Biomarkers and Mechanisms of Drug-Induced Vascular Injury in Non-Rodents

Abstract

In preclinical safety studies, drug-induced vascular injury can negatively impact candidate-drug selection because there are no obvious diagnostic markers for monitoring this pathology preclinically or clinically. Furthermore, our current understanding of the pathogenesis of this lesion is limited. While vasodilatation and increased shear stress appear to play a role, the exact mechanism(s) of injury to the primary target cells, smooth muscle (SMC) and endothelial cell (EC), are unknown. Evaluation of potential novel markers for clinical monitoring with a mechanistic underpinning would add value in risk assessment and risk management. This mini review focuses on the efforts and progress to identify diagnostic markers as well as understanding the mechanism of action in nonrodent drug-induced vascular injury.

Introduction

Drug-induced vascular injury (primarily arterial lesions) in animals has been a topic of intense discussion and debate in toxicology, clinical and regulatory circles. This has resulted in regulatory pressure on the pharmaceutical industry to provide data to confirm that a candidate drug is reasonably safe for administration to humans, when this toxicity is observed and the therapeutic index and/or safety margin is either low or negative, even though several approved drugs are known to cause vascular injury in rodent and nonrodent preclinical safety studies without any reported incidence in humans (Sabota, 1989). Development of potentially novel life saving therapies has therefore been hindered due to the lack of drug-induced vascular injury biomarkers to confirm the candidate drug safety for administration to humans. The pharmaceutical industry must therefore strive to develop noninvasive biomarkers of drug-induced vascular injury in animals and humans. Ultimately, this will improve risk assessment and risk management that could lead to a more effective and efficient development of candidate drugs.

There is an intellectual gap in our understanding of the pathogenesis and/or mechanism of action leading to the drug-induced vascular lesion and so the search for biomarkers of vascular injury must be balanced with rigorous scientific investigations aimed at elucidating this mechanism of action. In the dog, pharmacologically and structurally diverse compounds can cause decreases in mean arterial pressure (MAP), reflex tachycardia and vascular toxicity, particularly in the coronary arterial bed (Dogterom et al., 1992). However, recent data suggest a lack of concordance between drug induced vascular injury, profound increased heart rate (HR) and decreased MAP (Table 1). Interestingly, Minoxidil (an approved drug) and SB209670, an endothelin receptor antagonist (ETRA), cause a 6-fold increase in coronary blood flow and vascular lesions in dogs but the ETRA did not significantly increase HR or decrease MAP (Mesfin et al., 1989; Louden et al., 2000). These latter data suggest that localized changes in coronary blood flow dynamics, not systemic changes, may play a key role in the pathogenesis and/or mechanism of this lesion. It has been suggested that vasodilatation and inability of the pharmacologically sensitive coronary vascular bed to maintain tone are critical events involved in drug-induced vascular injury. Endothelial (EC) and smooth muscle cells (SMC) not only play an important role in mediating vascular tone but these cells are also the primary targets of drug-induced vascular injury. Therefore, studying EC and SMC damage in animal models of drug-induced vascular injury could improve our understanding of the mechanism of action as well as identify potentially novel biomarkers. Such mechanistically linked markers would add value if there is potential application in both preclinical and clinical studies.

This review focuses on the progress and future direction of investigative efforts in vascular biology aimed at identification and development of monitorable diagnostic markers with preclinical and clinical application. Additionally, the current concepts underlying the mechanism of action and biochemical pathways involved in mediating drug-induced vascular injury will be discussed.

Pathology

Vasodilators, irrespective of pharmacological class, produce similar macroscopic, microscopic and ultrastructural lesions, in the dog (Mesfin et al., 1987, 1989; Herman et al., 1989; Isaacs et al., 1989; Louden et al., 1998; Dogterom et al., 1992; Joseph, 2000; Albassam et al., 2001; Clemo et al., 2003). Macroscopically, lesions consist of petechial and ecchymotic hemorrhages predominantly on the epicardium of the right atrium and extramural coronary arteries of the atrioventricular groove. Occasionally, at high doses some compounds induce lesions in the left atrium, ventricular endocardium and papillary muscle with no involvement of the septum. Microscopically, lesions generally involve large, medium-sized and some small right and left branches of the extramural and rarely intramural coronary arteries. Lesions are common at branch points of medium to large arteries. Regardless of vessel size and/or location, drug-induced vascular lesions are characterized generally by multifocal hemorrhage, segmental medial necrosis/apoptosis with inflammation and perivascular edema secondarily (Figure 1). Breaks in the internal elastic lamina and subintimal edema are observed at sites of injury supporting the hypothesis of mechanical disruption of the vessel wall. Acute adventitial inflammation progressing to chronic inflammation and ultimately adventitial fibrosis is most often seen in longer-term studies. In general, atrial cardiomyocytes are unaffected and there is no evidence of myocardial infarction.

