Down-regulation of L-type Ca2+ channel transcript levels by 1,25-dihyroxyvitamin D3. Osteoblastic cells express L-type alpha1C Ca2+ channel isoforms.

Osteoblast Ca2+ channels play a fundamental role in controlling intracellular and systemic Ca2+ homeostasis. A reverse transcription-polymerase chain reaction strategy was used to determine the molecular identity of voltage-sensitive calcium channels present in ROS 17/2.8 osteosarcoma cells. The amino acid sequences encoded by the two resultant PCR products matched the α1C-a and the α1C-d isoforms. The ability of 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) and structural analogs to modulate expression of voltage-sensitive calcium channel mRNA transcripts was then investigated. ROS 17/2.8 cells were cultured for 48 h in the presence of either 1,25(OH)2D3, 1,24-dihydroxy-22-ene-24-cyclopropyl D3 (analog BT) or 25-hydroxy-16-ene-23-yne-D3 (analog AT), and the levels of mRNA encoding α1C were quantitated using a competitive reverse transcription-polymerase chain reaction assay. We found that 1,25(OH)2D3 and analog BT reduced steady state levels of α1C mRNA. Conversely, the Ca2+-mobilizing analog AT did not alter steady state levels of voltage-sensitive calcium channel mRNA. Since analog BT, but not analog AT, binds and transcriptionally activates the nuclear receptor for 1,25(OH)2D3, these findings suggest that the down-regulation of voltage-sensitive calcium channel mRNA levels may involve the nuclear receptor.

The balance between osteoclastic bone resorption and osteoblastic bone formation determines skeletal mass and composition. 1,25-Dihydroxyvitamin D 3 (1,25(OH) 2 D 3 ) 1 has long been appreciated as a hormonal modulator of osteoblast function and bone remodeling. 1,25(OH) 2 D 3 , classically considered a resorptive hormone, has the paracrine effects on osteoclasts that are believed to be mediated through separate membrane and nuclear osteoblastic 1,25(OH) 2 D 3 receptor systems (1). Paradoxically, 1,25(OH) 2 D 3 has anabolic effects on osteoblasts, stimulating production of the noncollagenous matrix proteins, osteopontin and osteocalcin (2)(3)(4). The exact mechanism by which the bone-forming and -resorbing aspects of the remodel-ing process are regulated and coupled remains unclear.
1,25(OH) 2 D 3 produces a number of rapid, membrane-initiated events such as alteration of calcium currents (5), activation of protein kinases (6), and stimulation of transcaltachia (7). Along with the rapid effects, 1,25(OH) 2 D 3 influences the transcription of many genes through binding to the nuclear receptor for 1,25(OH) 2 D 3 (nVDR) (8). The majority of the genes the expression of which is influenced by 1,25(OH) 2 D 3 are upregulated after hours to days of hormone treatment, with fewer demonstrations of negative regulation (reviewed by Hannah and Norman (9)).
The development of 1,25(OH) 2 D 3 analogs with selective activities on target cells has allowed the dual activities of 1,25(OH) 2 D 3 to be dissected. We previously reported the use of analogs that preferentially stimulate either nVDR-mediated or membrane-initiated pathways and showed that Ca 2ϩ influx is not required for the up-regulation of osteopontin and osteocalcin mRNA levels (10). Analogs that bind to the nVDR but fail to activate Ca 2ϩ influx have potential clinical importance because they stimulate bone formation with reduced hypercalcemic potential relative to 1,25(OH) 2 D 3 .
L-type voltage-sensitive calcium channels (VSCCs) on the plasma membranes of diverse cell types play a critical role in calcium signaling and cellular function. Excitable tissues such as skeletal muscle (11), cardiac muscle (12), and brain (13) express high levels of VSCCs that control excitation-contraction coupling and excitation-secretion coupling and initiate calcium-signaling events involved in intracellular signal transduction. L-type VSCCs also have been identified in other tissues including kidney (14), pancreas (15), and vascular smooth muscle (16). Structurally, these highly conserved channels consist of five distinct subunits (␣ 1 , ␣ 2 , ␤, ␥, and ␦) encoded by four genes (reviewed by Catterall (17)). The ␣ 1 subunit contains the pore through which Ca 2ϩ ions enter cells, and it can function in the absence of the other subunits (18). Potential phosphorylation sites on the ␣ 1 subunit exist for protein kinase A, protein kinase C, Ca 2ϩ /calmodulin-dependent protein kinase II, casein kinase, and cGMP-dependent protein kinase (17). Skeletal and cardiac muscle VSCCs share high sequence similarity (66%), while two of the four classes of brain Ca 2ϩ channels share ϳ75% amino acid identity with skeletal muscle (19). Further diversity is generated by alternative splicing, as is the case for the cardiac muscle ␣ 1 subunit (20).
