Bone and Joints
Diane Gunson1, Kathryn E. Gropp2 and Aurore Varela3, 1Novartis Pharmaceuticals Corporation, East Hanover, NJ, United States, 2Pfizer Inc., Groton, CT, United States, 3Charles River Laboratories, Inc., Senneville, QC, Canada
Abstract
The skeletal system (bones and joints) fulfills several needs in an animal. The skeleton serves as a scaffold on which muscles can act to produce locomotion, provides protection for vital structures, supplies an internal source for minerals, and harbors an environment that supports hematopoiesis. Due to slow and regimented turnover, the current state and the history of a bone coexist. The gross and microscopic structure of each bone depends on its location in the skeleton and the resulting profile of biomechanical forces, as well as on the species, sex, and age. Knowledge of how bone physiology varies due to each of these factors is key in recognizing and understanding bone pathology. In this chapter, basic bone and joint biology, including anatomy, physiology, recent advances in our knowledge of cell–cell signaling, and bone biomarkers are covered in the earlier sections. Animal models, both those found naturally and those generated by genetic or surgical manipulation, can offer great insight into a pathologic process as long as the strengths and limitations of the model are fully appreciated. A review of how bones and joints respond to injury is used to demonstrate the potential range of responses to injurious and restorative events, and the interrelationships between bone cells, is followed by a summary of mechanisms of bone toxicity by various prototypical substances. Many effects on bone in toxicity studies reflect the desired pharmacology of a therapeutic agent, even if the skeleton is not the intended site of action.
Keywords
Skeleton; joint; bone biomarkers; endochondral ossification; arthritis; fracture; bisphosphonate
Introduction
Toxicology and research studies in drug discovery and development generally are conducted with young animals that are skeletally immature and in an active state of growth. In contrast, many drug candidates are destined for use in an adult population in which skeletal growth has ceased. This means that compounds which affect bone growth can result in dramatic skeletal changes in test animals that may not be relevant to the intended human patient population. However, if children are included as an intended patient population, skeletal changes in nonclinical studies performed using immature animals become much more important for risk assessment. In such cases, juvenile studies in very young animals may be conducted purposely in addition to other standard nonclinical safety studies.
Choice of Species and Skeletal Sites
Toxicology studies to support nonclinical safety in drug development are mostly conducted in rats, mice, dogs, and cynomolgus monkeys. Sprague-Dawley or Han Wistar rats and CD-1 mice are the usual outbred stocks of rodents used for general toxicity studies in which skeletal toxicity may be identified. However, Long–Evans hooded rats are used to study skeletal pharmacology and pharmacokinetics when pigmented animals are required to examine the potential risk posed if compounds bind to melanin. Beagle dogs are the standard nonrodent species used to evaluate skeletal effects of test articles. However, cynomolgus monkeys are an alternative when dogs are not suitable models of human biological processes and/or responses, and are generally used in safety studies for human-derived biomolecules (which often are minimally functional or nonfunctional in nonprimate species).
Studies in drug discovery and basic research for skeletal diseases frequently are conducted in animals not used for safety testing. For example, nude mice may be used for oncology research in bone metastases where intact human tumor explants are placed in the subcutis or isolated tumor cells are injected intravenously. Purpose-bred mongrel dogs are frequently used instead of Beagles in non-GLP efficacy and early toxicity studies.
The choice of which skeletal site(s) to examine depends on the structure, metabolic activity, and biomechanical load. In general, skeletal sites are chosen to permit simultaneous evaluation of bones and bone marrow depots for hematopoiesis. Common sites include an axial long bone (often with a major diarthrodial joint in rodents), the sternum, and sometimes a rib (especially in large animals) or vertebra. In toxicity studies, the femorotibial (stifle/knee) joint along with a length of proximal tibia and distal femur is the standard sample for assessing the structures of bone and diarthroses (freely mobile joints); the sternum is used to examine bone and amphiarthroses (minimally mobile joints) as well as cell-rich bone marrow. Compared to these long bones, lumbar vertebrae in mice contain more cancellous (trabecular or “spongy”) bone, and this difference increases with age. In general, synarthroses (nonmobile joints, like skull sutures) are not evaluated specifically during skeletal toxicology studies.
Rats are generally 6–8 weeks of age at the start of nonclinical safety studies and 10–12 weeks old at the end of a 4-week study. At this age these animals still are growing rapidly, so compounds that affect the physis (growth plate) or bone modeling will cause detectable changes in the morphology of the skeleton. In fact, the growth plate of young rodents is a very sensitive detection system for compound-mediated effects on bone growth. It is important to ensure that long bones are examined microscopically, and that for this purpose they must be collected and fixed appropriately (see below).
Bone architecture in male and female Wistar Han and Sprague-Dawley rats and in many strains of mice exhibits sexual dimorphism. Female rats and mice have a greater volume of cancellous bone than males in the proximal tibia, while males have more cortical bone. In both sexes, the amount of cancellous bone in rodents of most strains decreases markedly in the transition from metaphysis to diaphysis in the long bones. Interestingly, hindlimb unloading for a 2-week period in Wistar Han rats results in greater bone loss in males than in females, with a marked decrease in the number of trabeculae in males. It is thought that the higher mass of trabecular bone in females may be related to estrogen levels and the higher mineral mobilization requirements for reproduction in females, especially during lactation.
The growth plate of a normal rat will become narrower as the period of rapid bone growth ends. In addition, the physis of a young intact male rat will be slightly wider than that of an age-matched, young intact female rat during the period of rapid bone growth. These normal age and sex differences can be readily appreciated by comparing long bone physes from control rats at the end of the dosing portion of a 1-month toxicity study with those that are collected from animals which are a few weeks older at the end of a subsequent 2- to 4-week recovery phase.
Dogs, usually purpose-bred Beagles, are generally 6–9 months of age at the start of an efficacy or toxicity study. At this age, young dogs exhibit very little growth in the length of limb long bones even though growth plates may not be completely closed (see Table 23.1). Thus, examination of the standard bone samples in this species, sternum and distal femur, may not reveal compound-related effects on growth plates, metaphyseal new bone quality and quantity, or remodeling bone. Examination of a rib with the costochondral junction as an additional bone specimen will provide an area where active endochondral ossification is still in progress in dogs well into their second year, and possibly longer.
Table 23.1
Age at Physeal Closure in Usual Laboratory Animals
| Species/Sites | Age at physeal closure | |
| CYNOMOLGUS MONKEYa | ||
| Male | Female | |
| Humerus proximal-distal | 6 years–3 years 5 months | 4 years 9 months–2 years 3 months |
| Radius proximal-distal | 5 years 3 months–5 years 3 months | 3 years 9 months–5 years 9 months |
| Ulna proximal-distal | 5 years–6 years 6 months | 4 years 6 months–5 years 9 months |
| Femur proximal-distal | 6 years–5 years 3 months | 4 years 9 months–4 years 9 months |
| Tibia proximal-distal | 5 years–5 years 3 months | 5 years–4 years 9 months |
| Fibula proximal-distal | 6 years–5 years 3 months | 4 years 9 months–4 years 9 months |
| DOGb | ||
| Humerus proximal-distal | 10/12 months–6/8 months | |
| Radius proximal-distal | 9/10 months–10/12 months | |
| Femur proximal-distal | 6/13 months–6/11 months | |
| Tibia proximal-distal | 6/11 months–5/8 months | |
| RATb | ||
| Humerus proximal-distal | 52 weeksc–6 weeks | |
| Radius proximal-distal | 8/14 weeks–104 weeksc | |
| Femur proximal-distal | 104 weeksc–15/17 weeks | |
| Tibia proximal-distal | 104 weeksc–>16 weeks | |
| RABBITd | ||
| Femur proximal-distal | 19/24 weeks | |
| Tibia proximal-distal | 22/32 weeks | |
| Fibula proximal-distal | 23/32 weeks | |
| GÖTTINGEN MINIPIGSe | ||
| Femur and lumbar vertebrae start to close at 25 and 21 months of age, respectively, with complete closure occurring around 3.5 years of age | ||
aFukuda S, Cho F, Honjo S. Bone growth and development of secondary ossification centers of extremities in the cynomolgus monkey (Macaca fascicularis). Exp. Anim. 1978:27(4):387–397.
bZoetis T, Hurtt ME. Species comparison of anatomical and functional renal development. Birth Defects Res. B Dev. Reprod. Toxicol. 2003 Apr;68(2):111–120.
cIn rats, rapid growth occurs between 1 and 5 weeks of age, but declines by skeletal maturity at 11.5–13 weeks. Until 26 weeks of age, longitudinal growth still continues, but it virtually ceases thereafter.
dKaweblum M, Aguilar MC, Blancas E, Kaweblum J, Lehman WB, Grant AD, Strongwater AM. Histological and radiographic determination of the age of physeal closure of the distal femur, proximal tibia, and proximal fibula of the New Zealand white rabbit. J. Orthop. Res. 1994 Sep;12(5):747–749.
eTsutsumi H, Katagiri K, Takeda S, Nasu T, Igarashi S, Tanigawa M, Mamba K. Standardized data and relationship between bone growth and bone metabolism in female Göttingen minipigs. Exp. Anim. 2004;53(Jul (4)):331–337.
Monkeys [usually cynomolgus (Macaca fascicularis)] often are 2–5 years of age at the start of a toxicity study. When sexually mature animals are required, the males will be in the upper end of this age range. If monkeys are selected using a particular weight range, as is often the case, the females usually are more mature than the males (Table 23.1); the practical implication is that females may have closed growth plates, while the males may still have open growth plates. The distal femur with its articular surface and the sternum are the standard bones examined in primates. However, as with dogs, rib samples may provide access to a more active site of endochondral bone growth, and the proximal tibia may provide a more uniform physis to evaluate than the distal femur.
Other species are sometimes used in specialized bone safety studies, including rabbits, guinea pigs, minipigs, and small ruminants (sheep or goats). Large species are utilized especially when seeking to assess the utility of orthopedic devices slated for human use. Other bone models of interest in these species typically are focused on improved understanding of basic biological processes, such as fracture healing or the pathogenesis of disease [e.g., osteoarthritis (OA)].
Structure, Function, and Cell Biology of Bone and Cartilage
The bone and cartilage in the skeleton consist of extracellular matrix, predominantly type I collagen in bone and type II collagen in cartilage. In bone, mineral crystals, mainly calcium hydroxyapatite, have been laid down to provide a strong scaffold.
Structure of Bones and Joints
The basic macroscopic topology of long bones is a cylindrical shell (cortex) comprised of compact (osteonal or Haversian) bone enclosing a cavity supported by an interior scaffold of cancellous (spongy) bone struts (Figure 23.1). Long bones occur in association with some type of joint surface (articular cartilage, fibrocartilage, etc.) at each end. Bone is anisotropic; its strength is variable depending on the orientation of the load applied, because its structure is not uniform. The cortex is covered by the periosteum on the outer surface and is lined by endocortical bone on the internal surface. Specific areas of cortical bone are termed “envelopes”: periosteal on the outer surface, cortical in the middle, and endosteal (endocortical and trabecular bone) lining the inside. The typical long bone contains four main regions (compartments): the epiphysis, the physis (growth plate), the metaphysis, and the diaphysis (shaft). In healthy animals, cavities in the epiphysis, metaphysis, and diaphysis of long bones tend to contain hematopoietic elements in rodents but often also harbor large amounts of white adipose tissue in larger animals, including dogs, primates, and minipigs. Nutrient blood vessels penetrate the cortex and branch to separately supply the epiphysis/physis and the metaphysis/diaphysis.
At the microscopic level, the osteons in the cortex are arranged as concentric lamellar circles of bone around central spaces (Haversian canals) containing blood capillaries and nerve fibers running parallel to the long axis of the bone (Figure 23.1) in most species. Cancellous bone in the epiphysis and metaphysis is composed of trabeculae which form a porous three-dimensional (3D) mesh that supports the cortex and the subchondral bone beneath the articular cartilage. The topology of trabecular bone has been described as an interconnected network of parallel plates or cylindrical rods. Trabeculae are categorized as primary, secondary, or tertiary spongiosa. Trabeculae of the primary spongiosa are located closest to the growth plate and contain the densest number of plates, while tertiary trabeculae extend in small numbers into the diaphysis as long, thin projections. Primary spongiosa are comprised more of calcified cartilage than bone.
The four cell types in bone [osteoclast, osteoblast, osteocyte, and bone-lining cell (BLC)] can be present in a single microscopic field at an active bone remodeling site. Osteoclasts are large cells that undertake bone resorption. Osteoclasts usually are multinucleated except in mice, where they are generally mononuclear. They are located in resorption pits (Howship’s lacunae) on the trabecular surface; osteoclast functional polarity is indicated by positioning of the nuclei away from the ruffled border lying on the bone surface. Osteoblasts, which secrete the protein matrix (osteoid) of new bone, line a forming surface of new bone in a single layer and are cuboidal when most active. Some osteoblasts will become BLCs on the newly completed bone surface; other osteoblasts are engulfed by osteoid during the bone formation process and become osteocytes within lacunae of osteons in the bone tissue. Osteocyte signaling by these embedded cells is important for bone maintenance.
