| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Centre for Endocrine and Diabetes Sciences,
Department of Medical Biochemistry and Immunology, School of Medicine, Cardiff University, Cardiff, UK;
Institut fuer Physiologie, Medizinische Fakultaet Carl Gustav Carus, Dresden, Germany; and
University of Glasgow, Department of Pathology, Royal Infirmary, Glasgow, UK
1Correspondence: Centre for Endocrine and Diabetes Sciences, School of Medicine, Cardiff University, Cardiff CF14 4XN, UK. E-mail: wmdjh{at}cardiff.ac.uk
ABSTRACT
Adenosine is known to stimulate interleukin (IL)-6 and vascular endothelial growth factor (VEGF) secretion from pituitary TtT/GF folliculostellate (FS) cells indicating that it is an important paracrine regulator of anterior pituitary function. This study demonstrates that rodent anterior pituitary cell lines produce extracellular adenosine that is able to increase intercellular gap junction communication in FS cells. Ecto-5'-nucleotidase (CD73), the enzyme that generates adenosine from AMP, was demonstrated by immunocytochemistry in 
20% of anterior pituitary cells, and some of these cells colocalized with prolactin and growth hormone. CD73 mRNA and protein were detected in GH3 and MMQ (somatotroph-lactotroph lineages) and TtT/GF cells, and enzyme activity was demonstrated by the conversion of exogenously added fluorescent ethenoAMP to ethenoadenosine. Adenosine production, as measured by HPLC, was detected in GH3 (1 µM/h) and MMQ (3 µM/h) but not in TtT/GF cells. Adenosine (EC50: 0.5µM) and NECA (universal adenosine receptor agonist; EC50 0.1 µM) stimulated connexin 43 (Cx43) mRNA and protein expression within 1–2 h in TtT/GF cells. Adenosine and NECA also stimulated gap junctional intercellular communication (as assessed by transmission of Alexa Fluor 488) by 6- to 8-fold in comparison with untreated TtT/GF cells. In cocultures of MMQ and TtT/GF cells, Cx43 expression in TtT/GF cells increased in proportion to the number of MMQ cells plated out. These data suggest that adenosine, formed locally in the anterior pituitary gland can stimulate gap junction communication in FS cells.—Lewis, B. M., Pexa, A., Francis, K., Verma, V., McNicol, A. M., Scanlon, M., Deussen, A., Evans, W. H., Rees, D. A., Ham, J. Adenosine stimulates connexin 43 expression and gap junctional communication in pituitary FS cells.
Key Words: CD73 • AMP • NECA • pituitary gland
THE PURINE NUCLEOSIDE, adenosine is formed by dephosphorylation of adenosine monophosphate (AMP), an intermediate molecule in a continual cycle of ATP formation and breakdown occurring in most cells (1)
. This cycle is mediated by a cascade of specific enzymes that occur within and outside of the cell. Within the cell, the major regulatory enzyme is adenosine kinase, which converts adenosine back to AMP, whereas outside of the cell the major enzyme is adenosine deaminase (AD), which converts adenosine to inosine. Adenosine transport may occur bidirectionally by facilitated diffusion or by secondary active transport mechanisms that use the sodium concentration gradient to cross the extracellular and intracellular boundaries (1)
. During hormone release and particularly metabolic stress, there is accumulation of nucleotides and nucleosides that leads to increased transport of ATP and related molecules out of the cell. During cell stress, the rate of ATP breakdown is increased such that adenosine accumulates extracellularly (2
, 3)
. Extracellular adenosine is able to activate the family of P1 (adenosine or A) receptors that comprise four subtypes: A1, A2a, A2b, and A3, the actions of which are mediated primarily through the stimulation (A2a and A2b) or inhibition (A1 and A3) of adenylate cyclase (4)
.
