PDGFR Inhibition Results in Pericyte Depletion and ... - SAGE Journals

Original Article

PDGFR Inhibition Results in Pericyte Depletion and Hemorrhage into the Corpus Luteum of the Rat Ovary

Toxicologic Pathology 2016, Vol. 44(1) 98-111 ª The Author(s) 2015 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0192623315613452 tpx.sagepub.com

Anthony P. Hall1, Susan Ashton2, Judith Horner1, Zena Wilson2, Jaimini Reens1, Graham H. P. Richmond2, Simon T. Barry2, and Steve R. Wedge2

Abstract The growth plate, ovary, adrenal gland, and rodent incisor tooth are sentinel organs for antiangiogenic effects since they respond reliably, quantitatively, and sensitively to inhibition of the vascular endothelial growth factor receptor (VEGFR). Here we report that treatment of rats with platelet-derived growth factor receptor beta (PDGFRb) inhibitors that target pericytes results in severe ovarian hemorrhage with degeneration and eventual rupture of the corpus luteum. Evaluation of the growth plate, adrenal gland, and incisor tooth that are typical target organs for antiangiogenic treatment in the rodent revealed no abnormalities. Histologically, the changes in the ovary were characterized by sinusoidal dilatation, increased vessel fragility, and hemorrhage into the corpus luteum. Immunocytochemical staining of vessels with alpha smooth muscle actin and CD31 that recognize pericytes and vascular endothelium, respectively, demonstrated that this effect was due to selective pericyte deficiency within corpora lutea. Further experiments in which rats were treated concurrently with both PDGFRb and VEGFR inhibitors ablated the hemorrhagic response, resulting instead in corpus luteum necrosis. These changes are consistent with the notion that selective pericyte loss in the primitive capillary network resulted in increased vessel fragility and hemorrhage, whereas concomitant VEGFR inhibition resulted in vessel regression and reduced vascular perfusion that restricted development of the hemorrhagic vessels. These results also highlight the utility of the rodent ovary to respond differentially to VEGFR and PDGFR inhibitors, which may provide useful information during routine safety assessment for determining target organ toxicity. Keywords platelet-derived growth factor receptor (PDGFR), vascular endothelial growth factor receptor (VEGFR), angiogenesis, ovary, corpus luteum, pericyte, endothelial cell

Introduction Angiogenesis, the de novo production of new blood vessels from a preexisting vascular bed, is essential for a number of physiological processes such as the development of an organism (Carmeliet et al. 1996; Ferrara et al. 1996), the maintenance of angiogenic-dependent tissues, and wound healing (Howdieshell et al. 2001). It also drives a variety of pathological conditions such as cancer (Hanahan and Weinberg 2000) and age-related wet macular degeneration. In this latter disease, treatment strategies aimed at inhibition of angiogenesis limit further vision loss or in some patients may even restore lost vision (Augustin, Scholl, and Kirchhof 2009). Treatment with angiogenic inhibitors, however, may be accompanied by adverse effects, which in the context of preclinical safety assessment may manifest as target organ pathology. In the growing rodent, for example, treatment with vascular endothelial growth factor receptor (VEGFR) inhibitors results in a constellation of target organ toxicity in angiogenic-dependent tissues such as the growth plate, ovary, incisor

tooth, and adrenal gland characterized by growth plate dysplasia (Gerber et al. 1999; Wedge et al. 2005; Hall, Westwood, and Wadsworth 2006), ovarian (and uterine) atrophy (Ferrara et al. 1998; Wulff et al. 2002; Wedge et al. 2005), incisor tooth dental dysplasia (Patyna et al. 2008; Fletcher et al. 2010; Hall, Westwood, and Wadsworth 2006), and adrenal gland cortical necrosis and hemorrhage (Patyna et al. 2008). Taken together, this pattern of target organ toxicity may be regarded as a preclinical signature of antiangiogenic effect (Hall 2005). In the ovary, the development of the corpus luteum from the avascular Graafian follicle is a highly regulated process that 1

AstraZeneca, Drug Safety and Metabolism, Alderley Park, Macclesfield, Cheshire, UK 2 AstraZeneca, Oncology iMed, Alderley Park, Macclesfield, Cheshire, UK Corresponding Author: Anthony P. Hall, Department of Pathology, Safety Assessment UK, AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK. Email: [email protected]

Hall et al. involves the coordinated recruitment and migration of both endothelial cells and pericytes (reviewed in Robinson et al. 2009). The principal growth factors driving this process are vascular endothelial growth factor (VEGF) and plateletderived growth factor (PDGF) whose cognate receptors are expressed on endothelial cells and pericytes, respectively (Hall 2006). In the rodent ovary, vascularization of the newly formed corpus luteum occurs within a few (typically 4) days and represents an intense period of angiogenesis. In the primate, it has been shown that more than 80% of the dividing cells in the corpus luteum are endothelial cells (Wulff et al. 2001). Antiangiogenic therapy, and especially anti-VEGF therapy, is an important treatment modality. This has been underscored by the successful approval in the past 10 years of a number of VEGF inhibitors for the treatment of cancer including bevacizumab, sorafenib, and vandetanib (reviewed in Lodish 2013) and the approval of ranibizumab (Lucentis) for the treatment of the wet form of macular degeneration (Augustin, Scholl, and Kirchhof 2009). The development of VEGF signaling inhibitors was accompanied by an interest in platelet-derived growth factor receptor (PDGFR) inhibitors since it was believed that targeting mural cells (pericytes and vascular smooth muscle) may destabilize mature vessels and render them more susceptible to antiangiogenic therapy (Erber et al. 2004). In order to explore this possibility, we identified several small molecule PDGFR inhibitors for the treatment of cancer in human beings. During the preclinical safety testing of these new molecules, we noted that PDGFR inhibitors induced a unique profile of toxicity that was quite distinct from the effect typically seen with VEGFR inhibitors. In this article, we describe the preclinical effects of these compounds on the key target (angiogenic dependent) organs. Using the ovary as a model for angiogenesis, we examined the effects of cotreatment with VEGFR and PDGFR inhibitors in order to better understand the mechanism of toxicity. Finally, we imaged pericytes and endothelial cells immunocytochemically in order to determine the effect of PDGFR inhibition on pericyte and endothelial cell recruitment into the developing corpus luteum.

