Abstract
Peroxisome proliferator-activated receptor (PPAR)-γ modulators, a class of antidiabetic drugs, have been associated with cardiovascular risks in type 2 diabetes in humans. The objective of this study was to explore possible cardiovascular risk biomarkers associated with PPAR-γ in rodents that could provide an alert for risk to humans. Normal, myocardial infarction-induced heart failure (HF) or Zucker diabetic fatty (ZDF) rats were used. Rats (n = 5–6) were treated with either vehicle or rosiglitazone (RGZ; 3 or 45 mg/kg/day p.o.) for 4 weeks. Biomarkers for potential cardiovascular risks were assessed, including 1) ultrasound for cardiac structure and function; 2) neuroendocrine and hormonal plasma biomarkers of cardiovascular risk; 3) pharmacogenomic profiling of cardiac and renal tissue by targeted tissue low-density gene array representing ion channels and transporters, and components of the renin-angiotensin-aldosterone system; and 4) immunohistochemistry for cardiac fibrosis, hypertrophy, and inflammation (macrophages and tumor necrosis factor-α). HF was confirmed by increase in cardiac brain natriuretic peptide expression (p < 0.01) and echocardiography. Adequate exposure of RGZ was confirmed by pharmacokinetics (plasma drug levels) and the pharmacodynamic biomarker adiponectin. In normal or HF rats, RGZ had no negative effects on any of the biomarkers investigated. Similarly, RGZ had no significant effects on gene expression except for the increase in interleukin-6 mRNA expression in the heart and decrease in epithelial sodium channel β in the kidney. In contrast, echocardiography showed improved cardiac structure and function after RGZ in ZDF rats. Taken together, this study suggests a limited predictive power of these preclinical models in respect to observed clinical adverse effects associated with RGZ.
Introduction
The peroxisome proliferator-activated receptors (PPARs) are a group of nuclear receptors that function as transcription regulators of metabolic pathways in carbohydrate, lipid, and protein metabolism. The PPAR family is composed of three proteins: PPAR-α, PPAR-δ, and PPAR-γ. PPAR-γ is the molecular target of thiazolidinediones (TZDs), which upon its activation exerts metabolic benefits in type 2 diabetes (T2D) management (Duan et al., 2008; Krentz, 2009). In addition to medical benefits of improved insulin sensitivity by TZDs, potential cardiovascular benefits also were suggested in trials with pioglitazone (Lago et al., 2007; Robinson, 2008). However, large-scale clinical studies also have suggested that TZDs increase risks of heart failure (HF) in patients with T2D (Robinson, 2008; Zinn et al., 2008). Furthermore, clinical studies and meta-analyses of rosiglitazone (RGZ) clinical trials failed to establish clear cardiovascular benefits and in fact raised concerns that RGZ may increase cardiovascular events risk, especially cardiac death (Robinson 2008), although this matter is still subject to controversy (Lago et al., 2007; Nissen and Wolski, 2007; Khanderia et al., 2008). In addition, it has been well documented that treatment with TZDs (e.g., RGZ) is associated with fluid retention, edema, and exacerbation of heart failure, probably due to renal mechanism (Staels, 2005; Yang and Soodvilai, 2008).
Because of these recent concerns about TZD-associated cardiovascular risks, the U.S. Food and Drug Administration issued an enhanced warning (“black box”) restricting the use of TZDs in patients with congestive HF (Tanne, 2007). Furthermore, the Food and Drug Administration has issued a final guidance for industry, requesting more rigorous evidence that new T2D treatments in general are not associated with an unacceptable increase in the risk of cardiovascular events (http://www.fda.gov/Cder/Guidance/8576fnl.pdf). Contrary to the clinical experience, experimental studies with PPAR-γ agonists have suggested potential cardiovascular protection, including in HF models. For examples, treatment of normal, diabetic, or obese Zucker rats or mice with the selective PPAR-γ activator RGZ reduced myocardial ischemic injury (Yue et al., 2001; Khandoudi et al., 2002; Sidell et al., 2002; Johns et al., 2005; Mersmann et al., 2008). Similar results were reported with pioglitazone that showed cardiac protection in a mouse myocardial infraction model (Shiomi et al., 2002; Frantz et al., 2004). It is however noteworthy that opposite results had been reported with RGZ, including increased mortality in rats subjected to myocardial infarction (MI) (Lygate et al., 2003).