Ultrastructurally, drug-induced vascular lesions are characterized by medial smooth muscle damage, with disruption of the basal laminae, and the appearance of lucent or dense cytoplasmic inclusions such as, large vacuoles, lipid droplets, myelinosomes and autophagosomes. In the absence of overt smooth muscle cell necrosis, one of the earliest changes observed is degeneration of smooth muscle cells as illustrated by their irregular profiles, basal laminal disruption, plasmalemmal attachment, plaque contraction and accumulation of cytoplasmic inclusion bodies. Intramedial erythrocytes and inflammatory cells are located within smooth muscle cell basal laminae and smooth muscle cell necrotic debris together with viable or necrotic inflammatory cells, erythrocytes (many of which show erythrocytolysis) and occasionally fibrin deposits. Disruption (or separation of the elements) of the internal elastic lamina is also observed occasionally. The intima may also be substantially disrupted, as shown by the presence of leukocyte cell attachment to the luminal surface of the endothelium, trans-intimal inflammatory cell migration that shows insinuation between the endothelium and the subjacent internal elastic lamina and inter-endothelial cell gaps.

Physiological Biomarkers of Drug-Induced Vascular Injury

Mean Arterial Pressure (MAP) and Heart Rate (HR)

In the dog, vascular toxicity is often times 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 (Dogterom et al., 1992). For example, Minoxidil a potent antihypertensive agent causes canine coronary arterial lesions and significantly lowers MAP and profoundly increase HR (Mesfin et al., 1989). Other compounds such as Hydralazine produced this trilogy of effects in the dog (Table 1). Interestingly, these effects are observed with compounds from diverse pharmacologic classes. These data suggests that the toxicity may be mediated in part by the hemodynamic changes in the vascular bed rather than the nature of the chemicals. However, recent experiences with ETRAs, a novel class of vasoactive agents, suggest that in comparison to Minoxidil profound hemodynamic changes (MAP and HR) are not a prerequisite for development of coronary arterial lesions in the dog (Louden et al., 2000; Stephan-Gueldner and Inomata, 2000; Albassam et al., 2001). For the ETRAs, lack of concordance with HR and MAP suggests that monitoring of possible vascular hazard in humans is not possible 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 response 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 (Albassam et al., 1999).

Regional Blood Flow

Several reports provide direct and indirect evidence suggesting that drug-induced vascular injury is associated with vasodilatation and increased blood flow and these changes precede arterial damage. 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 cardiac blood flow and arterial lesions in the coronary vasculature (Mesfin et al., 1989; Humphrey and Zins, 1984). Vasodilatation, increased regional coronary blood flow (CBF) and coronary arterial lesions have been observed with other structurally and pharmacologically diverse agents such as (±)-(1S, 2R, 3S)-3-(2-carboxymethoxy-4-methoxyphenyl)-1-1,3,4-methylenedioxyphenyl)-5-(prop-1-yl-oxy (indane-2-carboxylic acid (SB209670) (Louden et al., 2000), hydralazine (Chelly et al., 1986; Mesfin et al., 1987) and 2-cyano-1-methyl-3-[4-(4-methyl-6-oxo-1, 4,5,6-tetrahydro-pyridazin-3-yl) phenyl] guanidine (SK&F94836) (Isaacs et al., 1989). Minor but sustained increases in HR and the accompanying 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 (Louden et al., 2000). It has been hypothesized that local increases in CBF is the basis for selective coronary arterial lesion 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 pharmacologic mimicry of adenosine (i.e., adenosine agonists) by compounds such as imazo-dan (CI-914), (R)-N- (2,3-dihydro-1H-inden-1-yl) adenosine (CI-947) and N- (2,2-Diphenylethyl) adenosine (DPEA) all induce coronary arterial lesions in dogs and increase coronary arterial blood flow (Bacchus et al., 1982; Macallum et al., 1991; Metz et al., 1991; Albassam et al., 1998). 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. (Humphrey and Zins, 1984; Mesfin et al., 1989; Louden et al., 2000; Jones et al., 2003).

