Drug-Induced Vascular Injury—A Search for Biomarkers



Discussion Document for the NonClinical Studies Subcommittee

of the Advisory Committee on Pharmaceutical Science


The Expert Working Group on Drug-Induced Vascular Injury September 9, 2002




Expert Working Group Members:

William Kerns-Pharma Consulting Inc. (Co-Chair)

Lester Schwartz-GlaxoSmithKline (Co-Chair)

Kerry Blanchard-Boehringer-Ingleheim

Scott Burchiel-University of New Mexico

Robert Johnson-Schering-Plough

Fred Miller-NIEHS

Prakash Nagarkatti-Virginia Commonwealth University

Don Robertson-Pfizer

Paul Snyder-Purdue University

Active Contributors:

Eric Fung-Ciphergen

Mike Lawton-Pfizer

Calvert Louden- Astra-Zeneca

Heath Thomas-GlaxoSmithKline

Anthony Ward-BD Biosciences

FDA Liaisons:

James MacGregor-NCTR

Frank Sistare-CDER

David Essayan-CBER

Thomas Papoian-CDER



Drug-Induced Vascular Injury — A Search for Biomarkers



Introduction. Arterial and venular injury is a relatively common hazard identified during nonclinical toxicity testing within certain pharmacological classes of drug candidates. Drugs that induce vascular lesions in animals present a safety assessment dilemma to toxicologists, physicians and regulators wishing to assess the safety of new medicines for humans. This dilemma is confounded by the gaps in our knowledge regarding pathogenesis of drug-induced vascular injury in animals and importantly, the absence of validated nonclinical or clinical methods for monitoring vascular integrity noninvasively.

Contrary to past thinking, nonclinical experience with new and novel pharmaceuticals suggests that vascular injury is not always associated with systemic changes in blood pressure and heart rate, rendering these parameters of little value, clinically. Variation in species responsiveness to vasoactive and non-vasoactive agents and marked differences in reactivity of selected vascular beds, taken together with contributions from numerous reactive cell types (e.g., endothelium, vascular smooth muscle, and inflammatory/immune cells) all add complexity to defining mechanism(s) and defining robust biomarkers.

Experimental evidence suggests that drug-induced vascular injury in animals can occur via one or more of the following overlapping mechanisms:

Historically, toxicology studies in normal animals have not predicted the occurrence of the drug-induced hypersensitivity vasculidities, which are by far the most common drug-induced adverse vascular responses observed in humans. In general, drug-induced hypersensitivity angitis of humans is not recognized in toxicology studies in healthy animals. Conversely, drug-induced vascular lesions recognized in animals are not known to occur in humans, however this could be related to our lack of knowledge and methods for monitoring in humans. Until these methods become available, it will be difficult to make claims of species specificity. Confusion in histological terminology and lack of consistency in characterizing the key components of vascular damage have resulted in communication gaps between pathologists, toxicologists, regulators and clinicians, all who share responsibility for ultimately making a judgment on human risk. Poor understanding of pathogenic mechanisms, limitations of animal models and lack of valid biomarkers for nonclinical and clinical monitoring has elevated this issue to a shared level of concern between FDA and the pharmaceutical industry.



EWG Mandate. To evaluate the current state of knowledge of drug-induced vascular injury in animals and humans and to identify opportunities for the discovery and validation of predictive biomarkers of acute vascular injury that could be used in nonclinical drug safety programs and eventually in Phase I/II clinical programs.


White Paper Objectives. The purposes of this white paper are:

    1. review the current status of the field
    2. present current understanding on mechanisms of drug-induced vascular injury
    3. identify opportunities for practical biomarkers and high potential research areas
    4. annotate emerging technologies holding promise in the search for new markers
    5. identify opportunities that may be met through collaborative work between public and private partners
    6. confirm that this is an issue that needs resolution






















    1. Introduction

Limited understanding of the mechanisms by which drugs cause vascular injury has been a significant barrier to progress in the development of reliable biomarkers of vascular injury. This facet is further confounded by studies indicating multiple and distinct mechanisms operating in these syndromes. This is not entirely surprising because a wide variety of drugs, with varied pharmacologic effects that have been implicated, could potentially act on different cells but trigger a common down-stream event leading to vascular injury. A review of the literature indicates that at least three major mechanistic pathways may contribute to drug-induced endothelial and smooth muscle compromise, resulting in vascular injury in animals. These mechanisms can be broadly classified into the following categories:

Such mechanisms are not mutually exclusive and in fact there could be significant overlaps between these routes of injury.

      1. Biomechanical Injury
      2. The relevance and uncertain extrapolation of drug-induced arterial toxicity in rats, dogs and other species to humans continues to be an obstacle to drug-discovery and development, as well as a concern to regulatory scientists. The issue of arterial toxicity particularly in the dog is further complicated by the occurrence of spontaneous coronary arterial lesions that are reported to mimic drug-induced vasculopathy and the inability of modern science and technology to identify clear and unequivocal markers, suitable for assessment and monitoring of potential drug-related vascular injury in animals and humans. Historically, in the dog, vascular toxicity has been associated with profound cardiovascular hemodynamic changes in mean arterial pressure (MAP) and heart rate (HR), which have been used as surrogate markers of potential vascular toxicity in man at therapeutic doses . However, recent experiences with a novel class of vasoactive agents, endothelin receptor antagonists (ETRAs) and others, suggest that profound hemodynamic changes (MAP and HR) are not a pre-requisite for development of coronary arterial lesions in the dog . For this class of molecules, hemodynamic monitoring as a method to assess the potential hazard to humans is not possible and extrapolation of the potential risk to the human population can only be made on the basis of canine responses and the therapeutic index in the dog . This species-sensitive response in the dog has yet to be supported by scientific data that describes a unique mechanism of vascular toxicity that is not applicable to man at therapeutic concentrations.

        A number of structurally diverse pharmacological agents can induce arterial lesions in rats and dogs, and the mesenteric and coronary arteries, respectively, are primarily affected. The reasons for this tissue-specific selectivity are not clear but may be related to increased sensitivity in these respective vascular beds because spontaneous arterial lesions develop at these sites in both species . The beagle dog used routinely in toxicology studies is considered to be an extremely sensitive species to drug-induced coronary arterial lesions. For example, adenosine agonists and an ETRA will induce coronary arterial lesions in dogs and monkeys ). However, in the dog lesions occur after a very short duration of treatment at lower systemic exposure compared to the monkey. The mechanisms of this selective species response are ill-defined but may be related to localized changes in blood flow as demonstrated with adenosine , an endogenous substance released from the heart, adenosine agonists, and a number of structurally diverse vasoactive pharmacological agents . In the rat, there is indirect evidence linking increases in localized mesenteric blood flow and development of drug-induced mesenteric arterial lesions . Therefore, exploration of the contributions of localized changes in blood flow (BF) and subsequently changes in shear and hoop stress, to drug-induced arterial injury is worthy of investigation in both animals and humans.

        Flow-mediated mesenteric arterial lesions in the rat

        Within the mesenteric arteries, physiological activation of specific dopaminergic receptor subtypes (DA1) by fenoldopam leads to marked vasodilation and mesenteric arterial lesions . In order to test the vasodilation hypothesis, fenoldopam was administered with DA1 receptor antagonist and this treatment regimen prevented induction of hemorrhagic lesions in rats . Other studies have also reported that co-administration of the vasoconstrictor methoxamine attenuated the fenoldapam induced splanchnic arterial lesions. The vasodilation hypothesis was also tested using SK&F 95654, a potent Phosphodiesterase Inhibitor (PDEIII) which causes marked reduction in mean arterial pressure (~40 %) and induction of mesenteric arterial lesions. Co-administration of arginine vasopressin, a vasoconstrictor, reversed the PDE hypotensive effect in a dose-dependent manner and prevented induction of mesenteric arterial lesions .

        When administered at doses that induce mesenteric arterial lesions, SK&F 95654 a potent PDEIII inhibitor and Minoxidil a potent K+ channel opener produced long-lasting increases in mesenteric blood flow for 5-7 hours post dosing. These data indicate that prolonged vasodilation by certain agents will contribute to vascular injury.

        Flow-mediated coronary arterial lesion in dogs.

        Several vasoactive agents that lower blood pressure and increase heart rate induce regional, right coronary arterial damage . The basis for this selective coronary arterial lesion in dogs may be related to localized increases in coronary blood flow because under normal physiological conditions, localized control and regulation of coronary blood flow is mediated 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 CI 914, CI 947 and N- (2,2-Diphenylethyl) adenosine (DPEA) all induce coronary arterial lesions in dogs and increase coronary arterial blood flow . Therefore, it is now well accepted that administration of adenosine (A1) agonists as a pharmacological class are associated with coronary arterial lesions in dogs. A "class" effect for coronary arterial lesions has also been ascribed to K+ channel openers and ETRAs because these agents cause profound increases in regional blood flow.

        Several reports from studies in dogs indicate that extreme increases in regional CBF flow precede coronary arterial damage. For example minoxidil, a long lasting vasodilator, when given to dogs at cardiotoxic doses induced a 6-10 fold increase in regional cardiac blood flow and this sustained increase in flow resulted in damage to the coronary vasculature . Other structurally and diverse pharmacologic agents such as SB 209670 , hydralazine , SK&F 94836 and milrinone which are vasoactive have been reported to increase CBF and induce coronary arterial lesions in dogs.

        Other studies in dogs have shown medial hemorrhage and necrosis of extramural coronary arteries of the right atria that were associated with minor but sustained increases in heart rate (10-30 beats/min) and slight decreases in mean arterial pressure (10-15 mm Hg) after infusion of an endothelin receptor antagonist for 5 days . These changes were further shown to be associated with six-fold increases in regional blood flows and a disproportionate distribution of ETB receptors in the right coronary arteries of dogs .

        Blood flow and Biomechanical Injury

        It has been postulated that mesenteric and coronary arterial lesions develop because of marked vasodilation, increased blood flow, increased shear (endothelial cells) and hoop stress (vessel wall). Hoop stress or tension is the product of transluminal pressure and the vessel radius divided by wall thickness. Increased shear stress leads to inter-endothelial breaks gradual breakdown of vessel wall integrity, breaks in the internal elastic lamina and hemorrhage . It has also been speculated that vasodilation changes normal laminar flow to turbulent flow that imitates shear stress and this leads to a cascade of effects in the vessel wall culminating in hemorrhage and necrosis. It is not clear to what extent similar changes occur in humans given pharmacologically-active doses of certain vasodilators. It is also not clear that such changes in vessel wall tension in the absence of concurrent receptor mediated cellular alterations would result in the observed vessel injury.

      3. Toxicity following direct pharmacological and/or chemical perturbation.
      4. The target cell of drug-induced vascular toxicity is presumed to be endothelial and/or smooth muscle cell. Development of vascular lesions has been linked to the pharmacologic effect of the drug because the pharmacologic target is often located on these cells. For example ETRAs as a class cause arterial lesions in dogs because the drug target, endothelin receptors, are located on endothelial and smooth muscle cells . ET receptors on EC and SMC regulate vascular tone through opposing vasoconstrictor and vasodilator effects . Therefore, an upset of this delicate balance resulting in loss of vasoconstriction or vasodilation might lead to vascular damage. For example, it was shown that the disproportionate distribution of vasoconstrictive receptors which when antagonized within the right atrium and right coronary arteries predisposes these sites to pharmacological vasodilation resulting in six-fold increases in regional blood flow, but minimal changes in heart rate and blood pressure . A similar observation has been reported in the rat for DA1 receptors that are the target for fenoldopam . Other pharmacological targets associated with vascular injury that are located on EC and SMC cell include adenosine receptors, K+ channels, and the cGMP-inhibitable isoenzyme of phosphodiesterase . Vascular toxicity then may result from pharmacological interaction with the molecular target that initiates a series of interactive cascades among cellular and non-cellular components leading to eventual damage. Contributing to toxicity may be the sustained alterations by these agents of EC components normally responsible for targeting interactions with the immune system.

        Allylamine and β-Amino-proprionitrile (β–APN) are chemical substances that are direct acting vascular toxicants . Allylamine causes medial hypertrophy and sub-intimal proliferation and β–APN causes aneurysms of the aorta, coronary and mesenteric vasculature in man . Based on the different morphologic presentations of vascular injury induced by these agents, it is reasonable to assume that the pathophysiology is quite different from drug-induced lesions that tend to present a similar morphologic picture.

      5. Immune-mediated Injury

Although the precise mechanisms by which immune cells mediate injury to endothelial cells (EC) and smooth muscle cells is not clear, there is evidence that drugs and biologics can activate the immune system which in turn can cause EC injury. Moreover, drugs may modulate the EC directly in such a way as to trigger an inflammatory response and an immune attack directed against EC and smooth muscle. While the immune effectors mediating acute drug-induced vascular injury have not been extensively identified, several types of vasculitis as seen in chronic autoimmune diseases and a number of infections have been well characterized. Thus, the latter may serve as a model to address the possible mechanisms involved in acute and chronic phases of drug-induced vascular injury. Such disease-related vascular injuries have also been discussed below to provide additional clues on how drugs may induce immune-mediated vascular injury. The immune-mediated vasculitis can involve multiple effector mechanisms as described below:

Mediators of immediate hypersensitivity

Neutrophils. Neutrophils have been widely associated with vascular injury. Chemotactic agents including complement components, factors from fibrinolysis, activate them and kinin systems and chemokines produced by endothelial cells and by other leukocytes all activate neutrophils. Chemotactic stimuli facilitate their binding to the endothelial cells. Neutrophils carry a variety of granules containing hydrolases, myeloperoxidase, proteinases, and muramidase. In certain forms of vascular injury, antibodies against myeloperoxidase have been reported. Immune complexes can bind Fc receptors on neutrophils and activate them to release enzymes which could cause endothelial cell injury. Cytokines at the inflammatory site may contribute significantly to the generation of reactive oxygen species (ROS), particularly by PMN. The vascular endothelium is a major target of oxidant stress, playing a critical role in the pathophysiology of several vascular diseases and disorders. Specifically, oxidant stress increases vascular endothelial permeability and promotes leukocyte adhesion, which trigger alterations in endothelial signal transduction and redox-regulated transcription factors . Histopathological analysis of drug-induced vascular injury in the skin has shown two distinct patterns . The first type consists mainly of neutrophils and the second, predominantly mononuclear cells.

Mast Cells. Mast cells have been well characterized for their role in induction of allergies. Although their precise role in endothelial cell injury is not clear, several recent studies suggest that they may play a key role in endothelial cell functions and injury. For example, mouse or human mast cells can produce and secrete vascular permeability factor/vascular endothelial cell growth factor (VPF/VEGF) which can potently enhance vascular permeability and induce proliferation of vascular endothelial cells . Mast cells play a critical role in regulating the expression of EC adhesion molecules, ICAM-1 and VCAM-1, and thereby enhance leukocyte-mediated inflammatory response . In mercuric chloride-induced vasculitis in rats, the role of mast cells was implicated . In SK&F 95654 induced mesenteric vasculitis in rats, enhanced local degranulation of mast cells was noted .