It is very intriguing that the cardiac and renal liabilities observed in diabetic patients treated with TZDs have not been sufficiently apparent in experimental models used to establish their glycemic control benefits. The failure to identify potential biomarkers that could have aided in identification of potential cardiorenal risks by TZDs could have resulted from limited deliberate investigations for such biomarkers. Therefore, in the present study we embarked on a more detailed investigation aimed to elucidate possible biomarkers of cardiorenal risks associated with TZDs. We used the prototypic PPAR-γ agonist RGZ in three groups of rats, including two groups of rats that carry cardiovascular risks, namely, post-MI HF and Zucker diabetic fatty (ZDF) rats, the latter representing several metabolic manifestations of T2D. We investigated pharmacogenomic, physiological, and biochemical safety and efficacy biomarkers in relation to RGZ treatments (3 or 45 mg/kg/day) that represent clinically relevant and supraclinical RGZ exposures in each model.
Materials and Methods
Animals.
Adult male Lewis rats and ZDF rats (strain code 370), purchased from Charles River Laboratories, Inc. (Wilmington, MA), were used for all studies. Rats were singly housed, cared and subjected to surgical operation with an approved procedure according to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publication 85-23, revised 1996) as approved by the Institutional Animal Care and Use Committee.
Rat Model of Heart Failure.
HF was surgically induced by ligation of the left coronary artery as described previously (Pfeffer et al., 1979). In brief, rats (at 9 weeks of age) were anesthetized with pentobarbital 50 mg/kg i.p. while body temperature was maintained at 37°C during the surgical procedure. An incision was made in the left side of the chest to expose the fourth intercostal space. A 6-0 surgical suture was placed under the left coronary artery and tied to complete occlusion of the artery. In sham-operated animals, the coronary artery was not ligated. In HF animals, approximately 50% survived, of which approximately 50% were able to meet the inclusion criteria for ultrasound studies as described below. Animals were monitored and cared for daily as per Institutional Animal Care and Use Committee regulations. Four weeks after surgery, echocardiography was performed.
Echocardiography.
Cardiac structure and function were monitored by echocardiography before and 4 and 8 weeks after surgery using the Vevo 770 high-resolution in vivo microimaging system (VisualSonics, Toronto, ON, Canada) as described previously (Bartha et al., 2008). Echocardiography was done under anesthesia (inhalation of 2% isoflurane in oxygen). A high-frequency (17.5-MHz) real-time microvisualization scan-head (model RMV-716; VisualSonics) was used to obtain B-mode-guided parasternal short axis M-mode images for cardiac function and structure, including ejection fraction (EF), fractional shortening, left ventricular internal diameter (LVID; systolic and diastolic), left ventricular volume (LVV; systolic and diastolic), stroke volume, cardiac output, and heart rate (HR). The images were analyzed using the system imaging software according to the manufacturer's specification (VisualSonics).
Drug Treatment and Terminal Tissue Collection.
Prespecified inclusion criteria for drug treatment was set at left ventricular ejection fraction of 30 to 40% at 4 weeks postcoronary ligation. Vehicle [aqueous solution containing 2% Tween 80 (w/v) and 0.5% methylcellulose (w/v)], 3 mg/kg RGZ, or 45 mg/kg RGZ was administered by oral gavage daily for 4 weeks in HF rats, in parallel with vehicle; the 45-mg/kg oral dose of RGZ was administered in studies on ZDF rats. The lower dose of 3 mg/kg RGZ was selected as a clinically relevant dose for efficacy in rats (Yue et al., 2001; Kim et al., 2006; Sharabi et al., 2007; Pfützner et al., 2008), and the higher dose of 45 mg/kg represented the supraclinical RGZ exposure aimed to addressed the cardiorenal liability in these models.
Animal body weight was measured weekly throughout the study. At the end of the 4-week treatment, rats were euthanized by terminal bleeding under anesthesia (vide supra). Heart, lung, and kidneys were rapidly removed, blot-dried by filter paper, and weighed. The kidneys and heart (approximately 2–3 mm of the apex portion) were immediately frozen in liquid nitrogen and stored at −80°C for mRNA expression analysis. The remaining heart tissues were fixed with 10% neutral buffered formalin and embedded in paraffin for histological analysis.
