 –38). Treatment with Pio resulted in an increase in body weights that was evident as early as 3 days after commencement of treatment and was sustained throughout the 28-day treatment period. The Pio-induced increase in body weight was due in part to an accumulation of fat. Given the inverse relationship between obesity and insulin sensitivity, there was no apparent loss in the Pio-induced increase in insulin sensitivity, as seen by the sustained reduction in insulin levels at day 19 (Table 1). Additional studies (data not shown) using the euglycemic-hyperinsulinemic clamp technique showed that the Pio-induced increase in whole-body glucose disposal and decreased hepatic glucose output were sustained at day 28 of treatment. Concomitant with the Pio-induced increase in body weight, there is an increase in food consumption that cannot entirely be accounted for by the increase in body weight because food consumption per kilogram body weight is ⬃27% greater in the Pio group than in the controls after 11 and 19 days of treatment. Given the dramatic increase in IBAT mass in the Pio-treated animals (Fig. 2), it is tempting to attribute the increase in food consumption to an increase in energy expenditure. However, unlike thermogenic compounds, such as ␤3-adrenoDIABETES, VOL. 50, AUGUST 2001

C.J. DE SOUZA AND ASSOCIATES

FIG. 3. Transverse MRI sections from the thoracic (top) and abdominal (bottom) regions of a lean and an obese rat delineating the thoracic SC (tSC), abdominal SC (aSC), and visceral (V) fat mass.

ceptor agonists (39), Pio does not increase whole-body energy expenditure, as demonstrated from indirect calorimetry studies (21). Thus, the increased caloric intake appears to be preferentially mobilized toward energy storage, as indicated by the greater food efficiency in the Pio group (Fig. 1B). This anabolic effect of Pio is further supported at the cellular level by the observations that Pio treatment increases citrate synthase (an indicator of carbon entry into the Krebs’ cycle) and fatty acid synthase activity in the adipocytes, indicative of a greater flux of carbon into lipids (Table 3). With 14 days of Pio treatment, there was no appreciable increase in the mass of the SC and RP fat depots, whereas the OV depot increased by 20% (⬃3.2 g) (Fig. 2A) even though body weights were 11% greater (411 ⫾ 10 vs. 456 ⫾ 10 g, control vs. Pio; P ⬍ 0.05, ANOVA). After 19 days of treatment, the increase in body weights in the Pio-treated DIABETES, VOL. 50, AUGUST 2001

animals was accompanied by further increases in the mass of the OV and RP, but the mass of the inguinal SC white fat depots remained unchanged (Fig. 2B). However, the MRI data, which measures total SC fat, indicate that after 24 days of treatment, there is a clear and significant increase in the whole-body SC fat mass as well as in the SC mass in the abdominal region. This finding and similar observations reported by others (23,24) support the concept that the anabolic effects of Pio are not entirely restricted to the major fat pads but are probably more diffuse in nature. Moreover, the MRI data reveal regional differences in the Pio-induced adiposity, in which greater lipid accumulation occurs in the visceral compared with the abdominal SC fat mass, such that the V:aSC ratio is greater in the Pio-treated animals. Although this increase in visceral fat mass might seem contradictory to the improved insulin sensitivity, the quantitative changes in the fat mass have to be analyzed 1867

PIOGLITAZONE AND ADIPOSE REMODELING

TABLE 3 Biochemical parameters of ovarian fat after 14 days of treatment Parameter Ovarian fat depot (g) Per mg wet tissue wt DNA (ng) Protein (␮g) Citrate synthase (mU) FAS (pmol/min) Per depot DNA (ng) Protein (␮g) Citrate synthase (mU) FAS (pmol/min) Citrate synthase:FAS

Lean

Fatty control

Fatty Pio

1.8 ⫾ 0.3*

16.3 ⫾ 1.5

19.6 ⫾ 11.1*#

64.6 ⫾ 10.0* 16.5 ⫾ 0.7* 0.24 ⫾ 0.02

33.3 ⫾ 2.9 9.4 ⫾ 0.7 0.18 ⫾ 0.01

46.6 ⫾ 2.4# 14.3 ⫾ 0.5* 0.66 ⫾ 0.04*#

29 ⫾ 3

65 ⫾ 4*#

30 ⫾ 1 115 ⫾ 23* 28 ⫾ 4* 409 ⫾ 57* 53 ⫾ 8* 8.0 ⫾ 0.6

554 ⫾ 89 921 ⫾ 87*# 150 ⫾ 12 282 ⫾ 23*# 2,933 ⫾ 211 13,116 ⫾ 1,414*# 455 ⫾ 34 6.5 ⫾ 0.4

