Molecular Characterization of the [alpha]1 ... - Wiley Online Library

2 downloads 1636 Views 414KB Size Report
Dental Research Annual Meeting, Orlando, Florida, U.S.A.,. March 1997 and the .... ALP levels were determined by a histochemical tech- nique modified to ..... P 1995 Multiple cal- cium channel transcripts in rat osteosarcoma cells: Selective.
JOURNAL OF BONE AND MINERAL RESEARCH Volume 14, Number 3, 1999 Blackwell Science, Inc. © 1999 American Society for Bone and Mineral Research

Molecular Characterization of the ␣1 Subunit of the L Type Voltage Calcium Channel Expressed in Rat Calvarial Osteoblasts* JUAN C. LOZA,1,2 LILLIAN C. CARPIO,1 PETER G. BRADFORD,1,3 and ROSEMARY DZIAK1

ABSTRACT Voltage-activated calcium channels (VACCs) regulate extracellular calcium influx in many cells. VACCs are composed of five subunits. The ␣1 subunit is considered the most important in regulating channel function. Three isoforms of this subunit have been described: skeletal, cardiac, and neuroendocrine. It was the purpose of the present study to determine the molecular identity of the ␣1 subunit of the VACCs in rat calvarial osteoblasts and to study the nature of the regulation of these channels as a function of cellular growth. We also attempted to identify which isoform of the ␣1 subunit of the VACCs mediates the effects of epidermal growth factor (EGF) on osteoblastic cell proliferation. Reverse transcription-polymerase chain reaction was used to detect the isoforms of the VACCs that are expressed in osteoblastic cells. These analyses showed that the proliferative state of the cell and the time in culture influence RNA expression. The only ␣1 subunit detected in osteoblasts corresponds to the cardiac isoform. In additional experiments, the effects of EGF on cytosolic calcium and osteoblast proliferation were determined. For these experiments, the synthesis of the different isoforms of the VACCs was selectively blocked by antisense oligonucleotides prior to EGF stimulation. These studies showed that the cardiac isoform mediates the effects of EGF on cytosolic calcium and cellular proliferation in rat calvarial osteoblasts. (J Bone Miner Res 1999;14:386–395)

INTRODUCTION

F

is an important second messenger involved in a number of cellular events, i.e., excitation, contraction, and secretion.(1) There are two sources of ionic calcium. The intracellular calcium pool is regulated by the production of inositol trisphosphate, which in turn activates the inositol trisphosphate receptor/channel, and by a mechanism called calcium-induced calcium release (which involves the ryanodine receptor/channel).(2) Calcium influx REE IONIC CALCIUM

*Parts of this work were presented at the 75th International Association of Dental Research/26th American Association of Dental Research Annual Meeting, Orlando, Florida, U.S.A., March 1997 and the 76th International Association of Dental Research Annual Meeting, Nice, France, June 1998.

from the extracellular environment constitutes the second source of ionic calcium. Membrane channels sensitive to changes in membrane potential are responsible for this influx of calcium from the extracellular milieu. Based on their electrophysiological and pharmacological characteristics, voltage-activated calcium channels (VACCs) are described to be of the L, T, N, P, Q, or R type.(3) N, P, Q, and R type VACCs have been described in neurons. T and L type VACCs have been described in skeletal and cardiac muscle, neurons, and endocrine cells. L type VACCs have been described in nonexcitable cells, and they are often considered the most physiologically relevant to study because of their ability to move calcium very effectively and their high degree of regulation.(4) A number of studies have shown that osteoblastic cells express the L type VACCs.(5–11) In our laboratory, we have reported that rat calvarial osteo-

1

Department of Oral Biology, State University of New York at Buffalo, School of Dental Medicine, Buffalo, New York, U.S.A. Department of Restorative Dentistry, State University of New York at Buffalo, School of Dental Medicine, Buffalo, New York, U.S.A. Department of Pharmacology and Toxicology, State University of New York at Buffalo, School of Medicine and Biomedical Sciences, Buffalo, New York, U.S.A. 2 3

386

ISOFORMS OF CALCIUM CHANNELS IN OSTEOBLASTS blasts express both T and L type VACCs.(12) The expression of the L type current was dependent on the time in culture and the proliferative state of the cells. The role L type VACCs play in osteoblast physiology could be complex since different cellular states have been associated with the expression of L type VACCs. In the extensive studies conducted in other systems, VACCs have been shown to be composed of five subunits: ␣1, ␣2, ␤, ␦, and ␥. Of these five subunits, ␣1 contains the pore, the gating system, and the voltage-sensing sequences. This subunit has been fully cloned and sequenced.(3) It is known that the ␣1 subunit is organized into four homologous repeat units (I–IV). Each unit is made of six transmembrane domains, connected with different length linkers. Three isoforms have been described for this subunit, depending on the tissue of origin. The isoforms can be S or skeletal, C or cardiac, and D or neuroendocrine.(3) In osteoblastic cells, the isoform expression varies depending on the cell line used. Barry et al.(13) found that rat osteosarcomal cell line UMR-106 expresses all three isoforms. However, Meszaros et al.(14,15) found that the MC 3T3 and ROS 17/2.8 osteoblastic cell lines only express two splice variants of the C isoform. Since we have previously stated that rat calvarial osteoblasts express L type VACCs depending on the time in culture and the proliferative state of the cell, we wanted to determine which isoforms of the ␣1 subunit of the VACCs were expressed in primary cultures of rat osteoblastic cells. Epidermal growth factor (EGF) is a mitogen for osteoblastic cells.(16) We found that L type VACCs mediate the effect of EGF in cytosolic calcium and in osteoblastic cell proliferation.(17,18) Permeabilization of plasma membrane with Streptolysin O allows introduction of antisense oligonucleotides into a cell with the purpose of blocking specific mRNA translation but still maintaining cell viability.(19) This technique has been used to determine the role of VACCs and stretch-activated channels in osteoblastic cell physiology.(13,20) Barry et al.(13) found that in UMR-106 cells, parathyroid hormone (PTH) induces increases in cytosolic calcium, which are mediated by VACCs. Antisense oligonucleotide inhibition showed that the PTH effect on cytosolic calcium is dependent on the D isoform of the ␣1 subunit of the channel. In order to determine which isoform mediates the EGF effects on rat calvarial osteoblasts, in the present study we blocked the expression of the different ␣1 isoforms of L type VACCs by the antisense oligonucleotide inhibition technique prior to EGF stimulation of the cells. Therefore, the objective of the present study was to determine which isoforms of the ␣1 subunit of the VACCs are expressed in normal rat calvarial osteoblasts and to determine which isoforms are responsible for mediating growth factor–induced cellular proliferation.