Pharmacokinetics and Plasma Drug Concentration Measurement.
For pharmacokinetic (PK) analysis, platelet-poor plasma was prepared from both Lewis and ZDF rats after seven daily dosings of 45 mg/kg p.o. RGZ at 0, 2, 6, and 24 h after the last dosing (n = 5). Plasma drug levels were determined at trough (24 h after the last dosing). Concentrations of RGZ in plasma samples were determined by liquid chromatography/tandem mass spectrometry methods (by internal protocols). In brief, an aliquot of 150 μl of acetonitrile containing rosiglitazone as the internal standard (500 ng/ml) was added to each 50-μl plasma sample. The mixture was vortexed at 1200 rpm for 2 min, and protein pellets were spun down at 5700 rpm for 20 min at 4°C. The supernatant (10 μl) was injected into a high-performance liquid chromatograph for analysis. Chromatographic separation was performed using a 1200 LC system (Agilent Technologies, Santa Clara, CA), which was equipped with a Unison UK-C18 column (2.0 × 30 mm i.d., 3 μm; Imtakt Corporation, Kyoto, Japan). The mobile phase consisted of 10 mM ammonium acetate in water (solvent A) and 10 mM ammonium acetate in acetonitrile (solvent B). A gradient with a flow rate of 1 ml/min was initiated at 100% solvent A, increased to 100% solvent B at 3 min, and kept at 100% solvent B for another 1 min. The retention time for RGZ and the internal standard was 2.02 and 2.11 min, respectively. The ionization was carried out using atmospheric pressure chemical ionization, and a tandem mass spectrometric transition of 358.1 → 135.1 was used to monitor RGZ. The low quantitation limit was 1 ng/ml. PK parameters were calculated with WinNonlin (Pharsight, Mountain View, CA).
Histology and Immunohistochemical Analysis.
Fixed heart specimens (n = 5–6) were transversely sectioned across the ventricles for histological analyses. The extent of fibrosis in the heart was quantified using Gomori's trichrome stain. Specifically, for analysis of fibrosis and left ventricular hypertrophy in the rat hearts, five transverse sections were collected between 5 and 6 mm after the ligation plane in each animal. Five mosaic whole heart transverse photomicrographs were taken with an Axio Imager microscope (Carl Zeiss Inc., Thornwood, NY) equipped with an automatic motorized stage. The extent of fibrosis and left ventricular hypertrophy in these photomicrographs was then quantified by Axio Imager in a blinded manner. For analysis of myocardial fibrosis, the area of the collagen (green) staining was quantified. Left ventricular hypertrophy was assessed on the same tissue sections; the area of the left ventricle was assessed using Axio Imager software (Car Zeiss Inc.). Macrophages infiltration and TNF-α expression in the heart were evaluated as representatives of inflammation markers (Kherani et al., 2004). For the immunohistochemistry study, mouse monoclonal anti-ED1 (CD68) antibody (1:100 dilution; Abcam Inc., Cambridge, MA) was used to assess macrophages in the myocardium, and rabbit polyclonal anti-TNF-α antibody (1:100 dilution; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was used to probe TNF-α protein. Vectastain ABC systems (Vector Laboratories, Burlingame, CA) were used for immunohistochemistry detection.
Plasma Biomarker Analysis.
Immediately after euthanasia (by CO2 inhalation and pneumothorax), blood samples were collected by cardiac puncture into a Vacutainer K2 EDTA tube (7.2 mg, 4 ml; BD Biosciences, San Jose, CA) and centrifuged at 2500g at 4°C for 15 min (Sorvall RT 6000D centrifuge; DuPont, Wilmington, DE). Plasma was collected and stored at −80°C. Adiponectin was analyzed using a Mouse/Rat Adiponectin Enzyme-Linked Immunosorbent Assay kit purchased from B-Bridge International, Inc. (Mountain View, CA). BNP assay was conducted using an Assay Max Rat BNP-32 (rBNP-32) Enzyme-Linked Immunosorbent Assay kit purchased from AssayPro (St. Charles, MO). Insulin was analyzed using a Rat Insulin RIA kit purchased from LINCO Research (St. Charles, MO) according to the manufacturer's specifications. Aldosterone was determined using Coat-A-Count solid-phase 125I RIA kit purchased from Diagnostic Products (Los Angeles, CA).
