Current status and future directions for diagnostic markers ... - IOS Press

15

Cancer Biomarkers 1 (2005) 15–28 IOS Press

Current status and future directions for diagnostic markers of drug-induced vascular injury D.A. Brotta,b,∗ , H.B. Jonesc, S. Gouldc , J.P. Valentinc , G. Evansc, R.J. Richardson b and C. Louden a,b a

Global Safety Assessment, AstraZeneca Pharmaceuticals, Wilmington, DE, USA Toxicology Program, Department of Environmental Health Sciences, The University of Michigan, Ann Arbor, MI, USA c Global Safety Assessment, AstraZeneca Pharmaceuticals, Alderely Park, UK b

Abstract. Recently, there has been an increased incidence of vascular toxicity in pre-clinical toxicology studies. This is of concern because of the uncertain relevance and extrapolation of this finding to humans. In dogs, profound heart rate (HR) and mean arterial pressure (MAP) changes were considered surrogate markers for drug-induced vascular injury until the early 1990s when endothelin receptor antagonists (ETRA) did not significantly alter HR or MAP but induced identical lesions in the coronary arteries of dogs. Thus significant alterations in HR and MAP were found not to be a prerequisite for this lesion. Clinically, the potential for vascular injury coupled with the lack of an unequivocal non-invasive diagnostic marker is an issue of concern to pharmaceutical companies and the regulatory authorities. Therefore, qualification and validation of biomarkers as diagnostic tools for drug-induced vascular injury would add great value to risk management and expedite the drug development process. This review focuses on the status, progress and future trends in vascular biology aimed at identification and development of diagnostic markers that are specific, sensitive and possess potential utility in both a pre-clinical and clinical setting.

1. Introduction The occurrence of spontaneous arterial lesions complicates the issue of drug-induced vascular pathology identified in routine histological examination (dogs, rats and monkeys) of pre-clinical studies [24,54]. Species differences, site specificity, sensitivity and dose response add other challenging aspects and complexity to the issue. For example, the coronary vasculature is the most common site for lesion development in the dog [13,37,44,46], while in the rat the mesenteric vascular bed is affected [66]. Interestingly, endothelin receptor antagonists (ETRA) induced vascular lesions in the dog [2,30,38,62] and monkey [1] but not in rats [62]. ∗ Corresponding

author: David A. Brott. Current address: AstraZeneca Pharmaceuticals, 1800 Concord Pike, CRDL121, PO Box 15437, Wilmington, DE 19850-5437, USA. Tel.: +1 302 885 5188; Fax: +1 302 886 2341; E-mail: [email protected].

Fenoldopam, a dopaminergic agonist, induced vascular lesions in the rat but not in the dog or monkey [31], which suggests a species specific/sensitive response. Macroscopically in dogs, lesions commonly consist of petechial hemorrhages in the right atrium, atrial appendages and coronary groove with infrequent involvement of the left atrium and no lesions in the ventricles or septum. Histologically, the medium-sized to large extramural coronary arteries are primarily affected (Fig. 1). Acute arterial lesions are characterized by multifocal segmental or circumferential hemorrhage, smooth muscle cell necrosis/apoptosis, and perivascular edema. Lesions often progress to chronic perivascular inflammation with loss of normal arterial architecture. In rats, the mesenteric and splanchnic arterial beds are primarily affected. Arterial lesions in the rat are characterized by marked medial necrosis and hemorrhage, widened subendothelial space filled with

ISSN 1574-0153/05/$17.00 © 2005 – IOS Press and the authors. All rights reserved

16

D.A. Brott et al. / Current status and future directions for diagnostic markers of drug-induced vascular injury

Fig. 1. Hematoxylin and eosin staining of dog coronary arteries illustrating control appearance and the features of acute and chronic drug-induced vascular injury.

edema fluid and red blood cells [48,66]. The vascular endothelium remains intact in dogs and rats. Currently, there are no known, unequivocal, noninvasive diagnostic markers for monitoring vascular injury in pre-clinical toxicology or clinical studies. This could be related to our lack of knowledge and/or methods for monitoring. In addition, there is an increase in the number of candidate drugs that have identified vascular injury as a hazard and thus a safety issue. This observation in pre-clinical toxicology studies complicates risk assessment because a comparable vascular insult has not been identified in humans exposed to compounds that produce this lesion in dogs or rats [60, 61]. Thus, drug-discovery, development and regulatory scientists are faced with a dilemma when reviewing scientific data generated in pre-clinical toxicology studies that report vascular injury for the following reasons: (1) no diagnostic markers exist for monitoring (2) inconsistent nomenclature used to characterize the lesion (3) species differences exist in location of affected vessels (4) occurrence of lesions at doses with very low or no safety margins (5) lack of a dose-response relationship occurs in affected species and (6) tendency for lesions to occur in one (either rodent or non-rodent) species, but rarely both. This situation has led to an urgent and imperative need for a strong scientific partnership between academic institutions, regulatory agencies and pharmaceutical industries. These groups would pool their resources in order to enhance our knowledge of pathogenesis and/or identify biomarkers as diagnostic tools with pre-clinical and clinical application. In recognition of this, the Food and Drug Administration (FDA) convened “The Vasculitis Expert Working Group” to evaluate the current state of knowledge, discuss and identify promising areas of non-clinical scientific research aimed at developing biomarkers as potential diagnostic tools for pre-clinical and clinical application, which could ultimately aid in risk assessment

Fig. 2. Von Willebrand factor (vWF) immunohistochemistry of dog coronary arteries and rat mesenteric arteries. Control arteries showing vWF only in the endothelial cells. Treated arteries showing vWF in the endothelial cells and the smooth muscle cell layer.

and risk management of novel therapies associated with this hazard [64]. The ultimate goal is to identify a diagnostic marker of drug-induced vascular injury for use in preclinical and clinical studies. The ideal marker should be (1) specific and sensitive, (2) mechanistically linked to lesion formation, (3) altered before morphological evidence of cell injury (4) reversible, i.e. if treatment is withdrawn biomarker values would return to baseline and progressive injury would be avoided. Therefore, research activities in drug-induced vascular injury have focused on advancing our knowledge in four different areas (1) identification of specific and sensitive diagnostic markers of vascular injury, (2) determination of the value and significance of markers of endothelial cell (EC) and vascular smooth muscle cell (VSMC) injury, coagulation and inflammation, (3) elucidation of the mechanism(s) of drug-induced vascular injury (4) utilization of new technologies, such as genomics,


17

Table 1 Hemodynamic changes caused by drug-induced vascular toxicants Pharmacological class Potassium channel Opener (PCO)

