Mechanisms and Biomarkers of Cardiovascular Injury Induced by Phosphodiesterase Inhibitor III SK&F 95654 in the Spontaneously Hypertensive Rat

Abstract

The cardiovascular injury of the type III selective PDE inhibitor SK&F 95654 was investigated in SHR. Twenty-four hours after a single sc injection of 100 or 200 mg/kg of the drug, rats exhibited cardiomyocyte necrosis and apoptosis, interstitial inflammation, hemorrhage and edema, as well as mesenteric arterial hemorrhage and necrosis, periarteritis, EC and VSMC apoptosis, EC activation, and MC activation and degranulation. Elevated serum levels of cTnT and decreased cTnT immunoperoxidase staining on cardiomyocytes were detected in the drug-treated rats. Serum levels of α ₂-macroglobulin and IL-6 were significantly elevated following drug treatment. NMR spectral patterns of urine samples are significantly different between the drug-treated and control rats. These results indicate that measurement of serum cTnT, acute phase proteins, and cytokines as well as metabonomic urine profiles may serve as potential biomarkers for drug-induced cardiovascular injury in rats. Increased expression of CD63 on MC (tissue biomarker of MC), of nitrotyrosine on MC and EC (an indirect indicator of NO in vivo), and of iNOS on MC and EC (source of NO) suggest that NO produced by activated and degranulated MC as well as activated EC play an important role in SK&F 95654-induced mesenteric vascular injury.

Introduction

Identification of sensitive and reliable biomarkers for drug-induced vascular injury in animals has become an important issue in both drug-development and regulatory processes (Kerns et al., 2005). Type III PDE inhibitors such as SK&F 95654, have been reported to induce varying degrees of splanchnic vascular injury in different strains of rats, including Wistar (Joseph et al., 1996), Wistar derived (Westwood et al., 1990), Fisher 344 (Sandusky et al., 1989, 1991), and SD (Hanton et al., 1995, Zhang et al., 2002a). Little is known about the impact of preexisting hypertension on cardiovascular toxicity resulting from the PDE III inhibitor. The SHR strain was selectively bred from the Wistar-Kyoto (WKY) rat strain and is predisposed to hypertension, myocardial infarction, and mesenteric periarteritis nodosa (Okamoto, 1969). In previous studies, we demonstrated that the SHR strain was more sensitive than the WKY strain to the cardiotoxicity of anthracyclines (Herman et al., 1985, 1988, 1994) and the doxorubicin-induced cardiotoxicity in SHR was related to immune reactions (Zhang et al., 1993). Moreover, our previous study showed that male SHR were more susceptible than WKY rats to the cardiotoxicity of radiocontrasts and the contrast mediums-induced toxicity was associated with apoptosis (Zhang et al., 1999). We reasoned that male SHR could be a suitable animal model for accelerating and amplifying SK&F 95654-induced cardiovascular injury and associated biomarker responses.

In a previous study in SD rats, SK&F 95654 was shown to produce (1) cardiac injury consisting of cardiomyocyte necrosis and inflammation; (2) mesenteric vascular injury characterized by arterial hemorrhage and necrosis, and vascular inflammation; and (3) cellular events manifested by MC activation and degranulation, EC activation, and EC and VSMC apoptosis (Zhang et al., 2002a).

The objectives of the present study were (1) to characterize the cardiovascular injury induced by SK&F 95654 in SHR; (2) to compare the SK&F 95654-induced cardiovascular injury in hypertensive rats to those in normotensive SD rats; (3) to search for potential biomarkers of drug-induced cardiovascular injury using accessible serum and urine samples; and (4) to explore possible pathogenesis of drug-induced mesenteric injury.

Materials and Methods

Animals

Thirty-seven male SHR rats (10 weeks old) were supplied by Harlan Laboratories, Inc. (Indianapolis, IN). The experiments began after a 2-week acclimation period and continued until these rats attained the required age (12 weeks old). All rats were housed separately in an environmentally controlled room (18–21°C, 40–70% relative humidity) with a 12-hour light/dark cycle. Animals were fed Certified Purina Rodent Chow #5002 (Ralston Purina Co., St. Louis, MO) and water ad libitum. The experimental protocol was approved by the Institutional Animal Care and Use Committee, Center for Drug Evaluation and Research, FDA and conducted in an AAALAC (Association for Assessment & Accreditation of Laboratory Animal Care) accredited facility. All procedures for animal care and housing were in compliance with the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996).

