Novel Vascular Lesions in Mice Given a Non-Peptide Vitronectin Receptor Antagonist

Abstract

Novel vascular lesions were observed in mice given an αvβ3, αvβ5 receptor antagonist (SB-273005) for up to 3 months. Vascular smooth muscle cell (VSMC) necrosis was observed in aorta and renal hilar arteries approximately 6 hours after dosing followed by loss of VSMC, adaptive medial thickening by VSMC hypertrophy and deposition of PAS-positive matrix and collagen. Renal hilar and arcuate arteries developed delayed and transient fibrinoid necrosis and inflammation. Vascular regeneration was not evident following drug-withdrawal after 3 days of dosing. Vascular lesions were associated with necrosis, regeneration and fibrosis of heart, kidney and spleen consistent with initial ischemic injury followed by tissue repair. VSMC toxicity was likely not related to integrin antagonism because lesions were not observed with related compounds and no vascular changes were observed in other preclinical species. In vitro studies failed to demonstrate a direct toxic effect of SB-273005 on VSMC or unique species sensitivity of murine VSMC. In conclusion, SB-273005 caused VSMC necrosis in aorta and renal arteries of mice. Lesions did not progress or recover, but there was medial hypertrophic adaptation even with continued dosing. This is considered direct species-specific VSMC toxicity of unknown mechanism and unrelated to vitronectin receptor antagonism.

Introduction

Although drug-induced vascular injury is encountered frequently in preclinical drug development, the mechanisms involved are poorly understood. Kerns et al. (2005) proposed a classification of drug-induced vascular injury based upon mechanism of injury. These include biomechanical injury secondary to shear and/or hoop stress, direct pharmacological and/or chemical–mediated injury, and alterations secondary to an immunological response and/or inflammation (Kerns et al., 2005). Several compounds that cause vascular toxicity in preclinical species, such as minoxidel and phosphodiesetrase III inhibitors, are also vasoactive effecting a decrease in systemic blood pressure and reflex tachycardia, (Louden and Morgan, 2001). Although systemic hemodynamic changes may be present, hemodynamic changes localized to specific vascular beds, such as the coronary arteries in dogs and mesenteric arteries in rats, may play a greater role in the pathogenesis of these vascular lesions through alterations in shear and possible hoop stress. Other proposed mechanisms are thought to play minor roles. Confirmed drug-induced direct chemical vascular toxicity is uncommon, while immunological and inflammatory mechanisms, the predominant causes for drug-induced vascular injury in humans, are rare in animals (Louden and Morgan, 2001; Kerns et al., 2005).

Most reported cases of drug-induced vascular injury have focused on rat and dog with few descriptions of vascular toxicity in mice (Elwell and Mahler, 1999; Moyer et al., 2002). The present paper describes novel vascular lesions and secondary ischemic changes in mice treated with an integrin antagonist, SB-273005, a potent orally active non-peptide vitronectin (αvβ3, αvβ5) receptor antagonist. In 14-day and 3-month oral toxicity studies, acute and chronic vascular changes were observed in the thoracic aorta and renal arteries with transient ischemia affecting myocardium, renal tubules and spleen.

These changes were considered species-specific and unrelated to pharmacology because limited toxicity with no evidence of cardiovascular effects was observed in rat and monkey, species pharmacologically responsive to SB-273005, in studies up to 9 months in duration at systemic exposures up to 100 times those achieved in mice. Additionally, vascular lesions were not observed in mice given either structurally similar or distinct compounds of equal pharmacological potency at higher exposures. Therefore, a 14-day oral time-course study was conducted to characterize the histopathogenesis of these changes and in vitro studies examined and compared direct effects of SB-273005 on primary vascular smooth muscle cells (VSMC) from mice and monkey.

Materials and Methods

Test Substance

SB-273005 is a potent antagonist of the closely related integrins, human α_vβ₃ (Ki = 1.2 nM) and α_vβ₅ (Ki = 0.3 nM); mouse α_vβ₃ (Ki = 0.19 nM), SB-273005 (Figure 1) was synthesized at SmithKlineBeecham Pharmaceuticals. SB-273005 was administered orally to mice in 1% methylcellulose, adjusted to pH 3.5 (Sigma, St. Louis, MO).

Animals

CD-1 outbred albino mice {Crl:CD-1(ICR)BR} obtained from Charles River Laboratories, Raleigh, NC, were used in these studies. To minimize use of animals, only males were selected for the 14-day time-course study based on a slightly higher incidence of cardiac lesions in males in the 3-month study. Prior to randomization into study groups, mice were evaluated for suitability based on clinical observations, detailed clinical examination and body weight. Each selected animal was identified by a number tattooed on the tail and/or by implanted transponder. Mice were housed individually in stainless-steel cages.

Following randomization, mice were acclimated to local housing conditions for 5 days. On Day 1 of dosing, mice were approximately 12 weeks old. The environmental controls were set to maintain temperature within the range of 64–79°F and relative humidity within the range of 30–70% with a 12-hour light and 12-hour dark cycle. Mice were fed Certified Rodent Diet #5002 (PMI Feeds, Inc.) ad libitum. Filtered tap water was available ad libitum from an automatic watering system.