MATERIALS AND METHODS Cell culture All experiments were conducted on osteoblast-enriched cell preparations obtained by sequential collagenase digestion of Sprague-Dawley 21-day fetal rat calvaria.(21) It has

387

been shown that the cell populations recovered after the third and fourth 15-minute digestions have osteoblastic characteristics, such as expression of alkaline phosphatase (ALP), osteocalcin, and the ability to form collagen and mineralized nodules in culture.

Membrane potential recordings Freshly isolated rat calvarial osteoblasts were seeded and cultured as described.(18) Type I collagen-coated coverslips were placed in 6-well plates, 10,000 cells/ml, 1.5 ml/well. The cell cultures were maintained with BGJb media supplemented with 10% fetal calf serum (FCS); experiments were done 1–8 days from isolation. Prior to recording the membrane potential, and in order to obtain a more stable gigaseal, cells were placed in a calcium-free HEPES buffer (containing [in mM] 120 NaCl, 30 Mannitol, 3 K2HPO4, 1 MgSO4, 30 HEPES with the pH adjusted to 7.4 with NaOH, and supplemented with 0.1% bovine serum albumin and 0.5% glucose). This incubation partially detached the osteoblastic cell. Cells were immediately placed in an inverted microscope, where a perfusion chamber allowed exchange of buffers. Once in the chamber, the cells were washed for at least 5 minutes in a buffer containing (in mM) 130 NaCl, 10 CaCl2, 5 KCl, 1 MgCl2, 10 HEPES with the pH adjusted to 7.4 with NaOH, and 6 glucose. At this point, the membrane potential measurements were started. After a baseline recording of the membrane potential of the osteoblastic cell was obtained, the buffer was changed to a depolarizing buffer containing (in mM) 80 NaCl, 10 CaCl2, 55 KCl, 1 MgCl2, 10 HEPES with the pH adjusted to 7.4 with NaOH, and 6 glucose. All recordings were in the whole cell configuration of the patch clamp technique.(22) For this, borosilicate glass was pulled and coated with Sylgard to obtain electrodes (resistance of the electrodes varied from 2 to 8 M⍀). These electrodes were filled with a solution containing (in mM) 70 KCl, 30 K2SO4, 10 NaCl, 3 MgCl2, 2.5 ATP.Na2, 200 ␮M EGTA, 123 ␮M CaCl2, and 10 HEPES with the pH adjusted to 7.4 with NaOH. The electrode was connected to the headstage of a patch clamp amplifier, which in turn was connected to a paper tracer. After obtaining the whole cell patch clamp, the amplifier was set to the current clamp mode and current injections were delivered through the stimulator to test membrane resistance.

Cytosolic calcium measurements and depolarization-induced changes in cytosolic calcium levels All experiments were done on individual attached cells. For this purpose, freshly isolated rat calvarial osteoblasts were seeded in 6-well plates (10,000 cells/ml, 1.5 ml/well, in BGJb media supplemented with 10% FCS). Each well contained a round glass coverslip. After 48 h of incubation, the media was replaced in one third of the cultures with fresh BGJb media supplemented with 0.01% FCS, one third received BGJb media with 0.01% FCS and 1 ␮M cycloheximide (CHX), and the last third of the cultures received

388 BGJb media with 10% FCS. After 24 h of incubation, the cells were loaded with Fura-2 AM. For this manipulation, each well was washed twice with 1 ml of HEPES-buffered salt solution (HBSS) containing (in mM) 120 NaCl, 30 Mannitol, 1.25 CaCl2, 3 K2HPO4, 1 MgSO4, 30 HEPES with the pH adjusted to 7.4 with NaOH, and supplemented with 0.1% bovine serum albumin and 0.5% glucose. Following the wash, 5 ␮M Fura-2 AM in HBSS was added and the cells were incubated for 30 minutes in a 5% CO2 humidified incubator at 37°C in the absence of light. Afterward, each well was washed twice with HBSS, and then cells were transferred to a chamber in the stage of an inverted microscope (1 M 35; Zeiss, Jena, Germany), with 1 ml of a modified HBSS (above described HBSS where 1 mM MgCl2 replaced 1 mM MgSO4, and 5 mM KCl replaced 3 mM K2HPO4). A dual spectrophotometer and a photomultiplier tube (Fluorolog II; Spex Industries, Inc. Edison, NJ, U.S.A.) were connected to the microscope. Both spectrophotometer and photomultiplier tube were controlled by a PC-compatible computer and the appropriate software (DM3000 CM; Spex Industries, Inc.). The cells were excited at 340 and 380 nm wavelengths, and the emission fluorescence was recorded at 505 nm. At the end of the experiment, the ratio of fluorescence at 340 nm over fluorescence at 380 nm (R value) was obtained. This R value was used as an approximation of cytosolic calcium measurement because of the previously described problems in the calibration procedures associated with the use of Fura-2 to measure cytosolic calcium measurement.(23) Experiments were conducted for 5 minutes. Once a cell was focused and a stable baseline cytosolic calcium level was recorded, the modified HBSS was exchanged for a high potassium modified HBSS, which had 55 mM KCl instead of 5 mM, and 70 mM NaCl instead of 120 mM. This high potassium modified HBSS also contained 5 ␮M Bay K 8644, a known calcium channel agonist.