Clonal ROS 17/2.8 osteosarcoma cells possess functional Ltype VSCCs assessed using whole cell and single channel electrophysiological recording techniques (5). We previously found that these L-type VSCCs in osteoblasts were stimulated within milliseconds by 1,25(OH) 2 D 3 , a membrane-initiated event that does not require the nVDR. Activation of VSCCs was characterized by a shift in the threshold of activation toward the resting membrane potential and a prolonged open time (5). The molecular identity of the VSCCs present in ROS 17/2.8 osteo-blastic cells has not been previously investigated. Additionally, because the extent and ability of 1,25(OH) 2 D 3 to rapidly increase plasma membrane Ca 2ϩ permeability is directly related to the density of functional VSCCs, we investigated the long term ability of 1,25(OH) 2 D 3 to modulate transcript levels encoding the ␣ 1 subunit of the VSCC.
In this report, partial cDNA cloning of the L-type VSCC ␣ 1 subunit in ROS 17/2.8 osteoblastic cells is presented. Nondegenerate reverse transcription-polymerase chain reaction (RT-PCR) primers were designed to amplify and sequence the IV S3-IV S4 region of the channel, an isoform-specific region. Amplimer sequencing demonstrated that the deduced amino acid sequence of the ␣ 1 subunit is homologous to two VSCC isoforms, ␣ 1C-a and ␣ 1C-d . Furthermore, quantitative PCR was used to study the effects of 1,25(OH) 2 D 3 and analogs on the levels of L-type VSCC mRNA. It was found that 1,25(OH) 2 D 3 down-regulates the ␣ 1C transcript levels at a physiological dose (1 nM), and vitamin D analog studies suggest that this downregulation involves the nVDR.
RNA Isolation-Total RNA was extracted from growth phase ROS 17/2.8 and primary cells using guanidinium thiocyanate as described previously (1). Briefly, cells were washed twice with ice-cold balanced salt solution, dissolved in 4 M guanidinium thiocyanate, and centrifuged over a cesium chloride gradient.
RT-PCR Primer Selection for Cloning-The PCR primers were selected based on conserved regions of the ␣ 1 subunit such that nonconserved, isoform-specific sequences would be amplified during the reaction. The PCR primers were designated as PR 1/PR 2 as reported previously (20), with PR 1 being the sense primer. PR 1 and PR 2 are oligonucleotides of 25 and 26 base pairs (bp), respectively, designed to amplify a 903-bp fragment corresponding to bases 3159 -4062 of the rabbit skeletal muscle isoform (␣ 1S ), bases 3552-4455 of the rabbit cardiac isoform (␣ 1C ), and bases 4524 -5427 of the rat brain isoform (␣ 1B ), each relative to a translational start site. Detection of the neuroendocrine (␣ 1A ) isoform, if present, was also possible using PR 1/PR 2 primers.
RT-PCR Amplification of ROS 17/2.8 and Primary Osteoblast cDNA Pool and Cloning-cDNA was reverse transcribed from ROS 17/2.8 mRNA by random hexamer priming using an RNA PCR kit (Perkin Elmer, Branchburg, NJ). On completion of the reverse transcriptase reaction, primers were added, and PCR was carried out for 40 cycles as follows: 30 s at 94°C for denaturation; 1 min at 60°C for primer annealing; and 2 min at 72°C for polymerization. The amplimers were separated by electrophoresis through 1.8% agarose gels in Tris borate buffer to assess fragment size. The PR 1/PR 2 amplimer was then ligated into the PCR2000 vector (TA Cloning, San Diego, CA) using T4 ligase, and this mixture used to transform Escherichia coli (23). Plasmid DNA was isolated from the cultures using an alkaline lysis method and digested with EcoRI to confirm the presence of the desired insert (ϳ900 bp). Inserts were sequenced using the Sequenase Version 2.0 kit (U. S. Biochemical Corp.).
