| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Medicine (R.C., D.G., D.M., B.J., J.A., A.M.), Queen’s University Belfast, Royal Group of Hospitals, Belfast BT12 6BJ, Northern Ireland, United Kingdom; Academic Endocrine Unit (N.L., R.T.), Nuffield Department of Medicine, Oxford Centre for Diabetes, Endocrinology, and Metabolism, Churchill Hospital and Medical Research Council Functional Genomics Unit (D.B.), Department of Human Anatomy and Genetics, University of Oxford, Headington, Oxford OX3 0BP, United Kingdom; and Department of Medicine (D.T.O., L.T.), University of California at San Diego, La Jolla, California 92093
Address all correspondence and requests for reprints to: Ann McGinty, Ph.D., Department of Medicine, Mulhouse Building, RGH Grosvenor Road, Belfast BT12 6BJ, Northern Ireland, United Kingdom. E-mail: a.mcginty{at}qub.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Elevated expression of Cox-2 has been reported in a number of human cancers, including colon and breast, and evidence supports a causal role for this enzyme in tumor development (2). With respect to neuroendocrine tumors (NETs), it has been reported that, although absent from normal adrenal medulla, Cox-2 is expressed in pheochromocytoma tissue (3). In this study, Salmenkivi and co-workers (3) presented evidence that Cox-2 may be involved in promoting pheochromocytoma malignancy. Cox-2 has also been reported to be up-regulated in gastrointestinal and bronchopulmonary carcinoids, particularly at the tumor advancing edge, leading to speculation that it may be involved in carcinoid progression (4).
The aim of the current study was to employ an adrenal chromaffin (neuroendocrine) cell (PC12) model of forced Cox-2 overexpression (PCXII) (5, 6) to identify Cox-2-dependent gene products that might be implicated in the oncogenic/inflammatory actions of this enzyme in the setting of neuroendocrine neoplasia. Differential array analysis identified 59 putative Cox-2-dependent gene products, one of which was the neuroendocrine marker chromogranin A (CGA). CGA is a 48-kDa acidic polypeptide found together with catecholamines in the secretory vesicles of the adrenal medulla (chromaffin granules) and postganglionic sympathetic axons (7). CGA both co-resides and is co-released with catecholamines; indeed, we have reported that CGA is required for vesicular biogenesis, neurotransmitter storage and release, and regulation of blood pressure (8). Evidence has also been presented that CGA is a prohormone that gives rise to a number of biologically active peptides as a result of proteolytic cleavage (7). Real-time PCR, immunoblotting, and immunocytochemical analysis confirmed that CGA expression was Cox-2 responsive. Cox-2-dependent CGA up-regulation was also shown to lead to an increase in the known bioactive indices of this polypeptide, namely chromaffin granule biogenesis and dopamine production. CGA promoter-driven reporter gene expression studies provided evidence that prostaglandin E2 (PGE2)-dependent up-regulation required a proximal cAMP-responsive element (CRE; –71 to –64 bp). This study is the first to demonstrate that Cox-2 up-regulates both CGA expression and bioactivity in a cell culture model of Cox-2-expressing neuroendocrine neoplasia.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For experiments, PCMT and PCXII cell lines were routinely subcultured from confluent T75 flasks at a split ratio of 1:10 to achieve 70% confluency at the time of total RNA isolation and cell lysis/detachment. Cells were incubated in DMEM, 0.5% human serum (HS)/fetal calf serum (FCS), containing 2.5 mM IPTG (restriction media) (5, 6) for 18–24 h before all experiments.
Microarray analysis
For microarray analysis, total RNA was isolated using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA) according to the manufacturer’s instructions. The RNA obtained was then further purified using RNeasy Mini kit and any DNA contamination removed using RNase-free DNase I (QIAGEN, Valencia, CA), also according to the manufacturer’s instructions.
