Sage Journals: Discover world-class research

Abstract

This review discusses the regulation of growth plate chondrocytes by vitamin D₃. Over the past ten years, our understanding of how two vitamin D metabolites, 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃, exert their effects on endochondral ossification has undergone considerable advances through the use of cell biology and signal transduction methodologies. These studies have shown that each metabolite affects a primary target cell within the endochondral developmental lineage. 1α,25-(OH)₂D₃ affects primarily growth zone cells, and 24R,25-(OH)₂D₃ affects primarily resting zone cells. In addition, 24R,25-(OH)₂D₃ initiates a differentiation cascade that results in down-regulation of responsiveness to 24R,25-(OH)₂D₃ and up-regulation of responsiveness to 1α,25-(OH)₂D₃. 1α,25-(OH)₂D₃ regulates growth zone chondrocytes both through the nuclear vitamin D receptor, and through a membrane-associated receptor that mediates its effects via a protein kinase C (PKC) signal transduction pathway. PKCα is increased via a phosphatidylinositol-specific phospholipase C (PLC)-dependent mechanism, as well as through the stimulation of phospholipase A₂ (PLA₂) activity. Arachidonic acid and its downstream metabolite prostaglandin E₂ (PGE₂) also modulate cell response to 1α,25-(OH)₂D₃. In contrast, 24R,25-(OH)₂D₃ exerts its effects on resting zone cells through a separate, membrane-associated receptor that also involves PKC pathways. PKCα is increased via a phospholipase D (PLD)-mediated mechanism, as well as through inhibition of the PLA₂ pathway. The target-cell-specific effects of each metabolite are also seen in the regulation of matrix vesicles by vitamin D₃. However, the PKC isoform involved is PKCζ, and its activity is inhibited, providing a mechanism for differential autocrine regulation of the cell and events in the matrix by these two vitamin D₃ metabolites.

The Growth Plate as a Model

Our understanding of the mechanisms of vitamin D₃ action has changed considerably over the past decade, due in large part to advances resulting from the study of endochondral ossification. The growth plate is an ideal model for this because of its unique anatomy [see (Howell and Dean, 1992) for a review]. The lack of a vasculature ensures that only one cell type, the chondrocyte, is present in the growth plate. Moreover, cells in the endochondral chondrocyte lineage are aligned in parallel columns, with the least mature cells at one end and terminally differentiated chondrocytes at the other end.

These cells can be subdivided into maturation zones that can be visualized under a dissecting microscope. The resting zone (reserve zone) contains fully committed chondrocytes that produce a type II collagen matrix rich in proteoglycan aggregates typical of articular cartilage. This zone serves two purposes in the post-fetal growth plate: It provides a mechanical buffer so that mechanical forces on the bone are appropriately dispersed, and it permits interstitial growth to occur in the growth plate proper. Cells in the resting zone are not proliferative. However, in response to signals that are only now beginning to be understood, cells at the base of the resting zone begin to proliferate and undergo a set number of cell divisions. Because of the morphology of the growth plate, the columns of the proliferating cells form a discrete horizontal band, which is called the proliferative cell zone. Once proliferation is complete, the growth plate chondrocytes enter into a maturation phase and prepare to undergo hypertrophy. This region is termed the zone of maturation, or the pre-hypertrophic cell zone. It is somewhat difficult to distinguish where this zone stops and the hypertrophic process begins. As a result, it is often examined together with the upper hypertrophic cell zone. As the cells become hypertrophic, there is a tremendous increase in cell size that must be accompanied by rapid remodeling of the extracellular matrix. In the lower hypertrophic cell zone, there is evidence of initial calcium phosphate crystal deposition in extracellular matrix vesicles, and in the calcifying cell zone, the presence of mineral in the matrix becomes evident. Once the cartilage has calcified, it is resorbed by chondroclasts and vascularized by endothelial cells. As a consequence of the vascularization process, bone marrow stromal cells migrate onto the calcified cartilage spicules. Osteoprogenitor cells differentiate into osteoblasts and form bone. Cartilage calcification also appears to be needed for the formation of bone marrow.

Differences in chondrocyte morphology have been used to advantage by several teams of investigators to divide the cells into groups by maturation state, enabling them to study the role of endochondral maturation state in cellular response to local and systemic regulatory factors. One approach is to separate the cells by centrifugal elutriation, because the more mature cells are larger and have a greater density (O'Keefe et al., 1989). Our laboratory has used a different approach, taking advantage of the visual differences between maturation zones. By using sharp dissection to separate cells in the uppermost resting zone (reserve zone) from those in the post-proliferative, pre-hypertrophic, and upper hypertrophic zones (collectively termed the “growth zone”), we have been able to study chondrocytes at two distinct states of endochondral development. Even though neither resting zone cells nor growth zone cells proliferate in vivo, both groups of cells are able to proliferate in culture. Therefore, the number of cells can be amplified sufficiently for meaningful studies to be performed. Moreover, they retain their in vivo phenotype in culture [see (Boyan et al., 1997a) for a review], allowing us to make interpretations of the in vitro results that have relevance to the growth plate in vivo.

It must be emphasized that most of the proliferative cell zone is discarded. Any proliferative zone cells remaining are present in the resting zone cell population. By discarding the proliferative cell zone, one eliminates the potential for cross-contamination of cells in the resting zone with cells in the post-proliferative growth zone. Although both resting zone cells and growth zone cells contain heterogenous populations of chondrocytes, this cell culture model ensures that cartilage cells at distinctly different states of endochondral maturation are compared. Similarly, cells in the lower hypertrophic and calcifying cell zones are discarded, thus limiting the potential for contamination of growth zone cells with osteoprogenitor cells and osteoblasts.

Chondrocytes exhibit another feature that has been critical to the vitamin D₃ story. These cells produce extracellular organelles called matrix vesicles. The matrix vesicles are membrane-bound and contain a variety of components, depending on the cells that produce them. The membrane phospholipid composition differs from that of the plasma membrane (Peress et al., 1974; Boyan et al., 1988a), and there is a differential distribution of membrane and cellular enzymes as well (Takagi et al., 1981; Boyan et al., 1988a,b). Other features of the matrix vesicles show that they are distinct from the plasma membrane, although freeze-fracture micrographs and other imaging techniques have shown that they are released into the matrix during matrix synthesis by a process that appears to involve plasma membrane “budding” from cytoplasmic extensions (Cecil and Anderson, 1978; Hale and Wuthier, 1987). Most important to this discussion is the fact that matrix vesicles do not contain DNA or RNA. As a consequence, they have been ideal for sorting out the mechanisms of vitamin D₃ action on the cells.

In the growth plate, the number of matrix vesicles produced by the cells varies with maturation zone. In addition, the biochemistry of the organelles is dependent on the cell that produces them, so it is not surprising that the composition of the matrix vesicles also varies as a function of chondrocyte maturation (Boyan and Ritter, 1984). Moreover, this difference in matrix vesicle composition is retained in matrix vesicles produced by chondrocytes in culture (Boyan et al., 1988a). Thus, matrix vesicles produced by pre-hypertrophic/upper hypertrophic chondrocytes contain greater levels of alkaline phosphatase (the matrix vesicle marker enzyme) and phospholipase A₂ specific activities than do matrix vesicles produced by resting zone cells (Schwartz and Boyan, 1988). Both of these enzyme activities increase with endochondral maturation in the growth plate in vivo (Wuthier, 1973). In addition, matrix vesicles produced by pre-hypertrophic/hypertrophic chondrocytes contain higher levels of metalloproteinases involved in the process of matrix remodeling associated with chondrocyte hypertrophy and mineralization than do matrix vesicles produced by resting zone cells (Dean et al., 1992).

Endocrine Regulation of the Growth Plate by Vitamin D

The growth plate is a major target for vitamin D₃ metabolites. Regardless of whether vitamin D₃-replete rats are injected with [³H]-25-(OH)D₃, [³H]-1α,25-(OH)₂D₃, or [³H]-24R,25-(OH)₂D₃, the highest levels of both [³H]-1,25-(OH)₂D₃ and [³H]-24,25-(OH)₂D₃ are found in the growth plate cartilage (Seo et al., 1996). The fact that 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ localize in the growth plate cartilage suggests that they play important physiological roles in this tissue. In the absence of vitamin D₃, the growth plate fails to mineralize, and the hypertrophic zone becomes enlarged (Dean et al., 1985, 1989). This condition, known as rickets, occurs because the chondrocytes continue to mature and enter hypertrophy, but the chondroclasts cannot resorb the hypertrophic cells, since the cartilage is not calcified. Vascular invasion does not occur, and the metaphyseal bone cannot form properly. The bone/cartilage interface is destabilized, and bowing and shortening of the long bones result.