Compound Minoxidil

Nicorandil Hydralazine

HR % change + 10 + 40 + 60 + 65 + 70

MAP % change −8 − 14 − 28 − 29 − 30

lesionsA − − + + +

Reference 43

0.5 mg/kg 1 mg/kg 3 mg/kg 24 ug/kg/min 12 mg/kg bid 24 mg/kg bid

+ 10 + 22 + 43 + 101 + 117 + 100

− 39 − 50 − 60 − 31 ND ND

−/+ + + ND + +

25

Dose 0.05 mg/kg/day 0.14 mg/kg/day 0.43 mg/kg/day 1.44 mg/kg/day 4.32 mg/kg/day

26 11

Adenosine agonist

Adenosine Cl-947

23.5 um ole/kg 5 mg/kg

+ 42 ND

− 23 ND

ND +

8 46

PDE inhibitors

Indolidan

2 mg/kg 2 mg/kg/day

+ 40–50 + 40–50 + 43

− 20–25 − 25–35 ND

+ + ND

55

50 ug/kg/min 0.2 mg/kg/day 1 mg/kg/day 5 mg/kg/day 10 mg/kg/day 20 mg/kg/day

+ 10–20 +3 −8 ND ND +1

−5 5 −5 ND ND −8

Milrinone Endothelin receptor antagonist (ETRA)

SB209670 ZD1611

+ ND ND −/+ + +

34 38 30

A −, Lesions not observed; −/+, lesion observed in less than half of the animals; +, lesions were observed in most animals; ND, not detemined.

proteomics and metaboiomics to characterize changes in expression of proteins causally associated with this lesion.

2. Hemodynamic parameters as biomarkers of vascular injury 2.1. Mean arterial pressure (MAP) and heart rate (HR) In the dog, vascular toxicity is often associated with profound cardiovascular hemodynamic changes in mean arterial pressure (MAP) and heart rate (HR), and these parameters have been used as surrogate markers to monitor potential vascular toxicity in man at therapeutic doses. For example, Minoxidil (a potent antihypertensive agent) produces coronary arterial lesions but significantly lowers MAP and profoundly increases HR. Other compounds such as, Hydralazine , Milrinone and Nicorandil can produce these effects in the dog (Table 1). Interestingly, they are observed following exposure to compounds from diverse pharmacological classes [14]. These findings suggest that the toxicity may be mediated by the hemodynamic changes in the vascular bed rather than the nature of the chemicals. However, recent experiences with endothelin re-

ceptor antagonists (ETRA), a novel class of vasoactive agents, suggest that in comparison to Minoxidil profound and monitorable systemic hemodynamic changes are not a prerequisite for development of coronary arterial lesions in the dog [1,30,38,62]. For the ETRAs, lack of concordance with changes in HR and MAP indicates that monitoring of possible vascular hazard in humans is not possible using relatively crude physiological measures alone and extrapolation of the potential risk to the human population can only be made on the basis that the dog is a very sensitive species. This species sensitive contention is supported by the fact that significantly higher systemic exposure and longer duration treatment are required for a less severe lesion to develop in the monkey compared to the dog. 2.2. Regional blood flow Several reports provide direct and/or indirect evidence suggesting that drug-induced vascular injury is associated with vasodilatation and increased blood flow and these changes precede arterial damage (Table 2). For example, the potassium channel opener (PCO) minoxidil, a long lasting vasodilator, given to dogs at doses associated with profound hemodynamic changes (vasodilatation) induced a 6–10 fold increase in regional coronary blood flow (CBF) and arterial le-

18

D.A. Brott et al. / Current status and future directions for diagnostic markers of drug-induced vascular injury Table 2 Fold-change in regional blood flow caused by drug-induced vascular toxicants Pharmacological class Potassium channel

ETRA

Compound

Dose

Minoxidil JTV-506 Cromakalim Nicorandil Hydralazine SB209670

1 mg/kg 10 ug/kg 10 ug/kg 0.2 mg/kg 1 mg/kg 50 ug/kg

Coronary blood flow (fold increase) 4–6 4 3.4 3.6 1.5 2.2(LA) 6.1(RA)

Reference 27,44 67

11 38

LA left atrial, RA right atrial.

sions in the coronary vasculature [27,44]. Vasodilatation, increased regional CBF and coronary arterial lesions have been observed with other structurally and pharmacologically diverse agents SB209670 [38], hydralazine [11,45] and SK&F94836 [28]. Minor but sustained increases in HR and accompanying minor decreases in MAP and a surprising 6–10 fold increase in coronary regional blood flow was observed with SB209670, an endothelin receptor antagonist that caused coronary arterial lesions in dogs. It has been hypothesized that local increases in CBF is the basis for selective coronary arterial lesions in dogs. Under normal physiological conditions localized control and regulation of coronary blood flow is mediated in part by endogenous adenosine in response to increased oxygen demand. It is therefore not surprising that pharmacological mimicry of adenosine (i.e., adenosine agonists) by compounds such as imazodan (CI-914), CI947 and N-(2,2-Diphenylethyl) adenosine (DPEA) increase coronary arterial blood flow and induce coronary arterial lesions in dogs [3,5,7,39,46]. Therefore, it is now well accepted that administration of adenosine (A1) agonists as a pharmacological class is associated with coronary arterial lesions in dogs. A “class” effect for dog coronary arterial lesions has also been ascribed to ETRAs and PCOs because these agents cause profound increases in regional blood flow [11,27,30, 38]. In the rat, dopaminergic receptor activation by fenoldopam causes vasodilatation and mesenteric arterial lesions. These arterial lesions were prevented when fenoldopam was co-administered with a dopaminergic antagonist or the potent vasoconstrictor methoxamine [32]. Similarly, the potent vasoconstrictor, arginine vasopressin prevented the mesenteric arterial lesions induced by the potent vasodilator SK&F 95654, a phosphodiesterase inhibitor [31]. Minoxidil, SK&F 95654 and fenoldopam are all associated with varying degrees of increased mesenteric arterial blood flow in

the rat for prolonged periods [32]. These data collectively indicate that drug-induced mesenteric arterial lesions occur in part because of sustained vasodilatation and the resulting increased blood flow. Given the common association between increases in regional blood flow and vascular lesions, exploration of the contribution of this effect to the mechanism of injury and the subsequent changes in the vascular wall is worthy of future investigations. However, in dogs and humans under physiological conditions CBF can increase significantly during exercise and as such cannot be the only factor responsible for vascular injury [5]. In summary, localized vasodilatation, inability of the vasculature to maintain tone and alterations in regional blood flow are critical events in the development of arterial lesions in dogs and mesenteric arterial lesions in the rat. However, non-invasive monitoring of regional blood flow is tedious, with a low throughput, requires specialized instrumentation, a highly skilled technologist and thus is not practical on a large scale clinically.

3. Current status and progress Vascular injury and repair are a dynamically active processes that involve interactions between cellular and non-cellular components of the vessel wall. Understanding that the unfolding of these interactive biochemical events and processes in injured and/or repairing blood vessels offers an opportunity for the development of potential biological markers suitable for measurement in serum and/or plasma is critical. Circulating elements of the blood, particularly mediators of coagulation, vascular endothelium, the sub-endothelial collagen, EC and SMC as well as adventitial fibroblasts play an active and pivotal role in initiating and modulating the response of the vessel wall to injury. These events can culminate in an acute inflammatory


19

Fig. 3. vWF levels in rats. Rat platelet poor plasma contained an average of 91% vWF and saline washed platelets contained 67% vWF. Rat platelets contain significant amounts of vWF.

response followed by chronic inflammation and lesion resolution. Therefore, evaluating mediators of the coagulation cascade, secretory elements and constituents of endothelial and smooth muscle cells, as well as inflammatory cytokines would appear to hold substantial promise in the search for potentially useful diagnostic markers. Measurement of these cells and/or tissuederived products in plasma should be considered primary targets in this search.

