Calcitonin Stimulates Multiple Stages of Angiogenesis by Directly Acting on Endothelial Cells

Angiogenesis, the de novo sprouting and remodeling of capillaries, plays a central role in a variety of physiologic and pathologic processes (1, 2). Tumor angiogenesis, which can be broken down into several overlapping phases (3), typically correlates with increased tumor growth and metastasis (4). The first phase begins when angiogenic cytokines activate vascular cytokines. The second phase occurs when activated endothelial cells begin to proliferate and migrate towards the mitogenic stimulus where they form a vascular sprout. This phase also requires the release of proteolytic enzymes, which degrade the basement membrane to facilitate the invasion of endothelial cells. In the third phase, the vascular sprout forms a lumen. Finally, periendothelial cells are recruited to the area to provide further support for the new vessel.

Prostate cancer cells secrete a variety of proangiogenic substances, including vascular endothelial growth factor (VEGF), interleukin-8 (IL-8), and extracellular matrix proteinases (5–8). It has been proposed that neovascularization begins in benign prostatic hyperplasia and keeps progressing in a stepwise fashion in proliferative premalignant and malignant stages (9, 10). Although a strong correlation exists between the degree of neuroendocrine differentiation and the metastatic potential of prostate cancer, the results of this study fail to establish a mechanistic link between increased expression of bombesin and neurotensin with increased angiogenesis of the metastatic prostate cancer (10). Our previous studies have shown that calcitonin, a neuropeptide, is secreted by primary prostate epithelial cells, and its secretion from cancer-derived primary prostate cells is several-fold higher than it is from cells of benign prostatic hyperplasia origin (11). High affinity calcitonin receptors have been identified in membrane preparations of primary tumors and of LNCaP and PC-3M cell lines (12, 13). Exogenously added calcitonin stimulates the proliferation of LNCaP cells by stimulating cyclic AMP accumulation and by increasing cytoplasmic Ca²⁺ transients, and it stimulates their invasion by activating protein kinase A–mediated mechanisms (14). In addition, PC-3M cells stably overexpressing calcitonin form highly vascularized tumor xenografts in nude mice, raising the possibility that calcitonin might induce the angiogenic process in malignant prostate endothelium either by directly acting on endothelial cells or by inducing the secretion of proangiogenic cytokines or growth factors.

Considering that the calcitonin receptor–like receptor (CRLR) is expressed by human endothelial cells (3), we examined the possibility that prostate tumor–derived calcitonin might directly activate angiogenesis by activating either the calcitonin receptor or the CRLR in endothelial cells. We studied the effect of calcitonin on human microvessel endothelial-1 (HMEC-1) cells in in vitro models of different phases of neovascularization, such as migration, proliferation, invasion, and lumen morphogenesis. Finally, we examined the effect of PC-3M-derived calcitonin on angiogenesis in nude mice using a dorsal skinfold assay. The results provide evidence for the direct, receptor-mediated effect of calcitonin on HMEC-1 cells and identify an important role for tumor-derived calcitonin in neovascularization.

Human microvessel endothelial cells (HMEC-1), provided by Dr. Fransisco Candal (Center for Disease Control, Atlanta, GA), were maintained in MCDB 131 formula (Life Technologies Invitrogen, San Diego, CA) supplemented with 10% fetal bovine serum (FBS), epidermal growth factor (EGF, 10 ng/mL), hydrocortisone (1 μg/mL), and l-glutamine (10 mmol/L). Aliquots of cells were preserved frozen between passages 1 and 5. For all experiments, HMEC-1 cells were used at passages 3 or below and harvested from a confluent dish.

Calcitonin is known to interact with calcitonin receptor as well as CRLR (15). In addition, accessory proteins such as receptor activity–modifying proteins (RAMP) regulate the turnover and specificity of these receptors (16, 17). Although evidence for the presence of CRLR in human endothelial cells exists (3), the presence of calcitonin receptor in these cells has not been reported. We examined the presence of transcripts for these receptors as well as RAMPs in HMEC-1 cells.