Experimental Procedures

SK&F 95654 (±-5-methyl-6-[4-(4-oxo-1,4-dihydropyridin-1-yl)phenyl]-4,5-dihydro-3(2H)-pyridazinone) was kindly provided by SmithKline Beecham Pharmaceuticals (now GlaxoSmithKline) through a material transfer agreement. Rats (n = 6–15) were administered saline, dimethyl sulfoxide (DMSO), 100 or 200 mg/kg SK&F 95654 (dose volume = 0.2 ml/100 gm of body weight) via a single sc injection.

Histopathological Studies

Rats were exsanguinated under pentobarbital sodium anesthesia (65 mg/kg) (ip) 24 hour after treatment with SK&F 95654, DMSO, or saline and the entire mesentery, heart, kidney, liver, pancreas, spleen and small intestine were collected. For preparation of mesenteric tissue sections, the entire small intestine with the mesentery was removed from the abdominal cavity and spread onto a wax-coated metal tray. Small intestine with the mesentery was divided into 3 segments in the shape of a ring, each of which was fixed on the tray using thumbtacks at the borderline between the intestine and mesentery. After fixation of the intestinal-mesenteric sample in 10% neutral formalin solution for 24 hours, most of small intestine was trimmed off and the mesenteric tissue with a very narrow edge of intestine was saved for embedding in paraffin in a special direction, in which the plane of unfolded mesenteric tissue was parallel to the cutting plane of a paraffin block. Five-μm-thick sections were mounted onto 7.5 × 5 cm glass slides and stained with hematoxylin-eosin. Selected tissue sections were stained with toluidine blue for MC granules, with Masson’s trichrome or Movat’s pentachrome for myocyte necrosis (American HistoLabs, Inc. Gaithersburg, MD).

Grading System for Vascular Injury

The severity of mesenteric vascular injury was ranked on a scale of 0–5 as described previously (Zhang et al., 2002a). This system is based on the degree of inflammation and the percentage of arteries showing hemorrhage: 0 = no alteration; 1 = minimal inflammation; 2 = mild inflammation; 3 = moderate inflammation and <33% of arteries with hemorrhage; 4 = marked inflammation and 34–66% of arteries with hemorrhage; and 5 = severe inflammation and >67% of arteries with hemorrhage.

Grading System for Cardiac Injury

The severity of acute cardiac injury was graded on a scale of 0–5 as described previously (Zhang et al., 2002a). This system is based on the degree of myocardial inflammation, necrosis, interstitial hemorrhage and edema: 0 = no alteration; 1 = minimal inflammation; 2 = mild inflammation and necrosis; 3 = multiple foci with inflammation and necrosis as well as mild-to-moderate hemorrhage and edema; 4 = confluent areas of inflammation and necrosis as well as severe hemorrhage and edema; 5 = diffuse, severe inflammation, necrosis, hemorrhage and edema as well as involvement of multiple chambers, papillary muscles, septum, and valves.

Method for Counting Mast Cells

The number of mesenteric tissue MC was counted in 3 toluidine blue stained slides per rat. The counting was performed in 50 randomly selected microscopic fields of the mesentery at a magnification of ×250. The resultant data were expressed as mean number ± SD of MC per mm² of tissue section. MC degranulation was identified based on evidence for extruded granules together with appearance of ruptured MC membranes. A ratio of the number of degranulated MC to the number of intact MC was expressed as mean percentage of the sum total MC ± SD per mm² of tissue section.

Immunohistochemical Studies

For the detection of nuclear DNA fragments of mesenteric EC and VSMC, as well as cardiomyocytes, VasoTACS in situ apoptosis detection kit (Trevigen, Inc., Gaithersburg, MD) was used. The procedures for TUNEL assay were described previously (Zhang et al., 2002a).

