Histopathology of Vascular Injury in Sprague-Dawley Rats Treated with Phosphodiesterase IV Inhibitor SCH 351591 or SCH 534385

Abstract

Histopathological and immunohistochemical studies were conducted to characterize vascular injuries in rats treated with phosphodiesterase (PDE) IV inhibitors SCH 351591 or SCH 534385. Sprague-Dawley rats were administered PDE IV inhibitors by gavage at a range of doses and times. The two PDE IV inhibitors induced comparable levels of vascular injury, primarily in the mesentery and to a lesser extent in the pancreas, kidney, liver, small intestine, and stomach. Mesenteric vascular changes occurred as early as one hour, progressively developed over twenty-four to forty-eight hours, peaked at seventy-two hours, and gradually subsided from seven to nine days. The typical morphology of the vascular toxicity consisted of hemorrhage and necrosis of arterioles and arteries, microvascular injury, fibrin deposition, and perivascular inflammation of a variety of blood vessels. The incidence and severity of mesenteric vascular injury increased in a time- and dose-dependent manner in SCH 351591- or SCH 534385-treated rats. Mesenteric vascular injury was frequently associated with activation of mast cells (MC), endothelial cells (EC), and macrophages (MØ). Immunohistochemical studies showed increases in CD63 immunoreactivity of mesenteric MC and in nitrotyrosine immunoreactivity of mesenteric EC and MØ. The present study also provides a morphological and cellular basis for evaluating candidate biomarkers of drug-induced vascular injury.

Introduction

Phosphodiesterase (PDE) IV inhibitors have been investigated as possible anti-inflammatory and anti-asthma drugs (Souness et al. 2000; Torphy and Undem 1991). However, the side effects of these drugs on the central nervous, gastrointestinal, and vascular systems (Losco et al. 2004; Lyoumi et al. 1998; Robertson et al. 2001; Slim et al. 2002, 2003) have posed an obstacle to drug development. A major obstacle is that it is not known if preclinical drug-induced vascular injury (DIVI) is relevant to clinical applications. DIVI has become a major discussion point for the FDA and for pharmaceutical companies.

The mechanisms by which PDE IV inhibitors induce vascular injury may be multifaceted. An increase in intracellular cAMP (Larson et al. 1996) and nitrative stress (Slim et al. 2003) were proposed to contribute to DIVI. Physical stress secondary to local vasodilation may be a significant factor leading to injury (Daguès, Pawlowski, Sobry et al. 2007; Kerns et al. 2005). A significant inflammatory component to DIVI has also been observed in multiple studies (Slim et al. 2002, 2003; Dietsch et al. 2006; Mecklenburg et al. 2006; Daguès, Pawlowski, Sobry et al. 2007). A study on PDE III-inhibitor–induced vascular injury revealed that inducible nitric oxide synthase (iNOS) and nitrotyrosine were locally produced by mast cells (MC) and endothelial cells (EC) (Zhang et al. 2006). Peroxynitrite formation was correlated with proinflammatory changes and the later development of necrotic and apoptotic cell death of mesenteric blood vessels (Zhang et al. 2006). Recently, a comprehensive review of nitric oxide (NO) and peroxynitrite in cardiovascular disease has suggested that peroxynitrite formation may contribute to the development of early vascular injury, followed by the formation of other reactive nitrogen species derived from the reaction of nitrite with proteins such as myeloperoxidase in inflammatory cells, secondarily contributing to nitrosative stress (Pacher et al. 2007). Evidence is growing that peroxynitrite formation is a common final pathway in drug-dependent toxicity (Denicola and Radi 2005).