Endothelial Cells

Endothelial cell activation has been implicated as an important event in the induction of vasculitis and vascular leak syndrome. Endothelial cell activation is considered as an immunological activation . Activated endothelial cells can function as antigen-presenting cells and participate in T cell mediated immune reactions . Activated endothelial cells up-regulate expression of major histocompatibility complex (MHC)-encoded molecules . MHC class I antigens on the activated endothelial cells interact with CD8 T lymphocytes, while endothelial MHC class II antigens interact with CD4 T lymphocytes in cell-mediated immune responses . Furthermore, activated endothelial cells express a number of immunologically relevant surface molecules, such as adhesion molecules of the immunoglobulin gene superfamily. In SK&F 95654 induced mesenteric vasculitis in rats, evidence for early endothelial cell activation was reported . Increased numbers of circulating endothelial cells have been detected during vascular injury such as during myocardial infarction and endotoxinemia . Thus, it is possible that circulating endothelial cells may serve as a biomarker for vascular injury.

Adhesion Molecules

Several endothelial cell adhesion molecules belonging to the Ig superfamily play a critical role in the interaction between EC and leukocytes. These include ICAM-1, ICAM-2, VCAM-1, and MAdCAM. These molecules are either expressed or inducible in EC. In addition, integrins present on leukocytes are involved in adhesion to EC. Also, E-selectin expressed on EC binds to the carbohydrate ligands on leukocytes. Adhesion molecules play a critical role in the recruitment of leukocytes in many forms of vascular injury. Increased expression of ICAM-1 has been reported in vasculitic lesions involving nerve and muscle as well as in vessels of SLE patients . Also, in patients with diffuse vasculitis, the skin vessels increased expression of E-selectin and ICAM-1 . The levels of soluble ICAM-1 have been shown to be higher in individuals who develop atherosclerosis there by suggesting that sICAM-1 may serve as a biomarker for such vascular lesions. ICAM-1 immunostaining on endothelial cells has been shown to be enhanced in the mesentery of rats exposed to SK&F 95654 .


Several drugs can cause vascular injury associated with anti-neutrophil cytoplasmic autoantibodies (ANCA) . These patients produce Abs primarily against myeloperoxidase or proteinase 3 . Although it is not clear that these antibodies are directly pathogenic, the levels of ANCA may serve as a marker for disease activity and response to drug withdrawal . There is strong evidence to suggest that hydralazine and propyl thiouracil can cause ANCA-positive drug-induced vascular injury . In addition to ANCA, anti-EC Abs (AECA) has also been detected in a variety of vascular diseases. AECA may upregulate expression of adhesion molecules such as E- selectin, ICAM-1 as well as cytokines and chemokines by EC, which in turn may recruit leukocytes .

T Cells

Rats treated with mercuric chloride develop necrotizing vasculitis which occurs in two phases: an early and a late phase. In the latter phase, T cells seem to play an important role inasmuch as treatment with monoclonal antiboidies to inhibit T cell function completely prevents the development of late vasculitis . Furthermore, infiltration of T cells into vascular lesions and elevated soluble IL-2 receptors in the serum have been reported in Kawasaki disease . In Takayasu arteritis, which exhibits vasculitis that involves the aorta and pulmonary arteries, the infiltrating cells consist of cytotoxic T cells and NK cells. They injure the vascular cells by producing perforin . Vascular injury caused by cytolytic T cells and NK cells has been well documented with the use of IL-2 and other cytokine therapies including GM-CSF, IL-1 and IL-4 . IL-2 was shown to activate T cells to express high levels of CD44, which interacts with EC and induces EC injury by producing Fas ligand and perforin . Activated cytolytic T cells and NK cells express Fas ligand which in turn can induce significant levels of apoptosis in EC . In giant cell arteritis, the vascular injury appears to depend on T cells that produce IFN-which activates macrophages . In a patient with acute drug induced cutaneous vasculitis, a 100-fold increase in plasma concentrations of IFN- was noted thereby suggesting that IFN- may up-regulate adhesion molecule expression on EC and activation of T cells .


Activated macrophages are known to participate in the process of instability and rupture of atherosclerotic plaque . In giant cell arteritis, the mechanisms of injury have been mostly attributed to the effector macrophages. Macrophages specialize in oxidative damage with lipid peroxidation attacking smooth muscle cells and matrix components. These macrophages also produce reactive oxygen intermediates which in combination with nitrogen intermediates, cause protein nitration of endothelial cells . In Kawasaki disease, activation of cytokines produced by macrophages elicit proinflammatory and prothrombotic responses in endothelial cells . In this disease, macrophage colony stimulating factor has been shown to play a critical role in the pathogenesis and can be used as an indicator for the risks of valvulitis and coronary arteritis .

    1. Research Opportunities
    1. Identification and development of mechanistically based novel biomarkers of drug-induced vascular injury that targets both the vascular cells and immune cells.
    2. Direct measurement of early and sustained alterations in regional BF using flow in affected areas of vascular injury probes, radioactive or color-coated fluorescent microspheres
    3. Measurement of the coronary arterial diameter by echocardiography .
    4. Evaluation of endothelial nitric oxide synthetase (eNOS) because it is elevated by increased shear stress. It has been suggested that increased shear elevates eNOS which is co localized with caveolae and this coupling transducer signals into EC as well as SMC cells in the vessel wall.
    5. Develop methodologies (in vitro followed by in vivo studies) for direct measurement of shear stress or biochemical assays to determine eNOS or caveolin-1 concentrations that may be elevated because of increased shear.
    6. Electron microscopy to evaluate the status of EC and SMC caveolae because disruption of this cellular organelle is associated with vascular pathology in caveolin-1 gene disruption in mice .
    7. Immunohistochemistry, Western Blotting, Laser capture dissection and microarray technology to evaluate gene and protein changes in target cells i.e., EC and SMC.
    8. For agents that are potentially direct vascular toxicants employing cytotoxicity assays with cultured EC and SMC at relevant concentrations that may be achieved in vivo toxicology studies.
    9. Identification of the mechanisms by which immune cells are activated directly by the drugs, which in turn may contribute to EC injury. This involves identification of the nature of effector cells (e.g., neutrophils, lymphocytes, monocytes, macrophages, basophils, eosinophils, mast cells) adhesion molecules that are upregulated on leukocytes, which facilitate the leukocyte-EC interactions, and the effector molecules produced by leukocytes involved in EC injury.
    10. In instances where drugs modulate the ECs in a way that would secondarily trigger an immune attack, it is important to identify the nature of molecular and phenotypic changes occurring in EC that trigger inflammation and immune attack.
    11. The studies on the role of immune components in EC injury can be facilitated by the use of pharmacologically immunocompromised, genetically immunodeficient animals, transgenic knockout animal models deficient in specific immune components, and knock-in animal models where the original strain has a pathologic defect that can be corrected by supplying the normal functional gene.
    12. Laser capture to isolate EC or inflammatory cells and DNA microarray analysis may help in identifying specific pathways of immune cell and EC activation.
    13. The mechanism of immune cell-mediated EC injury such as necrosis or apoptosis will be useful in identifying better biomarkers and to develop strategies to prevent EC injury.
    14. Role of anti-neutrophil cytoplasmic autoantibodies (ANCA) in drug-induced vascular injury.
    15. Analyze genetic risk factors, which may play a role in immune-mediated EC injury.
    16. Quantitative imaging of gene induction in living animals.
    1. Gaps
    1. Correlate blood flow changes in the dog using the various methods and different pharmacological agents to clearly establish what magnitude of blood flow increase causes vascular lesions.
    2. Technology is limiting for measuring blood flow in rat. A reproducible and validated model is not available and so the increased blood flow hypothesis in the rat has not been appropriately tested.
    3. Measurement of shear stress and biomechanical effects has not been attempted in rats or dogs.
    4. No available data on the biochemical (signal transduction pathway) effects of these drugs on EC or smooth muscle cells in vivo or in vitro. This evaluation may yield potential soluble markers measurable in serum and/or plasma.
    5. Lack of hard scientific data that provides an understanding of the biomechanical forces interplay in the vascular wall during vasodilation that leads to subsequent injury.
    6. Do shear stress and/or exaggerated pharmacology and vasodilation cause apoptosis of EC and SMC in the vessel wall?
    7. Understanding and being able to separate primary vs secondary effects.
    8. Understanding the nature of leukocyte-EC cell interactions during drug-induced vascular injury.
    9. To identify biomarkers of inflammation as they relate to drug-induced vascular injury.
    1. Summary

It is clear that a more complete understanding of the mechanisms of drug-induced vascular injury will accelerate development of solutions to managing risk. Directed approaches to understand mechanisms of vascular injury will provide a link between structural damage and derangement of specific cardiovascular functions. Testing of the prevailing hypotheses as an initial approach in a decision tree paradigm will provide focus and direction to the investigative activities. Secondly, unraveling the mechanism of toxicity will ultimately lead to identification of mechanistically relevant biomarkers because the investigative studies will link the deranged cardiovascular alterations to cellular, biochemical and molecular events. The ultimate product of this effort then would be identification of mechanistically linked, relevant and bridging biomarkers of vascular toxicity that has both nonclinical and clinical application. Employment of new technologies such as proteomics, metabonomics and transcriptomics should be utilized as appropriate in these scientific investigations. Use of specific gene-targeted transgenic and knock-out and knock-in animals should provide useful information on the precise role of immune cells in drug-induced vascular injury. It is also clear that a balance in the interaction between leukocytes and endothelial cells is critical for immune system and vascular homeostasis. In contrast, perturbations in this balance can contribute to vascular and immune system anomalies. Thus, understanding the effect of drugs on EC and immune cells would lead to development of better models to study vascular injury and further help identify biological markers that can evaluate and predict drug-induced vascular injury.


    1. Introduction: Why Vascular Injury Biomarkers?
    2. Data from animal toxicology studies are generally needed from drug developers to demonstrate for regulatory agencies a safe strategy for initiating human clinical trials. When drug-induced toxicities, such as vasculitis (or vascular injury), occur in animals, reasonable assurances for safety are needed from animal studies when appropriate clinical monitoring is not feasible. Among the many goals for performing animal toxicology studies are: 1) an identification of all dose-limiting and associated toxicities and their reversibility, and identification of the dose-exposure and dose-response relationships seen up to maximally tolerated doses, 2) an identification of a safe starting dose for human trials, and 3) a definition of an appropriately safe patient monitoring strategy for clinical trials. Data from animal studies are best applied to define the relationship between dose, drug exposure, therapeutic effect, and dose-limiting toxicity in order to evaluate appropriateness for clinical trials. In such animal studies involving extensive histopathological evaluations, however, the use of serum or plasma biomarkers is often limited, in routine analyses, as described by a working group of clinical pathologists in 1996 . This working group has recommended routine application of a core set of clinical endpoints for routine toxicity screening. These endpoints, together with hematology data, serve as reporters of diminished kidney function (e.g., BUN, creatinine), altered liver integrity (e.g., ALT, AST), altered general homeostasis (e.g., electrolytes, pH), or tissue response to injury (e.g., alterations in circulating cell populations). These same clinical endpoints have been used for routine monitoring for well over 25 years despite great strides in biological research over this same time frame. The integrity of most tissues is either not being monitored or only monitored on a limited basis and clearly vascular injury is not covered.

      New chemical entities from several pharmacologic classes have been developed that present with histopathological evidence of vascular injury that is clearly associated with drug administration in a dose and time dependent manner. Many of the earlier studies that have been published in the peer-reviewed literature describing such findings, were seen with agents at doses that produced dramatic effects on vasomotor tone, blood pressure and heart rate. A reasonable level of assurance was felt that if higher doses were avoided and cardiovascular parameters were monitored in dose escalation clinical trials, the vascular injury could be avoided in the clinic. However, many of the agents under development more recently have been associated with histopathological evidence of vascular injury in animal toxicology studies, without any easily measureable effects on vascular tone, blood pressure, or heart rate. In addition, in some cases the therapeutic index is not broad. [The therapeutic index is defined as the multiple between the exposure achieved at the highest dose at which no dose limiting toxicity is discovered, as compared to the exposure achieved at the lowest dose achieving therapeutic effect in the same species.] The lack of a practical clinical monitoring strategy often results in the sponsor and regulator sorting through speculative discussions surrounding relevance of animal findings to humans, and the inaccuracies of any defined exposure-response relationships during the proposed ensuing clinical phases of development. There is growing concern and recognition of the role of vascular injury and vascular inflammation as a risk for cardiovascular disease, myocardial infarction, and stroke . There are also experiences indicating that toxicities can sometimes be species specific and irrelevant to humans. There is a need and desire for the identification and validation of mechanism-based easily accessible biomarkers that can be measured in animal toxicology studies and integrated with histopathology studies to demonstrate their ability to herald the early onset and monitor progression or reversibility of drug-induced vascular injury across pharmacologic classes and across species, including humans.

    3. Scope of the problem and paths forward
    4. The pathogenesis of human idiopathic vasculitides are poorly understood and definitive and distinctive diagnoses of each presentation can be difficult. The underlying mechanisms, the earliest initiating events, the role and interplay of specific immune cell components involved, the vessels targeted, the progression of early disease to later forms, the availability of diagnostic biomarkers are all important areas under active investigation. Currently recognized biomarkers for human vasculitides are of limited value, tend to be focused on immune manifestations of the disease and may reflect late stages of disease. For example, anti-neutrophil cytoplasmic antibodies to neutrophil myeloperoxidase (pANCA) and to neutrophil anti-proteinase 3 (cANCA) are known to be late-appearing and are not present in all forms of the clinical vasculitides. Furthermore, these and other markers which are presently the subject of ongoing clinical research are of unknown and largely untested value to animal toxicology studies. However, in the case of atherosclerosis, another systemic inflammatory disease, several prospective clinical studies have shown that plasma levels of high-sensitivity C-reactive protein (hsCRP), an acute phase reactant, is a strong predictor of future myocardial infarction and stroke in otherwise healthy men and women . Measurement of hsCRP in humans may soon become routine as a screen for risk of cardiovascular events. For the many examples of drug-induced vascular injury that have been noted in animal models, there are likely to be differences among the models and a blurring of boundaries across mechanisms. It will be important to carefully characterize the presentation of histopathology by identifying the vessels and tissues affected and to link this data with biomarker alterations, and to understand pathogenesis, to understand the role of specific immune components in initiation and progression of disease. For pragmatic solutions, the identification of practical easily accessible diagnostic interspecies biomarkers is of prime importance. Biomarkers useful to drug and biologic development would ideally: be useful for defining specific pathogenic mechanisms, be easily accessible, be sensitive and specific to appear with the earliest manifestations of vascular injury, define the site and type of vessel affected, track with disease severity and regress as disease mitigates, be measurable with reagents that identify regions of proteins conserved across species including humans. It is also recognized that biomarkers whose performance attributes may fall short of these ideals, may nevertheless have some value in specific contexts. It is likely that because pathogenic mechanisms will be different across pharmacologic classes and even across species within a class, and that the presentation of pathology will vary by vessel and tissues affected, panels of biomarkers may be necessary to help detect all forms of vascular injury and distinguish mechanisms.

