Humoral hypercalcemia of malignancy

Before the existence of the syndrome of humoral hypercalcemia of malignancy was recognized, it was reported that in inbred aged Fisher rats testicular neoplasms of Leydig cell origin develop spontaneously (123). The tumors were capable of being passaged by serial subcutaneous transplantations. The tumors grew rapidly but did not metastasize. The rat Leydig cell cancer was discovered to cause hypercalcemia with hypophosphatemia, hypercalcuria, and hyperphosphaturia in rats (199, 211, 212). Later Stewart et al. found that in patients with malignant hypercalcemia two groups could be identified: one group with solid tumors but no bone metastases and a second much larger group of patients with metastases to bone (245). The hypercalcemic state of the first group of patients was termed humoral hypercalcemia of malignancy, indicating that some humoral factor secreted by the solid tumor would activate bone resorption and lead to hypercalcemia. Sica and co-workers found that the rat Leydig hypercalcemia tumor did not metastasize to bone and that the hypercalcemia could be reversed by removal of the tumor from the rats. This study verified that the rat Leydig cell cancer was a suitable model of the humoral hypercalcemia occurring in patients with solid tumors (238). The tumors were investigated by light and electron microscopy and verified to be of Leydig cell origin (212). No sex hormone production was found in the tumor. In vivo, human chorionic gonadotropin caused an acute rise in serum calcium in 3 to 5 hours in tumor-bearing hypercalcemic rats and the binding of chorionic gonadotropin was competed by luteinizing hormone. This indicates that an active chorionic gonadotropin/luteinizing hormone receptor is functionally expressed on the H-500 cells. Hereafter studies found that the mediator of this hypercalcemia induced cAMP-production by adenylate cyclase as seen in PTH-driven hypercalcemia, and that this was mediated through the PTH receptor (120, 217, 242). In 1987 PTHrP was discovered (249). This peptide (or peptides as there are three isoforms) resembles PTH and activates the PTHR1 similar to PTH. Using a two side immunoradiometric assay, it was shown soon thereafter that indeed PTHrP is the humoral factor produced by the H-500 Leydig tumor (74) and that neutralizing the peptide could abolish the hypercalcemia in Fisher rats transplanted with the H-500 tumor (103). Furthermore EGF has been shown to induce PTHrP release whereas testorone, dexamethasone and 1.25 vitamin D 3 downregulate PTHrP in H-500 cells (154). Thus the Rice H-500 Leydig testis cancer cells seem to be a valid, well-established model of humoral hypercalcemia. The H-500 cells differ from the normal Leydig cells as they are not polygonal, and most human Leydig cell tumors do not have central necrosis in contrast to the H-500 cell tumor. In human PTHrP has been reported to be expressed in the Leydig cell, although its function is unknown (9). All the studies, investigating the H-500 cells, used serial transplantations of the tumor. Simultaneously, with the in vivo characterization of the H-500 cells, the cells were also characterized in vitro. In primary cell culture the cells produce PTHrP and grow like cancer cells for up to 10 cell passages (154, 206).

Hypercalcemia of malignancy due to metastases to the bone

In the paper by Stewart and co-workers that defined HHM, the larger group of patient with malignant hypercalcemia were the patients with metastases to bone (245). The cancers that metastasize to bone are among the most frequent cancers, namely kidney, lung, breast, and prostate cancers (90). Prostate cancers, that metastasize to bone, generally form mixed osteolytic and osteosclerotic lesions. Substantial increases in bone resorption occur in this setting as assessed by biochemical markers such as pyridinoline and deoxypyridinoline (56, 119, 251). Interestingly, the markers of bone resorption may be lower in those with skeletal metastases of breast cancer than in patients with metastatic prostate cancer, indicating that more bone degradation is taking place in the latter (56). The PC-3 cell line was derived from a bone metastasis of a grade IV prostatic adenocarcinoma from a 62-year-old male Caucasian (127). The PC-3 cell line induced hypercalcemia in the human subject from which it originated as well as in animal models when inoculated subcutaneously (127). Furthermore, it was shown the PC-3 cells express a functionally active calcium sensing-receptor and that high calcium through the CaR induced PTHrP release in these cells when grown in vitro (223). With this in mind, we chose the PC-3 cell line as a well-established model system to investigate new signaling pathways for the CaR in malignant cells with metastases to bone.

Malignant cells expressing the CaR that are not involved in hypercalcemia of malignancy

The last cell model used in our studies is the astrocytoma cell line U-87. The U-87 cell line is derived from a malignant gliomas grade III from a 44 year old Caucasian female (198). The astrocytoma cells do not induce systemic hypercalcemia neither by a humoral factor nor by metastases to bone. In the literature there are two astrocytoma cell lines that have been reported to express a functional CaR, the U373 and the U-87 cell lines (43, 44). Thus, the astrocytoma represents a model of cancer with no effect on systemic calcium and only local changes in calcium along with the normal changes in systemic calcium that will potentially stimulate the CaR expressed on these cells. The U-87 astrocytoma cells were chosen to investigate the effects of CaR stimulation on the PTTG oncogene due to its robust CaR-induced PTHrP release and ion channel regulation (40, 44). To validate the findings on the levels of PTTG expression in the U-87 cell line, we used human primary tumor samples that had been collected at surgery in the Brigham and Women's Hospitals surgical department as well as primary astrocytes, and the T98G and U-343 astrocytoma cell lines. Before we used the tumor to investigate mRNA expression levels, another part of the same tumor had undergone histological classification by the department of pathology.

Materials

Antibodies: Polyclonal antisera against phosphorylated and nonphosphorylated ERK1/2, p38 MAPK, SEK1, AKT (ser473) and ATF-2 kinases were purchased from New England Biolabs (Beverly, MA). A rabbit polyclonal antiserum against iNOS was purchased from BD Transductions Laboratories (Lexington, KY). Neutralizing antibodies against EGFR and HB-EGF were obtained from R&D Systems (Minneapolis, MN). Polyclonal antisera to EGFR and PTTG (M-16) and a mouse monoclonal antibody against phosphotyrosine PY99 were purchased from Santa Cruz Biotech (Santa Cruz, CA).

Inhibitors and activators: Selective inhibitors of p38 MAPK (SB203580), MEK1 (PD98059), JNK (SP600125), PI3K (LY294002), EGFR kinase (AG1478), PDGFR kinase (AG1296), pan MMPs (GM6001), and PKC (GF109203X) were all obtained from Calbiochem-Novabiochem (San Diego, CA). The PKC activator, phorbol 12-myristate 13-acetate (PMA), was purchased from Biomol (Plymouth Meeting, PA). The choice of dose of the individual inhibitors was based on data from the literature, previous experience from the laboratory and from pilot studies on dose response curves of the individual inhibitors (15, 64, 68, 125, 136, 142, 156, 183, 185, 268, 279). Most of the signalling molecules modulated by the chemical inhibitors were also investigated by western blotting or their upstream or downstream signaling molecules in order to validate whether calcium would activate the signalling molecules that were blocked pharmacologically, providing some support for the specificity of the inhibitors.

Other: The enhanced chemiluminescence kit, Supersignal, was purchased from Pierce (Rockford, IL). Protease inhibitors were obtained from Boehringer Mannheim (Mannheim, Germany). TGF α , EGF, 17- β -estradiol, BMP-2, TGF β , TNF α , TRAIL, TTNP3, I3CRA, Ciglitazon, and PGJ2 were obtained from Calbiochem (La Jolla, CA, USA). All cell culture reagents to H-500, PC-3, and U-87 cells were purchased from GIBCO-BRL (Grand Island, NY) with the exception of FBS, which was obtained from Gemini Bio-Products (Calabasas, CA). NPS R-467 and NPS S-467 were donated by NPS Pharmaceuticals, Inc. (Salt Lake City, UT). Other reagents were from Sigma Chemical (St. Louis, MO).

Methods

Protocol for H-500 in-vitro cell culture

The Rice H-500 rat Leydig cell tumor was obtained from the National Cancer Institute-Frederick Cancer Research and Development Center DCT Tumor Repository (Frederick, MD). Male Fischer 344 rats (Harlan Sprague-Dawley, Inc., Indianapolis, IN) weighing 200 to 220 g (10 weeks of age) were used for all experiments. A fragment of the H-500 tumor was implanted subcutaneously in each rat, and the tumor was allowed to grow for 8 to 14 days, with the size of the tumor determining when the tumor was removed; the average tumor removed was around 0.5-1 cm in radius. The encapsulated tumor was cut open, and tumor pieces were excised with a minimum of other tissues from the animal, rinsed several times with cell culture medium (see below), minced into smaller pieces, and dispersed by repeated pipetting and several passages through a 22-gauge needle. Dispersed H-500 cells were subsequently plated in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 100 U/ml penicillin-100 μ g/ml streptomycin and grown at 37°C in a humidified 5% CO 2 atmosphere. Cells were passaged every 3 to 5 days using 0.05% trypsin-0.53 mM EDTA and used for experimentation within the first 10 passages. After the 10 passages the response to calcium at times would be either less robust or absent, indicating that CaR expression was downregulated or that the H-500 cells had undergone more substantial phenotypical changes. No contamination of host cells was identified as assessed by the presence of homogeneous morphology under light microscopy. Rats were handled in accordance with local institutional guidelines. In one instance an animal died, and in two cases the tumor ulcerated through the skin. No secondary tumors were detected in the animals.

Protocol for PC-3 cell culture

The PC-3 human prostate cancer cell line was obtained from the American Type Culture Collection (Rockville, MD). The cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 100 U/ml penicillin-100 μ g/ml streptomycin and grown at 37°C in a humidified 5% CO 2 atmosphere. Cells were passaged every 4-5 days using 0.05% trypsin-0.53 mM EDTA.

Protocol for human primary astrocytes and U-87 cell culture

Human primary astrocytes were purchased from Clonetics-Biowhittaker (Walkersville, MD), and U-87, T98G, and U-343 astrocytoma cells from the American Type Culture Collection (Manassas, VA); they were maintained in monolayer culture in DMEM supplemented with 10% fetal bovine serum and 100 U/ml penicillin-100 mg/ml streptomycin and grown at 37°C in a humidified 5% CO 2 atmosphere. Primary astrocytes were derived from a fetal source. The primary astrocyte cultures are more than 90% pure as certified by the company based on assessment of the expression of glial fibrillary acidic protein.

Specimen selection and tissue samples

Nine glioma samples were collected at the time of surgery in patients who underwent craniotomy for glioma resection with human study approval. All operations were performed by Dr. Peter Black at the Brigham and Women's Hospital in Boston. A senior neuropathologist at the hospital evaluated all specimens and classified them in accordance with World Health Organization standard criteria. At the time of surgery, all tissue specimens were immediately snap frozen and stored in liquid nitrogen for RNA isolation.

Infecting H-500 cells with CaR constructs in rAAV

High-efficiency gene transfer into H-500 cells was accomplished using a recombinant Adeno-Associated Virus (rAAV)-based method. The CaR sequence with a naturally occurring, dominant negative mutation (R185Q), as well as the same vector containing the cDNA for the beta galactosidase protein (referred to hereafter as BG) were under the control of a cytomegalovirus immediate-early (CMV-IE) promoter element, and packaged as previously described by our collaborator (292). The BG served as the control for non-specific effects of rAAV infection. Cells were seeded (1000 cells/well) in 96-well plates in 0.1 ml of growth medium and cultured overnight. About 1000 virus particles/cell (as optimized by pilot studies) were used to infect each well. Cells were washed once with serum-free RPMI 1640. Virus particles were then added, and the culture was incubated for 90 min in the serum-free medium at 37°C in a cell-culture incubator. Equal volumes of RPMI 1640 containing 20% serum were added to the cells to achieve a final serum concentration of 10%. The cells were then cultured for 48 h, and experiments with low and high calcium concentrations were performed. No studies were done on cell growth or response to calcium besides the PTHrP and mRNA expression measurements with infected H-500 cells.