All studies were conducted in accordance with current guidelines for animal welfare (Guide for the Care and Use of Laboratory Animals, 1996; Animal Welfare Act, 1966, as amended in 1970, as amended in 1976, as amended in 1985, 9 CFR Parts 1,2,3). Procedures used in these studies were reviewed and approved by the internal Institutional Animal Care and Use Committee.

Study Design

14-Day Time-Course Study

One-hundred-fifty-nine male CD-1 mice were divided into 5 groups. Three groups of mice were given orally by gavage either vehicle (1% (w/v) aqueous methylcellulose), 300 or 1000 mg/kg/day SB-273005 for either 3, 8, or 14 consecutive days and sacrificed and necropsied on days 4, 9, and 15, respectively (n = 15 mice/dose group/time point; 12 mice for standard histopathology, 3 mice for electron microscopy). Two additional groups (n = 12/group) were given either 300 or 1000 mg/kg/day SB-273005 consecutively for 3 days (Days 1 through 3), followed by an 11-day drug holiday and then necropsied on Day 15.

Aorta, kidney, heart, and spleen were collected, fixed in formalin, routinely processed in paraffin, sectioned and stained with hematoxylin and eosin (H&E), and examined microscopically. Selected specimens of aorta were subjected to Movat’s pentachrome stain (Movat, 1955). Aorta was collected in toto to the level of femoral artery bifurcation. Four sections of aorta [aortic arch, distal thoracic, proximal abdominal aorta and distal abdominal aorta] were processed for microscopic examination from each animal. All mice given 1000 mg/kg SB-273005 were terminated on Day 1 due to high morbidity and mortality. Only aorta, heart, and kidneys from 12 mice given 1000 mg/kg were processed for microscopic examination.

Immunohistochemistry

To detect apoptotic cells, TUNEL assays and immunohistochemical detection of activated caspase 3 was performed on additional paraffin sections prepared from selected samples of aorta using the Ventana Discovery System. For detection of activated caspase-3, sections were de-paraffinized, rehydrated and subjected to heat antigen retrieval (CC1 citrate buffer Ventana Medical Systems, Inc., Tucson, AZ). Non-specific staining was blocked with 3% hydrogen peroxide, avidin/biotin block and protein block (DAKO, Carpinteria, CA) spiked with 5% normal goat serum. Polyclonal rabbit IgG anti-cleaved caspase-3 antibody (Cell Signaling, Beverly, MA) was applied manually at a dilution of 1:250 for 1 hour. Controls included omission of primary antibody and substitution of primary with non-immune rabbit IgG (SouthernBiotech, Birmingham, AL) matched for concentration. Anti-cleaved caspase-3 antibody was detected with biotinylated goat anti-rabbit IgG (Vector Labs, Burlingame, CA) followed by streptavidin-horseradish peroxidase (HRP), 3,3 diaminobenzidine (DAB) and counterstained with hematoxylin.

The ApopTag^® peroxidase In Situ Apoptosis Kit (Intergen Company, Purchase, NY). was used to detect DNA strand breaks. Briefly, sections were deparaffinized, rehydrated then incubated in protease 1 (Ventana Medical Systems, Tucson, AZ) for 4 minutes. Sections were blocked with 3% hydrogen peroxide followed by equilibration buffer, then incubated in terminal deoxynucleotidyl transferase (TdT) followed by anti-digoxigenin biotinylated secondary (Innogenex, San Ramone, CA), streptavidin-HRP and DAB (Ventana Medical Systems) and counterstained with hematoxylin.

Electron Microscopy

Samples were collected from the aortic arches of 3 controls, and each of 3 mice given 300 mg/kg/day of SB-273005 consecutively for 3 or 14 days and terminated on Day 4 or Day 15, respectively. Samples were dissected into 0.5 mm³ pieces and fixed with 2.5% glutaraldehyde/2 % formaldehyde in 0.1M phosphate buffer, pH 7.3, for 24 hours. Tissues were post-fixed in 1% osmium tetroxide in buffer, dehydrated in a series of graded ethanol (50–100%) and propylene oxide, and embedded in epoxy resin. Approximately 0.5 μm thick sections of tissue were cut, collected on glass slides, and stained with toluidine blue. Regions of interest were identified by light microscopic examination of the stained sections. Subsequently, samples were selected for ultrastructural examination and sections approximately 80 nm thick were cut, collected on copper grids, and stained with uranyl acetate and lead citrate. Samples were examined with a JEOL-1200 EX transmission electron microscope operated at 80 kV and representative images collected using a Gatan MultiScan 794 digital camera.