The collagen fibrils in most normal, mature bone tissues have a lamellar orientation that is parallel to the contour of the bone surface. Under polarized light, the collagen fibrils have an alternating light/dark appearance (Figure 23.2). The regular organization of collagen fibrils in lamellar bone provides great strength. In rapidly formed bone, such as beneath the growth plate of a young animal or in the bony callus of a healing fracture, the collagen fibrils are randomly oriented because the osteoblasts are not aligned; such bony areas have a crosshatched or woven appearance under polarized light, and thus are termed woven bone. Woven bone is formed quickly during times of maximal osteoblast activity, such as during early fracture healing, but due to the random fibril orientation is not as strong as lamellar bone. Woven bone osteocytes are larger, more numerous, and randomly placed compared to the orderly arrangement of osteocytes in osteons of lamellar bone. When viewed in hematoxylin and eosin (H&E)-stained sections, woven bone is paler than mature lamellar bone.
Diarthrodial (freely mobile) joints are comprised of multiple structures. Articular cartilages and their underlying subchondral plates of epiphyseal bone (which together comprise the articular–epiphyseal complex) from two opposing joint surfaces bear the load. The joint capsule encloses the joint fluid (a lubricant), which is secreted by the synovial membrane lining the capsule. Ligaments, attaching bones to bones, and tendons, connecting skeletal muscle to bones, provide stability to joints. Highly loaded diarthrodial joints like the knee may possess additional shock-absorbing structures, such as the fibrocartilage menisci and pads of white fat. Microscopic synovial villi may project into specialized recesses, such as niches between the large fat folds.
Articular cartilage classically has four zones or layers containing hyaline cartilage and chondrocytes. These zones are (I) the smooth superficial zone, which minimizes friction in conjunction with the lubrication provided by synovial fluid; (II) the transitional zone; (III) the deep radial zone, which absorbs shock; and (IV) the calcified (“mineralized”) cartilage zone, which interfaces with the underlying subchondral bone of the epiphysis (Figure 23.3). At the top of the superficial layer is the thin acellular lamina splendens. The chondrocytes and collagen in the superficial layer run parallel to the joint surface, but they change their orientations in the transitional zone to run perpendicular to the joint surface in the deep radial and calcified cartilage layers (Figure 23.3). The tidemark denotes the boundary between the deep radial layer (III) and the calcified cartilage layer (IV). Articular cartilage is composed of aggrecan (the major proteoglycan in cartilage) and type II collagen.
The synovial membrane is a thin, highly vascular lining covering the inner surface of the articular capsule as well as the surfaces of intraarticular ligaments and tendons. It is usually composed of two layers: a thin internal surface layer that is two to three cells deep, and a deeper subintimal layer of loose or fibrous connective tissue. Synovial cells are of two kinds. Type A (macrophage-like or M) cells have a prominent Golgi apparatus, prominent cytoplasmic vesicles, and little rough endoplasmic reticulum, while type B (fibroblast-like or F) cells have a well-developed rough endoplasmic reticulum but a poorly developed Golgi apparatus. By electron microscopy, synovial cells may be seen to possess filopodia (surface membrane folds, used when the cells serve as opportunistic phagocytes) and, in certain species (rat, rabbit, and calf), cell junctions.
Formation of Bone
Endochondral Ossification
Long bones develop first as cartilage anlagen (primordia) and then grow in length at the physes (growth plates) by endochondral ossification. These plates usually are located at both ends of long bones, but sometimes additional ossification centers are present at other locations. The growth plate consists of a thick layer of cartilage located between the epiphysis and the metaphysis. Chondrocytes in the upper layers of the growth plate divide and move downward toward the lower layer of the plate, where ossification of the matrix occurs. The normal growth plate is arranged in layers (“zones”) horizontally and columns vertically (Figure 23.4).
Closest to the epiphysis is the resting zone (also termed the reserve or prehypertrophic zone) of chondrocytes, beneath which lies the zone of proliferation with its flat, compact chondrocytes stacked in neat rows. As they move toward the metaphysis, these chondrocytes enter the zone of hypertrophy, where each chondrocyte is large, approximately spherical, and has an apparent space around the cell. The cartilage matrix becomes calcified at the lower reaches of this zone, resulting in death of the entrapped chondrocytes. Capillaries from the bone marrow invade upward between the spicules of calcified cartilage matrix, and osteoblasts aggregate along these vessels and lay down a thin layer of osteoid on this calcified cartilage scaffold. Mineral (calcium hydroxyapatite) is deposited in the osteoid, which now becomes the primary spongiosa (Figure 23.4), the layer of bone just beneath the growth plate containing many thin trabeculae of woven bone deposited over calcified cartilage cores. This area is rich in osteoclasts, which act to decrease the number of calcified cartilage spicules and the overall number of trabeculae in the primary spongiosa. The layer of secondary spongiosa contains fewer trabeculae, which are thickened by the application of additional bone by osteoblasts. The woven bone in the primary spongiosa is covered by lamellar bone as the secondary and tertiary spongiosa (mature trabeculae) are formed. This progressive modeling of cartilage to bone provides the engine by which the bone grows in length.
As long bones develop, it is necessary to decrease the bone width from the physis to the diaphysis. This adjustment occurs in the “cut back” zone in the metaphysis in young growing animals, where osteoclasts are plentiful beneath the periosteum and osteoblasts are very active on the endocortical surface. Increases in bone width with age occur from growth at the periosteum as osteoblasts deposit osteoid on the outer surface of the existing cortex.
Intramembranous Ossification
Flat bones [like those of the calvarium (skull) and the scapula] and foci of woven bone are formed by intramembranous ossification. In this process, bone is laid down directly in the mesenchymal collagenous matrix rather than by transmutation of a preformed cartilage model. At such sites, osteoblasts are derived from mesenchymal stem cells, and arranged at random. Woven bone is normally present in few locations in the skeleton: on the surface of calcified cartilage as a result of normal bone modeling, immediately subjacent to the deepest zone of articular cartilage (in small amounts), and at sites of tendon and ligament insertion. It is also formed as part of the initial response to fracture.
Molecular Regulation of Bone and Cartilage Development
Bone and cartilage differentiation, growth, and mineralization are influenced by a variety of hormones, which are systemically available molecules, as well as growth factors, which are produced locally (Table 23.2). Many aspects of this process occur at the growth plate. For example, Indian hedgehog (Ihh) is a master regulator of bone development, serving to control endochondral ossification by coordinating chondrocyte proliferation and differentiation in the growth plate, and osteoblast differentiation in the primary spongiosa. Ihh is synthesized in the growth plate by proliferating chondrocytes as well as early hypertrophic chondrocytes, where a key function is to synthesize parathyroid hormone (PTH)-related peptide (PTHrP), a proproliferative factor for chondrocytes, notably those in the prehypertrophic and early hypertrophic zones. Key hormones include growth hormone (GH), thyroxine (T4), cortisone, and sex hormones (estrogen and testosterone). Essential signaling molecules include prostaglandin (PG) E2 and retinoic acid; critical nutrients are vitamin D3, vitamin K and, in some species (e.g., guinea pigs, monkeys, humans) vitamin C; and other major growth factors such as insulin-like growth factors (IGF) I and II, fibroblast growth factors (FGFs), platelet-derived growth factors (PDGFs), epidermal growth factor (EGF), and vascular endothelial growth factor (VEGF), and bone morphogenetic proteins (BMP). Many of these factors work in combination; for example, activation of GH receptors on osteoblasts enhances chondrogenesis in growth plates via IGF-I and IGF-II. All these ligands have regulatory activity in the growth plate and are essential for normal development of the cartilage centers that presage endochondral ossification. Many factors negatively impact chondrocyte activity when they are deficient, while some (e.g., retinoic acid) also may be toxic when present at excessive levels.
Table 23.2
Molecules that Modulate Collagen Biosynthesis in the Skeleton
| Factor | Collagens affected | Cellular targets | Type of response |
| GROWTH FACTORS | |||
| TGFβ1 | I, III | Fibroblasts, osteoblasts, hepatoblasts, dedifferentiated chondrocytes | Increases α1(I) + α1(III) mRNA level & stability, enhances translational activity of α2(I) mRNA |
| II | Chondrocytes | Suppression of Col2a1 expression | |
| II | Chondrocytes | Up-regulation of Col2a1 expression | |
| VI | Fibroblasts | Up-regulates α3(VI), down-regulates α1 + α2(VI) | |
| TGFβ3 | α1(I) | Fibroblasts | Stimulates α1(I) mRNA levels; down-regulates in presence of TGFβ1 |
| BMP-2, -3, -4, -7 | I (?) | Osteoblasts | Induces osteoblast differentiation, stimulation of collagen I mediated through Smads and Cbfa1β |
| II; X | Chondrocytes | Up-regulation of Col2a1 expression in chondrocytic lines | |
| PDGF | V | Gingival fibroblasts | Stimulation of collagen protein synthesis |
| EGF | I | Osteoblasts | Inhibits collagen synthesis |
| VEGF | I; IV | Osteoblasts | Promotes vascular invasion of hypertrophy zone |
| IGF-1 | II | Chondrocytes | Stimulates Col2a1 synthesis |
| CYTOKINES | |||
| IL-1 | I, III | Dedifferentiated. Chondrocytes, fibroblasts | Enhanced mRNA levels |
| II | Chondrocytes | Suppresses COl12al transcription | |
| IFNγ | I | Fibroblasts | Suppression of collagen synthesis + mRNA levels |
| II | Chondrocytes | Suppression of α1(II) mRNA levels | |
| TNFα | I | Fibroblasts | Suppression of α1(I) mRNA |
| HORMONES AND OTHER FACTORS | |||
| Vitamin D3 | I | Osteoblasts | Active form—1,25(OH)2D3—decreases α1(I) mRNA and protein levels |
| Glucocorticoid | α1(I) | Calvaria fibroblasts | Suppresses collagen production in long-term cultures, reduced transcription rate of α1(I), multifactorial complex effects |
| Growth hormone | I, III | Chondrocytes | Induces IGF in proliferative chondrocytes in growth plate |
| PTH | I | Fetal rat calvaria | Suppression of α1(I) mRNA levels |
| X | Hypertrophic chondrocytes | Reduced α1(X) mRNA levels | |
| Retinoic acid | I, III | Chondrocytes | Induces switch from α1(II) to α1(I), α1(II), and α1(III) mRNA in chondrocytes |
| X | Chondrocytes | Transient stimulation of α1(X) expression | |
| Ascorbate (Vitamin C) | All collagens | Fibroblasts, chondrocytes, osteoblasts, etc. | Enhances prolyl and lysyl hydroxylation stimulation of α2(I) mRNA transcription, enhanced mRNA stability |
| Vitamin K | I | Osteoblasts, chondrocytes | Induces osteocalcin and matrix Gla protein |
| TRANSCRIPTION FACTORS AND ONCOGENES | |||
| c-fos | I | Osteoblasts | c-fos Overexpression inhibits α1(I) mRNA levels in osteoblasts |
| Cbfa1/Runx2 | I | Osteoblasts | Induces transcription of Col1a1 in osteoblasts by binding to OSE2 elements in the promoter |
| X | Hypertrophic chondrocytes | Stimulates Col10a1 expression by binding to OSE2 elements in the Col10a1 promoter | |
| SOX9 | II, IX, XI | Chondrocytes | Induces Col2a1 and Col11a1 expression by binding to enhancer |
| Thyroxine | X | Chondrocytes | Induces Col X mRNA Suppresses clonal expansion & proliferation |
Abbreviations: BMP, bone morphogenetic protein; Cbfa1/Runx2, core binding factor A1/Runt-related transcription factor 2; c-fos, FBJ murine osteosarcoma viral oncogene homolog; Col, collagen; EGF, epidermal growth factor; IFN, interferon; IGF, insulin-like growth factor; IL-1, interleukin 1; OSE2, osteocalcin-specific element 2; PDGF, platelet-derived growth factor; PTH, parathyroid hormone; OSE2, osteoblast-specific element 2; SOX9, SRY (sex determining region Y)-box 9; Smad, combination name for transcription factors that exhibit homology to both the Caenorhabditis elegans protein Sma and the Drosophila melanogaster protein Mad; TGF, transforming growth factor; TNF, tumor necrosis factor; VEGF, vascular endothelial growth factor.
Table adapted in part from Dynamics of Bone and Cartilage Metabolism, second ed., (M.J. Seibel, S.P. Robins, and J.P. Bilezikian, eds). Elsevier, Burlington, 2006; Table 24.1 page 22, with permission.
During skeletal growth and repair, cell proliferation and differentiation in cartilage and bone are coordinated by cell–cell signaling. For instance, canonical Wnt signaling (via beta-catenin T-cell factor) is a key regulator of skeletogenesis by accelerating endochondral ossification and suppressing chondrocyte formation, leading to shortening of the growth plate and increased calcification of the hypertrophic zone. In particular, Wnt/beta-catenin signaling works in concert with Ihh signaling in the growth plate, with Ihh stimulating reserve zone chondrocytes to enter proliferation while Wnt acts downstream to promote osteoblast maturation. Similar controlling mechanisms of osteoblast proliferation and differentiation occur in adult mesenchymal progenitor cells during fracture repair, so initiation of bone formation in fracture repair initially requires Ihh signaling and later Wnt signaling in differentiated osteoblasts.
Regulation of Bone Maintenance
Bone metabolism or bone turnover is a dynamic and continuous remodeling process that is normally maintained in a tightly coupled balance between resorption of older (mature) or injured bone and the formation of new bone. Every section will contain both old and new bone, and thus offers a glimpse at the history of the bone. The elements that mediate the set point for this balance include physical forces (i.e., biomechanics); biochemical equilibria (e.g., mineral homeostasis); and molecular signaling. Bone remodeling is divided into two major categories: targeted remodeling in response to injury or changes in biomechanical loading, and stochastic remodeling in response to homeostatic mineral requirements.
Remodeling in large animals (dogs and monkeys) and humans is performed by the basic multicellular unit (BMU) of bone (Figure 23.5). The 3D shape of a BMU is an asymmetric bicone; the osteoclasts resorbing bone line the shorter end of the cone while the osteoblasts laying down new bone line the longer end and act to close the space first created by the osteoclasts. The BMU will be separated from the bone marrow by “BLCs” that have separated from the bone surface, forming a canopy. A blood vessel will be present adjacent to the canopy. The BLCs are thought to play roles in maintaining ion fluxes between fluid in the bone and marrow, initiating BMU activity, and regulating induction of hematopoiesis.
True bone remodeling is minimal in the cortex in rodents. As such, long bones in adult rodents exhibit three major structural features that are not observed in bones of nonrodents. First, cartilage cores persist in trabecular and cortical bone for months or longer in rats. Second, rodent cortical bone contains few osteons. Third, the process of forming and shaping the diaphysis occurs by periosteal bone resorption and endosteal intramembranous bone formation on secondary spongiosa in the “cut back” zone in the metaphysis.
Biomechanics
Physical forces acting on the skeleton constitute a major extrinsic influence on postnatal bone development and maintenance. Cartilage thickness in a mature joint is also affected by local stress and the environment created by physical activity.
Bone responds to an applied force (stress) by undergoing an architectural deformation (a relationship known as Wolff’s Law), which depends on the magnitude, frequency, and distribution of the load. Bone can deform because the collagen within it imparts tensile strength. In contrast, the hydroxyapatite mineral crystals within the collagen matrix impart compressive strength. Within the osseous matrix, osteocytes are thought to act in part as a network of mechanical sensory cells. The healthy skeleton is able to continually respond to mechanical stimuli by initiating or inhibiting bone remodeling to maintain bone structure for typical strains within a normal physiological range, a capacity known as the mechanostat. In the context of desirable drug actions and unwanted toxicity affecting bone and cartilage, the concept that mechanical factors may modulate the effects of circulating agents (both endogenous ligands and drugs), genetic programming, and disease on bones (and vice versa) is important.
Evaluation of Toxicity
Toxicity to bone and cartilage can be evaluated in several different fashions. This section explores the common options used for this purpose in conventional product discovery and development programs.
Physiologic Endpoints: Biochemical and Biomarker Evaluation
Biochemical markers of bone turnover [or bone turnover markers (BTMs)] along with hormones and clinical biochemistry parameters measured in serum or urine provide a rapid, reliable, and dynamic real-time in vivo assessment of overall skeletal metabolic activity. BTM immunoassays are noninvasive, simple, specific, and sensitive tools to assess bone metabolism. BTMs represent an important component of nonclinical studies to assess any adverse effects or pharmacological action of compounds on bone metabolism.
The majority of markers used in humans for the diagnosis and monitoring of bone disease and treatment compliance can be assessed in most laboratory animal species with validated analytical methods. Bone markers can be incorporated into routine toxicology studies whenever these data might be able to add perspective to the interpretation of the toxicology endpoints. The storage conditions for samples and the type of assay to be performed must be considered as potential sources of variation. The collection of blood during nonclinical studies should be performed under standardized conditions across all subjects, preferably in the morning (before 10 a.m.) as there are marked circadian variations in biomarker concentrations. In general, blood is collected by intravenous puncture, and serum aliquots are stored frozen until analysis.
Multiple bone formation markers are available. Bone-specific alkaline phosphatase (ALP), which leaks from the osteoblast plasma membrane, is the most specific serum marker for osteoblast activity, but in the absence of intercurrent liver disease total ALP can be used as a surrogate marker for bone-specific ALP, especially in rodents for which no immunoassays are available for the bone isoenzyme. Osteocalcin is a bone matrix protein produced by osteoblasts. Procollagen type I telopeptides, which are cleaved during collagen I formation, also represent a key measure for bone formation, with amino-terminal type I procollagen propeptide (PINP) appearing to be the more sensitive marker of bone formation in most species (Figure 23.6). Common bone resorption marker assays of osteoclast activity detect the deoxypyridinoline crosslinks or the N- or C-terminal telopeptide fragments of collagen type I such as deoxypyridinoline (DPD) and type I collagen telopeptide released from the amino (NTx) and carboxyl (CTx) termini. Measurement of tartrate-resistant acid phosphatase (TRAcP5b or TRAP5b), an enzyme expressed by activated (bone-resorbing) osteoclasts, provides a direct assessment of osteoclast numbers. Collagen type I is also present in other tissues such as skin, dentin, cornea, vessels, fibrocartilage and tendons, but the collagen type I turnover rate in nonskeletal tissues is much lower than the process occurring in bone; therefore, nonskeletal collagen type I metabolism contributes very little to the collagen breakdown products in the circulation. Due to the coupling of bone formation and resorption, it is recommended that a panel of bone markers be evaluated in serum samples to fully characterize the skeletal status with respect to these two competing physiologic processes. For nonclinical studies, such panels typically include at least two markers of bone formation (usually osteocalcin and PINP) and two markers of bone resorption (generally CTxI and DPD).
Calcium and phosphorus metabolism and the hormones that govern them are also important to consider. The serum calcium level is regulated closely by homeostatic control mechanisms and thus usually is maintained within the normal range. Since half of serum calcium is bound to protein, either unbound ionized calcium (Ca2+) should be measured or total serum calcium concentrations should be corrected for serum protein levels. The serum phosphorus level fluctuates more widely than does the serum calcium concentration, with phosphorus levels varying based on dietary intake, release from bone, and urinary excretion. Measuring hormones related to calcium metabolism (e.g., calcitonin, PTH and 1,25-dihydroxyvitamin D3) also may provide useful information regarding calcium utilization.
The effectiveness of these bone biomarkers in short-term and long-term pharmacological studies has been demonstrated in the rat, dog, and monkey. Increases in biochemical markers of bone turnover normally are consistent with histomorphometric indices (see below) of bone turnover. Depending on the class of compound tested and its mechanism of action, different patterns of response can be expected, ranging from an overall increase [bone-building anabolic agents such as PTH(1–34), PTH(1–84), or PTH analogues] or decrease [anti-erosive agents like bisphosphonates or inhibitors of receptor activator of nuclear factor kappa-B ligand (RANKL)] in blood levels of biomarkers to an uncoupling of the response in biomarkers for formation and resorption (antibodies to inhibit sclerostin, an anti-anabolic molecule).
Conventional Evaluation of the Skeleton: Design Considerations and Methods
Production of high-quality bone sections begins at necropsy. If there are pathologic changes in the physis (growth plate), or in very young animals, there may be sufficient skeletal weakness to permit artifactual separation of the epiphysis (end) from the bone shaft at the growth plate with routine handling at necropsy. Therefore, care should be taken to ensure gentle handling of bones and joints during dissection.
Appendicular (long) bones are more frequently assessed during toxicity studies than axial (vertebrae or pelvis) bones. However, evaluation of both bone types may be necessary to fully characterize a disease process or skeletal response to treatment. The most commonly measured compartments are the metaphysis (formed mainly of trabecular bone) and diaphysis (comprised entirely of cortical bone). The epiphysis and its articular surface may be evaluated in arthritides. If the axial skeleton is to be assessed, the most common sites are either a lumbar vertebra (all species) or the iliac crest (in large animals).
One of the primary challenges to the histopathological study of bone biology in experimental animals is obtaining comparable sections from all individuals in a study, and also reproducing this orientation consistently across subsequent studies. This challenge is particularly difficult with respect to major diarthrodial joints [e.g., femorotibial (stifle or “knee”) and tibiotarsal (hock or “ankle”)], which have curved surfaces that alter shape radically over relatively short distances. The key to attaining this goal is to follow a trimming protocol based on skeletal features that are readily identifiable and can be sampled reproducibly. For example, the physis of the proximal tibia (which is linear) is more uniform and easier to orient than is the physis of the distal femur (which is curved in 3D), so consistent sections of physis are easier to obtain if the tibia is taken.
In rodents, the typical bone and joint specimens are sternum and femorotibial joint. The sternum is commonly used to evaluate bone marrow, but the thin cortical walls and lack of many important bone structures prevents a complete analysis of bone. The femorotibial joint, oriented in a sagittal plane, contains readily obtainable regions of cancellous (trabecular) bone in the epiphyses, primary and secondary spongiosa in the metaphyses, and endocortical and periosteal bone in the “cut back” region of the metaphyses as well as demonstrating other important skeletal features such as the growth plate, subchondral plate, articular cartilage, patella, patellar tendon, cruciate ligaments, and joint capsule. That said, the femorotibial joint often is oriented in the frontal plane to increase the amount of articular cartilage available for evaluation. In older mice, the lumbar vertebrae cut in a frontal plane will have many more trabeculae than either the distal femur or proximal tibia.
In the larger animal species, either the sternum (to evaluate bone marrow) or the rib (at the costochondral junction) and a long bone are evaluated during general toxicity studies to examine the health of bone. In young animals, endochondral ossification is vibrant at the costochondral junction. The best way of collecting ribs of large animals at necropsy to avoid separation of the bone and cartilage at the costochondral junction is to remove and fix the whole (monkeys) or a portion (several ribs together, for dogs) of one thoracic wall. After fixation, one of the larger ribs can be processed, taking care that the same rib is sampled in each animal. Inclusion of a long bone (typically the proximal tibia, with its linear growth plate) is common; the frontal plane is preferred for trimming because the effect of weight bearing on bone and joint architecture can be appreciated by comparing the medial side versus the lateral side of the joint.
It is critical to proper histological processing to remove as much skeletal muscle and other soft tissue from bones and joints as possible prior to fixation. For nonclinical toxicity studies, bone specimens generally are fixed by immersion in neutral buffered 10% formalin (NBF) at room temperature for at least 24 hours. Fixation time will be longer for large animal specimens with dense cortical bone and may be shorter for thin specimens (3–5 mm thick) of primarily trabecular bone with significant amounts of marrow.
Routine histological processing of bone requires decalcification following fixation. The common approach for nonclinical toxicity studies uses a formic acid-based solution in which the bones are immersed for 1–5 days following fixation, until they become completely decalcified. A solution of up to 10% formic acid, which also contains formaldehyde (approximately 4%) and methanol (5% to 10% by weight) to continue fixation, is commonly used. Alternatively, decalcification may be undertaken using a chelating agent, such as ethylenediaminetetracetic acid. Chelation is a slower and gentler means of removing calcium, and so is employed when special molecular procedures (e.g., immunohistochemistry to detect a labile antigen) is a desirable endpoint. Decalcification by chelation requires multiple transfers of bony specimens into fresh chelating solution (typically one every 24 hours for at least 6 days, depending on the sample size). Production of bone slabs (3–5 mm thick, for large bones) using a diamond saw or cutting a small window through the cortex (for small bones) enhances penetration of the fixative/decalcification solution compared to processing completely intact bones, and so reduces the time needed for fixation. Decalcified bone specimens generally are embedded routinely in paraffin in nonclinical general toxicity studies.
The standard stain for toxicology studies, H&E, is a good choice for general bone examination (Figure 23.2) even though osteoid cannot be differentiated on decalcified sections. Toluidine blue (Figure 23.2) and safranin O (Figure 23.2) commonly are employed to assess the integrity of articular cartilage matrix. Visualization of osteoid requires undecalcified bone sections; the modified Goldner’s trichrome stain (Figure 23.5) is suitable for demonstrating osteoid and also permits polarization of collagen. Movat’s pentachrome (with modifications) will differentiate mineralized and unmineralized bone and cartilage in undecalcified sections. Von Kossa is commonly used for differentiating mineralized bone from unmineralized osteoid.
Special Techniques for Examining the Skeleton
Nonclinical general toxicity studies in which a skeletal lesion is anticipated or specialized studies dedicated specifically to the assessment of skeletal endpoints often incorporate additional endpoints for skeletal analysis. These may examine morphological features (e.g., bone structure, bone density) or bone function (e.g., biochemical uptake, biomechanical properties). A few of the more common tests used in nonclinical toxicity studies are reviewed briefly here.
Radiological Examination
Conventional radiology is an essential component in examining the skeleton, and provides a commonly available tool for bone analysis in vivo. Radiographs also can be performed ex vivo on isolated skeletal segments, often using a high-resolution, bench-top, cabinet-housed, high-resolution radiography system. It is a simple means for examining the size, shape, and density of the entire bony skeleton or large subdivisions thereof, in a stereotypical orientation. Joint contours may be assessed using radiology, as diseases affecting articular cartilage also can impact bone integrity; however, cartilage itself is radiolucent and so cannot be evaluated directly in standard radiographs. Radiography is the method of choice to scan the entire skeleton in vivo of small animals before scheduled termination to detect any lesions at skeletal sites that are not routinely harvested and that would not have been detected otherwise. Obtained at intervals in growing animals, serial radiographs provide an accurate record from which to measure bone growth and assess epiphyseal closure in juvenile toxicity studies. Radiographic data are of particular importance for carcinogenicity studies but also are warranted in subchronic and chronic nonclinical studies when direct or indirect effects on bone tissue are suspected (Figure 23.7). Prestudy radiographs in primates can be very useful to document preexisting findings that should not be considered to be a consequence of test article administration.
Specialized imaging technology may provide additional information besides a simple assessment of bone structure. Molecular imaging is the visualization, characterization, and measurement of biological processes at the molecular and cellular levels. The radiotracer approach is used for positron emission tomography (PET) and single photon emission computed tomography (SPECT). Nuclear bone scans such as plain scintigraphy and SPECT radiobisphosphonate bone scans following the administration of 99m-technetium-labeled ethylene diphosphonate may be advantageous as this agent is preferentially incorporated into the skeleton at metabolically active sites associated with bone formation. Bone scintigraphy reveals enhanced radiobisphosphonate deposition due to changes in blood flow or increased osteoblastic activity, and commonly are used to find “hot spots” of metastatic cancer throughout the skeleton. For PET imaging, 18F-fluoride (NaF) is being extensively used for assessing bone metabolism, and it has a similar uptake mechanism to 99mTc-MDP 18F-FDG (fluorodeoxyglucose) is another PET tracer that can be used as a sensitive functional biomarker indirectly in the skeleton. In general, PET images have a lower resolution compared to other bone-imaging methods, but it has a high molecular sensitivity (nanomolar) with unlimited depth penetration (Figure 23.8). Optical imaging (fluorescence and bioluminescence) techniques are highly sensitive at limited depths (few millimeters), rapid and easy to perform. Fluorescence is used to image the skeleton, with a different bone specific fluorophores (labeled alendronate or other molecule), which are incorporated in the calcified bone matrix at spots with high bone turnover and are therefore good indicators of bone remodeling in sites of bone damage.
Computed tomography (CT), a 3D conventional radiographic technique, and magnetic resonance imaging (MRI), which produces high-resolution 3D representations of skeletal and soft tissues, increasingly are utilized in nonclinical toxicity studies to assess the shape and quantity of various bone compartments.
Osteodensitometry: DXA, pQCT and Micro-CT
Bone mineral density (BMD) can be evaluated in vivo or ex vivo using dual energy X-ray absorptiometry (DXA) and/or peripheral quantitative CT (pQCT), or by micro-CT. DXA allows a two-dimensional (2D) assessment, while pQCT and micro-CT provide a 3D volumetric analysis (Figure 23.9). With a nominal resolution (pixel size) of a few microns, micro-CT provides measurements of microarchitecture such as relative bone volume, trabecular number, trabecular surface area, trabecular thickness and trabecular separation (i.e., static parameters traditionally assessed with bone histomorphometry) as well as volumetric and tissue BMD. Baseline or pretreatment data are important to calculate individual animal changes at each time point. The percent change from pretreatment baseline calculated for each individual animal provides a more powerful data set (versus absolute values at each occasion) and helps to overcome limitations due to individual animal variability and the small group sizes normally used in toxicity studies for large animal species. Bone densitometry is a primary endpoint of drug efficacy in both clinical trials and nonclinical studies, and also is employed as a means to quantify any adverse effects on bone mass.
Biomechanical Testing
Bone strength constitutes a critical endpoint in skeletal assessment, serving as the gold standard of bone quality. Biomechanical testing of bone evaluates the functional impact of alterations in bone turnover, bone mass, and bone geometry.
The mechanical competency of a bone depends on multiple factors. These include the direction and magnitude of forces applied, the bone geometry (dimension and shape), and its material properties. Destructive static tests, where the force is applied slowly and gradually until the bone breaks, are most commonly used for animal models and drug testing. Biomechanical strength testing of long bones (in 3- or 4-point bending or in torsion about the long axis), the femoral neck (via shear), and vertebrae (by compression) can be performed to evaluate whether or not treatment has affected any of the properties that define bone strength. The load transmitted through the specimen until failure is recorded by a transducer interposed in the system. Software derives the displacement versus force curves, generating for most biological specimens a typical curve profile.
Common biomechanical endpoints include the force to failure (or peak load), stiffness (extrinsic rigidity of the specimen), and the cumulative energy required to break a bone [area under the curve (AUC)]. Data are normalized for bone size, and calculations of parameters independent of the size of the specimens can be made using the cross-sectional moment of inertia obtained by pQCT scans as a means for defining bone tissue intrinsic parameters: ultimate stress, modulus (intrinsic rigidity of the material when data are corrected for the specimen size), and toughness (a property relating material characteristics to energy absorption, i.e., the AUC corrected for the size of the specimen). Changes in intrinsic properties may reflect an effect of test article treatment on bone quality.
Bone specimens dedicated to biomechanical testing typically are collected at necropsy and cleaned of soft tissue. Without fixation, intact bones are frozen at −20°C until testing. Bones are warmed to room temperature before being subjected to biomechanical stress.
Bone Histomorphometry
Historically, histomorphometric evaluation of bone tissue can be divided into three major analyses: evaluation of bone structure (static parameters), evaluation of bone formation (dynamic parameters), and evaluation of BMU morphology and function. Bone histomorphometry in nonclinical toxicity studies has evolved over the past half century so that parameters, techniques, and nomenclature have become standardized.
Ethanol-fixed, undecalcified, plastic-embedded sections are used instead of standard formalin-fixed, decalcified, paraffin-embedded sections. Decalcification makes evaluation of osteoid impossible. The type of data desired from a histomorphometric examination must be planned in advance since some endpoints require the in-life administration of marker compounds that collect in bone.
When performing a histomorphometric evaluation, the size and location of the region of interest (ROI) depends on the species and type of bone. Some portions of a bone compartment may need to be excluded from analysis if significant, region-specific biological differences exist. As an example, primary spongiosa are routinely excluded from analysis of metaphyseal trabecular bone because primary spongiosa are immature and may not have been exposed to a test-article for the entire duration of a nonclinical study. Similarly, if osteoclast activity is an important endpoint, then measurements should be avoided in regions where osteoclasts are rarely present (such as the mid body of a vertebra).
Static histomorphometry is the direct measurement of the amount of total tissue area, bone area, and bone surface present in the ROI but is increasingly determined using micro-CT. These endpoints may be collected from any nonclinical study, even in the absence of advanced planning for special skeletal endpoints, since in-life treatment with bone-homing compounds is not required to gather such data.
Dynamic bone histomorphometry, which is performed more commonly than static histomorphometry, measures bone formation in trabecular or cortical bone by examining the incorporation of “fluorochrome labels” (fluorescent compounds) administered two or more times during a study (Figure 23.10). Fluorochrome labels that deposit preferentially in bone are given parenterally prior to necropsy. Common agents used for this purpose include calcein green, alizarin red, and doxycycline. Agents are given intermittently to track the rate of new bone formation, which is measured as the distance between any two marker lines (Figure 23.10). Often, multiple agents that emit light at different wavelengths, and thus appear in images as lines of different colors, are utilized to make interpretation simpler.
The other type of histomorphometric evaluation is assessment of BMU cellularity and function on stained (modified Goldner’s trichrome, toluidine blue, etc.), undecalcified sections. In this approach, average osteoid width (and extent) are measured. The average orthogonal wall width (WWi) of recently completed BMUs also can be quantified as a means of defining the activation frequency (an indicator of the bone remodeling rate).
Animal Models
Models of Bone Loss
Animal models of bone loss can be produced in many fashions. Surgical procedures include removal of organs (e.g., gonadectomy in rodents and nonhuman primates, hypophysectomy or thyroparathyroidectomy in all species) that produce hormones which serve to either build bone (anabolic) or prevent its resorption (anti-catabolic) and joint manipulations that damage the articular cartilage or disrupt stabilizing structures (ligaments, menescii). Other techniques that may be used experimentally to induce bone loss in rodents include chronic immobilization (tail suspension or limb casting), advanced age, or husbandry adjustments (e.g., dietary or light-cycle manipulations). Use of a calcium-restricted diet enhances ovariectomy-induced bone loss in nonhuman primates and minipigs. Female ferrets can be induced to lose bone by exposure to shortened light cycles (8 hours on/16 hours off). Dogs do not lose bone readily after ovariectomy and are not recommended as a bone loss model, although they are suitable for fracture healing assessments and evaluation of bone anabolic agents. While the number of animals will vary depending upon the endpoints chosen for evaluation and the anticipated effect of a treatment, common group sizes in published reports range from 6 to 25. For instance, higher numbers of animals per group may be needed for histomorphometric studies, while lower numbers are acceptable for biomechanical endpoints and conventional histopathologic evaluation.
Models of Arthritis
Immune-mediated arthritides are inflammatory and erosive joint conditions that affect humans and many animal species as spontaneous or (in animals) genetically engineered and induced diseases. Rodent models of immune-mediated arthritis include adjuvant-induced arthritis (AIA), produced by injection of rats with Freund’s adjuvant with or without an additional antigen; collagen-induced arthritis (CIA), produced by immunization of rats, mice, or nonhuman primates with type II collagen from any of several species; and other immune-mediated arthritides, typically produced by injection of bacterial cell wall fragments or other immunogenic proteins or glycoproteins. Genetically engineered animal models of arthritis, in which rats or mice are specifically altered to express a fabricated gene that predisposes them to arthritis, have been generated using various proinflammatory molecules. One example of this strategy is the tumor necrosis factor-alpha (TNF-α) transgenic mouse, which develops a symmetrical polyarthritis that appears first in the tarsus. Similarly, rats engineered to express the HLA-B27 human leukocyte antigen (a major histocompatibility complex, type I antigen carried by many human patients with autoimmune diseases) and human β2-microglobulin also develop immune-mediated polyarthritis. Immune-mediated joint disease is characterized by diffuse synovial activation (leading to hyperplasia and hypertrophy), extensive infiltration by multiple leukocyte classes (especially lymphocytes and plasma cells), and production of cell-dense fibrovascular tissue (termed pannus) that tends to invade bone and cartilage. The extent of bone loss varies with the arthritis model. In general, AIA models produce rapid and profound bone destruction, while CIA models yield a more slowly progressing dissolution of bone that becomes pronounced over long periods. For this reason, CIA often is considered a preferred choice for investigating immune-mediated joint disease in humans [e.g., rheumatoid arthritis (RA)]. The extent of bone loss is a function of both time (chronic lesions are more severe and destructive) and genetic background (e.g., Lewis rats are more arthritis-prone than are F344 rats). In rodents but usually not primates (including humans) with arthritis, bone loss resulting from marked erosion of preexisiting bones may be obscured to some degree by florid periosteal production of new bone, which may in severe cases result in ankylosis of affected joints.
Osteoarthritis (OA) is another spontaneous disease afflicting humans and various animal species. This condition differs from immune-mediated arthritis (i.e., “inflammatory” arthritis) in that OA (or “noninflammatory” arthritis) tends to exhibit a greater degree of degenerative changes in bone and cartilage in conjunction with a substantially lesser degree of inflammatory cell infiltration. Cardinal features of OA include cartilage degeneration, fragmentation, and loss; degeneration of subchondral bone; and formation of osteophytes (knobs of periarticular new bone produced in an effort to restore some stability to damaged joints). Animal models of OA have been produced in several species (rats, rabbits, guinea pigs, and dogs) via surgical destabilization of the femorotibial joint, which ultimately results in cartilage matrix degeneration and cartilage erosion. Spontaneous OA occurs in many aging animals, including rodents (e.g., STR/ort mice), guinea pigs, and monkeys; these models have been used to study biochemical and molecular mechanisms. Intraarticular injections of iodoacetate (which kills chondrocytes) or systemic quinolone antibiotic treatment also cause cartilage degeneration. Genetically modified mice that over-express or lack bone-regulating molecules also yield OA-like lesions; for example, animals lacking soluble RANKL decoy receptor osteoprotegerin (OPG−/−) develop progressive, multifocal degenerative joint disease. A study of T-2 mycotoxin-induced articular cartilage degeneration in the rat has been likened to Kaschin–Beck disease, an endemic form of OA seen in China and considered to be due to an environmental toxin.
Many semiquantitative histopathologic grading schemes are available for macroscopic and microscopic assessment of animal models of immune-mediated arthritis and OA. The scoring schemes may use features that are common to similar conditions across species, or they may emphasize changes that are species-specific. While general morphology is evaluated on routine, decalcified H&E sections, additional sections are stained with either toluidine blue or safranin O to permit grading of cartilage matrix integrity. In general, interpretations of arthritis severity are based not only on pathology lesion scores but also on clinical findings (e.g., number of affected joints); bone densitometry (e.g., DXA or pQCT); or (for immune-mediated models, in which swelling is a prominent feature) joint volumetric measurements (plethysmography).
Study Design Considerations
In nonclinical toxicity studies, evaluation of changes in bone structure, function, and metabolism are evaluated using a number of complementary methods. Some in vivo endpoints, such as densitometry and biochemical markers, may be used in humans, so animal data is of clinical relevance. Inclusion of these specialized assessments in the study design will depend on the study objectives and the age of the test animals. Bone biochemical markers and densitometry (DXA or/and pQCT) measurements can be included easily in most toxicity study designs of 28 days’ duration or longer in rodents and 3 months’ duration or longer for large animals, both for the terminal necropsy and during the recovery phase. For juvenile toxicity studies, pQCT is of particular interest for in vivo growth evaluations over time as the evolving geometry of developing bone can be assessed readily with this technique. Pretreatment data are of importance for interpreting densitometry and biomarker data, especially for large animal studies, for two main reasons. These endpoints may be measured intermittently throughout the course of a study, thus giving an idea of progressive skeletal effects, and they can capture changes at much earlier time points (often when structural changes may not be present or quantifiable). Loss or gain in bone mass, alterations in bone geometry, and ultimately shifts in biomechanical properties are cumulative, so important changes that may be detected later reflect the net adjustments in bone metabolism that occurred over the entire study period.
Specialized endpoints are often applied using a tiered strategy. For such an approach to work, provision is made in the study protocol to retain serum or/and urine as well as frozen and fixed samples of specific bones for possible future densitometric and/or biomechanical and/or histomorphometric assessments. Typically, these special samples will be analyzed as a “second-tier” assessment only if the outcome of the “first-tier” analysis (e.g., routine macroscopic and microscopic pathology with conventional serum chemistry parameters) suggest that bone likely is a target or off-target tissue.
Responses to Injury
Common Responses of Bone and Cartilage
The numbers of ways in which bone and cartilage can respond to injuries are limited. These responses may include altered metabolism alone (as indicated by changes in the profile of skeletal biomarkers) or the presence of functional and/or structural changes in bone and cartilage.
An injured region transforms into a special sphere of influence where normal physiologic processes involved in bone growth are accelerated as a means of repair. This phenomenon has been termed a regional acceleratory phenomenon (RAP). Undifferentiated bone marrow mesenchymal stem cells differentiate into osteoblasts, which then produce new (woven) bone. If necrosis has occurred, osteoclasts also will be recruited to resorb the damaged bone. Following the initial healing response, bone remodeling occurs to reshape the microarchitecture of the repair, to conform to biomechanical requirements of the damaged area. The remodeling process is of much greater intensity in juveniles than in adults.
Cartilage has even fewer means than bone for responding to injury. The key feature of repair seen in injured cartilage is clonal proliferation (see below). The cartilage matrix near clonal clusters usually is less dense (stains less darkly) with toluidine blue or safranin O than that of normal cartilage, suggesting that the cartilage proteoglycan composition is altered.
Common Toxicant-Induced Responses in Bone
All the toxicant-related responses of bone following exposure to exogenous chemicals or biologics represent either a normal reparative response or a pathologic condition in which the normal healing process is interrupted. Entities causing metabolic bone disease alter normal bone remodeling, whereas the formation of periosteal or endosteal new bone is similar to the process observed in the formation of a fracture callus. Even neoplasia can be considered in a broad sense as a disruption in the normal processes of cell proliferation and differentiation.
Necrosis
Spontaneous bone necrosis can occur in laboratory animals. Zones of necrotic bone may be encountered in which there is very little response in the surrounding tissue, but areas of a few empty osteocyte lacunae may be present as an artifactual change and should not be interpreted as bone necrosis. Age-related, idiopathic osteocyte necrosis has been observed in normal bone from healthy rhesus monkeys. This change is virtually absent in young animals but is present to some degree in all mature individuals. Common sites to appreciate this finding include the distal femur, femoral head, proximal tibia and vertebral epiphyses. In one mouse study, bone necrosis was observed in up to 4.6% of the test animals, and the incidence in control Wistar rats is <5%. Bony portions of the menisci, which develop by endochondral ossification with age, are necrotic in the femorotibial joint in a large proportion of old rats and mice. Drugs that cause severe anemia, occlusive vascular disease, thrombosis, or pancreatic release of lipolytic enzymes have been shown to induce aseptic bone necrosis in many species, including rats and dogs. Classic bone necrosis not only involves the osteocytes but also the adjacent bone marrow and engenders a vigorous repair response from both osteoblasts and osteoclasts.
Targeted antiresorptive bone therapies, including bisphosphonates and RANKL inhibitors in humans, have been associated with a low incidence of aseptic osteonecrosis of the mandible. A common factor in the early case reports of this condition was a history of a recent invasive dental procedure. Other predisposing factors, including immunosuppression, periodontitis, and/or concomitant bacterial infection in the oral cavity, also seem to be important factors in development of the clinical presentation, and an association with localized areas of acidic pH in the buccal cavity has been reported.
Atypical fractures (e.g., spontaneous metaphyseal femoral fractures) have been associated with long-term use of bisphosphonates and proton pump inhibitors, although the pathophysiology is different for conditions produced by these two drug classes. In the case of bisphosphonates, deposition of the bisphosphonate into the bone matrix may alter biomechanical properties and in the case of proton pump inhibitors (e.g., omeprazole, esomeprazole, etc.), calcium absorption in the gastrointestinal tract is reduced.
Regional Acceleratory Phenomenon
When the RAP has been evoked, most processes in affected bone are quickened above their normal levels. RAP can often hasten longitudinal bone growth following a fracture, denervation, surgical procedure (e.g., periosteal stripping), or tumor. An example of RAP of relevance to nonclinical toxicity studies is local acceleration of the remodeling process to close the wound resulting from a bone biopsy. Unfortunately, the RAP means that intermittent biopsies of the same bone cannot be used to make comparisons of preexisting bony structures and any posttreatment effects. Note that the RAP involves soft tissue adjacent to bone, too. Common RAP features seen in soft tissues include localized fibrosis and chronic inflammation.
Inflammation and Fibrosis
Acute inflammation, formation of granulation tissue, and fibrous repair in the skeletal system do not differ from these processes as seen in other body systems. Inflammatory diseases that weaken or reduce skeletal strength can lead to pathologic fractures (i.e., breaks in bone that is weakened locally by a disease process). Therefore, adaptive reconstruction and reparative processes often predominate in the morphologic picture associated with such conditions. It is frequently necessary to study the progression of a particular lesion, especially by examining early stages, in order to fully understand their pathogenesis.
Fibrous tissue may be deposited as a substitute for bone when disease prevents normal ossification of the tissues filling bone defects. For normal repair, bone fibrosis is characterized by deposition of type I collagen. The fibrosis associated with hyperparathyroidism [i.e., fibrous osteodystrophy (FOD)] probably represents a response that is distinct from the fibrosis used to repair a chronic injury (Figure 23.11). In FOD, fibrous tissue is closely applied to bone trabeculae, particularly at resorptive sites, and contains more reticulin than type 1 collagen.
Growth and Mineralization of Bone
Longitudinal bone growth is determined by events in the cartilage of the growth plate. Extension of a bone’s length depends largely on the rate of cell division in the proliferating zone. Circumferential bone growth occurs in the metaphysis and is accomplished by osteoblast division and apposition at the perichondrial ring. Any drug that causes inanition, affects calcium absorption, or produces a negative nitrogen balance can reduce longitudinal bone growth due to insufficient quantities of amino acids to form osteoid and/or minerals to mineralize it. Whenever longitudinal growth slows and the growth plate begins to close, focal degradative changes may be observed in the cartilage of the physis and within the metaphyseal trabeculae, and these trabeculae also appear thickened because osseous tissue is deposited on their surfaces over a longer period. When cessation of growth is severe enough, there may be an attempt at premature growth plate closure; in such cases, a lateral plate of bone termed a transverse bridge is formed beneath the growth plate. If restoration of nutrients permits bone growth to be reinitiated, these transverse bone plates will be left behind in the metaphysis as the growth cartilage recedes from them (Figure 23.12). Such persistent bony plates are termed Harris growth arrest lines. The line is eventually remodeled to remove the bulging ridge. Growth plate closure also occurs when the bone matrix content of osteocalcin is reduced by treatment with warfarin, or by smoothened hedgehog (SHH) inhibition (Figure 23.13). Concurrent treatment with PTH, particularly by continuous infusion, can attenuate growth plate closure caused by SHH inhibition (Figure 23.13).
Normally, longitudinal growth and transverse growth are coupled, but different signaling molecules affect growth in these two directions. For example, GH governs the bone growth rate at both locations indirectly, and a GH deficiency (dwarfism) or excess (gigantism) causes a simultaneous and proportional change in the bone width and length. However, some locally active agents (e.g., sclerostin inhibition and PGE2) have the potential to alter differentially growth in different bone compartments or portions of the skeleton. If longitudinal growth is inhibited more than transverse appositional growth, then the bones will be short but have a more normal diameter. Such changes can be produced by compounds that induce rickets (which is characterized by abnormal endochondral ossification with thicker growth plates and unmineralized osteoid) or physeal dysplasia (see below). If transverse growth is inhibited more than longitudinal growth, the bone becomes long and thin. Bones with this characteristic appearance are seen in vitamin A toxicity in rats where pathologic fractures occur in the thin cortices.
The longitudinal growth rate determines the age and amount of trabecular tissue in the metaphyseal zones of different bones. A faster longitudinal growth rate results in an increased amount of cancellous bone of a lesser age (e.g., proximal tibia), whereas a slower growth rate gives less bone of a greater age (e.g., lumbar vertebra). Since an altered longitudinal bone growth rate changes the dynamics of metaphyseal bone modeling and remodeling, studying a drug that alters the quantity of metaphyseal bone in a young animal should determine whether there is an effect on longitudinal bone growth.
Increased Osteoid
Increased osteoid may occur by several mechanisms. These include increased numbers of formation surfaces, an increased thickness of osteoid because of a mineralization defect, or an increased rate of osteoid production. In some cases, osteoid is not mineralized properly because the matrix is abnormal; fluorosis, Solanum malacoxylon toxicosis, and penetrating radiation are examples of this phenomenon. In other cases, osteoblast function may be impaired (such as with some early types of bisphosphonate toxicity) so that osteoid matrix is deposited but remains unmineralized.
Increased Bone
An abnormal increase of bone volume or quantity in the skeleton (Figure 23.14) is reported as increased bone with a location identifier (i.e., the increase is trabecular or cortical). Areas of new, woven bone filling part of the medullary cavity of vertebrae, nasal turbinates and tibiae are a spontaneous lesion in aging animals, such as Han Wistar and F344 rats; this finding was previously described as hyperostosis. Aging F344 female rats develop increased bone, but with a somewhat different appearance: thick trabeculae composed of mature lamellar bone partially filling the medullary cavity of a long bone (typically the tibia). Other modifiers may be used to indicate the anatomical site or characteristic appearance, such as exostosis and enostosis to describe increased bone on the periosteal (for exostosis) and endocortical envelopes, respectively. Osteosclerosis has been used for increased bone density without alteration in the gross shape of the affected bone. Under physiological conditions, as an adaptive response, or in disease, bone in adults increases in amount by raising the incremental difference between resorption and formation during remodeling (positive bone balance). An example of this type of increased bone occurs with myelofibrosis, a condition that follows persistent bone marrow depression and has been reported to follow radiation injury or administration of certain toxic chemicals (e.g., benzene, lead acetate).
Increased bone in nonclinical studies may occur in growing animals treated with compounds that affect bone formation or remodeling. For instance, matrix metalloprotease (MMP) inhibitors reduce bone removal while permitting bone formation to continue. Increased bone in the metaphyses of young, fast-growing rats is a classic finding with such compounds (Figure 23.14). This usually appears as thickened metaphyseal trabeculae.
Neoplasia
In general, primary bone neoplasia in laboratory animals is similar to the phenomenon in humans except that a few morphologic variants of tumors have not been recognized in animals. The vast majority of primary malignant bone tumors in laboratory animals are osteosarcomas, with their typical osteolytic and poorly differentiated features (Figure 23.7). In contrast, primary bone tumors arising from the other major mesenchymal cell types (e.g., chondrosarcomas, fibrosarcomas, liposarcomas) are uncommon. The cause of these neoplasms is not often attributed to toxicants, although long-term treatment of rats with recombinant human PTH has been associated with induction of occult osteosarcomas (Figure 23.7). Osteomas containing C-type retrovirus particles may be found in mice as spontaneous lesions, and increased numbers of osteomas have been observed in mice treated for life with fluoride or fluoride-containing compounds. Chordomas are uncommon axial skeleton neoplasms that arise from residual foci of primitive notochord anywhere along the ventral regions of the vertebral column, but they most commonly involve the cranial (skull) and caudal (tail) portions. They are slow-growing but highly infiltrative masses, producing bony destruction and frequently extending into adjacent soft tissues. They have been reported as spontaneous lesions in humans, rats, mice, dogs, ferrets, and mink.
Bone Modeling Alterations
Bone modeling as a process occurs principally during growth, but continues to a small extent throughout life. Toxic responses that depend on altered bone modeling affect compact bone and lead to insufficient or excess accumulations of bone or to mechanically inappropriate architecture. Bisphosphonates interfere with osteoclastic resorption of bone, causing an abnormal metaphyseal contour (“clubbing”) in the “cut back” zone and accumulation of unresorbed bone trabeculae in the marrow cavity beneath the growth plate (Figure 23.15).
Reactivation of bone modeling can occur in adult animals, as in the formation of marginal osteophytes in association with joint instability and degenerative arthropathy. Anabolic agents like PTH and PGE2 add bone mainly by modeling-dependent bone gain (uncoupled de novo bone formation). The production of woven bone in the adult or the resorption of bone in association with invasive neoplasia can be considered to be renewed modeling activities. Osteoclast-mediated bone resorption may occur because T-lymphocytes secrete interleukin-1β (formerly known as osteoclast-activating factor). This cytokine has powerful proerosive and proinflammatory properties, and along with TNF-α (another master proinflammatory cytokine) controls many signaling pathways responsible for bone and joint damage in both “inflammatory” (i.e., RA) and “noninflammatory” (i.e., OA) arthritides.
Bone Turnover Alterations
Bone turnover is an ongoing process under the influence of many different factors, both systemic and local. Effects on this process are visualized most readily in the metaphysis. In the growing bone, once the primary spongiosa is formed, the number of trabeculae is reduced by osteoclasts and their thickness increased by osteoblasts. The scalloped cartilage cores and reversal cement lines within metaphyseal trabeculae are evidence of this activity.
Increased bone in the metaphysis may be produced by several mechanisms. If osteoclastic resorption of the metaphyseal trabeculae is inhibited, increased bone results (such as after bisphosphonate administration). Increased bone in the metaphysis is associated with exposure to sodium/glucose cotransporter-2 (SGLT-2) inhibitors (in rats fed a standard laboratory diet), yellow phosphorus, or lead. The amount of metaphyseal bone also may be increased if the biomechanical load is enhanced or anabolic agents are administered. Usually agents that reduce bone turnover give a transient bone gain.
Any factor, including drugs and toxicants, that increase the rate of bone turnover in the metaphysis will lead to decreased bone. This finding will develop rapidly as reduced numbers and thickness of bony trabeculae since there normally is a loss in the amount of cancellous bone during metaphyseal turnover. In the rat, increased metaphyseal turnover after ovariectomy leads to decreased bone, which can be blocked using antiresorptive agents such as bisphosphonates and RANKL inhibitors. Osteoporosis is a pathologic reduction in bone mass in which the remaining bone is not normal, as indicated by loss of trabecular connectivity in three dimensions. In contrast, osteopenia refers to the state in which overall bone mass is reduced, but remaining bone is qualitatively normal. With aging, the amount of bone removed during remodeling declines, but the amount of bone that refills the resorption pit (i.e., Howship’s lacunae) declines even more; therefore, there is accelerated bone loss over time even though there is decreased bone resorption. Any condition or drug that enhances this deficit may cause irreversible bone loss, including malnutrition, high endogeneous levels or exogenous doses of corticosteroids, and chronic metabolic acidosis (which causes preferentially cortical bone loss in rats).
Background Bone Lesions
Most background lesions in animal species used in toxicity studies are more common in older animals and/or are associated with previous bone trauma or infection. Chondromucinous degeneration in the synchondroses of the sternum is the most common nonclinical background finding in rodents. Similar changes can occur in other cartilage locations, including long bone growth plates as well as intervertebral discs. The incidence in older animals can be quite high. The cause is unknown, but the change generally has no relationship to xenobiotic exposures.
Evaluation of new bone production in toxicity studies must consider species-related incidences of spontaneous increased bone. In laboratory rodents, two spontaneous lesions associated with increased bone include enhanced amounts of intramedullary cancellous bone in aged rats, a change traditionally termed hyperostosis. This type of increased bone is uncommon but occurs as a generalized condition, affecting long bones, the sternum, and vertebrae. The other condition, fibro-osseous lesion (FOL) in mice, is seen commonly in old females of certain strains but is rarely seen in males. FOL can be seen in the sternum, facial bones, and long bones, where increased bone begins as a partial replacement of the marrow cavity with undifferentiated mesenchymal cells and matrix. Osteoblasts, osteoclasts, and fibroblasts can be recognized in the lesion, as can fibrous tissue, osteoid, and woven bone. The cause of FOL is unknown. Murine strains with FOL exhibit a high incidence of ovarian cysts and cystic adenomatous hyperplasia of the uterus, suggesting that abnormal regulation of estrogen may play a role. Dietary estrogens have been shown to hasten the development of increased bone and are considered to be one of the many factors predisposing C3H mice to the development of osteosarcoma. Previous trauma or infection can lead to focal areas of decreased or increased bone. Metastatic neoplasms of soft tissues that spread to bone cause focal bone loss that may be detected radiologically. Such metastases may occur by direct invasion, but usually follow hematogenous spread and seeding of the bone marrow.
Common Toxicant-Induced Responses in Joints
Toxicant-related responses in joints following exposure to exogenous chemicals or biologics may affect any of the many intraarticular tissues. Typical joint findings reflect altered disease presentation or progression (e.g., in arthritis models) following xenobiotic treatment or an altered incidence and/or severity of spontaneous changes (many of which are age-related).
Inflammation
Acute synovitis is characterized by periarticular edema, fibrin deposition within the synovial membrane or joint space, and an inflammatory infiltrate (composed principally of neutrophils) that is limited to the synovium. With acute exudative reactions, edema and fibrin deposition in the joint may be considerable, with only a limited inflammatory cell infiltrate. However, in short order neutrophils migrate from the synovium into the joint space; synovial fluid may contain many more neutrophils than may be seen in a synovial biopsy. Even minor joint disturbances, such as saline flushing or instilling the joint with contrast agents, cause an acute inflammatory response with alteration in the surface lining cells.
More substantial reactions representing an escalation of synovitis to arthritis are associated with extension of the inflammatory process into the periarticular soft tissues and/or a substantial inflammatory cell influx into the joint space or other intraarticular tissues (e.g., fat pads). As acute inflammation subsides, the synovial inflammatory infiltrate changes character and becomes composed of mononuclear cells, initially with macrophages predominating. Over time, a lymphoplasmacytic infiltrate characterizes the immune-mediated arthritides (i.e., activation of an acquired immune response). Even in the face of large lymphoid aggregates in or near the affected joint, germinal centers are rarely seen.
All joint inflammations are accompanied by an increase in intermediate-type and type B (fibroblast-like) synovial cells. These cells alter their phenotypes when challenged to become facultative phagocytes. Fibroblast-like synoviocytes secrete several proinflammatory cytokines, which can sustain arthritis for long periods, and also proteases that contribute significantly to cartilage matrix degradation.
Hyperplasia/Hypertrophy
Synovial cells also can undergo hyperplasia and hypertrophy in response to arthritogenic stimuli (Figure 23.16). When this occurs, the lining cells increase in number, become plump and rounded, and have numerous surface filopodia. The synovial membrane has a tremendous capacity to undergo villous proliferation as a feature of low-grade, persistent inflammation. Villi form when fibrin deposited within the joint space is organized by invading mesenchymal cells, much like an organizing thrombus. Initially, villi that develop by this mechanism appear avascular and hypocellular, but later they become vascularized. Destruction of the articular cartilage occurs with, or more likely because of, synovial inflammation, especially where pannus (see below) arising from the inflamed synovium extend over the cartilage surface. Fragments of articular cartilage that detach or are eroded from the joint surface may lodge in synovial crevices and become incorporated as part of villous proliferation. These cartilage fragments gradually lose their proteoglycans (as shown by their altered staining pattern in histologic sections) and putatively evolve into a nidus for provoking continued inflammation and articular cartilage destruction as the body attacks the liberated proteoglycans and/or type II collagen.
In addition to joint inflammation and villous proliferation, there may be secondary responses within the synovial membrane, especially in chronic arthritis. These changes include rare osteochondral nodules, deposition of hemosiderin pigment within synovial lining cells or macrophages of the synovial membrane or articular capsule, fibrosis and articular capsule hypertrophy. There may also be mineralization within the synovial wall and joint capsule.
Cartilage Degeneration
Too much pressure, or too little, leads to structural and biochemical alterations in articular cartilage. Excess pressure causes cells in the superficial cartilage zone to become necrotic and the adjacent matrix to degrade. In contrast, insufficient pressure leads to cartilage matrix fibrillation. Articular cartilage has a high proportion of both water and proteoglycans, and is viscoelastic and highly deformable. Deformation of articular cartilage due to pressure or loading can lead to cartilage creep (i.e., displacement). The rate of creep depends on how quickly water can be forced out of the solid collagen–proteoglycan matrix. Water loss from the matrix leads to a condensation of the cellular and fibrillar components of cartilage.
Chondromalacia is the macroscopic term used to describe rubbery articular cartilage. Using H&E staining and light microscopy, chondromalacic cartilage often appears normal. However, a loss of cartilage matrix can be identified microscopically by altered matrix staining patterns as visualized by special staining techniques with cationic dyes such as safranin O or toluidine blue. Chondrocyte necrosis may be easily observed in some cases using light microscopy to visualize empty lacunae (Figure 23.17) or pyknotic (dark, shrunken) nuclei within lacunae; due care must be taken to avoid over-interpretation of artifactually empty chondrocyte lacunae since some cell loss is to be expected when trimming and cutting bony specimens. Loss of intercellular matrix, starting superficially and then moving deeper, follows degeneration of chondrocytes. Normal collagen fibrils within the cartilage matrix become more prominent by light microscopy as the proteoglycan matrix is degraded. Cracks and crevices develop and fragment the remaining cartilage into multiple longitudinal clefts (Figure 23.18). In some cases, portions of cartilage may become dislodged and develop into free-floating bodies (or “joint mice”) within the joint cavity.
Whenever there is cartilage cell necrosis, or loss of the superficial cartilage layers, the remaining chondrocytes undergo cell division to produce daughter cells. The closely aggregated families of new cells are frequently referred to as clusters, cell clones, or chondrones (Figure 23.17). Under normal circumstances, surface cartilage is worn away by friction and small cartilage clones develop in the middle zone, appearing to migrate toward the surface to be shed into the joint space. The deep margin of cartilage is slowly transformed into bone (chondroid bone). Progressive tidemarks of calcified cartilage form a historical marker of changes in the osteochondral border. While a significant amount of regeneration can occur (more than has been previously appreciated), the restored articular cartilage does not function as well as native tissue.
Associated Subchondral Bone Changes
In certain disease conditions of the adult, the normally slow cartilage modeling process is so exaggerated that the bone contour is altered. This process has been divided into progressive, regressive, and circumferential reactions. In the progressive reaction, there is an initial interstitial proliferation of cartilage leading to an increased thickness of the articular cartilage. The base of the cartilage becomes progressively mineralized, and the deep cartilage layers are invaded by blood vessels and converted into bone, resulting in thickening of the subchondral bone plate. Since the initial reaction represents a growth process, it is followed by remodeling of the tissue. The regressive reaction can be seen whenever loss of the cartilage surface results in the subsequent secondary loss of underlying bone. The cause of regression is unknown, but it may be related to mechanical overloading. This change happens during the process of eburnation (i.e., full-thickness, often localized loss of articular cartilage); occasionally, cartilage regenerates over the resulting bare bone surface, but the organization and matrix composition of the new cartilage is abnormal. The circumferential reaction consists of marginal osteophytes. These arise at the boundaries of affected joints due to periosteal bone formation, ossification of tendon or ligament entheses (attachment sites), or upward and lateral growth of cartilage with progressive transformation into bone. Osteophytes may be described as axial if near the center of the joint or abaxial if they are located at the periphery.
Trauma
The response of articular cartilage to trauma is dependent largely upon the depth of the injury. Superficial lacerations evoke only a short-lived metabolic and enzymatic response, failing to provide sufficient numbers of cells or matrix to repair the smallest injury. However, when the injury penetrates the osteochondral border, the response is similar to that occurring in other vascularized tissues, and results in the formation of granulation tissue. This response is similar to that arising in fracture repair except that bone proliferation stops at the cartilage–bone junction, leaving fibrous tissue to bridge cartilage wound edges. The fibrous tissue undergoes progressive hyalinization and chondrification to produce a fibrocartilaginous mass. Continuous passive motion has been shown to assist conversion of fibrocartilage to a more hyaline type of cartilage.
Vascular Changes in Joints
Not only does the vascular reaction occur within the subchondral bone marrow, but synovial vessels also may be stimulated during cartilage remodeling. When this occurs, a fibrovascular membrane (or “pannus”) originating from the synovium becomes attached to and invades the cartilage. The response of the subchondral osseous trabeculae is probably due to altered biomechanical stresses and strains resulting in part from decreased shock absorption by the articular cartilage. Trabeculae beneath damaged cartilage vigorously remodel to become thicker. When the trabecular thickness becomes excessive, central Haversian-like canals develop and the subchondral bone acquires an appearance more like compact bone. The chondrification of fibrous tissue filling the subchondral marrow may be deficient, and the tissue exhibits traits that are more characteristic of myxomatous connective tissue. There may be focal bone resorption, surrounding trabeculae may form a wall around the myxomatous reactive tissue, and the lesion can develop into a pseudocyst (i.e., a cavity lacking an epithelial or endothelial lining).