Reverse transcription-polymerase chain reaction amplification of RNA expressing VACCs Freshly isolated rat calvarial osteoblasts were seeded in 35 mm culture dishes at 750,000 cells/dish, using BGJb media supplemented with 10% FCS. After culturing the cells for 3 days, total RNA was isolated using the guanidine isothiocyanate/CsCl method. First-strand cDNA was generated using Superscript RNAse H−RT (GIBCO BRL, Grand Island, NY, U.S.A.) according to product specifications. Reverse transcriptionpolymerase chain reaction (RT-PCR) of VACCs was carried out using primers that cover regions of low degeneracy of the ␣1 subunit.(24) The coding strand primer contained a linker sequence with a HindIII restriction site, followed by nucleotides corresponding to the rabbit cardiac sequence, encoding FFMMNF, from the IIIrd repeat unit 6th transmembrane domain. The sequence is: 5⬘-CGAAGCTTCTTCATGATGAACATCTT-3⬘. The noncoding strand primer contained a linker sequence with a BamHI site, followed by nucleotides corresponding to rabbit cardiac-calcium channel sequence complementary to FISFYM of the IVth re-

LOZA ET AL. peat unit, 6th transmembrane region. The sequence is: 5⬘GCGGATCCATGTAGAAGCTGATGAA-3⬘. Primers designed to amplify a 220 bp region of rat glutaraldehyde phosphate dehydrogenase were also included in the PCR mixture. This was done for two reasons. First, amplification of a second PCR product allowed us to test the RT-PCR. Second, the band intensity of the 220 bp product served as an internal standard, against which the intensity of the band for the VACC RT-PCR product could be compared. The PCR reaction was carried out using Taq DNA polymerase (GIBCO BRL) according to product specifications. The thermocycler was programmed for 35 cycles of 94°C for 30 s, 55°C for 1 minutes 30 s, and 72°C for 1 minute. The products were run in a 1.2% low melting point agarose gel to detect PCR amplification product and to recover the product for DNA sequencing. The PCR products were cut with HindIII and BamHI, ligated with similarly cut pGEM11 plasmid (Promega Co., Madison, WI, U.S.A.), and then transformed into competent Escherichia coli DH5 ␣. The pGEM11/VACC colonies were selected, and their plasmid DNA was recovered and sequenced using the dideoxy chain termination method.

ALP measurements using flow cytometry Freshly isolated and cultured rat calvarial osteoblasts were seeded in 35 mm culture dishes at 750,000 cells/dish, using BGJb media supplemented with 10% FCS. Media was changed every third day. Cells were recovered after 3, 6, or 9 days in culture by a 4-minute trypsin digestion, washed in phosphate-buffered solution, fixed in 70% ethanol, and if necessary stored at 4°C. ALP levels were determined by a histochemical technique modified to detect the enzyme levels by flow cytometry.(32) Fixed cells were centrifuged and suspended in 0.1 M Tris-HCl containing 0.1 mg/ml of Naphtol AS-MX phosphate (first diluted in 10 ␮l of dimethyl sulfoxide) and 0.6 mg/ml of Fast Red TR salt. Cells were incubated for 30 minutes in a shaking water bath, at 37°C, in the dark. After staining, cells were centrifuged and suspended in filtered phosphate-buffered solution. A Becton Dickinson FACScan flow cytometer (Mountain View, CA, U.S.A.) controlled by CellQuest (Macintosh-based flow cytometry computer program) was used to measure the amount of red fluorescence in the fixed cells. This fluorescence is directly related to the amount of ALP. The staining is specifically blocked by preincubating the fixed cells with 10 mM orthovanadate, therefore indicating that this staining technique is specific for ALP.

Inhibition of calcium currents using antisense oligonucleotides against different isoforms of the ␣1 subunit of the VACC To determine which isoform of the ␣1 subunit of the L type VACCs mediates the effects of EGF in osteoblastic cell proliferation and cytosolic calcium levels, the expression of the different isoforms was blocked by introducing an antisense oligonucleotide into the cell.(19)

ISOFORMS OF CALCIUM CHANNELS IN OSTEOBLASTS

389

FIG. 1. High potassium induced depolarization in a representative individual, attached rat calvarial osteoblast. Membrane potential is measured by setting the patch clamp amplifier to the current clamp mode. To detect the activation of conductances, 10 pA/mV current injections were delivered through the stimulator to test membrane resistance. A large downward reflection indicated high membrane resistance, and therefore few conductances were activated. After depolarization, the deflections got smaller, thus indicating that one or more conductances got activated.

For this, freshly isolated cells were grown to confluency in 75-cm2 flasks. Cells were then trypsinized and suspended in HBSS. Cells were pelleted by centrifugation and resuspended at 1,000,000 cells/ml in HBSS containing 20 U/ml of Streptolysin O and 100 ␮M antisense or sense oligonucleotides specific for the three known isoforms of the ␣1 subunit of the L type VACCs (Skeletal, Cardiac or neuroenDocrine).(13) Cells were incubated for 10 minutes in a shaking water bath at 37°C, centrifuged, and resuspended in BGJb media with 10% FCS. For the cytosolic calcium measurements, cells were suspended at 50,000 cells/ml, and 1.5 ml of the suspension was seeded in 6-well plates, where each well contained a round coverslip. For the proliferation assay, cells were seeded in 24-well plates at 1,000,000 cells/ well. After cells had attached to the glass coverslips or the bottom of the wells for 6 h, the media was changed to BGJb with 0.01% FCS. After 12 h, the cells seeded in the 24-well plate were treated with 50 nM EGF, and cellular proliferation was assessed by a 3H-thymidine incorporation assay as described by Stephan and Dziak.(16) Concurrently with the EGF stimulation of the cells used for the proliferation assay, the cells seeded in glass coverslips were treated with Fura 2-AM as described above for cytosolic calcium measurements, and 50 nM EGF was used to stimulate focused single cells after their baseline cytosolic calcium level had been determined.

