Stereological and serial section analysis of chondrocytic enlargement in the proximal tibial growth plate of the rat

Cellular basis of differential endochondral growth in Lake Malawi cichlids

BACKGROUND

The shape and size of skeletal elements is determined by embryonic patterning mechanisms as well as localized growth and remodeling during post-embryonic development. Differential growth between endochondral growth plates underlies many aspects of morphological diversity in tetrapods but has not been investigated in ray-finned fishes. We examined endochondral growth rates in the craniofacial skeletons of two cichlid species from Lake Malawi that acquire species-specific morphological differences during postembryonic development and quantified cellular mechanisms underlying differential growth both within and between species.

RESULTS

Cichlid endochondral growth rates vary greatly (50%-60%) between different growth zones within a species, between different stages for the same growth zone, and between homologous growth zones in different species. Differences in cell proliferation and/or cell enlargement underlie much of this differential growth, albeit in different proportions. Strikingly, differences in extracellular matrix production do not correlate with growth rate differences.

CONCLUSIONS

Differential endochondral growth drives many aspects of craniofacial morphological diversity in cichlids. Cellular proliferation and enlargement, but not extracellular matrix deposition, underlie this differential growth and this appears conserved in Osteichthyes. Cell enlargement is observed in some but not all cichlid growth zones and the degree to which it occurs resembles slower growing mammalian growth plates.

Secondary ossification center induces and protects growth plate structure

Growth plate and articular cartilage constitute a single anatomical entity early in development but later separate into two distinct structures by the secondary ossification center (SOC). The reason for such separation remains unknown. We found that evolutionarily SOC appears in animals conquering the land - amniotes. Analysis of the ossification pattern in mammals with specialized extremities (whales, bats, jerboa) revealed that SOC development correlates with the extent of mechanical loads. Mathematical modeling revealed that SOC reduces mechanical stress within the growth plate. Functional experiments revealed the high vulnerability of hypertrophic chondrocytes to mechanical stress and showed that SOC protects these cells from apoptosis caused by extensive loading. Atomic force microscopy showed that hypertrophic chondrocytes are the least mechanically stiff cells within the growth plate. Altogether, these findings suggest that SOC has evolved to protect the hypertrophic chondrocytes from the high mechanical stress encountered in the terrestrial environment.

Developmental and Evolutionary Allometry of the Mammalian Limb Skeleton

The variety of limb skeletal proportions enables a remarkable diversity of behaviors that include powered flight in bats and flipper-propelled swimming in whales using extremes of a range of homologous limb architectures. Even within human limbs, bone lengths span more than an order of magnitude from the short finger and toe bones to the long arm and leg bones. Yet all of this diversity arises from embryonic skeletal elements that are each a very similar size at formation. In this review article, I survey what is and is not yet known of the development and evolution of skeletal proportion at multiple hierarchical levels of biological organization. These include the cellular parameters of skeletal elongation in the cartilage growth plate, genes associated with differential growth, and putative gene regulatory mechanisms that would allow both covariant and independent evolution of the forelimbs and hindlimbs and of individual limb segments. Although the genetic mechanisms that shape skeletal proportion are still largely unknown, and most of what is known is limited to mammals, it is becoming increasingly apparent that the diversity of bone lengths is an emergent property of a complex system that controls elongation of individual skeletal elements using a genetic toolkit shared by all.

A journey through growth plates: tracking differences in morphology and regulation between the spine and the long bones in a pig model

BACKGROUND CONTEXT

The process of linear growth is driven by axial elongation of both long bones and vertebral bodies and is accomplished by enchondral ossification. Differences in regulation between the two skeletal sites are mirrored clinically by the age course in body proportions. Whereas long bone growth plates (GPs) can easily be discriminated, vertebral GPs are part of the cartilaginous end plate, which typically shows important species differences.

PURPOSE

The objective of this study was to describe and compare histologic, histomorphometric, and regulatory characteristics in the GPs of the spine and the long bones in a porcine model.

MATERIALS AND METHODS

Two- and six-week-old piglet GPs of three vertebral segments (cervical, thoracic, and lumbar) and eight long bones (proximal and distal radius, humerus, tibia, and femur) were analyzed morphometrically. Further, estrogen receptors, proliferation markers, and growth factor expressions were examined by immunohistochemistry.

RESULTS

Individual vertebral GPs were smaller in width and contained fewer chondrocytes than long bone GPs, although their proliferation activity was similar. Whereas the expression pattern of growth hormone-associated factors such as insulin-like growth factor (IGF)-1 and IGF-1 receptor (IGF-1R) was similar, estrogen receptor (ER)-ß and IGF-2 were distinctly expressed in the vertebral samples.

CONCLUSIONS

Vertebral GPs display differential growth, with measurements similar to the slowest-growing GPs of long bones. Further investigation is needed to decipher the molecular basis of the differential growth of the spine and the long bones. Knowledge on the distinct mechanism will ultimately improve the assessment of clinically essential characteristics of spinal growth, such as vertebral elongation potential and GP fusion.

Possible effects of an early diagnosis and treatment in patients with growth hormone deficiency: the state of art

Growth hormone deficiency (GHD) is a relatively uncommon and heterogeneous endocrine disorder presenting in childhood with short stature. However, during the neonatal period, the metabolic effects of GHD may to require prompt replacement therapy to avoid possible life-threatening complications. An increasing amount of data suggests the importance of an early diagnosis and treatment of GHD because of its auxological, metabolic, and neurodevelopmental features with respect to the patients diagnosed and treated later in life.

The available results show favourable auxological outcomes for patients with GHD diagnosed and treated with r-hGH early in life compared with those from patients with GHD who do not receive this early diagnosis and treatment. Because delayed referral for GHD diagnosis and treatment is still frequent, these results highlight the need for more attention in the diagnosis and treatment of GHD.

Despite these very encouraging data regarding metabolic and neurodevelopmental features, further studies are needed to better characterize these findings. Overall, the importance of early diagnosis and treatment of GHD needs to be addressed.

Cellular scale model of growth plate: An in silico model of chondrocyte hypertrophy

The growth plate is the responsible for longitudinal bone growth. It is a cartilaginous structure formed by chondrocytes that are continuously undergoing a differentiation process that starts with a highly proliferative state, followed by cellular hypertrophy, and finally tissue ossification. Within the growth plate chondrocytes display a characteristic columnar organization that potentiates longitudinal growth. Both chondrocyte organization and hypertrophy are highly regulated processes influenced by biochemical and mechanical stimuli. These processes have been studied mainly using in vivo models, although there are few computational approaches focused on the rate of ossification rather than events at cellular level. Here, we developed a model of cellular behavior integrating biochemical and structural factors in a single column of cells in the growth plate. In our model proliferation and hypertrophy were controlled by biochemical regulatory loop formed between Ihh and PTHrP (modeled as a set of reaction-diffusion equations), while cell growth was controlled by mechanical loading. We also examined the effects of static loading. The model reproduced the proliferation and hypertrophy of chondrocytes in organized columns. This model constitutes a first step towards the development of mechanobiological models that can be used to study biochemical interactions during endochondral ossification.

In vivo dynamic loading reduces bone growth without histomorphometric changes of the growth plate

This in vivo study aimed at investigating the effects of dynamic compression on the growth plate. Rats (28 days old) were divided into three dynamically loaded groups, compared with two groups (control, sham). A device was implanted on the 6th and 8th caudal vertebrae for 15 days. Controls (n = 4) did not undergo surgery. Shams (n = 4) were operated but not loaded. Dynamic groups had sinusoidal compression with a mean value of 0.2 MPa: 1.0 Hz and ± 0.06 MPa (group a, n = 4); 0.1 Hz and ± 0.2 MPa (group b, n = 4); 1.0 Hz and ± 0.14 MPa (group c, n = 3). Growth rates (µm/day) of dynamic groups (a) and (b) were lower than shams (p < 0.01). Growth plate heights, hypertrophic cell heights and proliferative cell counts per column did not change in dynamic (a) and (b) groups compared with shams (p > 0.01). Rats from dynamic group (c) had repeated inflammations damaging tissues; consequently, their analysis was unachievable. Increasing magnitude or frequency leads to growth reduction without histomorphometric changes. However, the combined augmentation of magnitude and frequency alter drastically growth plate integrity. Appropriate loading parameters could be leveraged for developing novel growth modulation implants to treat skeletal deformities.

The skeleton: a multi-functional complex organ: the growth plate chondrocyte and endochondral ossification

Endochondral ossification is the process that results in both the replacement of the embryonic cartilaginous skeleton during organogenesis and the growth of long bones until adult height is achieved. Chondrocytes play a central role in this process, contributing to longitudinal growth through a combination of proliferation, extracellular matrix (ECM) secretion and hypertrophy. Terminally differentiated hypertrophic chondrocytes then die, allowing the invasion of a mixture of cells that collectively replace the cartilage tissue with bone tissue. The behaviour of growth plate chondrocytes is tightly regulated at all stages of endochondral ossification by a complex network of interactions between circulating hormones (including GH and thyroid hormone), locally produced growth factors (including Indian hedgehog, WNTs, bone morphogenetic proteins and fibroblast growth factors) and the components of the ECM secreted by the chondrocytes (including collagens, proteoglycans, thrombospondins and matrilins). In turn, chondrocytes secrete factors that regulate the behaviour of the invading bone cells, including vascular endothelial growth factor and receptor activator of NFκB ligand. This review discusses how the growth plate chondrocyte contributes to endochondral ossification, with some emphasis on recent advances.

Three-dimensional in situ zonal morphology of viable growth plate chondrocytes: a confocal microscopy study

Longitudinal growth, occurring in growth plates with structurally distinct zones, has clinical implications in the treatment of progressive skeletal deformities. This study documents the three-dimensional morphology of chondrocytes within histological zones of growth plate using confocal microscopy combined with fluorescent labeling techniques. Three-dimensional reconstruction of Calcein AM-labeled chondrocytes was made from stacks of confocal images recorded in situ from 4-week-old swine growth plates. Three-dimensional quantitative morphological measurements were further performed and compared at both tissue and cell levels. Chondrocyte volume and surface area increased about five- and threefold, respectively, approaching the chondro-osseous junction from the pool of reserve cells. Chondrocytes from the proliferative zone were the most discoidal cells (sphericity of 0.81 ± 0.06) among three histological zones. Minimum and maximum cell/matrix volume ratios were identified in the reserve (11.0 ± 2.2) and proliferative zones (16.8 ± 3.0), respectively. Evaluated parameters revealed the heterogeneous and zone-dependent morphological state of the growth plate. Tissue and cellular morphology may have noteworthy contribution to the growth plate behavior during growth process. The ability to obtain in situ cell morphometry and monitor the changes in the growth direction could improve our understanding of the mechanisms through which abnormal growth is triggered.

Calcium/calmodulin-dependent protein kinase II activity regulates the proliferative potential of growth plate chondrocytes

For tissues that develop throughout embryogenesis and into postnatal life, the generation of differentiated cells to promote tissue growth is at odds with the requirement to maintain the stem cell/progenitor cell population to preserve future growth potential. In the growth plate cartilage, this balance is achieved in part by establishing a proliferative phase that amplifies the number of progenitor cells prior to terminal differentiation into hypertrophic chondrocytes. Here, we show that endogenous calcium/calmodulin-dependent protein kinase II (CamkII, also known as Camk2) activity is upregulated prior to hypertrophy and that loss of CamkII function substantially blocks the transition from proliferation to hypertrophy. Wnt signaling and Pthrp-induced phosphatase activity negatively regulate CamkII activity. Release of this repression results in activation of multiple effector pathways, including Runx2- and β-catenin-dependent pathways. We present an integrated model for the regulation of proliferation potential by CamkII activity that has important implications for studies of growth control and adult progenitor/stem cell populations.

New insights into function of the growth plate

The mammalian growth plate is a complex structure which is essential for the elongation of long bones. However, an understanding of how the growth plate functions at the cellular level is lacking. This review, summarises the factors involved in growth-plate regulation, its failure and the consequence of injury. We also describe some of the cellular mechanisms which underpin the increase in volume of the growth-plate chondrocyte which is the major determinant of the rate and extent of bone lengthening. We show how living in situ chondrocytes can be imaged using 2-photon laser scanning microscopy to provide a quantitative analysis of their volume. This approach should give better understanding of the cellular control of bone growth in both healthy and failed growth plates.

The osmotic sensitivity of rat growth plate chondrocytes in situ; clarifying the mechanisms of hypertrophy

Bone elongation is predominantly driven by the volume expansion of growth plate chondrocytes. This mechanism was initially believed to be "hypertrophy", describing a proportional increase of cell water and organelles. However, morphometrical analysis subsequently assumed the increase to be "swelling", resulting in a disproportionate increase of cell water (osmotically active fraction). Histological approaches were performed on fixed tissue, and for the "swelling" assumption to be valid, the osmotic sensitivity of living cells before and during volume increase should differ. To test this, analysis of images acquired by 2-photon laser scanning microscopy (2PLSM) were used to determine the osmotic sensitivity, and osmotically active/inactive proportions of in situ chondrocytes from 15 living rat growth plates exposed to varying media osmolarities ( approximately 0-580 mOsm). The dimensions of cell volume swelling in hypotonic media were different to the preferential lengthening seen in vivo, confirming the complexity of directional cell volume increase. Boyle-van't Hoff analysis of cell volume over the range of media osmolarity indicated no significant difference (Student's t-test) in the osmotically inactive fraction, 39.5 +/- 2.9% and 47.0 +/- 4.3% (n = 13) for proliferative and hypertrophic zones, respectively, or the sensitivity of volume to changes in media osmolarity (proliferative 15.5 +/- 0.8 and hypertrophic zone 15.5 +/- 1.2%volume . Osm). The osmotic fractions did not change as chondrocytes progress from proliferative to hypertrophic regions of the growth plate. Our data suggest cell volume increase by hypertrophy may play a greater role in cell enlargement than swelling, and should be re-evaluated as a mechanism responsible for growth plate chondrocyte volume increase and hence bone elongation.

Postnatal bone elongation of the manus versus pes: analysis of the chondrocytic differentiation cascade in Mus musculus and Eptesicus fuscus

Bones elongate postnatally by endochondral ossification as cells of the cartilaginous growth plate undergo a differentiation cascade of proliferation, cellular hypertrophy and matrix synthesis. Interspecific comparisons of homologous bones elongating at different rates has been a useful approach for studying the dynamics of this process. The purpose of this study was to measure quantitative stereological parameters of growth plates of the third digit of the manus and pes of the laboratory mouse, and make comparisons to chondrocytic performance parameters in the homologous bones of the big brown bat, Eptesicus fuscus, where extremely rapid postnatal elongation of bones of the manus is associated with skeletal modifications for powered flight. Measurements were made across all zones of forelimb and hindlimb autopod growth plates by dividing each growth plate into strata of equal height (from thirteen 200-mum-high strata in the metacarpus to five 40-mum-high strata in phalangeal bones of the pes). Results indicate that all chondrocytic performance parameters known to quantitatively contribute to the elongation potential of a growth plate change together. A significant finding was that in growth plates of the chiropteran manus, final hypertrophic cell size and shape were achieved early in the zone of hypertrophy, indicating that interstitial expansion of the growth plate resulting from the incremental chondrocytic height increase in the direction of elongation was completed soon after the transition from the cessation of proliferation to the initiation of hypertrophy. This is unlike what has been reported in most mammalian growth plates previously analyzed, but is the situation in the proximal tibial growth plate of rapidly growing frogs and precocial birds. This suggests that a similar adaptation for stabilization of a rapidly elongating bone has evolved independently in three widely separated groups that have in common rapid growth in limbs to be used for early active, powered locomotion.

Continuous Activation of Gαq in Osteoblasts Results in Osteopenia through Impaired Osteoblast Differentiation

We explored the role of G alpha(q)-mediated signaling on skeletal homeostasis by selectively expressing a constitutively active G alpha(q) (mutation of Q209L) in osteoblasts. Continuous signaling via G alpha(q) in mouse osteoblastic MC3T3-E1 cells impaired differentiation. Mice that expressed the constitutively active G alpha(q) transgene in cells of the osteoblast lineage exhibited severe osteopenia in cortical and trabecular bones. Osteoblast number, bone volume, and trabecular thickness were reduced in transgenic mice, but the osteoclasts were unaffected. Osteoblasts from transgenic mice showed impaired differentiation and matrix formation. In the presence of a protein kinase C inhibitor GF109203X, this impairment was not seen, indicating mediation by the protein kinase C pathway. We propose that continuous activation of the G alpha(q) signal in osteoblasts plays a crucial, previously unrecognized role in bone formation.

Finite element modeling of the growth plate in a detailed spine model

Very few computer models of the spine integrate vertebral growth plates to investigate their mechanical behavior and potential impacts on bone growth. An approach was developed to generate a finite element (FE) model of the lumbar spine and their connective tissues including the growth plate, which allowed a personalization of the geometry based on patients' bi-planar radiographs. The geometrical validation was performed by deforming meshed vertebrae to reference vertebral specimens and comparing geometrical indices. No significant difference was found between the measured parameters, with errors under 1% in 83% of the geometrical parameters. Mechanical validation was done by simulating loading cases on a functional unit representing experimental testing on cadaveric spines. The flexibility of the functional unit remained between expected ranges of motion, but was more linear than experimental data. The mechanical behavior of the growth plate was evaluated under various loading cases: greater stresses were located in the proliferative zone for the different spinal loading cases tested. This modeling approach is a useful tool to study the effect of mechanical stresses on bone growth.

Effects of TGF-beta1 and triiodothyronine on cartilage maturation: in vitro analysis using long-term high-density micromass cultures of chick embryonic limb mesenchymal cells

Endochondral ossification is initiated by differentiation of mesenchymal cells into chondrocytes, which produce a cartilaginous matrix, proliferate, mature, and undergo hypertrophy, followed by matrix calcification, and substitution of cartilage by bone. A number of hormones and growth factors have been implicated in this process. Using in vitro, long-term, high-density, micromass cultures of chick embryonic mesenchyme, that recapitulate the process of chondrogenesis, chondrocyte maturation, and hypertrophy, we have investigated the importance of a balance between proliferation and apoptosis in cartilage maturation, focusing specifically on the effects of transforming growth factor-beta1 (TGF-beta1) and the thyroid hormone, triiodothyronine (T3). Our results showed that TGF-beta1 stimulates proliferation, by week 2 of culture, and T3 inhibits proliferation by week 3. Cell size increases in cultures treated with T3. Collagen type X is expressed in all culture, and delay in matrix deposition is seen only in the cultures treated with TGF-beta1. T3 stimulates alkaline phosphatase activity, but not calcification. T3 enhances apoptosis, as seen by TUNEL staining, and internucleosomal DNA fragmentation. The results support the roles of T3 and TGF-beta in cartilage maturation, i.e., TGF-beta stimulates proliferation and suppresses hypertrophy, while T3 stimulates hypertrophy and apoptosis.

In vitro chondrocyte differentiation using costochondral chondrocytes as a source of primary rat chondrocyte cultures: an improved isolation and cryopreservation method

INTRODUCTION

Isolating and culturing primary chondrocytes such that they retain their cell type and differentiate to a hypertrophic state is central to many investigations of skeletal growth and its regulation. The ability to store frozen chondrocytes has additional scientific and tissue engineering interest. Previous work has produced approaches of varying yield and complexity but does not permit frozen storage of cells for subsequent differentiation in culture. Investigations of growth plate dysplasias secondary to defective osteoclastogenesis in rodent models of osteopetrosis led us to adapt and modify a culture method and to cryopreserve neonatal rat costochondral chondrocytes.

METHODS

Chondrocytes were isolated from dissected ribs of 3-day-old rat pups by collagenase, hyaluronidase, and trypsin serial digestions. This was done either immediately or after the isolation was interrupted following an initial protease treatment to allow the chondrocytes, still in partially digested rib rudiments, to be frozen and later thawed for culture. Cells were plated in flat-bottom wells and allowed to adhere and grow under different conditions. Choice of media permitted cells to be maintained or induced to differentiate. Cell growth was monitored, as was expression of several relevant genes: collagen types II and X; osteocalcin, Sox9, adipocyte FABP, MyoD, aggrecan, and others. Mineralization was measured by alizarin red binding, and cultures were examined by light, fluorescence, and electron microscopy.

RESULTS

Cells retained their chondrocyte phenotype and ability to differentiate and mineralize the collagen-rich extracellular matrix even after freezing-thawing. RT-PCR showed retention of chondrocyte-specific gene expression, including aggrecan and collagen II. The cells had a flattened, "proliferating zone" appearance initially, and by 2 weeks post-confluence, exhibited swelling and other salient features of hypertrophic cells seen in vivo. Collagen fibrils were abundant in the extracellular matrix, along with matrix vesicles. The switch to collagen type X as marker for hypertrophy was not rigidly temporally regulated as happens in vivo, but its expression increased during hypertrophic differentiation.

CONCLUSIONS

This method should prove valuable as a means of studying chondrocyte regulation and has the advantages of simpler initial dissection, yields of a purer chondrocyte population, and the ability to stockpile frozen raw material for subsequent studies.

Transient disturbance in physeal morphology is associated with long-term effects of nitrogen-containing bisphosphonates in growing rabbits

Bisphosphonates have clinical benefit in children with severe osteogenesis imperfecta or osteoporosis and potential benefit in children with Perthes disease or undergoing distraction osteogenesis. However, there is concern about the effects of bisphosphonates on the physis and bone length. In 44 growing rabbits, zoledronic acid caused a transient disruption of physeal morphology, retention of cartilaginous matrix in trabeculae and cortical bone of the metaphysis, and a minor decrement in tibial bone length at maturity.

INTRODUCTION

Data from growing animal models suggest that bisphosphonates cause retention of longitudinal cartilaginous septa at the chondro-osseous junction, extension of trabeculae to the metaphyseal-diaphyseal junction, and varying dose-dependent effects on longitudinal growth. However, there is a lack of data regarding effects of intermittent use of nitrogen-containing bisphosphonates on the physis and on tibial length in models reaching maturity.

MATERIALS AND METHODS

Contralateral tibias of juvenile rabbits were examined after right tibial distraction osteogenesis from two previous studies. Animals were randomized to receive 0.1 mg/kg zoledronic acid (ZA) IV at 8 weeks of age (ZA*1) or 8 and 10 weeks of age (ZA*2) or saline. Body mass was analyzed from 5 to 44 weeks of age; tibial length and proximal physeal-metaphyseal histology and histomorphometry were analyzed at 8-52 weeks of age.

RESULTS

Tibial length was 3% less at 14 weeks of age in the ZA*2-treated versus saline group (p<0.05) in both studies, and this difference persisted at maturity in the long-term study group (26 weeks of age, p<0.05). Total body mass gain from 5 to 26 weeks of age was 14% less in ZA*2-treated than saline animals (p<0.05). Rate of weight gain from 8 to 10 weeks of age was 76% less in ZA*2 compared with saline animals (p<0.05). Radiographs showed radiodense lines in the metaphyses of ZA-treated bones, corresponding to the number of doses. Histologically, lines resulting from the first dose of ZA contained longitudinal cartilaginous matrix cores surrounded by bone, whereas those from the second dose contained spherical cores of matrix caused by transient disruption of physeal morphology after the first dose of ZA. Resorption of these lines at later times was radiographically and histologically evident, but remnants of cartilaginous matrix remained in the cortical bone of ZA-treated animals.

CONCLUSIONS

ZA treatment within the final 13.5% of the rabbit tibial growth period caused a transient disruption in physeal morphology and resorption associated with retention of cartilaginous matrix and coinciding with a persistent 3% decrement in tibial length. Disruption of physeal morphology and potential loss of bone length should be considered when administering nitrogen-containing bisphosphonates to children before closure of the major physes.

The pathophysiology of the growth plate in juvenile idiopathic arthritis

Children with chronic inflammatory diseases, such as juvenile idiopathic arthritis (JIA), suffer from a variety of growth disorders. These range from general growth retardation to local acceleration of growth in the affected limb. These disorders are associated with the increased production of proinflammatory cytokines, which may influence growth through a local effect in the growth plates of long bones and/or systemic effects throughout the whole body. In this article we review these aspects and also discuss the evidence for interaction between the inflammatory cytokine and growth-signalling pathways.

A new histomorphometric method to assess growth plate chondrodysplasia and its application to the toothless (tl, Csf1(null)) osteopetrotic rat

The proliferation and hypertrophy of growth plate chondrocytes set the pace and pattern for growth of endochondral bones. Complex signaling pathways regulating chondrocyte differentiation during development and growth have been discovered in recent years, but as yet little is known about how chondrocytes are able to orient themselves to align properly with respect to the direction of bone growth. Histomorphometric methods developed for analysis of growth plates rely to a significant extent on assessments of the relative heights of the zones of proliferating and hypertrophic chondrocytes. In a growing number of osteopetrotic mutations, however, it is becoming apparent that growth plates lack clearly demarcated zones of chondrocyte differentiation, and they show other notable histological abnormalities that cannot be measured with standard approaches. This is particularly true of mutations in which osteoclasts are altogether absent. We therefore developed a new approach that measures the salient features of this type of chondrodysplasia and have applied it to the toothless (tl) rat. The tl rat has a frameshift mutation in the Csf-1 gene that renders it null, resulting in severe osteopetrosis. An accompanying pathology is a severe, progressive growth plate chondrodysplasia. We measured cell orientation, cell area, and local columnar organization as functions of distance from the upper margin of the growth plate, in addition to growth plate thickness and cell density. All these parameters were markedly abnormal in the tl rats, thus implicating Csf-1 not only in its well-established role in regulating osteoclastic bone resorption, but also in endochondral ossification. This approach should prove useful in distinguishing among growth plate chondrodysplasias, most especially in the growing number of osteopetrotic mutations having growth plates that lack the normal zonal organization and in which the chondrocytes are mis-oriented. In turn, detailed assessments of chondrocyte misorientation may give insights into how they normally are able to arrange themselves with such precision.

Histologic and dynamic changes induced by chronic metabolic acidosis in the rat growth plate

To understand better the pathophysiology of growth impairment in persistent metabolic acidosis, the morphology and dynamics of the growth plate were studied in young rats grouped as follows: rats that were made acidotic by oral administration of ammonium chloride for 14 d (AC), nonacidotic rats that were fed ad libitum (control [C]), and nonacidotic rats that were pair-fed with the AC group (PF). AC rats became markedly acidotic and growth retarded. The volume of newly formed bone per day (mean +/- SEM) was significantly lowered (P < 0.05) in AC rats (AC, 3.4 +/- 0.4; C, 8.4 +/- 0.6; PF, 6.4 +/- 0.5 mm(3)/d). Growth plate height was lower in AC rats (303.8 +/- 12.7 microm) than in either C (478.0 +/- 16.0 microm) or PF rats (439.0 +/- 21.4 microm). The processes of chondrocyte proliferation (assessed by bromodeoxyuridine labeling) and maturation (assessed by stereologic estimators of size and shape of chondrocytes and the volume of matrix per cell) were not impaired by acidosis. By contrast, the dynamics of hypertrophic chondrocytes were altered significantly: both cell turnover per column per day (AC, 4.4 +/- 0.4; C, 8.0 +/- 0.8; PF, 6.2 +/- 0.6) and linear velocity of advance of chondrocytes (AC, 5.7 +/- 0.5; C, 11.2 +/- 0.9; PF, 9.4 +/- 0.8 microm/h) were lowered significantly. The study presented here shows the inhibitory effect of metabolic acidosis on cartilage cell progression and endochondral bone formation. Finally, the data show that metabolic acidosis caused a marked shortening of the growth plate because chondrocyte turnover was affected to a greater extent than bone tissue formation.

Histomorphological and proliferative characterization of developing periosteal neochondrocytes in vitro

Periosteal chondrogenesis is relevant to cartilage repair and fracture healing. Cell proliferation is a limiting factor of cartilage production. We used an in vitro organ culture model to test the hypothesis that proliferative activity correlates with cell morphology. One hundred and four periosteal explants from 26 two-month old New Zealand rabbits were cultured for up to 42 days. They were analyzed histomorphologically, and immunohistochemically with proliferative cell nuclear antigen (PCNA). The periosteal neocartilage displayed a consistent zonal pattern of chondrocyte cell shapes. The flat cell zone from day 7 to 21, consisted of uniform-sized small spindle-shaped cells. The round cell zone, which appeared on day 14, consisted of variable-sized round cells averaging 510 +/- 250 microm2 in area. They were subdivided into small round (<510 microm2) and large round cells (>510 microm2). The proliferative index was highest in the small round cell group (32 +/- 6%), intermediate in the flat cell group (27 +/- 6%), and lowest in the large round cell group (20 +/- 7%) (P < 0.001). Furthermore, the proliferative indices in the round cell group were inversely proportional to cell size. Therefore, (1) there is a sequential progression of cell morphology during periosteal chondrogenesis, (2) cell differentiation is arrested prior to terminal differentiation for some cells and not for others, and (3) proliferative activity is strongly related to cell morphology. This organ culture model provides us with opportunities to study the regulation of terminal chondrocyte differentiation and the control of cell proliferation. This will contribute to our understanding of cartilage repair, fracture healing and growth plate physiology.

Collagen metabolism is markedly altered in the hypertrophic cartilage of growth plates from rats with growth impairment secondary to chronic renal failure

Skeletal growth depends on growth plate cartilage activity, in which matrix synthesis by chondrocytes is one of the major processes contributing to the final length of a bone. On this basis, the present work was undertaken to ascertain if growth impairment secondary to chronic renal insufficiency is associated with disturbances of the extracellular matrix (ECM) of the growth plate. By combining stereological and in situ hybridization techniques, we examined the expression patterns of types II and X collagens and collagenase-3 in tibial growth plates of rats made uremic by subtotal nephrectomy (NX) in comparison with those of sham-operated rats fed ad libitum (SAL) and sham-operated rats pair-fed with NX (SPF). NX rats were severely uremic, as shown by markedly elevated serum concentrations of urea nitrogen, and growth retarded, as shown by significantly decreased longitudinal bone growth rates. NX rats showed disturbances in the normal pattern of chondrocyte differentiation and in the rates and degree of substitution of hypertrophic cartilage with bone, which resulted in accumulation of cartilage at the hypertrophic zone. These changes were associated with an overall decrease in the expression of types II and X collagens, which was especially marked in the abnormally extended zone of the hypertrophic cartilage. Unlike collagen, the expression of collagenase-3 was not disturbed severely. Electron microscopic analysis proved that changes in gene expression were coupled to alterations in the mineralization as well as in the collagen fibril architecture at the hypertrophic cartilage. Because the composition and structure of the ECM have a critical role in regulating the behavior of the growth plate chondrocytes, results obtained are consistent with the hypothesis that alteration of collagen metabolism in these cells could be a key process underlying growth retardation in uremia.

Chondrocytes and longitudinal bone growth: the development of tibial dyschondroplasia

Growth plate cartilage is central to the process of bone elongation. Chondrocytes originating within the resting zone of the growth plate proceed through a series of intermediate phenotypes: proliferating, prehypertrophic and hypertrophic, before reaching a terminally differentiated state. Disruption of this chondrocyte maturational sequence causes many skeletal abnormalities in poultry such as tibial dyschondroplasia (TD), which is a common cause of deformity and lameness in the broiler chicken. Cell and matrix components of the growth plate have been studied in order to determine the cause(s) of the premature arrest of chondrocyte differentiation and retention of prehypertrophic chondrocytes observed in TD. Chondrocyte proliferation proceeds normally in TD, but markers of the differentiated phenotype, local growth factors, and the vitamin D receptor are abnormally expressed within the prehypertrophic chondrocytes above, and within, the lesion. Tibial dyschondroplasia is also associated with a reduced incidence of apoptosis, suggesting that the lesion contains an accumulation of immature cells that have outlived their normal life span. Immunolocalization studies of matrix components suggest an abnormal distribution within the TD growth plate that is consistent with a failure of the chondrocytes to fully hypertrophy. In addition, the collagen matrix of the TD lesion is highly crosslinked, which may make the formed lesion more impervious to vascular invasion and osteoclastic resorption. Recent studies have applied the techniques of differential display and semiquantitative reverse transcriptase-polymerase chain reaction to RNA obtained from discrete populations of growth plate chondrocytes of different maturational phenotypes. This strategy has allowed us to compare phenotypically identical cell fractions from normal and TD growth plates in an attempt to identify possible candidate genes for TD.

Different bone growth rates are associated with changes in the expression pattern of types II and X collagens and collagenase 3 in proximal growth plates of the rat tibia

Skeletal growth depends on endochondral ossification in growth plate cartilage, where proliferation of chondrocytes, matrix synthesis, and increases in chondrocyte size all contribute to the final length of a bone. To learn more about the potential role of matrix synthesis/degradation dynamics in the determination of bone growth rate, we investigated the expression of matrix collagens and collagenase 3 in tibial growth plates in three age groups of rats (21, 35, and 80 days after birth), each characterized by specific growth rates. By combining stereological and in situ hybridization techniques, it was found that the expression of matrix collagens and collagenase 3 was specifically turned on or off at specific stages of the chondrocyte-differentiation cycle, and these changes occurred as a temporal sequence that varied depending of animal growth rate. Furthermore, the expression of these matrix proteins by a growth plate chondrocyte was found to be sped up or slowed down depending of the growth rate. In addition to expression of types II and X collagen, collagenase-3 expression was found to constitute a constant event in the series of changes in gene expression that takes place during the chondrocyte-differentiation process. Collagenase-3 expression was found to show a biphasic pattern: it was intermittently expressed at the proliferative phase and uniformly expressed at the hypertrophic stage. An intimate relationship between morphological and kinetic changes associated with chondrocyte hypertrophy and changes in the expression pattern of matrix collagens and collagenase 3 was observed. Present data prove that the matrix synthesis/degradation dynamics of the growth plate cartilage varied depending on growth rate; these results support the hypothesis that changes in matrix degradation and synthesis are a critical link in the sequence of tightly regulated events that lead to chondrocytic differentiation.

Skeletal dysplasia and defective chondrocyte differentiation by targeted overexpression of fibroblast growth factor 9 in transgenic mice

Mutations in fibroblast growth factor receptor 3 (FGFR3) cause several human chondrodysplasias, including achondroplasia, the most common form of dwarfism in humans. From in vitro studies, the skeletal defects observed in these disorders have been attributed to constitutive activation of FGFR3. Here we show that FGF9 and FGFR3, a high-affinity receptor for this ligand, have similar developmental expression patterns, particularly in areas of active chondrogenesis. Targeted overexpression of FGF9 to cartilage of transgenic mice disturbs postnatal skeletal development and linear bone growth. The growth plate of these mice exhibits reduced proliferation and terminal differentiation of chondrocytes similar to that observed in the human disorders. The observations provide evidence that targeted, in vivo activation of endogenous FGFR3 inhibits bone growth and demonstrate that signals derived from FGF9-FGFR3 interactions can physiologically block endochondral ossification to produce a phenotype characteristic of the achondroplasia group of human chondrodysplasias.

Microtubules are potential regulators of growth-plate chondrocyte differentiation and hypertrophy

Terminal differentiation of growth-plate chondrocytes is accompanied by the acquisition of a spherical morphology and a large increase in cell volume. These changes are likely to be associated with rearrangement of the cytoskeleton, but little information on this aspect of chondrocyte hypertrophy is available. We report a role for microtubules in the control of chondrocyte maturation and hypertrophy. Chick growth-plate chondrocytes were fractionated into five maturationally distinct populations by Percoll density gradient centrifugation, and agarose gel differential display analysis was performed. We identified a 1200 bp cDNA fragment derived from a transcript that was most highly expressed in the hypertrophic chondrocytes. After cloning and sequencing, FASTA and BLAST analysis revealed 100% identity to chick beta7-tubulin. Differential expression was confirmed in a reverse transcription-polymerase chain reaction (RT-PCR) assay using specific primers for a 343 bp fragment from the 3' untranslated region of beta7-tubulin. Beta7-tubulin was upregulated three-fold in fully hypertrophic chondrocytes compared with the other four fractions, which all had similar levels of expression. Immunocytochemical localization of beta-tubulin in chick growth-plate sections demonstrated little staining in the chondrocytes of the proliferating zone, but intense cytoplasmic staining was present in the large hypertrophic chondrocytes. In cell culture studies, the addition of colchicine (10(-6) mol/L) resulted in a higher rate of [3H]-thymidine uptake (36.0%; p < 0.001), but lower amounts of alkaline phosphatase activity (69.1%; p < 0.001), collagen (49.1%; p < 0.01), and glycosaminoglycan (43.3%; p < 0.01) accumulation within the cell-matrix layer. Further evidence for the involvement of microtubules in chondrocyte differentiation and hypertrophy was obtained by morphological assessment of colchicine-treated growth-plate explant cultures. A partial failure of chondrocyte hypertrophy was observed, although collagen type X immunoreactivity was noted within the interstitial matrix. Further studies are required to identify the exact role of microtubules in chondrocyte hypertrophy, but the results presented here suggest that upregulation of beta-tubulin may be required for increased microtubule synthesis during changes in cell size during the hypertrophic process. In addition, as cell-matrix interactions are required for chondrocyte maturation, microtubules may promote the differentiated phenotype as a result of their role in Golgi-mediated secretion of matrix proteins.