      Genomic, proteomic, and metabonomic technologies are expected to assist with questions of interspecies relevance by helping to define pathogenic mechanisms and by identifying additional monitorable interspecies biomarkers. Such approaches are both hypothesis-generating and hypothesis-directed. When careful attention is paid to study design and pivotal comparators are built onto careful investigative observation, the distinction between hypothesis generation and investigation is subtle and the value of these technologies especially in biomarker discovery should not be underestimated. Careful analyses of genomic and proteomic biomarker alterations at sites of tissue injury could elucidate pathogenic mechanisms. Investigations of altered protein and gene expression in circulating immune cells could serve a practical role in identifying accessible and monitorable biomarkers. Investigations of serum and urine components using proteomic and metabonomic technologies are already beginning to yield useful results. Expansion of such approaches is expected. In addition to such broad based biomarker discovery approaches, the systematic measurements of alterations in specific endpoints as tabulated below are described based on an evaluation of the published literature. These endpoints are hypothesized to be associated with drug-induced vascular injury and are furthermore hypothesized to add value to investigations of drug-induced vascular injury in drug developmental regulatory studies. To further investigate the value of these potential biomarkers, collaborative approaches are envisioned with the design of common protocols and appropriate controls, technology experts contributing common assay reagents, and a strategy developed toward both analytical and biological validation with a bridge toward clinical utility optimized. The following tabulation of potential biomarkers has been derived based on a recognition of a common association of manifestations of vascular injury with either 1) inflammation, inflammatory mediators, and the appearance of certain acute phase reactants; 2) the appearance of adhesion molecules and integrins; 3) the observation of excessive endothelial and smooth muscle cell activation and injury including both necrotic and apoptotic forms; 4) the observed and hypothesized involvement of diverse immune cell components including neutrophils, basophils, mast cells, and T-lymphocytes. An attempt has been made to develop some consensus toward prioritizing the investigations of the associations of these endpoints with drug-induced vascular injury.


    5. Potential Biomarkers


Rationale: Biomarker Group/Class Section Author



Reagent-Test Availability

Key points, references, confounders and other Issues (species relevancy, timing, specificity)

Current Usefulness

(A =high B=mod)

Future Promise RD Status (A=high B=mod)

Endothelial Cell Targets (Thomas)

sE-selectin (CD62E)



Rat - N Dog - N Human - Y

  • expressed selectively on endothelial cells following activation by TNF, IL-1 and LPS
  • ligands include: sialyl-Lewis (sCD15), P-selectin glycoprotein ligand-1, cutaneous lymphocyte antigen
  • involved in leukocyte tethering and rolling velocity reduction during the initial stages of inflammation and following endothelial activation
  • is shed from the endothelium and circulates as sCD62E in several inflammatory diseases that affect the endothelium, including diabetes, atherosclerosis, smoking etc
  • in healthy individuals, the levels of circulating sCD62E are very low
  • potential marker with high sensitivity and high specificity in healthy volunteers and high sensitivity and moderate to low specificity in selected patient populations
  • human tests are available
  • no rat reagents or test available
  • proposal submitted for review to set up and validate rat ELISA test system. ~10K; per sample cost will depend on volume
  • clinical site in Ireland is willing to collaborate for Fenoldopan study
  • May serve as an early biomarker for vasculitis. The maximal transcript level of E-selectin at 3-6 hr in a murine model (Fries JWU, Williams AJ, Atkins RC, Newman W, Lipscomb MF, and Collins T: Expression of VCAM-1 and E-selectin in an in vivo model of endothelial activation. Am J Pathol 143:725-737, 1993)
  • High degree of structural and functional homology of E-selectin among human, mouse, and rabbit (Becker-Andre M, van Huijiduinen RH, Losberger C, Whelan J, Delamarter JF: Murine endothelial leukocyte-adhesion molecule-1 is a close structural and functional homologue of the human protein. Eur J Biochem 206:401, 1992) (Weller A, Isenmann S, Vestweber D: Cloning of the mouse endothelial selectins: Expression of both E- and P-selectin is inducible by tumor necrosis factor. J Bio Chem 267:15176, 1992)
  • Significant increase of soluble levels of E-selectin in patients with a variety of vasculitides (» 1-3-fold increase above control subjects)



Rationale: Biomarker Group/Class Section Author



Reagent-Test Availability

Key points, references, confounders and other Issues (species relevancy, timing, specificity)

Current Usefulness

(A =high B=mod)

Future Promise RD Status (A=high B=mod)

Endothelial cells comprise the sole cell type lining the lumen of all vasculature in the body with approximately 1013 cells covering 7 m2 in the average adult human (Holash, J et. al., Science 284:1994-1998 (1999)). They regulate nutrient and blood component traffic to all tissues in the body, and are also intimately involved in control of blood flow, hemostasis and inflammation. As such, endothelial cells and their secreted and/or synthesized proteins should be considered prime targets for providing potential biomarkers of vascular injury. Circulating endothelial cells shed from the vascular wall have been documented after vascular damage due to multiple etiologies (reviewed in Dignat-George, F and J. Sampol, European J Haematol 65:215-220 (2000). Of the secreted endothelial cell proteins, von Willebrand Factor (vWF) has received the most attention as a potential biomarker of vascular damage (Lip, G.Y.H. and A. Blann, Cardiovascular Research 34:255-265 (1997)).

Circulating Endothelial Cells (CEC)


Rat - Y Dog - Y Human - Y

  • may need to use a combination of surface markers to identify including CD31, acetylated lipids (DiI-Ac-LDL)
  • shown to increase in number after varied causes of vascular damage in people
  • potential tail vein blood-drawing high background levels in rats




Rat - Y Dog - Y Human - Y

  • Elevation of plasma and serum protein after varied causes of vascular damage in people, and after drug-induced vascular injury in rats( although increase is short-lived)
  • sequential blood drawing artifacts
  • vWF elevated in smokers


vWF pre-protein


Rat - N Dog - N Human - Y

  • Sensitive marker acute and progressive injury; not found in normal plasma, rapid turnover and short half-life, releases as a 1:1 ratio with mature protein
  • no extensive human experience, baboon only animal species with data.
  • Refs:Thrombosis and Haemostasis 80:, 1002-1007; 1998 vWF pro-peptide


endothelial microparticles (EMPs)


Rat - ? Dog - N Human - Y

  • A marker of endothelial cell activation in thrombotic thrombocytopeienic purpura
  • The number of EMPs can be measured by monoclonal antibodies CD31 and CD51 using flow cytometry
  • Elevated EMPs in the plasma correlate to increased expression of ICAM-1 and VCAM-1
  • CD31 (also called PECAM-1) is expressed on endothelial cells, platelets, and megakaryocytes. It plays a role in the interaction of the endothelium and platelets
  • CD51/CD61 (also called 23 integrin) is an endothelial cell adhesion molecule, which is mainly expressed on endothelial cells but is also expressed on leukocytes and macrophages. It plays a role in the interaction of endothelial cells and extracellular matrix.
  • Refs: Jimenez JJ, Jy W, Mauro LM, Horstman LL, and Ahn YS (2001). Elevated endothelial microparticles in thrombotic thrombocytopenic purpura: Findings from brain and renal microvascular cell culture in patients with active disease. Br J Haematol. 112:81-90. Ferrer L, Fondevila D, Rabanal RM, and Vilafranca M (1995). Immunohistochemical detection of CD31 antigen in normal and neoplastic canine endothelial cells. J Comp Pathol. 112:319-326. (related to CD31and CD51on endothelial cells and platelets)


Smooth Muscle Cell Targets



Rat - Dog - Human -

  • Caveolin-1, plasmalemmal invaginated vesicle, key role in cell signaling pathway
  • expressed on EC, SMC and fibroblasts; loss impairs nitric oxide and calcium signaling causing aberrations in endothelium-dependent relaxation, contractility and maintenance of myogenic tone. See immuno histochemistry slide for caveolin-1
  • expressed on EC, SMC and fibroblasts; loss impairs nitric oxide and calcium signaling causing aberrations in endothelium-dependent relaxation, contractility and maintenance of myogenic tone. See immunohistochemistry slide for caveolin-1
  • Refs: Thrombosis and Hemostasis, 77: 387-393,


Smooth muscle surrounding vessels is often seen to be disrupted during vascular injury; smooth muscle may be disrupted by hemodynamic factors, or may be a direct target of cell injury

smooth muscle alpha actin


Rat - Y Dog - Y Human - Y

  • Smooth muscle alpha actin expressed in vascular smooth muscle
  • Antibody available which cross reacts across mammals
  • unknown whether increases in circulation after vascular injury


White Blood Cell Targets (Burchiel)



Rat - Y Dog - Y Human - Y

  • Neutrophils and monocytes are oftentimes seen in areas of vascular damage
  • unclear whether inflammatory cells are the cause or result of vascular damage


WBC infiltrates are oftentimes seen in histopathologic lesions associated with vascular damage which is termed "vasculitis" - unclear whether inflammatory cells are the cause or result of vascular damage

Mast Cells and Basophils



Rat - Y Dog - Y Human - Y

  • Recent studies by Zhang, Hermann, and Sistare at CDER demonstrate that mast cells are oftentimes found associated with lesions in SHR Wistar-Kyoto rats; due to the rapid onset of vascular lesions in this species, a possible mechanism of vascular injury may relate to "pseudoallergy"
  • Recent studies by Zhang, Herman and Sistare at CDER demonstrated that increased mast cells are oftentimes associated with mesenteric vascular injuries in SD, SHR, and WKY rats treated with the PDE III inhibitor, SK&F 95654; the possible release of inflammatory mediators and cytokines from mast cell degranulation is suggested (Zhang J, Herman EH, Knapton A, Chadwick DP, Whitehurst VE, Korerner JE, Papoian T, Ferrans V, and Sistare FD: SK&F 95654-induced acute cardiovascular toxicity in Sprague-Dawley rats: Histopathologic, electron microscopic, and immunohistochemical studies. Toxicol Pathol 30:28-40, 2002)


Acute Phase Reactant Proteins (Blanchard)

C-reactive protein (CRP)


Rat - Y dog - Y Human - Y

  • C-reactive protein is relatively conserved in its response amongst species (rat, dog and man). CRP is believed to bind bacterial and fungal cell walls, resulting in opsonized particles and serving as a complement cascade attractant; CRP is elevated in animals (rats and dogs) with vascular lesions

A (for relevant species)

An acute phase inflammatory response, often elicited by endogenous pyrogens such as IL-1, IL-6 or TNFa, involves the secretion of a variety of liver proteins into the blood (Loeb and Quimby 1999). Not unexpectedly, this response varies amongst species and is a general marker of inflammation not specific to vasculitis. Two such acute phase reactants, namely c-reactive protein and serum amyloid A, have recently been reported to coincide with drug-induced vasculitis in the dog (Zhang 2002)

Serum amyloid A


Rat - N Dog - Y Human - Y

  • Serum amyloid A appears to respond as an acute phase protein in man and dog, but may not be an appropriate marker in rat
  • The function of serum amyloid A is not completely understood but may play a role in lipid metabolism during inflammation.

A (for relevant species)



Rat - Y Dog - Y Human - Y

  • Haptoglobin is a mucoprotein that traps hemoglobin from lysed red blood cells; it has been reported to be elevated in inflammatory disorders. Zhang J et al., Acute Phase Proteins (APPs) as biomarkers of host-response to drug-induced vasculitis in rats and dogs. 2002 FDA Scientific Forum, Abstract T3
  • haptoglobin has been found to be elevated in dogs with vascular lesions
  • Ref: Zhang J, Herman E, Holt G, Blanchard K, Ratajczak H, Knapton A and Sistare F: Acute phase proteins (APPs) as biomarkers of host-response to drug-induced vasculitic in rats and dogs. Toxicological Sciences. 66 (1-S):15, 2002

A (for relevant species)



Rat - Y Dog - Y Human - Y

  • An acute phase protein related to the release IL-6 in vasculitis
  • -fibrinogen is increased to 4.8 fold 4 hr after the administration of IL-6 in male Wistar rats (Geiger T, Andus T, Klapproth J, Hirano T, Kishimoto T, and Henrich PC (1988). Induction of rat acute-phase proteins by interleukin 6 in vivo. Eur J Immunol 18:717-721)
  • Fibrinogen is increased to 3.8 fold after the administration of IL-6 in human hepatocytes (Castell JV, Gomez-Lechon MJ, David M, Hirano T, Kishimoto T, and Heinrich PC (1988). Recombinant human interleukin-6 (IL-6/BSF-2/HSF) regulates the synthesis of acute phase proteins in human hepatocytes. FEBS Lett. 232:347-350)




Rat - Y Dog - N Human - N

  • this particular protein may be a respondent specifically in the rat and is not relevant for dogs and humans
  • some data demonstrate that this marker is elevated in rats with vascular lesions

A (for relevant species)

Inflammatory (Nagarkatti)



Rat - Y Dog - Y Human - Y

  • Interleukin-1 (IL-1) is an important mediator of the immune system, playing key role during infection, inflammation, cell-differentiation, tissue remodelling, and apoptosis
  • It may act as an activator of cardiovascular cells inasmuch as cells of the vessel wall and the heart can produce IL-1
  • IL-1 can be detected in the serum using ELISA; one plate costs ~$300.


Inflammatory cytokines and chemokines, much like acute phase reactant proteins, are nonspecific indicators of inflammation; Pober JS, Cotran RS. Cytokines and endothelial cell biology. Physiol Rev. 1990; 70: 427-51.



Rat - Y Dog - Y Human - Y

  • IL-6 is primarily produced by macrophages, TH2 cells and bone marrow stromal cells
  • There are reports suggesting that in some types of vasculitis, IL-6 may play a critical role
  • It induces synthesis of acute phase proteins in the liver
  • IL-6 can be detected in the serum using ELISA; One plate costs ~$300
  • Ref: HogenEsch H, Snyder PW, Scott-Moncrieff JCR, Glickman LT, and Felsburg PJ: Interleukin-6 activity in dogs with juvenile polyarteritis syndrome: Effect of corticosteroids. Clin Immunol Immunopathol. 77:107-110, 1995.




Rat - Y Dog - Y Human - Y

  • TNF is a major mediator of apoptosis as well as inflammation and immunity
  • TNF is the prototypic proinflammatory cytokine and endothelial cells are the principal cellular targets of its actions
  • It is found in the serum and can be detected using ELISA; one plate costs ~$300.