For indirect immunoperoxidase staining for monoclonal antibodies (mAb) cTnT, CD63, iNOS, and nitrotyrosine, serial sections of formalin-fixed, paraffin-embedded tissues were mounted onto glass slides coated with poly-L-Lysine. Pretreatment of the sections was performed with microwave irradiation in a pressure cooker with Glyca (for cTnT) or Citra (for CD63, iNOS, and nitrotyrosine) antigen retrieval solution (BioGenex, San Ramon, CA). After microwave treatment, slides were cooled in the solution for 20 minutes and then rinsed with distilled water. For blocking endogenous peroxidase activity, sections were incubated with 0.3% hydrogen peroxidase in methanol for 30 min and then with 5% normal horse serum for 30 minutes. Sections were incubated overnight at 4°C with primary mAb, cTnT/MCA470 (clone: T1/16) (Serotec, Inc. Raleigh, NC), CD63 (BD Biosciences, San Diego, CA; Cat#: 551458), iNOS (BD Biosciences, San Diego, CA; Cat#: 610328), and nitrotyrosine (Cell Science, Norwood, MA; Cat#: 5002) at a dilution of 1:100. After washing with phosphate-buffered saline (PBS), the sections were incubated with a biotinylated second antibody (Vector Laboratories, Burlingame, CA) for 1 hour and then incubated with avidin-biotinylated horseradish peroxidase complex (Vector) for 30 minutes. The peroxidase reaction was carried out with 0.05% 3′3′-diaminobenzidine in 0.1 M Tris-HCI buffer and 0.01% hydrogen peroxide for 5 minutes. Finally, sections were counterstained with hematoxylin.

For CD63, iNOS, and nitrotyrosine negative control staining, the primary mAb was omitted or substituted by mouse IgG₁ (BD Biosciences, Cat#: 557273), by mouse IgG_2a (BD Biosciences, Cat#: 550339) or by mouse IgG_2b (Cell Sciences, Cat#: CNCH003B-100), respectively, in the incubation step.

Clinical Chemistry Analysis

Terminal blood samples were taken 24 hours after a single sc injection of SK&F 95654 and collected from the inferior vena cava for clinical chemistry determinations (AniLytics, Inc., Gaithersburg, MD).

Immunoassay for cTnT

Blood samples were centrifuged immediately after collection and the sera were frozen at −40°C until assayed. Serum concentrations of cTnT were monitored in the laboratory of Dr. Nader Rifai (Boston Childrens Hospital) by immunoassay (Elecsys STAT) (Roche Diagnostics, Indianapolis, IN).

Measurement of Cytokines and α₂-Macroglobulin

The serum levels of TNF-α and IL-6 were measured in rat sera using sandwich ELISA kits from R&D system (Minneapolis, MN) according to the manufacturer’s instructions. The method of measurement of α ₂-M was developed in our laboratory (Boehringer Ingelheim Pharmaceuticals, Inc.). Microtiter plates (96-well) were coated with purified rabbit anti-rat α ₂-M (Strategic Biosolutions, Newark, DE) diluted in pH 9.6 carbonate buffer, and incubated overnight at approximately 4°C. Wells were emptied and coated with PBS with 1% bovine serum albumin (BSA) and 1% normal rabbit serum and incubated for 1 hour at 37°C. Wells were washed 4 times with 200 μl/well PBS + 0.05% Tween 20. Washing was repeated in between the steps that follow. The standard rat α ₂-M (Lee Scientific, St. Louis, MO) and samples were diluted in PBS + 1% BSA and were added to wells in duplicate and incubated for 1 hour at 37°C. Alkaline phosphate-conjugated rabbit anti-rat α ₂-M (Strategic Biosolutions) was added and incubated for 1 hr at 37°C. Para-nitrophenyl phosphate substrate (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was added and the subsequent reaction was quantitated at 405 and 650 nm with a SPECTRAmax 340 (Molecular Devices Corp, Sunnyvale, CA) reader when the most concentrated dilution of the standard reached 1.5 to 2. Each 650 nm reading was subtracted from the corresponding 405 nm value prior to fitting the 4-parameter standard curve and extrapolating the unknown concentrations from the curve. Prior titration of selected samples was done to determine the optimum dilution to be used in the analysis.

Metabonomic Analysis

At study termination, urine samples were collected from the bladder via a needle-tipped syringe and stored at −20°C. Analyses for spectral profiles of intermediary metabolites using NMR spectroscopy were performed by Pfizer Inc. (Ann Arbor Laboratories). NMR analyses were conducted as described previously (Robertson et al., 2000, 2001). Briefly, NMR spectra were obtained from each urine sample. Discrete regions of each spectrum (0.04 ppm) were integrated and the resultant measurements were used as the raw data for pattern recognition analysis. The regions of the spectrum containing intense signals from the DMSO vehicle and related products (3.14–3.18 and 2.72–2.75 ppm) were excluded from the analyses, as were regions that contained resonances from SK&F 95654 and related products. The data were analyzed using PCA, an unsupervised multivariate statistical method useful for reducing multidimensional data such as multiple NMR spectra to 2 or 3 dimensions that can readily be comprehended. Graphical representations allow for identification of individual samples whose NMR spectra differ from each other. In these representations, the magnitude of response, severity in the case of toxicity, is proportional to the distance from the control in PC space. Additionally, separation of centroids within the PCA is indicative of different NMR spectral patterns, which may involve differences in one or more analytes. The NMR and PCA data were generated with knowledge of treatment regimen, but without knowledge of the clinical or histopathology data.

Statistics

Differences in cardiac and vascular lesion scores between groups were compared using the Mann–Whitney test (nonparametric analysis of variance). Differences in the number of mast cells and percentage of mast cell degranulation between groups were compared using analysis of variance (ANOVA). Differences in the serum levels of clinical chemistry analytes were compared using Student’s t-test, while differences in cTnT, cytokine, and α ₂-M levels were compared using the Tukey–Kramer or Dunn’s test for multiple comparisons. A value of p ≤ 0.05 was considered statistically significant.

Results

Vascular Injury

Mesenteric vessels with hemorrhage were observed grossly in 2/6 and 6/10 SHR administered 100 and 200 mg/kg SK&F 95654, respectively. However, mesenteric arterial medial hemorrhage and necrosis were observed microscopically in all rats treated with SK&F 95654 (Figure 1A). In addition, inflammatory response was noted in the perivascular spaces of capillaries, venules, arterioles, small veins, and arteries of the mesentery (Figure 1B). TUNEL assay showed little or no positive reaction in mesenteric vessels from control rats (Figure 1C), in contrast, a positive nuclear reaction was identified in EC and VSMC of the mesentery from SK&F 95654-treated rats (Figure 1D). Venous thromboses were occasionally visible in 2/16 rats administered 100mg/kg or 200 mg/kg of SK&F 95654. Cellular alterations observed by light microscopy included EC activation and mesenteric tissue MC activation or degranulation. In control SHR, EC appeared flat in shape with a small nucleus. In SK&F 95654-treated rats, activated EC of the mesentery appeared cuboidal in shape with a large nucleus (Figure 1E). In control SHR, MC were characterized by their elongated shape with dense cytoplasmic granules, which occupied the cytoplasm to such degree as to obscure the nucleus (Figure 1F). However, in SK&F 95654-treated rats, numerous MC were widely distributed at the affected areas (Figure 1G). In some MC, the number of cytoplasmic granules was increased and the size of cell body was enlarged, which indicates MC activation (Figure 1H, inset). In addition, the cell membranes of some MC were ruptured and cytoplasmic granules were released to the outside of MC, which indicates MC degranulation (Figure 1H).

Tables 1 and 2 summarized SK&F 95654-induced vascular lesion scores and the numbers of MC and the percentages of MC degranulation in SK&F 95654-treated SHR, respectively. Both tables show significant differences following drug-treatment.

Immunoperoxidase staining showed that immunoreactivity of CD63, iNOS and nitrotyrosine in MC and vascular EC varied from little or no reaction in saline-treated rats to a strong positive reaction in 100 mg/kg SK&F 95654-treated rats, and to a moderately positive reaction in 200 mg/kg SK&F 95654-treated rats. Negative control immunoperoxidase staining for CD63 showed no reaction in MC (Figure 2A, inset). Positive immunostaining for CD63 showed that MC expressed weak positive reaction in saline-treated rats (Figure 2A), a very strongly positive reaction of CD63 in 100mg/kg SK&F 95654-treated rats (Figure 2B) and a moderately positive reaction in 200mg/kg SK&F 95654-treated rats (Figure 2B, inset). Negative control immunostaining for iNOS showed no reaction in MC and vascular EC (Figure 2C, inset). Positive immunostaining for iNOS showed that immunoreactivity of iNOS in MC and microvascular EC was stronger in 100 mg/kg SK&F 95654-treated rats than in 200 mg/kg-treated rats (Figs. 2D, 2D inset, and 2E). The same staining patterns in MC and EC were identified in immunostaining for nitrotyrosine. Increased immunoreactivity of nitrotrysine in MC was more conspicuous in 100 mg/kg SK&F 95654-tereated rats (Figure 2F, top inset) than in 200 mg/kg SK&F-treated rats. (Figure 2F, bottom inset). It is noteworthy that increased immunoreactivity of nitrotyrosine was widely distributed in the microvascular EC, particularly EC of capillaries (Figure 2G), venules (Figure 2G, inset), arterioles, and small veins (Figure 2H).