3-Month Oral Toxicity Studies

Male and female CD-1 mice, approximately 11 weeks of age at initiation of dosing (Charles River Laboratories, Raleigh, NC, 12/sex/group) were given vehicle (1% (w/v) aqueous methylcellulose) or 30, 100 or 300 mg/kg/day of SB-273005 (12/sex/group) by oral gavage for up to 91 consecutive days. Surviving mice were sacrificed and necropsied approximately 24 hours after the final dose. Endpoints evaluated included in-life animal observations, ophthalmology, body weight, food consumption, hematology, clinical chemistry, organ weights and macroscopic and microscopic observations. Tissues were collected and fixed in formalin, processed routinely in paraffin, sectioned and stained with H&E. A standard comprehensive list of tissues was examined microscopically. Selected sections were stained with Masson’s trichrome and subjected to the Period acid Schiff’s reaction (PAS). Only results from the microscopic evaluation are currently presented.

Comparative Effects of SB-273005 on Primary VSMC Adhesion and Viability

An in vitro model system of primary VSMC from mouse and monkey was established to investigate the direct effects of SB-273005 on VSMC, the apparent target tissue in mouse vasculature. The effects of increasing concentrations of SB-273005 on VSMC adhesion, cytotoxicity and apoptosis were observed in each cell line to assess species-specificity.

Reagents

Type II collagenase, elastase, penicillin G, streptomycin, and amphotericin B were purchased from Invitrogen (Carlsbad, CA). Heat-inactivated fetal bovine serum, fibronectin, vitronectin, CaCl₂ were purchased from Sigma (St. Louis, MO). Annexin-V-FITC, propidium iodide (PI), and binding buffer were purchased as a kit, Annexin-V-FITC Apoptosis Detection Kit, from BD Biosciences (San Diego, CA). Protein block and streptavidin HRP were purchased from Dako (Carpineria, CA), anti-alpha actin antibody was purchased from Boehringer-Mannheim (Mannheim, Germany), biotinylated anti-mouse antibody was purchased from Vector Labs. Low-density lipoprotein from human plasma, acetylated, 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl-indocarbocyanine perchlorate-acetylated-low density lipoprotein (DiI AcLDL) was obtained from Molecular Probes (Eugene, OR). SB-273005 was provided by GlaxoSmithKline Drug Dispensary, UM Safety Assessment, King of Prussia, PA. Dulbecco’s Modified Eagle Media (DMEM) with 4.5 mg/mL glucose (DMEM with high glucose), 0.5% trypsin with EDTA, Dulbecco’s Phospate-Buffered Saline (DPBS), and RPMI 1640 media were provided by Media Prep Services, GlaxoSmithKline, King of Prussia, PA. Include: acetylated LDL.

Explant Derivatives and Cell Culture

Aortas from untreated 10–12-week-old CD1 mice or from control cynomolgus monkeys were placed in DMEM with high glucose, 10% heat-inactivated fetal calf serum, 500 U/mL (200 mg/mL for mouse) penicillin G, 500 mg/mL (200 mg/mL for mouse) streptomycin, and 1.25 μg/mL amphotericin B (monkey only) on ice. Adventitial and luminal surfaces were scraped, then aortas were minced and incubated in Henk’s Balanced Salt Solution containing 100 U/mL Collagenase Type II, 0.5 mg/mL elastase (monkey only), 100 U/mL penicillin G, 100 mg/mL streptomycin, 0.25 μg/mL amphotericin B, 3 mM CaCl₂ at 37°C for four hours with gentle agitation. Cells were placed through a 70 μm nylon mesh filter, centrifuged at 500 g (1400 rpm), 4°C, resuspended in DMEM with high glucose, 10% heat-inactivated fetal calf serum, 500 U/mL (200 mg/mL for mouse) penicillin G, 500 mg/mL (200 mg/mL for mouse) streptomycin, and 1.25 μg/mL amphotericin B (monkey only) (complete media) and cultured at 37°C, 5% CO₂. Cells were positively identified as smooth muscle cells by immunohistochemical staining with smooth-muscle cell alpha actin and negative uptake of DiI AcLDL by flow cytometry (endothelial cell marker). Mouse and monkey smooth muscle cells were cultured in complete media, 37°C, 5% CO₂. All cells were used before passage 9.

Cell Adhesion Studies

Corning 96-well ELISA plates were precoated overnight at 4°C with 0.1 mL of human vitronectin (0.2 μg/mL in RPMI 1640 medium) or fibronectin (1.0 μg/mL or 10.0 μg/mL in RPMI 1640 medium). At the time of the experiment, the plates were washed once with RPMI 1640 medium and blocked with 3.5% bovine serum albumin in RPMI medium for 1 hr at room temperature. Cells were resuspended in RPMI 1640 containing 0.2% bovine serum albumin at a density of 1 × 10⁶ cells/mL. Fifty microliters of cell suspension was added to each well and incubated for 1 hr at 37°C, in the absence or presence of various doses of SB-273005 ranging from 0.1 nM to 1 μM. Following incubation, 25 μL of a 10% formaldehyde solution, pH 7.4, was added and the cells were fixed at room temperature for 10 min. The plates were washed 3 times with 0.2 mL of RPMI and the adherent cells were stained with 0.1 mL of 0.5% toluidine blue for 20 min at room temperature. Excess stain was removed by extensive washing with deionized water. The toluidine blue incorporated into cells was eluted by the addition of 0.1 mL of 50% ethanol containing 50 mM HCl. Cell adhesion was quantitated at an optical density of 600 nm on a microtiter plate reader (Titertek Multiskan MC, Sterling, VA).