PG's and LT's


Rat - Y Dog - Y Human - Y

  • RIAs and ELISAs exist for a variety of prostaglandin (PGs) and leukotrienes (LTs) and their major serum and urinary metabolites
  • these are nonspecific indicators of inflammatory disease are of unproven use in vasculitic


Chemokines (MIP-1, MCP, IL-8)


Rat - Y Dog - Y/N Human - Y

IL-8 (chemokine)

  • A marker of endothelial cell activation (Cockwell P, Tse WY, and Savage COS (1997). Activation of endothelial cells in thrombosis and vasculitis. Scan J Rhematol 26:145-150)
  • An activator of neutrophils (formerly known neutrophil activating protein-1) and a potent neutrophil chemoattractant
  • It plays a role in the transendothelial migration of neutrophils, particularly in post-capillary venules
  • It enhances the interaction of neutrophils with ICAM-1
  • Significant increase inn soluble levels of IL-8 in patients with vasculitis (Tesar V, Jelinkova E, Masek Z, Jirsa M Jr, Zabka J, Bartunkova J, Stejskalova A, Janatkova I, and Zima T (1998). Influence of plasma exchange on serum levels of cytokines and adhesion molecules in ANCAS-positive renal vasculitis. Blood Purif 16;72-80)

MCP-1, MIP-1, RANTES (chemokines)

  • MCP-1, MIP-1, and RANTES are synthesized and secrete by activated endothelial cells
  • MCP-1 is an activator of monocytes and a monocyte chemoattractant, thereby playing a role in the regulation of monocyte extravasation
  • MIP-1 and RANTES are selectively chemoattractant for T cells
  • Significant increase in MCP-1, MIP-1a, and RANTRES gene expression levels in patients with vasculitis (Wong M, silverman ED, and Fish EN (1997). Evidence for RANTES, monocyte chemotactic protein-1, and macrophage inflammatory protein-1a expression in Kawasaki disease. J Rheumatol 24:1179-1185)
  • various assays exist for chemokines and cytokines, including those that are protein-based as well as mRNA-based. The useful of these markers in vasculitis has not been established


Adhesion Molecules (Zhang)

sCD44-Hyaluronic Acid


Rat - Y Dog - Y/N Human - Y

  • Activation of T cells by Ag or stimulation of monocytes with inflammatory cytokines induces CD44 to bind to hyaluronic acid (HA), an adhesion event implicated in leukocyte-leukocyte and leukocyte-endothelial cell interactions (Lesley J, Hyman R, Kincade PW. CD44 and its interaction with extracellular matrix. Adv Immunol 1993; 54:271-335.)
  • CD44 expressed on activated T cells and NK cells is involved in endothelial cell injury and vascular leak (Rafi-Janajreh AQ, Chen D, Schmits R, Mak TW, Grayson RL, Sponenberg DP, Nagarkatti M, Nagarkatti PS. Evidence for the involvement of CD44 in endothelial cell injury and induction of vascular leak syndrome by IL-2. J Immunol. 1999; 163 :1619-27)
  • Also, activated endothelial cells also express CD44 (Griffioen AW, Coenen MJ, Damen CA, Hellwig SM, van Weering DH, Vooys W, Blijham GH, Groenewegen G. CD44 is involved in tumor angiogenesis; an activation antigen on human endothelial cells. Blood. 1997;90:1150-9)
  • may be unique isoforms expression by endothelial cells; thus, detection of soluble CD44 and HA in blood may constitute a useful biomarker for vasculitis




Rat - Y Dog - Y/N Human - Y

  • MAdCAM-1 is an endothelial cell adhesion molecule, which is specifically expressed on the high endothelial venules in lymphoid tissues
  • L-selectin and 47 integrin are receptors of MAdCAM-1
  • Involvement of L-selectin-mediated or 7 integrin lymphocyte rolling
  • In rat, 4 integrins on the surface of neutrophils interact with MAdCAM-1 on the endothelium, suggesting that rat neutrophils may use 4 integrins to mediate selective recruitment of neutrophils to sites of inflammation in vivo ( Davenpeck KL, Sterbinsky SA, and Bochner BS (1998)
  • Rat neutrophils express 4 and 1 integrins and bind to vascular cell adhesion molecule-1 (VCAM-1) and mucosal addressin cell adhesion molecule-1 (MAdCAM-1). Blood. 91:2341-2346
  • unique to Peyer's patch endolthelial cells; Berg EL, McEvoy LM, Berlin C, Bargatze RF, and Butcher EC (1993)
  • L-selectin-mediated lymphocyte rolling on MAdCAM-1. Nature. 366:695-698. Connor EM, Eppihimer MJ, Morise Z, Granger DN, and Grisham MB (1999).
  • Expression of mucosal addressin cell adhesion molecule-1 (MAdCAM-1) in acute and chronic inflammation. J Leukoc Biol. 65:349-355.




Rat - Y Dog - Y/N Human - Y

  • ICAM-1 serves as a marker of endothelial cell activation (Cotran RS (1987). New role for the endothelium in inflammation and immunity. Am J Pathol 129:407-413)
  • ICAM-1 on the activated endothelial cells as a ligand interacts with Mac-1 (CD11b/CD18), a leukocyte integrin receptor
  • ICAM-1 plays a role in margination and extravasation of lymphocytes at sites of inflammation or a localized immune response
  • Increased levels of soluble ICAM-1 in patients with vasculitis (Tesar V, Jelinkova E, Masek Z, Jirsa M Jr, Zabka J, Bartunkova J, Stejskalova A, Janatkova I, and Zima T (1998).
  • Up-regulated ICAM-1 expression on the endothelial cells of the mesenteric artery in rat treated with SKF (Zhang et al)
  • Refs: Gonzalez-Amaro R and Sanchez-Madrid F (1999). Cell adhesion molecules: selectins and integrins. Crit Rev Immunol. 19:389-429. Coll-Vinent B, Grau JM, Lopez-Soto A, Oristrell J, Font C, Bosch X, Mirapeix E, Urbano-Marquez A, and Cid MC (1997). Circulating soluble adhesion molecules in patients with classical polyarteritis nodosa. Br J Rheumatol. 36:1178-1183.




Rat - Y Dog - Y/N Human - Y

  • A marker of endothelial cell activation
  • VCAM-1 on the activated endothelial cells as a ligand interacts with VLA-4 (CD49d/CD29), leukocyte 41 integrin
  • VCAM-1 plays a role in the firm attachment and subsequent transendothelial migration of leukocytes
  • Increased levels of soluble VCAM-1 in patients with vasculitis (Ara J, Mirapeix E, Arrizabalaga P, Rodriguez R, Ascaso C, Abellana R, Font J, and Darnell A (2001). Circulating soluble adhesion molecules in ANCA-associated vasculitis. Nephrol Dial Transplant 16:276-285)
  • Refs: Ara J, Mirapeix E, Arrizabalaga P, Rodriguez R, Ascaso C, Abellana R, Font J, and Darnell A (2001). Circulating soluble adhesion molecules in ANCA-associated vasculitis. Nephrol Dial Transplant. 16:276-285. Gearing AJH and Newman W (1993). Circulating adhesion molecules in disease. Immunol Today. 14:506-512. Gonzalez-Amaro R and Sanchez-Madrid F (1999). Cell adhesion molecules: selectins and integrins. Crit Rev Immunol. 19:389-429.


Immediate Hypersensitivity -Pseudoallergen (Burchiel)



Rat - NR Dog - Y Human - Y

  • histamine is the main vasoactive mediator released from mast cells and basophils in dogs and humans


As discussed above for mast cell and basophil targets, current research indicated that psudeoallergy may be a mechanism of altered flow and permeability of vessels, as well as inflammatory mediator release; there are numerous products of activated mast cells and basophils that could be potential biomarkers; however, at this time none of these markers have been associated with vascular injury



Rat -Y Dog - N Human - N

  • serotonin (5HT) is the main vasoactive mediator released from mast cells and basophils in rats and other rodents


PG's and LT's


Rat - Y Dog - Y Human - Y

  • potential marker, but no data to date




Rat - Dog - Human -

  • potential marker, but no data to date




Rat - Dog - Human -

  • potential marker, but no data to date


Apoptosis (Nagarkatti)



Rat - Y Dog - Y/N Human - Y

  • FasL is a 40-kd type 2 transmembrane protein belonging to the tumor necrosis factor (TNF) family
  • ligation of Fas, a cell-surface protein transduces an apoptotic signal leading to cell death
  • FasL is expressed mainly on activated leukocytes but recent studies suggest that it may also be expressed on endothelial cells; membrane bound FasL is rapidly cleaved into a soluble form which can be detected in the serum (1)
  • Fas ligand has been implicated in endothelial cell injury (2,3). Currently there are no tests in the market to detect soluble FasL in the serum
  • Nagarkatti lab has published a paper on detection of mouse sFasL using ELISA (1), and are currently developing an ELISA for rat FasL. Zeytun A, Nagarkatti M, Nagarkatti PS
  • Growth of FasL-bearing tumor cells in syngeneic murine host induces apoptosis and toxicity in Fas(+) organs. Blood. 2000 ; 95 :2111-7. Janin A, Deschaumes C, Daneshpouy M, Estaquier J, Micic-Polianski J, Rajagopalan-Levasseur P, Akarid K, Mounier N, Gluckman E, Socie G, Ameisen JC. CD95 engagement induces disseminated endothelial cell apoptosis in vivo: immunopathologic implications. Blood. 2002; 99:2940-7. Rafi AQ, Zeytun A, Bradley MJ, Sponenberg DP, Grayson RL, Nagarkatti M, Nagarkatti PS. Evidence for the involvement of Fas ligand and perforin in the induction of vascular leak syndrome. J Immunol. 1998 ;161:3077.


Apoptosis may be a mechanism of target cell injury which may be detected through various protein-based and biochemical assays, cell-based flow cytometry, and immunohistochemical studies; it is also possible that peripheral blood leuokocytes may serve as surrogates for tissues undergoing apoptosis [Burchiel]



Rat - Y Dog - Y/N Human - Y

  • there are a series of caspases that are activated during apoptosis that can be detected using cells undergoing apoptosis
  • this assay would require having sufficient numbers of recovered apoptotic cells to perform the biochemical assays


Annexin V


Rat - Y Dog - Y/N Human - Y

  • Annexin V is a useful marker for human and rodent apoptosis
  • Annexin V binds to cell membranes that bleb during the apoptotic process exposing phosphatidyl serine residues
  • Annexin V techniques have not proven useful in dog studies




Rat - Y Dog - Y/N Human - Y

  • the TUNEL assay can be performed using both flow cytometry and immunohistochemistry
  • it relies upon DNA repair enzymes that will incorporate biotinylated UTP into DNA


Immune Mediated

(Miller and Zhang)

Drug-induced vascular injury in man was initially described over three decades ago and now over 50 chemical and biologic therapeutics and vaccines have been associated with these syndromes. These agents can induce a variety of vasculitic manifestations ranging from small vessel hypersensitivity vasculitis and leukocytoclastic vasculitis to distinct vasculitic syndromes such as Wegener's granulomatosis, polyarteritis nodosa, and Churg Strauss syndrome. The agents associated with vascular injury in humans do not share any obvious structural, pharmacological or immunological properties. The pathogenic mechanisms remain to be defined and appear to be multifactorial. Indirect evidence, however, implicates chronic cellular and humeral immune activation in genetically susceptible individuals. Some cases associated with urticaria, angioedema, anaphylaxis and anaphylactoid reactions may involve Type I immunoglobulin (Ig)-mediated or Type III hypersensitivity, or may be caused by pharmacological, non-allergic means. Many drug-induced vasculitis cases though are possible manifestations of immune complex disease. These syndromes can be life-threatening diseases, and while most cases respond to dechallenge (removal of the suspect agent), many require additional concomitant immunosuppressive therapy for full resolution. References: Cuellar ML. Drug-induced vasculitis. Curr Rheumatol Rep 2002 Feb;4(1):55-9; D'Cruz D. Autoimmune diseases associated with drugs, chemicals and environmental factors. Toxicol Lett 2000 Mar 15;112-113:421-32; Jain KK. Drug-induced cutaneous vascu-litis. Adverse Drug React Toxicol Rev 1993 Winter;12(4):263-76; Dubost JJ, Souteyrand P, Sauvezie B. Drug-induced vasculitides. Baillieres Clin Rheumatol 1991 Apr;5(1):119-38

Autoantibody (ANCA); Immune complex & anaphylatoxin



Rat - Y Dog - Y/N Human - Y

Myeloperoxidase (MPO) & Proteinase 3 (PR3)

  • Autoantibodies directed against cytoplasmic antigens of neutrophils (ANCA), especially those with specificity for MPO and PR3, are valuable markers for vasculitides.
  • ANCA are specific granule proteins of granulocytes and monocytes
  • The cytoplasmic (classic) cANCA is induced by antibodies directed against protainase 3 (PR3; PR-ANCA) while the perinuclear pANCA is induced by antibodies against myeloperoxidase (MPO; MPO-ANCA) (Gross WL, Schmitt WH, Crernok E (1993). ANCA and associated diseases: immunodiagnostic and pathogenetic aspects. Clin Exp Immunol 91:1-12) (Mayet WJ, Schwarting A, and Meyer zum Buschenfelde KH (1994). Cytotoxic effects of antibodies to protainase 3 (C-ANCA) on human endothelial cells. Clin Exp Immunol 97:458-465)
  • A positive correlation between MPO-ANCA and PR3-ANCA levels and increased sE-selectin, sP-selectin, sICAM-1, and sVCAM-1 was reported in patients with ANCA-associated vasculitis using the quantitative anti-MPO-ELISA and anti-PR3-ELISA (Ara J, Mirapeix E, Arrizabalaga P, Rodriquez R, Ascaso C, Abellana R, Font J, and Darnell A (2001). Circulating soluble adhesion molecules in ANCA-associated vasculitis. Nephrol Dial Transplant 16:276-285)
  • Cytokines (TNF-, IL-, increase PR3 expression in the cytoplasm of endothelial cells. A possible direct pathogenic effect of anti-PR3 antibodies, through ANCA-endothelial interaction, has been suggested in vasculitides (Mayet WJ, Crernok E, Szymkowiak C, and Gross WL, Meyer zum BUSCHENFELDE KH (1993). Human endothelial cells express proteinase 3, a target antigen of anticytoplasmic antibodies in Wegener's granulomatosis. Blood 82:1221-1229)
  • Direct cytotoxic effects of antibodies to PR3 on human vascular endothelial cells a PR3 antibody-mediated mechanisms of endothelial injury (Mayet WJ, Schwarting A, and Meyer zum Buschenfelde KH (1994). Cytotoxic effects of antibodies to protainase 3 (C-ANCA) on human endothelial cells. Clin Exp Immunol 97:458-465); Refs: Vorcheimer and Fuster: JAMA 2001, 286:2154; Lindmark, et al: JAMA 286:2107; Zhang et al: JAMA 286:2136

Blood Coagulation Factors (Zhang)



Rat - Y Dog - Y Human - Y

  • Important regulator of activated thrombin - TM converts thrombin from a procoagulant to an anticoagulant
  • is a constitutive membrane protein - therefore probably needs to be cleaved into soluble forms to be detected in circulation
  • increased circulating concentrations of an antigenic component has been reported with thrompocytopenic purpura, disseminated intravascular coagulation, atherosclerosis and inflammatory connective tissues diseases
  • in vitro studies suggest TM is in endothelial cell tissue culture supernatants suggesting comes from damaged endothelial cells; Other cell sources - synovial cells, rheumatoid patients, platelets, megakaryocytes, mesothelial cells, PMN's; unlike vWF, TM concentrations appear independent of inflammatory cytokines IL-1 and TNF.; unknown if assay works in humans- utility in rats [Schwartz]Boffa MC and Karmochkine M (1998).
  • Refs: Thrombomodulin: An overview and potential implications in vascular disorders. Lupus. Suppl 2: S120-S125. Boehme MWJ, Raeth U, Scherbaum WA, Galle PR, and Stremmel W (2000). Interaction of endothelial cells and neutrophils in vitro: Kinetics of thrombomodulin, intercellular adhesion molecule-1 (ICAM-1), E-selectin, and vascular cell adhesion molecule-1 (VCAM-1): Implication for the relevance as serological disease activity markers in vasculitides. Clin Exp Immunol. 119:250-254.