Measurements of PTHrP in media

PTHrP was measured using a two-site immunoradiometric assay (IRMA) (Nichols Institute Diagnostics, San Juan Capistrano, CA) that detects PTHrP (1-72) with a sensitivity of 0.3 pmol/L. At 3.4 pmol/L our inter- and intra-assay variance were 8.8% and 4.0% respectively. PTHrP assays were initiated immediately after removing the conditioned medium to minimize loss of peptide. PTHrP concentrations were calculated from a standard curve generated by adding recombinant PTHrP(1-86) to the treatment medium (i.e. unconditioned medium). Calcium with or without inhibitors had no effect in the PTHrP assay.

The effects of Ca 2+ o as well as MAPK and PKC inhibitors on PTHrP release were determined by seeding H-500 or PC-3 cells in 96-well plates (1×10 4 cells/well) in 0.1 ml of growth medium. After 48 h, the growth medium was removed and replaced with 0.1 ml of Ca 2+ -free DMEM containing 4 mM L-glutamine, 0.2% BSA, 100 U/ml penicillin-100 μ g/ml streptomycin, and 0.5 mM CaCl 2 . Two h later, this medium was removed and replaced with 0.225 ml of the same medium or that supplemented with additional CaCl 2 (to a final concentration of 2.5, 5.0, or 7.5 mM) for 6 h. In other experiments the medium was supplemented either with the kinase inhibitors described or with 7.5 mM CaCl 2 together with the same inhibitors, in the PC-3 cells the inhibitors were preincubated for 0.5 h before stimulation. In the experiment with ADP the concentrations used were 10 -7 , 10 -8 , and 10 -9 M. Six h later, the conditioned medium was collected for determination of PTHrP release. The 6 h time point was chosen for subsequent experiments because it yielded PTHrP values falling on the linear portion of the PTHrP assay, whereas at 4 and 24 h, PTHrP was at the lower or upper portion of the curve, respectively. The fold increase in PTHrP release at high calcium compared to low calcium did not vary over the first 24 h. Furthermore the 6 hour time point was also chosen since results from our laboratory previously had shown no change on proliferation of the H-500 cells in this time frame as assed by MTT assay a colometric assay (224).

Determination of HGF secretion

To study HGF secretion, U87 cells were grown to 70-75% confluence in complete growth medium in 24-well plates. They were then serum-starved overnight in growth medium minus FBS containing 0.2% BSA along with various concentrations of EGF. Medium samples were cleared by centrifugation, and HGF was measured in this conditioned medium with an ELISA. The ELISA employs a quantitative sandwich, enzyme-linked immunoassay technique, utilizing a monoclonal antibody specific for HGF that is bound to microtiter wells. Assay sensitivity was 125 pg/ml. Data are expressed as picograms per microgram protein.

Western blot analysis

For the determination of PTTG, iNOS, AKT, ATF-2, ERK1/2, p38 MAPK, SEK1, or phosphorylation, monolayers of H-500, PC-3, or U-87 cells were grown on six-well plates. Cells were incubated for 18 h in serum-free, Ca 2+ -free DMEM containing 4 mM L-glutamine, 0.2% BSA, and 0.5 mM CaCl 2 . This medium was removed and replaced with the same medium supplemented as described in the papers. At the end of the incubation period, the medium was removed, the cells were washed twice with ice-cold phosphate-buffered saline (PBS) containing 1 mM sodium vanadate and 25 mM NaF, and then 100 mL of ice-cold lysis buffer was added [20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 25 mM NaF, 1% Triton X-100, 10% glycerol, 1 mM dithiothreitol, 1 mM sodium vanadate, 50 mM glycerophosphate, and a cocktail of protease inhibitors]. The protease inhibitors were aprotinin, leupeptin, soybean trypsin inhibitor, pepstatin, and calpain inhibitor (10 mg/ml of each), all from frozen stocks, as well as 100 mg/ml of Pefabloc. The sodium vanadate, NaF, and Pefabloc were freshly prepared on the day of the experiment. The cells were scraped into the lysis buffer, sonicated for 5 sec, and then centrifuged at 6,000 × g for 5 min at 4°C. The supernatants were frozen at -20°C. After thawing, equal amounts of supernatant protein (100 μ g) were separated by SDS-PAGE. The separated proteins were electrophoretically transferred to nitrocellulose membranes (Schleicher and Schuell) and incubated with blocking solution (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, and 0.25% BSA) containing 5% dry milk for at least 1 h at room temperature. AKT, ERK1/2, p38 MAPK, SEK1, and ATF-2 phosphorylation were detected by immunoblotting using an 18-h incubation with 1:1000 dilutions of rabbit polyclonal antisera specific for phospho-AKT, phospho-ERK1/2, phospho-p38 MAPK, phospho-SEK1, or phospho-ATF-2, respectively. For iNOS and PTTG 1:2000 and 1:1000, respectively, dilutions of specific primary rabbit polyclonal antisera was used. Blots were washed for five 15-min periods at room temperature (1% PBS, 1% Triton X-100, and 0.3% dry milk) and then incubated for 1 h with a secondary goat anti-rabbit, peroxidase-linked antiserum (1:2000) in blocking solution. Blots were then washed again (5 × 15 min). Bands were visualized by chemiluminescence according to the manufacturer's protocol (Supersignal, Pierce Chemical). The same membrane was used after stripping (Restore Western Blot Stripping, Pierce) to measure non-phospho AKT, ERK1/2, p38 MAPK, and SEK1. Protein concentrations were measured with the Micro BCA protein kit (Pierce).

Immunoprecipitation

After serum starvation for at least 48 h, PC-3 cells were stimulated with 7.5 mM calcium and at the indicated time points, cells were washed with ice-cold PBS and lysed with immunoprecipitation buffer containing 150 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 0.2 mM sodium vanadate, and protease inhibitors (as described above). The cell lysate was centrifuged at 10,000 × g for 10 min. For immunoprecipitation, equal amounts of protein were incubated with polyclonal EGFR antibody overnight, and then incubated with protein A/G agarose beads for a further 1 h at 4°C. Bound immune complexes were washed three times with immunoprecipitation buffer containing protease and phosphatase inhibitors and detergents. The pellets were eluted by boiling for 5 min with 2 × Laemmli sample buffer. Supernatant proteins were separated by 7.0% SDS-PAGE, transferred to nitrocellulose membranes, and immunoblotted with monoclonal anti-phosphotyrosine antibody (PY99). The stripped membrane was then reblotted with EGFR antibody.

DNA synthesis

Two methods for measuring DNA synthesis were used: [ 3 H]thymidine and BrdU uptake. H-500 cells were seeded in 24-well plates. The cells were cultured for 72 h and then serum-starved for 4 h, after which calcium (0.5-7.5 mM) alone or with NPS R-467, NPS S-467, PD98059, SB203580, or LY294002 was added along with [ 3 H]thymidine (1 μ l/ml; 50 μ Ci/ml), and the cells were cultured again for 24 h in 1 ml media. Incorporation of [ 3 H]thymidine in H-500 cells was measured by removing the medium and lysing the cells with 0.5 ml of 10% trichloroacetic acid. The resultant DNA pellet was dissolved in 0.5 ml of 200 mM NaOH. Incorporated radioactivity was measured by counting in a scintillation analyzer on a 3 H program (Tri-carb 2900TR, Packard Bioscience, USA). The effect of calcium on [ 3 H]thymidine was verified by using BrdU.

U-87 cells were seeded in 96-well plates and transfection with siRNA was performed using a cocktail consisting of OptiMEM (Invitrogen), siPort lipid, and 20 nmol oligonucleotide that was designed in our laboratory. U-87 cells were pulsed with 5-bromo-2'-deoxyuridine (BrdU) for 4 h, and its incorporation was measured using a kit obtained from Roche Diagnostic (Indianapolis, IN).

TUNEL assay

The TUNEL reaction was performed to detect apoptosis. Cells were plated on coverslips and treated as described for the DNA synthesis assay. The cells were stained with an in situ cell-death detection kit, (Roche Diagnostics, Indianapolis, IN) following the manufacturer's recommendations. Briefly, cells were washed with PBS once and fixed with 4% formalin for 5 min. After fixation, the coverslips were dried and then stored at -80°C. After the cells were washed with PBS, they were permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate for 2 min on ice. Slides were rinsed twice with PBS and then incubated for 60 minutes at 37°C with terminal deoxynucleotidyl transferase (TdT) enzyme in reaction buffer. The slides were rinsed three times with PBS and mounted with Vectashield (Vector Laboratories, Burlingame, CA). Samples were analyzed by confocal and fluorescence microscopy. TUNEL-positive nuclei were detected by the bright color in condensed or ruptured nuclei. Two researchers blinded to the status of the samples counted the apoptotic cells as well as the total number of cells in eight different fields from two independent experiments. The total number of cells ranged from 26 cells to 105 cells per field.

Northern blot analysis

To study effects on the expression of iNOS, PTHrP, PTTG mRNA, we performed Northern blot analysis as described before (38). Briefly, cellular RNA was isolated (50) using the Trizol reagent (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. The RNA recovered was quantitated by spectrophotometry, and aliquots of 20 μ g of total RNA from treated cells were loaded on a formaldehyde agarose gel after denaturation. The gel was stained with ethidium bromide to visualize RNA standards and ribosomal RNA, in order to document equal loading of RNA from the various experimental samples. The RNA was then blotted onto nylon membranes (Duralon, Stratagene, La Jolla, CA). Blots were hybridized with a cDNA probe for PTTG, PTHrP, and iNOS and washed under high stringency conditions as described previously (156). An iNOS cDNA probe was prepared by one-step RT-PCR using 2 μg total RNA derived from H-500 cells using the following primers: 5'- TGC TAT TCC CAG CCC AAC AAC -3' (iNOS sense, 120-140), 5'- TTT TGC CTC TTT GAA GGA GCC -3' (iNOS antisense, 486-466). The PCR product was then subcloned into the TOPO TA cloning kit (Invitrogen) following the manufacturer's instructions and sequenced to confirm its homology with the corresponding region of the rat iNOS mRNA (NM_ 012611.1). A full-length human PTHrP cDNA probe (1.7 kb) was generously provided by Dr. E. Schipani, Massachusetts General Hospital, Boston, MA and a rat PTTG probe was donated by professor Shlomo Melmed, Cedars-Sinai Research Institute, Los Angeles, CA. Equal loading was also confirmed by reprobing the membranes with GAPDH cDNA. Specific radioactive signals were analyzed on a Molecular Dynamics, Inc. PhosphorImager (Sunnyvale, CA) with the ImageQuant program.