Targeted Tissue Low Density Gene Array.
Total RNA was prepared using the Total RNA Isolation kit (Promega, Madison, WI) followed by RNA cleanup using the RNeasy Plus Mini kit (QIAGEN, Valencia, CA) according to the manufacturers' protocols. RNA was quantitated using a NanoDrop 1000 spectrophotometer (NanoDrop, Wilmington, DE), and RNA integrity was assessed using a 2100 Bioanalyzer (Agilent Technologies). cDNA was prepared from purified RNA using the High-Capacity cDNA Reverse Transcription kit with RNase inhibitor (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. Quantitative real-time polymerase chain reaction was determined using TaqMan low-density arrays (TLDA) from Applied Biosystems. Thermal cycling was performed using an ABI Prism 7900 sequence detection system (Applied Biosystems).
A TLDA was custom-designed and ordered from Applied Biosystems (see Table 1 for the list of genes). Detailed experimental conditions on the TLDA analysis were described previously (Collins-Racie et al., 2009). The comparative CT method of relative quantification (Livak and Schmittgen, 2001) using averaged values for 18S and glyceraldehyde-3-phosphate dehydrogenase as the normalizer compared with the CT value of the target gene (ΔCT) was used. Relative quantification (or fold change) between different sample groups (e.g., sham-rosiglitazone versus sham-vehicle control) was then determined according to the 2−ΔΔCT method, where ΔΔCT = ΔCT treated sample − ΔCT control sample(s). The mean of the expression values for the sham-vehicle samples (n = 6) was used as the calibrator for these calculations.
Statistical Analysis.
Statistical comparisons were made by analysis of variance (with mean differences determined using Fisher's protected least significant difference) test, and values were considered to be significant when p < 0.05 for the experimental groups compared with the vehicle-treated or sham-operated group as indicated in figure legends and tables.
Results
Pharmacokinetics Analysis of RGZ in Lewis and ZDF Rats.
A PK study was conducted in Lewis and ZDF rats after vehicle or 45 mg/kg p.o. RGZ, daily dosing for 7 days, and plasma samples were collected at various times (0, 2, 6, and 24 h) after the last dosing for drug concentration measurement (Fig. 1). The Cmax, area under the curve, and t1/2 values were 39.6 ± 7.5 μg/ml, 488 ± 83 μg · h/ml, and 4.6 ± 0.9 h, respectively, in Lewis rats and 37.7 ± 5.7 μg/ml, 397 ± 75 μg · h/ml and 5.0 ± 0.9 h, respectively, in ZDF rats. No statistical difference on PK profiles was observed between the two rat strains.
Effects of RGZ on Physiological Parameters in Sham-Operated and HF Lewis Rats and ZDF Rats.
RGZ at 45 mg/kg resulted in significant increase in body weight (11.7% increase in sham-operated animals, p < 0.01 compared with vehicle treatment; 11.9% increase in HF rats, p < 0.05; or 23.0% increase in ZDF rats, p < 0.01; Table 2). RGZ also significantly increased the heart weight (12.2% in sham-operated animals compared with vehicle treatment, p < 0.05; 9.3% in HF rats, p < 0.05; and 12.9% in ZDF rats, p < 0.01), yet heart weight/body weight under RGZ treatment remained the same. Lung weight was significantly increased in HF rats but was not affected by RGZ treatments.
Effects of RGZ on Cardiac Structure and Function by Echocardiographic Analysis.
As illustrated in Table 3 and Fig. 2, five experimental groups were carried out in Lewis rats to study the effect of RGZ under HF conditions, including sham-operated animals treated with vehicle or 45 mg/kg RGZ, MI-induced HF animals treated with vehicle, 3 or 45 mg/kg RGZ daily for 28 days (n = 5–6). Representative ultrasound images on EF in sham-operated and HF rats are illustrated in Fig. 2A, and the quantitative data before and after vehicle/drug treatment are depicted in Fig. 2B. Although a marked decrease in EF was observed in the HF groups, no effect of RGZ was noted in any of the echocardiography variables in RGZ-treated HF rats at any dose (Fig. 2; Table 3). In normal Lewis rats, no significant difference was observed in RGZ treated groups in most functional variables, but it was noted that RGZ treatment (45 mg/kg/day) resulted in improved ultrasound measurements of stroke volume and cardiac output (Table 3). Treatment with RGZ in ZDF rats resulted in significant improvement in cardiac function, showing 11.4% increase in EF, 26.4% decrease in LVIDs, and 41.2% decrease in LVVs in ZDF rats (p < 0.05; Table 3). Thus, no evidence has been raised to indicate deterioration in echocardiographic parameters of cardiac structure and function in any of the experimental groups.