In the rat, dopaminergic receptor activation by fenoldopam (Morgan et al., 1983) or inhibition of type III phosphodiesterase isoenzyme (SK&F 95654) causes vasodilatation and mesenteric arterial lesions (Joseph et al., 1996). These arterial lesions were prevented when fenoldopam was co-administered with a dopaminergic antagonist or the potent vasoconstrictor methoxamine (Joseph et al., 1997). Similarly, the potent vasoconstrictor, arginine vasopressin prevented the mesenteric arterial lesions induced by the potent vasodilator SK&F 95654, a phosphodiesterase inhibitor (Joseph, 2000). Minoxidil, SK&F 95654 and fenoldopam are all associated with varying degrees of increased mesenteric arterial blood flow in the rat for prolonged periods (Joseph, 2000). 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.

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.

Biochemical Markers of Drug-Induced Vascular Injury

The ultimate goal is to identify a diagnostic marker of drug-induced vascular injury for utility in preclinical and clinical studies. The ideal marker should be (1) specific and sensitive, (2) mechanistically linked to pathology (3) altered very early, prior to morphological evidence of cell injury (4) return to baseline values when there is no further cell and/or tissue damage. Research activities in drug-induced vascular injury has focused on advancing our knowledge in different areas: (1) identification of specific and sensitive diagnostic markers of vascular injury, (2) determine the value and significance of markers of EC and SMC cell injury, and (3) determine the mechanism of action.

Biomarkers of Endothelial Cell Injury

Perturbations and/or injury to EC stimulate the release of several molecules and/or proteins including von Willebrand Factor (vWF), vWF pro-peptide (vWFpp), vascular endothelial growth factor (VEGF), endothelin (ET), caveolin-1 (cav-1), asymmetric dimethyl arginine (ADMA), and nitric oxide (NO). It has also been suggested that as a consequence of arterial damage EC are released from the site of injury and can be measured as a biomarker of vascular wall injury (Scicchitano et al., 2003; McFarland et al., 2004). Of the proteins, vWF is of most interest because it has been a clinical marker in humans for years and it can also be evaluated in various preclinical species (Vischer et al., 1997; Brott et al., 2005a).

vWF and vWFpp

In the dog, vWF is of particular interest as a biomarker of vascular injury because EC are the sole source of circulating plasma levels. Dogs treated with a potassium channel opener (PCO) caused plasma vWF to increase as early as 3 hours postdosing (Brott et al., 2005b). This observation in conjunction with other reports raise the possibility that the transient 2–6-hour increase in circulating plasma vWF could be a reporter of endothelial cell activation/perturbation prior to morphologic evidence of vascular damage (Newsholme et al., 2000). However, plasma vWF values returned to baseline or lower 24 hours postdosing when vascular damage was confirmed histologically (Brott et al., 2005b). These data clearly indicate that there is a temporal disconnect, between increases in plasma vWF, the time when morphologic evidence of vascular injury is present and progression of the lesion characterized by inflammation secondarily. Therefore, vWF is not a suitable biomarker for use in preclinical safety studies to monitor progressive vascular damage.

The transient increase in plasma vWF may suggest that it is an attractive predictive biomarker of impending vascular injury. This must be viewed with caution because use of vWF as a marker of vascular compromise in humans has several limitations: (1) in normal subjects, plasma vWF display a wide range of values, (2) fluctuations in plasma are small and overlap normal values in subjects with suspected vascular injury, and (3) vWF is released from the basolateral side of the EC and in injured vessels it may be trapped at the site of release and this could cause an underestimation of actual amounts released into circulation, (4) plasma vWF can also be affected by thrombogenic and inflammatory processes because human platelets are a rich source of vWF, (5) changes in clearance rate and blood groups can also affect plasma vWF, and (6) the long half-life and population variability in plasma levels complicates evaluation because plasma levels may only change modestly even when there is extensive EC damage (Vischer et al., 1997). In summary, it is safe to conclude that based on the available data, evaluating vWF by itself is an unreliable predictor or reporter of ongoing and/or progressive vascular injury in humans, dogs or rats.