RNA produced was quantified using NanoDrop ND-1000 (Nanodrop Technologies, Wilmington, DE), and RNA quality was checked using RNA 6000 Nano Assay (Nanodrop Technologies) on Agilent bioanalyzer 2100 (Agilent no. 5065-4476; Agilent Technologies, Santa Clara, CA). Commercially available high-density oligonucleotide rat genome arrays RG_U34A (Affymetrix, Santa Clara, CA; catalog no. 900249, analyzing approximately 7000 full-length sequences and approximately 1000 expressed sequence tag clusters) were used. All procedures and hybridizations were performed according to the GeneChip expression technical manual (Affymetrix) as previously reported (9). In brief, 10 µg RNA was converted to double-stranded cDNA with Superscript II (Invitrogen) using T7-(dT)24 primer (Affymetrix). The cDNA was then cleaned on a GeneChip sample clean-up column (Affymetrix). Biotin-labeled cRNA probes were prepared using an Enzo Bioarray high-yield RNA transcript-labeling kit (Affymetrix). The resulting cRNA was purified using a GeneChip sample clean-up. After the cRNA passed a quality control check on a bioanalyzer, cRNA was fragmented to about 200 bp (fragmentation buffer containing 200 mM Tris-acetate at pH 8.2, 500 mM potassium acetate, and 150 mM magnesium acetate at 94 C for 35 min). Fragmented cRNA (15 µg) was used in a 300-µl hybridization cocktail containing spiked controls (Affymetrix), 0.1 mg/ml herring sperm DNA (Promega, Madison, WI), and 0.5 mg/ml acetylated BSA (Invitrogen). A total of 200 µl of the hybridization cocktail was used on each chip and incubated at 45 C for 16 h in the hybridization oven, rotating at 60 rpm. After hybridization, the arrays were processed using a GeneChip Fluidics Station 400 according to recommended protocols (EukGE-WS2v4; Affymetrix) of double-staining and posthybridization washes. Fluorescent images were captured using gene Array Scanner 2500 (Affymetrix). Gene transcript levels were determined from data image files using GeneChip operating software (MAS5.0 and GCOS1.0; Affymetrix). Global scaling was performed to compare genes from chip to chip; each chip was normalized to the same target intensity (TGT = 100). Data quality controls met the Affymetrix quality assessment guidelines. For each replicate and each probe, log fold changes between PCXII and control samples PCMT were calculated using only normalized data classified as present by MAS5.0 software. RNA samples from two independent PCMT/PCXII experimental pairs were analyzed and gave similar results with respect to the profile of putative Cox-2-responsive gene products identified.
Real-time PCR
Total RNA was isolated from cells using an RNeasy mini kit (QIAGEN), quantified and reverse transcribed using a TaqMan reverse transcription kit (Applera, Norwalk, CT). Gene-specific primers to rat CGA (accession number BC087703, 3'-AGG AGC ATG GGA TTC CAC AG and 5'-CAC TGG GAC CTC TCT CAC TGC) and the housekeeping gene GAPDH (accession number AF017079, 3'-ACG GAT TTG GTC GTA TTG GG and 5'-CAG AGT TAA AAG CAG CCC TGG T) were designed using Primer Express software (Applera) and synthesized by Invitrogen Life Technologies. Real-time PCR was performed in triplicate using a SYBR Green MasterMix and an ABI PRISM 7000 Sequence Detection System (Applera). Each well contained the following reaction mix: 0.5 µl cDNA, 5 µl 10x Sensimix dT (Quantace, Watford, UK), 3.7 µl RNase-free water (QIAGEN), 0.4 µl forward primer, and 0.4 µl reverse primer. Universal cycling conditions were used (one cycle at 95 C for 15 min and 40 cycles at 90 C for 15 sec and 60 C for 60 sec). Relative gene expression was calculated using the comparative CT method (10). All values were normalized to the housekeeping gene GAPDH and expressed as a fold change of the control (PCMT) using the formula 2–
CT.
Cell lysate generation
Cell lysates were generated exactly as described previously (5, 6). In some experiments, the Cox-2-selective inhibitor NS-398 (100 nM) was added for 24 h and PGE2 (1 µM) for 6 h before lysis. Conditioned media were retained for analysis of secreted CGA and dopamine levels.
Immunoprecipitation
Conditioned media samples standardized for protein were immunoprecipitated with 0.25 µg anti-CGA (20085; DiaSorin, Stillwater, MN) in whole-cell lysis buffer on a rotating shaker at 4 C for 90 min. After this time, 20 µl protein A/G plus agarose bead (Santa Cruz Biotechnology, Santa Cruz, CA) suspension, which had been equilibrated in whole-cell lysis buffer, was added to each sample exactly according to the manufacturer’s instructions. Samples were then placed on a rotating shaker at 4 C for 60 min. After this time, samples were centrifuged in a bench-top centrifuge at 13,000 rpm for 1 min, and the supernatant was removed by aspiration. Protein A/G plus agarose beads were washed three times with cell lysis buffer. After the final wash, the supernatant was removed by aspiration and samples were resuspended in 4x Laemmli sample buffer (11) before electrophoresis.