Quantitation of VSCC Transcript Levels by RT-PCR-For quantitation of VSCC mRNA levels, a pair of primers designed to amplify a 246-bp region of the ␣ 1 subunit was chosen, designated as 2514/2759, and used previously by Iwashima et al. (24). An internal standard cDNA for competitive quantitative PCR was generated using PCR with an upstream primer of 40 bp designed to anneal to two 20-bp regions of the ␣ 1 subunit located 100 bp apart. PCR under low stringency conditions (50°C annealing) produced a cDNA containing sequences identical with those of the target (VSCC) cDNA, each of which is recognized by both ϩ2514 and Ϫ2759 primers, thus serving as an appropriate competitor to the endogenous target. The competitor cDNA (denoted M146) and the target cDNA are easily distinguished by size. PCR reactions using 20 ng of RNA template were performed for 35 cycles in the presence of serial dilutions of M146 (0.05-5.0 fg) and 2 Ci [␣-32 P]CTP using the following conditions: 30 s at 94°C for denaturation; 1 min at 60°C for primer annealing; and 2 min at 72°C for polymerization. Products were separated by electrophoresis on 5% polyacrylamide gels which were dried and exposed to film for 24 h.

RESULTS
PCR Amplification-We selected sequences within the ␣ 1 subunit of L-type VSCCs as the target for PCR amplification based on the extensive sequence data available in the literature. This region was carefully selected to include conserved and nonconserved sequences of various isoforms including those from rabbit cardiac muscle, skeletal muscle, neuroendocrine tissue, and brain. The sequence of each primer used in this study is listed in Table I. The amplification products of the PR 1 and PR 2 primers produced from ROS 17/2.8 cells are shown in Fig. 1, which shows two bands, one of 903 bp and one of ϳ870 bp; the identity of each is discussed below.
␣ 1C Isoforms in ROS 17/2.8 Cells and Primary Calvarial Osteoblasts-The amplimers were subcloned and sequenced with the deduced amino acid sequences shown in Fig. 2. The deduced protein product of the 903-bp cDNA amplified from ROS 17/2.8 cells is similar to the rabbit cardiac L-type Ca 2ϩ channel isoform, ␣ 1C-a , and the smaller product is homologous to ␣ 1C-d , which is shorter by 11 amino acids, presumably from alternative splicing. Both products share 89% sequence identity at the nucleotide level and 100% identity at the deduced amino acid level to the respective rabbit cardiac Ca 2ϩ channel isoforms. Partial sequencing of the primary osteoblast cDNA in this same region indicated the presence of ␣ 1C VSCC which was identical with the ROS 17/2.8 sequence at the nucleotide level (data not shown). This partial sequence did not allow identification of isoforms in the spliced region but did demonstrate the presence of the cardiac isoform in primary cultures of rat calvarial osteoblasts.
Quantitation of VSSC mRNA Levels following Exposure to

1,25(OH) 2 D 3 and Analogs-Quantitative competitive RT-PCR
was utilized to quantify the mRNA levels for the ␣ 1 subunit following 48 h of exposure to 1,25(OH) 2 D 3 , vitamin D analogs or to vehicle (0.1% absolute ethanol) followed by isolation of total RNA. The primers used for PCR were designed to amplify a 246-bp VSCC cDNA (target) and a 146-bp competitor cDNA (M146). It was determined that a suitable linear range for this 35-cycle reaction lies between 10 and 80 ng of target RNA (Fig.  3A). We routinely used 20 ng of RNA and between 0.2 and 3 fg of the competitor cDNA (Fig. 3B) to compare VSCC mRNA levels in each treatment group. The densitometrically measured ratio of competitor/target can be plotted as a function of increasing amounts of competitor; thus, the slope of the regression line is inversely proportional to the mRNA levels. Treatment with both 1,25(OH) 2 D 3 and analog BT, which binds the nVDR and does not acutely activate plasma membrane Ca 2ϩ influx, lowered mRNA levels of the VSCC ␣ 1C subunit. The concentrations used for these hormones were 1 nM for 1,25(OH) 2 D 3 and 10 nM for analog BT, concentrations shown in our previous studies to induce maximal Ca 2ϩ uptake and maximal transcription of target genes, respectively (10,22). Fig. 4 demonstrates the down-regulation of VSCC transcript levels by 1,25(OH) 2 D 3 (Fig. 4A) and analog BT (Fig. 4B) plotted using linear regression analysis of scanned autoradiographs. Based on the point of equivalence, at which the ratio of target to competitor is equal to 1, the relative amounts of VSCC ␣ 1C transcript present in the 1,25(OH) 2 D 3 -and analog BTtreated cells were 48 and 44% of vehicle control, respectively. As stated in the legend to Fig. 4, these experiments were repeated 5-7 times. The r 2 values for all data points available for each treatment group are 0.96 (1,25(OH) 2 D 3 ); 0.98 (analog BT), and 0.92 (vehicle control). In contrast, analog AT, which preferentially induces maximal Ca 2ϩ uptake at 1 nM without binding to the nVDR, does not alter VSCC mRNA levels relative to control mRNA at this concentration (Fig. 5). In three additional independent experiments, the values for mRNA levels from cells treated with analog AT or vehicle control were not significantly different from each other: vehicle (competitor/target) 2.13, 1.57, 3.06, mean ϭ 2.25 Ϯ 0.75; analog AT (competitor/target) 1.63, 2.08, 2.66, mean ϭ 2.12 Ϯ 0.52. DISCUSSION We cloned a portion of the cDNA encoding the ROS 17/2.8 L-type VSCC ␣ 1 subunit and found it to encode a protein product identical with two isoforms of the cardiac VSCC, ␣ 1C-a and ␣ 1C-d in the fourth transmembrane region. The ␣ 1C isoform was also identified in primary rat calvarial osteoblasts. It is not surprising that both of these isoforms exist in osteoblastic cells, given that they are found in a variety of tissues and cells such as lung (25), aorta (16), brain (19), and fibroblasts (26). While it has been known for some time that various cells of the osteoblastic lineage contain functional Ca 2ϩ channels, the molecular nature of these channels has only recently been studied in ROS 17/2.8 cells using an RT-PCR experimental strategy (27). Barry et al. (28) extended these studies to show that UMR-106 osteo-

FIG. 1. Agarose gel of PR 1/2 RT-PCR amplimers.
RT-PCR amplification product from primers PR 1/PR 2 were separated by electrophoresis on a 1.8% agarose gel containing ethidium bromide. The primers were designed to amplify a 903-base pair product which was found (upper band) along with a smaller product (lower band). The deduced amino acid sequence of each product is shown below (see Fig. 2).

FIG. 2. Deduced amino acid sequence of ROS 17/2.8 osteoblastic L-type Ca 2؉ channel isoforms. The osteoblastic sequence (␣ 1C-a )
is 100% homologous to the rabbit cardiac isoform (␣ 1C ) in this region. Additionally, the ROS 17/2.8 VSCCs contain a 33-bp deletion which gives rise to the alternatively spliced isoform, ␣ 1C-d . The deletion is shown by a series of periods below the full length sequence in bold face. Dashes indicate matches with the ␣ 1C-a sequence, and the location of the sequence within the putative structure of the ␣ 1 subunit is shown in brackets above the sequence. TM, transmembrane.

FIG. 3. Autoradiographs of quantitative RT-PCR products.
Quantitative RT-PCR products were separated by electrophoresis on a 5% polyacrylamide gel which was dried and exposed to film for 18 h at room temperature. sarcoma cells contain skeletal (␣ 1S ), cardiac (␣ 1C ), and neuroendocrine (␣ 1D ) isoforms of the L-type Ca 2ϩ channel in a small region of the ␣ 1 cDNA. With regard to ␣ 1C , we find evidence for two cardiac isoforms, ␣ 1C-a and ␣ 1C-d , in ROS 17/2.8 cells.
Interestingly, it has been reported that when UMR-106 cells are subjected to chronic mechanical strain, they acquire a calcium conductance that responds to hypotonic swelling and can be blocked by antisense RNA to ␣ 1C (29). Thus, regulation of Ca 2ϩ channel expression in osteoblasts may be complex, being affected by hormonal and mechanical interactions as well as the stage of cell differentiation (30).