Reverse Transcriptase-PCR

One-step RT-PCR (kit from Qiagen, Santa Clarita, CA) was used for determining the presence of: 1 PTTG transcript(s) using a pair of primers that would yield a 352-bp product spanning nucleotides 165 to 517 of the human PTTG cDNA (NM_004219). Primer sequences are; 1) sense - 5'-AGT TTC AAC ACC ACG TTT TGG C -3'; 2) antisense - 5'-GCT TTT CAA GCT CTC TCT CCT CG -3'; 2 PTHR1 transcript(s) using a pair of primers that would yield a 411-bp product spanning from nucleotides 181-592 of the rat PTHR1 cDNA (NM_020073). Primer sequences are 1) sense, 5'-CAG ATT TTC CTG CTG CAC CG-3'; 2) antisense, 5'-TGA ACT TGA GGC ACT CGC TGT-3'. In brief, we used the following procedure for RT-PCR: 2 μ g total RNA was mixed with a master cocktail containing RT-PCR buffer, sense and antisense PTTG primers, dNTPs, RNase inhibitor, and an enzyme mixture containing reverse transcriptase (Omniscript and Sensiscript) and HotStart Taq DNA polymerase at the concentrations recommended by the manufacturer (Qiagen) to a final volume of 50 μ l. The temperature-cycle protocol was as follows: 30 min at 50°C for RT reaction, followed by denaturation and activation of HotStart DNA polymerase for 15 min at 95°C, and PCR amplification (30 sec at 94°C, 30 sec at 58°C, and 1 min at 72°C for 40 cycles). A final extension for 10 min at 72°C was performed after the end of 40 cycles. In order to eliminate amplification from contaminating genomic DNA, we omitted RT as a negative control for the RT-PCR reaction for each sample. RT-PCR products were fractionated on 1.5% agarose gels. The presence of a 352-bp or 411-bp amplified product, respectively, was indicative of a positive PCR reaction arising from the presence of a PTTG- or PTHrP-related sequence within the cDNA.

Quantitative real-time PCR

To amplify human pituitary tumor transforming gene (PTTG1) (NM_004219), human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (J_02642), rat PTTG (NM_022391), rat vascular endothelial growth factor (VEGF) (NM_031836), rat inducible nitric oxide synthase (iNOS) (NM_012611.1) and rat GAPDH (M_17701) cDNA, sense and antisense oligonucleotide primers were designed based on the published cDNA sequences using the Primer Express ver. 2.0.0 (Applied Biosystems, Foster City, CA). Oligonucleotides were obtained from Genosys (Woodlands, TX). The sequences of the primers were as follows: 5'- CGG CTG TTA AGA CCT GCA ATA ATC -3' (human PTTG sense, 18-41), 5'- TTC AGC CCA TCC TTA GCA ACC -3' (human PTTG antisense, 119-99), 5'-TTC AAT GGC ACA GTC AAG GC-3' (human GAPDH sense), 5'-TCA CCC CAT TTG ATG TTA GCG -3' (human GAPDH antisense), 5'- ATG ACC CTG GCG TGA AGA TTT -3' (rat PTTG sense, 127-147), 5'- AAG CAG CAA CAG AGA CCA GAG C -3' (rat PTTG antisense, 227-206), 5'- AGC CTT GTT CAG AGC GGA GAA-3' (rat VEGF sense 500-520), 5'- TAA CTC AAG CTG CCT CGC CTT-3' (rat VEGF antisense 606-586), 5'-GAT TCA GTG GT CCA ACC TGC A-3' (rat iNOS sense, 621-641), 5'- CGA CCT GAT GTT GCC ACT GTT-3' (rat iNOS antisense, 738-718), 5'-TTC AAT GGC ACA GTC AAG GC-3' (rat GAPDH sense), and 5'-TCA CCC CAT TTG ATG TTA GCG-3' (rat GAPDH antisense). cDNA was synthesized with the Omniscript RT Kit (Qiagen, Valencia, CA) using 2 μ g total RNA in a 20 μ l reaction volume. For real-time PCR, the cDNA was amplified using an ABI PRISM 7000 Sequence Detection System (PE Applied Biosystems, Foster City, CA). The dsDNA-specific dye SYBR Green I incorporated into the PCR reaction buffer QuantiTech TM SYBR PCR (Qiagen, Valencia, CA) to allow for quantitative detection of the PCR product in a 25- μ l reaction volume. The temperature profile of the reaction was 95°C for 10 min, 40 cycles of denaturation at 95°C for 15 sec, annealing at 60°C for 30 sec, and extension at 72°C for 30 sec. An internal housekeeping gene control, GAPDH, was used to normalize differences in RNA isolation, RNA degradation, and the efficiencies of the RT. The size of the PCR product was first verified on a 1.5% agarose gel, followed by melting curve analysis. Inter-sample differences <0.5 cycles were accepted for calculating the average crossing point (Cp).

Microarray

Total ribonucleic acid (RNA) was quantified by measuring ultraviolet (UV) absorption ratio at 260/280 nm and checked for quality using a Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA, USA). Preparation of the biotin-labeled complementary RNA (cRNA) target was performed using the BioRobot 9604 (Qiagen, Valencia, CA, USA) and a PTC-225 DNA Engine Tetrad TM Cycler (MJ Research, Inc., Boston, MA, USA). Single-stranded cDNA was prepared from 2 μ g of total RNA using a T7-(dT) 24-oligonucleotide primer and Superscript TM II RNaseH-reverse transcriptase (200 U/ μ l). Included in this reaction was a mixture of six bacterial RNAs of known concentration for use as positive controls (2.5 pg/ml of araB/entF, 8.33 pg/ml of fixB/gnd and 25 pg/ml of hisB/leuB). Double-stranded cDNA was then generated with E. coli DNA polymerase I (10 U/ μ l) and RNase H (2 U/ μ l). After purification using a Qiagen QIAquick purification kit, the double-stranded cDNA served as a template to prepare biotin-labeled cRNA via in vitro transcription (IVT), performed in the presence of biotinylated nucleotides. The labeled-cRNA transcripts were purified using RNEasy columns (Qiagen) and assessed for quantity and quality using the same methods described above. The biotin-labeled cRNA was then randomly fragmented by incubating 2 μ g of the sample in the presence of magnesium for 20 min at 94°C. This resulted in fragmented target with a size range between 100 and 200 bases.

Hybridization and Scanning. The biotinylated cRNA target was hybridized to two ADME-Rat Expression Bioarrays (Motorola Life Sciences). For each array, 2 μ g of the fragmented target cRNA was added to 260 μ l of hybridization buffer, denatured, and then injected into hybridization chambers, sealed, and incubated for 18 h at 37°C while shaking at 300 rpm. Each array was rinsed in a stringent 46°C wash in 0.75× TNT solution for 60 min, followed by Streptavidin labeling for 30 min (RT), then the 0.75× TNT and 0.05% TNT in series. The slides were spin-dried in an Eppendorf 5810R centrifuge (2000 rpm for 3 min) (swinging bucket rotor). Processed arrays were scanned using an Axon GenePix Scanner, and array images were acquired and analyzed using CodeLink TM Expression Analysis Software.

Motorola cDNA chip data analysis. The "normalized intensity" probe data generated by the CodeLink Expression Analysis Software (Amersham) were exported into Microsoft Excel. The data were then separated into two classes: high dose and low dose. The Excel function TTEST (low dose data, high dose data, 2, 2) was applied to each gene probe. Probes with p-values greater than the PVAL-THRESH of 0.05 were eliminated. The Excel function AVERAGE (high dose data)/ AVERAGE (low dose data) was next applied. Probes with ratios between the RATIO-THRESH of 2.0 and 1/RATIO-THRESH of 0.5 were removed. The remaining probes were candidates for significantly changed mRNAs. These genes demonstrated acceptable p-values and exhibited at least a two-fold change between the high- and low-dose calcium treatments.

Messenger RNA silencing

U-87 cells were plated in 96-well plate with 60-70% confluency. Twenty four hours after plating, cells were transfected with either negative control or two different PTTG mRNA silencing oligonucleotides, purchased from Ambion (Austin, TX). For human PTTG RNA silencing (siRNA), we first tested two different siRNA oligonucleotide sequences designated as PTTG1.1 and PTTG1.2. The sense and antisense sequences used were: PTTG1.1, 5'-GAU CUC AAG UUU CAA CAC Ctt-3' (sense) and 5'-GGU GUU GAA ACU UGA GAU Ctc-3' (antisense); PTTG1.2, 5'-GUC UGU AA A GAC CAA GG GAtt-3' (sense) and 5'-UCC CUU GGU CUU UAC AGA Ctt-3' (antisense). For a negative control, we used the following oligonucleotides (Ambion): 5'-AGU ACU GCU UAC GAU ACG Gtt-3' (sense) and 5'-CCG UAU CGU AAG CAG UAC Utt-3' (antisense). Chemically synthesized annealed oligonucleotide of the above mentioned PTTG siRNA sequences were used. The 100-nm final concentration of siRNA sequences were used for transfecting the U87 cells. Efficacy of silencing was determined by real-time PCR of human PTTG gene 48 h post transfection, and PTTG1.2 siRNA sequence was found to have 70-80% efficiency in reducing human PTTG mRNA, compared with negative control, whereas PTTG (1.1) was less than 50% efficient. Therefore, we used PTTG (1.2) to study the role of PTTG in the growth of U87 cells.

Measurement of Ca 2+ i by fluorimetry in cell populations

Coverslips with H-500 cells were loaded for 2 h at room temperature with Fura-2/AM in 20 mM HEPES, pH 7.4, containing 125 mM NaCl, 4 mM KCl, 1.25 mM CaCl 2 , 1 mM MgSO 4 , 1 mM NaH 2 PO 4 , 0.1% bovine serum albumin (BSA), and 0.1% dextrose and then washed once with a bath solution (20 mM HEPES, pH 7.4, containing 125 mM NaCl, 4 mM KCl, 0.5 mM CaCl 2 , 0.5 mM MgCl 2 , 0.1% dextrose, and 0.1% BSA) at 37°C for 20 min. The coverslips were then placed diagonally in a thermostatted quartz cuvette containing the bath solution using a modification of the technique employed previously in this laboratory (12). In the experiment with agonists for other G-protein-coupled receptors, angiotensin II was used at 10 -8 M, ADP at 10 -8 M, thrombin peptide agonist at 10 -5 M, and carbachol at 10 -4 M. The angiotensin II, ADP, thrombin peptide agonist, and carbachol were added into the bath solution. Excitation monochrometers were centered at 340 and 380 nm, and emission light was collected at 510 ± 40 nm through a wide-band emission filter. The 340/380-excitation ratio of emitted light was used to estimate Ca 2+ i as described previously (12).

Biological NO imaging by DAF-FM Diacetate

H-500 Cells were plated on coverslips in a 6-well plate. After 72 h at 70-80% confluency, the cells were challenged with 0.5 mM or 7.5 mM calcium for 18 h in serum-free, Ca 2+ -free DMEM containing 4 mM L-glutamine and 0.2% BSA. The cells were then washed twice with RPMI 1640 media without phenol red. The cells were then loaded for 1 h with 1 ml RPMI 1640 (without phenol red) containing 10 μ M DAF-FM diacetate (4-amino-5-methylamino- 2',7'-difluorofluorescein diacetate) (Molecular Probe, Eugene, OR), 0.1% pluronic acid, and 1 mM probenecid at 37°C in a 5% CO 2 incubator with no light. Finally, the coverslips were washed twice in RPMI 1640 media without phenol red.

Direct visualization of NO production. The coverslips incubated in 37°C RPMI 1640 media without phenol red were placed horizontally under the microscope lens. Photomicrographs with the fluorescent NO indicator were acquired with a laser scanning confocal microscope (LSCM, Leica TCS-NT, Heidelberg, Germany) equipped with an argon-krypton laser at an excitation wavelength of 488 nm and a band-pass filter for 500-550 nm (250). Simultaneous visualization of cell morphology by differential interference contrast (DIC) microscopy was performed to confirm equal cell numbers on the coverslips. The fluorescence images were obtained as a 1024 × 1024 pixel frame. All other settings, including scanning speed, pinhole diameter, and voltage gain, remained the same for all experiments. The images were stored on magneto-optical storage devices.

Measurement of NO by fluorometry in cell populations. The coverslips were placed diagonally in a thermostatted quartz cuvettes containing 37°C RPMI 1640 medium without phenol red. Excitation monochrometers were centered at 490 nm, and emission light was collected at 520 ± 40 nm through a wide-band emission filter.

Statistics

The data are presented as the mean ± SEM of the indicated number of experiments equal to N. Data were analyzed by one-way ANOVA followed by Fisher's PLSD test or by Dunnett's multiple comparison test or Student's t-test when appropriate. A P value of <0.05 was considered to indicate a statistically significant difference unlikely due to chance.

Hormone systems and receptors involved in extracellular calcium homeostasis

PTH and PTH receptor

As mentioned above PTH, calcitonin, and 1,25(OH) 2 vitamin D 3 are the three major Ca 2+ o -regulating hormones, the so-called "calciotropic hormones" (260, 262). The overall action of PTH on calcium homeostasis is to increase the level of Ca 2+ o . Early studies showed an inverse sigmoidal relation between Ca 2+ o and PTH (21, 86, 228). This inverse sigmoidal relation suggests a negative feedback from Ca 2+ o to the parathyroid gland and subsequentially the release of PTH. The molecular mechanism for this feedback, the CaR, was indeed discovered in 1993 (24). PTH release is mainly inhibited by high calcium, low phosphorus and low 1,25(OH) 2 vitamin D 3 . Parathyroid hormone (PTH) is an 84 amino acid polypeptide. Its mRNA, in addition to encoding the mature peptide, also encodes a `pre' or signal sequence of 25 amino acids and a basic `pro' peptide of 6 amino acids. The PTH gene has been mapped to chromosome 11 (176). PTH exerts its actions through the PTH receptor (PTHR1), which like the receptors for peptide hormones such as secretin, vasoactive intestinal peptide, glucagon and calcitonin belongs to class B of the seven transmembrane receptors, also termed G -protein-coupled receptors (139). PTHR1 has an extracellular amino terminal of about 185 amino acids, seven transmembrane spanning helixes (approximately 290 amino acids), and an intracellular domain, the carboxy-terminal cytoplasmic tail, of 110 amino acids (160). The PTH peptide is thought to act on the PTHR1 mainly in kidney and bone, although the receptor has been found in many organs not involved in calcium homeostasis, such as rat aorta, heart, bladder, stomach, ileum, and pregnant uterus. Furthermore, knock out mice with no PTHR1 die intrauterine and studies have suggested an association with abrupt cardiomyocyte death, suggesting that this receptor is important for the normal intrauterine development of the heart (201). The PTH peptide shares significant amino acid homology with parathyroid hormone related peptide (PTHrP), especially in the first 13 residues. PTHrP exists in 3 isoforms in humans: 139,141, and 173 amino acids. PTHrP is frequently expressed in the same cells that express the PTH receptor, or in cells immediately adjacent to them. This spatial proximity of PTHrP to the receptor, together with the fact that little if any PTHrP circulates under normal physiological conditions, have suggested that PTHrP acts as a paracrine or autocrine factor. PTHrP has several important physiological roles encompassing promotion of calcium mobilization from bone during lactation (278) transport of maternal calcium across the placental membrane (141), and the regulation of chondrocyte growth and differentiation in the growth plates of developing long bones (144).

Vitamin D and Vitamin D receptor

The overall effect of vitamin D on calcium homeostasis is to increase the level of Ca 2+ o ; its actions are slower than the rapid effects of PTH. Vitamin D is not a vitamin as such, since a vitamin is generally defined as a compound that the body is not able to produce. In the skin 7-dehydrocholesterol is converted to previtamin D 3 by ultraviolet radiation (113). Previtamin D 3 is subsequently spontaneously converted to vitamin D 3 , cholecalciferol. Vitamin D from the diet is in two forms cholecalciferol and ergocalciferol. Vitamin D is hydroxylated in the liver to 25(OH)-(ergo- or) cholecalciferol. There is little evidence that the active forms of ergocalciferol and cholecalciferol differ in their mode of action, however, more is known about vitamin D 3 . Lastly 25(OH)-cholecalciferol is hydroxylated in the proximal tubular cells of the kidney to the active metabolite, 1,25(OH) 2 -cholecalciferol (1,25(OH) 2 vitamin D 3 ), by the enzyme 1-alpha-hydroxylase (also termed CYP27B1) (148) and the less potent metabolite 24,25(OH) 2 -cholecalciferol by the 24-hydroxylase enzyme (also termed CYP24) (271). The 1-alpha-hydroxylase is the enzyme that will determines the rate of production of 1,25(OH) 2 vitamin D 3 . 1-alpha-hydroxylase is upregulated by several factors, the most important are: high serum parathyroid hormone (PTH), hypocalcemia, hypophosphatemia and low 1,25 vitamin D 3 . 1,25(OH) 2 vitamin D exerts its effects on calcium homeostasis through the vitamin D receptor (VDR). The VDR is a sterol receptor that is mainly found in the nucleus of the cell. The receptor is widely expressed throughout the cells of the body, but the effects of the VDR on calcium homeostasis are mainly in the epithelial cells of the intestine and kidney. This was demonstrated in the VDR knock out mice. Bone formation and mineralisation were found by two groups to be severely impaired in VDR knockout mice (150, 309). However, when the mutant mice were maintained on a rescue diet (calcium/ phosphorous/ lactose enriched) they had a normal development, and the impaired bone formation and mineralization were rescued (8). As the nuclear receptor is expressed in most cells, including many that are not involved calcium homeostasis, 1,25(OH) 2 vitamin D has been found to have other physiological roles that are less well understood, including cytokine production, cell growth and differentiation and secretion of hormones (51).

Calcitonin and calcitonin receptor

The third and last calciotropic hormone is the calcitonin (CT). The overall effect of CT in calcium homeostasis is to reduce the level of Ca 2+ o . The parafollicular cells (C cells) of the thyroid gland produce and release the hormone. CT is a 32 amino acid peptide with an N-terminal disulfide bridge and a C-terminal prolineamide residue. CT receptors are seven transmembrane receptors in superfamily B, and share homology with PTHR1 (151). In the case of the human CT receptors at least five splice variants have been described (145). In pharmacological doses CT reduces bone resorption and renal reabsorption of calcium (267), but the significance of these actions in the normal calcium homeostasis is still debated. Data that CT may not be redundant in calcium homeostasis came from studies in mice. CT knockout mice exhibited no identifiable developmental defects at birth, and had normal baseline calcium values. However, the knockout mice responded more to exogenous human PTH as evidenced by a greater increase in serum Ca 2+ and urine deoxypyridinoline crosslinks, an effect that could be prevented by CT (111). Surprisingly the mice had higher trabecular bone volume. These findings support a hypothesis that CT inhibits bone resorption in the face of an acute hypercalcemic challenge but may also imply that calcitonin is important in regulation of normal bone formation. As described above there is a positive relationship between Ca 2+ o and the release CT. This positive feedback loop of calcium on CT secretion is most likely mediated by the CaR (82, 165).

Disease with mutations in the CaR

The nonredundant physiological importance of the calcium-sensing receptor in humans was proven by the discovery of diseases caused by mutations in the receptor that lead to either loss-of-function or gain-of-function of the protein (105). Heterozygous (i.e., the mutation is present in only one allele and the other allele carries wild type CaR) gain-of-function mutations cause autosomal dominant hypoparathyroidism (ADH). Most patients have asymptomatic hypocalcemia with relative or absolute hypercalciuria. Heterozygous loss-of-function mutations give rise to familial hypocalciuric hypercalcemia - FHH, also termed familial benign hypocalciuric hypercalcemia (FBHH) by somein which most patients have an asymptomatic form of hypercalcemia with relative or absolute hypocalciuria. The homozygous or compound heterozygous variant of inactivating CaR mutations, on the other hand, produce neonatal severe primary hyperparathyroidism (NSHPT), a severe, sometimes deadly disease if left untreated. The mouse models with heterozygous or homozygous inactivation of the CaR gene have phenotypes comparable to the respective human conditions (108). Thus one way in which the mutated receptors may exert their effects on calcium homeostasis is through a reduced level of expression of wild type receptors, as in the heterozygous knock out mice.

Clinically FHH is mostly a benign state of hypercalcemia with few or no symptoms. The diagnosis of FHH is made in a patient with a family history of mild to moderate hypercalcemia averaging approximately 2.75 mM (total calcium), with an inappropriately low rate of urinary calcium excretion. The best distinction between FHH and other forms of hypercalcemia is often made by determining the ratio of calcium clearance to creatinine clearance (Ca/Cr). A value below 0.01 is found in about 80% of cases of FHH, while a similar proportion of cases of primary hyperparathyroidism caused by hyperplasia or parathyroid adenoma have values higher than this (23). While benign and asymptomatic hypercalcemia is typical of most FHH families, two families have been found to have calcium concentrations that average 3.13 and 3.35 mM, causing a neonatal severe hyperparathyroid-like state in some affected infants (11, 12). Due to the lack of symptoms in most cases, patients with FHH are often undiagnosed until a routine blood sample shows an unexpectedly high serum calcium level or family screening is done due to the birth of a child with NSHPT (163). Patients with FHH usually have normal PTH levels, despite the hypercalcemia, although in some cases the PTH levels are elevated (32, 101).

Neonatal severe primary hyperparathyroidism is clinically overt in most cases within the first half year. The infant has, as indicated by the name, severe, symptomatic hypercalcemia driven by PTH, as well as bony changes characteristic of hyperparathyroidism. Infants with NSHPT may have hypotonia, polyuria, dehydration, and failure to thrive. The mass of the parathyroid glands in NSHPT is often increased dramatically. Pathological investigation reveals chief cell hyperplasia. The hallmark of the disease is the associated hyperparathyroid bone disease. This bone disorder often leads to multiple low energy fractures (65, 87). Fractures in the ribs can produce a "flail chest" syndrome that secondarily causes respiratory difficulties. Biochemical analysis reveals hypercalcemia, hyperparathyroidism, and relative hypocalcuria (55). In the literature, levels of serum total calcium are reported to be moderately elevated (e.g., 3-3.25 mM) to as high as 7.7 mM in the most severe cases (23). PTH levels are often observed to be 10-fold higher than the upper limit of normal. Early diagnosis is crucial, as untreated NSHPT can be a severe neurodevelopmental disease that may be lethal when not treated surgically with total parathyroidectomy (55). Lately a wider clinical spectrum for the condition has become apparent; and the growing availability of genetic testing of the CaR gene has documented that some infants have milder hyperparathyroidism and a distinctively less severe clinical presentation and natural history than that just described. In these patients with the calcimimetics (CaR agonists, reviewed later) in theory might represent a means of lowering the serum calcium concentration and determining whether the patient derived any symptomatic benefit, thereby potentially providing a long term, effective medical therapy.