Many of the studies examining the role of vitamin D₃ and its metabolites in growth plate physiology have used one of two models. The chick epiphyseal growth plate is particularly useful, since chicks can be made rachitic by the elimination of vitamin D from the diet and the incubation of newly hatched birds under incandescent, rather than fluorescent, light. Induction of rickets in rats requires a combination of a low-phosphate diet and a low-vitamin D diet (Simon et al., 1973). Many of the effects attributed to low levels of 1,25-(OH)₂D₃ or 24,25-(OH)₂D₃ in the rat may actually be a consequence of low serum phosphate, since many of the morphological characteristics of rickets can be seen in hypophosphatemic (Hyp) mice. In these animals, the hypertrophic cell zone is enlarged, but not calcified (Miao et al., 2000). Consequently, there is a reduction in osteoclasts, with a concomitant of hypertrophic cartilage to be resorbed. Recent studies have shown that phosphate can induce apoptosis of growth plate chondrocytes (Mansfield et al., 1999). Since hypophosphatemia is a feature of vitamin D₃-deficiency rickets (Anderson and Sajdera, 1976; Howell et al., 1978), it is possible that pathologic features in vitamin D-deficient rats are due to low phosphate, and not to low vitamin D₃. These observations also suggest that 1α,25-(OH)₂D₃ may exert some of its effects via changes in phosphate.

Rickets can be treated by increasing the serum Ca⁺⁺ concentration (Holtrop et al., 1986), in part because the matrix vesicles produced by rachitic chondrocytes contain competent nucleational sites for calcium phosphate deposition to occur if Ca⁺⁺ is present (Howell et al., 1978; Boyan and Ritter, 1984). Treatment of rickets with 1α,25-(OH)₂D₃ is particularly effective, since this vitamin D metabolite not only facilitates Ca⁺⁺ transport (Norman et al., 1992), but also modulates other aspects of the physiology of the growth plate chondrocytes, including increases in local phosphate, resulting in optimal restoration of function.

24R,25-(OH)₂D₃ is also an important modulator of growth plate chondrocytes in vivo. Treatment of vitamin D- and phosphate-deficient rats with 24R,25-(OH)₂D₃ can also heal rickets, whether it is given systemically or injected locally into the reserve zone of the epiphyseal growth plate (Atkin et al., 1985; Lidor et al., 1987a). Similarly, both systemic (Ornoy et al., 1978) and local (Lidor et al., 1987b,c) injection of 24R,25-(OH)₂D₃ enhances the healing of fracture callus in the chick. While fracture callus is not a growth plate per se, the cells within the callus undergo endochondral ossification, indicating that 24R,25-(OH)₂D₃ may be important for endochondral development. This hypothesis is borne out by studies with the mandibular condyle as a model. The mandibular condyle is what is termed a secondary growth plate. Unlike primary growth plates such as those found in the epiphysis or at the costochondral junction, secondary growth plates remain functional in adults, remodeling continuously in response to mechanical force. The mandibular condyle does not have an established resting zone. Instead, multipotent mesenchymal cells differentiate directly into chondroblasts, produce a cartilage-specific extracellular matrix, and then undergo hypertrophy. Numerous studies have shown that these cells are particularly sensitive to 24R,25-(OH)₂D₃ (Mirsky and Silbermann, 1984; Somjen et al., 1987). Moreover, during embryonic bone formation, in which bone forms on a cartilagenous anlage via endochondral ossification, 24R,25-(OH)₂D₃ has been shown to regulate bone growth (Schwartz et al., 1989).

Because the studies described above were done under conditions where further metabolism of 24R,25-(OH)₂D₃ to 1,24,25-(OH)₃D₃ could not be ruled out, the possibility that 24,25-(OH)₂D₃ was not the active vitamin D₃ metabolite involved remained unresolved. Certainly, in vivo and organ culture experiments did not identify a target cell for the metabolite. Autoradiographic evidence of specific receptors for 24,25-(OH)₂D₃ (Corvol et al., 1980) was strong presumptive evidence that the vitamin D₃ metabolite played a role in the endocrine regulation of the growth plate by vitamin D₃. However, a specific nuclear receptor for 24,25-(OH)₂D₃ was not identified. The issue of whether 1,25-(OH)₂D₃ or 24,25-(OH)₂D₃ was responsible for vitamin D₃ effects in the growth plate has also been clouded by reports that chondrocytes themselves could metabolize 25-(OH)D₃ to 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃, and 24,25-(OH)₂D₃ could be metabolized to 1,24,25-(OH)₃D₃ (Garabedian et al., 1978).

Despite these concerns, there is increasing evidence that both vitamin D₃ metabolites play a role in the regulation of endochondral ossification. In a recent study done in collaboration with Dr. David Howell, we showed that systemic 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ exert very different effects on rachitic rat growth plates. In rachitic animals, there is a significant increase in both neutral metalloproteinase and collagenase activities in the growth plate (Fig. 1) (Dean et al., 2000, 2001). 1α,25-(OH)₂D₃ has no effect on neutral metalloproteinase activity, but causes a decrease in collagenase activity. In contrast, 24R,25-(OH)₂D₃ decreases the activity of neutral metalloproteinases and has no effect on collagenase activity. If, indeed, 24R,25-(OH)₂D₃ had been converted to 1,24,25-(OH)₃D₃, we would have expected it to elicit a response more like that of 1α,25-(OH)₂D₃. Moreover, these in vivo results are comparable with those from our cell culture studies, showing that 1α,25-(OH)₂D₃ affects only collagenase in growth zone cells, and 24R,25-(OH)₂D₃ affects only neutral metalloproteinase activity in resting zone cells (Maeda et al., 2000a,b).

Autocrine Regulation of the Growth Plate by Vitamin D3 Metabolites

Using the rat costochondral cartilage cell culture model described above, we have been able to show definitively that 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ both regulate growth plate chondrocytes, but that each metabolite exerts its effects predominantly on cells at different states of endochondral maturation. The studies that demonstrated this are described in detail in several reviews (Boyan et al., 1992, 1997b, 2001) and are summarized only briefly here.

1α,25-(OH)₂D₃ causes a decrease in proliferation of both resting zone cells and growth zone cells. It also causes an increase in proteoglycan sulfation in both cell types. However, it exerts its effects on differentiation only in cells derived from the growth zone. These include an increase in specific activities of alkaline phosphatase and phospholipase A₂ and a decrease in the specific activity of collagenase. The effect of 1α,25-(OH)₂D₃ on alkaline phosphatase is specifically targeted to matrix vesicles, since alkaline phosphatase specific activity increased in these extracellular organelles, whereas activity in the plasma membranes is unchanged following a 24-hour exposure of growth zone cells to the vitamin D₃ metabolite. There is also an increase in matrix vesicle phospholipase A₂ and metalloproteinases that degrade proteoglycans. These effects are commensurate with pre-hypertrophic maturation, since these cells are preparing to calcify their matrix.

In contrast, 24R,25-(OH)₂D₃ modulates proliferation only in resting zone cells, and the 24R,25-(OH)₂D₃-dependent decrease in cell proliferation is accompanied by stimulation of resting zone cell differentiation and maturation. Matrix vesicle alkaline phosphatase specific activity is increased, but specific activity of phospholipase A₂ is decreased, as is metalloproteinase activity. This suggests that 24R,25-(OH)₂D₃ stimulates the initial events in endochondral differentiation, but its effects also protect the matrix from premature calcification. Exposure of resting zone cells to 24R,25-(OH)₂D₃ for 36 hrs or more results in down-regulation of responsiveness to 24R,25-(OH)₂D₃ and up-regulation of responsiveness to 1α,25-(OH)₂D₃, indicating that the cells undergo a maturational shift to a growth zone phenotype (Schwartz et al., 1995). Treatment of resting zone cells with 1α,25-(OH)₂D₃ for up to five days does not elicit a shift in maturation; the cells remain responsive to 24R,25-(OH)₂D₃.

Many of the effects of 24R,25-(OH)₂D₃ on matrix vesicles produced by resting zone cells and the effects of 1α,25-(OH)₂D₃ on matrix vesicles produced by growth zone cells are due to direct action of the vitamin D₃ metabolites on the organelles. When matrix vesicles that are isolated from the extracellular matrix of the cultured cells are incubated directly with the vitamin D metabolites, the cell-specific effects on enzyme activity are elicited. For example, 24R,25-(OH)₂D₃ stimulates alkaline phosphatase specific activity in matrix vesicles isolated from cultures treated for 24 hrs with the vitamin D₃ metabolite, but matrix vesicle phospholipase A₂ is decreased. Similarly, when matrix vesicles are isolated from cultures and then incubated directly with 24R,25-(OH)₂D₃, alkaline phosphatase specific activity is increased and phospholipase A₂ specific activity is decreased. 1α,25-(OH)₂D₃ causes an increase in both enzyme activities in matrix vesicles isolated from cultures incubated with the metabolite for 24 hrs. Similarly, matrix vesicles that are isolated from cultures that were not treated with 1α,25-(OH)₂D₃ exhibit increases in both enzyme activities when they are incubated directly with the metabolite. Furthermore, the magnitude of the direct effect is comparable with that of the effect observed in the matrix vesicles isolated from the treated cultures. Moreover, the direct effects, like the effects observed in matrix vesicles from treated cultures, are cell-maturation-specific. 24R,25-(OH)₂D₃ elicits a direct effect only on matrix vesicles isolated from resting zone cell cultures, and 1α,25-(OH)₂D₃ elicits a direct effect only on matrix vesicles isolated from growth zone cell cultures. This indicates that at least part of the effect of 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ does not involve traditional 1α,25-(OH)₂D₃ receptor (VDR) mechanisms. It also suggests that the growth plate chondrocytes produce their own vitamin D₃ metabolites, supporting the results from the previous in vivo studies described above.