RESULTS Membrane potential measurements Initial experiments were performed to show that rat calvarial osteoblasts are capable of membrane depolarization when placed in an extracellular buffer containing high potassium. The baseline membrane potential measured in osteoblastic cells ranged from −60 mV to −30 mV, with a mean value of −46 mV recorded in 10 cells. When these cells were exposed to a depolarizing buffer, the membrane potential changed to a mean value of −20 mV, with values ranging from −36 mV to −12 mV (n ⳱ 4). A representative membrane potential trace is shown in Fig. 1.

Cytosolic calcium measurement After showing that high extracellular potassium causes membrane depolarization, we wanted to determine if an

increase in cytosolic calcium was associated with the change in membrane potential. Single rat calvarial osteoblastic cells were loaded with the calcium indicator Fura 2-AM and exposed to high potassium modified HBSS. As shown in Fig. 2, the cell responded with an increase in cytosolic calcium when challenged by a depolarizing buffer. This increase in calcium was only observed when 5 ␮M Bay K 8644 was present in the depolarizing buffer, suggesting that VACCs are mediating this response. We have previously shown by patch clamping techniques that the expression of calcium currents is dependent on the proliferative state of the cell.(12) When the serum in the media was reduced to 0.01%, the expression of the currents increased. In the following experiments, we attempted to repeat this observation, using the above described depolarization-induced increase in cytosolic calcium increase as an indirect measurement of the expression of these currents. As shown in Table 1, when cells cultured in media supplemented with 10% FCS were screened for cytosolic calcium changes, 4 out of 11 (36%) cells responded with an increase in calcium as a response to a depolarizing challenge. Nine out of 11 (81%) cells cultured in media with 0.01% FCS responded with an increase in cytosolic calcium when exposed to high potassium buffer. To study if continued protein synthesis was required for the increased percentage of responding cells, 1 ␮M CHX was added to the media supplemented with 0.01% FCS. The percentage of cells responding with an increase in calcium upon depolarization decreased to 10% (1 out of 10) in the presence of CHX. Cell cycle progression has been described to require elevated levels of cytosolic calcium. To determine if there was a relationship between cell cycle progression and expression of VACCs, three specific cell cycle blockers were added to media supplemented with 10% FCS. The effects of the different blockers used in this experiment were confirmed by flow cytometry–based cell cycle analysis, which is achieved by labeling active DNA synthesis with bromodeoxyuridine, which is detected by anti–bromodeoxyuridine-FITC as green fluorescence. Total DNA is stained with propidium iodine and it is detected as red fluorescence. Green and red fluorescence were detected using a Becton Dickinson FACScan flow cytometer controlled by CellQuest (a Macintosh-based flowcytometry computer program). Analysis of the double fluorescence labeling allowed us to determine the cell cycle position of the cells being tested. Progression through the G1 phase was blocked using 10 ␮M lovastatin, through the S phase using

390

LOZA ET AL. TABLE 2. PERCENTAGE OF RAT CALVARIAL OSTEOBLASTS RESPONDING WITH AN INCREASE IN CYTOSOLIC CALCIUM WHEN EXPOSED TO DEPOLARIZING BUFFER

FIG. 2. High potassium induced increase in cytosolic calcium (R value) measured in a representative individual, attached rat calvarial osteoblast using Fura 2-AM. High potassium depolarizes the cell and induces changes in cytosolic calcium. TABLE 1. PERCENTAGE OF RAT CALVARIAL OSTEOBLASTS RESPONDING WITH AN INCREASE IN CYTOSOLIC CALCIUM WHEN EXPOSED TO DEPOLARIZING BUFFER Serum conditions

Responding/total

Percentage

10% 0.01% 0.01% + 1␮M CHX

4/11 9/11 1/10

36% 81% 10%

Treatment conditions

Responding/total

Percentage

10% FCS 10% FCS + 0.5 ␮g/ml aphidicolin 10% FCS + 10 ␮M lovastatin

2/5

40%

4/9

44%

3/9

33%

TABLE 3. PERCENTAGE OF RAT CALVARIAL OSTEOBLASTS EXPRESSING CALCIUM CURRENTS (DETERMINED BY PATCH CLAMP TECHNIQUES DESCRIBED IN REF. 12), FOLLOWING TREATMENT WITH SPECIFIC CELL CYCLE BLOCKERS Treatment conditions

Responding/total

Percentage

10% FCS 10% FCS + 0.5 ␮g/ml aphidicolin 10% FCS + 0.5 ␮g/ml lovastatin

4/10

40%

3/6

50%

1/4

25%

0.5 ␮g/ml aphidicolin, and through the M phase using 0.5 ␮g/ml of nocodazole. None of these blockers caused significant increases in the number of cells expressing calcium currents (as determined by patch clamp technology described previously(12) or in the number of cells responding with increases in cytosolic calcium when challenged with a depolarizing buffer (see Tables 2 and 3).

RT-PCR amplification of RNA expressing VACCs and ALP levels RT-PCR technology (using PCR primers designed to matched conserved regions of the ␣1 subunit of the ␣1 type VACC) amplified the expected 900 bp when RNA recovered from rat calvarial osteoblasts was used (see Fig. 3). A number of conditions were tested to determine if the expression of the ␣1 subunit varied based on the proliferative state of the cell or the time in culture of the rat calvarial osteoblasts. In freshly isolated cells, the presence of serum in the culture media did not affect the expression of the ␣1 subunit (Fig. 3). However, subcultured cells (cells grown to confluency in 75 cm2 flasks, and then seeded and handled the same way as freshly isolated cells) showed that the expression of RNA for the ␣1 subunit of the L type VACCs was increased when the cells were grown in serum-deficient media (Fig. 4). In these experiments, PCR amplification of a 220 bp region of the housekeeping gene GADPH was used as an internal standard. Under both serum conditions, the intensity of the 220 bp bands was similar, indicating that the PCR reaction worked correctly and that the differences in the expression of the VACCs were real and not due to different RT-PCR conditions.