Patients with autosomal dominant hypoparathyroidism have an inherited form of hypocalcemia and are often clinically asymptomatic, similar to most patients with FHH. As stated above the genetic background is a gain of function mutation of the CaR. A few patients may develop seizures, neuromuscular irritability and calcification of the basal ganglia. A paraclinical landmark of the disease is the mild to moderate hypocalcemia, with serum PTH levels that are inappropriately within the lower half of the normal range or frankly subnormal (as opposed to normals who respond vigorously to the hypocalcemia) (197). They often have relative or absolute hypercalciuria, with normal or elevated levels of urinary calcium excretion, respectively, despite the low serum calcium. Their renal calcium excretion has been found to be higher than that of patients with typical hypoparathyroidism in some studies, and, analogous to the diagnosis of FHH, a high urinary Ca/Cr ratio (e.g., the ratio of the calcium to creatinine concentrations in the urine) may be better than hypocalcemia alone for differentiating ADH from primary hypoparathyroidism. However, not all studies showed this difference in the level of urinary calcium excretion between idiopathic hypoparathyroidism and ADH (182, 298). It is important to prevent renal complications such as nephrocalcinosis, nephrolithiasis, and renal impairment. These renal complications are often iatrogenic and caused primarily by the hypercalciuria where the clinician has tried to correct the low serum calcium with calcium and vitamin D treatment. Therapy with calcium supplements and vitamin D metabolites should be administered only to patients with symptomatic ADH, and the aim should be to increase the serum calcium concentration only to a level that will make the patient asymptomatic (189). The development of the calcilytic class of drugs (a CaR antagonist, reviewed later) may, in theory at least, provide an optimal treatment of symptomatic ADH patients by "resetting" the parathyroid and kidney to maintain a more nearly normal serum calcium concentration without hypercalciuria.

Is CaR the only calcium sensing mechanism?

Recently an orphan receptor, GPRC6A, in the superfamily C of the 7TM's was deorphanized and was found to resemble the CaR in its amino acid homology and several of its pharmacological properties (287). GPRC6A, like the CaR, is sensitive towards certain L-amino acids, but unlike the CaR, GPRC6A is a sensor of basic amino acids (288). The CaR senses aromatic amino acids most effectively (58). It was subsequently shown that this receptor is also sensitive towards extracellular calcium (albeit at high concentrations) and calcimimetics (194), allosteric activators of the CaR, which may implicate GPRC6A as a second calcium-sensing receptor (CaSR2). Another possible mechanism underlying some calcium-induced effects, e.g. in osteoblasts, is the calcium-binding intracellular protein, calcyclin (273). Lastly it should be noted that calcium could also have effects through ion channels such as L-type calcium channels, this is indeed the case in calcium regulated release of calcitonin by C-cell of the thyroid in vitro (165). Although data that the CaR is crucial for the calcitonin release is quite solid the L-type calcium channel may play an additional role (72), the data suggest that the L-type calcium channel is a calcium sensing mechanism involved in the calcium homeostasis.

Cloning and structure

The CaR was discovered, by an expression cloning technique, in Xenopus laevis oocytes, using a bovine cDNA library. A single clone of 5.3-kb showing the same pharmacological properties as the Ca 2+ o -sensing apparatus was found in a cDNA library prepared from mRNA extracted from bovine parathyroid glands (24). Soon afterward, the nucleic acid hybridization technique was used to clone the CaR in humans (6, 81), rats (210), mice (115), and rabbits (31). The nucleotide sequences of these receptors are approximately 85% identical to the original bovine parathyroid CaR, while the amino acid sequences are more than 90% identical (using http://www.ncbi.nlm.nih.gov/BLAST/), implying that functionally important mutations in the receptor are discarded through evolution. The CaR belongs to family C II of the superfamily of the seven transmebrane receptors (26). The 7TMs constitute the largest group of cell surface membrane receptors, and they carry significant importance in clinical medicine, as approximately 50% of the current prescription drugs target the 7TMs. The human CaR is 1078 amino acid residues long and, like all 7TMs, has three structural domains ( Figure 1 ): a hydrophilic N-terminal end of 612 residues that is an unusually large extracellular domain (ECD) characteristic of the family C receptors of the 7TMs; a hydrophobic transmembrane domain (TMD) of 250 amino acids, and an intracellular domain (ICD) 216 amino acids long, which is the hydrophilic C-terminal end of the protein. The receptor is modified by N-linked glycosylation, which is important in promoting the cell membrane expression of the receptor. The cell surface CaR consists primarily of a dimer linked by covalent disulfide bonds involving two cysteine residues in each monomer (e.g., cys129 and cys131) (13, 69, 124). The location of the binding sites for Ca 2+ o in the CaR is, at present, somewhat of a conundrum. The ECD is thought to contain one or more binding sites for Ca 2+ o in the CaR, but a mutated CaR expressed without the ECD also responds to Ca 2+ o and other polyvalent cations, implying that the TMD also senses Ca 2+ o (116, 209).

Intracellular signaling apparatus of the CaR

There are five protein kinase C (PKC) (Figure 1) and two protein kinase A (PKA) phosphorylation sites in the CaR (14). Phosphorylation of these PKC sites inhibits activation of phospholipase C (PLC) by the CaR, a major downstream mediator of the receptor's biological responses; this phosphorylation thereby acts as a negative feedback system, since PKC is also activated by CaR and is often situated downstream of PLC. The ICD directly or indirectly binds the scaffolding proteins filamin-A, beta arrestin and caveolin (10, 107, 135, 195). All three also bind to signaling partners activated by the CaR. This binding to filamin-A was recently found to inhibit the degradation of the CaR (312).

The CaR, like other 7TMs, acts through G-proteins. The first step following activation of the receptor is the release of a guanosine diphosphate (GDP) bound to the G-protein complex of G αβγ . Subsequently, the G αβγ complex is bound to the ligand-bound 7TM and activated, which releases the complex from the receptor. Then GTP binds to G α in the G αβγ complex, and the activated G α is dissociated from G βγ . There are three main groups of G α subunits: 1) G α s which activates adenylate cyclase (AC) to produce cyclic adenosine monophosphate (cAMP) from ATP; 2) G α i , which inhibits AC and inward calcium channels and activates outward potassium channels; and 3) G α q , which activates PLC and thereby the production of inositol trisphosphate (IP 3 ) and diacylglycerol (DAG). Before the CaR was cloned, it was recognized that activation of parathyroid cells with high calcium induced PLC activity and thereby IP 3 production and also inhibited hormone-dependent cAMP production, suggesting that the CaR activated G α q as well as G α i (45, 97). Counter intuitively, two studies have suggested that the CaR is linked to the G α s signaling pathway in growth hormone- and ACTH-producing producing pituitary adenoma cells respectively (66, 219), although this has not been confirmed by others, it might be due to stimulation of a calcium stimulated adenylate cyclase not through G α s .

A wide range of intracellular signaling pathways of the CaR have been recognized using a heterologous system comprising human embryonic kidney (HEK) cells with or without stably transfected CaR (HEK-CaR) ( Figure 2 ). Many of these intracellular signaling pathways have also been shown to be active in CaR-mediated signaling in other cells naturally expressing the CaR. In parathyroid cells and HEK-CaR cells, the CaR activates phospholipase (PL) A 2 , C, and D (134). PLC produces IP3, which then activates the IP3 receptor in the endoplasmic reticulum, resulting in release of calcium from intracellular stores, thereby producing spikes in the cytosolic free calcium concentration (Ca 2+ i ). The elevation in Ca 2+ i is sustained by the opening of calcium-dependent calcium channels on the cell membrane. The first step in the production of inositol lipids is the activation of phosphoinositol 4-kinase, which converts PI to PI 4 P. The CaR has been shown to activate phosphoinositol 4-kinase through G α q in parallel to activation of PLC in HEK-CaR cells (117). Mitogen-activated protein kinases (MAPKs) are also activated by phosphorylation of their respective upstream kinases. MAPK is an important intracellular signaling pathway that often acts through the nucleus, e.g., in the regulation of the cell cycle, but it can also regulate events close to the cell membrane, such as release of peptides. Kifor et al. have shown that the CaR in HEK-CaR cells and parathyroid cells activates MAPKs. More specifically, the CaR activated extracellular signal regulated kinase (ERK1/2, also termed the p42/44 MAPK), which, in turn, phosphorylated and activated PLA 2 (136). Soon after that finding, Handlogten et al. showed that the CaR promotes ERK1/2 phosphorylation in HEK-CaR cells; as a control, they used HEK cells stably transfected with a dominant-negative CaR (Arg796Trp) (93). They also showed that the CaR activates PLA 2 through G α q , PLC, calmodulin, and calmodulin-dependent kinase but not through Gas or ERK1/2 in HEK-CaR cells. The difference in the results of these two studies regarding the role of ERK1/2 in PLA 2 activation remains to be clarified. In cells expressing the CaR at a lower level than in parathyroid cells and HEK-CaR cells, MAPK and phosphatidylinositol 3-kinase (PI3K), a classical prosurvival pathway, have both been found to be activated by the CaR. The role of the CaR and these pathways in testicular Leydig cell cancer, prostate cancer, breast tissue, and astrocytoma cells will be discussed in detail later.

CaR expression in primary hyperparathyroidism

The expression of the CaR in the body is highest in the chief cells of the parathyroid glands. The role of the CaR in parathyroid neoplasia has, therefore, been found to be extensive. Adenomas are the most common cause of primary hyperparathyroidism (HPT). In primary HPT, the calcium set point of the pathologic parathyroid gland(s) is shifted to the right, as in familial hypocalciuric hypercalcemia, but the alteration in Ca 2+ o -regulated PTH release is not due to loss-of-function mutations in the CaR gene (114). All human studies (70, 84, 131, 137, 303) but one (80) have found that the amount of the CaR protein expressed on the cell surface and/or the amount of its mRNA are downregulated in adenomas from patients with primary HPT.

In a mouse model of primary HPT it was found that the parathyroid glands proliferated and the CaR was downregulated before there was HPT biochemically (159). Whether this observation is an association or CaR downregulation is playing a part in the proliferation of the glands is discussed below.

CaR expression in secondary hyperparathyroidism

In secondary hyperparathyroidism the etiology of the disease is not the parathyroid gland but may be due to renal insufficiency or lack of vitamin D. In hyperplastic parathyroid glands from patients with secondary HPT (e.g., in patients with renal failure), as for primary HPT, the amount of the CaR protein expressed on the cell surface and the amount of its mRNA are downregulated (84, 162, 276). In a human study, the parathyroid glands of 56 subjects, 52 from uremic patients with secondary HPT and 4 from normal subjects, were studied for CaR content. This study showed an association of decreased CaR protein with enhanced proliferation of parathyroid cells in secondary HPT by immunochemical techniques (302). This observation suggests that the CaR may participate in the genesis of the hyperplasia. To study this in detail the field has used animal models.

In a rat study using a model of secondary HPT, 5/6 nephrectomized rats received a high-phosphate diet to produce uremic secondary HPT; downregulation of the CaR was found at both the RNA and protein levels, suggesting an association between the decrease in CaR expression and parathyroid proliferation as seen in humans (22). Ritter et al. (213), however, showed that parathyroid cellular hyperplasia precedes the downregulation of the CaR in the rat 5/6 nephrectomized model of secondary HPT. The same group found that feeding a low-phosphate diet to the rats with secondary HPT reversed the downregulation of the CaR and the secondary HPT, but again, this was preceded by normalization of PTH and parathyroid-gland growth (214). Lewin et al. (149) performed isogenic kidney transplantation in rats with uremic secondary HPT and found that the PTH levels normalized rapidly despite the continued downregulation of both the CaR and the vitamin D receptor in the same rat model. This study suggests that normalization of levels of the CaR and vitamin D receptor are not necessary to reestablish the normal set point of the parathyroid glands. In contrast to these findings, increased parathyroid mass and manifest hyperplasia were reported in CaR knock-out mice, a finding that favors the involvement of the CaR in the inhibition of parathyroid cellular proliferation (108). Recently, a Japanese report showed that the calcimimetic NPS R-568 reversed the downregulation in CaR expression on the parathyroid glands in the rat model of uremic secondary HPT (170). The cause-and-effect relationships of the CaR, parathyroid hyperplasia, and altered set point in the pathophysiology of primary and secondary HPT, however, require further study. In the parathyroid glands, there are two mRNA transcripts of the CaR; they arise from two different promoters (48). The transcripts differ in the reading of exon 1, and a study showed that one transcript (type 1A) is downregulated in parathyroid adenomas. The significance of the splice variants in parathyroid gland adenomas and other forms of neoplasia is presently unknown, but could be a part of the explanation of the conflicting data.