1,275 ⫾ 119*# 10.3 ⫾ 0.7*#

Data are means ⫾ SEM, with n ⫽ 7. P ⬍ 0.05 vs. control (*) and lean (#), as determined by one-way ANOVA and the SNK test. FAS, fatty acid synthase.

with respect to the qualitative changes occurring in parallel, i.e., the change in the cell size and insulin sensitivity of this new fat mass. Despite the minimal changes in the quantity of the different white adipose tissue depots after 14 days of treatment, there was a clear qualitative change reflected as an increase in cellularity. In the OV fat pad, total DNA levels as well as levels per milligram tissue were 62 and 40% higher in the Pio-treated rats, respectively, indicative of a greater number of smaller cells in this fat pad (Table 3). This increase in cellularity was further visually confirmed from micrographs of OV fat pad tissue, where clusters of small fat cells are clearly dispersed throughout the fat pad, after 14 days of treatment (Fig. 6). These cellular effects, as demonstrated here and by others (23,24), are in keeping with the known PPAR-␥–mediated adipogenic effects of Pio (40,41). This adipogenic effect, which results in the generation of small fat cells, might help explain the decrease in circulating FFA levels and the improvement in insulin resistance. Relative to large fat cells, smaller fat cells have been shown to take up greater amounts of glucose in the presence of insulin (42) and are more sensitive to the lipogenic effects of insulin (43). The biochemical data (Table 3) also indicate that the fat pads have an enhanced capacity for lipid storage, as evident TABLE 4 Adipose cell size distribution after 14 days of treatment Fat depot SC Control Pio RP Control Pio OV Control Pio

n

Small (40–90 ␮m)

Medium (90–150 ␮m)

Large (⬎150 ␮m)

3 6

18 ⫾ 7 32 ⫾ 2*

71 ⫾ 6 55 ⫾ 1†

11 ⫾ 1 14 ⫾ 1

4 6

20 ⫾ 1 37 ⫾ 3*‡

56 ⫾ 2 48 ⫾ 4

24 ⫾ 2 15 ⫾ 3*

6 6

18 ⫾ 2 56 ⫾ 2‡*

57 ⫾ 3 37 ⫾ 2*‡

24 ⫾ 2 7 ⫾ 1*‡

Data are means ⫾ SEM of the percentage of cells in the 40 –200 ␮mol/l diameter size range. *P ⫽ 0.07, †P ⬍ 0.05, and ‡P ⬍ 0.005 vs. control as determined by the Student’s t test. 1868

FIG. 4. White adipocyte cell size profiles from SC (A) and RP (B) fat depots after 14 days of treatment with either Pio (F, n ⴝ 6) or vehicle (䡺, n ⴝ 3– 4). Freshly isolated adipocytes from these depots were sized using a Coulter Mutisizer II particle analyzer, and the size distribution of cells between 20 and 200 ␮mol/l was expressed as a percentage of the total. Data are expressed as an average combined profile, fit with an adjacent-averaged curve.

from the large increase in fatty acid synthase, a ratelimiting enzyme in the biosynthesis of lipids. In parallel, the activity of citrate synthase is also increased, indicative of a greater flux of two-carbon intermediates into the Krebs cycle and the generation of energy via oxidative phosphorylation to support the increased rates of fatty acid biosynthesis. Furthermore, treatment of diabetic rats with a PPAR-␥ agonist for 7 days was found to result in a coordinate regulation of genes that promote lipid storage in the adipose tissue, lipid depletion, and glucose utilization in the skeletal muscle and decrease gluconeogenesis in the liver (44). The magnitude of the adipogenic effects appears to be depot specific, as demonstrated from fat cell size distribution profiles of freshly isolated adipocytes using a Coulter counter. This instrument provides a unique methodology for the assessment of cell size and number by sorting aliquots of freshly isolated adipocytes according to their size over a range of 20 –200 ␮m in diameter. The instrument is capable of generating cell size distribution profiles with a high degree of precision. In studies measuring mean fat cell size in normal and high fat–fed rats, the interanimal variability (n ⫽ 16), measured as the percent DIABETES, VOL. 50, AUGUST 2001