4. Acute vascular injury 4.1. Acute phase proteins/mediators of the coagulation cascade Injury to the vasculature activates hemostatic mechanisms designed to rapidly plug and repair breaks in the vascular wall. Validated assays for the many proteins that mediate this process are not available and limit extensive evaluation in pre-clinical toxicology models. Other markers, irrespective of vascular injury are altered non-specifically when the coagulation cascade is activated. For example, changes in acute phase (fibrinogen, haptoglobin and C-reactive protein (CRP) pro coagulation (prothrombin time, partial thromboplastin time) and fibrinolytic proteins (D-dimer) can be altered in plasma but are not unique to vascular injury and thus are not appropriate for utilization as markers. Endothelial cell adhesion molecules could potentially be useful markers because they are either expressed or inducible upon EC activation in the process

Fig. 4. vWF levels in dogs. A, Three untreated dogs bled at 0, 3, 6, 12 and 24 hours. There is no difference in vWF levels with repeat bleeding. B, canine platelet poor plasma contained an average of 197% vWF and canine saline washed platelets contained 15% vWF. Dog platelets contain minimal amounts of vWF compared to normal plasma concentrations.

of vascular injury. A list of potential markers is shown in Table 3. Because drug-induced vascular injury is not primarily an inflammatory lesion the value of adhesion markers as a diagnostic tool is questionable. In summary, due to the multiplicity of conditions in which they are observed, acute phase proteins, mediators of coagulation and/or adhesion molecules are not appropriate indicators of drug-induced vascular injury. 4.2. Inflammatory markers The known association between inflammation, chronic vascular injury such as atherosclerosis and the increased incidence of mortality due to sudden cardiovascular events in humans has raised the level of safety concern for compounds that are associated with vascular injury pre-clinically. In humans, CRP, an acute phase protein and a marker of inflammation has been heralded as a diagnostic marker of vascular disease [47]. However, CRP has not been useful as a predictor or a reporter of

20

D.A. Brott et al. / Current status and future directions for diagnostic markers of drug-induced vascular injury Table 3 Evaluation of various markers for drug-induced vascular lnjury

Treatment LesionA DDAVP − Endotoxin − PCO-1 − PCO-2 + ETRA +

FibrinogenB + + − − +

Potential acute markers PT APTT D-dimer CRP Haptoglobin − − − − ND + + + + ND + − − − − ND ND ND − − + − − − −

Adhesion markers/activated platelets ICAM-1 pSelectin L-Selectin Activated platelet ND ND ND ND ND ND ND ND − − − − − − − − − − − −

Dogs were treated with 5 ug/mL DDAVP, 2 mg/mL Endotoxin, 200 mg/kg ZD6169 (PCO-1), 240 mg/kg ZD6169 (PCO-2) or 300 mg/kg ZD1611. A Prescence of lesions were determined 24 hous post dosing. −, lesions were not present; +, lesions were present. B Parameters were evaluated pre-dose and 3, 6, 12 and 24 hous post-dose, +, values were changed in the animals compared to 0 hour value; −, values were not changed in animals compared to 0 hour; ND, parameter was not evaluated in this group of animals.

drug-induced vascular injury in rats, dog and/or monkeys [21]. Other markers such as interleukin-6, and other cytokines have been suggested but these require further study [40]. Adverse perturbation of the immune system with subsequent inflammation can cause damage to SMC and EC. Some compounds may interact directly with the vascular wall and stimulate a chemotactic and inflammatory response [9]. In this case, circulating soluble chemokines/cytokines released at the site of inflammation may be measurable in serum or plasma and constitute a potential diagnostic marker. However, the cellular interactions between adhesion molecules, the vascular wall and immune cells make it difficult to identify unique markers that are diagnostic for this complex process. The vasculature can become secondarily inflamed as a response by the host tissue to adjacent severe inflammation. For example, an acute neutrophilic response can be a major toxic effect with subsequent acute inflammation of the mesenteric arteries in the rat ([64]; Kerns, personal communication;). Monitoring of neutrophil counts and serum cytokines may serve as useful markers for this process that are both morphologically and pathophysiologically distinct from drug-induced lesions. In summary, the issues of specificity and sensitivity must be resolved before adhesion molecules, inflammatory markers and cytokines can be identified and then validated as vascular injury markers. 4.3. Endothelial cell markers Physiological and/or pathological perturbation of the EC causes expression of numerous cell surface receptors and secretion of many proteins with potential diagnostic utility [37]. Of these many proteins, (Table 4) and specifically related to vascular injury, von Willebrand Factor (vWF), von Willebrand Factor propeptide (vWFpp), endothelins (ET-1), soluble E-selectin, thrombomodulin and nitric oxide synthase (NOS) have

gained most attention. The relative strengths and weaknesses of these molecules as potential diagnostic markers are presented in Table 4. vWF, is a potential diagnostic marker of vascular injury because it is synthesized by EC and released into the plasma. This analyte was evaluated as a marker of acute arterial damage in rats after a vasotoxic dose of a potent dopaminergic vasodilator, fenoldopam [48]. Plasma vWF levels transiently increased in rats prior to morphologic evidence of vascular damage. However, vWF values returned to baseline 24-hours postdosing when vascular damage was present histologically. It is clear that increases in plasma vWF preceded fenoldopam-induced mesenteric arterial damage. The authors concluded that vWF was not a reliable diagnostic marker of vascular injury because increases were transient and minimal and there was a lack of concordance between severity of arterial damage and the magnitude of plasma elevations when lesions were apparent. In our studies using a similar experimental protocol we made similar observations but in addition, showed marked increases in vWF immunoreactivity in the wall of fenoldopam-damaged arteries (Fig. 2). These findings raise the possibility that in damaged arteries with increased permeability and loss of vascular integrity, vWF, which is released from the basolateral surface of EC, diffuses into the vessel wall instead of plasma and this may help to explain the lack of concordance between lesion severity and plasma vWF levels. In support of this conclusion, by comparison with controls, significant increases in plasma vWF at 2 and 6 hours post dosing were observed in fenoldopam treated rats with no histologic evidence of arterial damage [48]. In addition to plasma, platelets trapped in the vessel wall may be another element that could provide an explanation of vWF immunoreactivity in fenoldopamdamaged arteries. Rat platelets have an abundance of vWF (Fig. 3) that could be potentially released upon activation and which explain minor increases in plasma


21

Table 4 Potential biomarkers of endothelial cells Endothelial Biomarkers vWF

Reagents for Human Dog Rat Y Y Y

Comment

References

Increased prior to lesion formation in rats and dogs Decreased overtime, potentially due to vWF binding within lesion Does not correlate with lesion progression or severity

35

vWFPP

Y

Y

N

Secreted in equal molar concentration with vWF, in humans Some human antibodies cross-react with dog Has a short half-life Need to produce species specific cross-reactive reagents

63

CEC by flow cytometry CEC by CFU-E Endothelin Thrombomodulin sE-selectin Nitric oxide

Y Y Y Y Y Y

Y N Y Y N Y

Y N Y Y N Y

Would be a marker of end organ damage and lesion resolution Length of assay time and lack of dog and rat assays is of concern Does not seem to correlate with PCO induced vascular injury Plasma levels increase with endothelial cell injury Lack of dog and rat assay is of concern Not a specific marker of endothelial cell injury

41,57 20 6 49 19 35

CEC circulating endothelial cells. CFU-E colony-forming unit endothelial.