Effects of RGZ on Cardiac Fibrosis and Left Ventricular Hypertrophy.
Coronary artery ligation resulted in myocardial tissue loss and prominent fibrosis (Fig. 3). The amount of myocardial fibrosis in HF rats was 7.4 and 7.6% in vehicle- and RGZ-treated (45-mg/kg) groups, respectively, compared with the baseline (0%) in sham-operated animals (p < 0.01; n = 5–6). Likewise, the left ventricular mass was significantly increased in HF rats compared with sham-operated animals; however, RGZ treatment had no effect on left ventricular mass (Fig. 3).
Effects of RGZ on Myocardial Inflammation in HF Rats.
ED1-immunoreactive macrophages and TNF-α expression were significantly elevated in the heart after MI-induced HF, but no changes were observed in response to RGZ treatment (Fig. 4).
Effects of RGZ on Plasma Biomarkers in Sham-Operated and HF Lewis Rats and ZDF Rats.
As illustrated in Fig. 5, no effects of RGZ on plasma glucose or insulin were observed in any of these experimental groups. The levels of adiponectin, as expected, were significantly increased in response to RGZ treatment (3.1-fold in sham-operated rats, 2.5-fold in HF rats, and 1.9-fold increase in ZDF rats; p < 0.01). The levels of BNP were significantly increased in HF rats (2.2-fold increase over sham; p < 0.05), regardless of RGZ treatment. Although the baseline of BNP levels was considerably elevated in ZDF rats, RGZ had no effect on plasma BNP. Hematocrits were decreased in each of the experimental groups in response to RGZ treatment (Fig. 5) as reported previously (Zhang et al., 2005; Forcheron et al., 2009).
TLDA Analysis on Gene Expression in Response to RGZ in Sham-Operated and HF Lewis Rats and ZDF Rats.
As shown in Fig. 6A, several key marker genes, including ACE, ANP, BNP, and PPAR-γ were significantly elevated in the heart after HF, but RGZ treatment affected only a few genes, including significantly reduction of ACE mRNA expression in both sham-operated (44% reduction; p < 0.05) and HF (46% reduction; p < 0.05) animals, and reduction of TNF-α mRNA in sham-operated animals (28% reduction; p < 0.05). RGZ augmented PPAR-γ mRNA expression in the heart of both sham-operated and HF rats. In the kidney (Fig. 6A, bottom), SGK, Na+/K+ pump, and ROMK1 mRNA expression was significantly elevated in rats with HF, suggesting increased activity of aldosterone. It should be noted that RGZ caused no significant up- or down-regulation in the transport-related ion channels or pumps. Thus, no evidence has been raised to indicate activation by RGZ of an aberrant renal mechanism that could augment salt and water retention.
To study the effect of RGZ on gene expression in rats with diabetic background, comparative studies were conducted in ZDF rats treated with vehicle or RGZ (Fig. 6B). Only IL-6 mRNA expression was significantly elevated in response to RGZ in the cardiac tissue and a slight but significant decrease in ENaCβ mRNA expression in renal tissue in ZDF rats (Fig. 6B). Taken together, no evidence has been raised to indicate augmentation in salt and water retention by a renal mechanism in ZDF rats either.