Growth plate cartilage formation and resorption are differentially depressed in growth retarded uremic rats

To characterize the modifications of growth plate in individuals with growth impairment secondary to chronic renal failure, young rats were made uremic by subtotal nephrectomy (NX) and, after 14 d, their tibial growth plates were studied and compared with those of sham-operated rats fed ad libitum (SAL) or pair-fed with NX (SPF). NX rats were growth retarded and severely uremic. Growth plate height (mean +/- SD) was much greater (P<0.05) in NX (868.4+/-85.4 microm) than SAL (570.1+/-93.5 microm) and SPF (551.9+/-99.7 microm) rats as a result of a higher (P<0.05) hypertrophic zone (661.0+/-89.7 versus 362.8+/-71.6 and 353.0+/-93.9 microm, respectively). The increased size of the growth plate was associated with a greater number of chondrocytes and modifications in their structure, particularly in the hypertrophic zone adjacent to bone. In this zone, chondrocytes of NX animals were significantly (P<0.05) smaller (12080.4+/-1158.3 microm3) and shorter (34.1+/-2.5 microm) than those of SAL (16302.8+/-1483.4 microm3 and 37.8+/-2.0 microm) and SPF (14465.8+/-1521.0 microm3 and 36.3+/-1.8 microm). The interface between the growth plate cartilage and the metaphyseal bone appeared markedly irregular in NX rats. Kinetics of chondrocytes was also modified (P<0.05) in the NX rats, which had lower cell turnover per column per day (5.4+/-0.9), longer duration of hypertrophic phase (89.0+/-15.2 h), and reduced cellular advance velocity (7.4+/-2.2 microm/h) compared with SAL (8.0+/-1.6, 32.1+/-6.7 h, and 11.3+/-2.7 microm/h) and SPF (7.2+/-1.1, 34.8+/-5.1 h, and 10.1+/-2.5 microm/h). Cell proliferation was no different among the three groups. Because the growth plates of SPF and SAL rats were substantially not different, modifications observed in the NX rats cannot be attributed to the nutritional deficit associated with renal failure. These findings indicate that chronic renal failure depresses both the activity of the growth plate cartilage by altering chondrocyte hypertrophy and the replacement of cartilage by bone at the metaphyseal end. The two processes are differentially depressed since cartilage resorption is more severely lowered than cartilage enlargement and this leads to an accumulation of cartilage at the hypertrophic zone.

Changes in cell, matrix compartment, and fibrillar collagen volumes between growth-plate zones

To define the contributions of changes in cell, matrix compartment, and fibrillar collagen volumes to longitudinal bone growth, we measured the differences in cell, pericellular/territorial matrix and interterritorial matrix volumes, and fibrillar collagen concentrations between the upper proliferative and lower hypertrophic zones of the proximal tibial physes of six miniature pigs. The mean numerical density of cells decreased from 110,000 cells/mm3 in the upper proliferative zone to 59,900 cells/mm3 in the lower hypertrophic zone. The mean cell volume increased nearly 5-fold (from 1,174 to 5,530 microm3), and the total matrix volume per cell increased 46% (from 8,040 to 11,760 microm3/cell) between the upper proliferative and lower hypertrophic zones. Both the pericellular/territorial matrix volume per cell and the interterritorial matrix volume per cell increased between the upper proliferative and lower hypertrophic zones; the pericellular/territorial matrix volume per cell increased 61% (from 4,580 to 7,390 microm3/cell), whereas the interterritorial matrix volume per cell increased 26% (from 3,460 to 4,370 microm3/cell). The total increase in mean cell volume of 4,356 microm3 exceeded the total increase in mean matrix volume per cell of 3,720 microm3; the total mean pericellular/territorial matrix volume per cell increased more than the total mean interterritorial matrix volume per cell (2,810 compared with 910 microm3/cell). Fibrillar collagen concentration was greater in the interterritorial matrix than in the pericellular/territorial matrix in both zones and increased in both matrix compartments between the upper proliferative and lower hypertrophic zones. The amount of fibrillar collagen per cell also increased in both matrix compartments between the upper proliferative and lower hypertrophic zones (from 1,720 to 3,100 microm3/cell in the pericellular/territorial matrix and from 1,490 to 2,230 microm3/cell in the interterritorial matrix; thus, the total amount of fibrillar collagen per cell increased from 3,210 to 5,530 microm3/cell). Growth rate was inversely related to the cell numerical density in the upper proliferative and lower hypertrophic zones and was directly related to interterritorial matrix volume per cell in the upper proliferative zone and to pericellular/territorial matrix volume per cell in the lower hypertrophic zone. These results show that cell enlargement contributes more to longitudinal bone growth than does increased matrix volume, that increased pericellular/territorial matrix volume makes a greater contribution to growth than does increased interterritorial matrix volume, and that the total amount of fibrillar collagen per cell increases between the upper proliferative and lower hypertrophic zones. The differences between the two matrix compartments in increase in volume, fibrillar collagen concentration, and amount of fibrillar collagen per cell strongly suggest that they differ not only in matrix organization but in rate of matrix accumulation and assembly and that these differences give the two compartments different roles in skeletal growth.

Differential growth by growth plates as a function of multiple parameters of chondrocytic kinetics

Differential elongation of growth plates is the process by which growth-plate chondrocytes translate the same sequence of gene regulation into the appropriate timing pattern for a given rate of elongation. While some of the parameters associated with differential growth are known, the purpose of this study was to test the hypothesis that eight independent variables are involved. We tested this hypothesis by considering four different growth plates in 28-day-old Long-Evans rats. Temporal parameters were provided by means of oxytetracycline and bromodeoxyuridine labeling techniques. Stereological parameters were measured with standard techniques. For all four growth plates, the calculated number of new chondrocytes produced per day approximated the number of chondrocytes lost per day at the chondro-osseous junction. This suggests that the proposed equations and associated variables represent a comprehensive set of variables defining differential growth. In absolute numbers, the proximal tibial growth plate produced about four times as many chondrocytes per day as the proximal radial growth plate (16,400 compared with 3,700). In the proximal tibia, 9% of growth is contributed by cellular division; 32%, by matrix synthesis throughout the growth plate; and 59%, by chondrocytic enlargement during hypertrophy. In the more slowly elongating growth plates, the relative contribution to elongation from cellular enlargement decreases from 59 to 44%, with a relative increase in contribution from matrix synthesis ranging from 32% in the proximal tibia 49% in the proximal radius. This study suggests that differential growth is best depicted as a complex interplay among cellular division, matrix synthesis, and cellular enlargement during hypertrophy. Differential growth is best explained by considering a set of eight independent variables, seven of which vary from growth plate to growth plate. Thus, this study confirms the importance of cellular hypertrophy during elongation and adds to our understanding of the importance of locally mediated regulatory systems controlling growth-plate activity.

Cell cycle analysis of proliferative zone chondrocytes in growth plates elongating at different rates

Regulation of postnatal growth of long bones occurs in multiple levels of chondrocytic activity, including stem cell proliferation, proliferative zone cycling, and regulation of changes in chondrocytic shape during hypertrophy. The differentiation sequence of chondrocytes is the same in all growth plates, but rates of elongation at a single point in time and over a period of time differ widely among individual growth plates, which suggests that the rates of sequential gene activation and suppression in this phenotypic pattern can vary. The purpose of this study was to investigate, directly and in vivo, parameters of the cell cycle of proliferative chondrocytes in growth plates growing at widely different rates at a single point in time in order to analyze the relationship between cell cycle time, including the duration of each phase of the cell cycle (G1, S, G2, and M), and the rate of growth. The experimental design used repeated pulse labeling with bromodeoxyuridine and was analyzed using a regression model of time of pulse label with increasing labeling index. Total cell cycle time was calculated as the inverse of the slope of the relationship of the labeling index and the time between labels. The y intercept was the calculated labeling index at time zero. Multiple comparison contrasts were used to test for individual differences among four growth plates with growth rates ranging from approximately 50 to 400 microns per 24 hours from 28-day-old rats. The estimate of total cell cycle time for the proximal tibial growth plate was 30.9 hours. Cell cycle times for the other three growth plates were 34.0, 48.7, and 76.3 hours for the distal radius, distal tibia and proximal radius, respectively. Although the times for the proximal tibia and distal radius did not differ significantly, all other times were significantly different (p < 0.05). Almost all differences in total cell cycle time were attributable to significant differences in the length of the G1 phase. The S phase was estimated at 3.4-6.1 hours; the G2 phase, at 3.0 hours; and the M phase, at 0.5-0.6 hours. The current study suggests that regulation through cell cycle parameters, specifically in the G1 phase, may be involved in overall regulation of differential postnatal long bone growth. It has previously been established that increase and shape change of cellular volume during hypertrophy may be regulated at the level of individual growth plates and that both are significant in understanding differential growth of long bone at this level. By demonstrating that chondrocytes in the proliferating zone have different cell cycle times that are regulated primarily through differences in the duration of G1, this study suggests that, in addition to systemic controls of chondrocyte proliferation, local controls may modulate rates of proliferation of individual growth plates and thus may be another locally mediated regulator of differential growth.

Reduced physeal area and chondrocyte proliferation in Pasteurella multocida toxin-treated rats

Pasteurella multocida toxin depresses weight gain in rats and pigs. It also affects tissues with rapidly dividing cells. In the present study, we investigated the role of this protein toxin on chondrocyte growth in vivo. Rats were divided into a single- or multiple-dose group and were given, respectively, either a single injection (0.15 or 0.6 micrograms/kg toxin subcutaneously) or multiple injections (0.01-0.2 micrograms/kg subcutaneously) of toxin. Bone (humerus) and other selected tissues were stained for bromodeoxyuridine immunoreactivity (BrDU-IR) in order to gauge cell proliferation. Physeal area was measured in rats from the multiple-dose group. Serum from single- and multiple-dose groups were tested for tumor necrosis factor alpha (TNF-alpha) and interleukin 6 (IL-6) activity using a bioassay system. Decreased weight gain, feed intake, and feed efficiency were observed in single- and multiple-dose groups of rats. Decreased BrDU-IR indices were present in the resting and proliferative zone chondrocytes of the humeral physis in rats from the multiple-dose group, as was decreased physeal area. Increased serum IL-6 bioactivity was present in rats after 24 hours, and no changes in TNF-alpha bioactivity were seen in any group. No alterations in BrDU-IR were seen in rats fed restricted (80% of control) diets. These studies show that sublethal doses of toxin decrease weight gain and affect growth of long bones through suppression of chondrocyte proliferation. These effects may be mediated by direct binding of the toxin to target cells or IL-6 but are not associated with altered feed intake or TNF-induced cachexia.

Calbindin-D28K and -D9K and 1,25(OH)2 vitamin D3 receptor immunolocalization and mineralization induction in long-term primary cultures of rat epiphyseal chondrocytes

Rat epiphyseal plat chondrocytes were grown on glass slides, as nonadhering monolayer cultures for up to 6 weeks. Chondrocyte growth, differentiation and maturation, matrix formation and mineralization, and the temporospatial distribution of the vitamin D-dependent calcium-binding proteins, calbindin-D9K and -D28K, and the 1,25(OH)2D3 receptor (VDR), were all monitored. Chondrocytes became confluent in 2.5 weeks, differentiated to acquire a chondrocyte (polygonal) morphology, produced extracellular matrix, and finally formed a true monolayer mineralizing cartilaginous tissue, with all the stages of chondrocyte development within a single culture. beta-Glycerophosphate promoted initial matrix mineralization in 4 weeks and accelerated cell differentiation. High nominal calcium and ascorbic acid were needed for abundant matrix formation. VDR occurred at all differentiation stages, in the nuclei and nucleoli and in the cytoplasm. Calbindin-D28K and -D9K were not coexpressed. Calbindin-D28K was found in prechondroblasts, chondroblasts, and in newly differentiated chondrocytes. It was cytoplasmic in prechondroblasts and subsequently also in the perinuclear region and in nuclei, suggesting migration to the nuclear chromatin. Calbindin-D28K was nuclear only in newly differentiated chondrocytes in vitro and was not found in mature chondrocytes. In contrast, calbindin-D9K was present in the cytoplasm of mature and hypertrophic chondrocytes only. It was first in the cell body and eventually migrated within and to the far end of long cell processes with a decreasing cytoplasmic concentration showed by decreased immunostaining intensity, and ultimately hypertrophy of chondrocytes in culture. These in vitro patterns of calbindins-D and VDR accurately reflect their in vivo distributions. The genomic action of vitamin D, in vitro, resulted in the synthesis of nuclear VDR and calbindins-D.(ABSTRACT TRUNCATED AT 250 WORDS)