Cartilage Dysplasia and Tissue Mineralization in the Rat Following Administration of a FGF Receptor Tyrosine Kinase Inhibitor

Abstract

PD176067 is a reversible and selective inhibitor of fibroblast growth factor receptor tyrosine kinase, and was in preclinical development as an angiogenesis inhibitor for the treatment of solid tumors. A 14-day oral toxicity study of PD176067 in young female rats (7 weeks old) was conducted at doses of 2.5, 5, and 10 mg/kg/day (15, 30, and 60 mg/m², respectively). Skeletal changes, and vascular and soft tissue mineralization were observed as primary drug-related toxicities. To determine if these changes are specific to young, rapidly growing animals with increased vascular and osseous development, PD176067 was administered to mature (11 months old) rats. Female rats received PD176067 by gavage for 14 days at doses of 2.5, 5, and 10 mg/kg/day and necropsied on day 15. Clinical signs of toxicity were seen at ≥5 mg/kg and one death occurred at 10 mg/kg. Physeal dysplasia (distal femur, proximal tibia, sternum) occurred in all drug-treated animals and was characterized by dose-related increased thickness of the zones of chondrocyte proliferation and hypertrophy, and marked thickening of the zone of ossification. Cartilage hyperplasia was characterized by proliferation of chondrocytes along margins of the synchondrosis and subperiosteum of sternebrae. Serum phosphorus levels increased 47% and 166% at 5 and 10 mg/kg, respectively. Mineralization of cardiac myocytes, aorta, various arteries, renal tubules, and gastric mucosa and muscularis was seen at 10 mg/kg, and consistent with the presence of calcium-phosphorus deposition. Physeal changes occurred at similar plasma PD176067 exposures in young and mature rats (AUC ≥ 4.83 μg · hr/mL). PD176067 produced morphologically similar lesions in young and adult rats.

Introduction

Solid tumors require angiogenesis (formation of new blood vessels) for continued growth and it is believed that inhibition of angiogenesis can serve as a therapeutic strategy for treating cancer (Folkman et al., 2001). In order for a solid tumor to grow beyond 2–3 mm, sufficient vascularization is required for the delivery and removal of metabolic products (Folkman, 1971). By inhibiting the formation of new blood vessels, tumor regression may occur due to increased tumor cell death in the presence of decreased tumor cell growth and division (Pluda, 1997). Angiogenic factors, such as acidic and basic fibroblast growth factors (FGFs) and vascular endothelial growth factors (VEGFs) stimulate endothelial cells to grow, divide, migrate, and differentiate, resulting in vascular development (Folkman and Klagsbrun, 1987; Neufeld et al., 1999; Cross and Claesson-Welsh, 2001). FGFs and VEGFs produce their effects by binding to cell surface receptors that have tyrosine kinase activity. These kinases subsequently phosphorylate tyrosine residues on proteins in an ATP-dependent fashion, resulting in progression of various signal-transduction pathways. Binding of FGF to its receptor (e.g., FGF receptor-1) results in receptor dimerization and autophosphorylation. Activation of various intracellular signaling pathways subsequently occurs, including Ras, Src, PI3K, and PLC pathways (Cross and Claesson-Welsh, 2001). The net result is that cell surface binding of growth factors results in initiation of signal-transduction pathways leading to endothelial cell proliferation and vascular development (Cross and Claesson-Welsh, 2001).

The utility of targeting angiogenesis in cancer therapy was validated with the anti-VEGF monoclonal antibody, bevacizumab (Avastin). Bevacizumab binds VEGF and prevents the interaction of VEGF with its tyrosine kinase receptors. Administration of bevacizumab to patients with metastatic renal and colorectal cancers resulted in significant clinical activity (Yang et al., 2003; Hurwitz et al., 2004). Bevacizumab was approved in 2004 by the U.S. FDA for use in 5-fluorouracil-based chemotherapy in patients with metastatic carcinoma of the colon or rectum.

PD176067 (Figure 1) is a small molecule that demonstrated inhibition of tumor growth in mice and was in preclinical development as an angiogenesis inhibitor for the treatment of solid tumors. PD176067 is a reversible, selective ATP-competitive inhibitor of FGF receptor tyrosine kinase, with in vitro IC₅₀ values of 2–9 nM against human FGF receptor-1 (Parke-Davis, 2000). Depending upon the in vitro assay used, inhibition of VEGF receptor tyrosine kinase is also observed, although at higher concentrations than for FGF receptor tyrosine kinase. For example, PD176067 has an in vitro IC₅₀ value of 16 nM against human VEGF receptor-2, and is 19-fold less potent with inhibiting in vitro endothelial cell proliferation dependent upon VEGF stimulation in comparison to basic FGF (Parke-Davis, 2000). PD176067 is structurally related to a previously described compound of the pyrido[2,3-d]pyrimidine class that is an inhibitor of FGF receptor tyrosine kinase (Mohammadi et al., 1998).

To evaluate the nonclinical toxicity of PD176067, young female rats (approximately 7 weeks old) were given bidaily oral doses at 2.5, 5, and 10 mg/kg/day for 14 days. This dosing regimen resulted in physeal dysplasia, and vascular and soft tissue mineralization. Although the mechanism(s) for elicitation of these lesions was not known, it was initially believed that the lesions resulted from inhibition of angiogenesis due to the intended pharmacology of the drug. A second study was conducted to determine whether PD176067 produced similar lesions in mature rats. The initial hypothesis was that drug-related lesions would not occur, or would occur to a limited extent in mature animals in which skeletal growth and neovascularization are typically reduced.

Materials and Methods

Test Compound

PD176067 was synthesized by Pfizer Global Research and Development, Ann Arbor Laboratories as a HCl salt. Dosing suspensions were prepared in aqueous 0.5% methylcellulose at concentrations of 0.25 to 1 mg/mL.

Animals

In the initial 2-week study, young female Sprague–Dawley (Crl:CD^® (SD)IGS BR) rats obtained from Charles River Laboratories were approximately 7 weeks old and weighed 157 to 180 grams at the start of dosing. The second rat study utilized mature female Wistar rats obtained from Charles River Laboratories that were approximately 9 to 11 months old and weighed 334 to 574 grams at the start of dosing. Strain differences were due to animal availability. Clinically acceptable animals were assigned to dose groups for each study using a computer-assisted randomization procedure to ensure equivalent distribution of body weights. Animals were housed individually in stainless steel wire mesh cages. Standard procedures/conditions were applied for animal care, feeding, and maintenance of room, caging, and environment. Animals were fed powdered Lab Diet (5002) Certified Rodent Diet ad libitum in stainless steel food containers. Water was supplied ad libitum via an automatic watering system. Animals were fasted overnight prior to scheduled euthanasia. The studies were conducted in accordance with Canadian and U.S. guidelines for animal welfare (Animals for Research Act of Ontario, 1980, and as amended in 1989, and guidelines of the Canadian Council on Animal Care; National Research Council Guide for the Care and Use of Laboratory Animals, 1996). The procedures used were reviewed and approved by the Pfizer Institutional Animal Care and Use Committee.

Experimental Design

Animals in both studies were dosed twice daily by gavage (BID), approximately 6 hours apart, at total daily doses of 2.5, 5, or 10 mg/kg/day (15, 30, or 60 mg/m²/day, respectively) for 14 days. Drug was administered in a dose volume of 5 mL/kg/dose (10 mL/kg/day total dose volume). Control animals received vehicle alone. The required volume of vehicle or drug suspension for each animal was based on the most recent individual body weight. Dosing was initiated on day 1 and animals were necropsied on day 15.

The study involving young rats was comprised of 11 rats/drug-treated group and 5 rats for the vehicle control. The first 5 rats/drug-treated group were designated for toxicological assessment and the last 6 rats/drug-treated group were designated as a satellite group for plasma drug level determination only. The study involving mature rats was comprised of 5 animals/group (for both control and PD176067). Plasma drug levels were measured in all drug-treated mature rats.

Experimental Procedures

All animals were observed at least 3 times daily for clinical signs of toxicity. Physical examinations were performed pretest and during week 2. Body weights were recorded pretest (for randomization), and prior to dosing on days 1, 7 or 8, and 14. Individual food consumption was recorded weekly in young animals only.

The following serum biochemistry parameters were evaluated at scheduled termination from fasted young rats: alanine aminotransferase, albumin, albumin/globulin (calculated), alkaline phosphatase, amylase, aspartate aminotransferase, bilirubin, calcium, chloride, cholesterol, creatine kinase (total), creatinine, glucose, globulin (calculated), phosphorus, potassium, protein (total), sodium, triglycerides, and urea nitrogen. On day 14, prior to dosing, blood samples (approximately 1 to 2 mL) were collected from nonfasted mature rats and serum analyzed for calcium, phosphorus, and alkaline phosphatase (total and individual isoenzymes—liver, intestine, and bone).

For toxicokinetic analyses, blood samples (approximately 1 mL heparinized blood) were collected after 14 days of dosing at 1, 2.5, 6, 7.5, 12, and 24 hours after the first daily dose. The 6-hour sample was collected prior to the second daily dose. In order to maximize the use of animals, 3 blood samples were collected from each animal. The first 2 or 3 animals/group for plasma drug level determination were bled at 1, 6, and 12 hours postdose. The second 3 animals/group were bled at 2.5, 7.5, and 24 hours postdose. The blood was centrifuged at approximately 4°C, and the plasma separated and stored at ≤ −20°C. Plasma concentrations of PD176067 were determined using a validated LC/MS/MS method using positive ion Turbo IonSpray with multiple reaction monitoring and a lower limit of quantitation of 5 ng/mL. Toxicokinetic parameters of maximum plasma concentration (Cmax), time-to-Cmax (tmax), and area under the plasma concentration-time curve (AUC (0–24)) were calculated. For purposes of toxicokinetic analyses, the 24-hour time point was considered equivalent to time zero. Following blood sampling, satellite animals were euthanized and discarded.

All surviving animals were euthanized on day 15 by carbon dioxide asphyxiation and complete necropsies were performed. Representative samples of the following tissues were collected at necropsy, fixed in 10% buffered formalin, and processed for light microscopy: aorta, bone (including distal femur, proximal tibia, and sternum), bone marrow, heart, kidneys, gastrointestinal tract, mesenteric artery, and pancreas. Tissues examined microscopically were stained with hematoxylin and eosin, and the Von Kossa calcium stain was used on select tissues.

Statistical Analysis

Arithmetic means and standard errors were calculated for body weight, weight change, food consumption, and clinical laboratory data. Treatment comparisons were performed on rank-transformed data using 1-factor analysis of variance (ANOVA) focusing on a dose-trend test sequentially applied at the 2-tailed 1% significance level. Statistical analyses were conducted separately within each study.

Results

Clinical Signs

Clinical signs of toxicity did not occur at 2.5 mg/kg in either young or mature rats. Reduced feces, salivation, urine staining, and red staining of the pelage were noted sporadically at 5 mg/kg in young rats. A significant number of clinical signs occurred in young rats at 10 mg/kg that included chromodacryorrhea, salivation, red staining of the muzzle and/or pelage, urine staining, reduced feces, hunched posture, paresis, dyspnea, dehydration, and/or thin body condition. Subsequently, 2 animals at 10 mg/kg were euthanized moribund on day 14. There were no effects on body weight or food consumption at 2.5 or 5 mg/kg. Young animals at 10 mg/kg lost an average of 26.5 grams in body weight (15%) during week 2 and this was associated with a 47% decrease in food consumption (data not shown).

Clinical signs occurred in a dose-dependent fashion in mature rats at ≥5 mg/kg, primarily in week 2. Chromodacryorrhea, red staining of the muzzle, fecal/urine staining, and fecal changes (soft, reduced, absent) were observed at ≥5 mg/kg. Animals at 10 mg/kg also exhibited diarrhea, mucoid feces, salivation, tremors, hypoactivity, and rough pelage. One animal at 10 mg/kg was found dead on day 15. This animal exhibited clinical signs of toxicity consistent with other animals at 10 mg/kg, along with hypoactivity, and had total body weight loss of 24% from day 1. Mean body weight loss of 22% was seen by day 14 in mature rats at 10 mg/kg (data not shown), and was likely due to decreased food consumption (not measured), suggested by reduced and/or absent feces in these animals. In summary, the clinical toxicity of PD176067 was similar between young and mature rats, with deaths at 10 mg/kg and no clinical signs at 2.5 mg/kg.

Serum Chemistry

Serum chemistry changes in young rats are presented in Table 1. A trend for increased serum phosphorus was seen at 2.5 and 5 mg/kg (increased 7% and 9%, respectively); although the changes were not statistically significant. Serum phosphorus was not evaluated at 10 mg/kg due to insufficient sample volume. Calcium was significantly decreased 9% at 10 mg/kg, which correlated with a decrease in serum albumin (12%). No other changes in serum chemistries were considered toxicologically significant in these animals.

In mature rats, serum phosphorus levels increased 47% and 166% in the 5 and 10 mg/kg groups, respectively, and calcium was increased 5% at 10 mg/kg (Table 2). The 36% decrease in the intestinal isoenzyme component of alkaline phosphatase at 10 mg/kg is consistent with decreased food consumption (Evans, 1996). No other changes in alkaline phosphatase (total, and liver and bone isoenzymes) were seen. Serum chemistry changes were not seen at 2.5 mg/kg.

Histopathology

The incidence of drug-related microscopic findings observed in young and mature rats is presented in Table 3. Vascular and soft tissue mineralization was present in both studies. Minimal to moderate mineralization in the media and subintima of the aorta (Figure 2A), the subintima of mesenteric arteries and/or arterioles in the pancreas were seen in young rats at ≥5 mg/kg and mature rats at 10 mg/kg. Positive Von Kossa staining indicated that the tissue mineralization was due to calcium-containing deposits (Figure 2B). One mature rat that died on day 15 also had mineralization of cardiac myocytes and transmural fibrinoid necrosis of the aorta. Minimal to moderate mineralization of the gastric fundic mucosa occurred in young and mature rats at 10 mg/kg, and was associated with gastric discoloration noted grossly. Gastric mineralization was more extensive in mature rats at 10 mg/kg and also included mineralization of submucosal arterial intima and subintima, and muscularis. Renal changes observed in mature rats at 10 mg/kg included mineralization of renal tubular basal lamina and subintima of renal arteries. Drug-related gross or microscopic changes were not seen in young rats at 2.5 mg/kg.

Abnormal endochondral ossification induced by PD176067 and classified as physeal dysplasia was present in the distal femur and proximal tibia growth plates in young rats at ≥5 mg/kg and mature rats at all doses. In young rats, progressive thickening of the physis was dose-related and severity of physeal dysplasia was minimal, mild, and mild to moderate at 2.5, 5, and 10 mg/kg, respectively. There was an approximately 4-fold increased thickness of cartilaginous growth plates in young rats at 10 mg/kg compared to age-matched controls (Figure 3A and 3C). In young rats, the proliferating zone had increased numbers of flattened chondrocytes aligned in columns and was twice the thickness as compared to control rats. There was increased depth of the zone of chondrocyte hypertrophy, and marked thickening of the zone of ossification (primary spongiosa). Disorganization of the distal columns of hypertrophic chondrocytes, variable enlargement of the perichondrial lacunae, and increased numbers of primary spongiosa with retention of cartilaginous cores and thin rims of osteoid lining the trabeculae were present (Figure 3C). Capillary loops were present within the tubes of mineralized cartilage matrix (Figure 3E). However, there was decreased resorption or amalgamation of primary spongiosa into secondary spongiosa.

In mature control rats, the narrow, inactive physes were sealed by a layer of bone along the zone of ossification, and the metaphyses contained a few thick trabeculae (Figure 3B). With administration of PD176067 to mature rats, there was enhanced proliferation and maturation of chondrocytes. There was increased prominence of the stacks of proliferating chondrocytes and an approximately 2-fold increased thickness of the zone of hypertrophy (Figure 3D). The metaphysis contained irregular thick trabeculae with retention of chondrocytes and cartilage cores (Figure 3F).