CaR expression in non-parathyroid adenomas

The CaR has been shown to be expressed in human gastrinomas, with the levels of expression varying in the eight samples investigated by quantitative PCR (83). Gastrinomas are notorious for being difficult to diagnose, and a widely used diagnostic method is to infuse calcium intravenously and study its effect on the gastrin levels. Itami et al. incubated CaR-expressing gastrinoma cells in vitro with calcium and observed an enhanced release of gastrin that could not be obliterated by the calcium channel blocker nifedipine (122). Also favoring a role for the CaR in the calcium-induced release of gastrin in gastrinomas is the finding of a small Japanese study of healthy humans that the calcimimetic KNR568 increased serum gastrin levels transiently (118). In a clinical study investigating the effects of oral calcium and peptone (amino acids, type II CaR agonists) in women with absorptive hypercalciuria and in normal controls, Bevilacqua et al. found that the effects of calcium and peptone on gastrin release were increased in the hypercalciuric group, indicating the presence of more-responsive CaR in these patients than normals (16). Taken together, these data strongly suggest that CaR is the mediator of the calcium-induced gastrin release. In a related tumor, the insulinoma, the CaR is likewise reported to be expressed (132, 140). Normal human β -cells of the pancreas also express the CaR and respond to an increase in Ca 2+ o by decreasing insulin release in response to low and high glucose levels in one study (243) and to increase insulin release in other studies (88). Two studies found that the responses to an increase in Ca 2+ o were opposite in insulinomas and normal β -cells, i.e., an increase in Ca 2+ o leads to an increase in insulin release in insulinomas (133, 291). The effect of Ca 2+ o seems to be mediated through the CaR, as nifedipine did not change the action of Ca 2+ o (140). In a study on human pituitary adenomas the CaR has been reported to expressed in nonfunctioning and growth hormone adenomas (219). The type I agonists of the CaR induced increase in cAMP, and in growth hormone producing adenomas calcium augmented the effect of growth hormone releasing hormone on growth hormone release from the adenoma cells. These data shows that the CaR is active in at least three endocrine adenomas. Whether a functional CaR on these adenomas is important in the development and progression of the diseases is unknown. It could be speculated from the data on pituitary adenomas that the CaR somehow modulates the effect of growth hormone releasing hormone on growth hormone release; this could have clinical implications in the diagnosis and perhaps treatment of the diseases. Whether the effect of a CaR-specific agonist, such as AMG 073, on gastrin, insulin, and growth hormone release will be useful in clinical tests for adenomas remains to be tested. CaR expression has also been found on premalignant adenomas, e.g. from the colon; the literature on these adenomas will be described in the section on malignant tumors below.

CaR and PTHrP in prostate cancer

Prostate cancer is known to metastasize frequently to bone. PTHrP has been described as a central mediator of osteolysis and growth of bone metastases (90). Sanders et al. reported the expression of CaR mRNA and protein in two human prostate cancer cell lines, PC-3 and LnCaP (223). The PC-3 cells responded to stimulation with Ca 2+ o and type I agonists with a dose-dependent response, approximately two-fold increase in PTHrP release. To further prove that the CaR is involved in the Ca 2+ o -evoked PTHrP release, they infected the PC-3 cells with the dominant-negative Arg185Gln CaR in a recombinant adeno-associated viral (rAAV) vector. The infected cells showed a blunted response to Ca 2+ o as compared with that of cells infected with empty vector. This result strongly supports the claim that the CaR is the mediator of the Ca 2+ o -induced PTHrP release in PC-3 cells. In the extracellular bone matrix, there are embedded growth factors (90). One of these growth factors, transforming growth factor β (TGF β ), was found to act additively with Ca 2+ o on the release of PTHrP by the PC-3 cells. This leads to a potentially destructive cycle, with Ca 2+ o and TGF β as stimulators of PTHrP release, owing to the release of these two substances from the resorbed bone. Of note, the concentration of Ca 2+ o in the bone microenvironment in HHM can reach much higher levels than those of systemic Ca 2+ o . In fact, the levels of Ca 2+ o in the immediate vicinity of resorbing osteoclasts can reach concentrations as high as 8 to 40 mM (239). We found in a later study that, in same PC-3 cell line, MAPKs are involved in the CaR-mediated PTHrP release (301). Serendipitously, we found that the CaR activates ERK1/2 through a triple membrane passing signaling pathway (TMPS) in the PC-3 cells. 7TMs use the epidermal growth factor receptor (EGFR) or other tyrosine kinase receptors as an intermediate signaling molecule in a process termed TMPS (290). A brief description of TMPS; the 7TMs activate intracellular signaling molecules, e.g., src or reactive oxygen species (274), which, in turn, activate a membrane-bound metalloprotease (200, 299) to cleave membrane-bound heparin-binding EGF (HB-EGF) or TGF α . The released HB-EGF then activates its receptor (the EGFR), which subsequently phosphorylates its own intracellular domain and activates downstream proteins such as MAPKs (155). ERK1/2 is maximally activated by the CaR at 30 minutes in PC-3 cells. These studies used NPS R-467 to show that the CaR activated the ERK1/2 pathway. Another non-trivial approach could have been the infection with a virus containing a dominant negative CaR. This method has been used in later studies on the phosphorylation of p38 MAPK (308). Inhibiting the Ca 2+ -induced phosphorylation of ERK1/2 by inhibiting metalloproteinases (GM6001), neutralizing the extracellular domain of the EGF and the EGFR with antibodies, or inhibiting the EGFR with a specific inhibitor (AG1478) all showed that the CaR utilizes the EGFR as a downstream signaling molecule in the activation of ERK1/2. The platelet-derived growth factor receptor (PDGFR) inhibitor, AG1296, had no effect on Ca 2+ o -induced ERK1/2 phosphorylation, showing that the CaR specifically used the EGFR in the TMPS mechanism. The inhibitors of MMP and EGFR described above had the same effect on Ca 2+ o -induced PTHrP release; furthermore, the potency of the MEK1/2 inhibitor, PD98059, on PTHrP release induced by Ca 2+ o was the same as that of the EGFR inhibitor (AG1478). These results suggest that CaR-mediated PTHrP release occurs through the EGFR and ERK1/2 signalling pathways in PC-3 cells. Similar results showing that the CaR uses EGFR as a mediator in high Ca 2+ o -induced ERK1/2 phosphorylation were also observed in H-500 and HEK-CaR cells as well, suggesting that this may be a widely used pathway by the CaR (157, 264). The participation of the EGFR pathway in CaR signaling makes the EGFR a promising drug target for treatment of prostate cancer, since targeting the CaR itself could jeopardize calcium homeostasis. The PC-3 cells are a model of malignant hypercalcemia as described initially. Another choice could have been the breast cancer cell line MDA-MB-231. They have also been shown to induce malignant hypercalcemia due to bone metastases with a functionally active CaR, Calcium induces PTHrP release from these cells (225). The PC-3 cells were chosen since the evidence for CaR-mediated PTHrP release was stronger, as the calcium induced PTHrP release in PC-3 cells could be abolished by the dominant negative CaR described above, whereas involvement of the CaR in MDA-MB-231 cells was suggested by effects of type I agonist of the CaR. To my knowledge it was never investigated whether the level of CaR protein in prostate tumors is correlated to malignancy and or bone metastasis. This could be interesting because an increased expression in higher malignancy grades would be suggestive of CaR as an active player in the progression of the diseases.

CaR and PTHrP in breast cancer

Breast tissue is active in calcium homeostasis during pregnancy and lactation and must be able to concentrate the calcium in milk; these observations have led to studies investigating a possible role for the CaR in breast tissue. Cheng et al. found CaR expression in normal, fibrocystic, and ductal carcinoma mammary epithelial cells (46), and Sanders et al. found the CaR in the human breast cancer cell lines MDA-MB231 and MCF-7 (225). Stimulation of MCF-7 cells with calcium and neomycin induced Ca 2+ i spikes, as measured by Fura-2 (187). Confocal microscopic studies showed that the CaR and calbindin-D-28k co-localized when the CaR was stimulated, suggesting that calbindin-D-28k, which binds calcium, is recruited to the CaR. Stimulating the MDA-MB231 cells with calcium and the type I CaR agonist neomycin induced release of PTHrP, with a maximal, two-fold increase (225). In normal breast duct epithelial cells, in contrast, stimulation of the CaR leads to decreased PTHrP release (278). Highly speculatively this could suggest that a switch in the function of the CaR takes place during the malignant transformation. Also, stimulation of the human MCF-7 breast cancer cell line with 100 μ M of the type II agonist NPS R-568 or calcium decreased the estrogen receptor α protein, whereas both CaR ligands upregulated transcription of the receptor. The estrogen receptor a is thought to play an important role in the function and growth of breast cancer (126), as estrogen receptor positive breast tumors display a greater tendency to grow in the bone and this finding could enhance the progression of the disease. Of note in this study, the effect of the calcium ionophore A23187 was opposite to that of calcium, indicating that calcium did not act in a nonspecific manner as a result of its influx through calcium channels. Although the doses of NPS R-568 were very high in this study, these data indicate that the CaR, in addition to its effect on PTHrP release, modulates the important estrogen receptor a in breast cancer. Breast cancer and prostate cancer are known to metastasize to bone, and the metastases are often mixed osteolytic and osteosclerotic lesions (90). Tumors that stimulate new bone formation, through a complex stimulatory effect by growth factors and other agents on the osteoblast, are thought to result in osteosclerotic metastases. This new bone formation may indeed cause hypocalcemia although this is not common. Breast cancer cells metastatic to bone will be exposed to a locally high level of Ca 2+ o , which, through stimulation of the CaR, can, in turn, induce two promalignant features, PTHrP release and estrogen receptor a regulation. Thus, the stimulation of PTHrP release by the CaR constitutes a vicious cycle in bone metastases of the breast and prostate cancer cell lines. Growth factors in addition to PTHrP that can induce such vicious cycles are endothelin-1, vascular endothelial growth factor, IL-8, and IL-11 (49). Bone itself stimulates the tumor by releasing insulin-like growth factors and TGF-b as described above. Secreted factors transmit the interactions between tumor and bone and provide novel therapeutic targets to interrupt the vicious cycle of bone metastases. Investigation of the effects of the CaR on the effects and/or production of endothelin-1, vascular endothelial growth factor, IL-8, and IL-11 might further clarify the role of the CaR in these malignant cycles.