Immunohistochemistry

Cells were grown on chamber slides to ~80% confluency in complete media, 37°C, 5% CO₂. All immuhistochemistry was performed at room temperature unless stated otherwise. Slides were incubated in 0.3% hydrogen peroxide in methanol for 10 minutes. Slides were then incubated in protein block for 45 minutes, followed by a 30-minute incubation in 2.5 μg/mL mouse anti-smooth muscle cell alpha actin or isotype-matched control. Three 5-minute washes in tris buffer containing 0.15M NaCl (TBS) were performed. Biotinlyated anti-mouse IgG was diluted 1:200 and slides were incubated for 15 minutes followed by three 5-minute washes TBS. Slides were then incubated in streptavidin-HRP for 15 minutes, washed 3 times 5 minutes in TBS, then detection was performed using DAB substrate according to the manufacturer’s directions. Slides were counterstained for 5 minutes in hematoxylin, dehydrated through a graded series of alcohols, cleared in xylene then coverslipped.

Apoptosis and Cytotoxicity Assays

Fluorescence-based apoptosis and cytotoxicity assays were carried out using flow cytometry. Apoptosis was detected with Annexin-V-FITC (to detect phosphatidyl serine translocation) in conjunction with the cell-impermeant DNA-binding fluorescent dye PI to test for cytotoxicity. Then 3.5 × 10⁶ cells in 2 mL of complete media were placed into each well of 6-well cell culture dishes or 6-well culture dishes coated overnight at 4°C with either vitronectin (0.2 μg/mL) or fibronectin (0.4 μg/mL) and incubated in a humidified chamber at 37°C, 5% CO₂ overnight. Cells were gently washed with DPBS, and then placed into DMEM with high glucose. SB-273005 was added at concentrations varying from 0 to 100 μM in 0.1% DMSO for 6, 12 or 24 hrs. Cells were then collected and placed into 100 μL 1X Binding Buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl₂), 5 μL AnnexinV-FITC, and 2 μL PI, then vortexed briefly and incubated in the dark for 15 minutes, then an additional 400 μL 1X binding buffer was added to each sample, and samples were analyzed within 45 minutes. A 4-hour treatment with 100 μM menadione was run with each assay as a positive apoptosis/cytotoxicity control for the system. Statistical analyses were performed on the ranked data with parametric one-way ANOVA (Milliken, 1984; SAS/STAT User’s Guide, 1990). Pair-wise comparisons between treated group and vehicle control were performed by Dunnett’s multiple comparisons test when overall tests showed significance (Dunnett, 1965; SAS/STAT User’s Guide, 1990).

Results

In-life Observations

14-Day Time-Course Study

All mice given a single dose of 1000 mg/kg died or were sacrificed in moribund condition approximately 4 to 6 hours postdose and 2 mice given 300 mg/kg/day were found dead on Day 5 and 7. Morbidity/mortality was preceded by decreased activity and low body posture.

3-Month Oral Toxicity Study

The deaths of 1 control and 5 mice given ≥30 mg/kg/day during the first 3 days of treatment were unexplained. These mice were not necropsied and were replaced. Thereafter, no drug-related deaths were observed.

Microscopic Observations

Temporal changes in drug-related microscopic findings were observed in aorta, renal artery, spleen, heart and kidney (Tables 1 and 2; Figures 2–8).

Aorta

Degenerative changes of the tunica media of the aorta were evident in all SB-273005 treatment groups (≥30 mg/kg/day). The most severe necrotic changes of the aorta were observed in mice given 1000 mg/kg/day and sacrificed approximately 6 hours postdose. Minimal to moderate medial smooth muscle cell necrosis was observed in arch, caudal thoracic and mid-abdominal aortic segments in the majority of mice examined (compare Figures 2A and 2B). In consequence, elastic laminae collapsed and did not show the crinkled appearance commonly observed in unaffected tissue. In mice given 300 mg/kg/day and terminated on Day 4, VSMC changes, characterized by necrotic cellular debris, segmental loss of smooth muscle cells and smooth muscle cell hypertrophy (with rare mitotic figures) were noted in the aortic arch of all mice (Figure 2C). In approximately half of the animals, the lesions extended to the caudal thoracic segment.

In mice terminated on Day 9 or Day 15 (following either 8 or 14 consecutive doses), aortic changes were similar in nature consisting of smooth muscle cell hypertrophy and segmental loss of smooth muscle cells replaced by amorphous extracellular matrix (staining green with Movat’s pentachrome) separating elastic fibers (Figures 2D and E). Aortic changes persisted after 3 months of dosing and included smooth muscle cell hypertrophy (Figure 2F), increased deposition of PAS-positive extracellular matrix and collagen (staining blue with Masson’s trichrome; Figure 3), all of which contributed to a considerable thickening of the media. The incidence of findings was dose-dependent (Table 2). Inflammatory cell infiltrates were not associated with aortic degenerative changes at any time point.