Biomarkers for detecting DIVI remain to be discovered and characterized. Although metabonomic profiling of urine has been demonstrated to be useful in the assessment of PDE IV-inhibitor–induced vascular injury in rats (Robertson et al. 2001; Slim et al. 2002), reliable biomarkers using serum or plasma have not been developed in preclinical models (Losco et al. 2004). Multiple technologies (toxicogenomics, genomics, proteomics, metabonomics, bioinformatics, etc.) have been proposed as tools for seeking vascular injury biomarkers (Kerns et al. 2005; Zhang et al. 2002). However, the high cost and technically demanding nature of these methods make it difficult to incorporate them in routine toxicology studies (Ishii et al. 2007). The evaluation of candidate biomarkers identified by the various “omics” methods requires direct comparison between biomarker response and histopathology findings. Therefore, histopathology remains a gold standard method in the search for vascular injury biomarkers, which may then lead to an understanding of the mechanisms underlying DIVI.

The objectives of the present study were to characterize the morphological features of vascular injury in rats treated with PDE IV inhibitors SCH 351591 or SCH 534385, to provide clues concerning possible mechanisms responsible for the development of drug-induced vascular injury, and to provide the morphological basis for selecting candidate biomarkers for PDE IV-inhibitor–induced vascular injury.

Materials and Methods

Animals

Sprague-Dawley rats (ten to twelve weeks old) were obtained from Charles River Laboratories (Wilmington, MA, USA) and housed separately in an environmentally controlled room (18°C–21°C, 40%–70% relative humidity) with a twelve-hour light/dark cycle. Rats were fed Certified Purina Rodent Chow #5002 (Ralston Purina Co., St. Louis, MO, USA) and water ad libitum. The experimental protocol was approved by the Institutional Animal Care and Use Committee, Center for Drug Evaluation and Research, FDA and conducted in an AAALAC-accredited facility (National Research Council 1996).

Study Design

Rats were treated with SCH 351591 or SCH 534385 (Schering-Plough Research Institute, Lafayette, NJ, USA) by oral gavage at a range of doses and various time points. The present study consisted of four study groups.

Study 1 (mixed dose response and time course):Rats were administered SCH 351591 at 3, 4.5, 6, 9, 20, 40, 80, or 160 mg/kg/day (mkd) for one to nine days. Rats were treated with saline for one, two, three, or nine days as a control. Animals were sacrificed twenty-four hours after the last dose for all except 80 and 160 mg/kg, where animals were sacrificed early.

Study 2 (time course):Rats were administered SCH 351591 at 4.5 mkd for one to ninety-six hours. Control rats were saline treated for twenty-four, forty-eight, or seventy-two hours.

Study 3 (time course):Rats were administered SCH 351591 at 20 mkd for one to seventy-two hours. Control rats were saline treated for twenty-four, forty-eight, or seventy-two hours. For both time-course studies, the rats in the one- to twenty-four-hour groups were sacrificed at the indicated time after a single dose. The animals in the longer time groups were dosed daily and sacrificed twenty-four hours after the last dose.

Study 4:Rats were treated with SCH 534358 or saline at 20 or 40 mkd for three days.

Histopathological Studies

The entire mesentery, heart, kidney, liver, pancreas, spleen, and small intestine were collected at necropsy. Three fragments of the mesentery were fixed in 10% formalin solution, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) for histopathological evaluation. To visualize MC granules, staining was done with toluidine blue. Fibrin deposits within blood vessel wall or in extracellular space were visualized with phosphotungstic acid hematoxylin (PTAH) stain. Vascular elastic fibers and smooth muscle cells were stained using Movat’s pentachrome.

Immunohistochemical Studies

Formalin-fixed, paraffin-embedded tissue sections were used for immunohistochemical studies. TUNEL assay and indirect immunoperoxidase staining for CD63 and nitrotyrosine were performed as described previously (Zhang et al. 2006).

