INTRODUCTION

Parathyroid hormone (PTH)¹-related peptide (PTHrP) was initially recognized as a tumor product that is responsible for most instances of the syndrome of humoral hypercalcemia of malignancy, a common metabolic complication of cancer (1, 2, 3). In this circumstance, PTHrP is elaborated by tumors in a sufficient quantity to cross-react with the classical PTH receptor in bone and kidney, resulting in the prototypical biochemical and clinical features of humoral hypercalcemia of malignancy. It is now known that the PTH and PTHrP genes arose by duplication and are two members of a small gene family (3). What remains from this common origin is a similar intron/exon organization and a short stretch of highly homologous sequence at the N terminus of each mature peptide. These N-terminal sequences appear to share a single G protein-coupled receptor, the PTH/PTHrP receptor (4). This is an interesting and unusual example of ligand-receptor coupling, because the functions of PTH and PTHrP appear to be so remarkably different (3). PTH has a highly restricted pattern of expression and functions as a classical systemic peptide hormone that is charged with maintaining the serum calcium and phosphorus concentrations within the physiological range. In contrast, the PTHrP gene is expressed in a wide variety of normal tissues and cell types and appears to function in an autocrine/paracrine fashion (3). The PTH/PTHrP receptor is also widely expressed, often in a hand-in-glove pattern by cells immediately adjacent to sites of PTHrP expression, another point favoring autocrine/paracrine function (5). Thus, as presently envisioned, the specificity of signaling of this small peptide family is entirely a function of the spatial and temporal expression of the two ligands and their shared receptor. PTHrP appears to be subject to multiple posttranslational processing steps that could generate additional products, but receptors for these peptides have not been identified (3). PTHrP does not normally circulate.

PTHrP is as widely expressed in fetal as it is in adult tissues, and a number of recent experiments in mice indicate that PTHrP functions as a developmental regulatory molecule. For example, targeted disruption of the murine PTHrP gene results in a form of lethal chondrodysplasia that seems to reflect an accelerated program of chondrocyte differentiation (6), whereas targeted overexpression of PTHrP to skin and breast results in a delay/failure of developmental programs in these tissues (7, 8). PTHrP also has been implicated in a number of local regulatory systems in adult tissues, including excitable tissues like smooth muscle. PTHrP is uniformly expressed in smooth muscle throughout the organism, and in a number of these sites the gene has been found to be induced by mechanical stretch (3, 9, 10, 11). In all smooth muscle systems, PTHrP has been found to be a relaxant (3). The functional significance of these findings is most easily appreciated in expandable tissues such as the uterus, bladder, and chicken oviduct shell gland, in which a stretch-induced relaxant would allow accommodation of mechanical filling or occupancy (9, 10, 11).

PTHrP and the PTH/PTHrP receptor have been identified in discrete neuronal populations in the central nervous system, but there is as yet very limited information as to the regulation of PTHrP gene expression or potential PTHrP effects at these sites (3, 12, 13). Here, we report that both PTHrP and its receptor are expressed in primary cultures of rat cerebellar granule cells. PTHrP gene expression in these cells is activity-dependent and is controlled by depolarization-induced calcium entry via the L-type voltage sensitive Ca²⁺ channel acting through a Ca²⁺/calmodulin kinase pathway.