Total RNA (2 μg) prepared from HMEC-1 cells was reverse transcribed into cDNA with 1 μg oligo dT primers in a 20-μL reaction volume using Invitrogen Superscript cDNA synthesis kit (Life Technologies Invitrogen). The primers were designed to be specific for calcitonin receptor, CRLR, RAMP1, RAMP2, and RAMP3, and not to cross-hybridize with other known sequences as previously described (3, 18). The forward and reverse primer sequences were calcitonin receptor, 5′-AAGCTTTTGCTTCTATTGAGCTGTGCC-3′ and 5′-GAATTCCTCAGTGATCACGATACTGTG-3′; CRLR, 5′-GTAATGTTAACACCCACGAGAAAG-3′ and 5′-ATCCCCAGCCAAGAAAATAATAC-3′; RAMP1, 5′-GCAGCATCCTCTACCCCTTCATC-3′ and 5′-TCGGCTACTCTGGACTCCTGTG-3′; RAMP2, 5′-CATCCCCTTCCTCATCACTCTTG-3′ and 5′-GGCTTCCATTCCCAGCAATC-3′; and RAMP3, 5′-GCCAGTGGAGGAAAATGTGATAAG-3′ and 5′-AGGAACCAGAGATGGGGTCAAG-3′.

PCR was done in 50-μL volume with 20 mmol/L Tris-HCl (pH 7.4, 25°C), 50 mmol/L KCl, 1.5 mmol/L MgCl₂, 0.1% Triton X-100, 200 μmol/L each of four deoxynucleotide triphosphates, 1 μmol/L of each primer, cDNA derived from the eqivalent of 200 ng total RNA, and 2.5 units of Taq Polymerse (Life Technologies Invitrogen). Samples were subjected to 35 cycles in the iCycler (Bio-Rad, Hercules, CA). Cycle variables were as follows: the initial denaturation step at 95°C for 5 minutes; the repeat cycles consisted of annealing at 55°C, 50°C, and 55°C for calcitonin receptor, CRLR, for RAMPs, respectively, for 40 seconds followed by extension at 72°C for 1 minute and denaturation at 94°C for 30 seconds; the last extension time was lengthened to 10 minutes. Amplified cDNAs were fractionated on 1.2% agarose gels in 89 mmol/L Tris, 89 mmol/L borate, 2.5 mmol/L EDTA and pH of 8.0. After staining with ethidium bromide, gels were photographed under UV light.

Membrane preparation. HMEC-1 cells in 100-mm dishes were detached with 0.05% EDTA in PBS, harvested by centrifugation at 200 × g for 10 minutes, suspended in cytoplasmic lysis buffer [10 mmol/L HEPES, 10 mmol/L NaCl, 1 mmol/L KCl, 5 mmol/L NaHCO₃, 1 mmol/L CaCl₂, 0.5 mmol/L MgCl₂, 1 mmol/L phenylmethylsulfonyl fluoride (PMSF), 100 units/mL aprotinin, 5 mmol/L EDTA], and homogenized by 50 strokes in dounce homogenizer. Membrane fraction was obtained by centrifuging homogenates at 15,000 × g for 7 minutes. Protein concentration was determined by Bio-Rad assay, and the membranes were aliquoted and stored frozen at −70°C.

Radioligand binding assay. The ¹²⁵I-mono-iodinated sCT of high specific activity was prepared as previously described (12). Membranes (1 μg protein per tube) were incubated with increasing concentrations of ¹²⁵I-sCT in the assay buffer (25 mmol/L Tris, 10 mmol/L MgCl₂, 1 mmol/L EGTA, 0.25 mmol/L PMSF, 0.85 mg/100 mL aprotinin, 0.85 mg/100 mL leupeptin) for 1 hour at 4°C. The total assay volume was kept at 50 μL. The radioactivity of the membrane fraction was determined after extensive washings. The binding data was analyzed by nonlinear regression and scatchard analysis (Prism, GraphPad, San Diego, CA).