Discussion
The present study provides a detailed assessment of pharmacogenomic, physiological, and biochemical biomarkers in relation to treatment with the PPAR-γ activator RGZ on cardiac and renal safety biomarkers in experimental rodent models of HF and T2D. Our study is unique by the maintained treatment with RGZ over a prolonged period and explored clinical and supraclinical exposures to ensure sufficient duration for drug impact on organ structure, function, and pharmacogenomics at steady state. Although our PK studies were conducted in naive rats, our parallel studies showed a similar PK profile for RGZ in sham-operated and HF rats induced by aortocaval fistula (Wang and Winaver et al., unpublished observations). We also confirmed proper drug exposures under different experimental conditions, including normal, HF, and ZDF rats. RGZ engagement with its target PPAR-γ was evident by the biomarker adiponectin, an acceptable biomarker of PPAR-γ activation that showed 2- to 3-fold increase in RGZ-treated groups (Fig. 5). Adiponectin response to RGZ in our study conforms with preclinical (Sharabi et al., 2007) and clinical observations, showing a 2.4- to 2.6-fold increase in both nondiabetic and diabetic patients (Kim et al., 2006; Pfützner et al., 2008), suggesting relevant PPAR-γ stimulation. In addition, the up-regulation of PPAR-γ mRNA adds further support for RGZ engagement with its target (Sommer and Wolf, 2007).
It was also noted that no effects of RGZ on plasma glucose or insulin were observed in any of our experimental groups. The lack of RGZ effect on insulin and glucose in ZDF rats was probably caused by the late-stage diabetes marked by insulin deficiency as reported previously (Forcheron et al., 2009). To validate this notion, additional separate groups of younger Lewis and ZDF rats (9 weeks old) were dosed with either vehicle or 45 mg/kg RGZ, and plasma glucose and insulin levels were monitored at day 7; after RGZ treatment, insulin levels were significantly reduced by 32% (Lewis rats, n = 6; p < 0.05) and 57% (ZDF rats, n = 6; p < 0.01). Glucose was decreased in parallel by 16% (p < 0.05) and 29% (p < 0.01), respectively, compared with vehicle treatment (X. Wang, Y. Zhan, X. Liu, and J. Winaver, unpublished observation). Thus, our data confirmed previous observations (Forcheron et al., 2009).
Echocardiography variables and plasma BNP levels (standard biomarkers of HF) confirmed the presence of HF in Lewis rats subjected to coronary ligation, whereas the elevated BNP levels in intact ZDF rats are compelling cases for cardiac distress in this model, too. Chronic treatment with RGZ in HF rats was not associated with deterioration in cardiac structure and function based on echocardiographic parameters (Table 3). Histological assessment (collagen/fibrosis and cardiac hypertrophy) also failed to show any untoward effect of RGZ after chronic high-dose exposure (Fig. 3). In contrast, treatment with RGZ in ZDF rats resulted in significant improvement in cardiac structure and function by means of echocardiographic variables such as increase in EF and decrease in LVIDs and LVVs (Table 3). It is also interesting to note that the plasma level of BNP, a biomarker of cardiac distress and heart failure, was significantly elevated in ZDF rats, and there was a trend of reduction in response to RGZ treatment, although not statistically significant (Fig. 5B). In addition, our study demonstrated significant elevation of plasma aldosterone in ZDF rats (supporting a compromised cardiovascular function and cardiac distress), but no further elevation of this hormonal biomarker of HF in response to RGZ treatment of ZDF rats was noted (Fig. 4), which is in agreement with a previous report (Fredersdorf et al., 2009).
Inflammation is another important cardiovascular risk, which is also known to be associated with MI-induced HF (Kherani et al., 2004). The levels of macrophage infiltration and TNF-α expression were selected to represent inflammation in congruency with the known human condition. Our data confirmed substantial macrophage infiltration and TNF-α expression postcoronary ligation-induced HF as described previously (Kherani et al., 2004), yet neither of these inflammation elements was affected by RGZ treatment (Fig. 4).
Pharmacogenomic studies revealed considerable model effects, including, as expected, elevated cardiac expression of ACE, ANP, and BNP mRNA in HF rats. In regard to the TLDA used to profile major ion channels and transporters in the kidney, we were prompted by a previous report suggesting a role for ENaCs in PPAR-γ-associated fluid retention (Guan et al., 2005). In our study, we found increased renal expression of SGK, Na+/K+ pump, and ROMK1 mRNAs in HF rats but no significant effects by RGZ (Fig. 6). RGZ failed to show any deterioration in these key renal mechanisms associated with salt and water retention based on gene expression. The only exception is the reduction in cardiac expression of ACE and increase in PPAR-γ mRNA in Lewis rats (Fig. 6A) and an increase in IL-6 mRNA in ZDF rats (Fig. 6B), in response to RGZ treatment. It remains to be explored whether the mRNA levels correspond to respective protein levels and function.