Neoplasia
Synovial sarcomas are the primary neoplasms that arise in joints. Synovial sarcomas typically have a biphasic cell population of spindle cells (usually in sheets) and epithelial-like cells in H&E-stained sections. However, one typically must examine numerous samples to appreciate the characteristic features of these tumors, which include slit-like spaces or a pseudoglandular pattern. Using immunohistochemical techniques, cells within synovial sarcomas show a biphasic pattern of intermediate filament protein expression, with the spindle cells containing vimentin and the plump epithelial-like cells staining for cytokeratin. Fibrosarcomas rarely develop in the periarticular tissues and can invade joints directly. These can be differentiated from synovial sarcomas because fibrosarcomas do not express cytokeratins.
Background Lesions
Spontaneous OA occurs in several species, including cynomolgus and rhesus monkeys, guinea pigs, mice, and rats. It is typically slowly progressive, and may be present at a high incidence microscopically in older animals. A severe unilateral destructive OA of the hip (idiopathic chrondolysis) in young cynomolgus monkeys has been reported, but the etiology is unknown (Figure 23.19).
Mechanisms of Toxicity
Although the skeletal system is exposed to circulating xenobiotics, it is not known to have a significant role in biotransformation of such substances. Agents reported to have a direct toxic effect on bone and/or cartilage are somewhat limited. It may be that we have not yet learned to evaluate the early primary toxic effects of drugs, chemicals, or other environmental agents on hard tissues. A list of agents and toxic mechanisms causing skeletal pathology are given in Table 23.3.
Table 23.3
Common Mechanisms of Bone and Joint Toxicity
These represent known actions that have been well studied. There is a wide divergence in the way many agents produce their effects, which can be observed at different organizational levels (e.g., molecular to whole animal).
Abbreviations: BAPN, β-aminoproprionitrile; DKK-1, Dickkopf-related protein 1; FGF-23, fibroblast growth factor 23; GH, growth hormone; IGF-1, insulin-like growth factor 1; MMP, matrix metalloproteinase; PTH, parathyroid hormone; RANKL, receptor activator of nuclear factor-κB ligand.
Table adapted from Handbook of Toxicologic Pathology, second ed. W. M. Haschek, C. G. Rousseaux, and M. A. Wallig, eds. (2002) Academic Press, Table 24.1, pp. 467–470, with permission.
Bone Toxicity
Primary Toxicity
Aseptic necrosis of the femoral head caused by alcohol consumption or corticosteroid administration is not due to direct drug-induced necrosis of osteocytes, but rather to compromised local circulation in a region of bone that is anatomically vulnerable due to its hemispherical shape and considerable range of motion. In contrast, cyclophosphamide, a cytotoxic antineoplastic agent, directly influences mitotic division of bone cells, thereby inhibiting bone formation and growth. Antibiotics of the quinolone class (e.g., nalidixic acid, ciprofloxacin) have been shown to cause chondrolysis.
Bisphosphonates, which are antiresorptive agents, are rapidly incorporated into mineralized bone and calcified cartilage. When osteoclasts take up bisphosphonates released from mineralized bone and cartilage by acid digestion, their function is impaired. Toxic intracellular metabolites such as ApppI (triphosphoric acid 1-adenosin-5′-yl ester 3-[3-methylbut-3-enyl] ester), a cytotoxic ATP analog that promotes apoptosis are produced. The osteoclast may undergo apoptosis, have a shortened lifespan, and exhibit morphologic changes such as the formation of giant osteoclasts.
Secondary Toxicity
Since skeletal tissues do not play a major role in biotransformation or elimination of potentially toxic substances from the body, most exogenous chemicals exert their effects on the bone via secondary mechanisms. It has been suggested that secondary toxic effects may be mediated by alterations in blood flow within bones. Increased oxygen tension is thought to stimulate bone resorption locally, while decreased oxygen tension causes bone to accumulate. The close similarity of some drug-induced bone reactions to those observed in spontaneous diseases mediated by a vascular reaction supports the argument for a vascular pathogenic mechanism. Conversely, such reactions could be a manifestation of the limited number of ways in which skeletal tissues can respond to injury.
An example of a tissue reaction that might be mediated by a vascular response is the subperiosteal formation of new bone in dogs following the administration of PGE2. The drug-induced gross and microscopic lesions are nonspecific but are very similar to those observed in hypertrophic (pulmonary) osteo(artho)pathy (HPO). Substances capable of inducing secondary bone changes usually are hormones, vitamins, or minerals, or affect the metabolism of skeletal-regulating agents. The long-term effects of PTH, corticosteroids, somatotropin, calcitonin, bisphosphonates, and vitamin D and its metabolites on bone resorption and formation are similar in quality but vary in magnitude. When new remodeling units are activated to increase bone turnover (e.g., by PTH or thyroid hormone), the initial bone resorption enlarges the remodeling space and leads to a temporary, nonprogressive bone loss. Conversely, the many agents that depress activation of bone turnover (e.g., calcitonin, increased dietary calcium and vitamin D, estrogens, nonsteroidal antiinflammatory agents, and bisphosphonates) reduce resorption first, leading to a small, nonprogressive net gain in bone.
Some agents act on several or all of the sequential remodeling stages (e.g., corticosteroids, PTH), whereas others act on only one particular stage. This role of a given agent with respect to regulating the turnover process will determine whether or not a specific modality of timed drug delivery is effective in eliciting a skeletal response. Since bone turnover events represent a sequence of different biological activities, intermittent, brief, or continuous delivery of the same agent may give distinct skeletal responses. The classic example of this phenomenon is the anabolic action of intermittent PTH administration versus the high bone turnover observed with continuous PTH delivery. Skeletal responses to pharmacologic agents may diminish with time as tolerance develops. Diminution of the skeletal response is seen with longer-term usage of bone anabolic agents such as PTH; the response plateaus because the additional bone activates the negative mechanical usage feedback loop to rid the skeleton of unneeded bone. One probable factor responsible for different drug doses causing varied biological responses is the differential action of drugs on the modeling and turnover stages. For example, a very small dose of 1,25-dihydroxy vitamin D3 reverses the endochondral growth abnormality associated with vitamin D deficiency, but high doses can produce markedly increased osteoid in the metaphysis.
It is important to recognize that skeletal cell-to-cell and cell-to-matrix interactions involve both bone cells and cells of the bone marrow and are controlled by both systemic and local factors. Therefore, cell functional activities and cellular interactions represent the final common pathway of toxic action rather than a basic mechanism of toxic action. Local events are influenced by circulating agents such as hormones, but the effects of circulating agents are also determined by local factors such as mechanical usage. In addition, the processes active in bone disease are not necessarily the same processes operating in physiologically normal states.
Certain diseases appear to affect one bone envelope (periosteal, Haversian, endosteal) more than others. In general, the endosteal bone envelope is more reactive and responsive than the Haversian envelope because endosteal (endocortical and trabecular) bone possesses a greater cumulative surface area, has cells with higher metabolic activity, and experiences higher bone turnover. The surface-to-volume ratio and turnover rate is three times higher in trabecular bone than cortical bone in both humans and dogs. In the rapidly growing young rat, the endocortical surface of the metaphysis is predominantly formative, while the periosteal metaphyseal surface is predominantly resorptive as the diameter of the bone is progressively reduced from wide at the physis to relatively narrow at the diaphysis. Certain drugs also appear to affect particular envelopes, and there are species differences as well. For example, PGE2 administration in dogs causes greater subperiosteal proliferation than endosteal proliferation of woven bone (Figure 23.20). In contrast, in rats endosteal proliferation is greater following PGE2 treatment. Mice treated with estrogens produce much more marrow cancellous bone than do other animals. Bone loss following ovariectomy in rats and nonhuman primates is much greater than in dogs.
Age also affects the responsiveness of the skeletal system. It is well recognized that fracture healing is less vigorous in old animals than in young individuals; the same is true concerning the response of the skeleton to circulating toxicants. The modeling and turnover processes present beneath the growth plates of growing animals respond to xenobiotics to a greater extent than do those in cortical or cancellous bone of older animals. In chronic studies, drug toxicity may influence (or be influenced by) the incidence and severity of spontaneous lesions occurring in the animal strain being used. For example, chronic studies with nitrofurazone show a drug-related effect that greatly increases the distribution and severity of age-related degenerative cartilage changes in rats.
Toxic Effects on Endochondral Ossification and Longitudinal Bone Growth
The active physis of rapidly growing long bones in rats is often affected by both small molecule and biologic pharmaceutical agents. In the context of drug safety risk assessment, it is important to keep in mind that many test article-related effects on the growing physis represent on-target pharmacology and may not be relevant to human safety. Decisions regarding the relevance of such data depend on the proposed indication, exposure margin, and skeletal maturity of the intended population of human patients.
Decreased longitudinal bone growth is a common finding in toxicity studies conducted in rapidly growing young rodents and may be secondary to decreased food intake (inanition) or occur as a consequence of test article activity. There are two main categories of disruption to endochondral ossification: physeal dysplasia and physeal dystrophy. While physeal dysplasia and physeal dystrophy reflect very different toxicologic mechanisms, both lead to functional impairment of the physis and decreased bone production and impaired growth.
Increased Thickness, Physis (Physeal Dysplasia)
Toxicity leading to increased physeal thickness is most notably represented by inhibitors of tyrosine kinases (e.g., anti-VEGF agents), MMPs, activin receptor-like kinase 5, and FGFs. Any substance capable of interrupting the transition from hypertrophic cartilage to the calcified cartilage spicules of primary spongiosa, whether by inhibition of angiogenesis, reduced vascular penetration, chondrocyte cell death, or a combination of these processes, can cause physeal dysplasia. Long bone physes are widened, sometimes dramatically, with disorganized and expanded chondrocyte columns, particularly in the hypertrophic zone (Figure 23.21). Transverse fractures through the widened (and weak) dysplastic physis can occur through any cartilage zone. When exposure to these inhibitors ceases, these findings usually resolve to a remarkable degree in short order, although some slight residual chondrocyte column disorganization may persist. Physeal dysplasia is most readily appreciated in the young, rapidly growing rat; it may not be readily apparent in older rodents with little or no longitudinal bone growth. This process has also been reported in skeletally immature nonhuman primates.
Decreased Thickness, Physis (Physeal Dystrophy)
In decreased physeal thickness, longitudinal bone growth is impaired due to decreased chondrocyte proliferation in the physis, which is characterized by a thin, relatively inactive physis from which fewer and shorter primary spongiosa are produced (Figure 23.22). Secondary spongiosa may be thin if osteoblast function is impaired (e.g., with glucocorticoids), or may be thick if osteoblast function is relatively unchanged. Agents causing this toxicity are most notably represented by inhibitors of the GH/IGF-1 axis, such as glucocorticoids, and somatostatin. Decreased physeal thickness with thicker secondary spongiosa occurs with inanition in rapidly growing rats that have significant weight loss due to the impact of reduced energy availability on the GH/IGF-1 axis. The ultimate manifestation of decreased physeal thickness is physeal closure, which occurs prematurely if the condition develops is manifested in young animals.
Toxic Effects on Bone Modeling and Turnover
While bone modeling and bone turnover are processes of bone growth and bone homeostasis, respectively, both processes share common molecular pathways. Thus, both processes may be affected by a large array of pharmacologic agents.
Toxicologic Effects on Osteoclasts
Effects of Impaired Osteoclast Formation in the Metaphysis
Antibodies to the Wnt pathway inhibitor Dickkopf-related protein 1 (DKK-1) decrease osteoclast formation. Decreased numbers of osteoclasts subjacent to the zone of hypertrophic chondrocytes impair the normal elimination of extraneous calcified cartilage spicules. Primary and secondary spongiosa trabeculae numbers are increased, leading to a markedly increased trabecular bone volume in the metaphysis. Other pharmacologic agents with similar effects on osteoclasts include RANKL inhibitors.
Bisphosphonates are rapidly incorporated into calcified cartilage and mineralized bone, rendering these extracellular hard tissues resistant to acid digestion by osteoclasts. The number of surviving calcified cartilage spicules in the proximal metaphysis (retained primary spongiosa) increases markedly (Figure 23.15). Exposure to bisphosphonates causes characteristic changes to osteoclasts and erosion surfaces. Osteoclasts are typically enlarged with increased numbers of nuclei. Trabeculae of bisphosphonate-treated bone may have slightly undulating surfaces associated with these enlarged osteoclasts, which may reflect the inability of the osteoclast to produce a resorption cavity of normal (more deeply scalloped) size.
Cathepsin K inhibitors produce direct antiresorptive effects on osteoclast function. Cathepsin K, an intracellular enzyme that digests type I collagen, is highly expressed in osteoclasts as a means of degrading demineralized bone matrix. Bone resorption is reduced to a similar degree as with bisphosphonates, but bone formation is enhanced because osteoclast crosstalk to osteoblasts remains intact. Retained primary spongiosa and increased trabecular bone are morphologic features of bones from cathepsin-K deficient mice. Conversely, Wnt inhibition decreases ossification, resulting in reduced trabecular bone formation in the metaphysis of growing animals.