Fig. 5. Plasma vWF levels of dogs treated with a potassium channel opener (PCO). Blood was collected pre-dose, 3, 6, 12 and 24-hours post-dose. PCO treatment caused a transient increase in vWF levels.

vWF following repeated venipuncture in saline-treated rats. By contrast, in dogs, vWF is not increased following repeated venipuncture, as dog platelets do not have an appreciable quantity of vWF (Fig. 4). Based on these data it is reasonable to conclude that the process of repeated venipuncture in rats and dogs does not damage EC appreciably and in saline treated rats the platelet plug at the injection site is the likely source for the increase in plasma vWF. Therefore, we conclude, that in fenoldopam treated rats, EC in damaged mesenteric arteries are the likely source of increased plasma vWF at two and six hours post dosing before the onset of morphologic evidence of damage and thus, supports vWF as a valid biomarker.

In dogs, the likely source of plasma vWF is EC and so we evaluated plasma from dogs treated with a PCO to determine the potential quality of this analyte as a diagnostic marker of acute vascular injury (Fig. 5). Single dose administration of a PCO caused vWF values to increase greater than 60% as early as 3 hours post-dosing and gradually declined, returning to baseline values at 24 hours post-dosing. The ETRA caused an increase in plasma vWF 6 hours or later, post-dosing (data not shown). Overall, plasma vWF values increased in five of six treated dogs at doses known to induce coronary arterial injury. It is well recognized that in the normal vasculature, vWF is localized to endothelial cells and damage presumably causes leakage and consequent in-

22


of Dr. Robert Montgomery, Department of Pediatrics, Medical College of Wisconsin, Milwaukee, WI and the Blood Center of Southeastern Wisconsin, Milwaukee, WI). Standard curves are comparable for human vWF propeptide, normal pooled and platelet poor dog plasma (Fig. 6). Using this assay, our preliminary data shows that exaggerated physiological and/or pathological perturbation of EC can cause significant increases in plasma vWF and vWFpp. 4.4. Smooth muscle cell markers

Fig. 6. vWF propeptide (vWFpp) standard curves. Closed boxes, human pooled plasma; open triangles, platelet poor canine plasma; and open diamonds, canine pooled plasma.

creased plasma levels. We evaluated damaged arteries of varying severity to determine if changes of vWF localization occur as a result of vascular injury. Immunohistochemical assessment showed a direct relationship between lesion severity and accumulation of vWF extracellularly within the media of damaged arteries (Fig. 2). Ultrastructural investigation did not provide indication of myocardial capillary endothelial damage that would result in leakage of vWF. These findings in dog plasma and damaged arteries support our observations as well as the published literature in the rat [48]. In summary, those involved in studying drug-induced vascular injury, (rats and dogs) would be well advised to evaluate vWF in order to assess acute transient effects before impending damage, although for purposes of monitoring progressive injury, vWF is unreliable. Mature vWF, is secreted primarily by endothelial cells in the dog and therefore elevation in plasma is highly suggestive of vascular endothelial compromise. After synthesis the pro vWF molecule undergoes endoproteolytic cleavage that causes release of mature vWF and vWF propeptide (vWFpp). Therefore, we wished to determine if there was a concomitant increase in plasma vWFpp in dogs treated with a single dose of PCO, multiple doses of ETRA, DDAVP or Endotoxin. Pro-peptide levels were measured by an enzyme-linked immunosorbent assay using specific antibodies that recognize both dog and human vWF propeptide (courtesy

In our studies, evaluation of smooth muscle α actin (SMA) immunohistochemically shows decreased immunoreactivity in injured SMC of damaged arteries only (Fig. 7). Loss of SMA immunoreactivity specifically at the site of vascular damage raises the possibility that this protein qualifies as a potential diagnostic marker of drug-induced vascular injury. Conceivably, SMA could be released in circulation as a result of arterial damage. SMA has been evaluated in urine to determine its value as a marker of chronic renal damage [9]. Therefore, evaluation of this analyte in urine, serum or plasma in an animal model of drug-induced vascular injury should be considered, but is challenging for the following reasons (1) analytical methods for assessment in serum/plasma are currently not available (2) significant quantity of SMA in plasma/serum would indicate that cell damage/death has occurred and (3) the kinetics associated with release of SMA into plasma and cell death are unknown. This marker has potential utility as a biomarker of drug-induced vascular injury but needs further characterization.

5. Novel mechanistically linked markers While vasodilatation and increased shear stress appear to play a role in drug-induced vascular injury, the exact mechanism of SMC and EC injury/death is unknown. There are gaps in our knowledge and understanding of the biochemical events and mediators that induce this injury and elucidating the specific biochemical pathways could yield potentially novel and mechanistically linked diagnostic markers. Our current approach is to dissect and identify the biochemical pathway(s) likely to be engaged in injury and/or death of SMC and EC subjected to increased shear stress as a result of excessive vasodilatation. We hypothesize, that measurable amounts of soluble proteins are released from endothelial and/or smooth muscle cells or the vas-


23

Fig. 7. Smooth muscle alpha actin (SMA) immunohistochemical staining of canine coronary arteries. Control sample showing high expression within the smooth muscle cell layer. Endothelin receptor antagonist-induced arterial injury. Note, loss of SMA within the smooth muscle cell layer.