Material and Method Test Compounds Four test compounds were synthesized (AZ12585313—a PDGFR inhibitor, AZ12253678—a pan class III receptor tyrosine kinase [RTK] inhibitor, AZD2171 [Cediranib]—a VEGFR inhibitor, and AZD2932—a VEGFR and PDGFR inhibitor) by AstraZeneca Research and Development, Alderley Park, U.K., for oral gavage dosing of rats at a dose volume of 5 ml/kg unless otherwise stated. AZ12585313 was prepared as a suspension in 0.5% w/v hydroxypropyl methylcellulose and 0.1% w/v polysorbate 80 to achieve a final dose of 2 mg/kg QD–25 mg/kg BID. AZ12253678 was prepared as a suspension in water containing either 30% w/v Capitsol in 40% v/v polyethylene glycol 400 or 0.5% w/v hydroxypropyl methylcellulose and 0.1% w/v polysorbate 80 to achieve a final dose of

99 Table 1. Summary of Experiments. Study Details

Compound

Specificity

Study 1 9-Day discovery study in AZ12585313 HW rat 17-Day discovery efficacy AZ12585313 study in Nu rat Study 2 1-Month study in AP rat AZD12253678

PDGFRa and b PDGFRa and b Pan class III PDGFR Study 3 9-Day discovery study in AZ12585313 þ PDGFRa and HW Rat AZD2171 b þ VEGFR Study 4 1-Month study in HW Rat AZD2932 PDGFRa and b þ VEGFR Note. VEGFR ¼ vascular endothelial growth factor receptor; PDGFRa ¼ platelet-derived growth factor receptor alpha; PDGFRb ¼ platelet-derived growth factor receptor beta; AP rat ¼ female Alpk:APfSD; HW rat ¼ Wistar Hannover rat; Nu rat ¼ nude rat.

2.55–38.3 mg/kg QD. AZD2932 was prepared as a suspension in water containing 0.5% w/v hydroxypropyl methylcellulose and 0.1% w/v polysorbate 80 to achieve a final dose of 4.92–49.2 mg/kg/QD.

Animals and Treatment All studies were performed in accordance with the standards of animal care and ethics described in ‘‘Guidance on the Operations of the Animals (Scientific Procedures) Act 1986’’ issued by the U.K. Home Offices, so that any clinical expression of toxicity remained within a moderate severity limit as described by the guidelines agreed with the U.K. Home Office inspector. Female Wistar Hannover rats (HW rats), substrain Alpk HsdBrlHan.WIST (Alderley Park Breeding Unit), or female Alpk:APfSD (Wistar derived; AP rat) were used for safety studies, whereas female nude rats, strain Hsd:RH-Foxn1rnu (Nu rats; Harlan laboratories, UK Ltd, Blackthorn, Bicester, Oxfordshire, OX25 1TP), were used for efficacy studies. Rats were group housed appropriate for each study and were acclimatized for at least 6 days before the treatment was started. Animal rooms were illuminated in a 12-hr light/dark cycle, and temperature and humidity were controlled within the limits of 21 C + 2 C and 55% relative humidity (RH) + 15% RH. Pelleted RM1 (E) SQC rodent diet and drinking water were freely available. The animals used were within an age range of 6–10 weeks at the start of dosing.

Study Design Discovery studies. All discovery safety studies were performed in the HW rat unless otherwise stated, using 2–5 female animals per group (summarized in Tables 1–7). Animals were dosed with compound by oral gavage for 9 days. AZ12585313 was dosed at 2 mg/kg QD–25 mg/kg BID, whereas the combination of AZ12585313 and AZD2171 was dosed at 25 mg/kg QD and 2.5 mg/kg QD, respectively. In the nude rat efficacy study, animals were dosed with AZ12585313 at 12.5–25 mg/kg QD for 17 days.

100

Toxicologic Pathology 44(1)

Table 2. Compound Specificity Cell Phosphorylation IC50 Values (μM).

pPDGFRa pPDGFRb pKDR pc-kit pflt-3 pCSF-1R

AZ12585313

AZ12253678

AZD2171

AZD2932

0.003 0.004 >10 0.018 >1 0.200

0.015 0.004 0.283 0.006 0.007 0.001

0.010 0.006 0.002 0.003 >1 0.625

0.007 0.004 0.007 0.007 0.007 0.140

Note. pPDGFRa ¼ phospo-platelet-derived growth factor receptor alpha; pPDGFRb ¼ phospo-platelet-derived growth factor receptor beta; pKDR ¼ phospho-KDR; pc-kit ¼ phospho-c-kit; pflt-3 ¼ phospho-flt-3; pCSF-1R ¼ phospho-CSF-1R.

1-month safety studies. All 1-month studies were performed to a standard format of either 3 or 4 groups of female HW or AP rats consisting of 5–9 animals per group per study (summarized in Tables 1–7). Animals were dosed with compound by oral gavage at concentrations appropriate to each study. AZD2932 was dosed at 4.92 mg/kg QD for 28 days or 49.2 mg/kg QD for 21 days, and AZ12253678 was dosed at 2.55–38.3 mg/kg QD for 28 days. Clinical signs were recorded postdosing and a physical examination was performed at least once weekly.

Necropsy and Histology On the scheduled day of necropsy, animals were sacrificed by administration of halothane anesthesia or carbon dioxide and then exsanguinated. Animals were necropsied according to the standard operating procedures (SOPs) where any macroscopic abnormalities were recorded. Depending upon study design, a limited number of tissues were harvested, principally to include the female reproductive tract, incisor tooth, and femorotibial joint in order to assess the effects of these compounds on angiogenesis. Tissues were fixed in buffered 10% formalin, processed to wax blocks, and then sectioned and stained with hematoxylin and eosin (H&E) for examination by light microscopy according to the standard SOPs. H&E slides were examined, and pathology data were recorded in a good laboratory practice (GLP)-compliant manner using an electronic data capture system (PLACES) for the 1-month studies.

Immunohistochemistry Four-micrometer thick sections were cut and mounted on SuperFrost Plus1 (Thermo Fisher Scientific, Runcorn, UK) electrostatically charged glass slides. Sections were allowed to dry at 37 C overnight in an incubator. For immunocytochemical staining of alpha smooth muscle actin (aSMA; for the demonstration of pericytes), sections were dewaxed in xylene and rehydrated through graded alcohols to water. Sections were placed onto a Labvision Immunostainer (Thermo Fisher Scientific, UK) where the remainder of the immunocytochemical staining was performed. Sections were first washed in Tris-buffered saline with 0.1% Tween

(TBST) followed by blockade of endogenous peroxidase with 3% hydrogen peroxide in TBST for 10 min. After a buffer wash, nonspecific Ig-binding sites were blocked for 10 min using a background blocker with casein (A.menarini diagnostics, Berkshire, UK). This was followed by primary antibody incubation with a monoclonal mouse anti-aSMA (1:1,000, Sigma, Gillingham, UK; cat# A2547) for 30 min. Following another wash in TBST, sections were then incubated with a ready-to-use peroxidase-labeled secondary reagent, Mouse EnVision-HRP (K4007; Dako, Ely, UK) for 30 min. To allow visualization of aSMA, sections were incubated with 3,30 -diaminobenzidine (A.menarini Diagnostics) and then counterstained with Carazzi’s hematoxylin. All dilutions were in TBST and procedures performed at room temperature. For immunocytochemical staining of CD31 (Pecam-1; for demonstration of endothelial cells), sections were placed onto the Ventana Discovery XT Staining Module. All steps were performed on this staining platform using Ventana Medical Systems, Inc. (Tucson, Arizona, UK) validated reagents, including deparaffinization, antigen retrieval, antibody incubation, and detection. Briefly, a standard cell conditioning 1 (CC1) heatinduced epitope retrieval protocol was applied. Sections were next incubated with a rabbit polyclonal anti-rat CD31 (1:200, cat# 250590; Abbiotec, York, UK) for 60 min followed by the OmniMap anti-rabbit HRP detection system (cat# 760-4311; Ventana Medical Systems Inc.) for 32 min. A counterstain procedure using hematoxylin (cat#760-2021; Ventana Medical Systems Inc.) was followed. All stained sections were dehydrated, cleared, and mounted. Appropriate negative and positive controls were utilized.

Kinase Selectivity Cell receptor phosphorylation was used as the key measure of potency. AZ12585313 inhibited the phosphorylation of PDGFR alpha (PDGFRa), PDGFR beta (PDGFRb), and c-Kit (Table 2). In brief, cells were placed in low serum or serumstarved conditions for 16–24 hr. Compounds were serially diluted in DMSO for a concentration range and added to the cells. The cells were then incubated for 30–120 min. Where required, cell lines were stimulated with the appropriate ligand to induce receptor phosphorylation. After 5–10 min, the cells were either lysed or fixed and processed using in-house receptor-specific ELISA, IGEN, or PACE assays.

Results Overview of Experiments and Compound Specificity A number of small molecule tyrosine kinase inhibitors were identified (Table 1) with high affinity for the PDGFR receptor (Table 2). AZ12585313 was developed as a specific high-affinity PDGFR inhibitor which showed approximately equipotent activity at both PDGFRa and PDGFRb receptors and approximately 5-fold less activity at the c-Kit receptor (Ple, Jung, Ashton, Hennequin, Laine, Lambert-van der Brempt, et al. 2012).

Hall et al.

101

Table 3. Effects of AZ12585313 on the Ovary and Adrenal Gland in the HW Rat. Corpus Luteum—Cystic Hemorrhagic Degeneration Number of Animals Dose (mg/kg) per Group Min 2 QD 1 BID 3 QD 6 QD 12.5 QD 12.5 QD 25 QD 6 BID 12.5 BID 25 BID

5 5 5 5 5 2 2 5 5 5

Corpus Luteum—Cystic Dilatation Adrenal Glands—Cortical Hemocysts

Mild

Mod

Severe

Min

Min

— — — — — — 1 — — b 2

— — — — — — — — — —

— — — — — — — — — 3

— — — 1 — — — — — —

— — — — — — — — — 1

— — 1 — 4 1 — 4 a 5 —

Note. HW rat ¼ Wistar Hannover rat. a Two animals affected unilaterally. b One animal affected unilaterally.

Table 4. Effects of AZ12585313 on the Ovary and Incisor Tooth in the Nude Rat.

Dose (mg/kg) 12.5 QD 25 QD 12.5 BID

Number of Animals per Duration Group (Days) 5 5 5

17 17 17

Cystic Hemorrhagic Degeneration

Corpus Luteum Necrosis

Min

Mild

0 1 4

1

AZ12253678 was developed as a pan class III RTK inhibitor, showing high affinity for both the PDGFRa and b, Flt3, c-Kit, and CSF1R receptors but lacked VEGFR2 (KDR) activity. AZD2932 was developed as another mixed kinase inhibitor that showed approximately equal activity at the PDGFR, VEGFR2, c-Kit, and Flt-3 receptors (Ple, Jung, Ashton, Hennequin, Laine, Morgentin, et al. 2012). All 3 molecules were evaluated in a number of early preclinical safety and efficacy studies (summarized in Table 1) in order to evaluate the antiangiogenic effects of PDGFR inhibition alone versus combined PDGFR and VEGFR inhibition. The results of these experiments have been summarized in Tables 3–7.

Study 1: Antivascular Effects of AZ12585313 (PDGFR␣ and b inhibitor) Ovary In order to investigate the antivascular effects of PDGFR inhibition in vivo, HW rats were treated with AZ12585313 (PDGFRa and b inhibitor) at 2 mg/kg QD–25 mg/kg BID for 9 days (Table 3). Histological assessment was focused primarily on tissues that were known to be responsive to angiogenic inhibition in the rodent, namely, the ovary, adrenal glands,

epiphyseal growth plate, and incisor teeth. In these tissues, treatment with VEGFR inhibitors had previously been shown to induce reliable signature changes of antiangiogenesis characterized by growth plate dysplasia (Gerber et al. 1999; Wedge et al. 2005; Hall, Westwood, and Wadsworth 2006), ovarian (and uterine) atrophy (Ferrara et al. 1998; Wulff et al. 2002; Wedge et al. 2005), incisor tooth dental dysplasia (Patyna et al. 2008; Fletcher et al. 2010; Hall, Westwood, and Wadsworth 2006), and adrenal gland cortical necrosis and hemorrhage (Patyna et al. 2008). Treatment with a PDGFRb inhibitor was therefore expected to induce a similar anti-angiogenic effect, however, in the ovary, treatment with AZ12585313 resulted in a profound anti-vascular effect consisting of a dose-responsive and treatment-related minimal to severe cystic haemorrhagic dilatation/degeneration of the corpus luteum (Table 3). This change appeared grossly as a nodular enlargement of the ovary characterised by abnormal multilocular areas of red discolouration (which corresponded to enlarged corpora lutea) (Figure 1A). In the most severely affected animals, the ovary was grossly enlarged and discoloured (Figure 1A lower inset). Cystic enlargement correlated microscopically with degeneration, compression, necrosis and rupture of the corpus luteum leading to loss of the ovarian cortex and haemorrhage into the interstitial space (Figure 1B). In less severely affected animals changes were characterized by single cell necrosis of luteal cells which evolved into cystic dilatation and enlargement with haemorrhage into the central lumen (Figure 1C). In newly formed, basophilic corpora lutea, the earliest morphological change was characterised by dilatation of fine walled capillaries/sinusoids, some of which appeared to have a generalized reduction in the numbers of spindle-shaped interstitial cells (later confirmed as pericytes) (Figure 1D). In ovaries that were not severely affected, residual normal appearing corpora lutea were present. Finally in one animal, treatment with AZ12585313 at 25 mg/kg/day BID resulted in atrophy of the uterus and vagina consistent with anoestrus.

102


Table 5. Effects of AZ12253678 on the Ovary, Growth Plate, Adrenal, and Incisor Tooth in Female AP Rats. Ovary—Cystic Hemorrhagic Degeneration of Corpora Ovary— Femur—Metaphyseal Adrenal Glands—Cortical Number of Lutea Necrosis Hyperostosis Hemocysts Animals Min Min Mild Mod Severe Dose (mg/kg) per Group Min Mild Mod Severe Mild Mod 2.55 QD 25.5 QD 25.5 QD 38.25 QD

6 3 4 4

1

1

4 3 4 4

1 3

2 3 3

Incisor—Dental Dysplasia Min

Mild

Mod

1 1 0

1 1 4

1

2 2 1

Note. AP rat ¼ female Alpk:APfSD.

Table 6. Effects of AZ12585313and AZD2171 in Female HW Rats.

Dose (mg/kg) AZ12585313 at 25 mg/kg/BID and AZD2171 at 2.5 mg/kg/QD

Ovary—Corpus Luteum Number of Necrosis Animals per Min Mild Group 2

1

1

Ovary— Cystic Corpus Luteum

minimal to severe cystic hemorrhagic degeneration of the ovarian corpus luteum (Table 5; Figure 1 supplement) that appeared identical to the change induced by AZ12585313. Again this was accompanied by mild to severe hemorrhage, but additionally in 4 animals, mild to moderate cortical necrosis was observed following the rupture of degenerate corpora lutea.

Mild 1

Note. HW rat ¼ Wistar Hannover rat.

Femorotibial Joint, Adrenals, and Incisor In the femorotibial joint and adrenal glands, treatment with AZ12585313 resulted in minimal or no antiangiogenic effects. Histological examination of the femorotibial joint revealed no significant effects, whereas the adrenals exhibited minimal antiangiogenic effects characterized by minimal adrenal cortical hemocysts in a single animal treated at 25mg/kg BID. To further investigate the effects of AZ12585313, we decided to examine the incisor teeth from xenograft-bearing nude rats which had been treated with AZ12585313 at 12.5 mg/kg QD, 12.5 mg/kg BID, and 25 mg/kg QD for 17 days in order to determine drug efficacy. The results from these experiments showed minimal cystic degeneration of the corpus luteum at 12.5 mg/kg BID (which correlated with that seen in HW rats treated at the same dosage) accompanied by mild necrosis of the corpus luteum in 1 animal but no other effects (Table 4).

Study 2: Antivascular Effects of AZ12253678—A Pan Class III Inhibitor Ovary In order to further evaluate the antiangiogenic effects of PDGFR inhibition, female AP rats were treated with a second PDGFR inhibitor, AZ12253678, which had high affinity for PDGFRa, PDGFRb, Flt3, c-Kit, and CSF1R but significantly lacked VEGFR2 activity. Treatment at 2.55–38.3 mg/kg QD for 28 days resulted in a treatment-related and dose-responsive

Femur, Adrenal Glands, and Incisor Teeth Examination of the femur, adrenal glands, and incisor teeth, however, revealed differences in the response to AZ12253678 (pan class III RTK inhibitor) compared to AZ12585313. In line with a minor antiangiogenic effect, treatment with AZ12253678 (pan class III RTK inhibitor) resulted in minimal growth plate dysplasia in the femur and minimal cortical hemocysts in the adrenal glands. Interestingly, examination of the incisor tooth revealed minimal to moderate dental dysplasia (Table 5). Changes in the femur presented as increased metaphyseal bone mass (hyperostosis) characterized by increased thickness of predominantly secondary trabeculae. No significant differences were noted in the growth plate hypertrophic zone. Adrenal hemocysts were characterized by degeneration, necrosis, and loss of zona fasciculata cells often accompanied by a minimal inflammatory cell infiltrate with pooling of red blood cells in dead space. Finally, dental dysplasia noted in the incisor teeth was characterized by disorganization and loss of odontoblastic epithelium with secondary abnormal deposition of degenerate dentine (Figure 2). Molar teeth, which do not need a continuously growing vasculature, were normal.

Study 3: Antiangiogenic Effects of Cotreatment with a PDGFR and VEGFR Inhibitor We hypothesized that the antivascular effect of AZ12585313 in the ovary was due to the pharmacological inhibition of PDGFRb in pericytes and that mural cell recruitment was critical for normal vascularization of the corpus luteum. We therefore decided to investigate this mechanism of toxicity by co-dosing rats with AZ12585313 and AZD2171 (cediranib), which are specific inhibitors of the PDGFR and VEGFR receptors, respectively (Brave et al. 2011). Two HW rats were treated with AZ12585313 at 25 mg/kg BID and AZD2171 at

5 1 1 1 1

6

Moderate

1

Severe

2 1 3 3 1 6 6 4.92 QD 49.2 QD

Note. HW rat ¼ Wistar Hannover rat. a These groups received the polyethylene glycol/Capitsol formulation b These groups will received the 0.5% w/v HPMC/0.1% w/v polysorbate 80 formulation

1

1

2

1

Min Mild Moderate Severe Moderate Mild Moderate a

b

Min

2.5 mg/kg QD for 9 days. These dosages were chosen because AZ12585313 had previously been shown to produce mild to severe ovarian cystic hemorrhagic degeneration of the corpus luteum in all treated animals (Table 1). Similarly, AZD2171 had previously been shown to be potently antiangiogenic in mice and rats treated at 0.75–6 mg/kg QD (Wedge et al. 2005), resulting in epiphyseal growth plate dysplasia and ovarian atrophy. The results of this experiment showed that the hemorrhagic phenotype noted with AZ12585313 (PDGFR inhibitor) monotherapy was completely ablated with AZD2171 co-therapy (VEGFR inhibitor) and replaced by minimal to mild degeneration of the ovarian corpus luteum characterized by avascular coagulative necrosis of the central zone of the corpus luteum accompanied by an almost complete absence of spindle-shaped interstitial cells (endothelial cells/pericytes; Table 6).

Study 4: Antiangiogenic Effects of AZD2932 (VEGFR and PDGFR Inhibitor)

Mild Number of Animals per Group Min Dose (mg/kg)

Femur—Growth Plate Dysplasia Adrenal Glands— Hemorrhagic Necrosis Adrenal Glands—Cortical Hemocysts Ovary—Increased Follicles Ovary—Reduced Corpora Lutea

Table 7. Effects of AZD2932 on the Ovary, Adrenal Gland, Femur, and Incisor in Female HW Rats.

Min Mild Moderate

103

Incisor Teeth— Dental Dysplasia

Hall et al.

In order to further validate the effects of combined PDGFR and VEGFR inhibition on the ovary of the rat, we decided to investigate the effects of a previously developed molecule, AZD2932, which showed mixed activity at the PDGFR and VEGFR2 receptors (approximately 1:1 inhibition of PDGFRb and VEGFR2). HW rats were treated with AZD2932 at 4.92 mg/kg QD for 28 days and 49.2 mg/kg QD for 21 days, and the ovary, femur, adrenal glands, and incisor tooth were examined histologically in order to validate the effects of combination of PDGFR/ VEGFR inhibition. The results of this analysis confirmed the effects previously noted with the AZ12585313 (PDGFR)/ AZD2171 (VEGFR) combination study (i.e., ablation of the hemorrhagic phenotype induced by PDGFR inhibitor monotherapy) and were consistent with dominance of the VEGFR effect, that is, reduced numbers of corpora lutea secondary to reduced vascular perfusion and vessel regression (Table 7). Detailed histological analysis of the ovary revealed changes consistent with ovarian atrophy characterized by a minimal to moderate reduction in the size and number of corpora lutea. Residual corpora lutea appeared abnormal with areas of single-cell necrosis and coagulative central necrosis, occasional areas of mineralization of necrotic luteal cells, and a generalized reduction in interstitial cells (pericytes and/or endothelial cells; Figure 3). No animals were observed with cystic hemorrhagic degeneration of the corpora lutea. In the adrenal glands, minimal to severe cortical hemocysts were noted together with severe necrosis in 1 animal. In the femorotibial joint, minimal to moderate epiphyseal growth plate dysplasia was noted, which was characterized by the thickening of the growth plate (due to retention of hypertrophic chondrocytes) and metaphyseal hyperostosis (due to thickening and fusion of secondary trabeculae) with a concomitant loss of primary trabeculae (Figure 4). Finally, in the incisor teeth, minimal to moderate incisor tooth dental dysplasia was noted. All of these changes were considered secondary to VEGFR inhibition.

104


Figure 1. (A) Gross photograph of an ovary from a Nude rat treated with AZ12585313 (PDGFRa and b inhibitor) at 12.5 mg/kg BID compared to a control ovary (upper inset). The ovary is enlarged with abnormal nodular areas of red discolouration which correlate histologically with areas of minimal cystic haemorrhagic dilatation and degeneration of corpora lutea. Lower inset shows a severely enlarged ovary from a HW rat treated at 25 mg/kg BID for 9 days. (B and C) Sub-gross Aperio whole image slide scans of H&E stained ovaries from HW rats treated with AZ12585313 (PDGFRa and b inhibitor) at 6 and 25 mg/kg BID showing severe (B) and minimal (C) cystic haemorrhagic dilatation/degeneration of the corpora lutea. In (B), the ovary is severely dilated with haemorrhage which occupies all the dead space and replaces the ovarian cortex and medulla. Residual abnormal corpora lutea are present (arrowheads) with areas of haemorrhage into the central lumen. In (C) the corpora lutea are abnormal with dilated and haemorrhagic central cavities. (D) Higher power photomicrograph of figure 1B (boxed area - original magnification x20) showing an abnormal corpus luteum with several areas of dilated sinusoids (*). Note the general absence of interstitial cells (pericytes/ endothelium) (arrow heads) throughout the corpus luteum.

␣SMA and CD31 Staining in Normal Ovaries To further evaluate the mechanism of PDGFR-induced degeneration and hemorrhage into the corpus luteum, we stained ovaries treated with AZ12585313 with aSMA and CD31 (PECAM-1). aSMA is a marker of pericytes, vascular smooth muscle, and myofibroblasts, whereas CD31 is a specific marker of vascular endothelium.

The ovarian cortex of normal cycling ovaries in rats and mice normally consists of a number of ovarian follicles and corpora lutea in varying stages of development that cycle every 4–5 days, synchronous to the estrous stage of the animal. In newly formed (basophilic) corpora lutea, CD31 staining revealed staining of flattened spindle cells (Figure 5), whereas aSMA staining revealed strong staining of spindle to stellateshaped pericytes with fine cytoplasmic processes that extend

Hall et al.

105

Figure 2. Photomicrograph (original magnification 10) of an hematoxylin and eosin–stained incisor from (A) a control rat and (B) a Wistar Hannover rat treated with AZ12253678 (pan class III inhibitor) at 38.3 mg/kg QD for 28 days. In the upper panel, note regular pseudostratified odontogenic epithelium (arrows) with fine intraepithelial capillaries that perfuse the epithelium. In the lower panel, there is generalized loss and disorganization of odontogenic cells and associated vasculature with the secondary deposition of abnormal dentine (g) and loss of fine parallel dentine tubules.

to contact several adjacent sinusoids (Figure 6). Endothelial cells and pericytes appeared to comigrate rapidly and in tandem and were most likely derived from the preexisting vascular supply in the theca interna of the ovarian follicle (Figure 6B and C). Older corpora lutea showed intense CD31 staining of plumper-shaped endothelial cells with a centripetal gradient of staining that became weaker toward the center of the corpus luteum (Figure 5A and B). aSMA staining generally showed strong staining throughout the corpus luteum (Figure 6A and B). In corpora lutea which had central areas of degenerate and necrotic cells, dense aSMA staining was present characterized by increased size and numbers of intensely staining aSMA-positive cells (Figure 6A inset). As expected, avascular ovarian follicles showed no staining with either CD31 or aSMA.

␣SMA and CD31 Staining in Ovaries Treated with AZ12585313 In ovaries treated with AZ12585313 at 25mg/kg BID for 9 days, no significant differences were noted in CD31 staining. Morphologically normal corpora lutea were associated with relatively normal numbers of plump endothelial cells (Figure 7A and B), whereas degenerate corpora lutea were associated with weak staining of normal flattened endothelial cells (Figure 7B and C). Staining the same corpora lutea with aSMA showed that, in relatively normal corpora lutea, reduced numbers of pericytes were present (Figure 8A and B), whereas in degenerate corpora lutea characterized by areas of angiectasis or cystic dilatation and hemorrhage, a significant decrease, or in some cases, an almost complete absence of pericytes was noted (Figure 8B

106


Figure 3. Subgross Aperio whole-image slide scans of hematoxylin and eosin–stained ovaries from (A) a control Wistar Hannover rat (HW rat) and (B) an HW rat treated with AZD2932 (vascular endothelial growth factor receptor and platelet-derived growth factor receptor inhibitor) at 49.2 mg/kg QD for 21 days. In the treated ovary, the total number and size of corpora lutea are reduced together with an increase in interstitial cells. At higher magnification (C) areas of central necrosis (*) within individual corpora lutea are visible and (D) in residual corpora lutea, there is a generalized loss of spindle-shaped interstitial cells (pericytes and endothelial cells).

Figure 4. Subgross Aperio whole-image slide scans of hematoxylin and eosin–stained femorotibial joint from (A) a control Wistar Hannover rat (HW rat) and (B) a HW rat treated with AZD2932 (vascular endothelial growth factor receptor and platelet-derived growth factor receptor inhibitor) at 49.2 mg/kg QD for 21 days. In the treated femorotibial joint, the epiphyseal growth plate is significantly increased in thickness due to the retention of hypertrophic chondrocytes. Higher-power images (C; original magnification 4) and (D; original magnification 10) detailing area of hypertrophic chondrocytes, loss of primary trabecula (*), and hyperostosis within the metaphysis due to thickening and fusion of secondary trabeculae (arrows).

and C). In other areas, degeneration and loss of luteal cells resulted in the development of microcysts which lacked both an endothelial and a pericyte lining. Finally, in some corpora lutea (Figure 8B and C), thinning and loss of aSMAþ theca interna cells were apparent (Figure 8B) which most likely contributed to the rupture of the corpus luteum and secondary hemorrhage into the ovary.