Biologically, pro-vWF is cleaved to produce mature vWF and vWF pro-peptide (vWFpp). These two proteins with different biological activity are released in equimolar concentrations in plasma and an increased level of both proteins predicts and reports endothelial cell activation/injury. However, vWFpp has several advantages over the mature peptide as a specific and sensitive marker of vascular injury: (1) EC is the sole source of this protein unlike vWF that is released from platelets in humans and rats but not dogs, (2) mild vascular injury causes a significant elevation, since concentration in normal plasma is very low, (3) vWFpp has a short half-life in dogs and humans and so persistent elevation in plasma is suggestive of progressive vascular injury, and (4) it does not appear to bind platelets or get trapped in the damaged vessel wall. In this regard, measurement of plasma vWFpp could be a useful, sensitive and specific biomarker of drug-induced vascular injury (Borchiellini et al., 1996).

This hypothesis was tested in a canine model of sustained activation and/or perturbation of the vascular endothelium using endotoxin (Figure 2). A single, low-dose of endotoxin caused a profound increase in plasma vWFpp as early as 3 hours postdose and this increase was sustained for greater than 24 hours. Desmopressin (1-deamino-8-darginine [DDAVP]) also increases plasma vWF pro-peptide), however, the pattern is markedly different from endotoxin (van Mourik et al., 1999). Endotoxin causes sustained increases over 24 hours versus DDAVP that causes a transient and short-lived increase returning to near baseline values within a few hours. The cumulative data suggest that regardless of the experimental protocols, physiological (DDAVP) or pathological (Endotoxin) perturbation of the EC cells in vivo causes measurable increases in plasma vWFpp. Localized (drug-induced) coronary injury is expected to cause a smaller increase in vWFpp when compared to systemic activation that occurs with endotoxin or DDAVP. There is strong evidence that supports vWFpp as a marker of EC activation/perturbation in other species. For example, in humans physiological stimulation of EC causes a transient short-lived increase, while pathological stimulation caused a sustained and prolonged increase for almost 24 hours. A similar response was observed in a baboon model of disseminated intravascular coagulation (DIC) (Vischer et al., 1997). Recent reports suggest that measurement and analyzing the vWF:vWFpp ratio in humans allows discrimination between chronic and acute phases of EC perturbation and/or activation (van Mourik et al., 1999).

On the basis of these findings in dogs, humans and baboons cleavage products of pro-vWF (vWF, vWFpp) should be included as a hematology parameter to assess EC perturbation, activation and/or injury when vascular toxicity is suspected. Furthermore, analysis of vWF:vWFpp ratio may discriminate and/or differentiate acute from chronic and progressive vascular injury.

Biomarkers of the Nitric Oxide Pathway

Caveolin

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 is obviously an intellectual gap 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 as a result of excessive vasodilatation. We hypothesize, that measurable, soluble proteins as potential diagnostic markers of drug-induced vascular injury are released from endothelial and/or smooth muscle cells or the vascular 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 caveolin-1.