The presence of A1 (5)
and A2 receptors (6)
in the anterior pituitary has been well described, and there have been numerous functional studies indicating their relevance to hormone secretion and cellular growth (6
, 7
8
9
10
11)
. Recent studies showed that adenosine receptors in the pituitary gland are selectively expressed with hormone expressing endocrine cells possessing mainly A1 receptors and "nonendocrine" folliculostellate (FS) cells possessing predominantly A2b receptors (12)
. We have also shown that adenosine, acting at A2b receptors and via the protein kinase C and p38 MAPK pathways, is a potent inducer of IL-6 and vascular endothelial growth factor (VEGF) secretion in TtT/GF FS cells (13)
. Although it is clear that adenosine can activate the anterior pituitary gland by stimulating adenosine receptors, the actual source of adenosine in this scenario is unknown. The possibility that adenosine could accumulate in the anterior pituitary gland was suggested in 1995 (14)
when a pituitary cytotrophic factor, indistinguishable from adenosine, was shown to be transported by retrograde blood flow to the brain where it stimulated dopamine release. The cell types within the anterior pituitary gland that produce and secrete adenosine have not yet been determined.
The promotion or inhibition of growth in the anterior pituitary gland may be regulated by intercellular communication that is channelled through gap junctions. Gap junctions are formed by the docking of two connexin hemichannels (connexons) (15)
on adjacent cells resulting in full channels that allow transmission of molecules up to 1–1.2 kDa in size. Connexons are composed of proteins called connexins, with connexin 43 (Cx43) being the most prominent subtype within the anterior pituitary gland (16
, 17)
. Gap junctions in the anterior pituitary gland are formed mainly between individual FS cells or between FS cells and lactotrophs or somatotrophs. Some years ago, it was shown that cell growth is inversely related to the degree of cell communication (18
, 19)
and exposure to mitogens leads to a rapid, but transient, inhibition of cellular communication before cells undergoing cell division (20
, 21)
. Cx43 redistribution has also been associated with growth control in adrenal cells (22)
and increased expression leads to increased proliferation and differentiation of osteoblast-like cells (23)
. Cx43 hemichannels also appear to be a transducer of survival signals in response to extracellular cues (24)
. On the other hand, Cx43 expression inhibits growth and is independent of gap junction formation (25
, 26)
.
It is possible that the trafficking of small molecules, such as adenosine or ATP, through gap junctions may exert regulatory influences on cell proliferation and differentiation. Since ATP decreases intercellular communication through gap junctions in astrocytes (27)
and ATP depletion in human renal proximal tubule cells activates gap junction hemichannels (28)
, it is conceivable that adenosine too may have actions at gap junctions. In addition, it has been shown that cell to cell signaling associated with connexin expression is due to enhanced ATP secretion (29)
, a process involving movement across connexin hemichannels that is inhibited by connexin mimetic peptides (30)
. We hypothesized that adenosine, produced locally in the pituitary gland, might influence Cx43 expression and modify intercellular transmission directly through gap junctions and indirectly via connexin hemichannels.
MATERIALS AND METHODS
Cell culture materials and reagents were from Invitrogen (Paisley, UK), Autogen Bioclear (Calne, UK), Sarstedt Ltd. (Leicester, UK), and Sigma-Aldrich (Poole, UK). 5'-N-ethylcarboxamidoadenosine (NECA), adenosine, and enzyme inhibitors such as erythro-9-(2-hydroxy-3-nonyl)-adenine (EHNA), iodotubericidin (ITU), and 
ß-methylene-adenosine diphosphate (AOPCP) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Rat GH3 and MMQ cell lines were in-house, and TtT/GF cells were kindly provided by Professor Kinji Inoue (Department of Regulation Biology, Saitama University, Urawa, Japan). Specific antisera to ecto-5'-nucleotidase (CD73), AD, and ß-actin (goat polyclonal) were from Santa Cruz Biotechnology (Santa Cruz, CA, USA), Cx43 (rabbit polyclonal) from Zymed Laboratories, and rat and human prolactin and human growth hormone (GH) were from the National Institute of Arthritis, Diabetes, and Kidney Diseases (Bethesda, MD, USA).