MATERIALS AND METHODS

Primary cultures of granule neurons about 90% pure were prepared from postnatal day eight (P8) rat cerebella by established methods (14). Timed pregnant female Sprague-Dawley rats or timed Sprague-Dawley litters were obtained from Charles River Breeders (Raleigh, NC). Rats were maintained in the Yale Division of Animal Care facilities on a standard diet of rat chow and water ad lib and a 12-h light-dark cycle. P8 rat pups were sacrificed by decapitation, and cerebella were quickly removed, dissected free of meninges and blood vessels, and placed in ice-cold Krebs' buffer (120 mM NaCl, 5 mM KCl, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, 1.3 mM glucose, 25 mM NaHCO₃, 0.3% (w/v) bovine serum albumin). After each cerebellum was collected, tissue was gently triturated (10 cycles with a Pasteur pipette) and then further dissociated by a two-step process using trypsin (0.3 mg/ml) for 15 min at 37 °C, followed by deoxyribonuclease I (40 µg/ml) with soybean trypsin inhibitor (0.3 mg/ml). The resulting cell suspension was then washed in culture medium (Dulbecco's modified Eagle's medium, obtained from Life Technologies, Inc., with 10% heat-inactivated horse serum, 0.5% glucose, 2 mM L-glutamine, and KCl added to reach a final concentration of 25 mM). Cells were plated at a density of 7-8 × 10⁵/cm² onto plastic Petri dishes coated with 10 µg/ml poly-D-lysine. For immunofluorescence analyses, cells were cultured at the same density but upon eight-well glass slides (Nunc) coated with 50 µg/ml poly-D-lysine, which was found to be necessary for reliable cell adherence. Cultures were maintained in a standard 37 °C humidified incubator at 95% oxygen and 5% carbon dioxide.

Cells were studied at 8-12 DIV. Experiments were either carried out in culture medium or, in cases where certain media components were to be left out, in synthetic, serum-free medium with the same ionic makeup of Dulbecco's modified Eagle's medium (1.8 mM CaCl₂, 25 mM KCl, 1.6 mM MgSO₄, 110 mM NaCl, 0.78 mM Na₂HPO₄, 3.3 mM glucose, 3.6 mM L-glutamine, 4.4 mM NaHCO₃). In the synthetic medium, ionic strength was maintained constant when KCl concentration was changed by use of choline chloride.

We evaluated several established techniques such as the Percoll cushion method (15) or trypsinizing and replating mature cultures and found that these procedures did not reliably produce separations of neurons and glia of high purity. We therefore used glutamate exposure to differentially kill granule cells but spare glia in 8-day-old cultures (16). Mature cultures (8 DIV) were washed three times in synthetic medium with 5 mM K⁺ (5K) without added Mg²⁺, since magnesium is known to block the N-methyl-D-aspartate (NMDA) receptor channel (17) and by this action protect neurons from the toxic effects of glutamate (16). Cells were then treated in the same synthetic media with 100 mM glutamate added for 1 h at room temperature. These conditions are somewhat more stringent than those used previously by de Erausquin et al. (16), where the cells were treated for 15 min with 50 mM glutamate, but these investigators employed a higher concentration of Ca²⁺ in their treatment medium (2.3 mM versus 1.8 mM). After the toxic exposure, cultures were washed twice in synthetic medium and returned to culture-conditioned medium in the incubator for an additional 24 h. After this time, loss of neurons was confirmed by direct observation under phase microscopy as well as by immunofluorescence analysis as described below, and sister cultures were harvested for RNA.

Medium (25 mM K⁺) conditioned for 6 h by cells after 8 DIV was assayed for PTHrP content using an immunoradiometric assay for PTHrP-(1-74) with a sensitivity of 1 pM (18). Media were collected in duplicate from three different primary cultures, and results were expressed as a function of cell protein content using the Bradford assay (19).

Cells were grown on eight-well glass slides as described above and subsequently fixed for 45 min in 2% paraformaldehyde, 75 mM phosphate buffer, pH 7.4. After three 5-min washes in phosphate-buffered saline (PBS; 0.8% (w/v) NaCl, 0.02% (w/v) KCl, 0.115% (w/v) Na₂HPO₄, 0.02% (w/v) KH₂PO₄) with 0.2% BSA (PBS-BSA), cells to be stained for glial fibrillary acid protein (GFAP) or PTHrP were permeabilized with 0.1% Triton X-100 (Eastman Kodak Co.) in PBS. Cells were then washed three more times in PBS-BSA and blocked with 100% normal goat serum for 20 min at room temperature. Optimal antibody concentrations were established by conventional dilutional analysis. Rabbit polyclonal antineuronal cell adhesion molecule (NCAM; Chemicon, Temecula, CA) was used at 1:400, and rabbit polyclonal anti-GFAP (Chemicon) was used at 1:2500. These antibodies were diluted in PBS-BSA with 5% normal goat serum. Incubations were carried out in a humidified box at 4 °C for 24 h. To examine for PTHrP immunoreactivity, cultures were pretreated with colchicine (0.1 µg/ml) for 7 h. A 1:1000 dilution of polyclonal rabbit anti-human PTHrP-(1-34) was used (Peninsula Laboratories, Belmont, CA); PTHrP-(1-34) has the same amino acid sequence in rats and humans (20, 21). After incubation, slides were washed four times in PBS-BSA prior to the addition of the secondary antibody, which consisted of a 1:400 dilution of a rhodamine-conjugated goat anti-rabbit IgG (Chemicon) that was added to slides for 1 h at room temperature, followed by three washes in PBS-BSA and a final PBS rinse. Coverslips were mounted with SlowFade Light (Molecular Probes, Eugene, OR), and slides were viewed using an epifluorescent microscope (Olympus BX50 with BXFCA attachment).

RNA was prepared, and RNase protection analysis was carried out as described (13, 22) using RNA probes prepared from a 343-base pair PvuII-BglI rat PTHrP cDNA fragment, a 230-base pair Sau3A-BamHI rat cyclophilin cDNA fragment, and a 259-base pair fragment of the rat PTH/PTHrP receptor DNA. Cyclophilin was added as a loading control in each assay using a probe prepared with a reduced specific activity because of the high relative abundance of cyclophilin mRNA (22). Each sample was assessed using 20 µg of total RNA.

To minimize PTHrP in the media prior to the assay, cultures (8 DIV) were preincubated in 5K medium for 1.5 h and then switched to fresh 5K medium for another 1.5 h. After the 3-h preincubation, cells were changed to synthetic 5K medium with 0.5 mM isobutylmethylxanthine, placed in the incubator for 10 min, and then switched to synthetic 5K medium with 0.5 mM isobutylmethylxanthine and PTHrP-(1-36) at concentrations from 10¹¹ to 10⁶ M in duplicate for another 5 min. At the end of the incubation, the medium was aspirated and replaced with ice-cold n-propanol. Cells in n-propanol were stored at 70 °C overnight prior to collection of supernatants for cAMP assay. n-Propanol supernatants were dried under vacuum and assayed for cAMP using a commercially available radioimmunoassay kit (Biomedical Technologies, Inc., Stoughton, MA). Results were normalized to the total protein in each plate as measured by Bradford assay. The PTHrP-cAMP dose response was estimated by use of nonlinear least squares analysis (Deltagraph Pro 3.5; DeltaPoint, Inc.) using the Michaelis-Menten equation of enzyme kinetics describing a rectangular hyperbola. Forskolin (5-50 µM) was also used to stimulate cAMP, which peaked at ~900 pmol/mg of protein under these conditions. This finding compares well with the 1000 pmol/mg of protein reported by Wroblewska et al. (23) using 30-100 µM forskolin in this culture model.

A ⁴⁵Ca flux assay was used to provide information about intracellular Ca²⁺ flux and was a modification of previously published methods (24, 25). Cells in culture were subjected to a 3-h preincubation in 5K culture medium, and experimental incubations were carried out in the appropriate synthetic medium. Studies were performed using 35-mm² dishes of cells in triplicate unless otherwise noted. After preincubation, the supernatant was aspirated and replaced with synthetic medium, which had been equilibrated for at least 3 h in a 37 °C incubation. Dishes were placed on a 37 °C warming tray during all manipulations outside the incubator and then were returned to the incubator for the designated time period. In order to preserve any factors that might have been released into the medium during the incubation and to minimize trauma to the cells, ⁴⁵Ca was added by quickly aspirating the medium from each plate and mixing it in a microcentrifuge tube with an aliquot of 0.5 mCi of ⁴⁵Ca (445 mCi/mmol, Amersham) diluted in synthetic medium such that the aliquot composed 5% of the final volume. This mixture was then gently returned to the plate, and the cells were replaced in the incubator. For the assay, the medium was aspirated, and the cells were rapidly washed four times with ice-cold wash buffer (5.4 mM KCl, 1.6 mM MgSO₄, 3.3 mM glucose, 0.5 mM EGTA, 154 mM choline chloride at pH 7.4) and flooded with 500 µl of 5% trichloroacetic acid for 45 min. Duplicate aliquots of supernatant (100 µl) were added to vials containing 4 ml of aqueous scintillation fluid (Optifluor; Packard Instruments, Meriden, CT) and counted for 10 min each in a liquid scintillation counter. Results from duplicate tubes were averaged, and the background counts were subtracted. A pilot study demonstrated that ⁴⁵Ca uptake varied linearly with incubation time over the 2-20 min examined (data not shown). We selected for use a 10-min incubation, as have others using this model system (25). Results were corrected for decay of the isotope and normalized for total protein in each plate, as determined by Bradford assay, to allow for direct comparison of experiments.

To assess L-type voltage-sensitive calcium channel (L-VSCC) flux, the specific antagonist nitrendipine (NTR) was employed. Micromolar concentrations of NTR are nonspecific, since 10⁷ M has been shown to inhibit ⁴⁵Ca flux through NMDA receptor-evoked single channel activity in excised outside-out patches of mouse cerebellar granule cells (26), and 10⁶ M NTR has been reported to directly block Na⁺ channels (27). In our experiments, NTR concentrations of 10¹⁰ to 10⁸ M appeared equivalent, inhibiting 25K-stimulated ⁴⁵Ca flux by about 20% (from 117 ± 7 to 92 ± 6 pmol of ⁴⁵Ca/mg of protein/10 min). We therefore selected 10⁹ M to define L-VSCC flux. To allow adequate time for equilibration, NTR was added 15 min before ⁴⁵Ca.

Nitrendipine was a generous gift of Dr. Howard Rasmussen (Medical College of Georgia). Bay K-8644, veratridine, monensin, KN-62, calmidazolium, forskolin, NMDA, and W-7 were obtained from Research Biochemicals, International (Natick, MA). All other reagents not specified were obtained from Sigma. Stock solutions of some reagents with poor aqueous solubility were made in ethanol at 1000 × final concentrations unless otherwise noted. An equal volume of vehicle was added in each case to control samples as well.

In order to compare each experimental point with every other, results were analyzed by one-way analysis of variance. Significance was determined by Fisher's protected test of least significant difference with a 95% confidence interval. Data were expressed ± S.E. Statistical analyses were performed using Statview II (Abacus Concepts).

RESULTS

At 8 DIV the cerebellar granule cell cultures (Fig. 1, top panel) were populated by clusters of uniform, small neurons interconnected by fine processes overlying a bed of confluent, flat glia. The neurons were NCAM-positive, and the glia were GFAP-positive (data not shown). Examination of total RNA obtained from cells grown under 25 mM K⁺ (25K) conditions using RNase protection analysis confirmed that transcripts for both PTHrP and its receptor were present (Fig. 2, top lane). Immunofluorescent anti-PTHrP staining of colchicine-pretreated cultures was primarily localized to the somata of the granule neurons and, to a lesser extent, the interconnecting delicate processes (Fig. 1, bottom panel). Solitary bright neurons were occasionally visualized on the substrate outside of the aggregates, whereas only faint background, uninhibitable fluorescence was localized to the glial cells underlying the neuronal clusters. After cultures were pretreated with glutamate to deplete neurons and examined 24 h later, the neuronal clusters and associated processes were completely disrupted (data not shown). RNA analysis showed that in these neuron-depleted (glial-enriched) cultures, PTHrP mRNA, as well as that of its receptor, was eliminated (Fig. 2, bottom lane).

The assay for PTHrP-(1-74) was performed upon medium from cultures depolarized for 6 h with 25K in three separate experiments. PTHrP was detected at concentrations of 2.5-6.0 pM (0.6-1.3 (± 0.1) fmol/mg of protein; limit of detection 1 pM). Medium from 4 mM K⁺ (``choline control'') possessed undetectable levels of PTHrP, as did glia-enriched cultures. These granule cell cultures also produced cAMP in response to exposure to PTHrP-(1-36) in a dose-dependent manner. The base-line cAMP production for these cultures was 30.7 ± 5.4 fmol/mg of protein, with maximal stimulation of 68.6 ± 8.2 fmol/mg of protein (p < 0.001). The half-maximum stimulation by PTHrP occurred at about 16 nM. For comparison, 25K produced only a small increase in cAMP (~9 fmol/mg of protein), and maximal cAMP stimulation with forskolin (50 µM) was 874 ± 52 fmol/mg of protein (results of three experiments performed in duplicate).

Cultures maintained in the depolarizing 25K medium exhibited high levels of PTHrP mRNA (Fig. 3A). When this medium was substituted by 5K medium with osmolality maintained constant by the addition of choline chloride, PTHrP mRNA decayed slowly over 3-4 h (Fig. 3, A). Replacement of conditioned with fresh (25K) medium, which would remove culture-derived growth and other factors, did not produce a similar decay in mRNA (Fig. 3A, right part). The loss of PTHrP mRNA in 5K medium was not due to cell injury, as shown by the slow return of PTHrP mRNA back to control levels by reexposure to 25K medium (Fig. 3A, left part).

The dynamics of PTHrP induction are shown in more detail in Fig. 3B. In order to visualize the PTHrP induced by 25K stimulation, cells in all subsequent experiments were pretreated for 3-4 h in 5K so that PTHrP mRNA levels prior to stimulation would be minimal. After 1 h of depolarization, faint but detectable mRNA was present, with continued increases up to a plateau at 4 h, equal to the base line. Exposure of cultures to graded depolarization reliably induced PTHrP mRNA at 15K, but not at 10K (Fig. 3C). However, depolarizing conditions alone were not sufficient for PTHrP mRNA induction, since treatment of cultures with strongly depolarizing concentrations of the Na⁺ channel agonists veratridine (0.5-25 µM), glutamate (0.1-10 mM), monensin (5 µM), or the Na⁺,K⁺-ATPase inhibitor ouabain (1 mM) did not induce PTHrP mRNA. Use of the potent glutamate agonist, NMDA (100 µM), also failed to induce PTHrP mRNA in cultures maintained at 5K but did activate expression in those maintained at 15K (data not shown).

The abundant level of PTHrP mRNA in depolarized granule cells required Ca²⁺ in the extracellular medium, as shown by an experiment in which cultures were pretreated with 5K for 4 h and then reexposed to 25K with varying medium [Ca²⁺] (Fig. 4). Elimination of Ca²⁺ in the medium resulted in no detectable PTHrP mRNA. A sharp Ca²⁺ concentration-dependent response curve was observed under 25 mM K⁺ conditions, such that minimal PTHrP was detected with 0.6 mM Ca²⁺ present, while at 1.2 mM Ca²⁺ large increases were appreciated (Fig. 4). Cells maintained in zero-added Ca²⁺ for 4 h remained viable, since they demonstrated a robust PTHrP mRNA response when placed in 1.2 mM Ca²⁺ (``add-back''). This Ca²⁺-dependent PTHrP mRNA induction was virtually completely blocked by the addition of 10⁹ to 10⁶ M NTR (Fig. 5A). Direct assessment of Ca²⁺ flux using ⁴⁵Ca confirmed that an appreciable (~15-20%) amount of Ca²⁺ flux was blocked by 1 nM NTR at 25K, while under 5K conditions no NTR-sensitive flux was observed (Fig. 6). Further, this NTR-sensitive calcium flux remained constant for at least 3 h (Fig. 6). The addition of nanomolar concentrations of NTR directly to 25K cultures also produced a decrease in PTHrP mRNA, with the same slow decay kinetics as observed after a change to nondepolarizing (5K) conditions (data not shown).

Confirmation that PTHrP mRNA induction was critically dependent on Ca²⁺ flux through L-VSCCs was obtained by use of the specific L-channel agonist, Bay K-8644 (Fig. 5B). We found that BAY K-8644 alone was ineffective under nondepolarizing conditions (data not shown). However, if cultures were maintained in mildly depolarizing (15K) medium, Bay K-8644 exposure in the presence of Ca²⁺ was associated with an increase in PTHrP mRNA. Omission of extracellular Ca²⁺ abolished this effect. The calcium ionophore A23817 could not be evaluated, since this compound proved extremely toxic to granule cell neurons.