Rat - Y Dog - Y/N Human - Y

  • The activated endothelial cells release vWF into the plasma and subendothelial matrix through exocytosis of the cytoplasmic Weibel-Palade bodies
  • Exocytosis of vWF from Weibel-Palade bodies is associated with the mobilization of P-selectin from Weibel-Palade bodies to the surface of endothelial cells
  • High levels of vWF in patients with vasculitis
  • In endotoxin-treated or balloon-injured rats, the immunostaining content of intracellular vWF was increased to three-fold, indicating that endothelial cell injury leads to increased vWF levels (Reidy MA, Chopek M, Cao S, McDonald T, and Schwartz SM (1989). Injury induces increase of von Willebrand factor in rat endothelial cells. Am J Pathol 134:857-864)
  • Up-regulated vWF expression on the endothelial cells of the mesenteric artery in rats treated with SKF (Zhang et al)
  • High molecular weight procoagulant product of endothelium - is a constituent of platelets (alpha granules) and mRNA of vWF is found in platelets, vWF can also be obtained from isolated platelets in vitro
  • backers suggest that circulating vWF is specific marker of endothelial damage and contribution from platelets is minimal. Platelet vWF after discharge remains bound to platelet surface
  • component of Wiebel-Palady bodies in endothelial cells (storage intracellular) also secreted constitutively secretion agonists - thrombin, IL-1, vasopressin
  • Clinical evidence of increased circulating vWF - inflammatory and atherosclerotic vascular diseases, after surgical procedures - bypass surgery, infarction with ischemic heart disease, thromboembolic disease, rheumatoid arthritis, increased with acute phase reactants; sensitive marker - perhaps too sensitive 18 hr half life [Schwartz].
  • Refs: Newsholme SJ, Thudium DT, Gossett KA, Watson ES, and Schwartz LW (2000). Evaluation of plasma von Willebrand factor as a biomarker for acute arterial damage in rats. Toxicol Pathol. 28:688-693. Cid MC, Monteagudo J, Oristrell J, Vilaseca J, Pallares L, Cervera R, Font C, Font J, Ingelmo M, and Urbano-Marquez A (1996). von Willebrand factor in the outcome of temporal arteritis. Ann Rheum Dis. 55:927-930.


Tissue Factor


Rat - Dog - Human -

  • A procoagulant, not expressed in normal endothelial cells
  • Cytokine-activated endothelial cells increase TF procoagulant activity (Cotran RS (1987). New role for the endothelium in inflammation and immunity. Am J Pathol 129:407-413)
  • A marker of endothelial cell activation
  • Increased levels of soluble TF in patients with vasculitis (Jurd KM, Stephens CJM, Black MM, and Hunt BJ (1996). Endothelial cell activation in cutaneous vasculitis. Clin Exp Dermatol 21:28-32)
  • Refs: Bach FH, Robson SC, Ferran C, Winkler H, Millan MT, Stuhlmeier KM, Vanhove B, Blakely ML, van der Werf WJ, Hoffr E, de Martin R, and Hancock WW (1994). Endothelial cell activation and thromboregulation during xenograft rejection. Immunol Rev. 141:5-30. information about time course of TF, TM, vWF in vitro study)




Rat - Y Dog - Y Human - Y

  • An anticoagulant
  • Cytokine-activated endothelial cells decrease tPA anticoagulant activity and increase plasminogen activating inhibitor-1 (PAI-1) procoagulant activity
  • Increased levels of soluble tPA in patients with vasculitis (Jurd KM, Stephens CJM, Black MM, and Hunt BJ (1996). Endothelial cell activation in cutaneous vasculitis. Clin Exp Dermatol 21:28-32)
  • Refs: Cockwell P, Tse WY, and Savage COS (1997). Activation of endothelial cells in thrombosis and vasculitis. Scand J Rhematol. 26:145-150. Jurd KM, Stephens CJM, Black MM, and Hunt BJ (1996). Endothelial cell activation in cutaneous vasculitis. Clin Exp Dermatol. 21:28-32


Hemodynamic (Schwartz)

vascular bed pressure - flow or imaging


Rat - Y Dog - Y Human - Y

  • vascular flow and pressure changes may or may not be predictive of vascular injury – the techniques needed to measure flow in animals are invasive and as such are not true biomarkers
  • in humans there are new imaging techniques that may be useful to measure flow changes in regional vasculatures


New Technologies



Rat - Y Dog - Y Human - Y

  • techniques such as Power Blotting, 2D gel electrophoresis and 2D HPLC combined with SELDI, LC/MS/MS and MALDI techniques may be useful to define profiles of proteins that may be associated with vascular injury;
  • this work is more hypothesis-generating than hypothesis-driven and is at an early stage of development
  • the hope is that novel protein markers may be discovered through global scanning or proteins




Rat - Y Dog - Y/N Human - Y

  • techniques such as gene array (Affymetrix and other arrays) may be useful to define profiles of genes that may be associated with vascular injury
  • this work is more hypothesis-generating than hypothesis-driven and is at an early stage of development
  • the hope is that novel gene markers may be discovered through global scanning and that quantitative PCR technique can be developed to monitor gene biomarkers


Initial studies suggest that metabonomics NMR profiling and computertechniques may be useful to non-invasively identify animals with vascularinjury; however, the observed changes in biofluids may be secondary or

tertiary to the mechanism of vascular injury or may be related to

physiological changes or non-vascular toxicity confounding interpretation of

the data.



Rat - Y Dog - Y Human - Y

  • initial studies suggest that metabonomics NMR profiling and computer techniques may be useful to identify animals with vascular injury; however,
  • these techniques rely on nonspecific changes in kidneys and other organs that may occur far downstream from the toxicologic events and often do not allow for mechanistic interpretation




    1. Toxicogenomics
      1. Background
      2. Toxicogenomics is an emerging discipline that uses genomic tools including gene expression profiling technologies to address problems of toxicological significance. Several recent publications on toxicogenomics have been published and can be reviewed for a more detailed discussion of its general principles . As applied to vascular injury, toxicogenomics can contribute to a mechanistic understanding of the lesion, molecular basis of species differences, aid in classifying various types of vasculopathy, development of gene-based screens to improve selection of compounds for drug development, and identification of more sensitive and specific biomarkers. To date no reports specifically describing the use of toxicogenomics techniques for vascular injury in preclinical animal models have been published. However, there have been numerous publications from the field of cardiovascular disease demonstrating that arterial tissue can be characterized using expression profiling approaches, and therefore the mechanistic understanding of vascular injury should be a realistic goal .

        There are at least two ways that toxicogenomic approaches can facilitate the identification of vascular injury biomarkers. First, by monitoring genes that are differentially expressed in blood vessels isolated from animals treated with compounds that induce vascular injury, it might be possible to narrow the search for candidate biomarkers by specifically focusing on genes that encode cell-surface or secreted proteins. These genes encode potential circulating biomarkers that could be further characterized directly in plasma or serum using immunoassays or other diagnostic methods. Second, surrogate biomarkers might also be identified by profiling circulating leukocytes, which can be easily obtained from whole blood. Alcorta et al. have reported that gene expression changes in circulating leukocytes from patients with a variety of renal diseases, including small vessel vasculitis (ANCA disease), can be clustered according to disease type. The latter approach will be particularly useful for characterizing vascular injury in humans.

      3. Platform(s)
      4. There are numerous technologies and platforms that can be used for genome-wide expression profiling. While there are several ways to categorize genomic approaches, most relevant for the study of vascular injury are so-called "closed" and "open" profiling systems. Closed systems typically use a microarray-type format and are based on pre-defined sets of genes. Such genes could be selected to address a specific hypothesis (e.g. oxidative-stress array), to characterize certain tissues (e.g. kidney array), diseases (e.g. breast cancer array), or profile an entire genome (e.g. rat genome array). Microarrays are generally based on cDNA clones, PCR-amplified cDNA fragments, or chemically synthesized oligonucleotides that are "spotted" or affixed to a solid support such as a glass slide or a nylon membrane. These arrays can be obtained from a variety of commercial vendors or constructed in house using commercially available or homemade equipment. Another type of array—the GeneChipÔ from Affymetrix—is based on oligonucleotides that are synthesized directly on a glass surface. Regardless of the platform used, experiments typically involve isolating RNA from control and treated tissues or cell lines, labeling the RNA, and hybridizing it to the array. Depending on the specific platform, absolute or differential expression levels of the genes represented on the arrays are then calculated and used for statistical analysis. Because mechanisms of vasculitis are not known, it is prefered to use the genome-wide microarrays that are generally available for rat, mouse, and human. Although these closed systems are evolving as the genomes become better defined, genes critical to the pathogenesis of vascular injury could be overlooked.

        In contrast to the "closed" systems, "open" expression-profiling systems do not begin with a pre-selected set of genes. Instead, the differences between two populations of RNA are determined directly, generally using some combination of RT-PCR, gel electrophoresis, and DNA sequencing. Examples include SAGE (Serial Analysis of Gene Expression), differential display, and subtractive hybridization. Kits for many of these techniques are commercially available, and several biotechnology companies use proprietary versions to build expression databases. One advantage of open methods for characterizing vascular injury is that, depending on the platform’s sensitivity, most of the genes directly involved in the lesion should be identified using these techniques; this would include some genes that have not been previously identified and therefore would not be represented on microarrays. Another advantage is that they may be better suited for characterizing vascular injury in species whose genomes have not been extensively sequenced, such as the dog. Limitations of the open methods include their technical difficulty, time-consuming nature, and expense.

      5. Knowledge Gaps/Limitations

      There are a number of technical issues that must be considered when working with blood vessels. Certainly one of the most significant is that most blood vessels are small, making it difficult to obtain sufficient quantities of RNA for most profiling methods. This problem is of course more acute in rodents than dogs. The development of increasingly sensitive amplification techniques offers promise that starting RNA amounts will not be a limitation in the future. Moreover, like most tissues, blood vessels are complex tissues comprised of multiple cell types and can be found in close association with surrounding connective tissues such as fat, pancreas or lymph nodes. When interpreting expression data from vascular tissue, this complexity can make it difficult to distinguish the contributions of endothelial cells from vascular smooth muscle cells (VSMCs), infiltrating leukocytes, and surrounding tissues. One possibility is to perform in situ hybridization or immunohistochemistry to localize the cellular source of specific transcripts of interest, assuming the necessary antibody and cDNA reagents are available or can be generated. Another is to use laser capture microdissection (LCM) to isolate "pure" populations of endothelial cells, VSMCs, or other targeted cell populations . Inherent to the lack of mechanistic understanding of vascular injury, good in vitro models are not available, limiting the ability to generate higher-throughput gene-based in vitro screens for vascular injury-inducing potential.

      As with all large-scale expression profiling experiments, data analysis and interpretation remains a key challenge. Lists of genes that are up- and down-regulated in tissues from animals with vascular injury will easily be generated, but how can we separate cause from effect? This task will be facilitated by solid experimental designs that include time courses (i.e. take samples at early time points to capture potential initiating events), dose-response (i.e. include a low dose that does not cause toxicity to help separate pharmacologically mediated changes in gene expression from toxicological ones), and careful choice of positive and negative controls (i.e. generating expression data from animals treated with compounds that cause inflammation but not vascular injury will facilitate the search for more specific biomarkers). Furthermore, bioinformatic approaches to link differentially expressed genes to altered metabolic and signaling pathways and well-designed and focused follow-up studies are critical to confirm new hypothesis that global gene expression approaches might generate.

      Ultimately, the success of identifying biomarkers for vascular injury and understanding mechanisms of vascular injury may require application of multiple technologies, including genomics, proteomics, metabonomics, flow cytometry, imaging, etc. This creates the additional challenge of combining disparate data from these various approaches and integrating them to allow cross-platform querying and extraction of biological knowledge to gain a more holistic understanding of vascular injury.

    2. Proteomics
      1. Background
      2. Proteomics technologies are broadly characterized as tools designed to examine the expression of proteins in a given set of samples. Types of information derived from proteomics technologies include protein identity, quantity, interactions, structure, and post-translational modification. The nature of the proteome presents a series of challenges that limit the ability of any single technology to completely assay and characterize it. First and foremost is the heterogeneity of proteins. Proteins can be large or small, globular or compact, hydrophobic or hydrophilic, basic or acidic. Post-translational modifications such as glycosylation, lipidation, or phosphorylation can have dramatic effects on how the proteins behave in biochemical assays. A second challenge of the proteome is the extraordinary wide dynamic range of protein expression, which is estimated to be one billion fold . This presents an extraordinary challenge to any detection method. A third challenge of the proteome is the absence of a method of direct amplification. While the polymerase chain reaction can amplify DNA and RNA, no such technique exists for proteins (although there are promising approaches in development). These challenges together create a large opportunity for technology development, as well as niches for the myriad of technologies that have already been developed. For the researcher studying vascular injury, these technologies can provide powerful tools to discover biomarkers or to understand fundamental mechanisms. To date there have been no published reports detailing proteomic–derived biomarkers for vascular injury, however a number of laboratories are currently pursuing the approach.

      3. Platforms (Strengths/Weaknesses)

      This document cannot, of course, list all proteomic technologies, which are constantly being developed and refined. Instead, it will focus on several well-established techniques that are broadly aimed at expression profiling (although some have additional applications such as studying post-translational modification). Consequently, a discussion of high-throughput crystallography, as an example, will not be presented here. Several themes will develop. First, all proteomics technologies consist of some permutation of separation followed by detection. Second, no single proteomics technology in its current state can possibly hope to characterize the entire proteome. It is therefore the recommendation that each researcher apply multiple approaches to the study of the proteome.