Overall, our study did not show significant differences between RGZ and vehicle in normal (sham-operated) and HF Lewis rats or in ZDF rats that present diabetic/metabolic conditions frequently used in diabetes drug discovery and development. This overall result should raise questions in regard to the translational medicine predictive value of these models or the particular species at large. In rats, MI-induced HF captures many key pathophysiological features present in patients, but it differs in important respects. The preclinical studies are conducted on young adult rats, whereas the prevalent human myocardial infarction that is followed by progression to heart failure occurs in late adulthood and often over a background of multiple confounding factors and medications associated with their management. Likewise, ZDF rats present several important features of T2D and complex metabolic derangement, including hyperlipidemia, insulin resistance, and hyperinsulinemia (Schmidt et al., 2003) and neuroendocrine compromises (as shown in our study for BNP and aldosterone). The contribution of HF as a risk factor over a background of T2D-like metabolic derangements, a suspected confounding factor in patients that have encountered cardiorenal adverse effect while treated with RGZ could not be explored in the present experimental setting because of extremely high mortality in ZDF rats subjected to coronary ligation-induced MI.
In an attempt to place the results of our study in clinical perspective of the cardiorenal adverse effects noted in RGZ-treated diabetic patients, we must consider the rather infrequent rate of cardiorenal adverse effects noted in patients exposed to RGZ and also to other TZDs (e.g., pioglitazone). Because the preclinical studies were conducted on small number of animals, exposed to the compound for few weeks, the likelihood to observe events that require prolonged durations (months and years) and large populations (thousands to tens of thousands) could have contributed to the lack of preclinical biomarkers for the rather infrequent postmarketing adverse effects noted in patients. In addition, the risk factors chosen in our study to represent the scope of potential biomarkers that might represent the adverse events observed in patients, although plausible, might not encompass the gamut of genomic, biochemical, metabolic, and physiological aberrations that humans with T2D encounter. Nevertheless, the lack of clear signal in Lewis rats (with or without HF) or ZDF rats, chronically treated with RGZ at clinical and probably supraclinical exposures in humans, may suggest that studies in rodents might be inadequate to manifest the cardiorenal risks observed in patients. Likewise, the usually short phase I clinical studies conducted in normal human volunteers might not have provided adequate warnings to such risks either because of the lack of the metabolic dispositions, the shorter exposure to the compound, or the limited size of the phase I cohorts to sort out uncommon cardiorenal risk predispositions to PPAR-γ modulators.
In summary, our report, the most comprehensive search for safety markers for RGZ-induced cardiovascular complications in rodents, aimed to omit individuals with predispositions for PPAR-γ liabilities in future clinical trials, suggests that studies with PPAR-γ agonists per se, or as part of a pan-PPAR modulator, might require chronic studies in other species, possibly including nonhuman primates, a species more representative of humans. An equally intriguing result from our study is the finding that some of the cardiovascular parameters were actually improved after RGZ treatment in the congestive heart failure rats. This may indicate that TZDs can exert beneficial cardiovascular actions, and unraveling their mechanisms may be of therapeutic importance.
Acknowledgments
We thank Michael Hansen and Narayanan Hariharan for helpful discussion; Kathy Laws and Elizabeth Lonie for determining plasma concentrations of RGZ; and members of BioResources at Pfizer Collegeville facility for excellent technical support on animal care and maintenance.
Footnotes
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
doi:10.1124/jpet.110.167635.
-
ABBREVIATIONS:
- PPAR
- peroxisome proliferator-activated receptor
- TZD
- thiazolidinedione
- T2D
- type 2 diabetes mellitus
- HF
- heart failure
- RGZ
- rosiglitazone
- MI
- myocardial infarction
- ZDF
- Zucker diabetic fatty
- EF
- ejection fraction
- LVID
- left ventricular internal diameter
- LVV
- left ventricular volume
- HR
- heart rate
- PK
- pharmacokinetic
- TNF
- tumor necrosis factor
- BNP
- brain natriuretic peptide
- TLDA
- tissue low-density gene array
- ACE
- angiotensin-converting enzyme
- ANP
- atrial natriuretic peptide
- SGK
- serum and glucocorticoid-regulated kinase
- ROMK
- renal outer medullary potassium
- IL
- interleukin.
- Received March 1, 2010.
- Accepted June 1, 2010.
- Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics