Nonclinical Safety Evaluation of Sunitinib: A Potent Inhibitor of VEGF,PDGF,KIT,FLT3,and RET Receptors

Abstract

Sunitinib malate (SUTENT) is a multitargeted receptor tyrosine kinase (RTK) inhibitor that is approved multinationally for the treatment of imatinib-resistant/-intolerant gastrointestinal stromal tumor and advanced renal cell carcinoma. This paper characterizes the organ toxicity of sunitinib in Sprague-Dawley rats and cynomolgus monkeys, and the reversibility of any treatment-induced effects. Rats and monkeys received sunitinib (0–15 and 0–20 mg/kg/day, respectively) orally on a consecutive daily dosing schedule for thirteen weeks or on an intermittent daily dosing schedule for up to nine months. Clinical observations and laboratory parameters were recorded. Necropsy was conducted following treatment/recovery periods, and histologic examinations were performed. In rats, sunitinib was generally tolerated at 0.3 and 1.5 mg/kg/day, and findings were reversible. In monkeys, the level at which there were no observed adverse effects was 1.5 mg/kg/day, and findings were similarly reversible (except for uterine/ovarian weight changes and skin pallor). Data suggest that inhibition of multiple RTK pathways may induce pharmacologic effects on organ systems in nonclinical species. Key pharmacologic effects of sunitinib included reversible inhibition of neovascularization into the epiphyseal growth plate, and impaired corpora lutea formation and uterine development during estrus. Similar observations have been noted with this class of RTK signaling inhibitors and are consistent with pharmacologic perturbations of physiologic/angiogenic processes associated with the intended molecular targets.

Introduction

Angiogenesis—the formation of blood vessels—is essential for the development of the vascular system in the growing embryo, as well as for wound healing and reproductive function in healthy adults. Vascular endothelial growth factor (VEGF) signaling is required for normal fetal development, as shown by the inactivation of a single allele of VEGF (Carmeliet et al. 1996; Ferrara et al. 1996). Angiogenesis also plays a pivotal role in several pathologic conditions, particularly tumorigenesis and metastasis (Folkman 1974). One of the most potent positive regulators of angiogenesis is VEGF (Ferrara 2004), and the importance of platelet-derived growth factor (PDGF) is increasingly recognized (Erber et al. 2004). Growth factors such as these—or the specific cell-surface receptors to which they bind (receptor tyrosine kinases; RTKs)—may be overexpressed or mutated in various types of cancer, leading to disruption of the normal angiogenic balance and creation of a proangiogenic environment (Ferrara et al. 2003; Franklin et al. 2002; Igarashi et al. 2002; Jones and Cross, 2004; Ono and Kuwano 2006). Many anti-cancer therapies are now directed against tumor-related angiogenesis, with specific targets including inhibition of VEGF and PDGF signaling. Strategies in development include monoclonal antibodies directed against VEGF or its receptors (e.g., bevacizumab), and small-molecule RTK inhibitors that act intracellularly to prevent autophosphorylation and activation of downstream growth-promoting signaling cascades (e.g. imatinib mesylate, sorafenib tosylate, and sunitinib malate). There is evidence that dual inhibition of VEGF and PDGF pathways provides greater antitumor efficacy than inhibition of either target alone (Bergers et al. 2003; Erber et al. 2004).

Sunitinib (SU11248; SUTENT;^® Pfizer, Inc., New York, NY, USA) is an oral, multitargeted RTK inhibitor that selectively and potently inhibits class III and V split kinase domain RTKs, including VEGF receptors (VEGFR-1, -2, and -3), PDGF receptors (PDGFR-α and -β), stem cell factor receptor (KIT), FMS-like tyrosine kinase-3 receptor (FLT3), the receptor for macrophage colony-stimulating factor (CSF-1R), and glial cell-line–derived neurotrophic factor receptor (rearranged during transfection; RET) (Abrams et al. 2003; Mendel et al. 2003; Motzer et al. 2006; Murray et al. 2003; O’Farrell et al. 2003; Pfizer Inc., data on file). Sunitinib has been approved multinationally for the treatment of gastrointestinal stromal tumor (GIST) after intolerance to or disease progression with imatinib therapy, and for the first-line treatment of advanced renal cell carcinoma (RCC) (Rock et al. 2007).

We report here the nonclinical safety profile of sunitinib as characterized in Sprague-Dawley rats and cynomolgus monkeys (Macaca fascicularis). The objectives of this work included the characterization of systemic exposure and target organ toxicity following oral administration of sunitinib for up to six months in rats and for up to nine months in monkeys, and verification of the reversibility of any treatment-induced effects.