Discussion We have studied the effects of a number of small molecule inhibitors with differential activity versus PDGFRa, PDGFRb, VEGFR, and pan class III RTKs. AZ12585313 showed high specificity for the PDGFR receptor, whereas AZ12253678 and AZD2932 showed a more mixed profile with additional specificity for other class III RTKs and/or VEGFR2.

Hall et al.

107

Figure 5. (A) Subgross Aperio whole-image slide scan showing CD31 immunocytochemical staining of a control (untreated) ovary with a number of old and new corpora lutea—note that the avascular (Graafian) follicles show no staining (*). (B) Higher-power image (original magnification 10) of a corpus luteum showing centripetal inward migration of more spindle-shaped endothelial cells and (C) increased magnification of a corpus luteum from another control animal (original magnification 10; inset digital enlargement original magnification 20) to demonstrate weak staining and lower cell density of flattened endothelial cells in a newly formed corpus luteum.

Treatment of female rats with AZ12585313 (PDGFRa and b inhibitor) resulted in a severe hemorrhagic phenotype in the ovary, which we attributed to PDGFRb inhibition. Examination of the growth plate, adrenal gland, and rodent incisor revealed no other changes that might suggest a significant antiangiogenic effect. These results therefore came as an initial surprise as prior experience, together with an extensive literature database, had shown that pharmacological inhibition of the VEGF signaling axis reliably and robustly resulted in ovarian atrophy in normal cycling rats (Ferrara et al. 1998). In addition, VEGF inhibition is reliably associated with a spectrum of ontarget pharmacological effects characterized by epiphyseal growth plate dysplasia, adrenal cortical hemocysts, and incisor tooth dental dysplasia (Gerber et al. 1999; Wedge et al. 2005, Ferrara et al. 1998; Wulff et al. 2002; Patyna et al. 2008; Fletcher et al. 2010). In a separate study, Sleer and Taylor (2007) demonstrated that microinjection of AG1295 (a PDGFR inhibitor) into the ovary resulted in hemorrhage in 3 of the 16 rats. Commonly, microinjection delivers a high dose of compound and the process damages the tissues. However, these data lend support to our original hypothesis that this phenotype

was due to vascular destabilization resulting from inhibition of the PDGFR receptor. The first indication that this effect might be due to vascular destabilization due to reduced pericyte recruitment came from careful examination of the ovaries which revealed that in some corpora lutea, dilatation of the sinusoids could be discerned, accompanied by a reduction in the numbers of interstitial spindle cells in H&E-stained sections (presumptive pericytes and endothelial cells). This change appeared to be the earliest effect induced by PDGFR inhibition, as it was restricted to newly formed basophilic corpora lutea. These results appeared to be strongly reminiscent of the phenotype induced in developing embryos by PDGF/PDGFRb gene deletion, which is characterized by vessel hyperdilation and hemorrhage resulting from pericyte insufficiency (Hellstrom et al. 1999). Repeating these experiments with AZ12253678, a pan class III RTK inhibitor that lacked VEGFR2 activity, resulted in very similar changes in the ovary, strongly indicating that inhibition of the PDGFR receptor was responsible for the effect and that this could be attributed to failure in adequately recruiting mural supporting cells (principally pericytes) leading to vessel

108


Figure 6. (A) Sub-gross Aperio whole image slide scan of the same ovary as in Figure 5 showing aSMA immunocytochemical staining. Strong aSMA staining is present throughout the ovary except for (Graafian) follicles (*). Inset shows a corpus luteum from another control animal with a central area of necrotic cell debris and strong staining of a focus of aSMAþ cells (B and C) High power image from area photographed in figure 5B and 5C showing centripetal migration and recruitment of pericytes. Pericyte recruitment from the theca is rapid and appears to be in tandem with endothelial cells. A high cell density (exceeding normal numbers) of pericytes is present in early corpora lutea with a decrease in cell density as the vasculature matures.

fragility in vascularizing corpora lutea. Treatment with AZ12253678 also resulted in small changes in the growth plate (dysplasia) and adrenal gland (hemocysts) as well as minimal to moderate incisor tooth dental dysplasia indicative of a minimal antiangiogenic (VEGF-like) effect. These effects most likely reflect the multikinase inhibitory action of AZ12253678 and support the notion that the growth plate (and possibly the adrenal gland and rodent incisor) can act as an angiogenic sentinel as it responds predictably and sensitively to inhibition through a number of different antiangiogenic mechanisms (Gerber et al. 1999; Brown et al. 2005; Hall, Westwood, and Wadsworth 2006). Treatment with either AZD2932 (VEGFR and PDGFR inhibitor; Ple, Jung, Ashton, Hennequin, Laine, Morgentin, et al. 2012) or a combination of AZD2171 (Cediranib; VEGFR inhibitor; Wedge et al. 2005) and AZ12585313 (PDGFRa and b inhibitor), which were dosed to achieve approximately 1:1 inhibition of the PDGFR and VEFGR receptors, resulted in avascular necrosis and a reduction in the number and size of corpora lutea in the ovary—essentially identical changes to

those induced by VEGFR inhibition alone (Ferrara et al. 1998) and similar to those induced by Sunitinib, which inhibits both VEGFR and PDGFR receptors (as well as c-Kit, FLT3, and RET receptors; Patyna et al. 2008). These experiments therefore confirmed that during ovarian angiogenesis, VEGFR inhibition was dominant to PDGFR inhibition, consistent with the sequential cascade of endothelial vascularization followed by pericyte stabilization during angiogenesis. Analysis of pericytes and vascular endothelial cells in normal ovaries and those treated with AZ12585313 (PDGFRa and b inhibitor) using CD31 and aSMA immunohistochemistry provided an interesting insight into how pericytes and endothelial cells migrate and function during normal angiogenesis. CD31 is specific for endothelium, whereas aSMA stains pericytes, myofibroblasts, and vascular smooth muscle, all of which express PDGFRb (Hall 2006). CD31 staining was relatively weak in newly vascularizing corpora lutea, most likely as a result of low expression during migration (RayChaudhury et al. 2001). In more mature corpora lutea, CD31 staining was intense with an abundance of plump (active) endothelial cells.

Hall et al.

109

Figure 7. (A) Subgross Aperio whole-image slide scan showing CD31 immunocytochemical staining of an ovary from a rat treated with 25 mg/kg BID of AZ12585313 (platelet-derived growth factor receptor alpha and platelet-derived growth factor receptor beta inhibitors) for 9 days (upper inset represents boxed area in (B)—digital enlargement from original 20 magnification, and lower inset represents boxed area in (C)—digital enlargement of original 20 magnification). (B) Higher-power view (original magnification 5) showing CD31 staining of a morphologically normal corpus luteum (*) and a degenerate corpus luteum (**) with flattened endothelium. (C) Higher-power view (original magnification 20) showing very weak staining of flattened endothelium from an early degenerate corpus luteum.