Immunoblot analysis
Samples standardized for protein were separated by SDS-PAGE using 5 and 8% acrylamide stacking and resolving gels, respectively, and immunoblotted for Cox-2 (anti-Cox-2, 1:200, catalog no. SC-1745; Santa Cruz Biotechnology), CGA (anti-CGA, 1:200, catalog no. SC-1488; Santa Cruz Biotechnology; or 20085, 1:200; DiaSorin), or chromogranin B expression (anti-CGB, 1:200, catalog no. SC-18235; Santa Cruz Biotechnology). Densitometric analysis was carried out by means of Molecular Analyst software, version 1.3 (Bio-Rad, Hercules, CA) using an imaging densitometer (Bio-Rad).
Immunocytochemistry
Immunocytochemical analysis was carried out exactly as described previously (12). Primary antibodies, namely anti-Cox-2 (SC-1745, 1:100; Santa Cruz Biotechnology), anti-CGA (20085, 1:500; DiaSorin), and anti-NSE (1:50) (13) were diluted in PBS containing 0.5% BSA fraction V, 0.5% Triton X-100, and 0.015 M sodium azide (antibody diluent). Colon adenocarcinoma, brain cortex, and adrenal medulla sections were employed as positive control tissue for the expression of Cox-2, NSE, and CGA, respectively. Negative controls routinely included were a secondary antibody only (antibody diluent) control and the appropriate serum control (1:500). Secondary antibodies used were rabbit antigoat fluorescein isothiocyanate (FITC) (F0233; DakoCytomation, Glostrup, Denmark) for anti-Cox-2 and a biotinylated donkey antirabbit-streptavidin-FITC system (RPN 1232; Amersham Biosciences, Little Chalfont, UK) for anti-CGA and anti-NSE. The slides were viewed by fluorescence microscopy and independently scored on the basis of fluorescence intensity by two observers blinded to their identity using a four-mark scale (– to +++).
Transmission electron microscopy
Growth medium was replaced with restriction medium 18–24 h before detachment of cells using trypsin/versene (1:1, vol/vol). Cells were fixed with 3% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 10 h, washed with buffer, and treated with 1% osmium tetroxide in 0.1 M sodium cacodylate buffer for 1 h. Samples were washed and en bloc mordanted with 0.5% uranyl acetate in acetate buffer at pH 5.0, dehydrated in ethanol, and embedded in epoxy resin. Ultrathin sections (70–90 nm) were counterstained with lead citrate. Measurements were made on micrographs (20 per cell line) printed at a final magnification of x20,000. Two independent observers, blinded to the identity of the micrographs, counted dense-core secretory (chromaffin) granules, and a mean value for each cell line was calculated (14).
Dopamine measurement
Levels of dopamine in PCMT and PCXII cell lysates (intracellular) and conditioned media (extracellular) were measured by means of reverse-phase HPLC, as described previously (15).
CGA promoter-reporter constructs and expression plasmids
For all CGA promoter constructs, promoter fragment positions are numbered relative to the major transcriptional (cap) site as +1. For example, pXP1200 contained 1200 bp of the mouse CGA promoter (5'-flanking region) fused to a luciferase (Luc) reporter in the promoterless luciferase reporter vector pXP1. Construction of mouse CGA promoter/luciferase reporter plasmids pXP1200, pXP77, and pXP61 was as described previously (16, 17, 18). A site-directed mutation of the CGA proximal CRE domain mutating the functional CRE site [–71 bp]5'-TGACGTAA-3'[–64 bp] to [–71 bp]5'-CATCACC-3'[–64 bp] was generated in plasmid pXP100 (mutations M13) as previously described (16).
Transfection and luciferase assay
For transfection experiments, PCMT cells were subcultured into 24-well plates. Sequentially truncated CGA promoter constructs (16, 17, 18) were transfected as indicated into cells using Transfast (Promega) exactly according to the manufacturer’s instructions. pHook3-LacZ was routinely employed to normalize for transfection efficiency and the pHook3 vector (Invitrogen) to control for vector-dependent effects. Cells were placed in DMEM, 0.5% HS/FCS 48 h after transfection, and 6 h after the indicated treatments were washed twice with ice-cold PBS, scraped into reporter lysis buffer (Promega), and assayed for ß-galactosidase and luciferase activities using the dual-luciferase reporter assay system according to the manufacturer’s instructions (Promega).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Microarray analysis identified CGA as a putative Cox-2-dependent gene product. mRNA levels for CGA were 2-fold higher in PCXII compared with PCMT cells. Given that the parental PC12 cell line was derived from a transplantable pheochromocytoma, an adrenal medullary NET (19), and that CGA is a well-defined and clinically relevant biomarker for the development of these tumors (20), we chose to further investigate the potential regulation of CGA expression by Cox-2. Real-time PCR analysis was next carried out to investigate whether CGA was indeed up-regulated by forced Cox-2 expression. This analysis confirmed that CGA mRNA expression was 3.2-fold higher in PCXII cells when compared with their control counterparts.