Materials and Methods

Five oral toxicity studies are described: a two-week investigative study in rats to evaluate adrenal and pancreatic toxicity was carried out following toxicity observations in an earlier study; a thirteen-week study in rats followed by a six-week recovery period; a six-month study in rats followed by an eight-week recovery period; a thirteen-week study in monkeys followed by a six-week recovery period; and a nine-month study in monkeys followed by an eight-week recovery period.

Animals and Husbandry

In the two-week investigative rat study, female Sprague-Dawley rats aged fifty-seven days and weighing 189–242 g at the start of the treatment period were supplied by Charles River Laboratories Italia SpA (Calco [Lecco], Italy). In the thirteen-week rat study, male and female Sprague-Dawley rats aged forty-three days—the males weighing 164–197 g and the females weighing 134–171 g at the start of treatment—were supplied by Charles River Laboratories Italia SpA (Calco [Lecco], Italy). In the six-month rat study, male and female Sprague-Dawley rats aged forty-six to fifty-two days—the males weighing 146–220 g and the females weighing 137–191 g at the start of the treatment period—were supplied by Charles River Laboratories, Inc., (Portage, MI, USA).

In the thirteen-week and nine-month monkey studies, juvenile to adult male and female cynomolgus monkeys—the males weighing 2.5–3.6 kg and the females weighing 2.3–3.2 kg—were supplied by Charles River BRF, Inc. (Houston, TX, USA).

In all studies, standard procedures and conditions were applied for animal care, feeding, maintenance of the rooms, caging, and environment. The procedures for housing and handling adopted throughout the thirteen-week rat study were in strict compliance with the Organization for Economic Cooperation and Development (OECD) principles of Good Laboratory Practice (GLP) and Italian guidelines for laboratory animal welfare. The other studies were conducted in compliance with the Food and Drug Administration GLP regulations (Code of Federal Regulations, Title 21, Part 58).

Drug

Sunitinib L-malate salt was supplied by Pharmacia and Upjohn (Milan, Italy and Kalamazoo, MI, USA) for the two-week and thirteen-week studies, and by Pfizer Global Research and Development (Kalamazoo, MI, USA) for the six- and nine-month studies. All dose levels are represented as the free base equivalent mg/kg dose. Sunitinib was formulated in an aqueous vehicle and administered to study animals via oral gavage.

Study Design

Rat Studies

In the two-week investigative study, rats given sunitinib at 15 or 45 mg/kg/day were euthanized at scheduled times (days 2, 4, 7, and 14; recovery on day 28), and samples of adrenal glands and pancreas were retained in Karnovsky’s fixative. The tissues were post-fixed in osmium tetroxide solution and then embedded in Epon–Araldite after appropriate dehydration. Adrenal glands and pancreas were sectioned at 0.5 μm thickness using a Leica UCT ultramicrotome, stained with toluidine blue, 1% Azure II and 1% borax, and examined by light microscopy to define areas of interest. Ultrathin sections of 70–90 nm were obtained and mounted on copper/palladium grids, stained with 10% aqueous uranyl acetate and 2.5% lead acetate, and examined using a Hitachi 7100 transmission electron microscope at 75 kV. Digital images were captured using an AMT digital camera.

In the thirteen-week study, twenty rats/sex/group were to be treated with sunitinib 0, 1.5, 5, or 15 mg/kg/day on a consecutive daily dosing schedule for thirteen weeks. In the six-month study, twenty-five rats/sex/group were to be treated with sunitinib 0, 0.3, 1.5, or 6 mg/kg/day on an intermittent daily dosing schedule (five consecutive twenty-eight-day dosing cycles, each separated by seven days off treatment) for six months. The first fifteen rats/sex/group were to be euthanized at the end of the treatment period; the remaining animals were to be maintained for a six- to eight-week recovery period.

Clinical signs and behavior were observed daily. Body weight was recorded during the pretest period and at least weekly thereafter. Ophthalmoscopic examinations were performed pretest and toward the end of the treatment and recovery periods. Laboratory examinations (hematology, clinical chemistry, and urinalysis) were performed as follows: on days 25, 91, 109, and 134 in the thirteen-week study; in the six-month study, blood samples were collected on days 98, 168, 169, and 170 of the treatment period and on days 7, 57, and 58 of the recovery period, and urine samples were collected on day 165 of the treatment period and day 53 of the recovery period. Blood samples for assessment of levels of systemic exposure to sunitinib and its primary metabolite, SU12662 (as measured by the area under the plasma drug concentration–time curve over a twenty-four-hour dosing interval [AUC_(0–24)]), were also collected (in six rats/sex in the active treatment groups and three rats/sex in the control group) at various time points throughout the study. Necropsy was performed at the end of the treatment and recovery periods; organ weights were recorded, and histologic examinations were performed.

Monkey Studies

In the thirteen-week study, six monkeys/sex/group were to be treated with sunitinib 0, 2, 6, or 20 mg/kg/day on a consecutive daily dosing schedule for thirteen weeks. In the nine-month study, seven monkeys/sex/group were to be treated with sunitinib 0, 0.3, 1.5, or 6 mg/kg/day on an intermittent daily dosing schedule (eight consecutive twenty-eight-day dosing cycles each separated by seven days off treatment) for nine months. Two (thirteen-week study) or three (nine-month study) monkeys/sex/group were to be maintained for a six- to eight-week recovery period following the end of the dosing period; the remaining animals were euthanized.

Toxicity was evaluated by assessing changes in clinical signs and behavior (at least daily), body weight (pretest, then weekly), and ophthalmoscopic examinations (pretest, and during dosing/recovery periods). Electrocardiography was also performed in the thirteen-week study (pretest, three times during treatment, and at the end of the recovery period). Laboratory examinations (hematology, clinical chemistry, and urinalysis) were performed as follows: pretest, two or three times during treatment, and three times during recovery in the nine-month study; in the thirteen-week study, blood samples were collected pretest, three or four times during treatment, and once during recovery, and urine samples were collected pretest and at the end of dosing and recovery. Blood samples for the assessment of systemic exposure to sunitinib and SU12662 were also collected (in three or four monkeys/sex/group) at various time points throughout the study. Necropsy was conducted at the end of the treatment and recovery periods; organs were weighed, and histologic examinations were performed.

Results

Tolerability

Rat Studies

In the thirteen-week study, treatment-related mortality occurred in nine of twenty male and three of twenty female rats receiving high-dose sunitinib (15 mg/kg/day). Owing to the poor general condition and high incidence of mortality in this group, treatment was stopped on day 62 (week 9), and most animals were euthanized; the last four rats/sex were maintained for the six-week recovery period.

In the six-month study, treatment-related mortality occurred in three rats in the high-dose group (6 mg/kg/day), one rat in the mid-dose group (1.5 mg/kg/day), and one rat in the low-dose group (0.3 mg/kg/day).

Monkey Studies

In the thirteen-week study, high-dose sunitinib (20 mg/kg/day) caused severe toxicity and resulted in early deaths (seven monkeys in total). Treatment was stopped on day 67 (week 10) after dose reduction (to 12 mg/kg/day) and temporary dosing cessation. Two monkeys/sex from the high-dose group were allowed to continue into the recovery phase.

In the nine-month study, the highest sunitinib dose (6 mg/kg/day) caused severe toxicity and resulted in early deaths (because of poor physical condition) in three monkeys. Consequently, dosing was stopped on day 168 following five dosing cycles.

Clinical Signs of Toxicity

Rat Studies

Treatment-related clinical signs included persistent yellowish discoloration of the fur in most rats given the two highest doses (5 and 15 mg/kg/day) in the thirteen-week study, and of the body surfaces in rats given 6 mg/kg/day in the six-month study. Additionally, signs of discolored (dark yellow) urine occurred in the six-month study in rats given ≥ 0.3 mg/kg/day. Broken incisors occurred in the thirteen-week and six-month studies at similar doses (≥ 5 mg/kg/day and 6 mg/kg/day, respectively). Rats given 15 mg/kg/day in the thirteen-week study suffered moderate to marked loss of body weight but recovered partially in the last study week after powdered food was supplied; some rats administered this dose had ruffled fur, decreased activity, and soft stools in the last three weeks of treatment.

Monkey Studies

Clinical indications of severe toxicity in monkeys included anorexia, weight loss, emesis, soft to watery feces, pale skin, decreased activity, hunched posture, hypothermia, discolored mouth or gums (pale, bloody), and lip or mouth lesions at 20/12 mg/kg/day in the thirteen-week study, as well as signs of dehydration and gingival ulceration at 6 mg/kg/day in the nine-month study. Some of these clinical signs (pale skin, soft feces, decreased body weight, and decreased food consumption) also occurred in male monkeys given 6 mg/kg/day in the thirteen-week study. In addition, pale and bloody gums were observed in monkeys given 6 mg/kg/day and in one female given 2 mg/kg/day in the thirteen-week study. A modest reduction in heart rate was observed at all dose levels (≥ 2 mg/kg/day) in the thirteen-week study. No clinical signs were observed at the end of the recovery phase of the thirteen-week or nine-month studies in monkeys.

Clinical Chemistry and Hematology Laboratory Assessments

Treatment-related changes in laboratory assessments are summarized in Table 1. It was notable that in the thirteen-week monkey study, none of the hematologic changes was present at the end of the recovery period. Furthermore, in the nine-month monkey study, treatment-related changes occurred only in the high-dose group (6 mg/kg/day).

Macroscopic and Organ Weight Findings

Rat Studies

Several macroscopic findings were observed at necropsy in rats in the thirteen-week study, involving mainly unscheduled deaths and animals in the high-dose group (15 mg/kg/day) that were euthanized at the end of the treatment period. These findings were consistent with the microscopic findings and included alterations in hemolymphopoietic organs, adrenal glands, gastrointestinal (GI) tract, common bile duct, teeth, and reproductive organs. In the mid-dose group (5 mg/kg/day), yellowish discoloration of the stomach and testes, broken incisors, and a single instance of an enlarged jejunum were observed. A marked increase in adrenal weight was observed in animals that were given 15 mg/kg/day in the thirteen-week study and were euthanized prior to scheduled terminations. Decreases in thymus weight occurred in rats given ≥ 5 mg/kg/day in the thirteen-week and six-month studies. A moderate decrease in splenic weight occurred in rats given 5 mg/kg/day in the thirteen-week study and in males given 6 mg/kg/day in the six-month study. A slight decrease in prostate and uterine weights occurred in the respective genders given 5 mg/kg/day in the thirteen-week study. After the recovery periods, only the decrease in thymus weight persisted in females that were given 5 mg/kg/day in the thirteen-week study.

Monkey Studies

Macroscopic findings observed at necropsy during or at the end of the dosing phase were consistent with microscopic findings and included the following: discolored adrenal glands; gingival and/or lingual depressed red foci or ulcerations; and red foci in the stomach or intestines with or without abnormal content with 20/12 mg/kg/day in the thirteen-week study and with 6 mg/kg/day in the nine-month study. Red foci also occurred in the liver and lungs with 20/12 mg/kg/day in the thirteen-week study. Red foci found in the GI tract of one monkey given high-dose treatment were identified by light microscopy as areas of necrosis accompanied by erosion/ulceration. These areas were microscopically consistent with cytomegalovirus infection and were considered to be indicative of impaired defense mechanisms in this animal. None of these findings was present at the end of the recovery phase. Decreases in thymus and spleen weights occurred with ≥ 5 mg/kg/day in the thirteen-week study and 6 mg/kg/day in the nine-month study. Increases in adrenal gland weights occurred with 20/12 mg/kg/day in the thirteen-week study and 6 mg/kg/day (males only) in the nine-month study. Decreases in ovarian and uterine weights occurred at all dose levels (≥ 2 mg/kg/day) in the thirteen-week study, but only with the 6 mg/kg/day dose in the nine-month study. At the end of the recovery period, ovarian weight change persisted in the 20/12 mg/kg/day dose group, but uterine weight change was still evident in all groups (≥ 2 mg/kg/day) in the thirteen-week study.

Histologic and Electron Microscopic Evaluations

Treatment-related histopathologic findings for the thirteen-week and six-month rat studies and the thirteen-week and nine-month monkey studies are summarized in Table 2. All of these changes were reversible or were undergoing reversal at the end of the treatment-free recovery periods. The histologic findings considered to be the primary nonclinical effects of sunitinib treatment were bone marrow hypocellularity, lymphoid atrophy (thymus, spleen, lymph node), tibial physeal dysplasia, pancreatic/salivary gland acinar cell degranulation, adrenal gland necrosis/hemorrhage, GI effects, incisor dental caries (rat only), and reproductive organ effects (ovary, female reproductive tract). Treatment-related histopathologic changes are described below and are illustrated in Figures 1A–1F and Figures 2A–2F, with electron micrographs in Figures 3A–3D.

Adrenal necrosis and hemorrhage were generally associated with extended treatment at elevated dose levels and occurred in both monkey studies at ≥ 6 mg/kg/day and in the thirteen-week rat study at 15 mg/kg/day (Figure 1B). This change occurred primarily in the cortex of the adrenals and was characterized by necrosis and loss of cortical cells, with blood-filled spaces that widened interepithelial cords or formed lake-like areas that replaced cortical cells; these findings are further represented at the ultrastructural level in Figure 3B from a rat treated with sunitinib for fourteen days at 45 mg/kg/day. This change was not evident in rats or monkeys at the end of the recovery period.

Bone marrow hypocellularity occurred in all rat and monkey studies and was characterized principally by depletion of erythroid precursors and, to a lesser extent, myeloid precursors at ≥ 1.5 mg/kg/day (Figure 1D). The effect on the bone marrow was generally more pronounced in both species after consecutive daily administration of the drug compared with cyclic daily administration. At the end of the recovery periods of six to eight weeks in both rats and monkeys, bone marrow was either completely repaired or showed distinct evidence of ongoing recovery.