To test this hypothesis, we incubated resting zone and growth zone cells with [³H]-25-(OH)D₃ and determined their ability to produce radiolabeled 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ (Schwartz et al., 1992a). Not surprisingly, both metabolites were produced in a time-dependent manner, and the levels that were secreted were comparable with the concentrations that elicited effects in the cultured cells as well as in isolated matrix vesicles. Moreover, production of 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ was regulated by TGF-β1, dexamethasone, and the metabolites themselves in a cell-specific manner. Subsequent studies have shown that resting zone cells and growth zone cells possess both the 1α-hydroxylase and the 24R-hydroxylase, and activities of these enzymes are modulated by TGF-β1, 1α,25-(OH)₂D₃, and 24R,25-(OH)₂D₃ in a complex interrelated manner, suggesting that the secreted metabolites can act back on the cells via autocrine mechanisms through cyclical regulatory pathways (Pedrozo et al., 1999a; Schwartz et al., 2002).

The physiological significance of the locally produced vitamin D₃ metabolites in vivo remains unclear. It is evident that 24R,25-(OH)₂D₃ is an important regulator of cell function, since the 24-hydroxylase knock-out mouse lacks a mandible and fails to mineralize its osteoid (St.-Arnaud, 1999). The growth plates of these animals appear to be normal empirically, but they have not yet been subjected to more stringent analysis. At least some of the pathology observed in this mouse can be attributed to high circulating levels of 1,25-(OH)₂D₃. It may be that local production of 24,25-(OH)₂D₃ is not as important in fetal and neonatal animals as it is in adolescent growth plate development. Despite these unresolved concerns, the cell culture studies indicate that local production of both 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ plays a role in fine-tuning chondrocyte function in the growth plate.

Mechanisms of 1,25-(OH)2D3 and 24,25-(OH)2D3 Action in the Growth Plate

Several questions have emerged over the past five to ten years concerning the mechanisms involved in the action of vitamin D₃ metabolites in the growth plate. Although receptors for both 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ were demonstrated by radioautography (Corvol et al., 1980), only the nuclear receptor for 1,25-(OH)₂D₃ was identified (Klaus et al., 1994). How could 24,25-(OH)₂D₃ work? 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ both exerted cell-specific effects on matrix vesicles isolated from the extracellular matrix of the cell cultures, where traditional VDR-mediated mechanisms were impossible. What kind of mechanism was involved? Many of the effects of the metabolites were targeted to the matrix vesicles and were not seen in plasma membranes from which the matrix vesicles were supposedly derived. How was this possible? The effects of 1α,25-(OH)₂D₃ were primarily on growth zone cells, whereas the effects of 24R,25-(OH)₂D₃ were almost exclusively on resting zone cells. How was this possible if they both have traditional 1,25-(OH)₂D₃ VDRs? The studies described below have begun to provide answers to these questions.

Specific membrane receptors for 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃

Growth zone and resting zone cells differ from one another in several ways. Not only is the phospholipid composition of the cytoplasmic membrane different in each cell type, but the phospholipid compositions of the matrix vesicles differ as well (Boyan et al., 1988a). These matrix vesicles are isolated from the cultures by differential centrifugation of the trypsin-digested extracellular matrix following removal of any cells. They exhibit a phospholipid composition typical of matrix vesicles isolated by collagenase digestion of the extracellular matrix of the growth plate. Moreover, the composition is distinct from that of vesicles found in the conditioned media of the cells, and it differs from that of the cytoplasmic membrane of its parent cell. These differences suggested that the fluidity of the membranes might also vary, and this proved to be the case. Since the two vitamin D₃ metabolites are lipophilic and have distinctly different charge densities, we hypothesized that they exerted their cell-specific and organelle-specific effects by modulating fluidity of the membrane. Indeed, fluidity of the cytoplasmic and matrix vesicle membranes was affected in a metabolite-dependent manner, and the effect was specific to the membrane being examined (Swain et al., 1993). Thus, at least part of the specificity of the effect could be explained by straightforward physical-chemical interactions.

These studies also pointed to phospholipids and their metabolism as a key component of the regulatory process. Not surprisingly, 1α,25-(OH)₂D₃ caused a rapid increase in the release of [¹⁴C]-arachidonic acid from pre-labeled growth zone cells, but had no effect on this parameter in resting zone cells (Schwartz et al., 1990; Swain et al., 1992). In contrast, 24R,25-(OH)₂D₃ caused a rapid decrease in arachidonic acid release, but had no effect on this parameter in growth zone cells. Re-incorporation of the released [¹⁴C]-arachidonic acid was also affected in a cell-specific manner. The rapid retailoring of membrane phospholipids, particularly the production of lyso-phospholipids, could account for the change in fluidity, but it also suggested that membrane-associated signal transduction pathways might play a role.

Related studies in our lab indicated that resting zone cells and growth zone cells also managed Ca⁺⁺ ions differently (Langston et al., 1990; Schwartz et al., 1991). Basal levels of the ions differed, and the change in Ca⁺⁺ flux in response to 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ was different. Whereas resting zone cells responded to 24R,25-(OH)₂D₃ with a rapid influx of ⁴⁵Ca⁺⁺, growth zone cells responded with a rapid efflux of ⁴⁵Ca⁺⁺. Perhaps even more importantly, Ca⁺⁺ flux in resting zone cells was unaffected by 1α,25-(OH)₂D₃, and 24R,25-(OH)₂D₃ had no effect on the rapid flux of Ca⁺⁺ in growth zone cells. This specificity of the Ca⁺⁺ response indicated a receptor-mediated mechanism.

Rapid changes in Ca⁺⁺ were also noted in chick intestinal epithelium in response to 1α,25-(OH)₂D₃ (Norman, 1997). Similarly, 1α,25-(OH)₂D₃ caused a rapid stimulation of voltage-gated Ca⁺⁺ channels in osteoblasts (Farach-Carson et al., 1991). To understand the phenomenon better, these investigators used analogues of 1,25-(OH)₂D₃ that had reduced affinity to the 1,25-(OH)₂D₃ nuclear VDR (1,25-nVDR). In a series of elegant studies, they showed that the activation of the voltage-gated Ca⁺⁺ channels was due to analogues with low affinity to the 1,25-nVDR, whereas analogues with high affinity to this receptor elicited the antiproliferative effect of the metabolite (Farach-Carson et al., 1993).

These studies in other systems supported our own observations pointing to a new class of receptors. We showed that the protein kinase C (PKC) signal transduction pathway was differentially regulated by 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ in growth plate chondrocytes (Sylvia et al., 1993). This was an amazing observation. 1α,25-(OH)₂D₃ causes a rapid increase in PKC activity in growth zone cells, but has no effect on PKC in resting zone cells. In contrast, 24R,25-(OH)₂D₃ causes a rapid increase in PKC in resting zone cells, but has no effect on PKC in growth zone cells. The effects of each metabolite are stereospecific. 1β,25-(OH)₂D₃ has no effect on PKC in growth zone cells (unpublished data), and 24S,25-(OH)₂D₃ has no effect on PKC in resting zone cells (Schwartz et al., 2000).

In both types of chondrocytes, the PKC involved in the responses of the cells was PKCα, yet the mechanisms by which the vitamin D₃ metabolites elicited their effects were very different. The PKC response to 1α,25-(OH)₂D₃ in growth zone cells was independent of new gene transcription or protein synthesis, whereas the PKC response to 24R,25-(OH)₂D₃ in resting zone cells involved both processes. The effect of 1α,25-(OH)₂D₃ was maximal at 9 min and had resolved completely by 90 min. The effect of 24R,25-(OH)₂D₃ was not maximal until 90 min, and activity was still elevated at 270 min. Part of the effect of 1α,25-(OH)₂D₃ involved translocation of cytosolic PKCα to the plasma membrane, but 24R,25-(OH)₂D₃ did not elicit translocation at all. Both metabolites caused an increase in existing plasma membrane PKCα when they were incubated directly with plasma membranes from their target cells (Sylvia et al., 1996), and this effect was rapid, occurring within 3 min. Both metabolites elicited an increase in diacylglycerol (DAG) production, but the effect of 1α,25-(OH)₂D₃ was rapid, whereas the effect of 24R,25-(OH)₂D₃ was delayed. Only 1α,25-(OH)₂D₃ caused an increase in inositol 1,4,5-trisphosphate production. The differences in the rapid response of each cell to its target metabolite of vitamin D₃ are discussed in greater detail in reference to Fig. 6 below.

In addition to the effects noted above, both 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ caused an increase in production of new PKC in their target cells, and this was targeted to their plasma membranes and matrix vesicles. As expected, the increase in the plasma membrane PKC was due to the PKCα isoform, but the increase in the matrix vesicle enzyme was due to the PKCζ isoform (Sylvia et al., 1996). Differential distribution of PKC isoforms in response to 1α,25-(OH)₂D₃ has also been noted in kidney cells (Simboli-Campbell et al., 1994). In the kidney cells, 1α,25-(OH)₂D₃ stimulated PKCα in the cell membranes, due in part to new PKCα production and in part to translocation of cytosolic PKCα. At the same time, PKCβ was translocated to the nuclear membrane.

Differential distribution of PKC isoforms between the cell and matrix vesicle membranes in the growth plate has particular significance. PKCα is sensitive to phospholipid and Ca⁺⁺ (Newton, 1995), but PKCζ is an atypical form of PKC and is insensitive to both phospholipid and Ca⁺⁺ (Liyanage et al., 1992). When matrix vesicles were incubated directly with their target cell-specific metabolite, PKCζ was inhibited. The non-target cell metabolite had no effect in either instance. Matrix vesicles are phospholipid-rich membranes, and the organelles are found in a Ca⁺⁺-rich environment. Thus, it is of interest that the isoform that was sensitive to the vitamin D₃ metabolite was insensitive to these two parameters. In addition, activity of PKCζ was decreased, whereas that of PKCα in the plasma membrane was increased (Sylvia et al., 1996), suggesting that the cells use this strategy to differentiate the autocrine actions of the secreted metabolites in the matrix from those in the cells.

The remarkable specificity of the target cell response, regardless of the membrane examined, also pointed to the probability of specific, and different, receptors for each metabolite. Results from studies with analogues of 1α,25-(OH)₂D₃, similar to those used for the study of the voltage-gated channels in osteoblasts, supported the hypothesis that the receptors were membrane-associated and suggested that the action of the hormone on the receptor might be related to downstream genomic events (Boyan et al., 1997c; Greising et al., 1997; Schwartz et al., 1997).

We hypothesized that if a membrane receptor to 1α,25-(OH)₂D₃ existed, it would be conserved in much the same manner as the 1,25-nVDR. Nemere and Norman and co-workers reported that rapid transcaltachia in chick intestinal epithelium could be blocked by an antibody generated to a 1α,25-(OH)₂D₃-binding protein isolated from the basal lateral membrane (Nemere et al., 1994). The hypothesis was proved to be correct. The antibody (Ab99) blocked the 1α,25-(OH)₂D₃-dependent effects on PKC in growth zone cells, as well as the effects of the 1α,25-(OH)₂D₃ analogue [1α-(hydroxymethyl)-3β-hydroxy 20-epi-22-oxa-26,27-dihomo vitamin D₃ (analogue 3a)] on this enzyme (Fig. 2) (Nemere et al., 1998). As expected, [³H]-1α,25-(OH)₂D₃ bound the membranes with high specificity, and Ab99 identified a 65-kDa protein on Western blots of the matrix vesicles (Nemere et al., 1998). Ab99 also blocked the physiological effects of 1α,25-(OH)₂D₃ on the cells (Pedrozo et al., 1999b), indicating either that these effects were mediated exclusively by the membrane receptor or that the effects of the membrane receptor were somehow affecting subsequent nuclear events, including those that involved the 1,25-nVDR. The fact that an antibody generated to a chick peptide recognized a 1α,25-(OH)₂D₃ binding protein in rat chondrocytes is remarkable, and highlights the functional importance and conservation of the epitope(s) involved.

Studies with the antibody also showed that the membrane receptor for 1α,25-(OH)₂D₃ was distinct from the receptor for 24R,25-(OH)₂D₃. Ab99 did not block the effects of 24R,25-(OH)₂D₃ on PKC in resting zone cells, nor did it block the effects of analogues that exhibit 24R,25-(OH)₂D₃-like effects on these cells (Pedrozo et al., 1999b). Matrix vesicles from resting zone cells exhibited specific binding for [³H]-24R,25-(OH)₂D₃. All of this pointed to two independent membrane receptors. However, matrix vesicles from resting zone chondrocyte cultures also bound [³H]-1α,25-(OH)₂D₃, and matrix vesicles from growth zone cells bound [³H]-24R,25-(OH)₂D₃, indicating that both types of receptors were present on both cell types. Scatchard analysis showed that the numbers of receptors for each ligand differed, but fewer receptors alone could not account for the ability of the cells to limit the PKC response to only one of the metabolites.

Role of phospholipases C and D in mediating the cell-specific PKC response

The observation that 1α,25-(OH)₂D₃ caused rapid increases in DAG and inositol 1,4,5-trisphosphate production in growth zone cells suggested that phospholipase C (PLC) might play a role in the mechanism of PKC activation. This hypothesis was supported by the observation that 1α,25-(OH)₂D₃ could stimulate PLC in other systems (Lieberherr et al., 1989). DAG binds to PKC, stimulating enzyme activity (Lapetina et al., 1985). Not surprisingly, when growth zone chondrocytes were treated with the DAG, dioctanoylglycerol (DOG), there was a rapid increase in PKC, with peak activity occurring at 9 min (Sylvia et al., 1998). In addition, DOG enhanced the PKC response to 1α,25-(OH)₂D₃. Moreover, when the chondrocytes were incubated with 1α,25-(OH)₂D₃ in the presence of U73122, a specific inhibitor of phosphatidylinositol-specific PLC (PI-PLC), the effect of the vitamin D₃ metabolite on PKC was blocked (Fig. 3).

In contrast, PLC was not involved in the response of resting zone cells to 24R,25-(OH)₂D₃. U73122 had no effect on 24R,25-(OH)₂D₃-stimulated PKC activity (Fig. 3 ). In addition, as noted for growth zone cells, DOG stimulated PKC in resting zone cells, but maximal increases in activity were not observed until 90 min. Moreover, 24R,25-(OH)₂D₃ caused delayed production of DAG and no production of inositol 1,4,5-trisphosphate, providing further evidence that PLC was not involved.

It was possible that DAG was produced in response to the vitamin D₃ metabolites by pathways other than through the action of PLC on phosphatidylinositol. Phosphatidylcholine-specific PLC can also result in DAG production (Exton et al., 1991). However, when the chondrocytes were treated with D609, a specific inhibitor of phosphatidylcholine-specific PLC (Muller-Decker, 1989), PKC activity was unaffected, whether or not the cells were stimulated with their target vitamin D₃ metabolite. These results indicated that PI-PLC was responsible for the rapid increase in PKC in growth zone cells in response to 1α,25-(OH)₂D₃, but they did not explain the delayed increase in PKC in resting zone cells in response to 24R,25-(OH)₂D₃.

Phospholipase D catalyzes the first step in an alternative pathway of phosphatidylcholine metabolism, ultimately causing an increase in DAG (Martin, 1988). Both resting zone cells and growth zone cells express mRNAs for two PLD isoforms, PLD1 and PLD2, although mRNA levels and basal activity of PLD are higher in resting zone cells. To assess the role of PLD in the response of the growth plate chondrocytes to the vitamin D₃ metabolites, we used two different inhibitors of PLD, wortmannin (Carrasco-Marin et al., 1994) and erythrodihydrosphingosine (EDS) (Franson et al., 1992). Both inhibitors caused decreases in basal PLD activity in the cells (Sylvia et al., 2001). Similarly, basal PKC activity was decreased in both cell types. However, only PKC stimulated by 24R,25-(OH)₂D₃ was blocked by PLD inhibition. Neither inhibitor affected 1α,25-(OH)₂D₃-dependent PKC activity in growth zone cells, whereas both wortmannin (Fig. 4) and EDS (data not shown) blocked 24R,25-(OH)₂D₃-dependent PKC activity in resting zone cells.

These observations clearly demonstrate that the cell-specific regulation of PKC by 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ involves phospholipid metabolism. In growth zone cells, 1α,25-(OH)₂D₃ causes rapid hydrolysis of phosphatidylinositol by PLC, resulting in production of DAG, and DAG stimulates PKC. In resting zone cells, 24R,25-(OH)₂D₃ activates hydrolysis of phosphatidylcholine, initiating an alternate pathway of DAG production. The increase in DAG leads to an increase in existing PKC.

Part of the effect of 24R,25-(OH)₂D₃ on PKC involves activation of existing enzyme, and part of the response involves new PKC synthesis (Sylvia et al., 1993). However, a nuclear receptor for 24R,25-(OH)₂D₃ has not yet been identified. Analogues that mimic the effect of 24R,25-(OH)₂D₃ in resting zone cells, but do not bind to the 1,25-nVDR, also increase PKC with maximal effects at 90 min, and both activation of existing plasma membrane PKC and new protein synthesis are involved (Boyan et al., 1997c; Greising et al., 1997). This suggests that the increase in existing PKC may result in new gene expression via the PKC signal transduction pathway and may not require a specific 24R,25-(OH)₂D₃ nuclear receptor per se.