Cardiac Injury

The myocardial lesions observed in SK&F 95654-treated SHR were significantly more severe than those found in saline- or DMSO-treated SHR (Table 1). Hearts from saline-or DMSO-treated SHR appeared normal (Figure 3A). In contrast, myocardial necrosis, interstitial inflammation (Figures 3B & 3C), interstitial hemorrhage, and edema were observed in SHR treated with SK&F 95654. Myocardial necrosis occurred mainly in the ventricles, the apex, the septum, and the papillary muscles, and rarely in the atria. Necrosis was characterized by the presence of hypercontraction bands with inflammatory cell infiltration (Figures 3B & 3C). TUNEL assessment showed no apoptosis in cardiac myocytes in control SHR (Figure 3D). In contrast, apoptotic myocytes were observed in SK&F 95654-treated SHR (Figure 3E). Immunoperoxidase staining for cTnT showed that the antibody against cTnT stained the regular cross-striations suggestive of the I bands of myocytes, but did not stain the intercalated disks of myocytes in control rats (Figure 3F). In contrast, there was no positive reaction in necrotic myocytes. A weakly positive reaction was observed in the vicinity of cardiac necrosis in SK&F 95654-treated SHR (Figure 3G). Compared with no staining or weak staining of cTnT in the lumen of intracardiac vessels in control SHR, the lumen of capillaries, venules, and arterioles expressed strongly positive reactions in SK&F 95654-treated SHR (Figure 3H).

Routine Clinical Chemistry

Serum concentrations of phosphorus and blood urea nitrogen (BUN) were significantly decreased and glucose increased in drug-treated SHR compared to those in saline-or DMSO-treated SHR (Table 3). Serum concentration of alanine aminotransferase (ALT) was elevated in 100 mg/kg SK&F 95654-treated rats, while elevation of aspartate aminotransferase (AST) was detected in 100 or 200 mg/kg SK&F 95654-treated rats compared to that obtained from DMSO-treated SHR (Table 3).

Serum Concentrations of cTnT, α₂-M, and IL-6

Serum levels of cTnT were significantly higher in SKF 95654-treated SHR (either 100 or 200 mg/kg) than in saline- or DMSO-treated SHR (Figure 4). SHR given 100 or 200 mg/kg SK&F 95654 had significantly higher serum levels of α ₂-M (Figure 5). Serum levels of IL-6 were not significantly elevated and serum levels of TNF-α did not change in SHR treated with SK&F 95654 compared to those found in saline treated SHR (Figure 6).

Metabonomic Analysis

A divergent urinary spectral pattern was identified by a PCA in the drug-treated rats. No overlap in the patterns in metabolic space, as defined by the first 2 principal components, was seen between drug groups and control groups (Figure 7). Two 100 mg/kg treated SHR, had markedly different trajectories than other SHR in the group and all 200 mg/kg treated SHR suggestive of a different biochemical effect. The major urinary biomolecular constituents responsible for the pattern separation in drug-treated rats included higher concentration of creatine and lower concentrations of hippurate, 2-oxoglutarate, succinate, and citrate (Table 4).

Discussion

SK&F 95654-induced Vascular Injury in SHR

SK&F 95654-induced mesenteric vascular injury in the SHR rat was characterized by arterial hemorrhage and necrosis as well as vascular inflammation, accompanied by mesenteric tissue MC activation and degranulation, EC activation, vascular EC apoptosis, and VSMC apoptosis. The pathological alterations induced by SK&F 95654 in the present study are similar to those observed in a previous study that utilized normotensive SD rats (Zhang et al., 2002a). However, the incidence and severity of vascular lesions was greater in SHR (mean score of 5 or 4 at doses of 100 or 200 mg/kg) than in SD rats (mean score of 3.3 or 3 at doses of 100 or 200 mg/kg).