FIG. 3. RT-PCR amplification of VACCs from total RNA isolated from freshly isolated rat calvarial osteoblasts, which had been in culture for 3 days. Agarose gel contains 1.2% agarose.

When RNA was isolated at different times in culture, the expression of the VACCs (Fig. 5) was highest after 3 days in culture, the levels decreased significantly at 6 days in culture and increased slightly at 9 days. Again the intensity of the bands for GADPH were similar throughout the three time points screened, thus indicating that the differences were real and not due to different RT-PCR conditions. ALP levels were measured to establish possible correlations between the stage of differentiation of the cells and VACC expression levels. As shown in Fig. 6, the ALP levels were high at day 3, experiencing a decrease at day 6 in culture, and recovered at day 9. For this last time point, VACC levels did not recover as much as ALP; however, the pattern of expression for this osteoblastic marker and the VACC levels were similar as a function of time in culture. The 900 bp RT-PCR products in the above gels were

ISOFORMS OF CALCIUM CHANNELS IN OSTEOBLASTS

391

FIG. 4. RT-PCR amplification of VACCs from total RNA isolated from subcultured rat calvarial osteoblasts, which had been in culture for 3 days. Internal control is housekeeping gene GADPH. Agarose gel contains 1.8% agarose.

FIG. 5. RT-PCR amplification of VACCs from total RNA isolated from freshly isolated rat calvarial osteoblasts, which had been in culture for 3, 6, and 9 days. Internal control is housekeeping gene GADPH. Agarose gel contains 1.8% agarose.

FIG. 6. Measure of the relative of ALP levels (mean ± SD from triplicate cultures) in osteoblastic cells as a function of time in culture. Experiments were repeated at least three times; the graph is representative of one of them.

cut out of the gel, recovered, digested with BamHI and HindIII, ligated to pGEM-11 cut with the same enzymes, and transformed into DH5 ␣. Positive clones were identified by white/blue selection and ampicillin resistance. DNA sequencing (data not shown) showed sequence identity > 95% when compared with the published sequences of VACCs from cardiac origin, splice variant a.(24)

FIG. 7. EGF-induced increase in cytosolic calcium (R value) in individual osteoblastic cells attached to a glass coverslip and treated with Streptolysin O and three different antisense oligonucleotides specific for the three known isoforms of the ␣1 subunit of the L type VACCs. (A) A control cell treated with Streptolysin O only, (B) a cell treated with antisense specific for the cardiac isoform, (C) a cell treated with antisense specific for the skeletal isoform, (D) a cell treated with antisense specific for the neuroendocrine isoform.

Antisense oligonucleotide inhibition studies Subsequently, we wanted to determine which isoform of the ␣1 channel subunit mediates the EGF effects on cytosolic calcium and osteoblastic cell proliferation. Specific antisense oligonucleotides, synthesized to anneal to a specific region of the mRNA for the different isoforms of the ␣1 subunit, were introduced into the intracellular environment by permeabilizing the osteoblastic cell plasma membrane with Streptolysin O. Cell viability (determined by trypan blue staining and 3H-uridine incorporation assay; data not shown) or the response of rat calvarial osteoblasts to EGF was not affected by the permeabilization process. For these studies, parallel cultures were used. As shown in Fig. 6A, the calcium response to EGF was not affected by Streptolysin O. The effects on cell proliferation are shown in Fig. 8. Cells treated with Streptolysin O and EGF showed a significant increase in cell proliferation when compared with cells treated with Streptolysin O alone. The antisense oligonucleotide specific for the cardiac isoform of the channel subunit was the only oligonucleotide that blocked the EGF and high potassium effects on cytosolic calcium (Fig. 7B and Table 4). Similarly, cells treated with the antisense oligonucleotide to the cardiac isoform showed a partial but siginificant inhibition of EGF-induced

392

LOZA ET AL.

TABLE 4. PERCENTAGE OF RAT CALVARIAL OSTEOBLASTS (TREATED WITH DIFFERENT ANTISENSE OLIGONUCLEOTIDES) RESPONDING WITH AN INCREASE IN CYTOSOLIC CALCIUM WHEN EXPOSED TO 50 NM EGF Antisense oligonucleotide used Control Cardiac Skeletal Neuroendocrine

Responding/total

Percentage

25/53 2/18 7/18 7/17

47% 11% 39% 41%

proliferation when compared with controls or cells treated with the oligonucleotides only. Use of the antisense oligonucleotides to the other isoforms of the ␣1 subunit of the VACCs or sense oligonucleotides (data not shown) did not inhibit the calcium response (Figs. 7C and 7D) or the proliferative response to EGF (Figs. 8B and 8C).

DISCUSSION In this paper, we studied the nature of the regulation of VACC expression in rat calvarial osteoblasts. Cytosolic calcium measurements were used to indirectly determine the presence of L type VACCs in osteoblastic cells. High extracellular potassium caused membrane depolarization of the osteoblasts and induced increases in cytosolic calcium. This event occurred only in the presence of Bay K 8644, a known calcium channel agonist. This finding suggests that the observed cytosolic calcium raise was due to L type VACC activity. When osteoblastic cells were maintained in media with 10% FCS, 36% of the cells screened responded with a depolarization-induced increase in cytosolic calcium. However, 81% of the cells maintained in 0.01% FCS responded with such an increase in cytosolic calcium. The increase in cell responsiveness required new protein synthesis, because 1 ␮M CHX reduced the percentage of cells responding to 10%. We have previously shown by patch clamp techniques that cells under the same conditions in media containing 10% FCS do not express L type VACCs until 7 or 8 days in culture.(12) In the present study, 36% of the cells screened by depolarization responded at 3 days. This could be explained by differences in manipulation of the cells. In our previous study, cells were pretreated for 15 minutes in a no-calcium-added buffer, to enable the osteoblasts to partially detach and make patch clamping possible. This manipulation could also affect expression of VACCs. In the present study, cells were loaded with Fura 2-AM and no further manipulation was done prior to cytosolic calcium measurements. This suggests a relationship between function of L type VACCs and cellular attachment, which is the case for a large number of events in signal transduction.(25) We reported that 81% of the osteoblasts in media supplemented with 0.01% FCS express L type VACCs, which mediate the depolarization-induced increase in cytosolic calcium. Only 36% of the cells maintained in 10% serum

FIG. 8. Osteoblastic cell proliferation assays. Proliferation is measured by 3H-thymidine incorporation (expressed as CPM). Cells were treated with Streptolysin O and three different antisense oligonucleotides specific for the three known isoforms of the ␣1 subunit of the L type VACCs: (A) cardiac, (B) skeletal, and (C) neuroredocrine. Control groups are cells treated with Streptolysin O only. EGF concentration used was 50 nM. Data used to obtain graph are mean ± SD from triplicate cultures. Statistics: analysis of variance. *Statistically different from control group, 95% confidence level. **Statistically different from EGF group, 95% confidence level. Experiments were repeated at least three times; presented graphs are representative of one of them.

containing media were capable of such a response. There is thus an apparent inverse relation between serum levels and expression of L type VACCs. We propose that cells maintained in 0.01% FCS become quiescent and lose factors

ISOFORMS OF CALCIUM CHANNELS IN OSTEOBLASTS required for osteoblast activity. As a response to this loss, the osteoblasts are required to prepare themselves for the eventual return of those factors. One such preparation could be the expression of L type VACCs. Reports in the literature indicate that the expression of VACCs is related to cell cycle progression. Smooth muscle cells,(26) GH4 pituitary cells,(27) and fibroblasts(28) have been shown to require L type currents for the transition through G1/S and/or G2/M. In osteoblastic cells, it has been difficult to correlate observed cytosolic calcium changes due to PTH stimulation and the position of the tested cell in the cell cycle.(29,30) If VACCs were needed for progression through a specific phase of the cell cycle, proliferating osteoblasts blocked at that specific point of the cell cycle would show an increase in the expression of VACCs. The fact that cell cycle blockers specific for G1, S, and M did not affect the expression of VACCs in osteoblastic cells growing in 10% FCS containing media suggests that VACCs are not required for transition through any of these phases. The possibility exists that VACC expression is required for G0/ G1 transition; however, there is no known blocker for this transition. It could be that osteoblasts in 0.01% FCS place themselves in G0 and that they need to express VACCs to move into G1. This transition to G1 only occurs after growth factors stimulate the cells. We have found that the increased expression of L type VACCs in osteoblasts maintained in 0.01% FCS media is a process that requires new protein synthesis. This could mean that new VACCs need to be synthesized or that intracellular factors need to be first synthesized so they can regulate the existing VACCs with inhibition of the currents. To investigate the first possibility, we undertook the task of studying the expression of L type VACCs at the RNA level. RT-PCR amplification studies indicate that expression of L type VACCs in rat calvarial osteoblasts depends on the proliferative state of the cell. This is in accordance with previous published data from our laboratory on the electrophysiological properties of the cell.(12) We found that serum deprivation in subcultured cells increased the RTPCR product (Fig. 4). However, this was not observed in freshly isolated cells (Fig. 3). We suggest that these differences are due to a larger degree of heterogeneity in the freshly isolated cell preparations, which might mask the effects of serum on the expression of VACCs. We have previously reported that it took 7 days for rat calvarial osteoblasts to express L type VACCs.(12) At the RNA level, RT-PCR product correspondingto the ␣1 subunit of L type VACCs was detected as early as 3 days in culture. It should be noted that osteoblastic cell culture conditions were different between these two studies. To recover enough RNA, the cell concentration had to be increased 50 times in the RT-PCR experiments. Although this did not change the effects of serum deprivation on VACC expression, it did affect the expression of VACCs as a function of time in culture. The expression of VACCs as a function of time in culture shows a significant decrease of the channel after 6 days in culture, with levels recovering partially after 9 days. The highest level of expression of these VACCs at the RNA level is at 3 days in culture, when they have been shown to

393

mediate EGF-induced osteoblastic cell proliferation. The decrease in expression at 6 days could be due to downregulation mechanisms associated with transition from proliferation to differentiation. Since VACCs have been associated with osteoblast differentiation markers,(31) it could be that at this point new events activate production of VACCs but with the purpose of mediating the differentiation process. This is partially supported by our ALP measurements (Fig. 6), which show a slight decrease in ALP at day 6 followed by a large increase at day 9. Similar factors might be responsible for the recovery of VACC levels and the ALP increase observed after 9 days in culture. Little is known about the regulation of VACCs at the RNA level in osteoblastic cells. Hormonal regulation is a very important issue when trying to understand osteoblast physiology, but vitamin D3 is the only hormone to this date that has been shown to have regulatory effects on the RNA levels of L type VACCs.(14) The role that VACCs play in osteoblastic differentiation still remains unclear, and it is an area that will be the focus of our future attention. RT-PCR technology was used to amplify a 900-bp fragment corresponding to a region of the ␣1 subunit of the L type VACCs spanning from the last transmembrane segment of the third domain to the last transmembrane segment of the fourth domain. The expression pattern of this subunit was followed by the intensity of this 900 bp fragment. Also, this fragment was recovered, cloned, and sequenced to determine which isoforms of the ␣1 subunit are expressed in rat calvarial osteoblasts. The only clone recovered in this study corresponds to the cardiac isoform, splice variant a. We followed this with our studies using antisense oligonucleotide to selectively block the different isoforms of the ␣1 subunit that could be expressed in rat calvarial osteoblasts. These experiments suggest that the cardiac isoform is the only isoform responsible for mediating the effects of EGF on cytosolic calcium and osteoblastic cell proliferation. Two other reports have looked into the isoforms expressed by osteoblastic cells. Barry et al.(13) working with rat osteosarcomal osteoblastic cell line UMR 106 found all three isoforms of the ␣1 subunit of the L type VACCs. The functional studies that followed the identification of these isoforms assign a role to the neuroendocrine isoform (D isoform) in mediating the effects of PTH on cytosolic calcium in these cells. Meszaros et al.(14) carried out studies similar to ours and to Barry et al.(13) They found that rat osteosarcomal osteoblastic cell line ROS 17/2.8 and rat calvarial osteoblasts expressed the cardiac isoform of the subunit, with splice variants a and d detected in the ROS 17/2.8 cell line. This group has also reported that mouse osteoblastic cell line MC3T3 expressed a variation of the cardiac isoform.(15) The RNA levels of this isoform were observed to increase during differentiation and after treatment with ascorbic acid. Our results and those reported by Meszaros et al.(14) are in agreement with respect to the type of isoform of the ␣1 subunit expressed in L type VACCs in osteoblastic cells. However, Barry et al. found that all three isoforms known to exist are expressed in UMR 106 osteoblasts. Additionally, they found that PTH effects on cytosolic calcium are mediated by the neuroendocrine isoform.