Exponentially growing HMEC-1 cells were plated at a density of 2 × 10⁵ cells per well in a 6-well plate for 24 hours. The growth culture medium (MCDB 131 supplemented with 10% FBS, EGF, 10 ng/mL, hydrocortisone, 1 μg/mL, and l-glutamine, 10 mmol/L) was then replaced with serum-free MCDB 131 medium, and the cells were transfected with one of the following small interfering RNA (siRNA) duplexes against calcitonin receptor: 5′-GCACUGCCCUGACUACUUCUU-3′ and 5′-P-GAAGUAGUCAGGGCAGUGCUU-3′, 5′-GAGAAGGUCUGUACUGCAAUU-3′ and 5′-P-uUGCAGUACAGACCUUCUCUU-3′, and 5′-CAAAUGCUAUGACCGGAUUUU-3′ and 5′-P-AAUCCGGUCAUAGCAUUUGUU-3′.

After overnight incubation in serum-free medium, the transfectants received the growth medium containing 10% FBS for additional 24 hours. The cells were then harvested to analyze calcitonin receptor mRNA abundance by real-time PCR.

The PCR was done in a 50-μL reaction using 5 μL of the diluted first-strand cDNA template, optimized amounts of primers, and 25 μL of the 2× IQ SYBR Green supermix (Bio-Rad). The reactions were cycled at 50°C for 2 minutes, 94°C for 5 minutes, and 40 cycles of 94°C for 30 seconds, 57°C for 1 minute, and 68°C for 2 minutes. Template concentrations were adjusted to bring PCR products of all samples in the linear range of amplification. The level of transcripts for each gene was normalized to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in the same sample. All PCR reactions were done in duplicates and at three doses of the cDNA template. C_t values were determined, and the fold change in mRNA abundance was calculated as described previously (19).

The sequences of specific primer pairs are as follows: calcitonin receptor, 5′-ATCGAGGTCCAAACCACCGTGAAGC-3′ and 5′-CAGTATTGACAAGGAGCACCCTG-3′ and GAPDH, 5′-ACGCCGCATCTTCTTGTGC-3′ and 5′-ACAGCCGCATCTTCTTGTGC-3′.

The assay examined the formation of vascular networks by HMEC-1 cells (20). The cells were plated in 8-well chambers at a density of 4 × 10⁴ cells per well and cultured in endothelial growth medium for 4 hours. The cells were then treated with conditioned media (CM) of PC-3M cells or various concentrations of calcitonin (0-100 nmol/L) in serum-free medium for 12 hours. The cells were then washed and stained for H&E with Diff-Quick staining kit (Dade Behring Diagnostics, Aguada, Puerto Rico). The network formation was observed under an inverted phase contrast microscope (Nikon, Tokyo, Japan), and the images were captured with a video graphic system (Retiga 1300 Digital Output Camera). The number of network projections in ten 100× fields were counted for four independent experiments in each group, and the data were obtained using the following formula: angiogenesis = a × b where a = the number of branch points per 100× field and b = the total number branches per point.

Transwell migration assay. Migration assays were done in transwell tissue culture plates (6.5 mm and 8-μm pore size, Becton Dickinson and Co., Franklin Lakes, NJ). The bottom of the transwell chamber was coated with 10 mg/mL of collagen I (Sigma, St. Louis, MO), and uncoated sites were blocked with 10% bovine serum albumin (BSA). Approximately 1 × 10⁵ cells per 100 μL were added into each well and allowed to migrate for 6 hours. The cells were then fixed with methanol and stained with crystal violet. The migrated cells in five random 100× fields were counted. Each experiment was done in triplicates, and the experiment was repeated twice. The results are expressed as mean number of migrated cells per field ± SE.

Endothelial cell proliferation assay. The growth/proliferation of HMEC-1 cells was measured using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay kit (American Type Culture Collection, Manassas, VA). Exponentially growing HMEC-1 cells were plated at the density of 8 × 10³ cells per well of 96-well plates and grown in complete MCDB131 medium for 24 hours. The complete medium was then replaced with 100 μL of basal MCDB131 medium (containing 0.1% BSA, 4 mmol/L l-glutamine, 100 IU/mL penicillin G, and 100 μg/mL streptomycin). MTT solution was added after 24 hours and incubated at 37°C for 4 hours, the reaction was terminated with the addition of 100-μL solubilization/stop solution, and incubation was continued at 37°C to completely dissolve the formazan product. The plates were then analyzed for absorbance on an ELISA plate reader (Bio-Rad) at 595 nm.

Cell invasion assay. These experiments were conducted in 24-well, two-compartment, Matrigel invasion chambers (Becton Dickinson, Bedford, MA) as described before (14). Exponentially growing HMEC-1 cells were serum starved for 24 hours with basal MCDB131 medium, harvested, and seeded at a density of 25 × 10³ cells per well in the insert of Matrigel invasion chamber. The lower chamber received the chemoattractant medium, which consisted of 90% serum-free MCDB131 medium and 10% conditioned medium from the cultures of PC-3M cells expressing constitutively active Gsα protein (21). The cells also received calcitonin or other agents, which were added to the insert. The incubations were carried out for 24 hours, after which the Matrigel (along with noninvading cells) was scraped off the inside bottom of the insert using cotton swabs, and outer side of the insert was fixed and stained using Diff-Quick staining kit. The number of cells migrated on the outer bottom side of the insert were counted under the microscope in at least six randomly selected fields (magnification, 100×). The final results were expressed as mean number of invaded cells per 100× field ± SE.

Growth correction. To rule out the possibility that cells migrated during early part of the 24-hour incubation period could proliferate during the remaining treatment period, we determined the growth rate of HMEC-1 cells under identical culture conditions as described before (14). This correction was then applied to the final results.

In vitro tube morphogenesis on Matrigel. The vessel morphogenesis assay was done as previously described (3). HMEC-1 cells were adjusted to a density of 3 × 10⁴ cells in 200 μL in MCDB 131 (Life Technologies Invitrogen) medium supplemented with 0.5% FBS and either calcitonin or VEGF and added to the wells of eight-well chamber slides precoated with growth factor–depleted Matrigel (7 μg/mL; Becton Dickinson, Bedford, MA). The slides were incubated at 37°C for 18 to 20 hours. The medium was then removed gently without disturbing newly formed tubules. The images were captured at a 40× magnification on Nikon Optiphot II microscope. Total tube length of each well was measured using IP Lab Image Analysis Program (Scanalytics, Inc., Arlington, VA).

The effect of PC-3M-secreted calcitonin on in vivo angiogenesis was examined using dorsal skinfold assay (22–25). Calcitonin-secreting activity of PC-3M cells was modulated by stably transfecting them with either constitutively active vector expressing calcitonin or hammerhead anti-calcitonin ribozymes. Each modified cell line had vector controls.

Overexpression of calcitonin. The full-length human calcitonin cDNA cloned from primary prostate tumor specimens was inserted in to pcDNA3.1 plasmid (Research Genetics, Carlsbad, CA) in the sense direction (26). The orientation of the insert was confirmed by sequencing at the institutional DNA sequencing facility. CT-pcDNA3.1 construct was then transfected into PC-3M cells and G418-resistant colonies were selected as described before (21).