Immunoblotting analysis with two different anti-CGA antisera indicated significantly higher CGA protein expression in the Cox-2-expressing cells (Fig. 1A
, second and third panels). Anti-CGA SC-1488 (second panel) is raised to an epitope mapping at the C terminus of CGA of human origin, whereas anti-CGA 20085 is raised to the native bovine protein. As can be seen, both antisera exhibited similar immunospecificity for CGA/CGA-derived fragments; endogenous proteolytic processing of CGA (and granins in general) has been extensively reported and characterized (7). Lysates were also blotted for CGB expression (Fig. 1A, bottom panel), to confirm the specificity of the observed Cox-2 dependence of CGA expression (i.e. negative control) and the equivalency of protein loading.
|
), control (vehicle-treated) PCXII cells exhibited increased levels of CGA when compared with PCMT cells. Importantly, PCMT cells treated with PGE2 exhibited significantly higher levels of CGA compared with their vehicle-treated counterparts, indicating that CGA expression is prostaglandin inducible in these cells. Conversely, PCXII cells treated with NS-398 showed a decreased level of CGA compared with their vehicle control; indeed, these levels were similar to that of vehicle-treated PCMT cells. This double proof, i.e. prostaglandin-dependent induction/pharmacological inhibition, provides additional evidence that in this model expression of CGA is positively modulated by Cox-2 in a PGE2-dependent manner.
Increased levels of intracellular CGA leads to increased secretion of this protein
Because of the secretory nature of CGA, medium conditioned by PCMT and PCXII cells was also analyzed for levels of this protein to investigate whether increased intracellular CGA expression in PCXII was reflected in increased secretion of this neuroendocrine biomarker. After immunoprecipitation and immunoblotting with anti-CGA antisera, densitometric analysis confirmed that secreted levels of CGA were higher (mean 2.40-fold) in the conditioned medium of PCXII cells (Fig. 2).
|
) (14). PCXII cells had, however, significantly higher numbers of chromaffin granules, with a 1.5-fold increase when compared with PCMT cells (Table 1). Because catecholamines are co-stored in, and co-released from, these dense-core secretory granules together with CGA (7, 19), and the predominant catecholamine produced by PC12 cells is dopamine (19), intracellular and extracellular levels of dopamine were next assessed in control and PCXII cells. As might have been expected given the increased number of chromaffin granules and the increased secretion of CGA in PCXII cells (Fig. 2), both intracellular and extracellular levels of dopamine were significantly higher in this cell-line, 3.2- and 3.4-fold, respectively (Table 1).
|
). Similarly, the anti-CGA antisera was shown to immunostain adrenal medulla, whereas the secondary antibody-only (antibody diluent) control was assessed as negative (–) and the rabbit serum control was assessed to be weakly positive (–/+; data not shown). Significantly, a clear difference between the intensity of staining for CGA between PCMT and PCXII cells was observed (Fig. 3). Although the PCMT cells stained positively (+), as would be expected given that CGA is a neuroendocrine marker and this is a rat neuroendocrine cell line, the PCXII cells stained more intensely positive (+++) for CGA than their control counterparts. Significantly, strong NSE positivity (+++) was observed in both the PCMT and PCXII cells (Fig. 3). Indeed, no distinguishable difference in the intensity of this staining was detected between the two cell lines. Because NSE expression was unaffected by forced expression of Cox-2, it would appear that this enzyme is exerting a specific effect on CGA and is not responsible for ubiquitous modulation of neuroendocrine marker expression/differentiation.