Renal Arteries

Renal vascular changes were evident in all SB-273005 treatment groups (≥30 mg/kg/day). In all examined mice given 1000 mg/kg/day and terminated approximately 6 hours postdosing, there was minimal to severe medial necrosis involving renal and/or arcuate arteries (compare Figure 4A and 4B). In mice given 300 mg/kg/day and terminated on Day 4, VSMC degeneration in renal arteries was noted in ~25% of mice and was characterized by loss of VSMC and attenuation of the media (Figure 4C). Minimal periarterial inflammation was noted in one mouse. In contrast, renal arterial changes observed in 50% of mice given 300 mg/kg/day and terminated on Day 9 or Day 15 (following either 8 or 14 consecutive doses) were characterized additionally by fibrinoid necrosis with minimal to severe periarterial inflammation (Figures 4D and 4E). After 3 months of continuous dosing, renal arteries still showed significant loss of VSMC and attenuation, although there was no active degeneration or remarkable perivascular inflammation (Figure 4F).

Heart

Cardiac changes were evident in all SB-273005-treatment groups. In mice given 1000 mg/kg/day and terminated approximately 6 hours postdosing, minimal acute focal myocardial necrosis was noted in one mouse. In mice given 300 mg/kg/day and terminated on Day 4, myocardial lesions were noted in 75% of the mice consisting of minimal to severe, acute to subacute multifocal myocardial necrosis involving the left ventricle and interventricular septum (Figure 5A). Necrosis and/or degeneration of VSMC in coronary arteries was also observed in 3 mice. In mice given 300 mg/kg/day and terminated on Days 9 or 15 (following either 8 or 14 consecutive doses), multifocal cardiac fibrosis was evident in 50% or less of the mice. Fibrotic foci were composed of highly cellular, immature connective tissue, considered a reparative response to preceding necrosis, with no evidence of continued myofiber degeneration. In mice given compound for 3 months, mature fibrous connective tissue (aniline blue positive with Masson’s trichrome) was arranged in narrow bands extending from base to apex in the middle or subendocardial region of ventricular wall and septum (Figure 5B).

Kidney

Renal tubular changes were observed in mice given ≥300 mg/kg/day SB-273005. Minimal to severe tubular necrosis and/or degeneration was seen in most mice on Days 1, 4, and 9 characterized by the presence of necrotic cells sloughing into tubular lumens and tubular casts composed of hyaline/granular material. In addition on Day 4 and 9, tubular regeneration was characterized by increased mitotic figures and tubules lined by prominent basophilic epithelium. Tubular regeneration was the most common finding in mice on Day 15 (given 3 or 14 doses; Figure 5C) and often associated with fibrinoid vascular necrosis and inflammation (Figure 5D). However, renal tubular findings in mice dosed for 3 months were not different from those observed in controls.

Spleen

Splenic changes were evident in mice given ≥100 mg/kg/day SB-273005. Acute splenic changes, characterized by minimal to mild fibrinoid degeneration localized to perifollicular red pulp with marginal zone depletion were evident in 8 of 12 mice given 300 mg/kg/day and terminated on Day 4 (Figures 6A and 6B). Similar findings were noted in mice given 300 mg/kg/day and terminated on Day 9. On Day 15, perifollicular congestion, marginal zone depletion and early fibrosis were the most common observation (Figure 6C). No splenic changes were observed in mice given 3 consecutive daily doses of 300 mg/kg/day and terminated on Day 15. Perifollicular fibrosis and irregular follicular cell depletion (marginal zone and/or periarteriolar lymphoid sheath) were observed in mice given compound for 3 months (Fig. 6D).

Immunohistochemistry

Examination of selected aortic sections stained for activated caspase 3 and TUNEL revealed no staining of nuclei on Day 1 of dosing and on Day 4 only occasional apoptotic VSMC were seen in aortas of mice given SB-273005 (Figure 7). Therefore, apoptosis was not considered an important mechanism of VSMC loss.

Electron Microscopy

Ultrastructural examination of aortic arches from mice given 300 mg/kg/day and terminated on Day 4 revealed degenerative and adaptive changes of VSMC and endothelial cell hypertrophy (Figure 8). Necrotic VSMC containing numerous autophagic vacuoles and electron-dense myofilamentous inclusions were frequently observed. No degenerative changes were observed in endothelial cells. Contact areas between viable VSMC and elastic laminae, were highly irregular in drug-treated mice characterized by disrupted cellular processes with deposition of amorphous extracellular matrix (consistent with proteoglycans) (Figure 8B). In mice treated for 14 days and terminated on Day 15, endothelial cell hypertrophy and subendothelial deposition of extracellular matrix were observed, and large areas of the media were devoid of VSMC. Deposition of proteoglycans and collagen at the interface of the elastic laminae and VSMC was greatly increased causing marked distortion of VSMC shape (Figures 8C and 8D).