C.J. DE SOUZA AND ASSOCIATES

FIG. 5. White adipocyte cell size profiles from OV fat depots after 14 (A) and 28 (B) days of treatment with either Pio (F, n ⴝ 6) or vehicle (䡺, n ⴝ 6). Freshly isolated adipocytes from these depots were sized using a Coulter Mutisizer II particle analyzer, and the size distribution of cells between 20 and 200 ␮mol/l was expressed as a percentage of the total. Data are expressed as an average combined profile, fit with an adjacent-averaged curve.

coefficient of variation, was ⬍1% (B.F. Burkey, unpublished data). In the studies reported herein, the distribution of fat cells from obese controls was unimodal, with a single peak at 137 ␮m in the SC depots and 147 ␮m in the RP and OV depots (Figs. 4 and 5). A similar unimodal distribution of fat cells is observed in lean animals but with a peak at 85–90 ␮m in the RP and OV fat pads (data not shown). In keeping with its known adipogenic effects, 14 days of treatment with Pio induced the appearance of a new peak in the 70-␮m range, indicative of a generation of new, small fat cells. However, there were clear depotspecific differences in the adipogenic response to Pio, with a rank order of responsiveness of OV ⬎ RP ⬎ SC. This differential response might, in part, be a reflection of the depot-specific differences in PPAR-␥ adipogenic responsiveness to TZDs because expression of the receptor is not known to differ between adipose depots (12,45). Although Pio stimulated adipogenesis, it also resulted in a decrease in the number of large fat cells in the OV and RP depots. This decrease in large fat cell numbers may be a reflection of an increase in the efflux of lipids from these insulin-resistant adipocytes in the presence of lower circulating insulin levels. Once again, there was a differential DIABETES, VOL. 50, AUGUST 2001

depot-specific response of these large fat cells to Pio, which might in part reflect depot-specific differences in insulin responsiveness. As insulin levels decline with Pio treatment, the large insulin-resistant adipocytes from one depot would be expected to release lipids and shrink faster than its counterpart in a less insulin-responsive depot. The decrease in the number of large adipocytes might also reflect a TZD-induced increase in mature adipocyte apoptosis given that PPAR-␥ agonists have been demonstrated to induce apoptosis in adipocytes (23) and other tissues (46,47). Furthermore, depot-specific differences in gene expression susceptibility to apoptosis have also been reported. Expression of cellular inhibitor of apoptosis protein-2 is higher in human omental fat than SC fat (13), and the omental fat is more susceptible to apoptotic stimuli than SC fat (48). Given the known detrimental relationship between hyperlipidemia and peripheral tissue insulin resistance, the qualitative and quantitative changes in adipose tissue reported here could also account for the Pio-induced decrease in skeletal muscle insulin resistance (35) by decreasing circulating levels of lipids (49). This decrease in plasma lipid levels might also account for the TZDinduced improvement in ␤-cell function (50,51) because chronic hyperlipidemia has been demonstrated to impair ␤-cell function (52–54). It must be noted that although the adipose tissue remodeling reported in these studies and by others (30) might explain the improvements in insulin sensitivity after chronic treatment with TZDs, they cannot account for the improvement in glycemic control that has been reported to occur within as little as 12 h after commencement of treatment with these agents (55). In humans, similar restructuring of adipose depots have been observed, but the shift in body fat appears to be inverse to that seen in animal studies. Treatment with troglitazone or Pio results in an increase in whole-body adiposity, but the shift in distribution of fat is such that the visceral fat decreases while the SC fat increases and the V:aSC ratio decreases (28 –30). One factor that could account for this species difference is food intake. The rodents in this study were fed ad libitum, which, coupled with the increase food efficiency, results in the accumulation of large amounts of fat not seen in the human studies where patients are given nutritional guidelines to follow. Hence, whereas the lipid accumulation effects of the treatment of patients with TZDs was greater in the SC fat depot than visceral fat depot, the nature of these quantitative changes are yet to be assessed. Based on the results of our animal studies, accompanying the quantitative changes, one would also expect to see qualitative changes in the nature of these adipose stores resulting in a more insulin-sensitive depot. Whether the beneficial effects of adipogenesis will persist with prolonged TZD treatment or whether increased adipocyte storage of lipids will eventually result in an overall increased population of larger insulin-resistant adipocytes remains to be determined. In conclusion, these studies demonstrate that whereas body weights and/or whole-body adiposity and adipose distribution remain important determinants in the pathology of obesity and insulin resistance, the quality and 1869