Further support of a critical role for L-VSCC Ca²⁺ conductance in PTHrP mRNA expression was obtained by use of the alkaloid veratridine, which depolarizes cells by directly activating voltage-sensitive Na⁺ channels (28). Although prolonged exposure to veratridine in the 0.5-25 µM range did not injure granule cells, no PTHrP mRNA was induced (data not shown). Measurements obtained using ⁴⁵Ca flux demonstrated that veratridine exposure over this concentration range induced increases in intracellular ⁴⁵Ca to a greater extent than did K⁺ alone. Cultures exposed to veratridine at 3.5 µM were found to have an intracellular ⁴⁵Ca flux equal to that of 25K (Fig. 6, right), but this treatment did not induce PTHrP mRNA (Fig. 7). At this concentration of veratridine, none of the ⁴⁵Ca flux was NTR-sensitive (Fig. 6, right). Add-back of 25K medium proved that these cells remained viable and capable of making abundant PTHrP mRNA (Fig. 7).

Three different inhibitors of the calmodulin/CaM kinase pathway were employed to study the signaling pathway between L-VSCC-mediated Ca²⁺ flux and PTHrP mRNA induction. Cultures were pretreated in 5K and subsequently exposed to 25K medium for 5 h along with the Ca²⁺/calmodulin pathway inhibitors calmidazolium (1 µM), W-7 (50 µM), or KN-62 (10 µM). Under these conditions, each of these compounds completely blocked induction of PTHrP mRNA (Fig. 8).

DISCUSSION

The cerebellar granule cell is rich in PTHrP mRNA in vivo (13). Because these cells develop postnatally, they can be successfully harvested in large numbers and maintained for weeks in vitro. As a result, this system, in which >90% of the neurons are granule cells, has been extensively used to study the biology and development of neurons (29, 30). Our cultures were of typical mature morphological appearance, containing predominantly interconnected NCAM-positive neuronal clusters overlying a bed of GFAP-positive, flat glia. The results of immunoradiometric assay of conditioned medium confirmed that these cultures make and release PTHrP, in contrast to the negative analyses of conditioned medium obtained from neuron-depleted cultures. The cellular localization of PTHrP was identified by the robust anti-PTHrP immunofluorescence of the granule cell neurons and their processes, absent in the flat astrocytes. Confirmation of a neuronal source for PTHrP was provided by the neuron depletion experiment (Fig. 2). Prior in situ hybridization studies suggested these same conclusions in vivo (12, 13). Direct study of primary cultures of cortical type I astrocytes by others has also shown that glia do not make PTHrP (31).

Our results also show that mRNA for the common PTH/PTHrP receptor is expressed only by neuron-containing cultures. The addition of PTHrP directly to the culture medium stimulated cAMP in a dose-dependent manner, as expected for the adenylate cyclase-coupled PTH/PTHrP receptor. The V_0.5 of ~16 nM, which we observed for PTHrP-(1-36) is similar to the ~30 nM previously reported using a cortical astrocyte culture model (31). These affinities are much lower than the ~0.5 nM previously reported for this receptor in renal membranes (32) and likely reflect the differences in peptide concentrations to which each receptor is exposed in vivo (33); the cerebellar receptor is bathed in high concentrations of PTHrP released locally by adjacent cells, whereas the renal receptor responds to nanomolar concentrations of the circulating hormone, PTH. In sum, these observations indicate that PTHrP is made by granule cells in vitro; the peptide is released into the medium, and it is capable of generating cAMP by acting through the granule cell PTH/PTHrP receptor.

PTHrP mRNA was identified in the brain by the initial tissue survey performed after the isolation of PTHrP and its cDNA (3). PTH/PTHrP-like biological activity and PTHrP mRNA were subsequently identified in multiple regions of the rat brain, and the mRNA was localized to discrete neuronal populations in the hippocampus, cerebral cortex, and cerebellum by in situ hybridization histochemistry (13). A subsequent survey revealed PTHrP mRNA to be expressed in an even greater variety of neurons and also demonstrated that the PTH/PTHrP receptor mRNA was widely expressed in the nervous system, although in a pattern that was anatomically discrete from the PTHrP pattern (12). Curiously, this second study did not identify PTH or PTH/PTHrP receptor mRNA in cerebellar granular cells, a discrepancy for which we have no explanation. Many of the neurons that express the PTHrP gene are known to be rich in L-VSCCs, excitatory amino acid receptors, and Na⁺,K⁺-ATPase (13). PTH/PTHrP receptor mRNA has also been reported by two groups in cultured cortical type I astrocytes (31, 34), but it did not seem to be present in cerebellar glia populating neuron-depleted cultures.