Mitogen-activated protein (MAP) kinase appears to play a role in this process. Others have shown that 1α,25-(OH)₂D₃ stimulates MAP kinase in intestinal epithelial cells (de Boland and Norman, 1998). MAP kinase regulated through the PKC pathway acts by phosphorylating transcription factors like AP-1 (Cobb, 1999). Analysis of preliminary data from our laboratory indicates that 24R,25-(OH)₂D₃ activates MAP kinase in resting zone cells, and that 1α,25-(OH)₂D₃ activates MAP kinase in growth zone cells, suggesting that this pathway may be involved in signal transduction leading to genomic events in both cell types.

Clearly, such a pathway would be important for 24R,25-(OH)₂D₃, if indeed a nuclear receptor does not exist. Attempts to identify a nuclear receptor for this vitamin D₃ metabolite in chick fracture callus, which is enriched in chondrocytes, failed to do so, although specific membrane-binding of 24R,25-(OH)₂D₃ was shown (Kato et al., 1998). For 1α,25-(OH)₂D₃, the importance of the PKC/MAP kinase pathway is less evident, since chondrocytes have nuclear receptors for 1α,25-(OH)₂D₃. It is possible that the PKC/MAP kinase pathway acts as a fine-tuning mechanism for the traditional VDR-mediated mechanism, or that it modulates distinctly different actions that work in concert with the effects of the 1,25-mVDR.

The role of phospholipase A₂

Further modulation of the differential responses of resting zone chondrocytes and growth zone chondrocytes to the vitamin D₃ metabolites is provided by the phospholipase A₂ (PLA₂) signal transduction pathway. 1α,25-(OH)₂D₃ stimulates PLA₂ in growth zone cells, but not in resting zone cells (Schwartz and Boyan, 1988). In contrast, 24R,25-(OH)₂D₃ inhibits PLA₂ in resting zone cells, but has no effect on PLA₂ in growth zone cells. A major product of the action of PLA₂ is arachidonic acid, which is itself a substrate for cyclooxygenase, an enzyme responsible for prostaglandin biosynthesis (Fletcher, 1993). Not surprisingly, 1α,25-(OH)₂D₃ causes an increase in PGE₁ and PGE₂ production by growth zone cells, whereas 24R,25-(OH)₂D₃ causes a reduction in PGE₁ and PGE₂ production by resting zone cells (Schwartz et al., 1992b). The opposite effects of the two metabolites on PLA₂ activity and subsequent metabolism of arachidonic acid are, again, remarkable. Even when purified PLA₂ is incubated directly with the metabolites, 1α,25-(OH)₂D₃ stimulates the enzyme and 24R,25-(OH)₂D₃ inhibits activity (Swain et al., 1992). This suggests a major role for PLA₂ in the differential responses of the two cell types.

The differential effects of the metabolites on PKC involve the cell-specific modulation of PLA₂ by the target metabolite, thereby changing the amount of endogenous arachidonic acid. Results from studies with activators to increase PLA₂ activity and inhibitors to decrease PLA₂ activity support this hypothesis. Stimulation of PLA₂ with mellitin causes a small decrease in PKC in control cultures of resting zone cells and blocks the effects of 24R,25-(OH)₂D₃ on PKC activity (Fig. 5) (Boyan et al., 1997b). When growth zone cells are treated with mellitin, PKC in control cultures is increased, and these effects are synergistic with the increase in activity caused by 1α,25-(OH)₂D₃. Inhibition of PLA₂ activity has the opposite effect on the cells. In resting zone chondrocytes, PKC activity is stimulated, and in growth zone cells, PKC activity is decreased.

The products of PLA₂ action on phospholipids also exert regulatory effects on the cells. Arachidonic acid causes an increase in PKC activity in growth zone cells and a decrease in PKC activity in resting zone cells. Arachidonic acid can act as a secondary messenger by binding to peroxisome proliferator activator receptors (PPARs) in the nucleus, resulting in gene transcription (Bocos et al., 1995; Lin et al., 1999; Tessier-Prigent et al., 1999; Bonazzi et al., 2000).

In addition, arachidonic acid can be further metabolized as indicated above. The effects of both 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃ on PKC activity appear to be mediated by prostaglandin rather than by leukotriene production, since inhibition of lipoxygenase does not block the effects of either vitamin D₃ metabolite on PKC (Boyan et al., 1999). However, inhibition of cyclooxygenase-1 (Cox-1) with either resveratrol (Schwartz et al., 2000) or indomethacin (Helm et al., 1996) prevents the action of 24R,25-(OH)₂D₃ on PKC activity in resting zone cells. Analysis of preliminary data indicates that inhibition of Cox-1 also prevents the action of 1α,25-(OH)₂D₃ on PKC in growth zone cells. In contrast, inhibition of Cox-2 with NS-398 (Schwartz et al., 2000) has no effect. Since Cox-1 is constitutively expressed in both chondrocytes, these observations suggest that modulation of PLA₂ activity is the rate-limiting step in the responses of resting zone cells to 24R,25-(OH)₂D₃ and of growth zone cells to 1α,25-(OH)₂D₃.

Once the prostaglandin is produced, it, too, has differential effects on the cells that contribute to the specific responses of resting zone and growth zone chondrocytes to their target metabolite. PGE₂ exerts its effects on the cells via a class of membrane receptors consisting of four isoforms, EP1-EP4 (Negishi et al., 1995). Resting zone and growth zone cells express EP1 and EP2, as well as a variant of EP1, EP1v (Okuda-Ashitaka et al., 1996; Del Toro et al., 2000). The physiological effects of PGE₂ in resting zone cells are mediated by both EP1 and EP2. Similarly, the responses of the cells to 24R,25-(OH)₂D₃ are mediated by both receptors. EP1 regulates PKC and alkaline phosphatase activities, whereas both receptors mediate the effects of 24R,25-(OH)₂D₃ on proteoglycan production. Exogenous PGE₂ inhibits PKC, but stimulates alkaline phosphatase (Schwartz et al., 1998), raising the issue of how these two apparently discordant responses can be regulated by the same receptor in a positive manner by 24R,25-(OH)₂D₃. The action of PGE₂ on its receptor results in increased cAMP production and consequent inhibition of PKC (Helm et al., 1996). Since 24R,25-(OH)₂D₃ causes a decrease in PGE₂ production, less PGE₂ is available to stimulate the EP1 receptor, reducing the inhibitory effect on PKC, thereby increasing the PKC-dependent stimulation of alkaline phosphatase (Fig. 6). Thus, the cross-talk between the pathways is critical.

In growth zone cells, only the EP1 receptor is involved in the action of 1α,25-(OH)₂D₃. 1α,25-(OH)₂D₃ stimulates PGE₂ production. Unlike the decrease in PKC noted in resting zone cells, PGE₂ acts back on EP1 to synergistically increase PKC and alkaline phosphatase activities in growth zone cells. The increase in cAMP due to the action of PGE₂ on EP1 has an inhibitory action on PKC. However, it also regulates the phosphorylation of MAP kinase (Kurino et al., 1996), providing another pathway to the nucleus and gene transcription, and this overrides the inhibition.

Summary

These studies have shown that the effects of the vitamin D₃ metabolites on growth plate chondrocytes involve traditional 1,25-nVDR-mediated mechanisms as well as membrane-associated signal transduction pathways. Resting zone chondrocytes are target cells for 24R,25-(OH)₂D₃, whereas growth zone cells are regulated by 1α,25-(OH)₂D₃. The specificity of the cell response is due to several factors, including the expression of specific membrane receptors for each metabolite and the differential regulation of phospholipid metabolism. 1α,25-(OH)₂D₃ mediates its rapid effects through PLC-dependent activation of PKC, as well as through the PLA₂ signal transduction pathway. 24R,25-(OH)₂D₃ mediates its effects through PLD-dependent activation of PKC, and through down-regulation of the PLA₂ signal transduction pathway. It is likely that PKC and PLA₂ modulate both rapid and delayed physiological responses to the vitamin D₃ metabolites, in part through MAP kinase-dependent regulation of gene transcription. Regulation of 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ production provides a mechanism for cellular control of extracellular matrix events through the direct action of the metabolites on matrix vesicle membrane receptors. Differential distribution of PKC isoforms between the cell membrane and the matrix vesicle membrane helps ensure that the extracellular organelle is regulated separately from the cell. Thus, 24R,25-(OH)₂D₃ and 1α,25-(OH)₂D₃ contribute to the overall homeostasis of the growth plate via endocrine mechanisms and promote appropriate differentiation and maturation of the cells in the growth plate via autocrine pathways.

Figure 1.
Neutral metalloproteinase (MMP) and collagenase content of growth plate cartilage from normal and vitamin D/phosphate-deficient (-VDP) rats. At the animals' death, the proximal tibial growth plate cartilage was removed, and neutral MMP and collagenase were extracted and then assayed on aggrecan and type 1 collagen substrates. All values are the mean ± SEM in enzyme units/gram wet weight tissue for n ≥ 7 samples. P < 0.05, normal vs. –VDP.

Figure 2.
Effect of antibody specific for the 1α,25-(OH)₂D₃ membrane receptor (Ab99) on protein kinase C (PKC) activity of growth zone chondrocytes treated with 1α,25-(OH)₂D₃ or analogue 3a. Confluent, fourth-passage growth zone chondrocytes were treated for 9 min with control media, 10^-8 M 1α,25-(OH)₂D₃ (left panel) or analogue 3a (right panel) in the absence or presence of Ab99 or rabbit immunoglobulin G1. At harvest, PKC specific activity in the cell layer was determined. Data represent the mean ± SEM of 6 cultures from one of two experiments yielding comparable results. P < 0.05, treatment vs. control; •p < 0.05, Ab99 vs. no Ab99.

Figure 3.
Effect of inhibiting PLC on protein kinase C (PKC) activity in resting zone (RC) and growth zone (GC) chondrocytes. Confluent, fourth-passage RC cells were treated with 10^-7 M 24R,25-(OH)₂D₃ for 90 min (left panel), and GC cells were treated for 9 min with 10^-8 M 1α,25-(OH)₂D₃ (right panel) in the presence or absence of the phospholipase C inhibitor U73122 (10 μM). At harvest, PKC specific activity in the cell layer was determined. Data represent the mean ± SEM of 6 cultures from one of two experiments yielding comparable results. P < 0.05, treatment vs. control; •p < 0.05, U73122 vs. no U73122.

Figure 4.
Effect of inhibiting PLD on protein kinase C (PKC) activity in resting zone (RC) and growth zone (GC) chondrocytes. Confluent, fourth-passage RC cells were treated with 10^-7 M 24R,25-(OH)₂D₃ for 90 min (left panel), and GC cells were treated for 9 min with 10^-8 M 1α,25-(OH)₂D₃ (right panel) in the presence or absence of the phospholipase D inhibitor wortmannin (10 μM). At harvest, PKC specific activity in the cell layer was determined. Data represent the mean ± SEM of 6 cultures from one of two experiments yielding comparable results. P < 0.05, treatment vs. control; •p < 0.05, wortmannin vs. no wortmannin.

Figure 5.
Effect of activating PLA₂ on protein kinase C (PKC) activity in resting zone (RC) and growth zone (GC) chondrocytes. Confluent, fourth-passage RC cells were treated with 10^-7 M 24,25-(OH)₂D₃ for 90 min (left panel), and GC cells were treated for 9 min with 10^-8 M 1,25-(OH)₂D₃ (right panel) in the presence or absence of the phospholipase A₂ activator melittin (0.3 μg/mL). At harvest, PKC specific activity in the cell layer was determined. Data represent the mean ± SEM of 6 cultures from one of two experiments yielding comparable results. *P < 0.05, treatment vs. control; •p < 0.05, melittin vs. no melittin.

Figure 6.
Mechanism of action of vitamin D₃ metabolites. Proposed pathways of action of 1,25-(OH)₂D₃ in growth zone chondrocytes (GC) and 24,25-(OH)₂D₃ in resting zone chondrocytes (RC). Phospholipase C (PLC), Phospholipase A₂ (PLA₂), arachidonic acid (AA), cyclooxygenase-1 (Cox-1), PGE₂ EP receptor (EP), phospholipase D (PLD), protein kinase C (PKC), protein kinase A (PKA).

Footnotes

Acknowledgements

The authors acknowledge the contributions of their collaborators, Drs. Ilka Nemere (Utah State University, Logan, UT), Gary Posner (Johns Hopkins University, Baltimore, MD), and Anthony Norman (University of California, Riverside), to the studies described in this review. They also acknowledge the efforts of the students, fellows, and staff who made these studies possible. The work was supported by US PHS grants DE-05937 and DE-08603, and by the Center for the Enhancement of the Biology/Biomaterials Interface at the University of Texas Health Science Center at San Antonio.

References

Anderson HC, Sajdera SW (1976). Calcification of rachitic cartilage to study matrix vesicle function. Fed Proc 35:148–153.

Atkin I, Pita JC, Ornoy A, Agundez A, Castiglione G, Howell DS (1985). Effects of vitamin D metabolites on healing of low phosphate, vitamin D-deficient induced rickets in rats. Bone 6:113–123.

Bocos C, Gottlicher M, Gearing K, Banner C, Enmark E, Teboul M, et al. (1995). Fatty acid activation of peroxisome proliferator-activated receptor (PPAR). J Steroid Biochem Molec Biol 53:467–473.

Bonazzi A, Mastyugin V, Mieyal PA, Dunn MW, Laniado-Schwartzman M (2000). Regulation of cyclooxygenase-2 by hypoxia and peroxisome proliferators in the corneal epithelium. J Biol Chem 275:2837–2844.

Boyan BD, Ritter NM (1984). Proteolipid-lipid relationships in normal and vitamin D-deficient chick cartilage. Calcif Tissue Int 36:332–337.

Boyan BD, Schwartz Z, Swain LD, Carnes DL Jr, Zislis T (1988a). Differential expression of phenotype by resting zone and growth region costochondral chondrocytes in vitro. Bone 9:185–194.

Boyan BD, Schwartz Z, Carnes DL Jr, Ramirez V (1988b). The effects of vitamin D metabolites on the plasma and matrix vesicle membranes of growth and resting cartilage cells in vitro. Endocrinology 122:2851–2860.

Boyan BD, Schwartz Z, Swain LD (1992). In vitro studies on the regulation of endochondral ossification by vitamin D. Crit Rev Oral Biol Med 3:15–30.

Boyan BD, Dean DD, Sylvia VL, Schwartz Z (1997a). Cartilage and vitamin D: genomic and nongenomic regulation by 1,25-(OH)₂D₃ and 24,25-(OH)₂D_3. In: Vitamin D. Feldman D, Glorieux FH, Pike JW, editors. San Diego, CA: Academic Press, pp. 395-421.

10.

Boyan BD, Sylvia VL, Dean DD, Schwartz Z (1997b). Effects of 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ on protein kinase C in chondrocytes are mediated by phospholipase A₂ and arachidonic acid. In: Vitamin D: chemistry, biology, and clinical applications of the steroid hormone. Norman AW, Bouillon R, Thomasset M, editors. Riverside, CA: University of California at Riverside, pp. 353-360.

11.

Boyan BD, Posner GH, Greising DM, White MC, Sylvia VL, Dean DD, et al. (1997c). Hybrid structural analogues of 1,25-(OH)₂D₃ regulate chondrocyte proliferation and proteoglycan production as well as protein kinase C through a nongenomic pathway. J Cell Biochem 66:457–470.

12.

Boyan BD, Sylvia VL, Dean DD, Pedrozo H, Del Toro F, Nemere I, et al. (1999). 1,25-(OH)₂D₃ modulates growth plate chondrocytes via membrane receptor-mediated protein kinase C by a mechanism that involves changes in phospholipid metabolism and the action of arachidonic acid and PGE₂. Steroids 64:129–136.

13.

Boyan BD, Sylvia VL, Dean DD, Schwartz Z (2001). 24,25-(OH)₂D₃ regulates cartilage and bone via autocrine and endocrine mechanisms. Steroids 66:363–374.

14.

Carrasco-Marin E, Alvarez-Dominguez C, Leyva-Cobian F (1994). Wortmannin, an inhibitor of phospholipase D activation, selectively blocks major histocompatibility complex class II-restricted antigen presentation. Eur J Immunol 24:2031–2039.

15.

Cecil RNA, Anderson HC (1978). Freeze-fracture studies of matrix vesicle calcification in epiphyseal growth plate. Metab Bone Dis Rel Res 1:89–95.

16.

Cobb MH (1999). MAP kinase pathways. Prog Biophys Molec Biol 71:479–500.

17.

Corvol M, Ulmann A, Garabedian M (1980). Specific nuclear uptake of 24,25-dihydroxycholecalciferol, a vitamin D₃ metabolite biologically active in cartilage. FEBS Lett 116:273–276.

18.

de Boland AR, Norman AW (1998). 1α,25-(OH)₂D₃ signaling in chick enterocytes: enhancement of tyrosine phosphorylation and rapid stimulation of mitogen-activated protein (MAP) kinase. J Cell Biochem 69:470–482.

19.

Dean DD, Muniz OE, Berman I, Pita JC, Carreno MR, Woessner JF Jr, et al. (1985). Localization of collagenase in the growth plate of rachitic rats. J Clin Invest 76:716–722.

20.

Dean DD, Muniz OE, Howell DS (1989). Association of collagenase and tissue inhibitor of metalloproteinases (TIMP) with hypertrophic cell enlargement in the growth plate. Matrix 9:366–375.

21.

Dean DD, Schwartz Z, Muniz OE, Gomez R, Swain LD, Howell DS, et al. (1992). Matrix vesicles are enriched in metalloproteinases that degrade proteoglycans. Calcif Tissue Int 50:342–349.

22.

Dean DD, Schwartz Z, Muniz OE, Carreno MR, Maeda S, Howell DS, et al. (2000). Vitamin D metabolites regulate cell maturation and matrix metalloproteinase activity in growth plate cartilage in vivo (abstract). Trans Orthop Res Soc 25:985.