In addition, in some corpora lutea, a centripetal gradient of staining was apparent, representing the inward migration of endothelial cells from established blood vessels within the theca (Bassett 1943). aSMA staining, on the other hand, revealed more uniform staining of pericytes, myofibroblasts, and low numbers of vascular smooth muscle cells. In a few corpora lutea, a gradient of staining was visible, but in general, pericytes were evenly distributed with uniform and intense staining. In all cases, pericytes were closely adjacent to endothelial cells. In favorable planes of section, fine cytoplasmic processes radiated from cell bodies to form intimate contacts with endothelial cells. The density of pericytes and the number and size of cytoplasmic processes suggested to us that the corpus luteum was particularly dependent upon pericyte vascular support. Finally, at the center of some corpora lutea, the foci of aSMA-positive cells were present around areas of cell debris suggesting perhaps a transitory phagocytic role for aSMAþ myofibroblasts (Nagao et al. 1986). Treatment of animals with AZ12585313 (PDGFRa and b inhibitor) revealed no changes in CD31 staining, indicating that PDGFR inhibition has no significant effect on endothelial cell migration. This is consistent with the fact that granulosa steroidogenic cells are the main source of VEGF in the luteinizing Graafian follicle (Wulff et al. 2000). aSMA staining, however,

revealed a significant reduction in the numbers of pericytes. In newly formed degenerate corpora lutea, there was a severe reduction, and in some cases, an almost complete absence of pericytes. This loss of pericytes coincided with areas of sinusoidal dilation and hemorrhage. These results accounted for the loss of spindle cells noted in the H&E-stained slides. They also confirmed that PDGFRb rather than PDGFRa inhibition selectively inhibited pericyte recruitment since pericytes express PDGFRb but not PDGFRa (Lamagna and Bergers 2006). aSMA staining also occasionally revealed a reduction in the numbers of thecal smooth muscle cells that surrounded corpora lutea and presumably contribute to the structural integrity of the corpus luteum (Young and McNeilly 2010). AZ12585313 (PDGFRa and b inhibitor) was developed as a specific PDGFR inhibitor to target the interaction of tumor vasculature with pericytes or to moderate loosening of cell contacts with endothelial cells, which would increase the sensitivity of nude capillaries to VEGF inhibition (Erber et al. 2004). In our experimental model, PDGFR inhibition alone was uniquely active as a single agent in the vascularizing corpus luteum. Newly developing (basophilic) corpora lutea appeared to be the most sensitive to PDGFRb inhibition. This is likely due to increased vessel fragility resulting from endothelial proliferation, and active pericyte recruitment and attachment, as these types of vessels are known to be hyperpermeable (Nagy et al.

110


Figure 8. (A) Subgross Aperio whole-image slide scan showing alpha smooth muscle actin (aSMA) immunocytochemical staining of the ovary shown in Figure 7, with upper inset of boxed area in B (digital enlargement original magnification 20) and lower inset of boxed area in C (digital enlargement original magnification 20). (B) Higher-power view (original magnification 5) of aSMA staining showing reduced numbers of pericytes in the morphologically normal corpus luteum (*) and almost complete absence of pericytes in degenerate corpus luteum (**). Theca interna also shows reduced numbers of SMAþ cells and rupture (arrow heads). (C) Higher-power view (original magnification 20) of a degenerate corpus luteum showing almost complete absence of pericytes and reduced pericytes in theca interna (arrow heads). Most vessels lack pericyte support (high-magnification inset), whereas in other areas, microcysts are developing which lack both a vascular endothelium and a pericytes (*).

2008). In older normal looking corpora lutea, or corpora lutea that evaded significant pericyte loss, vessel maturation due to pericyte attachment likely results in more resistant microvessels consistent with the known vascular stabilizing effects of pericytes (Hall 2006). These corpora lutea are more likely to persist despite PDGFRb inhibition and appeared as residual structures within otherwise degenerate ovaries. Consistent with this hypothesis, our results suggest that in human cancer, PDGFR monotherapy may provide only modest benefit in tumors with established vessels with more mature mural cell support. In addition, our data show that the rodents, as part of the safety evaluation of new chemical entities, can provide additional efficacy and mechanistic information. In some cases, these data can even be quantified (Wedge et al. 2005). Careful histological evaluation therefore might not only identify target organs and ascribe them empirically to the intended pharmacology but may be used to confirm the mechanism of action in normal tissues. Author Contributions Authors contributed to conception or design (SA, JH, ZW, JR, GR, RW, AH); data acquisition, analysis, or interpretation (SA, ZW, JR, RW, AH, SB); drafting the manuscript (AH); and critically revising

the manuscript (SA, JH, ZW, JR, GR, RW, AH, SB). All authors gave final approval, and agreed to be accountable for all aspects of work in ensuring that questions relating to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

References Augustin, A. J., Scholl, S., and Kirchhof, J. (2009). Treatment of neovascular age-related macular degeneration: Current therapies. Clin Ophthalmol 3, 175–82. Bassett, D. E. L. (1943). The changes in the vascular pattern of the ovary of the albino rat during the estrous cycle. Am J Anat 73, 251–78. Brave, S. R., Ratcliffe, K., Wilson, Z., James, N. H., Ashton, S., Wainwright, A., Kendrew, J., Dudley, P., Broadbent, N., Sproat, G., Taylor, S., Barnes, C., Silva, J. C., Farnsworth, C. L., Hennequin, L., Ogilvie, D. J., Jurgensmeier, J. M., Shibuya, M., Wedge, S. R., and Barry, S. T. (2011). Assessing the activity of cediranib, a VEGFR-2/3 tyrosine kinase inhibitor, against VEGFR-1 and members of the structurally related PDGFR family. Mol Cancer Ther 10, 861–73.