      Classically, two-dimensional gel electrophoresis (2-D PAGE) has been used to assay the proteome . Protein samples are first separated on the basis of pI using an isoelectric focusing gradient. These proteins are then further separated in a second dimension based on their migration in an electric field, which is generally a function of their hydrodynamic volume, which roughly correlates to their molecular weight. Detection can then be performed using any of a myriad of agents, including silver staining and fluorescent dyes. Individual spots on the gels can be excised, digested with trypsin, and tryptic fragment fingerprint matched against a database. Where indicated, tandem mass spectrometry can be performed and matched against the theoretical tandem mass spectra of peptide sequences to provide (in most situations) definitive identification. Its drawbacks include poor visualization of low abundance proteins, generally poor visualization of extremely basic or acidic proteins, poor visualization of hydrophobic proteins, and low throughput . Matching of 2D gels from different samples is time-consuming, although computer algorithms that aid this process are improving. Thus, this technique is useful for the researcher who wishes to perform differential profiling and has abundant sample volume but limited numbers of samples. It is not useful for the researcher with limited sample volume or large numbers of samples. This may limit the statistical validity of some of the conclusions derived from 2D gel studies. Recent advances in software used to align 2D gels, integrated platforms, and independent 2D gel service providers can overcome many of these limitations.

      Surface-enhanced laser desorption ionization and time-of-flight mass spectrometry (SELDI-TOF MS) is a proteomics technique with a broad range of applications, including protein expression profiling . Separation is performed using a combination of traditional elution chromatography (usually on spin columns) followed by retentate chromatography. Retentate chromatography is performed by incubating samples (or fractions of samples) on arrays containing chromatographic substrate on its surface (ref). Traditional chromatographic steps of binding and washing are performed, but what is detected by laser desorption TOF MS are the proteins that have been retained on the arrays. This technique can be used for differential profiling, particularly when samples are scarce (only 20 ul of serum is required) as well as when a relatively large number of samples need to be profiled. In addition to differential profiling, this technology can be used to study protein-protein interactions, post-translational modifications, and epitope mapping. Therefore, it may be quite useful in studying mechanisms of vascular injury in addition to identifying novel biomarkers. Its principle drawback is that biomarker identification of the biomarkers generally requires some form of off-line purification, although identification itself is possible using the array.

      A relatively novel technique, ICAT ™ for differential expression profiling requiring labeling of samples has recently been described and commercialized. In this method, the proteins in two samples to be compared are labeled separately on the side chains of their reduced cysteinyl residues using one of two isotopically different, but chemically identical sulfhydryl-reactive ICAT reagents (one being an isotopically "light" reagent, d(0), the other being a "heavy" reagent containing eight deuterium atoms on its carbon backbone, d(8)). The labeled protein mixtures are combined and enzymatically digested (e.g. with trypsin), and the labeled peptides are isolated by affinity chromatography, using the affinity tag (biotin group) that is part of the ICAT reagents. The selected peptides are separated by liquid chromatography and analyzed by tandem mass spectrometry . This technique assumes quantitative and equivalent labeling and recovery of analytes and labels only proteins containing cysteines. In addition, it is not well suited to the study of proteins that have been modified post-translationally. It may be a good technique for differential profiling of a limited number of samples, but is not high-throughput.

      Multidimensional liquid chromatography followed by tandem mass spectrometry is another technique for identification proteins contained in a complex sample . Samples for this procedure are generally globally digested, and then liquid chromatography using orthogonal separations principles is performed serially, and the individual peptides are then sequenced by tandem mass spectrometry. While useful in providing information of proteins contained within a sample, quantitation information is minimal and the procedure is notoriously low-throughput.

      Finally, a brief discussion of protein arrays in their many forms is warranted . There has been much interest in developing arrays on which hundreds to thousands of proteins or antibodies are immobilized, and then incubated with a sample and binding information for each of these analytes obtained simultaneously. This has, however, been difficult to achieve in practice, in large part because identifying reagents (i.e. purification of the thousands of proteins or antibodies) is difficult, immobilization strategies that retain protein function are difficult to achieve, and various protein-protein interaction reactions often have differing binding requirements (e.g. pH, salt, divalent cations, etc). Specific types of protein arrays include tissue arrays and multi-analyte western blots (refs). In the future, it can be expected that many types of arrays will be developed, most with specific types of applications rather than global protein expression profiling.

      These descriptions are not intended to be comprehensive, but rather insight into the major proteomic tools available to the researcher studying vascular injury. The vascular injury researcher should be aware of these diverse technology offerings. Some of these techniques are readily transferable to the individual researcher’s laboratory; others are best done in collaboration with experts versed in the use of the techniques. In addition, the vascular injury researcher will find many opportunities to participate in the improvement of existing techniques and the development of novel ones. Beyond the applications of protein expression profiling detailed here, the study of protein-protein interactions to delineate protein-interaction maps can be quite useful in studying the mechanisms underlying vascular injury. In addition, the study of post-translational modifications is a core application of many of these proteomics technologies and can therefore be useful in studying how different types of post-translational modification determine the function of the proteins involved in the mechanisms of vascular injury.

    3. Metabonomics
      1. Background and Principles
      2. Toxicants, by definition, disrupt the normal composition and flux of endogenous biochemicals in, or through, key intermediary cellular metabolic pathways. These disruptions alter, either directly or indirectly, the blood that percolates through the target tissues. This altered blood can then produce urine changes, either directly or indirectly, producing characteristic biomolecular traces. The diagnostic utility of any one trace biomolecule is limited due to the number of variables affecting it along its route to the urine and by the commonality of biochemical processes disrupted by toxicants. However, if a significant number of trace molecules are monitored, the overall pattern, or " fingerprint" produced may be more consistent and predictive than any one marker. Metabonomics combines the techniques of high-resolution nuclear magnetic resonance (NMR) and pattern recognition technology to rapidly generate these diagnostic metabolic fingerprints .

      3. Platforms
      4. Although, in principle, many analytical techniques could be used for metabonomic evaluations, high field NMR provides the distinct advantage of being able to simultaneously detect thousands of molecules in a bodily fluid with little sample preparation. This ability, allows an unbiased assessment of metabolic response . Proton (1H) NMR spectroscopy is capable of detecting individual atoms in soluble proton-containing molecules with a molecular weight of approximately 20 kD or less. Standard flow probes used with 600 MHz spectrometer are capable of detecting thousands of resonances in the urine from molecules at concentrations in the mid micromolar range or higher with a data acquisition time of a few minutes. Incorporation of cryoprobe technology will lower this limit significantly. The other half of the technology, pattern recognition, is considerably more diverse in the availability of platforms . However, principal component analysis (PCA) has been the most frequently used technique for small scale focused metabonomic studies . An exciting area of ongoing research is the linking of metabonomics technology with proteomic and toxicogenomic platforms allowing potential biomarker assessment from gene to protein to phenotype.

      5. Application of Technology to Drug-Induced Vasculopathy
      6. The application of metabonomic technology to drug-induced vasculopathy is a relatively recent development. To date most work has focused on PDE4- induced arteriopathy in the rat. Principal component analysis produced a clear urine spectral pattern separation between 8 of 11 rats with vascular lesions and 36 of 37 rats without lesions in samples collected after 3 or 4 days of treatment with CI-1018, a PDE4 inhibitor . Furthermore, in rats immunosuppressed with dexamethasone, the urine spectral pattern was still evident, suggesting the spectral changes were not simply an indirect index of inflammation . The specificity of the spectral changes produced in rats with vascular lesions was confirmed by Zhang et al , who demonstrated that spectral changes produced in response to renal lesions in SHR rats could be differentiated from spectral changes associated with vascular lesions produced by the same compound in WKY rats. Work with rolipram, demonstrated that the spectral changes precede vascular lesions in both time and dose .

      7. Advantages/Disadvantages
      8. Using urine as a sample matrix, a distinct advantage of the technology is that comprehensive toxicological information can be gained from a single animal including onset, peak and regression of toxicity without concerns over sample numbers or timing. Results can be readily correlated with concurrent clinical and clinical pathology assessments if necessary. Since sample collection is non-invasive, analyses are easily piggybacked onto other studies with minimal impact. Sample size is usually not an issue in rats as they produce 10 –20 mL of urine/day. Sample preparation is minimal requiring only addition of antibacterial agent, dilution in deuterated buffer and centrifugation to eliminate particulates - all of which are amenable to robotic processing in a microtiter plate format. The analyses themselves can be conducted relatively rapidly (200-300/day) and have a relatively low sample cost/analysis. Since NMR is non-selective, once a sample is analyzed, the spectra can be queried as needed to obtain different biological information without the necessity of rerunning the analyses. The primary disadvantages of the technology are the high initial capital costs of obtaining an instrument and the necessity for collecting samples in metabolism cages with cold collection capability as bacterial contamination of samples essentially eliminates the utility of the sample. Other disadvantages with the technology are not appreciably different than those associated with proteomic and genomic approaches. Since NMR is non-selective, any change in animal physiology or toxicity (other than the one of interest) may affect the urine NMR spectra, making it difficult to associate a pattern specifically with the lesion of interest. Additionally, it is often difficult to mechanistically associate bimolecular changes in urine to observed lesions. NMR is not particularly sensitive (relative to something like mass spec) but it is not subject to interferences and the utilization of longer acquisition times or the use of cryoprobe technology can reduce the sensitivity limitation. Given the wealth of data that can be generated and the uncertainty as to what it all means, bioinformatics limitations and information overload plague this technology as they do the other "omics" approaches.

      9. Knowledge Gaps:

      Metabonomics is a newer technology than proteomic and toxicogenomic approaches (at least to the toxicologist anyway). With that newness comes a certain amount of "black box" fear for those not familiar with NMR. However NMR, as a platform, is much older and more stable than platforms for other technologies, which should help reduce those fears accordingly. The multivariate statistical methods are much less user friendly and represent a real knowledge gap between the users and those assessing the significance of the generated data. Within the vascular injury application, the primary knowledge gap is the lack of a clear mechanistic relevance of the biomolecular changes driving the pattern separation. It seems clear that the urinary changes are not simply a reflection of inflammation, but it is difficult to understand how minimal to moderate pathologic changes in one vascular bed (e.g. the mesentery) could drive micro to millimolar urinary changes in molecules involved in intermediate metabolism. It seems likely that other factors may be involved, some of which, when elucidated, may lead us to a better understanding of the etiology of this troublesome lesion in rats which may lead to a clearer evaluation of the human clinical significance of these changes.

    4. Flow Cytometry
      1. Background and Principles
      2. There are numerous circulating cells present in the blood stream of both clinical and pre-clinical species. These cells are exposed to the pharmacologic agent of interest, and the possibility exists that during vascular toxicity, cells within the vessel wall may dislodge and enter the circulation. Information, both quantitative and functional, can be obtained by analyzing these circulating or dislodged cells.. There is considerable heterogeneity between arterial, venous, and capillary vascular endothelium, and even organ specific endothelial phenotypic alterations have been identified . This may allow discrimination of site-specific vascular injury, once appropriate markers have been identified. Flow cytometry has wide application in the study of circulating cells, which can be collected by minimally invasive means. It permits rapid simultaneous acquisition of multiple parameters, which are valuable for analyzing diverse functional or quantitative changes in individual cells . With the advent of eight- or nine-color sorters and/or analyzers permitting multicolor fluorescence analyses and the continuing development and commercialization of a wide range of fluorescent probes have made flow cytometry a unique technology to meet these needs. These capabilities coupled with sequential gating strategies permit analyses of extremely large numbers of cells leading to subsequent identification of even rare cells .

      3. Platforms
      4. Flow cytometry allows for rapid measurement of numerous characteristics on individual cells/particles moving single file in a fluid stream. Forward angle light scatter can distinguish cell/particle size whereas side angle light scatter can determine internal complexity. Furthermore, light emitted from fluorescently labeled probes or antibodies allows identification of a wide variety of cell surface, cytoplasmic and nuclear structures. Light scatter and fluorescent emission are then detected by photodetectors. Flow cytometers use two basic platforms, bench top analyzers and sorters. These are available from a variety of manufacturers. Both platforms can be used for high speed laser cell interrogation and analysis The biggest difference is that a sorter can be used to select cells of interest into a receptacle for further analysis or culture. Typically sorters are much more accommodating in design of lasers and optics compared to bench top analyzers, providing more flexibility in study design, and numbers of parameters that can be examined simultaneously.

      5. Application of Technology to Drug-Induced Vasculopathy
      6. The application of flow cytometry to the study of drug induced vascular injury has only begun recently. Flow cytometric analysis in relation to drug-induced vascular injury would be applicable in several areas 1) alterations of normally circulating cells such as white blood cells, 2) cells released from the vascular wall into circulation after vascular damage, 3) assessment of oxidative stress in circulating cells. Single cell suspensions of circulating leukocytes can be monitored as sentinels of drug-induced effects as it lends itself amenable to rapid analyses of thousands of cells . There are numerous commercially available antibodies to cell surface proteins on leukocytes in most species, allowing for immunophenotyping and assessment of activation markers. It is currently unknown how drug-induced vascular damage impacts these measurements. There are small numbers of circulating endothelial cells present in normal individuals, and the numbers have been shown to increase after various insults resulting in vascular damage . To date, most of these studies have utilized antibody linked magnetic beads to concentrate circulating endothelial cells for enumeration, and the use of flow cytometric quantitation is untested. There is in vivo evidence for oxidative stress in vascular endothelium and leukocytes of spontaneously hypertensive rats as compared to normotensive rats . Hypertension is well known to result in vascular damage and repair. This assay relies on the use of hydroethidine that is a reduced non-fluorescent precursor of the fluorescent molecule ethidium bromide. Hydroethidine, in the presence of superoxide radical, converts to ethidium bromide that can bind DNA and then fluoresces brightly. This assay could be applied to models of drug-induced vascular damage to assess either leukocytes or circulating endothelial cells.

      7. Advantages/Disadvantages
      8. Advantages. The equipment is widely available in both clinical and preclinical laboratory settings. There is an ever-increasing number of commercially available antibodies and fluorochromes applicable to flow cytometry. Additionally, antibodies can be made to most antigens providing increased utility to flow cytometric analyses in toxicology. Flow cytometric examination of cells can be done at rates of tens of thousands of cells per second, and can frequently be done in an automated fashion.

        Disadvantages. The capital costs for equipment, especially for sorters, are quite high. Qualified personnel are needed to run and interpret multi-parameter data. Antibodies and fluorochromes are key to fully utilizing the potential of flow cytometry in assessment of vascular damage. Despite the availability of antibodies for endothelial cells, because of the heterogeneity of endothelial cells, truly pan-endothelial antigens likely do not exist. This could also be an advantage once the localization of antibodies is mapped to specific vascular endothelium, potentially allowing identification of vascular damage location. Making new antibodies to current antigens without antibodies or previously unidentified antigens takes both time and money. Many of the currently available antibodies do not cross-react across multiple species.