Semiquantitative Grading System for Vascular Injury

The severity of mesenteric vascular injury was ranked on a scale of 0–5, mainly based on the percentage of arteries showing hemorrhage and necrosis, with assessment of microvascular injury and vascular inflammation. Scores of 1 or 2 indicate early inflammatory response, and scores of 3 to 5 were used for destructive vascular lesions, respectively. The specific criteria are: 0 = normal, no vascular injury; 1 = minimal inflammation manifested itself in plexiform vasculopathy (multiple channels in the media and inflammatory cell infiltration in the media and adventitia of small veins); 2 = mild acute microvascular injury (fibrin insudation, fibrin exudation, inflammatory cell infiltration, and edema in the perivascular spaces of the postcapillaries, venules, and arterioles), as well as increased severity and frequency of lesions in score 1, without significant evidence of necrosis; 3 = predominant arteriolar hemorrhage and necrosis in < 30% of mesenteric vessels, in association with moderate inflammation; 4 = predominant arterial hemorrhage and necrosis in 31%–60% of mesenteric vessels, in association with severe inflammation; and 5 = severe arteriolar or arterial hemorrhage and necrosis in > 60% of mesenteric vessels, accompanied with very severe inflammation.

Statistics

Differences in scores of mesenteric vascular lesions between groups were compared using the Kruskal-Wallis rank test. A value of p≤ .05 was considered statistically significant.

Results

Characteristic of PDE IV-inhibitor–induced Vascular Injury in Rats

Fully developed vascular injury was found at seventy-two hours after rats were administered SCH 351591. On gross examination, mesenteric vessel hemorrhage was discernible. On microscopic examination, medial hemorrhage, necrosis, and apoptosis of blood vessels were found mainly in the arteries and arterioles of the mesentery, followed by, in order of decreasing incidence, in the arteries and arterioles of the pancreas, in the arterioles of the kidney (at the hilum), the liver (at the hepatic triad), the small intestine (at muscular coat), and the stomach (at lamina propria mucosa and submucosa). Vascular lesions were not found in the coronary arteries, nor were any myocardial lesions identified.

Most representative vascular lesions induced by these PDE IV inhibitors were manifested in mesenteric blood vessels as follows: (a) medial hemorrhage, necrosis, and apoptosis in small- to large-sized muscular arteries (Figure 1A); (b) microvascular injury consisting of fibrin insudation (Figures 1Band 1C) and fibrin exudation (Figure 1D), demonstrated by PTAH staining; (c) EC activation (Figures 1E and 1F) and apoptosis; (d) hemorrhage of the arterioles, venules, and capillaries; and (e) perivascular inflammatory response in a variety of blood vessels (capillaries, venules, small veins, arterioles, and arteries) (Figures 1G and 1H).

Alterations in central lymphoid (thymus) and peripheral lymphoid organs (spleen and Peyer’s patches) were noted in rats treated with SCH 351591. On gross examination, the volume of the thymus and spleen decreased, whereas mesenteric lymph nodes and intestinal Peyer’s patches were enlarged compared with those in control rats. On microscopic examination, the cortex of the thymus became thinned and the T-cell zone was drastically reduced (Figures 2A and 2B). The TUNEL assay showed extensive apoptosis in the thymic cortex (Figures 2C and 2D). It was noteworthy that the number of MC and apoptotic lymphocytes of the cortex in the thymus increased with time and dose. In the spleen, the T-cell–dependent areas were also reduced (Figures 2E–2H). Microscopy also showed T-lymphocyte proliferation in mesenteric lymph nodes and at Peyer’s patches of the ileum (figures not shown).

Dose Response in Rats Treated with SCH 351591

Table 1presents the frequency and severity of overall lesion scores following treatment with SCH 351591 at a range of doses and times. Mesenteric vascular lesion scores increased in direct proportion to the administered doses over the range of 3 to 80 mkd. Rats treated with 160 mkd for one day had a lesion score of 1.7, which was comparable with the same score noted in the group that received 4.5 mkd for three days.

One rat died after two consecutive doses of 80 mkd SCH 351591. The remaining rats in the group were in moribund condition and were sacrificed less than three hours after the last dose. Two rats from the group treated with 9 mkd were removed from the study because of morbidity on day 6. Rats treated with 3 mkd SCH 351591 demonstrated mild body weight loss beginning on the third day and continuing to the ninth day (Figure 3). In rats treated with 9 mkd SCH 351591, body weight was dramatically lost from day 2 to day 9 (Figure 3). On gross examination, mesenteric vessel hemorrhage was detected in two of six rats in the 4.5 mkd group, in four of six rats in the 40 mkd group, and in three of six rats in the 80 mkd group.