PTHrP and CaR in humoral hypercalcemia of malignancy

The Rice H-500 rat Leydig cells are a primary cell culture of a rat Leydig cell tumor and represent a xenotransplantable model of HHM used in four of the studies in this thesis. The H-500 model is not the only transplantable model of humoral hypercalcemia of malignancy; additional such tumors include: the PTHrP producing cancer cell lines LC-6 established from human large-cell lung cancer (67), the human lung SCC cell line (RWGT2) from explanted metastatic tumor tissue (91), the C-26 cells that were originally isolated from a chemically induced mouse colon adenocarcinoma (61), and the Walker carcinosarcoma (WCS) 256, a rat mammary carcinoma cell line also producing PTHrP (106). The H-500 cell model was chosen as a model of HHM to investigate possible roles of the CaR, rather than any of the other models mentioned above, since Rizzoli et al. found that challenging Rice H-500 rat Leydig tumor cells with Ca 2+ o stimulated the release of a PTH-like bioactivity (215). Two additional studies published at almost the same time showed that the CaR was present on these cells and that type 1 agonists of the CaR such as calcium stimulated PTHrP release (30, 224). In the study by Sanders et al. they found that the CaR was expressed at both the protein and mRNA levels and that stimulation of the H-500 cells with Ca 2+ o resulted in a dose-dependent increase in PTHrP release by more than three-fold. The type I agonist neomycin had the same effect as Ca 2+ o on PTHrP release (224). Buchs et al. also showed that the presence of the ionophore ionomycin did not block the calcium-induced PTHrP release in H-500 cells (30). These results point toward the CaR functioning as the mediator of the calcium-induced PTHrP release. We later demonstrated that the response to Ca 2+ o of H-500 cells infected with a naturally occurring dominant-negative Arg185Gln CaR (11) construct in a rAAV vector (292) was shifted to the right and downward in comparison to the response of cells infected with b-galactosidase, a protein of the same size as the CaR ( Figure 4 ) (259), thereby showing that the Ca 2+ o -induced PTHrP release is, at least in part, through the CaR. Therefore, stimulation of the CaR by Ca 2+ o in HHM leads to an increase in PTHrP release. The released PTHrP then increases systemic Ca 2+ o . This process constitutes a malignant, vicious cycle, and pharmacological blockade of the CaR, its downstream signaling apparatus, or the secreted PTHrP may represent a fruitful therapeutic approach. The post-receptor mechanism by which the CaR stimulates PTHrP release seems to be through a de novo synthesized transcript, as the pan polymerase inhibitor, actinomycin D, inhibits the effect of Ca 2+ o on the expression of PTHrP mRNA as well as on protein release. It was shown subsequently that the effect of the CaR on PTHrP release seems to be through PKC, MAPK, and PI3K, as the effects of 7.5 mM Ca 2+ on PTHrP release were blocked by inhibiting PKC and the three MAPKs-MEK1/2 (upstream of ERK1/2), p38 MAPK, and c-Jun N-terminal kinase (JNK)-and by inhibiting protein kinase B (AKT) (255). The stimulation of the MAPKs by Ca 2+ o was independent of PKC, since the phosphorylation of ERK1/2, p38 MAPK, and stress-activated protein kinase ERK kinase 1 (SEK1) (the kinase upstream of JNK) by Ca 2+ o could not be blocked by the pan PKC inhibitor GF-109203X, despite the ability of the PKC-activator, phorbol 12-myristate 13-acetate, to phosphorylate ERK1/2 ( Figure 5 ). This may be explained by the finding that H-500 cells loaded with the calcium-sensitive dye, fura 2, did not exhibit Ca 2+ i spikes in response to high Ca 2+ o , which suggests either that the CaR is expressed at too low a level on the cell membrane to produce Ca 2+ i spikes (30) or that the CaR mediates its effects on MAPK in a PLC-independent pathway in H-500 cells. The same phenomenon, that CaR activation does not lead to Ca 2+ i spikes, was also reported in intact gastric mucosal cells that expressed a functionally active CaR (33). Filamin-A and beta-arrestin binds to the CaR (107, 195) and have been shown to interact with MAPKs (161), providing a possible explanations for the PLC-independent activation of MAPKs downstream of the CaR. In support of this hypothesis other receptors also coupled to the G α q like angiotensin II receptor type I have been shown to activate MAPKs independent of G α q through the scaffolding protein β -arrestin (5). In H-500 cells, the phosphorylation of the ERK1/2 was delayed, and ERK1/2 activation reached a maximum at 30 to 60 minutes and was sustained up to 2 hours before declining (Figure 5). In dispersed parathyroid and HEK-CaR cells, the phosphorylation of ERK1/2 by high Ca 2+ o was maximal at 10 to 30 minutes (136). This difference in the kinetics of MAPK activation may be explained by the fact that prolonged activation of ERK1/2 by EGF, in contrast to brief activation by PDGF, leads to translocation of phosphorylated ERK1/2 to the nucleus and to proliferation in Swiss 3T3 fibroblasts (172). As described later in this thesis, stimulating the CaR in H-500 cells indeed enhanced proliferation and protected the cells from apoptosis through the p38 MAPK and PI3K pathways but, surprisingly, not through the ERK1/2 pathway (255). The overall message from the studies investigating the possible role of the CaR as a mediator of calcium-induced PTHrP release by prostate, breast, and testis cancer is that the CaR induces PTHrP release from the tumor cells and that the effect of the CaR appears to be most pronounced in the H-500 cells, as the fold increase in PTHrP was the highest, at least in the models tested. One could speculate that this is due to primary culture versus cell line as the H-500 cells represent the former and the PC-3 and MDA-MB-231 the latter. In the nonmalignant murine Leydig TM3 cell line, indication of a cell membrane bound calcium sensing mechanism has been found (4), but this study did not test whether the CaR was present. Furthermore, the presence of the CaR in human testis or testis cancer has to my knowledge not been investigated. This could be highly interesting as the CaR might have the same function in the Leydig cells as in the GH producing pituitary adenoma, that is modulating the regulatory hormone effect on the Leydig cell; this hypothesis is of course highly speculative.

PTTG in testis Leydig cancer

The oncogenic roles of PTTG, along with its abundance in the testis, as well as the upregulation of PTHrP release and DNA synthesis by the CaR in our HHM model, led us to hypothesize that the CaR could play a role in the regulation of PTTG in H-500 cells. Our investigation using northern blot analysis revealed that PTTG was expressed in H-500 cells and that stimulation by Ca 2+ o induced an upregulation in the level of the PTTG (261). Real-time PCR with specific primers for PTTG showed that Ca 2+ o upregulated PTTG messenger RNA in a dose-dependent manner and reached a peak at 18 hours. H-500 cells infected with the dominant-negative CaR, Arg185Gln, exhibited a blunted response to Ca 2+ o as compared with H-500 cells infected with β -galactosidase. This suggests that the calcium-induced upregulation of PTTG was mediated through the CaR. To assess the specificity of the CaR for inducing PTTG mRNA, we used ADP, the ligand for another group of 7TMs, the purinergic receptors, some of which, like the CaR, are coupled to Gaq11. In a previous study, we showed that treatment of H-500 cells with ADP produced a rapid and transient increase in Ca 2+ i (259), thus confirming that H-500 cells express a functional ADP-sensitive receptor linked to the PI-PLC system that elevates Ca 2+ i in an agonist-dependent manner. Because ADP is degraded by ectonucleotidases within 15 to 30 minutes, we used instead ADPbS, a non-degradable agonist of the ADP receptor. We next examined whether the effect of activation of the purinergic receptor, like activation of the CaR, would increase PTTG expression. Whereas, high Ca 2+ o significantly increased the expression of PTTG mRNA (6.06 ± 2.18-fold increase), the cells treated with ADPbS (10 -6 M) did not show any change in their expression of PTTG mRNA (0.84 ± 0.06). Upregulation of PTTG has been shown to induce proliferation and angiogenesis (121, 191). We therefore investigated whether the high-Ca 2+ o -induced CaR-mediated action on PTTG expression also regulated the expression of the potent angiogenic factor, vascular endothelial growth factor (VEGF). We found, using real-time PCR, that high Ca 2+ o indeed upregulated the mRNA of the VEGF gene in H-500 cells stimulated for 18 hours as compared with low calcium. Prior to this study, the regulation of PTTG expression was known to be 1) inhibited by cyclosporine A and hydrocortisone in T lymphocytes and 2) upregulated by estrogen and bFGF and downregulated by a peroxisome proliferator-activated receptor γ agonist rosiglitazone in pituitary adenomas (98, 99, 247). Our study showed for the first time that a G α q11 -coupled 7TM, the CaR, upregulates PTTG expression in testis-derived rat Leydig cancer cells.

PTTG and astrocytoma cells

The finding the PTTG was upregulated by the CaR in the H-500 cells led us to hypothesize that CaR may be upregulating PTTG in most malignant, CaR-expressing cells. In order to test this hypothesis, we used the U-87 cells. The U-87 astrocytoma cell line was chosen to investigate the effects of CaR stimulation on the PTTG oncogene because of its robust response to calcium in PTHrP release and ion channel regulation (40, 44).

PTTG expression has been positively correlated with several malignancies, including breast, prostate, and thyroid cancer (18, 92). In 2003, PTTG expression was reported to be present in developing neurons and was potentially implicated as an important cell-cycle regulator in human neurogenesis (19). These reports led us to primarily investigate whether PTTG was expressed in astrocytomas, the most common primary tumor of the brain. Astrocytomas, particularly high-grade astrocytoma or glioma, carry a very poor prognosis. Using PCR, we showed that PTTG and its binding protein were expressed in human primary astrocytes, astrocytoma, and glioblastomas, suggesting that astrocyte-derived cells possess the complete functional apparatus for PTTG (263). A comparison of the expression levels showed a trend towards greater expression of PTTG in the malignant samples than in the astrocytes. Similar results was obtained by a German group a year later in a larger sample size. They used mRNA from normal brain as controls and found very low expression of the PTTG mRNA (36). PTTG has been implicated in angiogenesis and proliferation; thus secondarily we studied the proliferative role of PTTG in astrocytomas by designing a small interfering RNA against PTTG mRNA and investigating its effect on bromodeoxyuridine incorporation, a measure of DNA synthesis and proliferation. Silencing PTTG mRNA via the small interfering RNA methodology inhibited serum-induced incorporation of bromodeoxyuridine by approximately 50% in the U-87 astrocytoma cell line as compared with the incorporation by U-87 cells treated with a scrambled oligonucleotide ( Figure 8 ). The incorporation of bromodeoxyuridine was used to measure DNA synthesis and not [ 3 H]thymidine incorporation as used in paper III, because the bromodeoxyuridine incorporation assay requires fewer cells and hence less small interfering mRNA. As in paper III it would have added value to the paper if I had counted the cells as mentioned earlier. Furthermore, making a stable U-87 cell line with a PTTG loss-of-function mutation and implanting them back in to an animal would also have strengthened our hypothesis that PTTG is important in the growth of the tumor. In addition, the use of astrocyte cells from PTTG knock out animals (285) would also be a possible future path to investigating the role of the PTTG in astrocyte cells. Keeping this in mind, we still concluded that PTTG plays a role in regulating the growth of astrocytoma cells. To test our primary hypothesis that the CaR could regulate PTTG expression and to characterize the regulation of PTTG in astrocytomas, we stimulated the U-87 cells with agents known to be important in either the regulation of the astrocytoma cell cycle or of PTTG in other tissues using doses known to be effective based on the literature as well as data from our laboratory and pilot studies (when no known dose was considered effective). In these cells, in contrast to the H-500 cells, Ca 2+ o did not upregulate PTTG expression. Neither estradiol nor ciglitazone, a peroxisome proliferator-activated receptor γ agonist, both of which regulate PTTG expression in pituitary adenoma cells, regulated PTTG mRNA expression. The primary astrocytes as well as in the human U-87 cells and other astrocytoma cell lines have previously been shown to express this receptor and to be responsive to peroxisome proliferator-activated receptor g agonist (41, 171). Three different peroxisome proliferator-activated receptor γ agonists in suprapharmacologic doses were used in the pilot studies, and none affected PTTG mRNA expression. Estrogen has been found to increase the life of nude rats implanted with U-87 cells, suggesting the existence of an estrogen receptor in the tumor (196). Stimulating the U-87 cells with the two EGFR ligands, EGF and TGF α , upregulated PTTG at both the mRNA and protein levels. EGF and TGF α are known promalignant growth factors in astrocytomas. The effects of the two ligands were through the EGFR, as the EGFR inhibitor, AG1478, abolished the effect of EGF and TGF α on the PTTG transcript. Another promalignant growth factor, hepatocyte growth factor, upregulated PTTG in the U-87 cells. The fact that EGF, TGF α , and hepatocyte growth factor all induced PTTG upregulation but that recognized regulators of astrocytoma growth and PTTG expression, including calcium, had no such effect, points to the cell-selective regulation of the PTTG oncogene. Following the same line of argument, the lack of effect of calcium on PTTG regulation could also be explained by the fact that stimulation of the CaR in U-87 cells did not change their proliferative rate when they were challenged with calcium (44). This observation led us to speculate that PTTG is only regulated by compounds that also affect the cell cycle. This hypothesis is supported by data showing the IGF, a proliferatory cytokine in U-87 cells and U-87 tumor cells in vivo (234), also upregulated PTTG in U-87 cells (36).