      9. Knowledge Gaps

      It is currently unknown whether localized vascular damage, typical of the drug-induced vasculopathies, alters either normally circulating cells or dislodges vascular wall cellular components in a detectable fashion. Likewise the duration of increased or altered circulating cells after vascular insult is largely unknown, although it appears to return to normal as the symptoms subside . There is marked disagreement, in the literature, on the number of circulating endothelial cells present in normal individuals; these discrepancies may be related to differences in methods and/or antibodies utilized. It will be important to provide details regarding methodology for quantitation of cells when reporting on models of vascular damage. There is currently a need to more fully elucidate the phenotypic heterogeneity of the vascular endothelium in a localized manner, specifically in vascular beds where drug-induced vascular damage is common: coronary vascular bed in dogs, splanchnic vascular bed in rats. Localization of damage is currently only detected by histologic examination. Reagents allowing this localization should be pursued by techniques such as injection of phage-display peptide libraries that detect specific surface molecules in the vascular endothelium , or pursuing novel methods of culturing endothelial cells from specific vasculature .

    5. Bioinformatics and Multivariate Statistics
      1. Background
      2. "Omics" dominates the list of new techniques available to the researcher studying almost any biological system. Rapid developments in technology platforms coupled with high throughput techniques allow researchers to measure many analytes from many samples simultaneously, generating a staggering amount of data. This leads to the challenge of mining that data, archiving that data, and retrieving that data. The fields of bioinformatics and multivariate statistics, coupled with increasing computational power, present to the researcher solutions to their data-mining problems as well as opportunities for research and development of novel solutions to these problems. Because of this, computational biology should itself be thought of as a novel technology.

      3. Platforms
      4. The term bioinformatics in this discussion is reserved strictly for two tasks: first, to describe methods that mine primary and secondary sequences of nucleic acids and proteins as well as structural data for proteins to derive information regarding genes or proteins; and second, to archive and retrieve such data in a computationally efficient manner. Thus, bioinformatics encompasses strategies used to perform sequence alignment, to predict coding regions from genomic DNA, predict secondary structure of proteins, and so forth. It also encompasses databasing such information, such as is performed at the National Center for Biotechnology Information (NCBI) or SWISS-PROT. Another type of database is one of theoretical enzymatic digests of proteins contained in a protein database, which is searched with an actual enzymatic digest of an unknown sample to obtain the identity of the protein. An outstanding review text is written by Baxevanis . Vascular injury researchers have at their disposal a number of tools, many of which are continually being improved. For example, better prediction of coding regions in genomic data leads to a more complete catalog of genes. This in turn leads to a more complete database of enzymatic digests for protein identification. A more complete understanding of protein function and structure allows researchers to catalog protein domains and sites of specific types of post-translational modification, thus allowing a researcher to generate and test hypotheses regarding protein function more quickly. However, much remains to be done. For example, databases of theoretical translations do not account for real or theoretical post-translational modifications. This can lead to an incorrect protein identification or, more likely, the mistaken conclusion that one has found a heretofore unidentified protein. Algorithms that more accurately predict gene structure also remain to be developed. Bioinformatics can also be used to understand species differences in vascular injury (i.e. comparative genomics). For example, sequence alignment can reveal conserved domains that might dictate similar functions (and therefore similar responses to drugs) as well as non-conserved regions that might help predict divergent functions.

        This discussion will not include the problem of visualization of genomics or proteomics data. A good review can be found in Chakravarti . Perhaps of more interest to the researcher employing these "omics" technologies is statistical analysis of data generated by these techniques. For example, if one were to use a 10k cDNA microarray on 20 samples, there would be 200,000 data points. SELDI technology can easily generate data for 140,000 protein peaks for forty samples. Particularly in the most common employment of these "omics" technologies, there is a profound asymmetry between sample number (few) versus data points per sample (many). This presents a challenge for statisticians, in which the converse situation (many data points relative to sample number) is preferred. Generally, the goal is to find the set of variables (genes or proteins) that distinguish classes (such as with vascular injury versus without vascular injury) of samples from each other. Univariate analysis on a variable-by-variable basis is simple to perform, but in most complex biological conditions, it is unusual to find a single variable that is both sensitive and specific for a given condition. For example, a number of acute phase reactant proteins are increased in animal models of vascular injury, but these are generally not considered to be specific to vascular injury (?reference?). Using multiple variables simultaneously is a common method to overcome the limitation of single markers. Multivariable analysis is therefore an important component of any study aimed at identifying biomarkers of vascular injury. Multivariate analysis can be conveniently divided into two broad classes, supervised and unsupervised. Supervised learning requires that during the development of a classification model, information regarding outcome (class assignment) is available and is useful in developing a diagnostic. Unsupervised learning does not have such a requirement and is more powerful for data exploration. These classes of algorithms will be discussed in turn.

        Supervised learning techniques are generally described as classification algorithms, in which data for each of the samples is supplied to the algorithm along with the class assignment for each sample (classes are the various groupings of samples such as drug-treated versus untreated). The algorithm uses the class assignment to develop models that can be used to classify these samples (called the training set). In the ideal case, these models are then tested on different samples for which the identity is hidden (called the testing set). In general, the performance of the model on the test set is expected to be lower than that on the training set. The quality of the model is a function of the overall classification success rate as well as the decline in classification success rate from the training set to the test set of samples. It is not uncommon to develop a model that can classify the training set with high success but that does poorly on a test set – these models are generally described as being overfit. Examples of supervised learning techniques include support vector machines, neural nets, and decision trees. The vascular injury researcher is recommended to utilize algorithms that provide the underlying variables that contribute to the classification model, since these features can be concluded to be biomarkers for future workup.

        Unsupervised learning techniques are generally described as clustering algorithms. No class assignments are presented to the algorithm, which attempts to distribute the samples as far apart as possible in n-dimensional space. Samples that are more similar to each other will be closer to each other in these models, and the distance between samples is therefore a measure of dissimilarity. Common unsupervised learning techniques include self-organizing maps, hierarchal and k-means clustering, and principal component analysis. These types of techniques are particularly useful in finding hidden classes. Since no a priori assumptions are made about the categorization of samples, the distribution of the samples may reveal heretofore-unknown subgroups. A drawback of unsupervised learning techniques is that deconvolution of these models may not reveal a defined, small subset of biomarkers. Instead, one often gets a long list of co-regulated genes or proteins. This can be particularly problematic for discovery of biomarkers that are to be used as a framework for mechanistic studies of vascular injury.

        Statisticians are constantly developing refinements of these techniques as well as novel approaches. The challenge for the vascular injury researcher not versed in statistics will be to utilize the technique(s) that are most appropriate for the type of data being mined. Vascular injury researchers should seek out collaboration with statisticians who are versed in the biology of genes and proteins, so that intelligent decisions regarding algorithm selection and parameters are made. Moreover, it should be recognized that proper experimental design is the first step toward optimal statistical analysis; poor experimental design can lead to the inability to make statistically sound conclusions or, worse, incorrect conclusions from the data. Finally, all data that are supplied to these types of algorithms must be intelligently pre-processed (e.g. calibrated and normalized).

      5. Conclusions

    The technologies described in this section can be sources of biomarkers as well as provide insight into the mechanisms of vascular injury. The combination of these technologies will provide a synergistic effect, both from the standpoint of validating results across the various platforms, as well as providing the opportunity for "meta" statistical analysis i.e. the integration of information garnered from multiple techniques. It is recognized that the adoption of any one of these technologies is itself a resource heavy proposition, and to undertake more than one simultaneously creates even greater pressure on potentially limited resources. Therefore, it would be prudent for researchers across institutions to allocate resources and become centers of excellence in a specific subset of these technologies. The data can then be shared, archived, and mined. This will expedite the discovery and validation of biomarkers and enable a greater understanding of the mechanisms of vascular injury.

  2. References

1. Aebersold, R., and Goodlett, D. Mass spectrometry in proteomics. Chem. Rev 101:269-295; 2001.

2. Albassam, M. A., Metz, A. L., Gragtmans, N. J., King, L. M., Macallum, G. E., Hallak, H., and McGuire, E. J. Coronary arteriopathy in monkeys following administration of CI-1020, an endothelin A receptor antagonist. Toxicol Pathol 27:156-64.; 1999.

3. Albassam, M. A., Metz, A. L., Potoczak, R. E., Gallagher, K. P., Haleen, S., Hallak, H., and McGuire, E. J. Studies on coronary arteriopathy in dogs following administration of CI- 1020, an endothelin A receptor antagonist. Toxicol Pathol 29:277-84.; 2001.

4. Albassam, M. A., Smith, G. S., and Macallum, G. E. Arteriopathy induced by an adenosine agonist-antihypertensive in monkeys. Toxicol Pathol 26:375-80.; 1998.

5. Alcorta, D., Preston, G., Munger, W., Sullivan, P., Yang, J., Waga, I., Jennette, J., and Falk, R. Microarray studies of gene expression in circulating leukocytes in kidney diseases. Exp Nephrol 10(2):139-149; 2002.

6. Arnal, J. F., Dinh-Xuan, A. T., Pueyo, M., Darblade, B., and Rami, J. Endothelium-derived nitric oxide and vascular physiology and pathology. Cell Mol Life Sci 55:1078-87.; 1999.

7. Bacchus, A. N., Ely, S. W., Knabb, R. M., Rubio, R., and Berne, R. M. Adenosine and coronary blood flow in conscious dogs during normal physiological stimuli. Am J Physiol 243:H628-33.; 1982.

8. Baxevanis, A., and Ouellette, B. Bioinformatics: a practical guide to the analysis of genes and proteins: Wiley-Liss, New York; 2001.

9. Berne, R. M. The role of adenosine in the regulation of coronary blood flow. Circ Res 47:807-13.; 1980.

10. Berne, R. M., Knabb, R. M., Ely, S. W., and Rubio, R. Adenosine in the local regulation of blood flow: a brief overview. Fed Proc 42:3136-42.; 1983.

11. Blann, A. Endothelial cell activation, injury, damage and dysfunction: Separate entities or mutual terms? Blood Coagul Fibrinolysis 11 (7):623-630; 2000.

12. Boesiger, J., Tsai, M., Maurer, M., Yamaguchi, M., Brown, L. F., Claffey, K. P., Dvorak, H. F., and Galli, S. J. Mast cells can secrete vascular permeability factor/ vascular endothelial cell growth factor and exhibit enhanced release after immunoglobulin E-dependent upregulation of fc epsilon receptor I expression. J Exp Med 188:1135-45.; 1998.

13. Boor, P. J., Gotlieb, A. I., Joseph, E. C., Kerns, W. D., Roth, R. A., and Tomaszewski, K. E. Chemical-induced vasculature injury. Summary of the symposium presented at the 32nd annual meeting of the Society of Toxicology, New Orleans, Louisiana, March 1993. Toxicol Appl Pharmacol 132:177-95.; 1995.

14. Calabrese, L. H., and Duna, G. F. Drug-induced vasculitis. Curr Opin Rheumatol 8:34-40.; 1996.

15. Cambridge, G., Wallace, H., Bernstein, R. M., and Leaker, B. Autoantibodies to myeloperoxidase in idiopathic and drug-induced systemic lupus erythematosus and vasculitis. Br J Rheumatol 33:109-14.; 1994.

16. Castle, A., Carver, M., and Mendrick, D. Toxicogenomics: a new revolution in drug safety. Drug Discov Today 7(13):728-736; 2002.

17. Chakravarti, D., et al. Informatic tools for proteome profiling. Biotechniques suppl. March:4-15; 2002.

18. Chelly, J. E., Doursout, M. F., Begaud, B., Tsao, C. C., and Hartley, C. J. Effects of hydralazine on regional blood flow in conscious dogs. J Pharmacol Exp Ther 238:665-9.; 1986.

19. Cines, D. B., Pollak, E. S., Buck, C. A., Loscalzo, J., Zimmerman, G. A., McEver, R. P., Pober, J. S., Wick, T. M., Konkle, B. A., Schwartz, B. S., Barnathan, E. S., McCrae, K. R., Hug, B. A., Schmidt, A. M., and Stern, D. M. Endothelial cells in physiology and in the pathophysiology of vascular disorders. Blood 91:3527-61.; 1998.

20. Clozel, M., Gray, G. A., Breu, V., Loffler, B. M., and Osterwalder, R. The endothelin ETB receptor mediates both vasodilation and vasoconstriction in vivo. Biochem Biophys Res Commun 186:867-73.; 1992.

21. Corthals, G., et al. The dynamic range of protein expression: a challenge for proteomic research. Electrophoresis 21:1104-1115; 2000.

22. Detweiler, D. K. Spontaneous and induced arterial disease in the dog: pathology and pathogenesis. Toxicol Pathol 17:94-108.; 1989.

23. Dignat-Gorge, F., and Sampol, J. Circulating endothelial cells in vascular disorders: new insights into an old concept. Eur J Haematol 65:215-220; 2000.

24. Dogterom, P., Zbinden, G., and Reznik, G. K. Cardiotoxicity of vasodilators and positive inotropic/vasodilating drugs in dogs: an overview. Crit Rev Toxicol 22:203-41; 1992.

25. Dong, Q., Bernasconi, S., Lostaglio, S., De Calmanovici, R., Martin-Padura, I., Breviario, F., Garlanda, C., Ramponi, S., Mantovani, A., and Vecchi, A. A general strategy for isolation of endothelial cells from murine tissues. Characterization of two endothelial cell lines from the murine lung and subcutaneous sponge implants. Arteriosclerosis Thrombosis & Vascular Biology 17(8):1599-1604; 1997.

26. Drab, M., Verkade, P., Elger, M., Kasper, M., Lohn, M., Lauterbach, B., Menne, J., Lindschau, C., Mende, F., Luft, F. C., Schedl, A., Haller, H., and Kurzchalia, T. V. Loss of caveolae, vascular dysfunction, and pulmonary defects in caveolin-1 gene-disrupted mice. Science 293:2449-52.; 2001.

27. El-Deredy, W. Pattern Recognition Approaches in Biomedical and Clinical NMR Spectroscopy: A Review. NMR in Biomedicine 10:99-124; 1997.

28. Fung, E., and Enderwick, C. ProteinChip® clinical proteomics: computational challenges and solutions. Biotechniques suppl. March:34-41; 2002.

29. Fung, E., et al. Protein biochips for differential profiling. Curr. Opin. Biotech 12:65-69; 2001.

30. Gerrity, R. G., Richardson, M., Caplan, B. A., Cade, J. F., Hirsh, J., and Schwartz, C. J. Endotoxin-induced vascular endothelial injury and repair. II. Focal injury, en face morphology, (3H)thymidine uptake and circulating endothelial cells in the dog. Exp Mol Pathol 24:59-69.; 1976.

31. Glenny, R. W., Bernard, S., and Brinkley, M. Validation of fluorescent-labeled microspheres for measurement of regional organ perfusion. J Appl Physiol 74:2585-97.; 1993.

32. Gossett, K., Narayanan, P., Williams, D., Gore, E., Herzyk, D., Hart, T., and Sellers, T. Flow cytometry in the preclinical development of biopharmaceuticals. Toxicol Pathol 27(1):32-37; 1999.

33. Grefte, A., van der Giessen, M., van Son, W., and The, T. Circulating cytomegalovirus (cmv)-infected endothelial cells in patients with active cmv infection. J Infect Dis 167:270-277; 1993.