394 UMR 106 cells have been used extensively in studying the effects of PTH on osteoblastic cells. Primary osteoblast cultures are somehow more difficult to use for the study of PTH regulation, because of the degree of differentiation and the density of expressed PTH receptors. We suggest that all three isoforms of the ␣1 subunit are expressed in UMR 106 cells because of differences in the differentiation state of the cells and/or the transformation state resulting in loss of normal cell control processes. We have previously shown that EGF increases osteoblastic cell proliferation.(16) We extended this observation by showing that cytosolic calcium increases due to calcium influx through VACCs mediate the EGF-induced increase in proliferation.(17) The sequence of events triggered by EGF in rat calvarial osteoblasts has been described.(18) EGF initially activates a non–voltage-dependent calcium conductance, such as a receptor operated calcium channel or the T type VACCs. This initial calcium influx leads to changes in membrane potential that activates the L type VACCs, which are very efficient in mediating calcium influx inside the rat calvarial osteoblasts. We needed to identify the isoform responsible for EGF-induced calcium influx. Our antisense oligonucleotide inhibition studies allowed us to determine that the cardiac isoform of the ␣1 subunit of the L type VACCs is responsible for mediating the effects of EGF on cytosolic calcium and osteoblastic cell proliferation. The inhibition caused by the antisense oligonucleotide to the cardiac isoform of the ␣1 subunit in EGF-induced osteoblastic proliferation (Fig. 8A) was not complete. However, the inhibition of EGF-induced cytosolic calcium changes by this same oligonucleotide (Fig. 7B) was complete. This could be explained in two ways. First, EGF activates multiple signaling pathways; some of these pathways do not require changes in cytosolic calcium to induce cellular proliferation. Those calcium-dependent pathways blocked by the oligonucleotide mediate only a fraction of the EGF effects in proliferation. We have reported(17,18) that the use of L type–specific pharmacological antagonists, like nifedipine or verapamil, partially inhibits the EGFinduced osteoblasic cell proliferation. The results presented in this study are in agreement with our previous observations. A second consideration is the time frame used to carry out this experiment. The exposure of permeabilized cells to the antisense oligonucleotide was only for 10 minutes as the starting point of the experiment. The cells were then allowed to attach and recover from the oligonucleotide-Streptolysin O treatment for 6 h and then were serum deprived for an additional 12 h. At this time point, cytosolic calcium experiments were carried out and the proliferation assay was initiated. The assay lasted for an additional 24 h (36 h after the 10-minute treatment with oligonucleotides). Expression of the cardiac isoform of the ␣1 subunit of the VACCs could have recovered during this prolonged experiment, and therefore the inhibitiory effect could be reduced. In conclusion, we have described the nature of the regulation of the L type VACCs in rat calvarial osteoblasts at the cellular level by performing cytosolic calcium measurements and at the molecular level by detecting the RNA specific for the ␣1 subunit of this channel. The expression of

LOZA ET AL. these channels is dependent on the time in culture and on the proliferative state of the cell. These findings are in agreement with previous data presented by our laboratory on patch clamping measurements of the L type currents in osteoblastic cells. We also identified the cardiac isoform of the ␣1 subunit of the VACCs as the principal one expressed in rat calvarial osteoblasts and we were able to relate the effects of EGF on calcium influx and cellular proliferation to the function of this cardiac isoform. Collectively, these findings provide much information on the molecular characteristics of the VACCs expressed in rat calvarial osteoblasts and the role of VACCs as they relate to osteoblastic cell proliferation.