Epiphyseal growth plate thickening (physeal dysplasia) occurred in both rats and monkeys at ≥ 5 and ≥ 2 mg/kg/day doses, respectively. The change was more prominent after consecutive daily dosing and was characterized by an increased number of hypertrophic chondrocytes, which resulted in an expansion of the zone of hypertrophy (Figure 1F) and calcification of the growth plate in the longitudinal axis of the bone. This change was often associated with a reduction of bony trabeculae in the metaphysis, with the occurrence of focal areas of cartilage in the periostium of the femur or in the tibial diaphysis in some rats in the thirteen-week study. A transverse bony plate immediately subjacent to the growth plate with occasional necrosis of chondrocytes in the hypertrophic zone developed in monkeys in both studies. Growth plate thickening was primarily evident in the appendicular bones (e.g., femur, tibia) but was also evident to a lesser extent in axial skeletal bone such as the sternum. Repair of this change in both species was generally complete or in progress at the end of the recovery periods.

Ovarian follicular atresia/decreased corpora lutea, with or without ovarian degeneration and mineralization, occurred in rats in the thirteen-week study at ≥ 1.5 mg/kg/day and in monkeys at 6 mg/kg/day. Uterine effects of endometrial atrophy also occurred in rats in the thirteen-week study but were evident only at the higher dose of 15 mg/kg/day; in the monkey, this change was evident concurrent with the ovarian changes (6 mg/kg/day). Ovarian or uterine changes were not evident in the six-month rat study with cyclic daily treatment. Ovarian follicular atresia was characterized by degeneration of the follicles (Figure 2B) with or without mineralization and was often associated with fewer corpora lutea than normal. The uterine change was characterized by thinning of the endometrium with compaction of the stratum spongiosum, and was evident in the cervix of some monkeys. Evidence of ovarian effects at the end of the recovery period persisted only with the highest dose (15 mg/kg/day) in rats; the effects were completely reversed in the thirteen-week and nine-month monkey studies.

Caries of maxillary and mandibular incisor teeth occurred in rats given ≥ 5 mg/kg/day, but not in monkeys in either study. The caries were characterized by degeneration and necrosis of the dentine, as well as necrosis of the pulp, odontoblasts, and ameloblasts (Figure 2D). In some rats, there was secondary inflammation in the periodontal ligament and surrounding tissue. Although caries were still evident at the end of the recovery period, the incidence was much reduced, and the proximal portions of the teeth were devoid of change, indicating ongoing repair that is likely to have been incomplete because of insufficient recovery time.

Buccal and esophageal changes occurred in monkeys given ≥ 6 mg/kg/day in the thirteen-week and nine-month studies, but not in the rat studies. The buccal changes were characterized by epithelial cell necrosis with erosions and ulcerations that affected the gingiva. Atrophy of the esophageal and lingual epithelium also occurred and was characterized by thinning of the mucosal squamous epithelium with compaction, hypere-osinophilia, and nuclear pyknosis of the superficial layers that resulted in parakeratosis in the esophagus. These changes had reversed completely by the end of the recovery periods.

Pancreatic and salivary gland acinar cell degranulation occurred in rats and monkeys at ≥ 5 and ≥ 2 mg/kg/day, respectively. There was no effect on pancreatic islets in either species. Degranulation was more prominent in animals dosed continuously compared with those dosed cyclically and was characterized by a reduction in the number of zymogen granules within the acinar cell cytoplasm (see pancreas; Figure 2F). Acinar cell hypertrophy of the salivary gland, as well as vacuolation and apoptosis in the pancreas, were also evident in some rats in the thirteen-week study; ultrastructural changes in the pancreas taken from a rat treated with sunitinib for fourteen days at 45 mg/kg/day are represented in Figure 3D. These changes were not evident in either species at the end of the recovery period.

Increased splenic hematopoiesis was evident in rats given 6 mg/kg/day in the six-month study but was not evident in the monkey studies. This change was characterized by increased numbers of erythroid and myeloid precursors in the red pulp of the spleen. Complete reversal of the increased hematopoiesis occurred at the end of the recovery period.

Lymphoid atrophy occurred in the studies and was a consistent finding in rats and monkeys at ≥ 5 and ≥ 6 mg/kg/day doses, respectively. This change was characterized by lower than normal numbers of lymphocytes in the cortical thymus and, to a lesser extent, in the white pulp of the spleen and in the cortex of the lymph nodes. Lymphoid atrophy was completely reversed or showed partial reparation at the end of the six- to eight-week recovery periods in both species.

Intestinal changes occurred in rats given ≥ 5 mg/kg/day in the thirteen-week study (but not in the six-month study) and in monkeys given ≥ 6 mg/kg/day. In the small intestine, these changes were characterized by glandular hyperplasia that was sometimes associated with inflammation in the intestinal wall. Large intestinal (colon, cecum, rectum) changes occurred in the monkey and were characterized by a reduction in size and/or number of crypt goblet cell vacuoles, often associated with a reduction in lamina propria lymphocytes. These changes had reversed completely by the end of the recovery period.