Silencing of calcitonin expression. Two strands of oligonucleotide templates encoding each ribozyme against calcitonin cDNA having a BamHI and HindIII site at the 3′ and 5′ end, respectively, were synthesized at the Genemed Biosynthesis, Inc. (San Francisco, CA). Sense R4, 5′-GAAGATCTTC/CAAGGGCAG/TTTCGTCCTCACGCACTCATCAG/ATCTGGCT/CCGCTCGAGCGG-3′; Antisense R4, 5′-CTTCTAGAAG/GTTCCCGTCA/AAAGCAGGATGCCGTGAGAGTC/TAGACCGA/GGCGAGCUCGCC-3′; Sense R5, 5′-GAAGATCTTC/AGCTTCTAG/TTTCGTCCTCACGCACTCATCAG/ATCTGGCT/CCGCTCGAGCGG-3′; Antisense R5, 5′-CTTCTAGAAG/TCGAAGATC/AAAGCAGGATGCCGTGAGAGTC/TAGACCGA/GGCGAGCUCGCC-3′.

The oligonucleotides were purified by PAGE, annealed and cloned into BamHI and HindIII sites of ptv5 vector (27). The vector is a pUC-based plasmid carrying a U6 polIII promoter driving the transcription of the ribozyme. The transcription is terminated by a polIII polymerase termination signal (four thymidine residues) to generate ribozymes with very little extraneous sequences. Each ribozyme was cotransfected with plasmid pcDNA3.1-zeocin in to PC-3M cells, and zeocin-resistant colonies were selected after 4 weeks. As a control, vector ptv5 lacking the ribozyme template was transfected to generate PC-3M-v2 cell line.

The clones were then screened for calcitonin mRNA expression and calcitonin secretion by S1 nuclease assay and specific RIA, respectively, as previously described (26).

To validate the results of in vitro studies under in vivo conditions, we exposed skinfolds of nude mice to steady concentrations of calcitonin by placing PC-3M variant cells in transparent diffusion chambers as described before (28, 29). The diffusion chambers were prepared by cementing 0.45-μm membranes on both sides of the rim of the “O” ring (Millipore, Bedford, MA). The chambers were then sterilized by UV radiation for 20 minutes and wetted with sterile PBS. PC-3M-v, PC-3M-CT+, PC-3M-R4, or PC-3M-CT− cells (2 × 10⁶ cells per 150 μL of sterile PBS) were injected individually into each chamber. A dorsal air sac in an anesthetized mouse was prepared by injecting 10 mL of air in the skinfold. The air sac was then opened to implant the chamber underneath the skin, and then incision was sutured. After 10 days, the animals were sacrificed, and carefully skinned around the implanted chambers. The skinfold covering the chambers was carefully removed, stretched on a glass slide and photographed under visible light microscope (20×). New blood vessels on the skinfold were counted and their lengths were determined. Sterile small-animal surgical techniques were followed during the whole procedure, which was approved by the institutional Animal Care and Use Committee.

Data were expressed as mean ± SE and were statistically evaluated by one-way ANOVA followed by Newman-Keuls test. The difference was considered significant if P < 0.05.

Total RNA from HMEC-1 cells was tested for the presence of calcitonin receptor, CRLR, and RAMP-1 to RAMP-3 transcripts. We obtained predicted amplified products of the transcripts of following sizes: calcitonin receptor (666 bp), CRLR amplimers (392 bp), and RAMP-3 (192 bp; Fig. 1A). Transcripts of RAMP-1 or RAMP-2 were not amplified.

Because calcitonin receptor transcript was detected in HMEC-1 cells, we tested its functional presence by the radioligand receptor assay. The HMEC-1 membranes displayed saturable ¹²⁵I-sCT binding sites (Fig. 1B). Scatchard analysis of the binding data determined the K_d value of 3.2 nmol/L, confirming the presence of high-affinity calcitonin binding sites (Fig. 1C).

As revealed by the results of Fig. 2, increasing concentrations of calcitonin significantly stimulated network formation of HMEC-1 cells in a dose-dependent fashion. The effect of calcitonin on capillary-like formations was evident in the number of network formation per field as well as on the thickness of these tube-like structures.

Because angiogenesis involves multiple processes associated with neovascularization including migration of endothelial cells, their proliferation, invasion and lumen morphogenesis on Matrigel, we examined the effects of the CM of calcitonin-treated PC-3M cells in each of these phases.