PIOGLITAZONE AND ADIPOSE REMODELING

FIG. 6. Ovarian white adipose tissue morphology in obese controls (A) and in obese animals treated with Pio for 14 days (B). Pieces of fat samples from several areas of the OV fat depot were fixed, H&E stained, and sectioned, and micrographs were taken at a magnification of ⴛ50. Representative sections from two animals are shown.

nature of the fat is also essential in determining the metabolic risk of increased whole-body adiposity. ACKNOWLEDGMENTS

The authors thank Sherry De Vito and Gregory Argentieri for their excellent assistance with the white fat histology and Kevin Poirier and Jim Wasvary for assistance with the MRI measurements. 1870

REFERENCES 1. Seip M, Trygstad O: Generalized lipodystrophy, congenital and acquired (lipoatrophy). Acta Paediatr Suppl 413:2–28, 1996 2. Wickelgren I: Obesity: how big a problem? Science 280:1364 –1367, 1998 3. Brunzell JD, Hokanson JE: Dyslipidemia of central obesity and insulin resistance. Diabetes Care 22 (Suppl. 3):C10 –C13, 1999 4. Murphy JE, Zhou S, Giese K, Willams LT, Escobedo JA, Dwarki VJ: Long-term correction of obesity and diabetes in genetically obese mice by a single intramuscular injection of recombinant adeno-associated virus encoding mouse leptin. Proc Natl Acad Sci USA 94:13921–13926, 1997 DIABETES, VOL. 50, AUGUST 2001

C.J. DE SOUZA AND ASSOCIATES

5. Gavrilova O, Marcus-Samuels B, Graham D, Kim JK, Shulman GI, Castle AL, Vinson C, Eckhaus M, Reitman ML: Surgical implantation of adipose tissue reverses diabetes in lipoatrophic mice. J Clin Invest 105:271–278, 2000 6. Lowell BB: PPAR-␥: an essential regulator of adipogenesis and modulator of fat cell function. Cell 99:239 –242, 1999 7. Brun RP, Tontonoz P, Forman BM, Ellis R, Chen J, Evans RM, Spiegelman BM: Differential activation of adipogenesis by multiple PPAR isoforms. Genes Dev 10:974 –984, 1996 8. Willson TM, Cobb JE, Cowan DJ, Wiethe RW, Correa ID, Prakash SR, Beck KD, Moore LB, Kliewer SA, Lehmann JM: The structure-activity relationship between peroxisome proliferator-activated receptor gamma agonism and the antihyperglycemic activity of thiazolidinediones. J Med Chem 39:665– 668, 1996 9. Montague CT, O’Rahilly S: The perils of portliness: causes and consequences of visceral adiposity. Diabetes 49:883– 888, 2000 10. Despres J-P: The insulin resistance-dyslipidemic syndrome of visceral obesity: effect on patients risk. Obes Res 6:8S–17S, 1998 11. Gautier JF, Mourier A, de Kerviler E, Tarentola A, Bigard AX, Villette JM, Guezennec CY, Cathelineau G: Evaluation of abdominal fat distribution in noninsulin-dependent diabetes mellitus: relationship to insulin resistance. J Clin Endocrinol Metab 83:1306 –1311, 1998 12. Lefebvre AM, Laville M, Vega N, Riou JP, van Gaal L, Auwerx J, Vidal H: Depot-specific differences in adipose tissue gene expression in lean and obese subjects. Diabetes 47:98 –103, 1998 13. Montague CT, Prins JB, Sanders L, Zhang J, Sweter CP, Digby J, Byrne CD, O’Rahilly S: Depot-related gene expression in human subcutaneous and omental adipocytes. Diabetes 47:1384 –1391, 1998 14. Marette A, Mauriege P, Marcotte B, Atgie C, Bouschard C, Theriault G, Bukaoiecki