cular extracellular matrix during drug-induced vascular injury caused by localized vasodilatation, loss of arterial tone and increased CBF mediated by the potent vasodilator nitric oxide and nitric oxide synthase which is physically linked to and regulated by caveolae, specifically cav-1. Caveolae are plasmalemmal invaginations that play an important role in transcytosis, molecular transport and signal transduction [4,10,23,36,42,53,58]. They appear to be the focal point for compartmentalization, organization and modulation of signal transduction activities for many receptors and enzymes that are colocalized to caveolae. Cav-1 is not only the major structural protein of caveolae [53], but it also modulates the function of signal transduction particularly in EC and VSMC, the primary targets of drug-induced vascular injury. In addition, adenylyl cyclase (AC) and other components of the cAMP pathway [18,29,33,51, 65] and nitric oxide synthase (NOS) are co-localized with cav-1. We know that cAMP down-regulates cav-1 expression [50] and that cav-1 is a negative regulator of NOS and hence indirectly regulates nitric oxide (NO) production [17,22]. Therefore, we would expect that compounds that increase cAMP would down-regulate Cav-1, induce NOS activation and lead to continuous NO production. Nitric oxide, a free radical, plays a pivotal role in vasodilatation but excesses can lead to cell damage [16]. For example, in mice, targeted gene disruption of Cav-1 caused impairment of nitric oxide synthase, nitric oxide (NO) and calcium signaling in the cardiovascular system causing aberrations in endothelium-dependent relaxation, contractility and maintenance of vascular myogenic tone [15]. Cav-1 down-regulation was also associated with a 5-fold increase in systemic NO, but no major changes in protein production of NOS [68], suggesting that deregulation and/or increased activity of NOS were responsible for the marked increase in NO production. Cav-1 down-

regulation leads to activation of AC, increased cAMP levels (observed with the structurally and pharmacologically diverse compounds that cause vascular lesions; Table 5) and increased NOS activity which could lead to the generation of undesirable, potentially cytotoxic, super-physiological bursts of NO that are detrimental to both vascular smooth muscle and/or endothelial cells. Thus, sustained and prolonged NOS activation causes elevated NO levels causing profound localized vasodilatation and increased blood flow. It is possible therefore that drug-induced vascular injury could be caused by a synergistic effect of NO-induced vasodilatation and free radical induced cell injury. These data suggest that Cav-1 plays an important role in regulating NOS activity and is a potentially important influence in drug-induced vascular injury. Dissection of the relationship between loss of cav1, and increases in NOS activity, NO, vasodilatation and vascular injury will yield mechanistically linked biomarkers that will improve our capability for risk assessment of drug-induced vascular lesions. Therefore, a series of studies and experiments were conducted to determine if Cav-1 qualifies as a mechanistically linked diagnostic biomarker of drug-induced vascular injury and if decreased levels of Cav-1 precede lesion formation. 5.1. Caveolin-1 expression- in vivo It is clear from our studies with vWF that a useful diagnostic marker of progressive vascular injury is one that is expressed on EC and VSMC and undergoes compartmental shift from the damaged vascular wall to the circulation as the severity of injury progresses. To determine if Cav-1 qualified as such a marker, we determined localization and expression of this membrane protein in control dog coronary and rat mesenteric arter-

24

D.A. Brott et al. / Current status and future directions for diagnostic markers of drug-induced vascular injury Table 5 Primary and secondary targets of pharmacologically and structurally diverse compounds known to produce druginduced vascular injury Compound class βadrenergic PDE inhibitors Potassium channel openers Adenosine/agonists ETRA Dopamine receptor agonists

(i.e. Compound) isoproterenol 3-isobutyl-1-methylxanthine IBMX) ZD6169 cyclohexyladenosine ZD1611 fenoldopam

Primary target β receptor PDE III & IV SMC K+ channels A1 or A2 receptors ETA1 or ETB1 receptors DA1 & DA2 Receptors

Secondary target ↑cAMP ↑cAMP ↑cAMP ↑cAMP ↑cAMP ↑cAMP

Fig. 9. Western blot analysis of caveolin-1 in rat A10 smooth muscle cells. A10 cells express caveolin-1 that is lost in a dose dependent manner with Fenoldopam, minoxidil or IBMX treatment.

Fig. 8. Caveolin-1 staining of dog coronary and rat mesenteric arteries. Dog control arteries express high levels of caveolin-1 that is lost within the lesions of injured arteries. Rat control arteries express caveolin-1, albeit at a lower level than dogs, which is lost with fenoldopam treatment.

ies as well as drug-induced lesions of increased severity (Fig. 8). In normal arteries, regardless of species, Cav-1 is expressed exclusively on EC and VSMC. In damaged coronary arteries from dogs treated with either a PCO or an ETRA, we observed decreases and/or loss of cav-1 expression in EC and VSMC. Importantly,this decrease and/or loss of cav-1 expression was confined only to the site of EC and VSMC cell damage. Morphologically, cav-1 expression in normal-appearing (unaltered) regions of damaged arteries cav-1 was comparable to that in undamaged control arteries. The loss of cav-1 in injured arteries had a segmental distribution similar to that observed with hematoxylin and eosin stained sections. Similar findings, though less pronounced, were observed in the rat model of acute drug-induced (fenoldopam) mesenteric vascular injury. These data suggest that Cav-1 qualifies as a novel and potentially unique biomarker because (1) it is expressed only in EC and VSMC, targets of vascular injury (2) in contrast to vWF that partially compartmentalizes in the wall of severely damaged arteries, cav-1 loss undergoes com-

partmental shift from the vascular wall to the circulation allowing this analyte to be measured in serum and/or plasma (3) sustained and/or continuous elevation is suggestive of sustained and progressive injury. We rationalized that cav-1 is a novel, mechanistically linked diagnostic marker of vascular integrity based on cellular localization, expression and biological function. To better quantify loss of cav-1 as an indicator of vascular injury we determined if this event (loss of cav1) precedes EC and/or VSMC cell death and developed an in vitro cell system to address this. A smooth muscle cell line (A10) was selected because in vivo, SMC is the primary target of drug-induced vascular injury. Western blot analysis determined that these A10 SMC cells express cav-1 (Fig. 9). The broad-spectrum phosphodiesterase inhibitor (IBMX), minoxidil (a potassium channel opener) and fenoldopam (a dopaminergic agonist) that are known vascular toxicants in animals, decreased the level of cav-1 protein in vitro. In addition, morphometric analysis of immunofluorescence expression of cav-1 in vitro with cellomics supported our western blot data (data not shown). Since there was a loss of cav-1 in vivo as lesion severity increased, it was imperative to determine if this loss precedes smooth muscle cell death. To address this question, using cav-1 loss and apoptosis as end points, we performed flow cytometric analysis of A10 cells treated with pharmaco-


25

Fig. 10. Flow cytometric analysis of rat A10 smooth muscle cells. Caveolin-1 and apoptosis analysis of A10 cells treated with IBMX or minoxidil. Note, caveolin-1 loss occurs by 6 hours, but apoptosis did not occur until 24 hours. Therefore caveolin-1 loss preceded cell death.

logically diverse vascular toxicants. The flow cytometry data clearly showed that loss of cav-1 occurred at 6 hours post-dosing with either IBMX or minoxidil in the absence of cell death (Fig. 10). However, both of these compounds induced apoptosis at 24-hours. Therefore, loss of cav-1 could be an early event prior to cell death in the evolution of drug-induced vascular injury. These findings qualify cav-1 as a potentially novel mechanistically linked diagnostic marker. Future investigations will include (1) determination of basal levels of circulating cav-1 in control dogs (2) monitoring of circulating cav-1 in dogs and rats treated with known vascular toxicants (3) determination of the relationship between circulating cav-1 and lesion severity.