In both young and mature rats, the synchondroses of sternebrae were thickened by proliferating and hypertrophic chondrocytes (Figure 4A, 4B, and 4C). The cartilage hyperplasia elevated and extended along subperiosteal margins of the sternebral body with resemblance to endochondral ossification in young rats. (Figure 4D).

Within the metaphyseal region, bone marrow elements were decreased in young rats at 10 mg/kg (minimal) and at ≥5 mg/kg in mature rats (minimal to moderate). Bone marrow hypocellularity was considered secondary to physeal dysplasia and disruption of metaphyseal bone marrow microenvironment.

Toxicokinetics

Plasma toxicokinetic parameters of PD176067 in young and mature rats are presented in Table 4. Administration of PD176067 to young rats resulted in plasma concentrations that increased in a greater than dose-proportional fashion. Cmax occurred at 7.5 hours, which was 1.5 hours after the second dose. Similarly, plasma drug levels increased in a greater than dose-proportional fashion in mature rats, suggesting saturation of clearance mechanisms with increasing dose. This was most evident at 10 mg/kg in which AUC (0–24) and Cmax increased approximately 8-fold in comparison to the values at 5 mg/kg. PD176067 AUC (0–24) values were approximately 2- to 11-fold higher in mature rats than young rats. The difference in plasma drug levels between the 2 studies may have been due to strain and/or age differences in absorption, metabolism, and/or clearance.

Discussion

Oral administration of PD176067 in young rats (7 to 9 weeks old) for 2 weeks produced vascular and soft tissue mineralization, and growth plate changes. Initially, it was believed that growth plate development is dependent upon neovascularization and that vascular mineralization can occur secondary to alterations in bone metabolism and growth. Because PD176067 is an inhibitor of FGF receptor tyrosine kinases, it was hypothesized that these lesions arose due to the pharmacologic inhibition of angiogenesis. Therefore, it was proposed that drug-induced lesions would not occur, or would occur to a limited extent, in mature animals in which neovascularization is typically reduced. To test this hypothesis, PD176067 was dosed for 2 weeks in mature rats (9–11 months old).

Administration of PD176067 to mature rats at ≥2.5 mg/kg resulted in physeal dysplasia. This lesion was also observed in young rats at ≥5 mg/kg. PD176067 exposure levels in which physeal dysplasia occurred were AUC (0–24) ≥4.89 and ≥ 4.83 μg ·hr/mL in mature and young animals, respectively. These data suggest that there were no differences in susceptibility for lesion development between young and mature rats. Both young and mature female rats received PD176067 by oral gavage at 2.5, 5, and 10 mg/kg. Plasma drug levels increased in a greater than dose-proportionate fashion in both age groups, suggesting that saturation of clearance had occurred with increasing dose. This may have resulted from inhibition of cytochrome P450-mediated metabolism and/or inhibition of clearance. PD176067 has been demonstrated to inhibit various cytochrome P-450 isozymes, with IC₅₀ values for human CYP450 3A4, 2C9, and 2C8 of 2.21, 0.84, and 0.025 μM, respectively (Parke-Davis, 2000).

Chondrocytes produce basic FGF, which can modulate chondrocyte differentiation, and elicit vascular invasion and ossification of cartilaginous cords (Baron et al., 1994). In addition, hypertrophic chondrocytes produce VEGF, which acts to regulate cartilage remodeling, ossification, and angio-genesis in the growth plate (Gerber et al., 1999; Harper and Kalgsbrun, 1999). Oral administration to rats of an inhibitor of VEGF receptor tyrosine kinase (AstraZeneca’s ZD4190) resulted in a marked increase in the femoral physeal zone of hypertrophy (Wedge et al., 2000). Administration of a recombinant humanized anti-VEGF monoclonal IgG antibody (rhuMabVEGF) to young adult cynomolgus monkeys produced physeal dysplasia characterized by increased hypertrophied chondrocytes, subchondral bony plate formation, and inhibition of vascular invasion of the growth plate (Ryan et al., 1999).