LJ, Marceau P, Biron S, Madeau A, Despres JP: Regional variation in adipose tissue insulin action and GLUT4 glucose transporter expression in severely obese premenopausal women. Diabetologia 40:590 –598, 1997 15. Fried SK, Bunkin DA, Greenberg AS: Omental and subcutaneous adipose tissues of obese subjects release interleukin-6: depot differences and regulation by glucocorticoid. J Clin Endocrinol Metab 83:847– 850, 1998 16. Richelsen B, Pedersen SB, Moller-Pedersen T, Bak JF: Regional differences in triglyceride breakdown in human adipose tissue: effects of catecholamines, insulin, and prostaglandin E2. Metabolism 40:990 –996, 1991 17. Arner P: Differences in lipolysis between human subcutaneous and omental adipose tissues. Ann Med 27:435– 438, 1995 18. Wirth A, Steinmetz B: Gender differences in changes in subcutaneous and intra-abdominal fat during weight reduction: an ultrasound study. Obes Res 6:393–399, 1998 19. Janssen I, Ross R: Effects of sex on the change in visceral, subcutaneous adipose tissue and skeletal muscle in response to weight loss. Int J Obes Relat Metab Disord 23:1035–1046, 1999 20. Armellini F, Zamboni M, Mino A, Bissoli L, Micciolo R, Bosello O: Postabsorptive resting metabolic rate and thermic effect of food in relation to body composition and adipose tissue distribution. Metabolism 49:6–10, 2000 21. Burkey BF, Dong M, Gagen K, Eckhardt M, Dragonas N, Chen W, Grosenstein P, Argentieri G, de Souza CJ: Pioglitazone promotes energy storage and not expenditure in brown adipose tissue of obese fa/fa Zucker rats: comparison with CL 316,243. Metabolism 49:1301–1308, 2000 22. Wang Q, Dryden S, Frankish HN, Bing C, Pickavance L, Hopkins D, Buckingham R, Willams G: Increased feeding in fatty Zucker rats by the thiazolidinedione BRL 49653 (rosiglitazone) and the possible involvement of leptin and hypothalmic neuropeptide Y. Br J Pharm 122:1405–1410, 1997 23. Okuno A, Tamemoto H, Tobe K, Ueki K, Mori Y, Iwamoto K, Umesono K, Akanuma Y, Fujiwara T, Horikoshi H, Yazaki Y, Kadowaki T: Troglitazone increases the number of small adipocytes without the change of white adipose tissue mass in obese Zucker rats. J Clin Invest 101:1354–1361, 1998 24. Hallakou S, Doare´ L, Foufelle F, Kergoat M, Guerre-Millo M, Berthault MF, Dugail I, Morin J, Auwerx J, Ferre´ P: Pioglitazone induces in vivo adipocyte differentiation in the obese Zucker fa/fa rat. Diabetes 46:1393–1399, 1997 25. Takeda: Actose Prescribing Information 26. SmithKlein Beecham: Avandia Prescribing Information 27. Kelly IE, Walsh K, Han TS, Lean MEJ: Effects of a thiazolidinedione compound on body fat and fat distribution of patients with type 2 diabetes. Diabetes Care 22:288 –293, 1999 28. Miyazaki M, Mahankali A, Matsuda M, Mahankali S, Cusi K, Mandarino L, DeFronzo RA: Effects of pioglitazone on abdominal fat distribution and insulin sensitivity in patients with type 2 diabetes mellitus (T2DM) (Abstract). Diabetes 49 (Suppl. 1):A1245 29. Kawai T, Takei I, Oguma Y, Ohashi N, Tokui M, Oguchi S, Katsukawa F, Hirose H, Shimada A, Watanabe K, Saruta T: Effects of troglitazone on fat distribution in the treatment of male type 2 diabetes. Metabolism 48:1102– 1107, 1999