|
). Because PGE2 is a cAMP-generating stimulus, it was hypothesized that Cox-2 might effect CGA up-regulation via PGE2-dependent CRE activation. For the purposes of the current study, the mouse (CRE –71 to –64 bp) CGA promoter constructs pXP1200, pXP100, pXP77, and pXP61 and the site-directed CRE mutant reporter construct pXPM13 (16) were transfected into PCMT cells. Experiments confirmed that the cAMP-dependent ligands forskolin (1 mM) and nicotine (1 mM) both activated transcription of pXP1200 and pXP100 in PCMT cells (data not shown). Because the latter produced the greater reporter gene activity, this was used as a positive control stimulus in subsequent experiments. PGE2 (1 µM) stimulated luciferase activity (Fig. 4) in PCMT cells transfected with constructs containing an intact CRE, namely pXP1200 and pXP77, 9.3- and 5.8-fold over control (pXP1), respectively. Markedly lower prostaglandin-dependent, and indeed, nicotine-dependent, levels of reporter gene activity were, however, observed in cells transfected with CGA promoter constructs in which the CRE was either absent (pXP61: PGE2 2.6-fold; nicotine 3.0-fold) or functionally inactive (pXPM13: PGE2 2.1-fold; nicotine 1.3-fold). These results confirm that PGE2 requires a functional CRE to elicit CGA promoter activation. Taken together, these data support the contention that Cox-2 up-regulates CGA gene transcription in a PGE2-dependent manner via activation of a key proximal CRE.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The biological consequences of Cox-2-dependent CGA up-regulation in a NET-derived cell line are intriguing. It has been reported that, although absent from normal adrenal medulla, Cox-2 is up-regulated in pheochromocytoma tissue and may be involved in promoting an aggressive tumor phenotype (3, 31). Moreover, Cox-2 has been implicated in gastrointestinal and bronchopulmonary carcinoid progression (4) and found to be associated with a poor prognosis in mid-gut carcinoid patients (31). In light of these findings, a model could be proposed whereby CGA, or its peptide fragments, may have pro-tumorigenic actions. The precise biological function of elevated CGA in neuroendocrine, and indeed nonneuroendocrine, neoplasms is unclear. Limited studies in cell and animal models have provided contradictory evidence as to whether CGA promotes or inhibits tumorigenesis. It has been reported that CGA inhibits apoptosis (24). However, CGA or its fragments have also been demonstrated to inhibit cell proliferation and invasion (25), whereas forced overexpression of CGA in vivo has been demonstrated to reduce tumor growth (26, 27). Given that CGA levels have been shown to correlate with tumor burden and prognosis in a range of NETs (7, 20), it is likely that this protein functions as more than an inert biomarker of disease. Moreover, given that the identification of in vivo neoplastic processing products of CGA is an area of emerging interest (28), it is conceivable that the biological activity of peptides released as the result of tumor-dependent proteolysis remains to be fully characterized.
The findings of the current study also have specific relevance to the pathogenesis of pheochromocytoma. It has been reported that essential hypertension is characterized by elevated CGA levels and reduced levels of the catecholamine release-inhibitory CGA fragment catestatin (7). Furthermore, a recent study using CGA-deficient mice has confirmed that loss of CGA results in elevated blood pressure and that this is ameliorated by insertion of the human CGA gene or exogenous catestatin replacement (7, 8). These results lend support to the contention that CGA modulates blood pressure by a dual mechanism, namely, control of neurotransmitter storage and release (via its role in secretory granule biogenesis) and generation of the physiological inhibitor of this release, catestatin (via its proteolytic processing). In light of the findings of the current study, it may be speculated that pheochromocytoma-derived Cox-2 may, via CGA up-regulation (and possibly abnormal proteolytic processing) facilitate the increased catecholamine levels and hypertension characteristic of this particular neuroendocrine neoplasm (7, 29). Furthermore, given the importance of Cox-2-derived eicosanoids, particularly renal thromboxane, PGE2, and PGI2, to blood pressure homeostasis (30), pheochromocytoma-associated Cox-2 up-regulation could also potentially contribute to malignant hypertension via the direct vasoactive effects of its downstream products.
In conclusion, the current study provides evidence that Cox-2 up-regulates CGA expression in a neuroendocrine cell line. The relevance of this finding to NET development will provide a focus for future investigations.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online May 31, 2007
Abbreviations: CGA, Chromogranin A; Cox-2, cyclooxygenase-2; CRE, cAMP-responsive element; FCS, fetal calf serum; FITC, fluorescein isothiocyanate; HS, human serum; IPTG, isopropyl-1-thio-ß-D-galactopyranoside; NSE, neuron-specific enolase; NET, neuroendocrine tumor; PGE2, prostaglandin E2.
Received February 5, 2007.
Accepted for publication May 15, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CT method. Methods 25:402–408[CrossRef][Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