Caveolae are lipid rich plasma membrane invaginations of the cell that play an important role in transcytosis, molecular transport and signal transduction (Rothberg et al., 1992; Anderson et al., 1998; Shaul and Anderson, 1998; Bucci et al., 2000; Liu et al., 2002; McIntosh et al., 2002). Caveolae appear to be the focal point for compartmentalizing, organizing and modulating signal transduction activities for many receptors and enzymes that are co-localized to caveolae. Caveolin-1 (Cav-1) is not only the major structural protein of caveolae (Rothberg et al., 1992), but it also modulates the function of signal transduction particularly in endothelial (EC) and vascular smooth muscle cells (SMC), the primary targets of drug-induced vascular injury. In addition, adenylyl cyclase (AC) and other components of the cAMP pathway (Garcia-Cardena et al., 1997; Ishizaka et al., 1998; Rizzo et al., 1998; Labrecque et al., 2003; Yamaguchi et al., 2003) and nitric oxide synthase (NOS) are co-localized with cav-1. It is known that cAMP down-regulates cav-1 expression (Park et al., 2000) and cav-1 is a negative regulator of NOS and hence indirectly regulates nitric oxide (NO) production (Feron and Kelly, 2001; Goligorsky et al., 2002). Therefore 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 (Fattman et al., 2003). 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 myogenic tone (Drab et al., 2001). Cav-1 down-regulation was also associated with a 5-fold increase in systemic NO, but no major changes in protein production of NOS (Zhao et al., 2002). Suggesting that deregulation and/or increased activity of NOS were responsible for the massive increase in NO production. Thus, the consequence of cav-1 down-regulation leads to activation of AC, increased cAMP levels (which is observed with the structurally and pharmacologically diverse compounds that cause vascular lesions) increased NOS activity which could lead to the generation of undesirable, potentially cytotoxic, superphysiological/pathological bursts of NO that are detrimental to smooth muscle and/or endothelial cell health. Thus, sustained and prolonged NOS activation induces pathological NO levels causing profound localized vasodilatation and increased blood flow. Hence, drug-induced vascular injury could be caused by a synergistic effect of NO-induced vasodilatation and free radical induced cell injury. These data suggests that cav-1 plays an important role in regulating NOS activity and potentially drug-induced vascular injury.

Dissecting the relationship between loss of cav-1, and increases in NOS activity, NO, vasodilatation and vascular injury will yield mechanistically linked biomarkers that will substantially improve our evaluation and capability to assess and manage potential human risk. Cav-1 qualifies as a potential mechanistically linked diagnostic biomarker of drug-induced vascular injury and warrants further evaluation.

VEGF and ADMA

The endothelium is well recognized as the primary player in maintenance and regulation of vascular structure and tone. This is accomplished primarily through the proximity, physical and biochemical intracellular signaling between EC and SMC. Therefore, biomarkers of vascular toxicity must include protein and/or molecules that mediate EC and SMC function. Vascular endothelial growth factor (VEGF) and asymmetric dimethylarginine (ADMA) are secretory products that affect EC and SMC function and therefore could have utility as diagnostic markers of vascular injury. VEGF is secreted primarily from EC and proteins from this family interact with a set of cell-surface receptors that trigger autocrine and/or paracrine responses within the vessel wall and endothelium (Hariawala et al., 1996; Bouloumie et al., 1999).

For example, VEGF can cause dilation of isolated dog coronary arteries (Ku et al., 1993) and when administered exogenously, causes severe hypotension in conscious pigs (Hariawala et al., 1996) and rats (Yang et al., 1996). The biochemical pharmacology implicates VEGF activation of tyrosine kinase receptors that causes vasodilatation and consequently hypotension (Hennequin et al., 1999). The current thinking is, vasodilatation is due to VEGF induced release of NO, the potent endothelium-derived relaxant factor (Horowitz and Hariawala, 1995; Bussolati et al., 2001) or opening of potassium channels similarly to the pharmacologic action of PCO (Cuevas et al., 1991).

Circulating physiological levels of VEGF when neutralized by antibodies is associated with hypertension and clinically pre-eclampsia hypertensive patients have low serum levels of VEGF (Roberts et al., 1989; Roberts and Cooper, 2001; Pridjian and Puschett, 2002). These data collectively indicate that inhibition of VEGF and/or its tyrosine kinase receptors could lead to vasoconstriction and vascular pathology. If VEGF associated tyrosine kinase inhibition is associated with drug-induced vascular injury, unique biochemical markers of inhibited NO induced hypertension could add value. One such molecule is ADMA, a naturally occurring amino acid that circulates in plasma, is excreted in urine and is found in several tissues and cells (Boger, 2003).

ADMA, a member of a family of closely related proteins, is generated by degradation of methylated proteins and subsequent physiological protein turnover. It is cleared by excretion through body fluids but also enzymatically by dimethylaminohydrolase (DDAH) that is regulated in part by NO through a feedback mechanism (Boger, 2003). ADMA inhibits eNOS by competitive displacement of L-arginine, the physiological substrate, from the enzyme (Boger and Ron, 2005). It (ADMA) generated considerable more interest when shown to inhibit all three isoforms of nitric oxide synthase and ultimately NO production (Vallance and Leiper, 2004). Support for this idea is strengthened by clinical data that shows marked elevation of plasma ADMA levels in patients with diseases such as systemic and pulmonary hypertension, chronic renal failure and congestive heart failure, all characterized by dysregulation of NOS and decreased NO production (Gorenflo et al., 2001; Zoccali et al., 2001; Kielstein et al., 2002; Saitoh et al., 2003). Therefore, compounds that inhibit NOS, VEGF and/or VEGF tyrosine kinase receptors are likely to cause drug-induced vascular lesions mechanistically related to hypertension of which ADMA would be a good biomarker.