The mechanisms responsible for the increased SK&F 95654-induced vascular injury in SHR remain unclear. Excessive sustained vasodilation, increased blood flow, increased shear and hoop stress, direct pharmacological and/or chemical compromise, and immune-mediated mechanisms have all been suggested as pathogenic mechanisms for drug-induced vascular injury in normotensive rats (Hanton et al., 1995; Joseph et al., 1996, 2000; Kerns et al., 2005). Data from the present study suggest that the elevated blood pressure in SHR would exacerbate vascular injury. There is evidence from a variety of sources that for cellular responses the strain of rat can be important. VSMC and splenic mononuclear cells (or macrophages) from SHR have been found to produce high levels of NO (Pascual et al., 1993; Papapetropoulos et al., 1994; Xiao and Pang, 1996). NO has been suggested to mediate immune dysfunction in SHR (Pascual et al., 1993).

The basal (unstimulated) levels of plasma nitrite (a metabolite of NO), the basal levels of aortic cGMP (cyclic guanosine 3’, 5’-monophosphate) (an indicator of NO production) and the basal levels of iNOS protein expression are also higher in the SHR rat than in the WKY rat (Wu et al., 1996). The amount of cGMP generated in VSMC is greater in SHR exposed to exogenous or endogenous nitrovasodilators than in nontreated rats (Papapetropoulos et al., 1994). The relaxation responses of mesenteric arteries to nitrovasodilators are more pronounced in SHR compared to WKY rats (Tesfamariam and Halpern, 1988). The cGMP levels of the effluent collected from nitrovasodilator-perfused mesenteric arteries are very high in SHR (Fukuda et al., 1991). Rat serosal MC have been shown to be capable of releasing a NO-like factor derived from L-arginine, which leads to a significant increase in cGMP concentrations in response to a nitrovasodilator (Salvemini et al., 1991).

Taken together, these data led us to hypothesize that potential pathogenic mechanisms responsible for the drug-induced vascular injury in SHR may be involved in excessive local NO release. To verify our hypothesis, the present study employed peroxidase staining for mAb nitrotyrosine and iNOS. NO itself does not produce nitrotyrosine, but it is capable of forming the potent biological oxidant, peroxynitrite, by reaction with superoxide. Since the identification of nitrotyrosine residues within cells reflects the involvement of peroxynitrite in vivo, nitrotyrosine can be used as an indirect marker of NO-mediated cellular injury (Viera et al., 1999; Coleman, 2000; Xu et al., 2001) or acute inflammation (Kooy et al., 1995). Type II iNOS was reported to lead to sustained synthesis of NO in target cells (Heneka et al., 2000; Ikegami et al., 2002). The present study revealed that increased nitrotyrosine and iNOS were expressed on MC and EC, suggesting that per-oxynitrite (one of the metabolites of NO) and NO may be involved in the SK&F 95654-induced vascular injury. It is interesting that immunoreactivity of iNOS and nitrotyrosine at the affected area was greater in 100 mg/kg-treated SHR than in 200 mg/kg SK&F 95654-treated SHR.

Mouse mAb CD63 was used as a specific marker of MC in various rat tissues (Nishikata et al., 1992; Boros et al., 1995). In the present study, a strong positive reaction of CD63 on MC in 100 mg/kg SK&F 95654-treated rats and a moderate positive reaction in 200 mg/kg SK&F 95654-treated rats formed a striking contrast to little or no reaction on MC in control rats. These findings would be interpreted as indicating that MC underwent activation or degranulation, thus suggesting that CD 63 may serve as a marker for MC activation or degranulation. Recently, a direct correlation between MC degranulation and early vasculitis following HgCl₂challenge was detected in the rat model (Vinen et al., 2004). Whether rat peritoneal MC serve as a possible source of NO production is still controversial, with claims by Salvemini et al. (1991) and with disclaims by Coleman (2000). Increased nitrotyrosine and iNOS expression in MC in the present study seem to support the notion that SHR mesenteric MC could serve as a source of in vivo NO in response to SK&F 95654.