On microscopic examination, in the 3 mkd group for three days, no acute vascular injury was found except for MC activation/degranulation and EC activation. In the 3 mkd group treated for nine days, activation of MC and EC was not seen, but activated MØ were noted. In addition, polyarteritis nodosa appeared at this time point.

In the 4.5 mkd group treated for three days, characteristic lesions of arteriolar and arterial hemorrhage and necrosis were obvious, with cellular alterations suggestive of cell activation and inflammatory cell responses. Among these activated cells were EC, MC, and MØ. In contrast to the small size and flat shape in saline-treated control rats, EC became larger in size and appeared cuboidal in shape in drug-treated rats. MC of the mesentery were larger and had more abundant cytoplasmic granules in the vicinity of the affected microvasculature in drug-treated rats as compared to control rats (Figures 4A and 4B). Some cytoplasmic granules were found outside of the cytoplasm of MC and the cell membranes were ruptured, suggesting MC degranulation. Immunostaining for CD 63 showed a positive reaction on MC (Figures 4C and 4D). Cytoplasmic vacuoles and digested red blood cells were apparent in MØ surrounding the affected vessels (Figure 4E), indicating MØ activation and hyperactive phagocytosis. Immunoperoxidase for nitrotyrosine showed a positive reaction on MØ (Figure 4F) and EC (Figure 4H) compared with a negative reaction in control rats (Figure 4G). TUNEL assay demonstrated no EC and VSMC apoptosis in the mesenteric arteries from control rats. In contrast, apoptosis was observed in EC and VSMC in mesenteric artery (Figure 4I) and/or arterioles (Figure 4J) obtained from SCH 351591-treated rats.

In the 9 mkd group for nine days, arterial hemorrhage and necrosis, perivascular inflammation, and polyarteritis nodosa were simultaneously found. Marked proliferation of fibroblasts (Figure 4K) was associated with polyarteritis nodosa and mingled with inflammatory cells (lymphocytes, macrophages, and leukocytes) and edematous fluids.

Time Course Response in Rats Treated with SCH 351591

In rats treated with 4.5 mkd for one to seventy-two hours, vascular injury of the mesentery progressed with time. By ninety-six hours, the lesion score declined (Table 2). Immune cell activation of the mesentery and mild inflammatory response appeared as early as one hour, and these alterations preceded characteristic injury of hemorrhage and necrosis. At twenty-four to forty-eight hours, the degree of immune cell activation and inflammatory response became moderate; however, no arterial or arteriolar hemorrhage and necrosis was found. It was not until seventy-two hours that characteristic arterial hemorrhage and necrosis occurred. By ninety-six hours, activation of immune cells and inflammatory response had subsided, but polyarteritis nodosa developed.

In rats treated with 20 mkd groups for one to seventy-two hours, arterial and arteriolar hemorrhage and necrosis was more severe at seventy-two hours than at any other time points (Table 3). As expected, the lesions were more severe at earlier times than with the 4.5 mkd dose. In addition, polyarteritis nodosa was not observed at ninety-six hours as it had been in rats treated with 4.5 mkd (Table 4).

Table 4illustrates the developmental process of mesenteric vascular injury. In rats treated with SCH 351591, the ensuing vascular injury was mediated by immune cell activation (MC, EC, MØ), followed by an early change in permeability (fibrin exudation and edema), initial recruitment of inflammatory cells from the blood into the perivascular spaces, irreversible vascular lesions (hemorrhage, necrosis, and apoptosis), and final formation of polyarteritis nodosa.

Comparison of Vascular Toxicity and Immunotoxicity between SCH 351591 and SCH 534385

The histopathological morphology of the vascular toxicity was found to be indistinguishable in rats treated with SCH 351591 and SCH 534385. Both drugs induced (a) similar or identical characteristics of lesions (microvascular injury, arterial hemorrhage and necrosis, and vascular inflammation) (Figures 5A–5D); (b) activation of immune cells (MC, EC, MØ); (c) preferred locations of the mesentery, kidney, liver, and small intestine in decreasing order; (d) a significant atrophy of the thymus and spleen and loss of T-cells in the cortex of the thymus (Figures 5E and 5F) and thymus-dependent T-cells zone in the spleen (Figures 5G and 5H) via apoptotic deletion pathway; and (e) comparable vascular lesion scores (a score of 4.0 in rats treated with 20 mkd of SCH 351591 for seventy-two hours versus a score of 3.5 in rats treated with 20 mkd SCH 534385 for seventy-two hours) (Table 5). However, dissimilarity of other aspects of drug toxicity was also observed. In rats treated with SCH 534385, on gross examination, the small intestine appeared to be jelly-like and semi-transparent, the intestinal lumen was remarkably distended, the intestinal walls were thinned, and punctuate hemorrhage was readily recognized. On microscopic examination of the small intestine, fibrinoid necrosis of arterioles, severe inflammation, and vasodilatation of the muscular coat (the muscularis) were observed in SCH 534385-treated rats. In addition, on the outside of the small intestine, Peyer’s patches protruded abnormally. On microscopic examination, reactive hyperplasia of T-lymphocytes within Peyer’s patches was observed.

Discussion

Pathological Characteristics of SCH 351591- or SCH 534385-induced Vascular Injury in Rats

The present study demonstrated that SCH 351591 induced time- and dose-dependent mesenteric vascular injury in SD rats. Similar to PDE IV-inhibitor–induced vascular injury in cynomolgus monkeys (Losco et al. 2004), drug-treated rats showed a time course of injury from early microvascular (insudation and exudation) followed by arterial hemorrhage and fibrinoid necrosis, mainly in small- to large-sized mesenteric blood vessels and to a lesser extent in the pancreatic vessels; inflammatory infiltration within blood vessel walls; and edema in mesenteric connective tissue. In addition to necrosis and hemorrhage, another arterial VSMC alteration consisted of fibrin deposits in the walls. Polyarteritis nodosa was a late-stage event seen in some animals. This lesion has been reported as a late-appearing finding in studies with PDE III inhibitors and with nonspecific PDE inhibitors (Hanton et al. 1995).

SCH 534384, another second PDE IV inhibitor that has not been reported previously to induce DIVI in rats, resulted in vascular injury and lymphoid organ alterations similar to SCH 351591, suggesting that they may share the same mechanism of drug-induced toxicity. Both PDE IV inhibitors (SCH 351591 and SCH 534385) have similar vascular injury patterns to another PDE IV inhibitor, BYK169171 (Mecklenburg et al. 2006).

Activation of MC in Association with Vascular Injury

In the rat, peritoneal MC are widely distributed in the mesentery, closely associated with the microvasculature, and the number of MC is known to be higher than in other organs. Therefore, MC-released proinflammatory mediators and cytokines may directly lead to microvascular injury, followed by induction and arterial hemorrhage. However, the effect of PDE IV inhibitors on MC is controversial in the literature (Dent and Giembycz 1996; Souness et al. 2000). Rolipram, a prototypical PDE IV inhibitor, has been demonstrated to inhibit antigen-induced histamine release and leukotriene C₄production from mouse bone-marrow–derived MC (Torphy and Undem 1991) and to suppress MC degranulation in the guinea pig (Underwood et al. 1997). However, rolipram has also been shown to have a noninhibitory effect on antigen-induced histamine release from rat peritoneal MC (Frossard et al. 1981) and on the IgE-mediated release of histamine from human lung MC (Weston et al. 1997). At clinically relevant concentrations (1 or 10 μmol/L), rolipram had an insignificant effect on the IgE-mediated release of histamine, sulfidoleukotrienes, and prostaglandin D₂from cultured human MC, but at a concentration of 100 μmol/L, it significantly suppressed the release of these three mediators (Shichijo et al. 1998). The reason for the discrepancy is not obvious, but it may be related to the sources of MC, which were derived from different tissues and species (Souness et al. 2000), as well as the concentrations of PDE IV inhibitor used (Shichijo et al. 1998).

The present study, however, revealed that SCH 351591 and SCH 534385 induced MC degranulation at sites of inflammation, indicating the release of pro-inflammatory mediators from MC. Numerous MC were associated with the affected vessels of the mesentery, particularly with the microvasculature. Morphological changes in MC were conspicuous, including an increase in the cell size and in the cytoplasmic granules (indices of MC activation), as well as extrusion of cytoplasmic granules and rupture of cell membranes (indices of MC degranulation). Many MC in drug-treated rats were positively stained with a CD 63 mAb, which recognizes the AD1 antigen as a specific surface marker of MC (Boros et al. 1995; Nishikata et al. 1992). This protein is thought to be a receptor for the tissue inhibitor of metalloproteinase-1 (TIMP-1) (Jung et al. 2006), which was detected at a high level in serum following treatment with SCH 351591 (Weaver et al. 2008) and with another PDE IV inhibitor, CI-1044 (Daguès, Pawlowski, Guigon et al. 2007).

Activation of EC in Association with Vascular Injury

Morphological changes corresponding to the microscopic features of EC activation were also observed in rats treated with SCH 351591 or SCH 534385. The alterations of activated EC were characterized by protrusion of EC into the lumen of blood vessels, as well as hypertrophy (plump cuboidal appearance). In our previous studies with the PDE III inhibitor SK&F 956454, we reported several ultrastructural alterations of activated EC, including increased biosynthetic organelles (Golgi complex, rough endoplasmic reticulum, and ribosomes) and increased permeability (pinocytotic vesicles) (Zhang et al. 2002). Activated EC can increase secretion of vasoactive mediators, heighten vascular permeability (Ballermann 1998; Cotran 1989), allow accumulation of reactive oxygen species and activate a cascade of proteases (Bach et al. 1997), as well as induce the expression of adhesion molecules (E-selectin), chemokines (IL-8), and the pro-coagulant factor tissue factor (Bach et al. 1997). Uncontrolled EC activation can lead to loss by apoptosis (Bach et al. 1997), which was observed in the present study. It has been reported that PDE IV is the dominant PDE isozyme in rat pulmonary microvascular EC, where it most likely plays a major role in the regulation of cAMP metabolism (Thompson et al. 2002).

Activation of MØ in Association with Vascular Injury

At the site of inflammation, the number and size of MØ was increased. Many MØ displayed evidence of activation, such as phagocytosis of erythrocytes and other materials. It has been reported that MØ can be activated to synthesize many factors, such as fibroblast-activating factor and basic fibroblast growth factor (bFGF) (Ikegami et al. 2002). This is one likely explanation for the proliferation of fibroblasts composing the polyarteritis nodosa observed in rats treated with either of the two PDE IV inhibitors. MC have also been reported to be a major source of bFGF in chronic inflammation (Qu et al. 1995). Thus, activation of MØ and MC may contribute to fibroblast proliferation at a late stage of vascular injury.