Osteoblasts are generally considered to be nonexcitable cells that share few characteristics with excitable cells in skeletal and cardiac muscle. With regard to plasma membrane Ca 2ϩ permeability, an obvious difference between osteoblasts and the excitable tissues is the total number of the functional VSCCs expressed. ROS 17/2.8 cells express between 1000 and 2000 functional L-type Ca 2ϩ channels per cell (5) while differentiated myocytes contain at least 10 times this level (31). This difference in VSCC levels is also apparent at the mRNA level, in which standard Northern blotting techniques used routinely for excitable tissues are frequently incapable of detecting low levels of VSCC mRNA in cells such as ROS 17/2.8. Given our findings that osteoblasts express VSCC isoforms common to cardiac tissues, osteoblasts may represent a "hybrid" between excitable and nonexcitable cells that may possess unique characteristics that relate to function. One such function of osteo-blasts in their bone microenvironment is the ability to regulate the secretion of various matrix proteins in response to hormonal and mechanical stimuli.
Over the past decade, 1,25(OH) 2 D 3 has emerged as an important hormone in the physiology of many tissues. While generally considered a resorptive hormone in bone, it has anabolic effects on the expression of many extracellular matrix proteins that contribute to bone mass. In fact, the majority of the bone/extracellular matrix genes under regulation by 1,25(OH) 2 D 3 are up-regulated by the seco-steroid-including genes encoding alkaline phosphatase (32), osteocalcin (4), and osteopontin (2,3). In this study, we report that the ␣ 1C Ca 2ϩ channel expressed by osteoblastic cells is down-regulated at the mRNA level following exposure to 1 nM 1,25(OH) 2 D 3 for 48 h. This finding is one of a handful of genes negatively regulated by 1,25(OH) 2 D 3 which include aggrecan proteoglycan (33), Id (34), atrial natriuretic peptide (35), and collagen type I (36). In previous studies, it was shown that UMR-106 cells treated with 1,25(OH) 2 D 3 for 8 -24 h lost responsiveness to parathyroid hormone-induced increases in intracellular Ca 2ϩ (37). Although the contribution of L-type Ca 2ϩ channels was not directly tested in their study using dihydropyridine channel blockers, it can be inferred that the channel down-regulation seen here after long term exposure to 1,25(OH) 2 D 3 may contribute to the diminished response. The data presented in this article clearly indicate that Ca 2ϩ influx is unlikely to serve as the triggering signal for Ca 2ϩ channel down-regulation, since exposure to 1 nM analog AT, which induces maximal Ca 2ϩ influx in ROS 17/2.8 cells, does not alter ␣ 1C mRNA levels. It is therefore doubtful that the Ca 2ϩ -mobilizing effects of 1,25(OH) 2 D 3 alone are responsible for the decreased levels of VSCC mRNA. The finding that analog BT alone can negatively regulate the steady state levels of the Ca 2ϩ channel transcript strongly suggests that the genomic effects of 1,25(OH) 2 D 3 mediated through the nVDR may be responsible for our observations. The question remaining is whether or not the downregulation is a transcriptional one caused by direct interaction of the hormone-receptor complex with regulatory regions of the VSCC gene or whether a distinct mechanism exists. For example, our laboratory has noted in ion transport assays that VSCCs are maximally active in growth phase cells (50 -80% confluent), suggesting that VSCC expression and/or activity can be linked to growth and proliferation, which in turn may be influenced by 1,25(OH) 2 D 3 . Experiments are under way to determine definitively whether the effects of 1,25(OH) 2 D 3 on VSCC mRNA levels require a functional nVDR.
In summary, this study provides evidence for the identity of L-type Ca 2ϩ channel isoforms (␣ 1C ) in osteoblastic cells. Furthermore, the demonstration that 1,25(OH) 2 D 3 down-regulates the steady state levels of mRNA encoding VSCCs points out the interdependence of the rapid, membrane-initiated, and  ). B, the products from cells treated with 10 ng of analog BT (E) for 48 h are compared with controls (q). The increased slope in the 1,25(OH) 2 D 3 and analog BT plots indicate decreased amounts of VSCC ␣ 1 subunit mRNA present in these samples. In each graph, data points were averaged directly from two individual experiments in which all data points were available. The error bars for these two experiments are contained within the symbols. The trend was also reproduced in five separate experiments, which were not included in this graph because all data points were not available in each series. genomic effects of this hormone. By affecting the biosynthesis of the VSCCs at a genomic level, 1,25(OH) 2 D 3 can alter membrane permeability to Ca 2ϩ and thus the capacity of the osteoblast to respond to a variety of hormones unique to bone physiology.