Abstract
During this presentation, a variety of class effects were reviewed by their differing effects on bone, including inhibition of endochondral ossification, inhibition of the growth hormone-insulin-like growth factor 1 axis, promotion of bone formation, inhibition of bone formation, abnormal bone formation, promotion of bone resorption, inhibition of bone resorption, and bone necrosis.
Keywords
Inhibitors of endochondral ossification include any substance that acts to impede this intricate process. The morphologic findings are similar: increased thickness of the physis with increased numbers of hypertrophic chondrocytes which may be disorganized from their normally neat vertical stacks. Sequelae from this inhibition may include decreased longitudinal bone growth and potentially fractures through a relatively weak thickened growth plate. Vascular endothelial growth factor (VEGF) inhibitors are one of the best-known classes; by preventing neovascularization at the ossification front, no new bone is formed and the hypertrophic chondrocytes (normally the source of VEGF for this process) become much more numerous and distortion of a wider physis is present, along with a lack of penetrating new blood vessels at the ossification front. 1 Another common class is that of matrix metalloproteinase (MMP) inhibitors. Some MMPs found in bone (MMP9) affect endochondral ossification through impacting VEGF signaling, whereas others (MMP13) are responsible for digestion of the cartilage walls of hypertrophic chondrocytes.2,3 Inhibition of either type of MMP will lead to a thickened and disorganized growth plate, whether it be via disrupted VEGF signaling or impaired cartilage digestion. Fibroblast growth factor receptor (FGFr) inhibitors can also cause this finding, via inhibition of MMP13. 3 General cytotoxicity of the bone marrow compartment can have bystander effects on the process of endochondral ossification by wiping out the blood vessels and bone cells which are required for endochondral ossification. 4
Less potent first-generation bisphosphonates (such as etidronate) can also cause increased physeal thickness likely due to impairment of local vascularization and a direct inhibitory effect on osteoblasts and osteoclasts. 5
Another type of impedence to endochondral ossification occurs with the inhibition of the growth hormone-insulin-like growth factor 1 (GH-IGF1) axis. This axis is key to chondrocyte proliferation in the physis, and inhibition leads to a thinner growth plate (decreased thickness, physis) with few proliferating chondrocytes. Inhibitors of GH-IGF1 axis include corticosteroids (which also impair osteoblast function), FGF21 analogues/agonists (which decrease IGF1), and somatostatin (which inhibits the effects of GH).6,7 Somatostatin can lead to physis inactivity, characterized by a lack of primary and secondary spongiosa in the metaphysis and increased marrow adiposity. 8 Thicker secondary spongiosa may or may not be present. Decreased food consumption causing lower body weights or body weight gain may also decrease longitudinal bone growth, but it should be in proportion to these changes, and morphologic findings may not be present unless the decreased food consumption and body weight changes are severe. 9
An extreme effect on chondrocyte proliferation is caused by inhibition of the hedgehog (HH) pathway, which is another key pathway for chondrocyte proliferation and can cause closure of the physis; an example of physis closure using a Smoothened Hedgehog (SHH) inhibitor is shown. Partial rescue of the growth plate closure can be attained by administration of parthyroid hormone–related peptide (PTHrP), which is downstream of one of the HH pathways to increase chondrocyte proliferation. However, HH also impacts chondrocyte proliferation in a process independent of PTHrP, so complete rescue may not be attainable with this strategy. 10
Moving on to class effects on metaphysis, effects on bone formation were then presented. Increased bone formation through increased osteoblast function is most classically associated with intermittent administration of PTH, PTHrP, and PTH analogues. 11 A more recent modality of bone anabolism is sclerostin inhibition. 12 The wnt pathway maintains bone mass. Sclerostin inhibits the wnt pathway, thereby ultimately decreasing bone formation.
Both intermittent administration of PTH and inhibition of sclerostin work via favoring bone formation via the wnt pathway. 13 The morphologic appearance is of thicker and/or more numerous trabecular bone.
A special instance of increased bone formation occurs as an off-target effect of sodium glucose limited transport 2 (SGLT2) inhibitors. This effect is species-specific (rat) and diet-specific (glucose) and is characterized by increased trabecular bone in the deep metaphysis. 14 The sodium glucose limited transport 1 (SGLT1) pathway is inhibited, resulting in increased intestinal absorption of calcium. Inhibition of the SGLT1 pathway has also been associated with decreased osteoclast function, 15 which leads to increased primary spongiosa at the ossification front and irregularly shaped cortical bone in the metaphyseal cut-back zone. Substitution of fructose for glucose in the diet prevents inhibition of SGLT1 in rats. 14
Inhibition of bone formation is most associated with inhibition of the wnt pathway, 13 a common target in oncology. Primary spongiosa will not progress to secondary spongiosa, and osteoblasts are uncommon.