Cell culture
Rat GH3 and MMQ and mouse TtT/GF and AtT-20 cells were cultured in the following media: respectively, Ham’s F-10, 15% horse serum (HS), and 2.5% FCS; Ham’s F12 15% HS and 2.5% FCS; Dulbecco’s modified Eagle’s medium (DMEM) 10% heat-inactivated FCS and DMEM 10% FCS. For experimental purposes, MMQ cells were seeded into poly-L-lysine (70–150 kDa, 0.1 mg/ml) coated dishes or "Thermanox" coverslips (Invitrogen) (for immunocytochemistry).
To ascertain whether pituitary endocrine cells could directly affect Cx43 expression in FS cells, MMQ and TtT/GF cells were set up in a coculture system. In this system, products that are produced in the cell-conditioned media of the MMQ cells can pass through the polycarbonate membrane and affect the function of TtT/GF cells. TtT/GF cells (0.05x106 cells/cm2) were plated into 24-well plates and cultured overnight, before coincubating for 24 h with MMQ cells (0, 0.01, 0.05, and 0.25x106). The MMQ cells were plated into cell culture inserts (polycarbonate membrane 0.4 µm, Nunc) in 0.5 ml DMEM containing 1% FCS and EHNA (5 µM) and ITU (10 µM). A similar volume of fresh media was used for the TtT/GF cells. After 24 h, the culture media in the culture dish wells were harvested and adenosine levels measured (see below), and Cx43 expression in TtT/GF cells was determined by Western blot analysis (see below).
Immunocytochemistry
Rat pituitaries were taken from adult male Wistar rats that had previously been anesthetized with sodium pentobarbitone (50 mg/kg) and perfused with heparinized saline and freshly prepared 4% paraformaldehyde in 0.1 M PBS. Tissues were postfixed in paraformaldehyde and paraffin-wax embedded. Human pituitary was obtained at autopsy from patients with no evidence of endocrine dysfunction. Consent for use of tissues was appropriate, and the studies were approved by the Research Ethics Committee of Glasgow Royal Infirmary. Human tissues were also fixed in paraformaldehyde and wax embedded.
Vectastain avidin-biotin complex (ABC) kits (Vector Laboratories, Burlingame, CA) were used for immunocytochemistry and were carried out in paraformaldehyde fixed rat pituitary tissue and acetone-fixed anterior pituitary cell-line monolayers. Antigen retrieval was performed by heating in a microwave for 30 min in 10 mM citrate buffer pH6. Incubations with specific antisera were carried out overnight at 4°C (1 h at room temperature in the case of ß-actin) at the following concentrations: CD73 and AD (1:200) and rat and human prolactin and human GH (1:2000). Sections were then incubated for 90 min with the appropriate second antibody (Ab) coupled to biotin and then visualized with streptavidin-fluorescein or diaminobenzidine. Nuclei were stained with 4',6'-diamidino-2-phenylidole (DAPI) or with hematoxylin. For colocalization studies, vector blue was used in addition to diaminobenzidine and for the negative controls, nonimmune sera were used in place of the specific antisera.
Reverse transcriptase-polymerase chain reaction and quantitative reverse transcriptase-polymerase chain reaction
Total cellular RNA was prepared using TRIzol reagent (Invitrogen, Carlsbad, CA) and treated with RQ1 ribonuclease-free deoxyribonuclease (Promega, Madison, WI); 0.2 µg RNA were reverse transcribed using oligodeoxythymidilic acid [oligo(dT)15] for 1 h at 37°C, and the cDNA generated was subjected to polymerase chain reaction (PCR) amplification using primers specific for rat or mouse CD73 and AD. Primer sequences were designed using the Primer 3 software program and gene sequences obtained from GenBank and are shown in Table 1
. For the PCR reaction, 30–40 cycles of the reaction: 94, 60, and 72°C for 30 s, 1 min and 2 min, respectively, and a final extension step of 72°C for 10 min, were performed as appropriate. Amplified products were electrophoresed in 2% agarose and visualized with ethidium bromide.