Deletion of the murine FGF receptor-3 gene resulted in mice (FGFR-3−/−) that developed bone dysplasia characterized by expansion of proliferating and hypertrophic chondrocytes within the growth plate (Deng et al., 1996). The study by Deng et al. suggested that FGF receptor-3 regulates endochondral ossification by limiting chondrocyte proliferation in the growth plate; therefore, inhibition of receptor function results in chondrocyte proliferation.

Achondroplasia, the most common form of human dwarfism, arises due to a mutation in the FGF receptor-3 gene, leading to receptor overactivity. This “gain-in-function” mutation in the receptor leads to inhibition of chondrocyte proliferation and maturation within the growth plate (Aviezer et al., 2003). PD176067 administration to rats resulted in chondrocyte proliferation and cartilage formation, characterized by increased thickness of both the zone of proliferation and the zone of hypertrophy. Although the mechanism by which this occurs in rats is unknown, inhibition of FGF and/or VEGF dependent signaling pathways appears to be a component. In summary, the current and published data suggest that inhibition of growth factor signaling can lead to increases in chondrocyte proliferation and expansion of the hypertrophic zone, resulting in dysplastic growth of cartilage.

Mineralization of soft tissue and vasculature was evident in young and mature rats at 10 mg/kg. These changes were consistent with the presence of calcium-phosphorus depositions. A correlation between the presence of growth plate changes and tissue mineralization was not evident. Although physeal dysplasia occurred in all mature animals at ≥2.5 mg/kg, mineralization was only evident at 10 mg/kg. Also, 1 mature rat at 10 mg/kg had moderate physeal dysplasia but no soft tissue mineralization. In addition, whereas administration of rhuMabVEGF to young adult cynomolgus monkeys produced physeal dysplasia, tissue mineralization was absent (Ryan et al., 1999). Although the mechanism for the soft tissue mineralization produced by PD176067 is unknown, the etiology does not appear related to inhibition of angiogenesis. Mineralization was present in cardiac myocytes, basal lamina of renal tubules, and gastric mucosa and muscularis. These regions appear distinct from areas in which neovascularization may occur (i.e., capillary beds, microvasculature). Also, evidence of overt endothelial cell injury such as edema, hemorrhage, thrombus/fibrin deposition, cell debris, and/or inflammatory cells, were not observed.

PD176067 was administered to beagle dogs to determine whether the toxicity seen in rats occurs in a nonrodent species (Courtney et al., 2003; Datta et al., 2003). Oral administration of PD176067 to juvenile and adult dogs for up to 2 weeks at ≥40 mg/kg/day produced physeal dysplasia, tissue mineralization, and hyperphosphatemia. Tissue mineralization in dogs occurred independent of animal age and did not appear secondary to alterations in bone physiology (Courtney et al., 2003; Datta et al., 2003).

Tissue mineralization in rats following administration of PD176067 is similar to vitamin D toxicity due to dysregulation in serum calcium (Ca) and phosphorus (P) homeostasis (Grant et al., 1963; Rosenblum et al., 1977; Kamio et al., 1979; Mortensen et al., 1996). Markedly elevated levels of serum phosphorus were seen in mature rats at 10 mg/kg (increased 166%), and serum calcium was slightly increased at this dose. Although serum phosphorus values were not available in young rats at 10 mg/kg, a trend for increased levels was seen at 2.5 and 5 mg/kg. Hyperphosphatemia in the presence of normo- or hypercalcemia can result in an increased (Ca) × (P) product, which is associated with tissue mineralization (Block, 2000, 2001). Elevations in serum phosphorus were also seen in dogs administered PD176067 at dose levels associated with tissue mineralization (Datta et al., 2003).

In conclusion, oral administration of PD176067 to female rats for 14 days resulted in physeal dysplasia and soft tissue mineralization. These lesions were morphologically similar in young and mature rats, and do not appear to be related to inhibition of angiogenesis. Tissue mineralization was associated with elevated serum phosphorus levels and was consistent with calcium-phosphorus deposits. Because PD176067 produced skeletal lesions and soft tissue mineralization in rats and dogs, further study is warranted to determine whether this agent has the potential to be used for the treatment of cancer.