With some modalities, bone formation may be present but abnormal. One of the more common types of abnormal bone seen in toxicity studies is woven bone, and one of the most common causes is oncology agents that cause severe decreased hematopoietic cellularity of the bone marrow. 4 As the bone marrow recovers, cytokine signaling affects the marrow population of mesenchymal stem cells, which differentiate into osteoblasts. These signals may be so ubiquitous that osteoblasts are not oriented similarly and the collagen I produced does not have a lamellar orientation and is randomly oriented. 8 This difference in collagen I fibril orientation can be best appreciated using polarized light with which bone may appear as “woven” versus the normal “lamellar” orientation. Woven bone can be produced very quickly and is characterized by the lack of lamellar orientation with irregular and narrow trabecula with large and irregularly placed osteocyte lacunae. Woven bone is weak due to the random orientation of the collagen fibrils. In the case of a cytotoxic oncolytic agent, the bone may completely recover after the dosing regimen is complete, while with some bone-active therapies, woven bone may not resolve. 16
Mineralization defects also lead to abnormal bone. These defects can include defects in the matrix: mineral deficiencies (calcium, phosphorus) which cause rickets (osteomalacia) or mineral toxicoses (fluoride) which can cause increased but weakened bone, as well as inadequate collagen I production or crosslinking. 8 While differential staining (pale) on hematoxylin and eosin (H&E)-stained slides can strongly suggest a mineralization defect (including increased osteoid), quantitative assessment is not possible on decalcified tissue and assessment of undecalcified tissue is needed to accurately define unmineralized osteoid.
Dosing regimen can have profound effects on bone cellular response. The classic example is PTH; intermittent dosing leads to an anabolic effect, whereas continuous dosing (or exposure) leads to catabolism. 10 Intermittent dosing of PTH or PTHrP analogues causes positive bone balance due to increased bone formation over bone resorption and is the basis of the efficacy of teriparatide (Forteo) and similar biologics. Continuous exposure to PTH or PTHrP, such as found with hyperparathyroidism (primary or secondary) or with neoplasia (PTHrP), leads to highly increased bone turnover with bone resorption predominating (negative bone balance); osteoclast numbers increase markedly and can be found within trabecula (tunneling resorption or trabecular splitting), and marrow fibrosis is widespread. Another example of a differing bone response to dosing regimen is methotrexate, where a low dose causes decreased osteoclasts (anabolic) and a high dose leads to increased osteoclasts, decreased osteoblasts, and proliferating chondrocytes (catabolic). 8
A large class of bone-active therapeutic agents are those with anti-resorptive properties. Osteoclasts are either functionally impaired (bisphosphonates, cathepsin K-inhibitors) or not produced (receptor activator of nuclear factor kappa-Β ligand [RANKL] and Dickkopf-related protein 1 [DKK1] inhibitors).17-20 These agents are widely used to treat osteoporosis as well as other conditions where excess osteoclast activity is a problem (metastatic cancers, Paget’s disease, etc.). The hallmarks of these agents in nonclinical toxicity studies are primary spongiosa retained at the ossification front, abnormal formation of the metaphyseal cortex (clubbing at the cut-back zone), and either a lack of osteoclasts (RANKL or DKK1 inhibition) or osteoclasts with abnormal morphology (such as osteoclasts with increased nuclei, some apoptotic, with bisphosphonates, or osteoclasts with red cytoplasmic granules with cathepsin K inhibition). At high doses, bone formation may also be decreased (first-generation bisphosphonates).
Finally, the criteria for bone necrosis were discussed. Inexperienced pathologists may overcall this finding. When assessing bone necrosis, bone lacunae should be devoid of osteocyte nuclei or should contain evidence of pyknosis or cellular debris. The size of the affected bone should exceed 500 μm2. 21 It is extremely important to assess the structures surrounding the bone. If the necrosis is acute, there should also be necrosis of surface bone cells and the adjacent bone marrow and adipose tissue; hemorrhage may be present. In the repair stage, there should be a vigorous response by surface bone cells: osteoclasts should be resorbing the dead bone and osteoblasts should be laying down new bone to replace the dead tissue or to wall it off. When evaluating bone tissue stained en bloc with basic fuchsin, the staining should be paler than normal bone.
It is crucial to be cognizant of species, age, and sex differences in bone. Bone turnover in rodents is different (based on modeling) than the larger species (or humans) used in toxicity studies (based on remodeling and some modeling). 8 After the main period of growth in mice, long bones contain little trabecular bone; lumbar vertebrae are the preferred alternate site. The metaphyseal trabecular pattern is different between the sexes in rats (females have more uniform trabecular structure, and males have more trabecular bone near the cortices and less in the center of the metaphysis). 22 Growth plate thickness is age- and sex-dependent during rapid growth; at the same age in rats, growth plates will be thicker in males than in females as the males are growing faster. Controls should be age- and sex-matched.
In the larger animal species (primates and dogs), the growth plates in females close earlier than in males. The costochondral junction of the rib is a useful location to evaluate the growth plate in the larger animal species. 8
In all species, be aware of bone site differences (such as the faster growth rate of the distal femoral physis compared with the proximal femoral and proximal tibial physes).
The ossification front and metaphysis are areas of high metabolic activity, particularly in rapidly growing animals commonly used in toxicity studies.
Numerous pathways can affect bone cells; keep in mind the tremendous cross-talk between osteocytes, osteoblasts, and osteoclasts.
The International Harmonization of Nomenclature and Diagnostic Criteria (INHAND) terminology23 has been developed to promote microscopic nomenclature consistency among toxicologic pathologists and to prevent confusion of terms which may have different connotations in clinical use by emphasizing a descriptive approach. Toxicologic pathologists are strongly recommended to use this terminology in safety assessment.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