Comparative Effects of SB-273005 on Primary VSMC Adhesion and Viability VSMC Cultures

A qualitative assessment of positive immunoreactivity of aortic smooth muscle cells in culture with smooth muscle cell alpha actin was established for mouse and monkey and estimated to be approximately 90% of all cells staining positively. Endothelial cell contamination of mouse cultures appeared minimal by flow cytometry based on negative uptake of DiI-acetylated LDL following a four-hour incubation period (data not shown).

Adhesion Assays to Human Recombinant Vitronectin

Mouse and monkey cells attached and spread in wells pre-coated with human recombinant vitronectin within 1 hour. In both species, cell attachment was inhibited with increasing concentration of SB-273005. The IC₅₀’s obtained from the adhesion assays to human recombinant vitronectin for mouse and monkey were similar (monkey IC₅₀ = ~6 nM, mouse IC₅₀ = ~2.9 nM) (Figure 9A).

Effects of SB-273005 on Cytotoxicity and Apoptosis

Treatment with 100 μM menadione is a well characterized inducer of apoptosis and cytotoxicity and was established as a positive control for Annexin V and PI staining in both mouse and monkey VSMC by flow cytometry (Sakagami et al., 2000; Okayasu et al., 2001; Chiou et al., 2003). Following incubation of monkey and mouse VSMC with SB-273005 cultured on plastic substratum, cell detachment, qualitatively, appeared to be concentration-dependent and there was no increase in cytotoxicity nor apoptosis parameters in cells from either species (data not shown).

No effect on cytotoxicity, as measured by PI+/Annexin V+/− staining, was observed following incubation of monkey and mouse VSMC with SB-273005 on human recombinant vitronectin substratum (Figure 9B). Interestingly, a concentration-dependent increase in percent apoptotic VSMC was observed in monkey but not mouse in response to increasing concentrations of SB-273005 with an IC₅₀ of ~9.6 μM, as measured by changes in Annexin V+/PI− staining (Figure 9C).

To investigate whether another substratum would alter response of murine VSMC to SB-273005, a similar adhesion study was conducted with fibronectin substratum. Cell detachment appeared to be again concentration-dependent and not associated with increased apoptosis or cytotoxicity (data not shown).

Discussion

Unique arterial changes that have not been described for rat, dog or monkey were observed in aorta and renal arteries in mice given SB-273005, an orally active non-peptide α_vβ₃, α_vβ₅ receptor antagonist, for up to 3 months. In SB-273005-treated mice, VSMC necrosis of aorta and renal hilar arteries was observed acutely 6 hours after dosing, followed by continued degeneration and VSMC cell loss after 3 consecutive doses. There was no regeneration after drug-withdrawal. Aortic changes were similar in mice treated for 8 days or 3 months and characterized by VSMC loss, VSMC hypertrophy, deposition of amorphous extracellular matrix, increased collagen and elastic fiber collapse. These changes occurred in the absence of vascular hemorrhage, thrombosis or inflammation, but were associated with ischemic injury at terminal vascular beds in heart, kidney, and spleen.

It has been suggested that the mechanism of drug-induced vascular injury initiation in preclinical species could be deduced from pathologic features, affected vessel caliber and vascular bed (Louden and Morgan, 2001). Lesions affecting small and medium sized arteries are characterized by medial hemorrhage, necrosis and inflammation and have been considered a consequence of altered local hemodynamics mediated either directly or indirectly by compounds (Goligorsky et al., 2002; Yu et al., 2006; Louden et al., 2006). Hemodynamic changes effect alterations in blood flow, shear stress and vessel geometry. These biomechanical signals are sensed and transduced by both EC and VSMC affecting among others, numerous shear stress and stretch responsive genes (McCormick et al., 2003; Wasserman et al., 2004; Kakisis et al., 2004; Harrison 2005). In preclinical species, some vasodilators affect selected small and medium caliber arteries, e.g., coronary arteries of dogs and splanchnic vascular beds of rats increasing localized blood flow to these vascular beds (Kerns et al., 2005). Profound local vasodilation may lead to vessel overdistention, thereby increasing endothelial shear/VSMC hoop stress and disrupting the intimal barrier.

Microscopic changes initially consist of hemorrhage and plasma leakage through the disrupted intimal barrier, followed by endothelial and/or VSMC necrosis and secondary inflammation. Lesions induced by vasodilators are usually reversible. Vasodilators frequently also cause a decrease in systemic blood pressure and reflex tachycardia but these effects are generally considered unrelated to localized vascular injury in preclinical species. Examples of this type of injury occurring as a direct consequence of intended pharmacology of a drug include minoxidil (potassium channel opener), endothelin receptor antagonists, Fenoldopam (dopamine DA1 agonist) and phosphodiesterase (PD 3) inhibitors (Kerns et al., 2005; Louden et al., 2006).

Vascular injury induced by vasoconstrictive agents such as L-norepinephrine, endothelin, or thromboxane, are associated with hypertension and vascular changes including localized or widespread constrictions of small to mid-sized arteries, narrowing of the lumen, VSMC necrosis and hyaline thickening of the media (Louden and Morgan, 2001). Pulmonary hypertension may be induced in rats with monocrotaline causing endothelial necrosis, VSMC proliferation, migration and hypercontraction (Pullamsetti et al., 2005).