It is clear from these studies, that inhibition of NOS and/or decreased production of NO or down regulation of Cav-1 leading to sustained NOS activity and increased NO production will lead to vascular lesions, albeit through different mechanisms. In preclinical toxicology studies, compounds that inhibit VEGF associated tyrosine kinase receptors and NOS inhibitors are associated with hypertension and vascular lesions with a pathological pattern that is morphologically distinct from vascular lesions induced by vasodilators. Therefore, appropriate characterization of the vascular pathology is critical as it can provide a rationale for the mechanism and possibly the biomarker of choice.

Smooth Muscle Actin as a Biomarker of SMC Injury

A SMC specific protein, smooth muscle α actin (SMA), could potentially be released in circulation as a result of arterial damage. In our studies, evaluation of SMA immunohistochemically shows decreased immunoreactivity in injured SMC of damaged arteries (Brott et al., 2005a). In drug-induced vascular injury, 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. In urine, SMA has been evaluated to determine its value as a marker of chronic renal damage (Haas et al., 1999). Therefore, evaluation of urine, serum or plasma for this analyte in an animal model of drug-induced vascular injury should be considered. This could be 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 cell damage/death has occurred and the kinetics between release of SMA and cell death is unknown.

C-Reactive Protein and Inflammatory Markers

In drug development, clinical progression of compounds associated with vascular injury preclinically, has always been a concern because in humans, there is a well-recognized association between inflammation, chronic vascular injury (atherosclerosis) and cardiovascular mortality. In humans, c-reactive protein (CRP), an acute phase protein and a marker of inflammation has been heralded as a diagnostic marker of vascular disease (Mosca, 2002; Ridker, 2003). However, CRP has not been useful as a predictor or a reporter of drug-induced vascular injury in rats, dog and/or monkeys. Other markers such as interleukin-6, and other cytokines have been suggested (McDuffie et al., 2004; Zhang et al., 2004) but these require further study. Adverse perturbation of the EC and activation 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. In this case, circulating soluble cytokines released at the site of inflammation may be measurable in serum and/or plasma and hence constitute a diagnostic marker. However, the cellular interactions between adhesion molecules, the vascular wall and immune cells make it difficult to determine a unique marker that is diagnostic for this complex process (Schwartz et al., 1989). The vasculature can become inflamed as a secondary event, 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 [Kerns W.D., personal communication]. Additionally, neutrophilia and elevated CRP levels are useful diagnostic features of beagle pain syndrome characterized by an inflammatory perivascular lesion that often complicates interpretation of drug-induced vascular lesions in the dog. Monitoring of neutrophil counts and serum cytokines may serve as useful markers for this process that is morphologically and pathophysiologically distinct from drug-induced lesions. Since drug-induced vascular injury is not primarily an inflammatory lesion, specificity and sensitivity must be resolved before adhesion molecules, inflammatory markers and cytokines can be qualified and then validated as vascular injury markers.

Conclusion

Direct or in-direct chemical interaction with receptors, enzymes or ion channels within the vascular wall can result in profound hypertension, hypotension (systemic or local vasodilatation), or increased localized blood flow all of which can cause damage to SMC and/or EC resulting in vascular damage. In drug-induced vascular injury, the initiating site/cell of injury, EC or SMC layer is unclear. However, the evidence to date indicates that, hypertension induces vascular lesions with a pathological pattern that is morphologically distinct from vascular lesions induced by vasodilators. Therefore, histological characterization of the lesion, physiological measurements and biochemical endpoints of SMC and EC injury should be evaluated in order to identify biomarkers for monitoring drug-induced vascular injury. Since vascular injury involves multiple mediators and cell types, evaluation of a panel rather than a single biomarker may be more useful for assessing early, and severe progressive vascular injury.