DIABETES, VOL. 50, AUGUST 2001

30. Mori Y, Murakawa Y, Okada K, Horikoshi H, Yokoyama J, Tajima N, Ikeda Y: Effects of troglitazone on body fat distribution in type 2 diabetic patients. Diabetes Care 22:908 –912, 1999 31. Guide for the Care and Use of Laboratory Animals. Washington, DC, U.S. Printing Office, Dept. of Heath, Education, and Welfare Publication (National Institutes of Health), 1995, 85-23 32. Culebas JM, Moore FD: Total body water and the exchangeable hydrogen. I. Theoretical calculation of nonaqueous exchangeable hydrogen in man. Am J Physiol 232:R54 –R59, 1977 33. Friedman JM, Leibel RI: Tackling a weighty problem. Cell 69:217–220, 1992 34. Moller DE, Flier JS: Insulin resistance: mechanisms, syndromes and implications. N Engl J Med 325:938 –948, 1991 35. Sugiyama Y, Shimura Y, Ikeda H: Effects of pioglitazone on hepatic and peripheral insulin resistance in Wistar fatty rats. Arzneimittelforschung 40:436 – 440, 1990 36. Boden G: Perspectives in diabetes: role of fatty acids in the pathogenesis of insulin resistance and NIDDM. Diabetes 46:3–10, 1997 37. Pan DA, Lillioja S, Kriketos AD, Milner MR, Baur LA, Bogardus C, Jenkins AB, Storlien LH: Skeletal muscle triglyceride levels are inversely related to insulin action. Diabetes 46:983–988, 1997 38. Manco M, Mingrone G, Greco AV, Capristo E, Gniuli D, De Gaetano A, Gasbarrini G: Insulin resistance directly correlates with saturated fatty acids in skeletal muscle triglyceride. Metabolism 49:220 –224, 2000 39. Arch JRS, Wilson S: Prospects for ␤3-adrenoceptor agonists in the treatment of obesity and diabetes. Int J Obes Relat Metab Disord 20:191–199, 1996 40. Schoonjans K, Staels B, Auwerx J: The peroxisome proliferator activated receptors (PPARs) and their effects on lipid metabolism and adipocyte differentiation. Biochim Biophys Acta 1302:93–109, 1996 41. Spiegelman BM: PPAR-gamma: adipogenic regulator and thiazolidinedione receptor. Diabetes 47:507–514, 1998 42. Olefsky JM: The effects of spontaneous obesity on insulin binding, glucose transport, and glucose oxidation of isolated rat adipocytes. J Clin Invest 57:842– 851, 1976 43. Abbott WG, Foley JE: Comparisons of body composition, adipocyte size and glucose and insulin concentrations in Pima Indians and Caucasian children. Metabolism 36:576 –579, 1987 44. Willson TM, Way J, Mansfield T, Brown KK, Kiewer SA: Coordinate regulation of gene expression in multiple insulin-sensitive tissues by a PPAR␥ agonist. Diabetologia 43 (Suppl. 1):A24, 2000 45. Adams M, Montague CT, Prins JB, Holder JC, Smith SA, Sanders L, Digby JE, Sewter CP, Lazar MA, Chatterjee VKK, O’Rahilly S: Activators of peroxisome proliferator-activated receptor ␥ have depot-specific effects on human preadipocyte differentiation. J Clin Invest 100:3149 –3153, 1997 46. Bishop-Bailey D, Hla T: Endothelial cell apoptosis induced by the peroxisome proliferator-activated receptor (PPAR) ligand 15-deoxy-Delta12, 14prostaglandin J2. J Biol Chem 274:17042–17048, 1999 47. Chinetti G, Griglio S, Antonucci M, Torra IP, Delerive P, Majd Z, Fruchart JC, Chapman J, Najib J, Ataels B: Activation of proliferator-activated receptors alpha and gamma induces apoptosis of human monocyte-derived macrophages. J Biol Chem 273:25573–25580, 1998 48. Niesler CU, Siddle K, Prins JB: Human preadipocytes display a depotspecific susceptibility to apoptosis. Diabetes 47:1365–1368, 1998 49. Ikeda H, Taketomi S, Sugiyama Y, Shoda Y, Shoda T, Meguro K, Fujita T: Effects of pioglitazone on glucose and lipid metabolism in normal and insulin resistant animals. Arzneimittelforschung 40:156 –162, 1990 50. de Souza CJ, Yu JH, Robinson DD, Ulrich RG, Meglasson MD: Insulin secretory defect in Zucker fa/fa rats is improved by ameliorating insulin resistance. Diabetes 44:984 –991, 1995 51. Cavaghan MK, Ehrmann DA, Byrne MM, Polonsky KS: Treatment with the oral antidiabetic agent troglitazone improves ␤ cell responses to glucose in subjects with impaired glucose tolerance. J Clin Invest 100:530 –537, 1997 52. Sako Y, Grill VE: A 48-hour lipid infusion in the rat time-dependently inhibits glucose-induced insulin secretion and B cell oxidation through a process likely coupled to fatty acid oxidation. Endocrinology 127:1580 – 1589, 1990 53. Lee Y, Hirose H, Ohneda M, Johnson JH, McGarry JD, Unger RH: ␤-Cell lipotoxicity in the pathogenesis of non-insulin– dependent diabetes mellitus of obese rats: impairment in adipocyte-␤-cell relationship. Proc Natl Acad Sci USA 91:10878 –10882, 1994 54. McGarry JD, Dobbins RL: Fatty acids, lipotoxicity and insulin secretion. Diabetologia 42:128 –138, 1999 55. Colca JR, Wyse BM, Sawada G, Jodelis KS, Connell CL, Fletcher-McGruder BL, Palazuk BJ, Diani AR: Ciglitazone, a hypoglycemic agent: early effects on the pancreatic islets of ob/ob mice. Metabolism 37:276 –280, 1988

1871