Renal changes also occurred in the six-month rat and nine-month monkey studies at 6 mg/kg/day. The change was evident in the rat as a more pronounced chronic progressive nephropathy compared with controls, and in the monkey as increased glomerular mesangial matrix. It is noteworthy that at the end of the recovery period, the change had reversed completely in the monkey but was still evident in the rat. The reason for this disparity is not clear, but the well-recognized progressive nature of chronic nephropathy in the rat (Hard and Kahn 2004) might have confounded the potential reversibility of a renal change in this species.

Toxicokinetics

Across the studies, systemic exposure to sunitinib and SU12662 increased with increasing dose (Haznedar JO, manuscript in preparation). The total drug (sunitinib plus SU12662) AUC_(0–24) values associated with treatment-related signs of toxicity (e.g., histopathologic findings such as adrenal hemorrhage, pancreatic atrophy, or bone growth plate thickening) were ≥ 2097 ng*hr/mL in monkeys and ≥ 6407 ng*hr/mL in rats, as shown in the thirteen-week studies. These values are approximately 1.1 and 3.3 times greater, respectively, than the typical mean human exposure (1929 ng*hr/mL) observed at the recommended clinical dose of 50 mg/day (Pfizer Inc., data on file).

Discussion

In rats, sunitinib was generally tolerated at doses of 0.3 and 1.5 mg/kg/day, and most adverse findings were reversible or were in the process of reversal during the six- to eight-week recovery period. In monkeys, the no observed adverse effect level (NOAEL) was identified as 1.5 mg/kg/day, and findings were considered similarly reversible, except for uterine/ovarian weight changes and skin pallor that seemed likely to require a much longer time to resolve. The mechanism by which the skin pallor persisted after recovery remains unclear. Numerous cases of skin and hair color changes in patients undergoing sunitinib treatment have been reported (sunitinib prescribing information), potentially implicating effects of sunitinib on the function of KIT-positive melanocytes associated with hair follicles, as a potential mechanistic explanation for these findings (Moss et al. 2003). Clinical signs and mortality were typically associated with maximally tolerated doses of sunitinib, and tolerability appeared inversely proportional to the duration of dosing. Rats and monkeys given high oral doses of sunitinib appeared hypoactive and exhibited GI effects (diarrhea/loose stool, emesis, dehydration), sometimes associated with excessive salivation. Hypoactivity appeared to be a result of poor health in the animals at high doses, whereas the GI effects may have been associated with increased local concentrations of sunitinib in the GI tract. Oral dosing of another multitargeted RTK inhibitor, imatinib, caused emesis and diarrhea in monkeys, suggesting a local irritation effect (Cohen et al. 2002). In monkeys, death followed declining health, which was observed after prolonged incidences of emesis and loose stools, with concomitant decreases in food consumption and body weight; the ultimate cause of death could not be determined because of multi-organ system effects observed at the highly toxic dose levels.

Adrenal cortical congestion and hemorrhage were observed in monkeys, and hemorrhagic necrosis was observed in rats. Although the mechanism of sunitinib-induced adrenal toxicity is not fully understood, further investigation using electron microscopy (Figure 3) in a two-week rat study suggested that sunitinib induced dose- and time-related early impairment of the capillary microvasculature in the adrenal cortex. The study also suggested that the capillary lesions preceded the damage observed in the parenchyma (cortical cell necrosis). A comprehensive assessment of corticosterone/cortisol and aldosterone levels was not performed, but sparse sampling in this two-week rat study revealed compound-related aldosterone and corticosterone hormonal changes (initial decreases on Days 2 and 7 and subsequent increases at the end of treatment, Day 14). The hormonal changes on Day 2 occurred before the appearance of adrenal morphologic changes on Day 4 (data not shown). The mechanism of sunitinib-related adrenal toxicity is unknown, but the effect on hormone concentration prior to evidence of morphologic changes could be indicative of a multifactorial event that resulted in this finding. Although adrenal changes were not reported in monkeys given the anti-VEGF antibody bevacizumab for thirteen weeks (Ryan et al. 1999), the adrenal gland has been reported as being a particularly susceptible endocrine tissue to compound-induced lesions (Ribelin 1984).