23.

Dean DD, Schwartz Z, Muniz OE, Carreno MR, Maeda S, Howell DS, et al. (2001). Effect of 1α,25-(OH)₂D₃ and 24R,25-(OH)₂ D₃ on metalloproteinase activity and cell maturation in growth plate cartilage in vivo. Endocrine 14:311–323.

24.

Del Toro F Jr, Sylvia VL, Schubkegel SR, Campos R, Dean DD, Boyan BD, et al. (2000). Characterization of PGE₂ receptors (EP) and their role in 24,25-(OH)₂D₃-mediated effects on resting zone chondrocytes. J Cell Physiol 182:196–208.

25.

Exton JH, Taylor SJ, Augert G, Bocckino SB (1991). Cell signalling through phospholipid breakdown. Molec Cell Biochem 104:81–86.

26.

Farach-Carson MC, Sergeev IN, Norman AW (1991). Nongenomic actions of 1,25-dihydroxyvitamin D₃ in rat osteosarcoma cells: structure-function studies using ligand analogs. Endocrinology 129:1876–1884.

27.

Farach-Carson MC, Abe J, Nishii Y, Khoury R, Wright GC, Norman AW (1993). 22-Oxacalcitriol: dissection of 1,25(OH)₂D₃ receptor-mediated and Ca⁺² entry-stimulating pathways. Am J Physiol 265:F705–F711.

28.

Fletcher JR (1993). Eicosanoids. Critical agents in the physiological process and cellular injury. Arch Surg 128:1192–1196.

29.

Franson RC, Harris LK, Ghosh SS, Rosenthal MD (1992). Sphingolipid metabolism and signal transduction: inhibition of in vitro phospholipase activity by sphingosine. Biochim Biophys Acta 1136:169–174.

30.

Garabedian M, DuBois MB, Corvol MT, Pezant E, Balsan S (1978). Vitamin D and cartilage. I. In vitro metabolism of 25-hydroxycholecalciferol by cartilage. Endocrinology 102:1262–1268.

31.

Greising DM, Schwartz Z, Posner GH, Sylvia VL, Dean DD, Boyan BD (1997). A-ring analogues of 1,25-(OH)₂D₃ with low affinity for the vitamin D receptor modulate chondrocytes via membrane effects that are dependent on cell maturation. J Cell Physiol 171:357–367.

32.

Hale JE, Wuthier RE (1987). The mechanism of matrix vesicle formation. Studies on the composition of chondrocyte microvilli and on the effects of microfilament-perturbing agents on cellular vesiculation. J Biol Chem 262:1916–1925.

33.

Helm SH, Sylvia VL, Harmon T, Dean DD, Boyan BD, Schwartz Z (1996). 24,25-(OH)₂D₃ regulates protein kinase C through two distinct phospholipid-dependent mechanisms. J Cell Physiol 169:509–521.

34.

Holtrop ME, Cox KA, Carnes DL, Holick MF (1986). Effects of serum calcium and phosphorus on skeletal mineralization in vitamin D-deficient rats. Am J Physiol 251:E234–E240.

35.

Howell DS, Dean DD (1992). Biology, chemistry and biochemistry of the mammalian growth plate. In: Disorders of bone and mineral metabolism. Coe FL, Favus MJ, editors. New York: Raven Press, Ltd., pp. 313-353.

36.

Howell DS, Blanco LN, Pita JC (1978). Further characterization of a nucleational agent in hypertrophic cell extracellular cartilage fluid. Metab Bone Dis Rel Res 1:155–160.

37.

Kato A, Seo E-G, Einhorn TA, Bishop JE, Norman AW (1998). Studies on 24R,25-dihydroxyvitamin D₃: evidence for a nonnuclear membrane receptor in the chick tibial fracture-healing callus. Bone 23:141–146.

38.

Klaus G, Von Eichel BV, May T, Hugel U, Mayer H, Ritz E, et al. (1994). Synergistic effects of parathyroid hormone and 1,25-dihydroxyvitamin D₃ on proliferation and vitamin D receptor expression of rat growth cartilage cells. Endocrinology 135:1307–1315.

39.

Kurino M, Fukunaga K, Ushio Y, Miyamoto E (1996). Cyclic AMP inhibits activation of mitogen-activated protein kinase and cell proliferation in response to growth factors in cultured rat cortical astrocytes. J Neurochem 67:2246–2255.

40.

Langston GG, Swain LD, Schwartz Z, Del Toro F, Gomez R, Boyan BD (1990). Effect of 1,25(OH)₂D₃ and 24,25(OH)₂D₃ on calcium ion fluxes in costochondral chondrocyte cultures. Calcif Tissue Int 47:230–236.

41.

Lapetina EG, Reep B, Ganong BR, Bell RM (1985). Exogenous sn-1,2-diacylglycerols containing saturated fatty acids function as bioregulators of protein kinase C in human platelets. J Biol Chem 260:1358–1361.

42.

Lidor C, Atkin I, Ornoy A, Dekel S, Edelstein S (1987a). Healing of rachitic lesions in chicks by 24R,25-dihydroxycholecalciferol administered locally into bone. J Bone Miner Res 2:91–98.

43.

Lidor C, Dekel S, Hallel T, Edelstein S (1987b). Levels of active metabolites of vitamin D₃ in the callus of fracture repair in chicks. J Bone Jt Surg 69(B):132–136.

44.

Lidor C, Dekel S, Edelstein S (1987c). The metabolism of vitamin-D₃ during fracture-healing in chicks. Endocrinology 120:389–393.

45.

Lieberherr M, Grosse B, Duchambon P, Drueke T (1989). A functional cell surface type receptor is required for the early action of 1,25-dihydroxyvitamin D₃ on the phosphoinositide metabolism in rat enterocytes. J Biol Chem 264:20403–20406.

46.

Lin Q, Ruuska SE, Shaw NS, Dong D, Noy N (1999). Ligand selectivity of the peroxisome proliferator-activated receptor alpha. Biochemistry 38:185–190.

47.

Liyanage M, Frith D, Livneh E, Stabel S (1992). Protein kinase C group B members PKC-delta,-epsilon,-zeta and PKC-L(eta). Comparison of properties of recombinant proteins in vitro and in vivo. Biochem J 283:781–787.

48.

Maeda S, Dean DD, Sylvia VL, Luna MH, Boyan BD, Schwartz Z (2000a). Matrix metalloproteinase activity in chondrocytes is differentially regulated by 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ and mediated through protein kinase C (abstract). Trans Orthop Res Soc 25:996.

49.

Maeda S, Dean DD, Sylvia VL, Boyan BD, Schwartz Z (2000b). Metalloproteinase activity in growth plate chondrocyte cultures is regulated by 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ and mediated through protein kinase C. Matrix Biol 20:87–97.

50.

Mansfield K, Rajpurohit R, Shapiro IM (1999). Extracellular phosphate ions cause apoptosis of terminally differentiated epiphyseal chondrocytes. J Cell Physiol 179:276–286.

51.

Martin TW (1988). Formation of diacylglycerol by a phospholipase D-phosphatidate phosphatase pathway specific for phosphatidylcholine in endothelial cells. Biochim Biophys Acta 962:282–296.

52.

Miao D, Bai X, Panda DK, Karaplis AC, Goltzman D, McKee MD (2000). Cartilage abnormalities are associated with abnormal phex expression and with altered matrix protein and MMP-2 expression in Hyp mice (abstract). J Bone Miner Res 15(Suppl 1):S474.

53.

Mirsky N, Silbermann M (1984). Studies on hormonal regulation of the growth of the craniofacial skeleton: III. Effects of 24,25(OH)₂D₃ on cartilage growth in the mandibular condyle of suckling mice. J Craniofac Genet Dev Biol 4:303–320.

54.

Muller-Decker K (1989). Interruption of TPA-induced signals by an antiviral and antitumoral xanthate compound: inhibition of a phospholipase C-type reaction. Biochem Biophys Res Commun 162:198–205.

55.

Negishi M, Sugimoto Y, Ichikawa A (1995). Molecular mechanisms of diverse actions of prostanoid receptors. Biochim Biophys Acta 1259:109–120.

56.

Nemere I, Dormanen MC, Hammond MW, Okamura WH, Norman AW (1994). Identification of a specific binding for 1,25-dihydroxyvitamin D₃ in basal-lateral membranes of chick intestinal epithelium and relationship to transcaltachia. J Biol Chem 269:23750–23756.

57.

Nemere I, Schwartz Z, Pedrozo H, Sylvia VL, Dean DD, Boyan BD (1998). Identification of a membrane receptor for 1,25-dihydroxy vitamin D₃ which mediates rapid activation of protein kinase C. J Bone Miner Res 13:1353–1359.

58.

Newton AC (1995). Protein kinase C: structure, function and regulation. J Biol Chem 270:28495–28498.

59.

Norman AW (1997). Rapid biological responses mediated by 1α,25-dihydroxyvitamin D₃: a case study of transcaltachia. In: Vitamin D. Feldman D, Glorieux FH, Pike JW, editors. San Diego, CA: Academic Press, pp. 233-256.