Characterization of granule cell cultures with respect to PTHrP expression showed that abundant transcripts were present under maintenance (25K) conditions. Membrane depolarization was one critical requirement, as PTHrP mRNA decayed to undetectable levels over 3-5 h when cultures were switched to nondepolarizing (5K) conditions. Cell injury did not account for this decay, since restoration of full depolarizing medium (add-back) was associated with a return of abundant PTHrP mRNA. However, depolarization was clearly not always sufficient for PTHrP mRNA expression, since exposure of cultures in 5K medium with added veratridine, monensin, or ouabain over a wide range of depolarizing concentrations each failed to stimulate PTHrP mRNA in this model. The key to understanding these observations is that the Ca²⁺ ion is an absolute requirement for PTHrP mRNA induction; removal of Ca²⁺ from the 25K medium completely abolished the PTHrP signal under maintenance conditions. Further, only a specific form of Ca²⁺ flux, that through the L-VSCC, proved to be biologically relevant for PTHrP mRNA induction.

The results of the ⁴⁵Ca studies showed that although a large amount of intracellular ⁴⁵Ca accumulated under nondepolarizing conditions (5K; Fig. 6, left), no activation of the PTHrP gene was detected (Fig. 3C). In contrast, under depolarizing K⁺ conditions (15K), a Ca²⁺ flux was stimulated that was inhibitable by the specific L-VSCC antagonist, nitrendipine, and it was this flux that was associated with PTHrP mRNA induction. The non-L-VSCC flux in the granule cell system is mediated mostly by Ca²⁺ conductance through the NMDA receptor, since it is blocked almost entirely by the specific antagonist MK801 (25).² Further, the depolarization threshold for detectable PTHrP message occurred at about 15K, which is also the lowest K⁺ concentration at which an appreciable L-VSCC flux can be measured. Supporting evidence for the importance of the L-VSCC was also provided by the BAY K-8644 experiments, in which mild depolarization (15K), normally insufficient to induce substantive amounts of PTHrP mRNA, became effective. Prior study of granule cells by Kingsbury and Balazs (35) has shown that BAY K-8644 is effective at inducing L-VSCC flux only under media conditions of 15K. Under these conditions, the L-channel is phosphorylated (by action of a cAMP-dependent protein kinase (36, 37)), which, in turn, allows BAY K-8644 to conduct Ca²⁺. This fact is also the likely (but untested) explanation for the observation that NMDA receptor stimulation produced PTHrP mRNA only when added to 15K medium and not to nondepolarizing 5K medium, since the requirements for mildly depolarizing conditions for NMDA activation of its receptor have been observed in this system by others (25). Although veratridine treatment produced a large intracellular calcium flux, none of it was through the L-VSCC, and this same conclusion has been reached by other investigators studying different neuronal culture model systems (38, 39). It thus appears that the lack of L-VSCC-mediated Ca²⁺ flux explains why veratridine treatment does not induce PTHrP mRNA.

There are many routes by which Ca²⁺ can enter the cytoplasm, including VSCCs, the NMDA receptor, and other ionotropic receptors. Accumulating evidence suggests that highly localized differences in [Ca²⁺] occur at the subcellular level in neurons, which accounts for the specificity with which a Ca²⁺ flux can elicit a given response (40), an example being the L-VSCC-PTHrP relationship we report here. The L-VSCCs are generally distributed only over the cell body and proximal dendrites of neurons (41). This location, near the nucleus, explains why the L-VSCC is not involved in regulating transmitter release but rather is strategically placed to translate signals from the cell membrane into changes of gene expression (42). An excellent example of signaling specificity from the literature is the two pathways regulating c-fos gene expression: Ca²⁺ entry via the NMDA receptor and depolarization-induced Ca²⁺ entry via the L-VSCC. The NMDA pathway tracks through the serum response factor to the serum response factor on the c-fos gene, whereas the L-VSCC pathway works via calmodulin/CaM kinase to act on the cAMP response element on the c-fos gene (43).