6. Future directions to identify biomarkers of drug-induced vascular injury Applications of new technologies will undoubtedly aid in identification of valuable diagnostic markers of drug-induced vascular injury. Utilization of omics (genomics, proteomics, metabonomics) and flow cytometry platforms to evaluate proteins, cells and/or cell fragments may be useful. For example, flow cytometry has wide applications in the study of circulating

cells released from damaged arteries because it has the ability to discriminate the phenotypic variability in the EC population of arterial, venous and capillary beds thus allowing the identification of site-specific markers [12]. This is of particular value given the complexity of interactions and the heterogeneous cell population activated in drug-induced vascular injury. Flow cytometric techniques were employed to aid in identification and confirmation of the origin of circulating EC cells (CEC) using a limited genetic analysis that included positive and negative selection [41,57]. The data showed a considerable variation in both the phenotypes and genotypes of CEC when whole blood was analyzed using an expression panel of well-recognized EC antigens that included vWF and CD31. Transcriptional analysis showed considerable heterogeneity of sorted CEC used to monitor and assess onset and progression of drug-induced vascular injury in rats. The authors concluded that the phenotypic and genotypic heterogeneity was a reflection of the complex nature, site of origin and stage of differentiation of the released CEC. In summary, the use of immunophenotypic labeling, flow cytometric sorting and transcriptional profiling of CEC in whole blood qualify as a diagnostic marker of drug-induced vascular injury and warrants further

26


study. Another potential utility is to monitor lesion onset, progression and resolution. Available mechanistic evidence of drug-induced vascular injury supports the concept that there is a close association between desired pharmacological response, vascular toxicity and the molecular target. Therefore, inclusion of genomics in investigative protocols of drug-induced vascular injury will provide a valuable mechanistic link between gene and protein expression and the physiologic/pathologic outcome at the tissue and cellular levels. Arterial wall heterogeneity, cross contamination with adjacent surrounding tissues such as pancreas, liver, fat (mesentery) and myocardial tissue (coronary artery) may distort the significance of observed changes when interpreting transcriptional profiling of expressed genes in the primary target cells of drug-induced vascular injury. Given the segmental nature of the lesion, collection of the specific site of alteration for comparison with normal undamaged segments from the same vessel can be challenging. Even with these limitations, which may be eased by the use of laser microdissection microscopy of suitable tissue sections, genomics has the potential to provide information that will aid our understanding of genetic events and aid in identification of novel proteins that are unique to drug-induced vascular injury. Proteomics uses body fluids for detection of proteins over wide molecular-weight range. The twodimensional gel method is commonly used for proteomics, but surface-enhanced laser desorption ionization time-of flight mass spectrometry (SELDITOF/MS) and multidimensional liquid chromatography are additional methods. Proteomics will aid in the detection of novel diagnostic markers of drug-induced vascular injury because it can detect proteins across a wide range of molecular mass and greater than 10 kDa. However, albumin and other abundant proteins can interfere with the detection of low abundance proteins. A new method was recently published that allows the detection of 0.5–15 kDa small proteins and peptides [56]. Limitations for wider acceptance and use of this technology include:(1) difficulty in identification of the many proteins present on a gel (2) labor intensive evaluation of data and (3) low-throughput technology. Metabonomics is rapidly gaining acceptance as a powerful tool in the search for biomarkers of tissue injury because it can determine the biochemical profile of blood or urine samples using high-resolution nuclear magnetic resonance (NMR) spectrometry. For example, in urine, a unique metabolic profile was identified in rats given a PDE IV inhibitor (CI-1018) that caused

mesenteric arterial lesions [52,59]. Follow-up investigations using this high throughput technology should include an evaluation of pharmacologically and structurally diverse compounds that are known vascular toxicants in rats, dogs and monkeys.

7. Conclusions The realization that HR and MAP were flawed indicators of drug-induced vascular injury led to their exclusion as universal surrogate markers of drug-induced vascular injury in the 1990s. As a consequence, an urgent requirement emerged for new biomarkers of druginduced vascular injury with the ultimate goal of identifying a mechanistically linked marker that is altered prior to cell injury/death. The last ten to fifteen years have seen a dramatic increase in this research effort both within academia and the pharmaceutical industry including the convening of the “expert working group on drug-induced vascular injury” by the FDA. The expert working group outlined several potential markers of drug-induced vascular injury for further investigation and the need for proteomic/metabonomics research into the identification of new markers. Investigation of the various potential biomarkers has progressed, but considerably more work is needed to define a unique panel of biomarkers.

References [1]

[2]

[3]

[4] [5]

[6]

[7]

M.A. Albassam, A.L. Metz, N.J. Gagmans, L.M. King, G.E. Macallum, H. Hallak and E.J. McGuire, Coronary arteriopathy in monkeys following administration of CI-1020, an endothelin A receptor antagonist, Toxicol Pathol 27 (1999), 156–164. M.A. Albassam, A.L. Metz, R.E. Potoczak, K.P. Gallagher, S. Haleen, H. Hallak and E.J. McGuire, Studies on coronary arteriopathy in dogs following administration of CI-1020, and endothelin a receptor antagonist, Toxicol Pathol 29 (2001), 277–284. M.A. Albassam, G.S. Smith and G.E. Macallum, Arteriopathy induced by an adenosine agonist-antihypertensive in monkeys, Toxicol Pathol 26 (1998), 375–380. R.G. Anderson, The caveolae membrane system, Annu Rev Biochem 67 (1998), 199–225. A.N. Bacchus, S.W. Ely, R.M. Knabb, R. Rubio and R.M. Berne, Adenosine and coronary blood flow in conscious dogs during normal physiological stimuli, Am J Physiol 243 (1982), H628–H633. B. Battistini, P. D’Orleans-Juste and P. Sirois, Endothelins: circulating plasma levels and presence in other biologic fluids, Lab Invest 68 (1993), 600–628. R.M. Berne, The role of adenosine in the regulation of coronary blood flow, Circ Res 47 (1980), 807–813.

D.A. Brott et al. / Current status and future directions for diagnostic markers of drug-induced vascular injury [8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19] [20]

[21]

[22]

[23]