REFERENCES 1. Tsien R, Tsien R 1990 Calcium channels, stores and oscillations. Ann Rev Cell Biol 6:715–760. 2. Clapham D 1995 Calcium signaling. Cell 80:259–268. 3. Perez-Reyes E, Schneider T 1995 Molecular Biology of calcium channels. Kidney Int 48:1111–1124. 4. Hille B 1992 Calcium channels. In: Hille B (ed.) Ionic Channels of Excitable Membranes. Sinauer Associates, pp. 83–114. 5. Amagai Y, Kasai S 1989 A voltage-dependent calcium current in mouse MC3T3-E1 osteogenic cells. Jpn J Physiol 39:773– 777. 6. Caffrey J, Farach-Carson M 1989 Vitamin D3 metabolites modulate dihydropyridine-sensitive calcium currents in clonal rat osteosarcoma cells. J Biol Chem 264:20265–20274. 7. Chesnoy-Marchais D, Fritsch J 1988 Voltage-gated sodium and calcium currents in rat osteoblasts. J Physiol 398:291–311. 8. Duncan R, Misler S 1989 Voltage-activated and stretchactivated Ba+2 conducting channels in an osteoblast-like cell line (UMR 106). FEBS Lett 251:17–21. 9. Guggino S, Wagner J, Snowman A, Hester L, Sacktor B, Snyder S 1988 Phenylalkylamine-sensitive calcium channels in osteoblast-like osteosarcoma cells. J Biol Chem 263:10155– 10161. 10. Karpinsky E, Wu L, Civitelli R, Avioli L, Hruska K, Pang P 1989 A dihydropyridine-sensitive calcium channel in rodent osteoblastic cells. Calcif Tissue Int 45:54–57. 11. Morain P, Peglion J, Giesen-Crouse E 1992 Ca+2 channel inhibition in a rat osteoblast-like cell line, UMR 106, by a new dihydropyridine derivative, S11568. Eur J Pharmacol 220:11– 17. 12. Loza J, Stephan E, Dolce C, Dziak R, Simasko S 1994 Calcium currents in osteoblastic cells: Dependence upon cellular growth stage. Calcif Tissue Int 55:128–133. 13. Barry E, Gesek F, Froehner S, Friedman P 1995 Multiple calcium channel transcripts in rat osteosarcoma cells: Selective activation of alpha1D isoform by parathyroid hormone. Proc Natl Acad Sci USA 92:10914–10918. 14. Meszaros G, Karin N, Akanbi K, Farach-Carson M 1996 Down-regulation of l-type Ca+2 channel transcript levels by 1,25-dihydroxyvitamin D3. J Biol Chem 271:32981–32985. 15. Meszaros G, Karin N, Farach-Carson M 1996 Voltage-sensitive calcium channels in osteoblasts: Mediators of plasma membrane signaling events. Connect Tissue Res 35:107–111. 16. Stephan E, Dziak R 1994 Effects of genistein, tyrphostin and pertussis toxin on EGF-induced mitogenesis in primary culture and clonal osteoblastic cells. Calcif Tissue Int 54:409–413. 17. Loza J, Carpio L, Lawless G, Marzec N, Dziak R 1995 Role extracellular calcium influx EGF-induced osteoblastic cell proliferation. Bone 4 (Suppl):341S–347S. 18. Loza J, Marzec N, Simasko S, Dziak R 1995 Role of epidermal growth factor-induced membrane depolarization and resulting

ISOFORMS OF CALCIUM CHANNELS IN OSTEOBLASTS

19. 20.

21.

22.

23.

24. 25. 26.

27.

calcium influx in osteoblastic cell proliferation. Cell Calcium 17:301–306. Barry E, Gesek F, Friedman P 1993 Introduction of antisense oligonucleotides into cells by permeabilization with Streptolysin O. Biotechniques 15:10616–10620. Duncan R, Kizer N, Barry E, Friedman P, Hruska K 1996 Antisense oligodeoxynucleotide inhibition of a swellingactivated cation channel in osteoblast-like osteosarcoma cells. Proc Natl Acad Sci USA 93:1864–1869. Luben R, Wong G, Cohn D 1976 Biochemical characterization with paratohormone and calcitonin of isolated bone cells: Provisional identification of osteoclasts and osteoblasts. Endocrinology 99:526–534. Hamill O, Marty A, Neher E, Sakmann B, Sigworth F 1981 Improved patch-clamp technique for high resolution current recordings from cells and cell-free membranes. Pflugers Arch 391:85–100. Wiltink A, Maasen A, Herrmann-Erlee M, van der Meer J, van der Plas A, Willems P, Van Duijn B, Nijweide P, Ypey D 1993 Heterogeneity of intracellular calcium responses to parathyroid hormone and thrombin in primary osteoblast-like cells and UMR 106–01 cells: Correlations with culture conditions, intracellular calcium concentration and differentiation state. Cell Calcium 14:591–600. Perez-Reyes E, Wei X, Castellano A, Birnbaumer L 1990 Molecular diversity of l-type calcium channels. J Biol Chem 265:20430–20436. Clark E, Brugge J 1995 Integrins and signal transduction pathways: the road taken. Science 268:233–239. Kuga T, Kobayashi S, Hirakawa Y, Kanaide H, Takeshita A 1996 Cell cycle-dependent expression of L- and T-type Ca+2 currents in rat aortic smooth muscle cells in primary culture. Circ Res 79:14–19. Dolah F, Ramsdell J 1996 Maitotoxin, a calcium channel acti-

28.

29.

30.

31.

32.

395 vator, inhibits cell cycle progression through the G1/S and G2/M transitions and prevents CDC2 kinase activation in GH4C1 cells. J Cell Physiol 166:49–56. Kanzaki M, Shibata H, Mogami H, Kojima I 1995 Expression of calcium-permeable cation channel CD20 accelerates progression through the G1 phase in Balb/c 3T3 cells. J Biol Chem 270:13099–13104. Bizzarri C, Civitelli R 1994 Activation of the Ca+2 message system by parathyroid hormone is dependent on the cell cycle. Endocrinology 134:133–140. Wiltink A, Bos M 1995 S-phase independence of parathyroid hormone-induced calcium signaling in primary osteoblast-like cells. Cell Calcium 17:270–278. Guggino S, Lajeunesse D, Wagner J, Snyder S 1989 Bone remodeling signaled by a dihydropyridine- and phenylalkylamine-sensitive calcium channel. Proc Natl Acad Sci USA 86:2957–2960. Ziomek C, Lepire M, Torres I 1990 A highly fluorescent simultaneous azo dye technique for demonstration of nonspecific alkaline phosphatase activity. J Histochem Cytochem 38:437–442.

Address reprint requests to: Juan C. Loza, D.D.S., Ph.D Department of Oral Biology SUNY at Buffalo 320 Foster Hall Buffalo, NY 14214 U.S.A. Received in original form February 11, 1998; in revised form August 21, 1998; accepted October 21, 1998.