Components of MC such as chymase have been reported to induce VSMC apoptosis (Leskinen et al., 2001) and apoptosis of cardiomyocytes (Hara et al., 1999). An in vitro study has shown that MC activation or degranulation induced EC apoptosis (Latti et al., 2003). In addition, MC degranulation has been reported to upregulate P-selectin and ICAM-1 on EC and to have a general pro-inflammatory effect on communication between MC and EC (Torres et al., 2002). It was reported that MC degranulation was involved in the initiation or progression of vasculitis and arthritis (Johnston et al., 1998; Lee et al., 2002). Taken together, SHR may be intrinsically or genetically prone to produce NO with a subsequent cascade of cellular and biochemical event. In SHR, production of NO may result in activation of MC and macrophages, degranulation of MC, increase in VSMC cGMP, upregulation of endothelial adhesion molecules, and trigger vascular cell apoptosis, which ultimately leads to enhanced VSMC vasorelexation and to cause local vascular injury.

EC activation is an important event in the pathogenesis of vasculitis. The activated EC increase secretion of vasoactive mediators and heighten vascular permeability (Cybulsky et al., 2001). The activated EC also promote release of NO and the NO cascade can be affected by pro-inflammatory cytokines (Ballermann, 1998). In activated EC, regulated NO synthesis from endothelial NO synthase (eNOS) is unstable and thus dramatically reduced; however, NO production from iNOS tends to be high and sustained (Ballermann, 1998). In the present study, the high levels of vascular endothelial NO synthesis may also contribute to SK&F-95654-induced vasodilation and injury.

SK&F 95654-Induced Cardiac Injury in SHR

SK&F 95654-induced cardiac injury is characterized by myocardial necrosis and interstitial inflammation, hemorrhage and edema. These findings are consistent with those observed in SD rats administered SK&F 95654 (Zhang et al., 2002a). In addition, apoptosis of cardiac myocytes was identified in SHR. Compared to the cardiac lesions in SD rats (Zhang et al., 2002a), the lesions were more severe in SHR (4 and 5 at doses of 100 and 200 mg/kg in SHR vs. 2 and 3.8 in SD, respectively. The mechanisms responsible for the enhanced drug-induced cardiac injury in SHR are not fully understood. It has been reported that the noradrenaline and the serotonin content are greatly increased in SHR hearts compared to those of WKY rats (Okamoto, 1969). Tendencies to increase cardiac weight and to cause concentric hypertrophy were also reported in SHR (Okamoto, 1969).

Other intrinsic properties together with hypertension-enhanced afterload may contribute to the increased injury in SHR following treatment with SK&F 95654. Other PDE III inhibitors (SK&F 94120, isomazole, and indolidan) caused cardiac toxicity in rats and dogs, which may be due to exaggerated positive inotropic/vasodilator effects of the drugs (Harleman et al., 1986; Sandusky et al., 1989, 1991). The mechanisms by which vasodilating antihypertensive agents (hydralazine, minoxidil, etc.) produced myocardial necrosis in experimental animals were also attributed to exaggerated effects of these drugs: vasodilatation that results in a reflex tachycardia via the adrenergic neurotransmitter and hypotension that may reslt in hypoxemia (Balazs and Ferrans, 1978). In the SHR model, pharmacological hypotensive and vasodilative effects of SK&F 95654 may result in a reflex tachycardia and lead to a further increase in energy and oxygen utilization. An imbalance between oxygen supply and demand for hypertrophied and stimulated hearts in SHR could be responsible for the myocardial lesions induced by SK&F 95654.