We evaluated the role of the calmodulin pathway by use of agents that target different sites in this cascade. Calmidazolium and W-7 antagonize calmodulin itself, whereas KN-62 is a CaM kinase inhibitor that prevents the binding of calmodulin to CaM kinase, impairing a critical autophosphorylation (44). All inhibitors were found to block the induction of PTHrP mRNA. The concentrations used were selected based upon prior experience with inhibition of the CaM pathway in cell culture systems. However, these compounds, particularly calmidazolium, have been reported to directly reduce L-VSCC flux in neuronal (45, 46, 47) and nonneuronal systems (48), leading to concerns about their specificity. This question has been directly examined for calmidazolium and KN-62 in cerebellar granule cell cultures by use of ⁴⁵Ca flux (49), and the results showed that KN-62 in concentrations up to 10 µM had no effect on L-VSCC flux; calmidazolium at the 1 µM concentration we used reduced ⁴⁵Ca flux by <10% while completely inhibiting CaM. The lack of effects of KN-62 on L-channel inhibition have been confirmed in other neuronal systems as well (50, 51). Thus, use of multiple agents that target different sites clearly implicate the Ca²⁺/calmodulin cascade in PTHrP gene expression. Other examples of Ca²⁺ channel-mediated PTHrP gene regulation have not been reported, but the finding in cerebellar granule cells might well be relevant to the stretch-induced PTHrP mRNA response in smooth muscle since L-VSCCs and the calmodulin pathway have been implicated in mechanotransduction (52, 53).

There is very limited information as to what function(s) PTHrP might serve in the central nervous system, but several points merit consideration. First, parathyroid extracts were found to have biological effects outside bone and kidney as early as 1925 (3), leading to the general working hypothesis that PTH might have local actions in a number of sites, including smooth muscle and the central nervous system (3). While it seems increasingly clear that smooth muscle effects previously attributed to PTH correspond to actual PTHrP functions and the same may be largely true in the brain, there is evidence that PTH itself is produced in the central nervous system (54, 55), where most attention has focused upon the hypothalamus. In addition, a relative of the PTH/PTHrP receptor subfamily has recently been isolated from a rat cerebral cortex cDNA library and termed the PTH2 receptor because it responds only to PTH (56). The sites of expression of this receptor in the central nervous system remain unknown. Second, the PTH/PTHrP receptor can couple to both adenylyl cyclase and phospholipase C and is thus capable of signaling through either the cAMP or calcium pathway (57). In an initial study employing hippocampal neurons in primary culture, PTHrP was found to increase cAMP and to produce a sustained but not a transient increase in intracellular calcium that appeared to be channel-mediated (58). In the case of cerebellar granule cells, we have demonstrated the presence of the PTH/PTHrP receptor and its capacity to signal via cAMP but have not yet evaluated calcium signaling or more distal responses. An unexpected finding in the study of hippocampal neurons was the capacity of a C-terminal PTHrP peptide, PTHrP-(107-139), to elicit an increase in intracellular calcium (58). This region of the molecule has previously been implicated in the regulation of osteoclastic bone resorption, but nothing is known of its receptor (3). Third, PTHrP-(1-36) has been reported to induce protoplasmic astrocytes in culture to become more differentiated process-bearing cells, suggesting that PTHrP might have developmental regulatory effects in the nervous system similar to those reported in a number of other tissues (6, 7, 8, 31, 34). Finally, the survival-promoting effects of K⁺ depolarization for granule cells depend upon the presence of extracellular Ca²⁺ and in large part upon the induction of transmembrane L-VSCC fluxes (59). It is interesting to note that PTHrP mRNA was detected only under conditions promoting survival of granule cells. Since a similar role of the L-VSCC in cell survival has been observed for other cultured neurons (60, 61, 62), it will be interesting to determine whether PTHrP is a common element in those systems as well.