Hall et al. Brown, A. P., Courtney, C. L., King, L. M., Groom, S. C., and Graziano, M. J. (2005). Cartilage dysplasia and tissue mineralization in the rat following administration of a FGF receptor tyrosine kinase inhibitor. Toxicol Pathol 33, 449–55. Carmeliet, P., Ferreira, V., Breier, G., Pollefeyt, S., Kieckens, L., Gertsenstein, M., Fahrig, M., Vandenhoeck, A., Harpal, K., Eberhardt, C., Declercq, C., Pawling, J., Moons, L., Collen, D., Risau, W., and Nagy, A. (1996). Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 380, 435–39. Erber, R., Thurnher, A., Katsen, A. D., Groth, G., Kerger, H., Hammes, H. P., Menger, M. D., Ullrich, A., and Vajkoczy, P. (2004). Combined inhibition of VEGF and PDGF signaling enforces tumor vessel regression by interfering with pericyte-mediated endothelial cell survival mechanisms. FASEB J 18, 338–40. Ferrara, N., Carver-Moore, K., Chen, H., Dowd, M., Lu, L., O’Shea, K. S., Powell-Braxton, L., Hillan, K. J., and Moore, M. W. (1996). Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature 380, 439–42. Ferrara, N., Chen, H., Davis-Smyth, T., Gerber, H. P., Nguyen, T. N., Peers, D., Chisholm, V., Hillan, K. J., and Schwall, R. H. (1998). Vascular endothelial growth factor is essential for corpus luteum angiogenesis. Nat Med 4, 336–40. Fletcher, A. M., Bregman, C. L., Woicke, J., Salcedo, T. W., Zidell, R. H., Janke, H. E., Fang, H., Janusz, W. J., Schulze, G. E., and Mense, M. G. (2010). Incisor degeneration in rats induced by vascular endothelial growth factor/fibroblast growth factor receptor tyrosine kinase inhibition. Toxicol Pathol 38, 267–79. Gerber, H. P., Vu, T. H., Ryan, A. M., Kowalski, J., Werb, Z., and Ferrara, N. (1999). VEGF couples hypertrophic cartilage remodeling, ossification and angiogenesis during endochondral bone formation. Nat Med 5, 623–28. Hall, A. P. (2005). The role of angiogenesis in cancer. Comp Clin Path 13, 95–99. Hall, A. P. (2006). Review of the pericyte during angiogenesis and its role in cancer and diabetic retinopathy. Toxicol Pathol 34, 763–75. Hall, A. P., Westwood, F. R., and Wadsworth, P. F. (2006). Review of the effects of anti-angiogenic compounds on the epiphyseal growth plate. Toxicol Pathol 34, 131–47. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100, 57–70. Hasegawa, E., Oshima, Y., Takeda, A., Saeki, K., Yoshida, H., Sonoda, K. H., and Ishibashi, T. (2012). IL-27 inhibits pathophysiological intraocular neovascularization due to laser burn. J Leukoc Biol 91, 267–73. Hellstrom, M., Kalen, M., Lindahl, P., Abramsson, A., and Betsholtz, C. (1999). Role of PDGF-B and PDGFR-beta in recruitment of vascular smooth muscle cells and pericytes during embryonic blood vessel formation in the mouse. Development 126, 3047–55. Howdieshell, T. R., Callaway, D., Webb, W. L., Gaines, M. D., Procter, C. D., Jr., Sathyanarayana, Pollock, J. S., Brock, T. L., and McNeil, P. L. (2001). Antibody neutralization of vascular endothelial growth factor inhibits wound granulation tissue formation. J Surg Res 96, 173–82. Lamagna, C., and Bergers, G. (2006). The bone marrow constitutes a reservoir of pericyte progenitors. J Leukoc Biol 80, 677–81.

111 Lodish, M. B. (2013). Clinical review: kinase inhibitors: Adverse effects related to the endocrine system. J Clin Endocrinol Metab 98, 1333–42. Nagao, S., Nemoto, H., Suzuki, M., Satoh, N., and Iijima, S. (1986). Myofibroblasts in basal cell epithelioma: With special reference to the phagocytic function of myofibroblasts. J Cutan Pathol 13, 261–67. Nagy, J. A., Benjamin, L., Zeng, H., Dvorak, A. M., and Dvorak, H. F. (2008). Vascular permeability, vascular hyperpermeability and angiogenesis. Angiogenesis 11, 109–19. Patyna, S., Arrigoni, C., Terron, A., Kim, T. W., Heward, J. K., Vonderfecht, S. L., Denlinger, R., Turnquist, S. E., and Evering, W. (2008). Nonclinical safety evaluation of sunitinib: A potent inhibitor of VEGF, PDGF, KIT, FLT3, and RET receptors. Toxicol Pathol 36, 905–16. Ple, P. A., Jung, F., Ashton, S., Hennequin, L., Laine, R., Lambert-van der Brempt, C., Morgentin, R., Pasquet, G., and Taylor, S. (2012). Discovery of new quinoline ether inhibitors with high affinity and selectivity for PDGFR tyrosine kinases. Bioorg Med Chem Lett 22, 3050–55. Ple, P. A., Jung, F., Ashton, S., Hennequin, L., Laine, R., Morgentin, R., Pasquet, G., and Taylor, S. (2012). Discovery of AZD2932, a new Quinazoline Ether Inhibitor with high affinity for VEGFR-2 and PDGFR tyrosine kinases. Bioorg Med Chem Lett 22, 262–66. RayChaudhury, A., Elkins, M., Kozien, D., and Nakada, M. T. (2001). Regulation of PECAM-1 in endothelial cells during cell growth and migration. Exp Biol Med (Maywood) 226, 686–91. Robinson, R. S., Woad, K. J., Hammond, A. J., Laird, M., Hunter, M. G., and Mann, G. E. (2009). Angiogenesis and vascular function in the ovary. Reproduction 138, 869–81. Sleer, L. S., and Taylor, C. C. (2007). Platelet-derived growth factors and receptors in the rat corpus luteum: Localization and identification of an effect on luteogenesis. Biol Reprod 76, 391–400. Wedge, S. R., Kendrew, J., Hennequin, L. F., Valentine, P. J., Barry, S. T., Brave, S. R., Smith, N. R., James, N. H., Dukes, M., Curwen, J. O., Chester, R., Jackson, J. A., Boffey, S. J., Kilburn, L. L., Barnett, S., Richmond, G. H., Wadsworth, P. F., Walker, M., Bigley, A. L., Taylor, S. T., Cooper, L., Beck, S., Jurgensmeier, J. M., and Ogilvie, D. J. (2005). AZD2171: A highly potent, orally bioavailable, vascular endothelial growth factor receptor-2 tyrosine kinase inhibitor for the treatment of cancer. Cancer Res 65, 4389–400. Wulff, C., Dickson, S. E., Duncan, W. C., and Fraser, H. M. (2001). Angiogenesis in the human corpus luteum: Simulated early pregnancy by HCG treatment is associated with both angiogenesis and vessel stabilization. Hum Reprod 16, 2515–24. Wulff, C., Wilson, H., Largue, P., Duncan, W. C., Armstrong, D. G., and Fraser, H. M. (2000). Angiogenesis in the human corpus luteum: Localization and changes in angiopoietins, tie-2, and vascular endothelial growth factor messenger ribonucleic acid. J Clin Endocrinol Metab 85, 4302–9. Wulff, C., Wilson, H., Wiegand, S. J., Rudge, J. S., and Fraser, H. M. (2002). Prevention of thecal angiogenesis, antral follicular growth, and ovulation in the primate by treatment with vascular endothelial growth factor Trap R1R2. Endocrinology 143, 2797–807. Young, J. M., and McNeilly, A. S. (2010). Theca: The forgotten cell of the ovarian follicle. Reproduction 140, 489–504.