The vascular lesions associated with SB-273005 were unique from those described for either vasodilators or constrictors because large elastic arteries were affected rather than medium to small-sized arteries, though certain features support a role of hemodynamics in lesion development. The vascular lesions were located at sites of turbulent blood flow (ascending aorta, a major aortic bifurcation, and renal arteries) which are recognized to be predisposed to vascular injury (Boisseau, 2005). In addition, the ischemic lesions noted in kidney, heart, and spleen indicate altered vascular function and perfusion. Although hemodynamic effects of SB-273005 in mice have not been studied, these features suggest either a direct or indirect effect of SB-273005 hemodynamics and vascular perfusion. Interestingly, cardiovascular studies in rat and monkey did not reveal an effect of SB-273005 on blood pressure or heart rate (unpublished observations).

Direct chemical vascular injury can be induced by primary amine compounds, such as the industrial compound allylamine or the sweat pea (Lathyrus odoratous) toxin (β-aminoproprionitrile; β-APN; Boor et al., 1995). Allylamine in rat is likely specifically metabolized in VSMC of small to medium sized vessels and causes medial hypertrophy and subintimal proliferation of VSMC. β-APN inhibits lysyl oxidase, thereby interfering with collagen and elastin cross-linking in rat connective tissue and aorta. Chronic treatment of immature rats is associated with aortic aneurysms. Both compounds given concomitantly for 10 days caused acute VSMC necrosis in large elastic arteries and degeneration of mid-sized muscular arteries without damage to endothelial cells, medial hemorrhage or inflammation. Months after the 10-day treatment, the thoracic aorta showed considerable medial thinning with loss of viable VSMC, disorganized elastic fibers and aneurysms (Boor, 1997). Acute necrotic effects of SB-273005 on mouse aorta VSMC shows some similarities to the combined treatment of allylamine and β-APN, suggesting a possible direct effect of SB-273005 on VSMC, however, in vitro studies failed to demonstrate such an effect. Although VSMC regeneration was not a feature of SB-273005-induced vascular injury, VSMC loss did not progress and no thinning of the vascular wall or aneurisms were observed. Instead, there was an adaptive tissue response with VSMC hypertrophy and increased matrix and collagen deposition.

In addition, apoptosis was not a feature of the degenerative medial changes in SB-273005 in contrast with the reported role of apoptosis in the pathogenesis of vascular lesions induced by phosphodiesterase-inhibitors (Zhang et al., 2002; Slim et al., 2003).

The chronic arterial changes in mice given SB-273005 are strikingly similar to those described in a mouse model of Hutchinson-Gilford progeria syndrome (Varga et al., 2006). These mice carry a human bacterial artificial chromosome with the progeria mutation, that produces defective mutant laminin A protein. Lamin A is nuclear membrane protein that plays an important role in cytoplasmic-nuclear mechanotransduction. Transgenic mice develop progressive medial VSMC loss in large arteries, medial thickening with proteoglycan and collagen replacement, elastic fiber breakage and altered functional response to the administration of a vasodilator. The lack of associated secondary ischemic changes in internal organs and lack of inflammatory response in renal arteries may be related to the slow and progressive loss of VSMC in these transgenic mice. The authors propose that the mutated laminin rendered VSMC fragile and susceptible to progressive devitalization at arterial locations that experience mechanical stress and sites of turbulent flow, such as the aortic arch and bifurcations. Similarly, lesions in SB-273005 were located at sites of turbulent flow. SB-273005-induced hemodynamic changes, whether direct or indirect, would potentially exacerbate lesion development at sites of turbulent flow or alternatively, SB-273005 may directly alter or impair normal adaptive responses to turbulent flow.

Integrins play an important role in both EC and VSMC cell-cell and cell-matrix adhesion and through connections to the cytoskeleton, activate intracellular signaling pathways (Brassard, 1999; Hynes, 2002). Although α_vβ₃, α_vβ₅ integrins play a structural and functional role in the vasculature, it is considered unlikely that SB-273005-induced vascular changes are pharmacology-based. In the current study, lesions were limited to VSMC as endothelial cell loss or necrosis was not observed. In addition, these changes were not observed in chronic toxicity studies rat and monkey, at systemic exposures up to 100× that achieved in mouse in the presence of other pharmacologically mediated effects (growth plate hypertrophy with retained primary spongiosa secondary to osteoclast inhibition). Importantly, in vivo studies in mice given structurally similar and distinct vitronectin antagonists of similar potency achieving exposures equal to or greater than SB-273005 failed to induce any vascular changes (unpublished observations). Although SB-273005 is a highly selective integrin antagonist, an off-target hit on another integrin in mouse can not be excluded. However, there are no reports of vascular changes induced by other integrin antagonists in mice (Yamamoto, 1998) nor, to our knowledge, in integrin knock-out mice (Hynes, 2002; Reynolds et al., 2002; Hodivala-Dilke, 1999).