60.

Norman AW, Nemere I, Muralidharan KR, Okamura WH (1992). 1 beta, 25 (OH)₂-vitamin D₃ is an antagonist of 1 alpha,25 (OH)₂- vitamin D₃ stimulated transcaltachia (the rapid hormonal stimulation of intestinal calcium transport). Biochem Biophys Res Commun 189:1450–1456.

61.

O'Keefe RJ, Crabb ID, Puzas JE, Rosier RN (1989). Countercurrent centrifugal elutriation: high-resolution method for the separation of growth-plate chondrocytes. J Bone Jt Surg 71(A):607–620.

62.

Okuda-Ashitaka E, Sakamoto K, Ezashi T, Miwa K, Ito S, Hayaishi O (1996). Suppression of prostaglandin E receptor signaling by the variant form of EP1 subtype. J Biol Chem 49:31255–31261.

63.

Ornoy A, Goodwin D, Noff D, Edelstein S (1978). 24,25-dihydroxyvitamin D₃ is a metabolite of vitamin D essential for bone formation. Nature 276:517–520.

64.

Pedrozo HA, Boyan BD, Mazock J, Dean DD, Gomez R, Schwartz Z (1999a). TGF-β1 regulates 25-hydroxyvitamin D₃ 1α- and 24-hydroxylase activity in cultured growth plate chondrocytes in a maturation-dependent manner. Calcif Tissue Int 64:50–56.

65.

Pedrozo HA, Schwartz Z, Rimes S, Sylvia VL, Nemere I, Posner GH, et al. (1999b). Physiological importance of the 1,25-(OH)₂D₃ membrane receptor and evidence for a membrane receptor specific for 24,25-(OH)₂D₃. J Bone Miner Res 14:856–867.

66.

Peress NS, Anderson HC, Sajdera SW (1974). The lipids of matrix vesicles from bovine fetal epiphyseal cartilage. Calcif Tissue Res 14:275–281.

67.

Schwartz Z, Boyan BD (1988). The effects of vitamin D metabolites on phospholipase A₂ activity of growth zone and resting zone cartilage cells in vitro. Endocrinology 122:2191–2198.

68.

Schwartz Z, Soskolne WA, Atkin I, Goldstein M, Ornoy A (1989). A direct effect of 24,25-(OH)₂D₃ and 1,25-(OH)₂D₃ on the modeling of fetal mice long bones in vitro. J Bone Miner Res 4:157–163.

69.

Schwartz Z, Swain LD, Ramirez V, Boyan BD (1990). Regulation of arachidonic acid turnover by 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ in growth zone and resting zone chondrocyte cultures. Biochim Biophys Acta 1027:278–286.

70.

Schwartz Z, Langston GG, Swain LD, Boyan BD (1991). Inhibition of 1,25-(OH)₂D₃- and 24,25-(OH)₂D₃-dependent stimulation of alkaline phosphatase activity by A23187 suggests a role for calcium in the mechanism of vitamin D regulation of chondrocyte cultures. J Bone Miner Res 6:709–718.

71.

Schwartz Z, Brooks BP, Swain LD, Del Toro F, Norman AW, Boyan BD (1992a). Production of 1,25-dihydroxyvitamin D₃ and 24,25-dihydroxyvitamin D₃ by growth zone and resting zone chondrocytes is dependent on cell maturation and is regulated by hormones and growth factors. Endocrinology 130:2495–2504.

72.

Schwartz Z, Swain LD, Kelly DW, Brooks BP, Boyan BD (1992b). Regulation of prostaglandin E₂ production by vitamin D metabolites in growth zone and resting zone chondrocyte cultures is dependent on cell maturation. Bone 13:395–401.

73.

Schwartz Z, Dean DD, Walton JK, Brooks BP, Boyan BD (1995). Treatment of resting zone chondrocytes with 24,25-dihydroxyvitamin D₃ [24,25-(OH)₂D₃] induces differentiation into a 1,25-(OH)₂D₃-responsive phenotype characteristic of growth zone chondrocytes. Endocrinology 136:402–411.

74.

Schwartz Z, Posner GH, White MC, Greising DM, Sylvia VL, Dean DD, et al. (1997). Hybrid analogs of 1,25-(OH)₂D₃ modified on both the A-ring and C,D-ring side chain elicit rapid membrane-mediated effects on chondrocytes. In: Vitamin D: chemistry, biology, and clinical applications of the steroid hormone. Norman AW, Bouillon R, Thomasset M, editors. Riverside, CA: University of California at Riverside, pp. 381-382.

75.

Schwartz Z, Gilley RM, Sylvia VL, Dean DD, Boyan BD (1998). The effect of prostaglandin E₂ on costochondral chondrocyte differentiation is mediated by cAMP and protein kinase C. Endocrinology 139:1825–1834.

76.

Schwartz Z, Sylvia VL, Del Toro F, Hardin RR, Dean DD, Boyan BD (2000). 24R,25-(OH)₂D₃ mediates its membrane receptor-dependent effects on protein kinase C and alkaline phosphatase via phospholipase A₂ and cyclooxygenase-1 (Cox-1) but not Cox-2 in growth plate chondrocytes. J Cell Physiol 182:390–401.

77.

Schwartz Z, Pedrozo HA, Sylvia VL, Gomez R, Dean DD, Boyan BD (2002). 1α,25(OH)₂D₃ regulates 25-hydroxyvitamin D₃ 24R hydroxylase activity in growth zone costochondral growth plate chondrocytes via protein kinase C. Calcif Tissue Int (in press).

78.

Seo EG, Schwartz Z, Dean DD, Norman AW, Boyan BD (1996). Preferential accumulation in vivo of 24R,25-dihydroxyvitamin D₃ in growth plate cartilage of rats. Endocrine 5:147–155.

79.

Simboli-Campbell M, Gagnon A, Franks DJ, Welsh J (1994). 1,25-dihydroxyvitamin D₃ translocates protein kinase C beta to nucleus and enhances plasma membrane association of protein kinase C alpha in renal epithelial cells. J Biol Che m 269:3257–3264.

80.

Simon DR, Berman I, Howell DS (1973). Relationship of extracellular matrix vesicles to calcification in normal and healing rachitic epiphyseal cartilage. Anat Rec 176:167–179.

81.

Somjen D, Earon Y, Harel S, Shimshoni Z, Weisman Y, Harell A, et al. (1987). Developmental changes in responsiveness to vitamin D metabolites. J Steroid Biochem 27:807–813.

82.

St-Arnaud R (1999). Targeted inactivation of vitamin D hydroxylases in mice. Bone 25:127–129.

83.

Swain LD, Schwartz Z, Boyan BD (1992). 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ regulation of arachidonic acid turnover in chondrocyte cultures is cell maturation-specific and may involve direct effects on phospholipase A₂. Biochim Biophys Acta 1136:45–51.

84.

Swain LD, Schwartz Z, Caulfield K, Brooks BP, Boyan BD (1993). Nongenomic regulation of chondrocyte membrane fluidity by 1,25-(OH)₂D₃ and 24,25-(OH)₂D₃ is dependent on cell maturation. Bone 14:609–617.

85.

Sylvia VL, Schwartz Z, Schuman L, Morgan RT, Mackey S, Gomez R, et al. (1993). Maturation-dependent regulation of protein kinase C activity by vitamin D₃ metabolites in chondrocyte cultures. J Cell Physiol 157:271–278.

86.

Sylvia VL, Schwartz Z, Ellis EB, Helm SH, Gomez R, Dean DD, et al. (1996). Nongenomic regulation of protein kinase C isoforms by the vitamin D metabolites 1α,25-(OH)₂D₃ and 24R,25-(OH)₂D₃. J Cell Physiol 167:380–393.

87.

Sylvia VL, Schwartz Z, Curry DB, Chang Z, Dean DD, Boyan BD (1998). 1,25-(OH)₂D₃ regulates protein kinase C activity through two phospholipid-independent pathways involving phospholipase A₂ and phospholipase C in growth zone chondrocytes. J Bone Miner Res 13:559–569.

88.

Sylvia VL, Schwartz Z, Del Toro F, DeVeau P, Whetstone R, Dean DD, et al. (2001). 24R,25-(OH)₂D₃ regulates phospholipase D2 (PLD2) activity of costochondral chondrocytes in a metabolite specific and cell maturation dependent manner. Biochim Biophys Acta 1490:209–221.

89.

Takagi M, Parmley RT, Denys FR (1981). Endocytic activity and ultrastructural cytochemistry of lysosome-related organelles in epiphyseal chondrocytes. J Ultrastruct Res 74:69–82.

90.

Tessier-Prigent A, Willems R, Lagarde M, Garrone R, Cohen H (1999). Arachidonic acid induces differentiation of uterine stromal to decidual cells. Eur J Cell Biol 78:398–406.

91.

Wuthier RE (1973). The role of phospholipids in biological calcification: distribution of phospholipase activity in calcifying epiphyseal cartilage. Clin Orthop Rel Res 90:191–200.