Enhanced Osteoclast Recruitment/Function
The classic toxicologic example of increased osteoclast function is represented by continuous exposure to PTH or PTHrP. Bone turnover is markedly increased, especially with continuous PTH exposure, and can occur in inappropriate locations such as directly beneath the osteoid seam or in the middle of the trabecula (so-called “tunneling resorption,” Figure 23.23). Misplaced bone removal results in trabecular splitting and a paradoxical increase in trabecular connectivity in the face of decreased trabecular bone volume by this ultimately catabolic process.
Toxic Effects on Osteoblasts
Decreased Osteoblast Formation/Function
At high doses, first generation bisphosphonate compounds, including etidronate and clodronate, impair osteoblast function as well as osteoclast function, resulting in extremely low or effectively no bone turnover. Osteoid area and volume are augmented (increased osteoid), indicating a mineralization defect. Administration of such agents to young but skeletally mature dogs has been demonstrated to alter the structural integrity of the ribs and vertebral processes, leading to multiple spontaneous fractures. Increased signaling by peroxisome proliferator-activated receptor gamma (PPAR-γ) produced either directly by PPAR-γ agonists or indirectly by modulators of PPAR-γ activity (e.g., FGF-21 in mice) can have profound influence on the differentiation of mesenchymal bone marrow stem cells. Such increases may divert these stem cells toward adipocyte differentiation and away from osteoblast differentiation. Increased adiposity of bone marrow with reduced bone formation occurs as a result of this shift. The resulting decreases in trabecular and cortical bone volume in the mouse can be dramatic.
As noted above, many inhibitors of chondrocyte proliferation, such as glucocorticoids, also inhibit osteoblast function. If these inhibitors have a relatively short half-life compared to the dosing interval (i.e., a pulsatile exposure regimen), variation in the thickness of the secondary spongiosa may occur with sharply linear transverse bands (Harris growth arrest lines) in the proximal metaphysis of a growing animal (Figure 23.12). With respect to homeostatic bone remodeling, the decreased bone formation leads to a net bone loss since the steady state of bone resorption followed by equivalent bone formation is no longer in balance.
Increased cortical porosity and decreased trabecular bone volume are typical findings with bone loss due to glucocorticoids, which can progress to osteoporosis if exposure is chronic. These effects result from the ability of steroids of this class to impair osteoblast activity.
Increased Osteoblast Formation/Function
Intermittent exposure to PTH or its analogs increases bone formation by increasing osteoblast numbers. The outcome occurs in part due to inhibition of the Wnt antagonist sclerostin. Marrow fibrosis is another classic feature of continuous PTH exposure, and there is evidence that the cells responsible for this response are collagen-producing mesenchymal stem cells that are destined to become preosteoblasts.
Treatment with anti-DKK antibodies removes Wnt inhibition to osteoblast differentiation. In growing bone, the combination of reduced osteoclast formation (see above) and increased osteoblast numbers by inhibition of DKK-1 results in a metaphysis with greatly increased amounts of trabecular bone and a decreased marrow space resembling the appearance and structure of osteopetrotic bone. Treatment with Wnt inhibitors in animals with actively growing bones leads to a decrease in ossification beneath the growth plate, with a reduction in bony trabeculae within the metaphysis.
Aberrant Bone Production
Anticancer pharmaceuticals often have profound effects on bone marrow, with rapid production of woven bone appearing in the bone marrow space secondary to primary deleterious effects on hematopoietic progenitor cells. This new bone is typically noted in the middle of a segment of sternum, or in the deeper metaphysis. The mechanism for this response is thought to be stimulation from local growth factors upregulated in response to myelotoxicity. This woven bone rapidly resolves after cessation of treatment. Aberrant woven bone has also been reported in periosteal, endocortical, and trabecular bone regions in dogs and rats treated with PGE2. This molecule induces osteoblast differentiation via activation of the prostanoid EP4 receptor. A similar change is observed in trabecular bone of dogs treated with toxic doses of YM175, a bisphosphonate.
A well-known cause of abnormal bone formation is lathyrism in rats fed sweet pea (Lathyrus odoratus) or treated with its toxic agent, beta-aminopriopionitrile. This toxin inhibits lysyl oxidase, the enzyme responsible for the hydroxylation of lysine, the normal precursor to crosslinking of elastin and type I collagen fibrils. Reduced fibril crosslinking in connective tissue leads to an abnormal bone matrix and decreased bone growth and strength. Copper is a cofactor for lysyl oxidase, so copper deficiency can also depress lysyl oxidase activity. Pronounced periosteal bone proliferation (increased bone) at the tendon insertion sites of the adductus longus and pectineus on the femur is a classic finding. Distortion of other long bones and spine also occur.
Neoplasia
Prolonged stimulation of osteoblasts by some PTH analogs has been shown to produce a spectrum of osseous neoplasms in rats. Lesion types include osteosarcoma (Figure 23.7), osteoma, and osteoblastoma.
Toxic Responses in Joints
Inflammation
The underlying biochemical mechanisms involved in immune-mediated and nonimmune articular reactions to drugs are similar, since the production of local factors in both situations can lead to chondromalacia or degradation of articular cartilage. Systemic administration of small molecular weight peptidoglycans (typically of bacterial origin for animal arthritis models) can induce acute polyarthritis by a nonimmune-mediated process, an inflammatory response thought to be mediated by mast cell degranulation. Development of immune-mediated arthritis with exposure to peptidoglycans or other antigens requires that the antigenic material be deposited in the joint and the development of delayed hypersensitivity. Nonimmunoglobulin, T cell-derived molecules (e.g., lymphokines) that bind specifically to antigens are the effector molecules. One such protein, albeit incompletely characterized, has been designated “arthritogenic factor” because of its ability to sustain proliferative synovitis when instilled into the joint cavity. The joint lesion that develops following a single injection of arthritogenic factor persists for at least 4 weeks and does not require complement.
The chronicity of antigen-induced arthritis depends on the persistence of a sufficient amount of antigen in the affected joint. Antigen retention is mediated by antibody-dependent trapping. The electronic charge of the antigen also appears to determine the development of arthritis because antigen penetration into cartilage matrix depends on both molecular weight and charge. Immune complexes trapped in collagenous tissues within joints provoke an inflammatory response; the pathologic role of sequestered antigen in maintenance of the chronic inflammatory response lies in long-lasting leakage of antigen into surrounding tissue. Immune complexes can be phagocytosed by macrophages or synovial cells, which then produce IL-1 to perpetuate the inflammation. Alternatively, complement that has leaked into the joint cavity across the inflamed synovium is activated by immune complexes. Production of C3a and C5a components is a further means of inducing IL-1 secretion.
Systemically delivered therapeutic agents that are immune-stimulant compounds have been known to cause destructive arthritis in conventional studies in laboratory animals. Immune stimulants have been designed for therapy of certain systemic diseases, and adjuvants are crafted for a similar purpose with vaccines, to increase the immune response. Such compounds may cause lymphoid hyperplasia with prominent increases in B cells in lymph nodes and other lymphoid organs as well as less prominent T-cell increases in lymph nodes and spleen. It is likely that cytokine production is increased, especially IL-1. Joint lesions produced by immune-stimulant agents (Figure 23.24), which affect mainly the tarsus, carpus, and phalanges, are similar to those seen in animal models of RA, such as AIA. Approximately 10–14 days after introduction of the adjuvant (depending on the adjuvant and antigen), rats develop swollen paws with skin discoloration. Microscopic examination reveals periosteal and synovial inflammation with fibrin deposition in synovial membranes, expansion of joint spaces due to fluid exudation (and sometimes neutrophil accumulation), periosteal new bone formation, and marked osteoclastic erosion with numerous, prominent Howship’s lacunae. Visible erosion or loss of articular cartilage is limited to occasional joint surfaces adjacent to thickened synovium that was encroaching into the joint space (Figure 23.24), but the cartilage matrix of the remaining cartilage is degraded. Such synovium often contained inflammatory cells. Articular cartilage was intact and appeared normal in many joints with severe peri-articular reactions although matrix loss can be substantial in apparently normal cartilage in extensively inflamed joints. Cartilage loss would increase after a longer period of disease.
Other systemically delivered compounds cause joint swelling but with a more fibroplastic response. These include bleomycin, in which the fibroplasia is considered to develop because of activation of macrophages. The fibroplasia or “tendonitis” that develops during treatment with MMP inhibitors is thought to occur via blockade of tissue collagenases that play a role in the normal turnover of collagen in tendons, ligaments, and periarticular connective tissue. Rats treated with 6-sulfanilamidoindazole develop an acute exuberant synovitis and periarthritis with arteritis that can resolve in a few weeks. No concomitant changes in bone or cartilage have been demonstrated.
Degeneration
Degradative changes in articular cartilage leading to subchondral bone erosion are a feature of both inflammatory and degenerative joint disease (Figure 23.18). Loss of articular cartilage may result since normal matrix removal continues in the face of decreased synthesis. Lysosomal enzymes (collagenase, cathepsins, elastase, and arylsulfatase) are present in inflammatory, synovial, bone, and cartilage cells. Their release may cause degradation of proteoglycan (as in papain-induced or hypervitaminosis A-associated cartilage matrix degeneration). Free radicals like superoxide (
) or peroxide (H2O2) generated by local enzyme activity can depolymerize polysaccharides and degrade synovial fluid and cartilage. Synovial lining cells from rats with AIA generate H2O2 constitutively, whereas cells from corresponding areas of control rats do not.
The cartilage catabolic process also involves aggrecanase or neutral metalloproteases, which act on connective tissue macromolecules (collagenases, proteoglycanases). These enzymes are produced by synoviocytes and chondrocytes. Their secretion into the joint is induced by peptidic factors released from cells of the immune system. Among these mediators, IL-1 plays an important role. Since IL-1 is also produced by activated synovial cells, a sustained inflammatory condition does not seem to be necessary for it to be secreted in considerable quantities over an extended period. Degradation of articular cartilage does not necessarily require an exogenous source of enzymes arising from synovial cells or from neutrophils within synovial fluid. Neutral metalloproteases can be released from chondrocytes themselves as latent enzymes requiring activation, probably by serine proteases such as plasminogen activator. IL-1 induces the synthesis of cartilage neutral metalloproteases, stimulates the production of plasminogen activator, and promotes the destruction of cartilage matrix macromolecules. Tissue inhibitors of metalloproteases (TIMPs) are expressed constitutively in cartilage to prevent overactivity of the proteases. Catabolic states within cartilage are characterized by inactivation of TIMPs, resulting in activation of metalloproteases.
Cartilage Necrosis
In the case of drug toxicities, as seen with the quinolone class of antibiotics, the distribution of necrotic chondrocytes in articular cartilage is unique. Necrosis usually occurs only in animals in which the articular cartilage is still in the process of growth, and begins in the zona intermedia (transitional zone) of the articular cartilage. Necrosis of chondrocytes and subsequent matrix lysis produces a cleft seen grossly as a blister on the articular surface. This bleb quickly ruptures. Tendon ruptures have been associated with fluoroquinolones, which is thought to result from their cytotoxicity (demonstrated in vitro) to tenocytes.
Fibroplasia
Rats treated with marimastat and other potent MMP inhibitors, delivered continuously in subcutaneous pumps, develop swollen paws, often with reddening, within 2–3 weeks. Histologically there is fibroplasia, often accompanied by inflammation, in the subcutis, synovium, joint capsules, ligaments, tendons, tendon sheaths and peri-articular skeletal muscles.
The most sensitive site for developing fibroplastic change is the tarsus, followed by the patellar attachment to the quadriceps tendon and the carpal ligaments. Severely affected animals also have fibroplasia in the carpus, and often in the digits of both fore- and hindpaws. Fibroplasia in the joint capsule often extends into the joint space and is associated with substantial angiogenesis, forming “pannus.” Occasional foci of cartilage metaplasia and frequently new bone formation also are evident, either at the periosteal surfaces or within the new fibrous tissue. Bone resorption associated with abundant osteoclasts is sometimes present in the cortex of the metatarsals, tibia, and other bones. Inflammation, predominantly lymphocytic, is often present in the proximal part of the gastrocnemius muscle and the distal part of the quadriceps.
Summary
Test articles vary widely in the time needed (weeks to months) to produce an effect on bone and cartilage. The physis and the metaphysis are the most metabolically active sites, and examination of these domains in a fast-growing bone of young animals will give the greatest opportunity to detect xenobiotic-related toxic effects. Many chemical agents affect the skeletal system indirectly via their actions in mediating cell differentiation or modulating cell function. Therefore, skeletal effects may be reflected in the rates at which bone cells function. Evaluation of routine formalin-fixed, decalcified, H&E-stained sections is a suitable first-tier screen for skeletal toxicity, but mechanistic information and a deeper understanding of the pathogenesis often requires such special second-tier techniwues as assessment of bone biomarkers, bone density, bone biomechanical properties, and bone histomorphometry to uncover mechanisms by which skeletal alterations have occurred. Animal models, either naturally occurring or created as a result of genetic, surgical, and/or mechanical manipulation, are useful in assessing the efficacy or toxicity of potential therapeutic agents. Understanding the utility and limitations of each model with respect to predicting human responses is vital to interpretation of the data generated.