Bone marrow effects were characterized by hypocellularity of the erythroid and myeloid cell lineages in rats and monkeys, in most cases associated with dose-related decreases in organ weights (thymus and spleen) and lymphoid depletion of hemolymphopoietic organs (thymus, spleen, and/or lymph nodes). Additionally, hematologic findings (red and white blood cell parameters) in both species generally correlated, probably reflecting the histologic changes in the bone marrow and lymphoid organs. However, there was some difference in the effect on the hematopoietic components of the spleen between rats that were dosed on consecutive days (thirteen-week study) versus cyclically (six-month study), as evidenced by increased hematopoiesis in the six-month study. It seems likely that the cyclic daily dosing regimen enabled a splenic hematopoietic response during the dosing phase in the rodent. Bone marrow effects and associated changes in the hemolymphopoietic organs resolved in both rats and monkeys upon discontinuation of treatment. Although systemic stress likely occurred and contributed to the severity of the changes in the spleen and thymus in animals given high doses of sunitinib, a direct effect of sunitinib on the thymus and spleen also seems likely, given that such changes occurred in these organs in animals in good body condition. In addition, the potent inhibitory selectivity of sunitinib against the KDR (VEGF-2) receptor, which has been revealed to be present on hematopoietic cells (Ziegler et al. 1999), further supports a likely direct effect on cells in the spleen and thymus.

Sunitinib-related effects on the hematopoietic system following treatment were not unexpected and could be attributed to the inhibition of VEGFR-2 and/or other RTK targets. For example, dysregulation of the activity of several RTKs (e.g., KIT and PDGFR-β/fibroblast growth factor [FGF]-1 receptors) has been implicated in the progression of a variety of hematopoietic malignancies (Correll et al. 2006; Reilly 2003), suggesting that sunitinib-related effects on the hematopoietic system may involve multiple RTK pathways. Additionally, myelosuppression was a direct pharmacologic effect of imatinib (PDGFR, KIT, and Bcr-Abl inhibitor) administration in the rat, dog, and monkey (Cohen et al. 2002).

Reversible thickening (physeal dysplasia) of the growth plate, characterized by disrupted endochondral ossification of the long bones, was observed in sunitinib-treated rats and monkeys. Growth plate thickening has also been observed in monkeys following treatment with bevacizumab (Ryan et al. 1999). Normal ossification of the growth plate is dependent upon invasion of the growth plate by capillaries, which is necessary to initiate calcification, bone production, and bone resorption at the growth plate (Carlevaro et al. 2000). VEGF is an important factor for the growth of these capillaries (Ferrara 2001). The principal reason for the thickening of the growth plate is the accumulation of hypertrophic chondrocytes. The physiologic thickness of the growth plate is usually maintained by the synchronized orchestration of physeal cell proliferation and degeneration, which includes the cellular turnover of hypertrophic chondrocytes at the chondro-osseous junction. The importance of vascular-mediated contributions to physiologic growth plate morphology is highlighted by the development of the same growth plate morphology of increased numbers of hypertrophic chondrocytes after experimental interruption of the metaphyseal vascular supply (Trueta and Amato 1960). Sunitinib is an inhibitor of at least three RTKs believed to be involved in vascular growth or maintenance. The consistent finding of physeal dysplasia caused by an anti-VEGF antibody and by sunitinib suggests that sunitinib inhibits the pharmacologic effects of VEGF signaling on angiogenesis at the physeal growth plate of maturing animal bones.

The periosteal and diaphyseal cartilage likely occurred secondarily to the physeal dysplasia. Although the precise mechanism by which the periosteal cartilage developed was not examined in our studies, it is recognized that changes in the physical forces acting on bone can cause bone deformations and skeletal pathologies (Woodard and Jee 1991), and such biomechanical changes probably occurred because of the physeal dysplasia. Bone is usually formed on the periosteum through a process of intramembraneous ossification that does not involve a prior cartilaginous core (Thompson 2007), suggesting that the periosteal cartilage probably originated from cartilage in the physis as modeling or growth of the bone occurred. Additionally, the metaphysis progressively becomes a part of the diaphysis during normal bone growth (Woodard and Jee 1991), a process that may have resulted in unmineralized cartilage becoming a part of the diaphysis of the bone during bone modeling in the rat.