S.P. Bishop, P.C. Powell, N. Hasebe, Y.I. Shen, T.A. Patrick, L. Hittinger and S.F. Vatner, Coronary vascular morphology in pressure-overload left ventricular hypertrophy, J Mol Cell Cardiol 28 (1996), 141–154. P.J. Boor, A.I. Gotlieb, E.C. Joseph, W.D. Kerns, R.A. Roth and K.E. Tomaszewski, Contemporary issues in toxicology: Chemical-induced vasculature injurym Tox Appl Pharm 132 (1995), 177–197. M. Bucci, J.P. Gratton, R.D. Rudic, L. Acevedo, F. Roviezzo, G. Cirino and W.C. Sessa, In vivo delivery of the caveolin-1 scaffolding domain inhibits nitric oxide synthesis and reduces inflammation, Nature Medicine 6 (2000), 1362–1367. J.E. Chelly, M.F. Doursout, B. Begaud, C.C. Tsao and C.J. Hartley, Effects of hydralazine on regional blood flow in conscious dogs, J Pharmacol Exp Ther 238 (1986), 665–669. D.B. Cines, E.S. Pollak, C.A. Buck, J. Loscalzo, G.A. Zimmerman, R.P. McEver, J.S. Pober, T.M. Wick, B.A. Konkle, B.S. Schwartz, E.S. Barnathan, K.R. McCrae, B.A. Hug, A.M. Schmidt and D.M. Stern, Endothelial cells in physiology and in the pathophysiology of vascular disorders, Blood 91 (1998), 3527–3561. F.A.S. Clemo, W.E. Evering, P.W. Snyder and M.A. Albassam, Differentiating spontaneous from drug-induced vascular injury in the dog, Toxicol Pathol 31(suppl) (2003), 25–31. P. Dogterom and G. Zbinden, Cardiotoxicity of vasodilators and positive inotropic/vasodilating drugs in dogs: an overview, Crit Rev Toxicol 22 (1992), 203–241. M. Drab, P. Verkade, M. Elger, M. Kasper, M. Lohn, B. Lauterbach, J. Menne, C. Lindschau, F. Mende, F.C. Luft, A. Schedl, H. Haller and T.V. Kurzchalia, Loss of caveolae, vascular dysfunction, and pulmonary defects in Caveolin-1 gene-disrupted mice, Science 293 (2001), 2449–2452. C.L. Fattman, L.M. Schaefer and T.D. Oury, Extracellular superoxide dismutase in biology and medicine, Free Radical Bio and Med 35 (2003), 236–256. O. Feron and R.A. Kelly, The caveolar paradox: suppressing, inducing and terminating eNOS signaling, Circ Res 88 (2001), 129–131. G. Garcia-Cardena, P. Marasek, B.S.S. Masters, P.M. Skidd, J. Couet, S. Li, M.P. Lisanti and W.C. Sessa, Dissecting the interaction between nitric oxide synthase (NOS) and caveolin, J Bio Chem 272 (1997), 25437–25440. A.J. Gearing and W. Newman, Circulating adhesion molecules in disease, Immunol Today 14 (1993), 506–512. J. George, E. Goldstein, S. Abashidze, V. Deutsche, H. Shmilovich, A. Finkelstein, I. Herz, H. Miller and G. Keren, Circulating endothelial progenitor cells in patients with unstable angina: association with systemic inflammation, Eur Heart J 25 (2004), 1003–1008. P.S. Giffen, J. Turton, C.M. Andrews, P. Barrett, C.J. Clarke, K.W. Fung, M.R. Munday, I.F. Roman, R. Smyth, K. Walshe and M.J. York, Markers of experimental acute inflammation in the wistar han rat with particular reference to haptoglobin and c-reactive protein, Arch Toxicol 77 (2003), 392–402. M.S. Goligorsky, H. Li, S. Brodsky and J. Chen, Relationships between caveolae and eNOS: everything in proximity and the proximity of everything, Am J Physiol Renal Physiol 283 (2002), F1–F10. M. Haas, S.M. Meehan, M.A. Josephson, E.J. Wit, E.S. Woodle and J.R. Thistlethwaite, Smooth muscle-specific actin levels in urine of renal transplant recipients: correlation with cyclosporine or tacrolimus nephrotoxicity, Am J Kidney Dis 34 (1999), 69–84.

[24] [25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

27

H.A. Hartman, Spontaneous extramural coronary arteritis in dogs, Toxicol Pathol 17 (1989), 138–144. E.H. Herman, V.J. Ferrans, R.S. Young and T. Balazs, A comparative study of minoxidil-induced myocardial lesions in beagle dogs and miniature swine, Toxicol Pathol 17 (1989), 182– 192. S.J. Humphrey and F.J. Schwende, Cardiovascular interactions characterizing the vasodilator components of nicorandil in conscious dogs, J Cardiovasc Pharm 26 (1995), 810–821. S.J. Humphrey and G.R. Zins, Whole body and regional hemodynamic effects of minoxidil in the conscious dog, J Cardiovasc Pharmacol 6 (1984), 979–988. K.R. Isaacs, E.C. Joseph and G.R. Betton, Coronary vascular lesions in dogs treated with phosphodiesterase III inhibitors, Toxicol Pathol 17 (1989), 153–163. N. Ishizaka, K.K. Giendling, B. Lassegue and R.W. Alexander, Angiotensin II type 1 receptor: relationship with caveolae and caveolin after initial agonist stimulation, Hypertension 32 (1998), 459–466. H.B. Jones, A. Macpherson, G.R. Betton, A.S. Davis, R. Siddall and P. Greaves, Endothelin antagonist-induced coronary and systemic arteritis in the beagle dog, Toxicol Pathol 31 (2003), 263–272. E.C. Joseph, Arterial lesions induced by phosphodiesterase III (PDE III) inhibitors and DA1 agonists, Toxicology Letters 112–113 (2000), 537–546. E.C. Joseph, J.A. Rees and A.D. Dayan, Mesenteric arteriopathy in the rat induced by phosphodiesterase III inhibitors: an investigation of morphological, ultrastructural and hemodynamic changes, Toxicol Pathol 24 (1996), 436–450. L. Labrecque, I. Royal, D.S. Surprenant, C. Patterson, D. Gingras and R. Beliveau, Regulation of vascular endothelial growth factor receptor-2 activity by caveolin-1 and plasma membrane cholesterol, Mol Biol Cell 14 (2003), 334–347. C.S. Liang, A. Thomas, N. Imai, C.K. Stone, S. Kawashima and W.B. Hood, Jr., Effects of milrinone on systemic hemodynamics and regional circulations in dogs with congestive heart failure: comparison with dobutamine, J Cardiovasc Pharmacol 10 (1987), 509–516. G.Y. Lip and A. Blann, von willebrand factor: a marker of endothelial dysfunction in vascular disorders, Cardiovasc Res 34 (1997) 255–265. P. Liu, M. Rudick and R.G.W. Anderson, Multiple functions of caveolin-1, J Bio Chem 277 (2002), 41295–41298. b C. Louden and D.G. Morgan, Pathology and pathophysiology of drug-induced arterial injury in laboratory animals and its implications on the evaluation of novel chemical entities for human clinical trials, Pharmacol Tox 89 (2001), 158–170. C.S Louden, P. Nambi, M.A. Pullen, R.A. Thomas, L.A. Tiernery, H.A. Solleveld and L.W. Schwartz, Endothelin receptor subtype distribution predisposes coronary arteries to damage, Amer. J Pathol 157 (2000), 123–134. G.E. Macallum, R.M. Walker, N.J. Barsoum and G.S. Smith, Preclinical toxicity studies of an adenosine agonist, N-(2,2diphenylethyl) adenosine, Toxicology 68 (1991), 21–35. E. McDuffie, X. Yu, Y. Song and M. Albassam, Potential markers of acute coronary artery injury in dogs following administration of CI-1034, an endothelin A receptor antagonist, The Toxicologist 78 (2004), 1820a. D.C. McFarland, M.S. Scicchitano, R.A. Thomas, P.K. Narayanan, L.W. Schwartz and H.C. Thomas, 6-color flowsorting of rat circulating endothelial cells and TaqMan real – time PCE analysis, Cytometry 54A (2004), 65.