Potential Biomarker for SK&F 95654-Induced Cardiovascular Injury

Serum levels of cTnT were significantly increased in SHR following treatment with SK&F 95654. Elevation of cTnT levels may be attributed, in part, to the observed subendocardial necrosis and apoptosis. Immunoperoxidase staining for cTnT showed that reduction in the immunoreactivity with mAb against cTnT correlated well with the increase in serum levels of cTnT and the severity of myocardial injury. The serum levels of cTnT detected at 24 hours in SHR may not be maximal. Our previous studies with SK&F 95654-treated SD rats showed that serum levels of cTnT increased as early as 1–6 hours after administration of the drug, peaked at 8 hours, declined at 12–24 hours, and returned to normal levels 2 weeks postdosing (Zhang et al., in preparation). It has been reported that highly significant elevations in serum levels of cTnT were found 2, 4, or 6 hours after administration of isoprenaline in rats and the serum level of cTnT concentration was much lower 24 hours later (Bertinchant et al., 2000). Serum levels of cTnT have been reported to increase 6 hours following dosing with the β-agonist orciprenaline and to decrease significantly over a 24 hour period (Bertsch et al., 1997). Rats injected with a single dose of 500 μg/kg isoproterenol demonstrated that the serum levels of cTnT were highest at 3 hours and declined by 24 hours post-treatment (Herman et al., 2005). Thus, it appears that monitoring serum levels of cTnT can provide a useful marker of drug-induced cardiovascular injury.

Increased levels of acute phase reactant proteins in the circulation may represent one of the most important aspects of the initial host defense reaction to different noxious agents (Weimer et al., 1965; Strnad et al., 2000). Acute phase proteins are synthesized in the liver in response to pro-inflammatory cytokines (IL-6, IL-1 and TNF-α) released by activated MC, EC, and macrophages (Gentry, 1999; Zhang et al., 2002b). Alpha₂-macroglobulin and haptoglobin are among the most important acute phase proteins in rats and play critical roles in inhibition of proteases, binding free radicals, and prevention of fluid loss from the circulation (Gentry, 1999; Strnad et al., 2000). In traumatized rats, increases in cytokines (IL-6, IL-1, and TNF-α) preceded the elevation of α ₂-M and haptoglobin (Strnad et al., 2000). Significant increases in serum levels of IL-6, IL-1, and TNF-α were detected as early as 1 hour after introduction of trauma. The increase in cytokines was followed at 6–12 hours by a sustained increase in α ₂-M and haptoglobin that was maximal at 24–48 hours (Strnad et al., 2000). Experiments in malnourished SD rats with acute inflammation induced by injection of turpentine oil or lipopolysacharide showed that serum levels of α ₂-M and IL-6 increase as early as 6–9 hours (Lyoumi et al., 1998). Our earlier study of SK&F 95654-induced vascular injury in SD rats found an increase in plasma levels of α ₂-M and an early change in α ₂-M profile detected by two-dimensional gel proteomic analysis (Zhang et al., 2002b; Zhang et al., in preparation). In the present study, serum levels of α ₂-M were significantly higher in drug-treated SHR rats than in control rats. The increased serum levels of α ₂-M and IL-6 may herald early onset of drug-induced vascular inflammation, although our finding did not show elevation of TNF-α.

Metabonomic Analysis

The metabonomics approach has been demonstrated to be a reliable noninvasive method for detecting mesenteric vascular injury induced by the PDE IV inhibitor (CI-1080) (Robertson et al., 2001; Slim et al., 2002). In the present study, metabonomic urine profiles were different in SK&F 95654-treated SHR compared to control animals. The PCA scores plot showed significantly divergent patterns between drug-treated and control rats (Figure 7). The present study showed systematic difference in NMR spectral patterns, as indicated by the PCA scores plot (Figure 7) was clearly indicative of the divergent biological responses. The fact that there was no overlap between the treatment groups and control group suggests a substantial difference in the biochemical composition of the urine. This was confirmed by individual assessment of the spectra that revealed major changes in several urinary constituents relative to control animals (Table 4).

The severity of the urinary changes observed in the present study was similar to that in a previous metabonomic analysis of PDE IV inhibitor-induced vasculitis in rats (Robertson et al., 2001). This is not surprising given the complexity of the toxicological findings seen in this study. As metabonomics is by definition a systems assessment of metabolic changes, when dealing with multiple target organs, it is very difficult if not impossible to deconvolute and ascribe any specific urinary biomolecular change or pattern of changes to any one target organ without first establishing a mechanistic link. However, the ability to noninvasively detect divergent effects, regardless of their origin and with no a priori knowledge about what to expect, represents a significant advancement in the toxicologist’s armamentarium. Additionally, identification and relative quantitation of specific biochemical changes reveal true phenotypic changes that can serve to focus mechanistic work, particularly other “omic” investigations. Eventually, specific patterns of urinary biomolecular changes will be linked with specific target organ effects, but this goal still requires a good deal of effort to document.