Downstream signaling apparatus, deciding the fate of the CaR stimuli?

The CaR, like all other cell surface receptors, communicates messages from the surface of the cell to inside the cell. Signals modulating the CaR's activity are generated mainly through changes in systemic or local calcium homeostasis, and changes in Ca 2+ o will, through the CaR, generate activity in intracellular signaling pathways downstream of the CaR. In our model of HHM, I found that stimulating H-500 cells with high Ca2+ activated several major intracellular signaling pathways: the PKC, AKT/PI3K, and MAPK pathways ( Figure 9 ). These three signaling pathways may be seen as information highways inside the cell, although the full impact of the pathways remains to be understood. Kifor et al. first recognized that the CaR activated the MAPK pathway in parathyroid gland cells and HEK-CaR (136). Subsequently, MacLeod et al. described the requirement for MAPK in CaR-mediated PTHrP release in HEK-CaR cells (156). In accordance with these results, I showed that the ERK1/2, p38 MAPK, and SEK1/JNK pathways were required for CaR-mediated PTHrP release in H-500 cells, which naturally express the CaR although at a lower level than the parathyroid cells. Our results on the involvement of the ERK1/2 pathway in CaR signaling is particularly interesting because of the involvement of the ERK1/2 pathway in CaR-mediated PTHrP release but not cell proliferation, whereas p38 MAPK and AKT/PI3K were involved in both CaR-mediated proliferation and PTHrP release. Another intriguing observation regarding the CaR signaling pathways is that PKC seems to work in parallel to MAPKs, but the latter were not distal to PKC, since inhibition of PKC did not change calcium-stimulated ERK1/2, p38 MAPK, or SEK1 activation by high calcium. A valid concern is that the western blotting showing the activation of the MAPK and AKT pathways was carried out after challenging the cells with calcium. This, of course, alone does not prove that the CaR mediated these calcium-induced effects. It would have strengthened the argument if the western blots were performed by challenging the cells with a type II CaR agonist, or if the cells had been infected with DN CaR as in the studies on PTHrP release. But together with the data showing that the CaR-induced PTHrP release can be inhibited by selective inhibitors of the signaling pathways, these data in the conjunction with the literature mentioned above are highly suggestive of that the CaR is utilizing these pathways. Therefore, it is valid to state that the CaR signals through various kinase pathways that lead to different outcomes in our model of HHM. Such differences may be useful in a clinical setting, for example, if one cellular function regulated through a specific pathway is desired whereas another function mediated through a distinct CaR pathway is unwanted. Modulation of the latter could be a pharmaceutical target. One must also keep in mind that the outcome of a signal through a specific intracellular pathway activated by the CaR may depend on the type of cell as well as on whether it is normal or neoplastic. In future studies, the in vitro characterization of patient tumor cells may provide information that will allow the oncologist to design the most successful treatment regime.

The H-500 cells seems to be less dependent of embedded growth factors in the bone to sustain a high level of PTHrP release and unlimited growth as the cells kept their cancer phenotype for at least 10 cell passages in vitro. Therefore, the primary event in this model of HHM could be speculated to be the malignant transformation of the Leydig cells enabling the cells to grow unrestrained and produce PTHrP, and the factors released simultaneously with bone degradation and calcium release may play a lesser secondary role that seems to augment the pace of the development of the disease. PTHrP production seems to be a key factor for the H-500 cells, as H-500 tumor cells treated with antisense to PTHrP showed less growth and did not induce hypercalcemia in vivo (205). The use of a PTHrP neutralizing antibody or a PTHrP promoter approach in theory seems to be a very promising approach to treat rats with implanted H-500 cells. To my knowledge none of these studies have been performed.

The observation that the CaR upregulates the iNOS enzyme opens the door to an entirely new understanding of the CaR in normal physiology both in tissues with high levels of expression of the receptor, such as parathyroid gland and kidney, and in cell types with low expression of the CaR, such as cardiomyocytes and fibroblasts. Two reports have suggested that the regulation of the myogenic tone maybe dependent of a CaR expressed on the endothelial cells (181, 289). And this effect on myogenic tone seems to be through the NO system, as indicated by one study that found that downregulation of the CaR by siRNA decreased NO production in response to calcium in human endothelial cells (316), strongly suggesting that the CaR mediated these effects on NO. These studies support our initial hypothesis that CaR utilizes the NOS/NO system as a downstream signaling molecule. It would be interesting to study whether the effects of the CaR on iNOS and VEGF in H-500 cells will also be present in those tissues with particularly high or low levels of expression of the CaR. Furthermore, studying the role of NOS/NO downstream of the CaR in the parathyroid gland is an important to clarify whether this signaling pathway is important as indicated by the studies by Gardner et al. (78, 79) .

In the literature one report indicates the existence of the CaR by immunohistochemistry in testicular cells (169). It could be of interest to study whether the CaR is functionally expressed in the normal testicular cells and in particular the Leydig cells. The CaR may be sensing calcium in local fashion as described by Hofer et al. (110), and the functions that I have described in this thesis for the CaR in cell cycle and NO regulation could be investigated as they may be of importance in normal Leydig cells.

Regulation of PTTG by the CaR in Cancer

Since the cloning of PTTG in 1997 from pituitary tumor cells, our understanding of its structure, function, and clinical importance has made impressive progress (258). A great body of data has enabled us to understand the molecular and cell biological regulation of PTTG and its biological functions. Increasing evidence regarding the roles of PTTG in various cancers, including pituitary adenomas, is making PTTG a candidate marker of malignancy and a chemotherapeutic target (100, 222, 313). I found that the CaR upregulated the PTTG mRNA in the testicular H-500 cells, whereas this was not the case in astrocytoma U-87 cells; Moreover, the CaR stimulates growth in H-500 cells but not in U-87 cells. In U-87 cells I found TGF and HGF pathways to upregulate the PTTG, and others have found that IGF-1 is also contributing to the upregulation (36). These data indicate that the CaR uses PTTG in a proproliferative response. Hence further studies are needed to analyze whether the CaR utilizes the PTTG pathway to promote growth in the H-500 cells and other cancer cells with a functional CaR. If this hypothesis is valid, one could speculate that the PTTG mRNA or protein may be used as a marker of therapeutic effect of chemotherapy or even a drug target, especially when the CaR is stimulating growth.

CaR in cancer: Theoretical and clinical implications

In this thesis I propose that the calcium released during malignant bone degradation is a possible second candidate, in addition to TGF β , that may stimulate the tumor to increase its production of hypercalcemia-inducing agents-in our studies, PTHrP. Our findings in H-500 cells are summarized schematically in Figure 9. Stimulation of the CaR by Ca 2+ o leads to enhanced PTHrP release, through the PKC, MAPKs, and PI3K pathways. The oncogene PTTG, known to be necessary for normal testis development, is also upregulated when the CaR is activated. This may be merely an association that indicates that the rate of growth of the tumor cells is upregulated but may also be an important signaling pathway in the cell downstream of the CaR but proximal to proliferation. Thus PTTG may be a drug target in the vicious cycle of our HHM model. To validate this hypothesis, I could have used the siRNA approach as used in paper IV. Furthermore in paper II, I found that Ca 2+ o upregulates VEGF, perhaps the most potent growth factor in the stimulation of new vessel formation. I speculate that the VEGF upregulation may be mediated through the CaR. This has to my knowledge not been tested. In H-500 cells, stimulating the CaR likewise upregulated the constitutively active NOS enzyme, inducible NOS, but not the mRNA of endothelial or neuronal NOS. The upregulation of iNOS by the CaR seems to increase NO production, as high calcium concentration increased NO levels in the H-500 cells. A key characteristic of the cancer cell is its capacity to grow, and the growth of a tumor is decided primarily by the increase in its rate of proliferation and the reduction in its apoptosis. In the case of proliferation, my data could not confirm that the upregulated NO participated in the CaR-induced DNA synthesis. I found that stimulating the CaR in the H-500 cells leads to increased DNA synthesis through the p38 MAPK and PI3K signaling pathways, but not through the ERK1/2 signaling pathway or NO production. Last, stimulation of the CaR protects H-500 cells from undergoing programmed cell death. Thus, stimulation of the CaR induces several promalignant features in the Leydig tumor, and blocking the CaR seems to be a promising drug target in the treatment of this disease. On the other hand, since the CaR is so ubiquitously expressed, and especially because of its key role in calcium homeostasis, the positive effects of blocking the receptor pharmaceutically may be overshadowed by its adverse effects, even in cells in which the CaR has a documented promalignant role, such as H-500 cells. In PC-3 prostate cancer cells, we found that EGFR mediates some of the effects of the CaR on PTHrP release; thus, the EGFR is another possible membrane-bound target in the treatment of hypercalcemia-inducing cancers in which a functional CaR is expressed. Targeting the EGFR in colon cancer in which its growth was induced by prostaglandin, acting through its respective 7TM, has also had promising experimental results (185). The triple membrane spanning signaling or transactivation by CaR through the EGFR was latter shown also to be present in HEK-CaR cells, fibroblast and H-500 cells (157, 264, 266). The importance of this newly characterized pathway has not been established in parathyroid cells, which would be an important future study as the full story of CaR's signalling apparatus in the parathyroid cells has yet to be elucidated.

Although direct blocking of the CaR in malignant disease may be hazardous, I believe that understanding the mechanism of tumor growth is crucial for success in treatment of malignant diseases. Therefore, knowledge on the role of the CaR in malignant cells from this series of studies may be useful in selecting appropriate and optimal treatments of malignant hypercalcemia. The clinical treatment of malignant hypercalcemia follows two approaches: the first priority is treatment of the underlying malignant disease by a surgeon or an oncologist, and the second is correction of the hypercalcemia by rehydration and bisphosphonates, and sometimes glucocorticoids, in order to decrease the degradation of bone by osteoclasts. The use of bisphosphonates is in accord with my results, as well as with the work of Guise et al., which may be treating not only the hypercalcemia but also the underlying disease because of the decreased stimulation of the CaR or TGF β receptor on the tumor cell surface. On another note, a third effect of the bisphosphonates seems to be a potential direct inhibitory effect on cancer cell growth (49). This, of course, raises the question of whether the bisphosphonates should be introduced in a treatment regimen for patients with malignant tumors proven to be sensitive to bisphosphonates even before hypercalcemia is present. Although this potential effect has not been fully established, as improved survival of patients with bone metastasis treated bisphosphonates is still debatable (28). Most of the models described in this thesis are based on cell lines or primary cultures that are very homogeneous and in which it is possible to modify experimental variables one at a time. In the complex clinical situation with heterogeneous tumors, perhaps only some cells express the CaR, and there is great variety in patients' responses, e.g., their immune responses. Accordingly, the next step is to test my hypothesis in animal studies and, ultimately, in clinical trials.