Calcitonin directly increases the migration of endothelial cells. The results of Fig. 3A show that calcitonin stimulated migration of HMEC-1 cells in a dose-dependent fashion. Calcitonin-induced increases in the migration were 60% and 100% over the controls at 50 and 100 nmol/L calcitonin, respectively, and were statistically significant.

Effect of calcitonin on proliferation of HMEC-1 cells. At 10 nmol/L, calcitonin did not have any effect on proliferation of HMEC-1 cells. At 50 nmol/L, calcitonin induced a marginal increase in cell proliferation. Only at the highest tested concentration of 100 nmol/L, calcitonin induced a moderate but significant 50% increase.

Calcitonin increases the invasion of HMEC-1 cells through Matrigel. The results of Fig. 3C show that at 50 and 100 nmol/L, calcitonin caused significant, 75% and 128% increase in invasion of HMEC-1 cells, respectively. The results are corrected for possible cell proliferation during the experimental period.

When challenged with VEGF, HMEC-1 cells migrate throughout the Matrigel surface and align to form vascular cord–like structures (3). The micrographs presented in Fig. 4 show that HMEC-1 cells do not form cord-like structures in the absence of angiogenic stimulus. However, they formed these structures when treated with various concentrations of VEGF (Fig. 4A-B) and calcitonin (Fig. 4C-D). VEGF stimulated tube formation at 3 and 6 ng/mL concentrations (Fig. 4A-B). Similarly, calcitonin also increased these formations at all tested doses, and 100 nmol/L calcitonin caused over 10-fold increase in tube formations compared with vehicle controls (Fig. 4C). VEGF as well as calcitonin also increased the thickness (microvessel density) of these structures. (Fig. 4B-D).

To confirm that the effect of calcitonin on tube formation is specific and receptor-mediated, we silenced calcitonin receptor expression in HMEC-1 cells using siRNA against calcitonin receptor. Our earlier studies with other cell lines including prostate cell lines have shown that siRNA3 was most potent and reduced calcitonin receptor mRNA expression by >95%. siRNA2 was also effective but less potent than siRNA3. siRNA1 was completely ineffective and was used as a negative control. The results of Fig. 5A have shown that siRNA1 was ineffective in silencing calcitonin receptor gene expression. siRNA2 attenuated calcitonin receptor mRNA abundance by 75% whereas siRNA3 decreased it by 94% (Fig. 5A).

Sham-transfected HMEC-1 cells and those receiving siRNA1 formed almost equal number of tubes with a similar microvessel density (Fig. 5B). Treatment with 50 nmol/L calcitonin increased the number of tube structures by almost 7-fold, which was similar to the responses in previous experiments (see Fig. 4). However, HMEC-1 cells receiving either siRNA2 or siRNA3 displayed no tube formation at all either with/without calcitonin stimulation, suggesting that calcitonin-mediated tube formation by HMEC-1 cells cannot occur in the absence of calcitonin receptor. The pooled data from all experiments is presented in Fig. 5B.

To examine a possibility that calcitonin may stimulate the tube formation by inducing VEGF secretion by PC-3M cells, we tested the effect of calcitonin on tube formation in the presence or absence of anti-VEGF serum (80 ng/mL). The cells were pretreated for 1 hour with anti-VEGF serum before the addition of calcitonin. Anti-VEGF serum partially attenuated but did not abolish the stimulatory effects of calcitonin on in vitro angiogenesis (Fig. 6).

Modulation of calcitonin expression in PC-3M cells. The results depicted in Fig. 7A show the abundance of calcitonin mRNA in PC-3M variant cell lines and their controls. As expected, PC-3M cells expressing CT-pcDNA 3.1 plasmid (CT+) displayed markedly higher abundance of calcitonin mRNA than those expressing empty vector (PC-3M-v). Likewise, the abundance of calcitonin mRNA in cells expressing inactive ribozyme (R4) was similar to PC-3M-v, but CT− cells expressing potent ribozyme (R5) displayed almost complete absence of calcitonin mRNA (Fig. 7A). Consistent with the levels of calcitonin mRNA, PC-3M-CT+ cells secreted 4-fold greater levels of calcitonin compared with PC-3M-v (Fig. 7B). Again, secretion of calcitonin by R4 cells was similar to PC-3M-v, but that of PC-3M-CT− cells was undetectable (Fig. 7B).

Conditioned media of PC-3M cells stimulates the network formation by HMEC-1 cells: its modulation by calcitonin expression. To compare proangiogenic actions, PC-3M cell secretions, we collected the CM from cultures of PC-3M, PC-3M-CT+, PC-3M-R4, and PC-3M-CT− cells after 24 hours and tested first for in vitro angiogenic activity. The CM of PC-3M, PC-3M-v, and PC-3M-R4 cells exhibited similar network-forming activity (Fig. 7C). However, The CM of PC-3M-CT+ cells displayed 7-fold higher angiogenic activity. In contrast, the CM of PC-3M-CT− cells could not stimulate angiogenesis, suggesting the absence of factors stimulating the angiogenesis.

To test the biological significance of calcitonin-induced angiogenesis, we examined the effect of implantation of PC-3M-v, PC-3M-CT+, PC-3M-R4, and PC-3M-CT− cells on neovascularization in the skinfolds of nude mice. The secretions of PC-3M cell variants induced the formation of new vessels in skinfolds. However, the response seemed to vary with calcitonin expression in PC-3M cells. For example, angiogenesis in mice implanted with PC-3M-CT+ cells was observably greater than those receiving either PC-3M-v or PC-3M-R4 cells (Fig. 8). In contrast, mice receiving PC-3M-CT− cells did not display any angiogenic response. The length of new vasculature in skinfolds of mice treated with the PC-3M cell variants was also determined (Fig. 8).

Calcitonin is essential for the maintenance of calcium balance (15, 30). It has also been implicated in the regulation of several functions in reproductive and neuroendocrine tissues (31–35). Overexpression of calcitonin and calcitonin receptor has been reported in several malignancies including breast and prostate cancers (26, 36–38). Our earlier studies have shown that the expression of calcitonin in primary prostate tumors is epithelium specific (26). The number of cells expressing calcitonin and/or calcitonin receptor mRNAs correlate positively with the tumor grade, and cells of aggressive tumors coexpress calcitonin and calcitonin receptor mRNAs (26). We examined the potential role of calcitonin in angiogenic process because our recent results suggest that PC-3M-CT+ cells form larger xenografts with markedly higher vascularization than those formed by PC-3M-v cells. The present results for the first time show that HMEC-1 cells express calcitonin receptor and respond to calcitonin with angiogenesis. Moreover, modulation of calcitonin expression in PC-3M cells alters their ability to stimulate angiogenesis. These results strongly suggest that prostate-derived calcitonin functions as an important angiogenic factor.

Angiogenesis is multiphasic process involving migration of endothelial cells in tumor, their proliferation and invasion in lumen, and the formation of new blood vessels. We examined the effect of calcitonin on each of these processes individually. We have shown that calcitonin stimulated all these processes and led to the formation of vascular tubules on Matrigel. These actions of calcitonin were mediated by calcitonin receptor as silencing of calcitonin receptor expression completely abolished calcitonin response. However, the magnitude of calcitonin response in each of these processes was different. For example, the effect of calcitonin on proliferation of endothelial cells was marginal and significant only at the highest tested concentration of 100 nmol/L. In contrast, the effect of calcitonin on migration and invasion was stronger, increasing by 2- to 3-fold in response to 100 nmol/L calcitonin. However, the most potent action of calcitonin was on the network formation or tube morphogenesis on Matrigel, and calcitonin also increased the thickness of these microvessels. This action was significant even at the lowest tested concentration of 10 nmol/L. Considering that calcitonin could markedly stimulate this process at concentrations of the K_d of calcitonin receptor, it is likely that this effect has an important pathophysiologic significance (12). These actions of calcitonin were similar to VEGF, which is considered as principal growth factor for blood vessel formation (39). Because prostate cancer cells as well as endothelial cells were shown to express VEGF (39–41), we examined a possibility that VEGF may mediate proangiogenic actions of calcitonin by assessing calcitonin-induced tube formation in the presence of anti-serum to VEGF. This partially attenuated but did not abolish calcitonin action, suggesting that calcitonin may stimulate angiogenesis directly as well as indirectly.