34. Griffin, T., and Aebersold, R. Advances in proteome analysis by mass spectrometry. J Biol Chem 276:45497-45500; 2001.

35. Gristwood, R. W., Comer, M. B., Eden, R. J., Taylor, E. M., Turner, J. A., Wallduck, M., and Owen, D. A. In vivo pharmacological studies with SK&F 94836, a potent inotrope/vasodilator with a sustained duration of action. Br J Pharmacol 93:893-901.; 1988.

36. Gygi, S., et al Evaluation of two-dimensional gel electrophoresis-based proteome analysis technology. Proc. Natl. Acad. Sci 97:9390-9395; 2000.

37. Gygi, S., et al. Quantitative analysis of complex protein mixtures using isotope coded affinity tags. Nat. Biotech 17::994-999; 1999.

38. Hartman, H. A. Idiopathic extramural coronary arteritis in beagle and mongrel dogs. Vet Pathol 24:537-44.; 1987.

39. Haynes, B. F., Telen, M. J., Hale, L. P., and Denning, S. M. CD44--a molecule involved in leukocyte adherence and T-cell activation. Immunol Today 10:423-8.; 1989.

40. Henriksen, P., and Kotelevtsev, Y. Application of gene expression profiling to cardiovascular disease. Cardiovasc Res Apr;54(1):16-24; 2002.

41. Holmes, E., and Shockcor, J. Accelerated Toxicity Screening Using NMR and Pattern Recognition-Based Methods. Current Opinion in Drug Discovery and Development 3:72-78; 2000.

42. Humphrey, S. J., and Zins, G. R. Whole body and regional hemodynamic effects of minoxidil in the conscious dog. J Cardiovasc Pharmacol 6:979-88.; 1984.

43. Igarashi, H., Hatake, K., Shiraishi, H., Samada, K., Tomizuka, H., and Momoi, M. Y. Elevated serum levels of macrophage colony-stimulating factor in patients with Kawasaki disease complicated by cardiac lesions. Clin Exp Rheumatol 19:751-6.; 2001.

44. Janin, A., Deschaumes, C., Daneshpouy, M., Estaquier, J., Micic-Polianski, J., Rajagopalan-Levasseur, P., Akarid, K., Mounier, N., Gluckman, E., Socie, G., and Ameisen, J. C. CD95 engagement induces disseminated endothelial cell apoptosis in vivo: immunopathologic implications. Blood 99:2940-7.; 2002.

45. Joseph, E., Mesfin, G., and Kerns, W. Pathogenesis of arterial lesions caused by vasoactive compounds in laboratory animals. In: W. D. Kerns, Bishop, S.P (ed.), Cardiovascular Toxicology, Comprehensive Toxicology Series Vol 6, pp. 279-307: Elsevier Press, UK; 1997.

46. Joseph, E. C. Arterial lesions induced by phosphodiesterase III (PDE III) inhibitors and DA(1) agonists. Toxicol Lett 112-113:537-46.; 2000.

47. Joseph, E. C., Jones, H. B., and Kerns, W. D. Characterization of coronary arterial lesions in the dog following administration of SK&F 95654, a phosphodiesterase III inhibitor. Toxicol Pathol 24:429-35.; 1996.

48. Kerns, W. D., Arena, E., Macia, R. A., Bugelski, P. J., Matthews, W. D., and Morgan, D. G. Pathogenesis of arterial lesions induced by dopaminergic compounds in the rat. Toxicol Pathol 17:203-13.; 1989.

49. Kerns, W. D., Arena, E., and Morgan, D. G. Role of dopaminergic and adrenergic receptors in the pathogenesis of arterial lesions induced by fenoldopam mesylate and dopamine in the rat. Am J Pathol 135:339-49.; 1989.

50. Kerns, W. D., Roth, L., and Hosokawa, S. Spontaneous Idiopathic Polyarteritis. In: D. Dungworth, Mohr, U (ed.), Pathobiology of the Aging Dog Vol. 2, Chapter 15, pp. 118-126; 2001.

51. Kiely, P. D., Pecht, I., and Oliveira, D. B. Mercuric chloride-induced vasculitis in the Brown Norway rat: alpha beta T cell-dependent and -independent phases: role of the mast cell. J Immunol 159:5100-6.; 1997.

52. Leung, D. Y. The potential role of cytokine-mediated vascular endothelial activation in the pathogenesis of Kawasaki disease. Acta Paediatr Jpn 33:739-44.; 1991.

53. Liang, C. S., Thomas, A., Imai, N., Stone, C. K., Kawashima, S., and Hood, W. B., Jr. Effects of milrinone on systemic hemodynamics and regional circulations in dogs with congestive heart failure: comparison with dobutamine. J Cardiovasc Pharmacol 10:509-16.; 1987.

54. Lindon, J., et al Metabonomics: Metabolic processes studied by NMR spectroscopy of biofluids. Concepts in Magnetic Research 12:289-320; 2000.

55. Lindon, J., et al. NMR Spectroscopy of Biofluids. Annual Reports on NMR Spectroscopy 38:1-88; 1999.

56. Lotti, T., Ghersetich, I., Comacchi, C., and Jorizzo, J. L. Cutaneous small-vessel vasculitis. J Am Acad Dermatol 39:667-87; quiz 688-90.; 1998.

57. Louden, C., et. al. Coronary arterial lesions in dogs treated with an endothelin receptor antagonist. J. Cardiovasc. Pharmacol 31(Suppl. 1):S384-S385; 1998.

58. Louden, C., and Morgan, D. G. 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 Toxicol 89:158-70.; 2001.

59. Louden, C. S., Nambi, P., Pullen, M. A., Thomas, R. A., Tierney, L. A., Solleveld, H. A., and Schwartz, L. W. Endothelin receptor subtype distribution predisposes coronary arteries to damage. Am J Pathol 157:123-34.; 2000.

60. Lum, H., and Roebuck, K. A. Oxidant stress and endothelial cell dysfunction. Am J Physiol Cell Physiol 280:C719-41.; 2001.

61. Macallum, G. E., Walker, R. M., Barsoum, N. J., and Smith, G. S. Preclinical toxicity studies of an adenosine agonist, N-(2,2- diphenylethyl) adenosine. Toxicology 68:21-35; 1991.

62. MacPhee, M., Wiltrout, R., McCormick, K., Sayers, T., and Pilaro, A. A method for obtaining and culturing large numbers of purified organ-derived murine endothelial cells. J Leukoc Biol 55:467-475; 1994.

63. Majewski, S., Hunzelmann, N., Johnson, J. P., Jung, C., Mauch, C., Ziegler-Heitbrock, H. W., Riethmuller, G., and Krieg, T. Expression of intercellular adhesion molecule-1 (ICAM-1) in the skin of patients with systemic scleroderma. J Invest Dermatol 97:667-71.; 1991.

64. Meng, H., Tonnesen, M. G., Marchese, M. J., Clark, R. A., Bahou, W. F., and Gruber, B. L. Mast cells are potent regulators of endothelial cell adhesion molecule ICAM-1 and VCAM-1 expression. J Cell Physiol 165:40-53.; 1995.

65. Merkel, P. A. Drug-induced vasculitis. Rheum Dis Clin North Am 27:849-62.; 2001.

66. Mesfin, G. M., Piper, R. C., DuCharme, D. W., Carlson, R. G., Humphrey, S. J., and Zins, G. R. Pathogenesis of cardiovascular alterations in dogs treated with minoxidil. Toxicol Pathol 17:164-81.; 1989.

67. Mesfin, G. M., Robinson, F. G., Higgins, M. J., Zhong, W. Z., and DuCharme, D. W. The pharmacologic basis of the cardiovascular toxicity of minoxidil in the dog. Toxicol Pathol 23:498-506.; 1995.

68. Mesfin, G. M., Shawaryn, G. G., and Higgins, M. J. Cardiovascular alterations in dogs treated with hydralazine. Toxicol Pathol 15:409-16; 1987.

69. Metz, A. L., Dominick, M. A., Suchanek, G., and Gough, A. W. Acute cardiovascular toxicity induced by an adenosine agonist- antihypertensive in beagles. Toxicol Pathol 19:98-107; 1991.

70. Mutin, M., Canavy, I., Blann, A., Bory, M., Sampol, J., and Dignat-George, F. Direct evidence of endothelial injury in acute myocardial infarction and unstable angina by demonstration of circulating endothelial cells. Blood 93:2951-8.; 1999.

71. Nicholson, J., et al. Metabonomics, a platform for understanding drug toxicity and gene function. Nature: Reviews in Drug Discovery 1:153-161; 2002.

72. Nicholson, J., et al. Metabonomics: Understanding The Metabolic Responses Of Living Systems To Pathophysiological Stimuli Via Multivariate Statistical Analysis Of Biological NMR Spectroscopic Data. Xenobiotica 29:1181-1189; 1999.

73. Panegyres, P. K., Faull, R. J., Russ, G. R., Appleton, S. L., Wangel, A. G., and Blumbergs, P. C. Endothelial cell activation in vasculitis of peripheral nerve and skeletal muscle. J Neurol Neurosurg Psychiatry 55:4-7.; 1992.

74. Pasqualini, R., and Ruoslahti, E. Organ targeting in vivo using phage display peptide libraries. Nature 380:364-366; 1996.

75. Patel, V., et al. C-Reactive Protein: A 'Golden Marker' for inflammation and coronary artery disease. Cleveland Clinic J Med 68:521-534; 2001.

76. Pennie, W., Woodyatt, N., Aldridge, T., and Orphanides, G. Application of genomics to the definition of the molecular basis for toxicity. Toxicol Lett 120(1-3):353-358; 2001.

77. Pober, J. S., and Cotran, R. S. Immunologic interactions of T lymphocytes with vascular endothelium. Adv Immunol 50:261-302; 1991.

78. Praprotnik, S., Rozman, B., Blank, M., and Shoenfeld, Y. Pathogenic role of anti-endothelial cell antibodies in systemic vasculitis. Wien Klin Wochenschr 112:660-4.; 2000.

79. Rafi, A. Q., Zeytun, A., Bradley, M. J., Sponenberg, D. P., Grayson, R. L., Nagarkatti, M., and Nagarkatti, P. S. Evidence for the involvement of Fas ligand and perforin in the induction of vascular leak syndrome. J Immunol 161:3077-86.; 1998.

80. Rafi-Janajreh, A. Q., Chen, D., Schmits, R., Mak, T. W., Grayson, R. L., Sponenberg, D. P., Nagarkatti, M., and Nagarkatti, P. S. Evidence for the involvement of CD44 in endothelial cell injury and induction of vascular leak syndrome by IL-2. J Immunol 163:1619-27.; 1999.

81. Ridker, P. M. High-Sensitivity C-Reactive Protein. Potential Adjunct for Global Risk Assessment in the Primary Prevention of Cardiovascular disease. Circulation 103:1813-1818; 2001.

82. Robertson, D., et al. Metabonomic assessment of drug-induced vasculitis in rats. Toxicolological Sciences 60(1S):375; 2001.

83. Robertson, D., et al. Metabonomic Assessment of Vasculitis in Rats. Cardiovascular Toxicology 1:7-19; 2001.

84. Robertson, D., et al. Metabonomics: Evaluation of Nuclear Magnetic Resonance (NMR) and Pattern Recognition Technology For Rapid in Vivo Screening of Liver and Kidney Toxicants. Toxicological Sciences 57:326-337; 2000.

85. Seko, Y. Takayasu arteritis: insights into immunopathology. Jpn Heart J 41:15-26.; 2000.

86. Shapiro, H. M. Practical Flow Cytometry: Wiley-Liss Inc., New York.; 1995.

87. Shiohara, T., Sagawa, Y., and Nagashima, M. Systemic release of interferon-gamma in drug-induced cutaneous vasculitis. Lancet 339:933.; 1992.

88. Sinzinger, H., Rauscha, F., Fitscha, P., and Kaliman, J. Effect of high and low-dose aspirin on circulating endothelial cells. Vasa 17:84-88; 1988.

89. Slim, R., et al Effect of dexamethasone on the metabonomics profile associated with phosphodiesterase inhibitor-induced mesenteric vascular lesions in rats. Toxicology and Applied Pharmacology in press; 2002.

90. Solovey, A., Lin, Y., and Browne, P., et al Circulating activated endothelial cells in sickle cell anemia. NEJM 337:1584-1590; 1997.

91. Stagliano, N., Carpino, A., Ross, J., and Donovan, M. Vascular gene discovery using laser capture microdissection of human blood vessels and quantitative PCR. Ann N Y Acad Sci 947:344-349; 2001.

92. Steffen, R. P., Eldon, C. M., and Evans, D. B. The effect of the cardiotonic imazodan (CI-914) on myocardial and peripheral hemodynamics in the anesthetized dog. J Cardiovasc Pharmacol 8:520-6.; 1986.

93. Stephan-Gueldner, M., and Inomata, A. Coronary arterial lesions induced by endothelin antagonists. Toxicol Lett 112-113:531-5.; 2000.

94. Suzuki, H., Swei, A., Zweifach, B., and Schmid-Schonbein, G. In vivo evidence for microvascular oxidative stress in spontaneously hypertensive rats. Hypertension 25:1083-1089; 1995.

95. Uchiyama, T., and Kato, H. The pathogenesis of Kawasaki disease and superantigens. Jpn J Infect Dis 52:141-5.; 1999.

96. Ulrich, R., and Friend, S. Toxicogenomics and drug discovery: will new technologies help us produce better drugs? Nat Rev Drug Discov 1(1):84-88; 2002.

97. Vial, T., and Descotes, J. Clinical toxicity of cytokines used as haemopoietic growth factors. Drug Saf 13:371-406.; 1995.

98. Walter, G., Bussow, K., Lueking, A., and Gluler, G. High-throughput protein arrays: prospects for molecular diagnostics. Trends Mol. Med 8(6):250-253; 2002.

99. Weingand, K., Brown, G., Hall, R., Davies, D., and Gossett, K. Harmonization of animal clinical pathology testing in toxicity and safety studies. Fundamental Appl Toxicol 29:198-201; 1996.

100. Weyand, C. M., and Goronzy, J. J. Pathogenic mechanisms in giant cell arteritis. Cleve Clin J Med 69:SII28-32.; 2002.

101. Weyand, C. M., and Goronzy, J. J. Pathogenic principles in giant cell arteritis. Int J Cardiol 75 Suppl 1:S9-S15; discussion S17-9.; 2000.

102. Ye, S., Lavoie, T., Usher, D., and Zhang, L. Microarray, SAGE and their applications to cardiovascular diseases. Cell Res 12(2):105-115; 2002.

103. Zhang, J., Herman, E., Knapton, A., Chadwick, D., Whitehurst, V., Koerner, J., T, P., Ferrans, V., and Sistare, F. SK&F 95654-induced acute cardiovascular toxicity in Sprague-Dawley rats- histopathologic, electron microscopic, and immunohistochemical studies. Toxicol Pathol 30:28 - 40; 2002.