28 [42]

[43]

[44]

[45]

[46]

[47] [48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

D.A. Brott et al. / Current status and future directions for diagnostic markers of drug-induced vascular injury D.P. McIntosh, X.Y. Tan, P. Oh and J.E. Schnitzer, Targeting endothelium and its dynamic caveolae for tissue-specific transcytosis in vivo: a pathway to overcome cell barriers to drug and gene delivery, PNAS 99 (2002), 1996–2001. G.M. Mesfin, M.J. Higgins, F.G. Robinson and W.Z. Zhong, Relationship between serum concentrations, hemodynamic effects, and cardiovascular lesions in dogs treated with minoxidil, Toxicol Appl Pharmacol 140 (1996), 337–344. G.M. Mesfin, R.C. Piper, D.W. DuCharme, R.G. Carlson, S.J. Humphrey and G.R. Zins, Pathogenesis of cardiovascular alterations in dogs treated with minoxidil, Toxicol Pathol 17 (1989), 164–181. G.M. Mesfin, G.G. Shawaryn and M.J. Higgins, Cardiovascular alterations in dogs treated with hydralazine, Toxicol Pathol 15 (1987), 409–416. A.L. Metz, M.A. Dominick, G. Suchanek and A.W. Gough, Acute cardiovascular toxicity induced by an adenosine agonist-antihypertensive in beagles, Toxicol Pathol 19 (1991), 98–107. L. Mosca, C-reactive protein – to screen or not to screen? N Engl J Med 347 (2002), 1615–1617. S.J. Newsholme, D.T. Thudium, K.A. Gossett, E.S. Watson and L.W. Schwartz, Evaluation of plasma von willebrand factor as a biomarker for acute arterial damage in rats, Toxicol Pathol 28 (2000), 668–693. S. Ohdama, S. Takano, S. Miyake, T. Kubota, K. Sato and N. Aoki, Plasma thrombomodulin as a marker of vascular injuries in collagen vascular diseases, Am J Clin Pathol 101 (1994), 109–113. H. Park, Y.M. Go, R. Darji, J.W. Choi, M.P. Lisanti, M.C. Maland and H. Jo, Caveolin-1 regulates shear stress-dependent activation of extracellular signal-regulated kinase, Am J Physiol Heart Circ Physiol 278 (2000), H1285–H1293. V. Rizzo, D.P. McIntosh, P. Oh and J.E. Schnitzer, In situ flow activates endothelial nitric oxide synthase in luminal caveolae of endothelium with rapid caveolin dissociation and calmodulin association, J Bio Chem 273 (1998), 34724–34729. D.G. Robertson, M.D. Reily, M. Albassam and L.A. Dethloff, Metabonomics assessment of vasculitis in rats, Cardiovasc Toxicol 1 (2001), 7–19. K.G. Rothberg, J.E. Heuser, W.C. Donzell, Y.S. Ying, J.R. Glenney and R.G.W. Anderson, Caveolin, a protein component of Caveolae membrane coats, Cell 68 (1992), 673–682. Z. Ruben, P. Deslex, G. Nash, N.I. Redmond, M. Poncet and D.C. Dodd, Spontaneous disseminated panarteritis in laboratory beagle dogs in a toxicity study: a possible genetic predilection, Toxicol Pathol 17 (1989), 145–152. G.E. Sandusky, J.R. Means and H.K. Yeager, Coronary artery toxicity in beagle dogs given indolidan, a type III phosphodiesterase inhibitor, Fundam Appl Toxicol 21 (1993), 228–235.

[56]

[57]

[58]

[59]

[60] [61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

P. Schulz-Knappe, H.D. Zucht, G. Heine, M. Jurgens, R. Hess and M. Schrader, Peptidomics: the comprehensive analysis of peptides in complex biological mixtures, Comb Chem High Throughput Screen 4 (2001), 207–217. M. Scicchitano, R. Thomas, D. McFarland, P. Narayanan, H. Thomas, L. Tierney and L. Schwartz, Transcriptional phenotyping of circulating endothelial cells from sorted rat whole blood, Veterinary Pathology Meeting 40 (2003), 626. P.W. Shaul and R.G.W. Anderson, Role of plasmalemmal caveolae in signal transduction, Am J Physiol (Lung Cell Mol Physiol) 19 (1998), L843–L851. R.M. Slim, D.G. Robertson, M. Albassam, M.D. Reily, L. Robosky and L.A. Dethloff, Effect of dexamethasone on the metabonomics profile associated with phosphodiesterase inhibitor-induced vascular lesions in rats, Toxicol Appl Pharmacol 183 (2002), 108–116. J.T. Sobota, Review of cardiovascular findings in humans treated with minoxidil, Toxicol Pathol 17 (1989), 193–202. J.T. Sobota, W.B. Martin, R.G. Carlson and E.S. Fenestra, Minoxidil: right atrial cardiac pathology in animals and in man, Circulation 62 (1980), 376–387. M. Stephan-Gueldner and A. Inomata, Coronary arterial lesions induced by endothelin antagonists, Toxicology Letters 112–113 (2000), 531–535. U.M. Vischer J. Ingerslev, C.B. Wollheim, J. Mestries, D.A. Tsakiris, W.A. Haefeli and E.K. Kruithof, Acute von willebrand factor secretion from the endothelium in vivo; Assessment through plasma propeptide (vWf:AbII), Thromb and Haemost 77 (1997), 387–393. White paper, FDA. Drug-induced vascular injury – a search for biomarkers, (Sept 9, 2002) at www.fda.gov/ohrms/dockets/ac/ 02/briefing/3891B1 04 vewg.ncss.doc. T. Yamaguchi, Y. Murata, Y. Fujiyoshi and T. Doi, Regulated interaction of endothelin B receptor with caveolin-1, Eur J Biochem 270 (2003), 1816–1827. E.M. Yuhas, D.G. Morgan, E. Arena, R.P. Kupp, L.Z. Saunders and H.B. Lewis, Arterial medial necrosis and hemorrhage induced in rats by intravenous infusion of fenoldopam mesylate, a dopaminergic vasodilator, Am J Pathol 119 (1985), 83–91. S. Zhang, K. Sakai, H. Hasuo, K. Furuta, H. Ureshino, O. Shibata and K. Sumikawa, Systemic and coronary hemodynamic effects of JTV-506, a novel potassium channel opener, in conscious dogs: comparison with cromakalim and nicorandil, Res Commun Mol Pathol Pharmacol 105 (1999), 115–127. Y.Y. Zhao, Y. Liu, R.V. Stan, L. Fan, Y. Gu, N. Dalton, P.H. Chu, K. Peterson, J. Ross Jr. and K.R. Chien, Defects in caveolin-1 cause dilated cardiomyopathy and pulmonary hypertension in knockout mice, PNAS 99 (2002), 11375–11380.