To evaluate the biological significance of calcitonin expression by prostate cancer cells, we modulated calcitonin expression in PC-3M cells and examined its effect in an in vivo model of angiogenesis (22, 24, 42). The implantation of PC-3M cells induced neovascularization in mouse skinfold and that overexpression of calcitonin significantly increased this activity. In contrast, silencing of calcitonin expression abolished angiogenic response, reinforcing a possibility that calcitonin is a potent stimulator for angiogenesis and that tumor cell–derived calcitonin may play a key role in prostate cancer progression by stimulating intratumoral vascularization.

The cells of neuroendocrine phenotype are present in androgen-independent prostate carcinoma and are considered as early indicators for androgen independence (43–45). They are shown to secrete paracrine peptides such as bombesin, neurotensin, calcitonin, and others that may support androgen-independent growth of prostate cancer cells (26, 46–48). Recent evidence has identified very high correlation between neuroendocrine expression, microvessel density and VEGF expression in prostate carcinoma (9, 41), suggesting a potential role of neuroendocrine secretions in angiogenesis of prostate cancers. However, bombesin and neurotensin could not stimulate angiogenesis in a model using human umbilical vascular endothelial cells (10). The present results have shown that calcitonin potently stimulated angiogenesis in an in vivo model as well as an in vitro model using HMEC-1 cells. The specificity of calcitonin-induced angiogenesis was shown by the results that the silencing of calcitonin receptor in HMEC-1 cells completely abolished calcitonin-induced angiogenic response. In addition to its direct actions on endothelial cells, calcitonin activates other processes that promote angiogenesis. For example, calcitonin stimulates matrix metalloproteinase (MMP) expression of prostate cancer cells (48–53) and the secretion of activated MMP-2 and MMP-9 (14). Increased secretion of activated MMPs leads to the proteolytic degradation of extracellular matrix and the basement membrane, which enables endothelial cells to invade tumor and form new vessels (54). In addition, the present results that immunoneutralization with anti-VEGF serum significantly reduced calcitonin-induced tube formation raise a possibility that calcitonin may also stimulate the secretion of VEGF by PC-3M cells.

Considering the presence of calcitonin receptor, CRLR, and RAMP-3 in HMEC-1 cells, it is conceivable that other ligands for calcitonin receptor, CTR-RAMP3 or CRLR-RAMP-3 heterodimers may also induce angiogenic response. However, there is no evidence for the expression of amylin in prostate cancer and there seems to be controversy on the effect adrenomedullin on growth of prostate cancer cell lines (55–57). The results of this and other laboratories have reported the overexpression of calcitonin and calcitonin receptor in androgen-independent cases of prostate cancer (26, 52, 58, 59). This, when combined with the present results that PC-3M-derived calcitonin stimulates angiogenesis, raises a strong possibility for the role of prostate-derived calcitonin in tumor growth and angiogenesis. Possible interactions among calcitonin receptor, CRLR, adrenomedullin, and calcitonin in regulation of tumor angiogenesis need further examination.

In summary, the present results show for the first time the presence of calcitonin receptor in HMEC-1 endothelial cells. Moreover, PC-3M-derived calcitonin stimulates angiogenesis in in vitro as well as in vivo models of angiogenesis. Considering the presence of calcitonin in prostate tumor microenvironment, calcitonin may significantly influence tumor growth by regulating intratumoral vascularization.

Grant support: NIH grant CA96534 (G.V. Shah).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.