Peroxynitrite Formation Involved in PDE IV-inhibitor–induced Vascular Injury and Inflammation

Peroxynitrite, a potent biologic oxidant, is formed by a reaction of NO with superoxide anions (Viera et al. 1999). The term peroxynitrite actually contains the anionic form (ONOO⁻) (peroxynitrite anion) and the protonated form (ONOOH) (peroxynitrous acid), and both forms can perform direct oxidative reactions with target molecules (Denicola and Radi 2005). The greater toxicity and reactivity of peroxynitrite relative to NO and superoxide (₂O ^°) was reviewed very recently (Pacher et al. 2007). Peroxynitrite formation can initiate oxidative DNA damage and DNA strand breaks, leading to peroxynitrite-triggered apoptosis in EC and VSMC (Mihm et al. 2000; Li et al. 2004; Dickhout et al. 2005; Pacher et al. 2007). Much interest has been centered very recently on the key role for locally produced peroxynitrite in the pathophysiology of inflammation (Pacher et al. 2007). Increased peroxynitrite formation was reported to contribute to the increased EC expression of vascular cellular adhesion molecules (VCAM, P-selectin, and E-selectin) and to enhance interaction between neutrophils and EC (Sohn et al. 2003). In particular, peroxynitrite-induced L-selectin shedding and upregulation of CD11b/CD18 expression on neutrophils and peroxynitrite-induced slight increases in E-selectin and ICAM-1 expression on human coronary artery EC have been observed (Zhao et al. 2004; Zouki et al. 2001). Both activated EC and MØ expressed nitrotyrosine in the present study. Peroxynitrite from activated EC may make a primary contribution to inflammation, whereas peroxynitrite from activated MØ may be a secondary contributor to the pathological processes of developing vascular inflammation. It appeared in the present study that at least three types of activated cells (EC, MC, and MØ) and other leukocytes initiate, amplify, and finally execute vascular inflammation. The pathogenesis of vascular inflammation may involve the early activation of EC and MC, followed by activation/degranulation of eosinophils, myeloperoxidase activity of neutrophils and peroxynitrite formation of MØ, and eventually fibroblast proliferation within the vessel wall or in the vicinity of injured artery. Significant evidence of reactive oxygen species has been inferred in the mRNA responses to treatment with the PDE IV inhibitor CI-1044 (Daguès, Pawlowski, Sobry et al. 2007).

Morphological and Cellular Basis of Biomarkers for PDE IV-inhibitor–induced Vascular Injury

It is of interest to determine if activation of EC, MC, and MØ could provide useful clues for searching candidate bio-markers of PDE IV-induced vascular injury. It has been shown that mouse and human MC secrete vascular endothelial growth factor (VEGF), a vascular permeability factor, upon specific immunological activation, thus increasing vascular permeability to small and large molecules (Boesiger et al. 1998). VEGF, a MC-derived multifunctional cytokine, plays a particularly important role in microvascular hyperpermeability and subsequent plasma protein extravasation (Boesiger et al. 1998; Dvorak et al. 1995). It has been demonstrated that bFGF is located in both the cytoplasmic granules and extruded granules of MC, and thus bFGF can be released via MC degranulation (Qu et al. 1998). bFGF, a MC-derived polypetide, has been regarded as an important factor for fibroblast proliferation in chronic inflammation (Qu et al. 1995). Extracellular proteinase-related proteins (TIMP-1) may also be produced by activated MØ, venular SVMC, and fibroblasts. Consistent with roles of peroxynitrite formation in nitrosative stress, evidence of chemokines, cytokines, acute phase proteins, and dose- and time-dependent increases in serum levels of nitrite has been identified in the same study, part II (Weaver et al. 2008).

Finally, the present study revealed that PDE IV inhibitors induced apoptosis of T-cells and atrophy of the thymus and spleen in drug-treated rats. These findings suggest that T-lymphocytes were deleted in the thymic cortex and the thymic-dependent T-cell zone of the spleen via apoptotic pathways. The restriction of cell loss to T-cell areas is not consistent with stress-associated responses (Pruett et al. 2000). A detailed discussion of the suppressive effects of PDE IV inhibitors on the immune system is beyond the scope of this paper.

In conclusion, this study has shown that two PDE IV inhibitors cause drug-induced vascular injury in both a dose-and time-dependent manner. The lesions are similar to those caused by other PDE IV inhibitors. The data provide a strong foundation to evaluate candidate biomarkers that could be used in nonclinical studies and, potentially, to clinically evaluate the safety of drugs in these classes.