In vitro studies showed that in pharmacologically relevant species (mouse and monkey), cell attachment was inhibited with increasing concentrations of SB-273005. However, mouse cells were not uniquely sensitive compared with monkey. Furthermore, SB-273005-induced apoptosis was only observed in monkey VSMC cultures on vitronectin, a species where vascular lesions were not observed in chronic studies at exposures up to 100× of those achieved in mouse. Apoptosis or cytotoxicity was not observed at any concentration tested in mouse VSMC cultured on different substrata. The findings in monkey VSMC are an anticipated pharmacological effect of integrin antagonism as apoptosis or anoikis is a known consequence of integrin antagonism (Frisch and Screaton, 2001). Currently, anoikis seems more likely, since it appears that there is disruption of smooth muscle cells from associated elastic laminae. These in vitro results may have failed to reflect in responses because SB-273005-mediated effects may be endothelial-dependent. Moreover, in vivo toxicity differences of SB-273005 may be due to yet unexplored (local) metabolic differences, since onset of clinical signs occurred several hours after Cmax of the parent compound (unpublished observation).

In SB-273005-treated mice, apoptosis was not a feature of the degenerative medial changes in contrast with the reported role of apoptosis in the pathogenesis of vascular lesions induced by phosphodiesterase inhibitors (Zhang et al., 2002; Slim et al., 2003). From studies on atherosclerosis, it is well known that VSMC have an inherent potential to proliferate (Campbell et al., 1991). However, in SB-273005-treated mice, no regenerative efforts were observed and there was no proliferation of VSMC even when dosing was discontinued after only 3 consecutive daily doses. Similarly, no VSMC proliferation was observed in progerin mutant transgenic mice or rats treated with β-APN (Varga et al., 2006; Boor, 1997).

It is of interest that in the spontaneously hypertensive rat, thickening of the large arteries is due to VSMC hypertrophy while smaller resistance vessels thicken as a result of VSMC hyperplasia (Campbell et al., 1991). Currently, VSMC necrosis was also induced in mouse renal hilar arteries, resulting in chronic loss of VSMC. After the initial 3-day insult on VSMC, medial degenerative changes including fibrinoid necrosis and perivasculitis affecting the arcuate arteries ensued within the next 12 days, with or without continued dosing but were no longer present after 3 months. These findings are reminiscent of spontaneous age-and strain-dependent polyarteritis, commonly including renal arteries, or rat mercuric chloride-induced intestinal vasculitis (Elwell and Mahler, 1999; Plendel et al., 1996; Kiely et al., 1997). The pattern of vascular injury, i.e., acute onset and limited progression, is consistent with vascular injury induced by vasodilators (Kerns et al., 2005). Therefore, delayed degeneration and inflammation of renal arteries currently observed in mice may be secondary to vasodilation.

Minimal to severe acute degenerative changes in heart and kidney following 1 or 3 doses of SB-273005 were considered secondary to ischemia based on distribution of the lesions. This type of acute ischemic injury suggests that initial systemic exposure to SB-273005 in mice had profound vasomotor or hemodynamic consequences. In the heart, distribution of degenerative changes was consistent with location of terminal vascular beds of coronary arteries and in the kidney, necrotic foci were vasocentric and segmental. Over time, even in the face of continued dosing, these lesions became reparative in nature, and in some cases, appeared to reverse, which was most apparent after 3 months of continued dosing. In these mice, only strands of fibrotic tissue remained in the heart, replacing lost myocardial tissue. Renal observations in treated mice were similar to those of control mice.

The pattern of ischemic degeneration in heart and kidney is well known, however, no reports on splenic changes currently observed were found. The pattern of necrosis, hemorrhage and cellular depletion seen after 3 doses of SB-273005, is consistent with a vascular compromise of the marginal zone and adjacent red pulp. These structures are supported by capillaries that are derived from the central arteriole and the marginal sinus (Schmidt et al., 1993; Balázs et al., 2001). Most likely, the main splenic artery may have been affected acutely in a similar fashion as the aorta and renal artery, causing inadequate oxygenation of the splenic periphery. Degeneration in the spleen caused by 3 days of dosing was completely recoverable, in contrast to effects in other tissues. However, with continued dosing, chronic tissue damage was also observed in the spleen, indicating extended effects of SB-273005.

In conclusion, SB-273005 caused acute, transient necrosis of VSMC in aorta and renal arteries of mice. This is a mouse-specific lesion considered a direct effect of unknown mechanism on VSMC and unrelated to vitronectin (α_vβ₃, α_vβ₅) receptor antagonism. Hemodynamic changes contributed to lesion pathogenesis but were not considered the initiating cause of injury. Chronic adaptive changes of the aorta include medial thickening with VSMC hypertrophy and deposition of collagen and extracellular matrix while renal hilar and arcuate arteries developed delayed but transient fibrinoid necrosis and inflammation. Ischemic necrosis of cardiac, renal and splenic tissues in areas of dependent terminal arterial blood supply indicate profound secondary hemodynamic consequences in these tissues.