Decreased ovarian and uterine weights were observed in sunitinib-treated rats and monkeys, with decreased ovarian follicular development and uterine endometrial atrophy also observed in monkeys. The decreases in ovarian and uterine weights observed in the monkey study at 6 mg/kg/day correlated histologically with the absence of active corpora lutea, decreased numbers of developing follicles, and increased numbers of atretic follicles in the ovaries, and with thinning of the endometrium and dense stratum spongiosum in the uterus. Decreases in ovarian and uterine weights and impaired development of corpora lutea in the ovaries were also reported in monkeys after treatment with bevacizumab (Ryan et al. 1999). In that study, the failure of corpora lutea to develop appeared to be caused by a failure of the cyclic growth of blood vessels that normally occurs in the female reproductive tract. The cause of the endometrial atrophy seen in the present study, in addition to that in the oviduct, cervix, and vagina, could not be ascribed unequivocally to inhibition of VEGF signaling, since no uterine alterations were seen in the study in which monkeys were given bevacizumab. However, the endometrium is a complex tissue consisting of different cell types, and cyclic endometrial growth is dependent on its ability to regenerate a vascular capillary network. Regulation of endometrial angiogenesis is mediated by VEGF and its receptors, but this involves a complex interaction between cell types and is affected by both endocrine and paracrine pathways (Taylor et al. 2001). It is likely that multiple VEGF receptors are involved in normal endometrial development, and the expression of VEGF and its receptors in the endometrial tissue may exhibit cyclic variations (Punyadeera et al. 2006). Although systemic stress likely contributed to the development of ovarian and uterine changes at high doses in the rat and monkey studies conducted with sunitinib, the presence of these changes at dose levels not associated with indications of pronounced toxicity, and their persistence at multiple dose levels through the recovery phase, suggest that direct effects of sunitinib likely contributed to these changes.

Incisor dental caries occurred only in the rat in these studies and are likely to reflect specificity to growing teeth. There was no evidence of effects of sunitinib on nongrowing, premolar, or molar teeth. The underlying antiangiogenic action of sunitinib is a mechanism that possibly resulted in perturbation of dentine and enamel formation, rendering the incisors more susceptible to breaking, as these teeth grow continuously throughout the lifespan of rodents. This growth occurs through mechanisms involving both vasculogenesis and angiogenesis, as evidenced, respectively, by the presence of: (1) vacuoles in clusters of cells between and intercalated with blood capillaries, initially in proximity to the outer enamel epithelium; and (2) the loss of basal lamina, the proliferation of endothelial cells, and the presence of filopodia and lateral sprouting (Law et al. 2003; Manzke et al. 2005).

Alterations in the exocrine pancreas (acinar cell degranulation or atrophy) were observed in rats and monkeys, mainly with high-dose sunitinib. This finding has not been reported in monkeys treated with bevacizumab (Ryan et al. 1999), suggesting that the observed change is not likely to be related to inhibition of VEGF signaling. Multiple receptors on pancreatic acinar and duct cells and intracellular signaling pathways (some involving tyrosine kinases) are likely to regulate pancreatic physiology and may be altered in pathophysiologic states (Bi and Williams 2004; Satoh et al. 2005; Williams 2002).

The yellow discoloration of skin, fur, and/or urine was most likely caused by the yellow color of the test drug or metabolite(s). Sunitinib malate is a yellow to orange powder, and the wide tissue distribution of the drug or its metabolites and the recognition of the renal system as a primary route of excretion are consistent with the yellow discoloration evident in some tissues and the dark yellow color of the urine. The findings from the rat and monkey studies, including the assessment of clinical pathology parameters, did not support another cause of the yellow discoloration.

These findings suggest that, in the rat and monkey, inhibition of multiple important RTK pathways by sunitinib may result in pharmacologic effects on a variety of organ systems. Key pharmacologic effects of sunitinib include reversible inhibition of neovascularization into the epiphyseal plate during growth, and impaired formation of corpora lutea and uterine development during the estrous cycle. Similar toxic effects have been observed in nonclinical studies in rats and monkeys treated with other VEGF/PDGF signaling inhibitors, including bevacizumab and sorafenib (Ryan et al. 1999; Nexavar Summary of Product Characteristics 2007). For example, similar observations relating to physeal dysplasia of the tibial bone and impaired formation of corpora lutea have been noted in monkeys treated with the anti-VEGF antibody bevacizumab (Ryan et al. 1999). It was recently demonstrated that some capillaries (containing endothelial fenestrations and a relatively high level of expression of VEGFR-2 and VEGFR-3) in adult mouse pancreatic islets, thyroid, adrenal cortex, pituitary, small intestinal villi, choroid plexus, kidney, and epididymal adipose tissue are dependent on VEGF signaling for survival (Kamba et al. 2006). After the removal of inhibitory signals, capillaries in certain microvascular beds regrow, thus exhibiting VEGF-dependent plasticity. The time course of capillary regression begins within two weeks of treatment, with the loss of vessel patency, intraluminalfibrin deposition, and cessation of blood flow (Baffert et al. 2006); apoptosis and the loss of endothelial cells follow. Pericytes either retract from regressing endothelial cells onto adjacent normal vessels or undergo regression themselves, while empty sleeves of basement membrane remain temporarily at sites of endothelial cell regression and assist in capillary regrowth (Mancuso et al. 2006). The findings noted for this class of RTK inhibitor in the current study are consistent with pharmacologic perturbations of physiologic, cellular, or angiogenic processes resulting from interaction with the intended molecular targets.