|
Changes in Cx43 mRNA expression after NECA treatment were analyzed by comparative quantitative reverse transcriptase (QRT)-PCR (using MX-3000 machine, Stratagene, La Jolla, CA) in TtT/GF cells. Briefly, TtT/GF cells, plated into replicate 60 mm dishes at a density of 0.75 x 106 cells/dish, were incubated overnight and treated for 0, 1, 2, 4, 8, and 24 h with NECA. The primer sequences for Cx43 and PGK-1 and ß-actin (both house-keeping genes) are shown in Table 1
. The reaction mixture contained 1 µl (equivalent to 40 ng RNA) of cDNA, 12.5 µl QPCR SYBR green Mastermix (Stratagene), 2.5 mM MgCl2, and 500 nM of each primer in a volume of 25 µl. The annealing temperature was 60°C. The ratios of Cx43 to PGK-1 or ß-actin in treated and untreated cells were compared and a ratio value of 1:1 was assigned for untreated cells.
Adenosine production
The time-related production of adenosine in conditioned media from GH3, MMQ, and TtT/GF cells was measured. Briefly, GH3, MMQ (both 0.1xl06 cells/cm2), or TtT/GF cells (0.05xl06 cells/cm2) were plated out into 48-well dishes. The cells were incubated overnight, washed in the appropriate serum free media (see cell culture section), and preincubated in the same for 1 h. Experiments were carried out in the presence of the AD inhibitor, EHNA (5 µM) and iodotubericidin, (l0 µM), an adenosine kinase inhibitor, in a volume of 0.5ml. Conditioned media were collected at 30, 60, 120, and 180 min and acidified with 1 M perchloric acid (PCA; 0.025 ml) for 2 min and then neutralized with 0.5 M K3PO4 (0.05 ml) before storage at –20°C. Cells were counted at the end of the experiments, and the adenosine levels were corrected for unit cell number. Samples for adenosine measurement were derivatized and analyzed by fluorescent HPLC (see below).
CD73 enzyme activity
CD73 enzyme activity was measured in GH3, MMQ, and TtT/GF conditioned media. The cells were plated, at the same density as for the adenosine release experiments, into replicate wells of 12-well dishes and similarly treated before experimentation. Cells were incubated in a 2 ml volume of serum-free medium in the absence or presence of the CD73 inhibitor ß-methyleneadenosine-5'-diphosphate, AOPCP, 50 µM) or the alkaline phosphatase inhibitor levamisole (50 µM) for 15 min before the addition of 5 µM 1,N 6-ethenoAMP (a bioactive, fluorescent analog of AMP) for up to 2 h; 0.2 ml of conditioned media were removed at the following time points: 5, 15, 30, 60, and 120 min. from each well and acidified and neutralized as before. CD73 activity was measured as production of 1,N 6-etheno-adenosine from preadded 1,N 6-ethenoAMP and determined by fluorescent HPLC.
HPLC analysis
To measure adenosine, samples were derivatized and then analyzed by fluorescent HPLC (31)
. Lyophilized samples were reconstituted in 0.143 ml Krebs-Henseleit buffer [118.4 mM NaCl, 4.7 mM KC1, 25 mM CaCl2, 1.2 mM KH2PO4, 1.2 mM MgSO4, 25 mM NaHCO3, and l0 mM glucose, adjusted to pH 4 with 8 M NaOH], followed by the addition of 0.051ml citrate-phosphate buffer (pH 4) and 0.005 ml chloracetaldehyde (45% v/v). Samples were then incubated for 40 min at 80°C and rapidly cooled to 4°C before subjecting to HPLC analysis. For quantification of adenosine, standards of known concentrations were similarly treated.
CD73 enzyme activity was determined as breakdown of ethenoAMP to ethenoadenosine. HPLC analysis of etheno-compounds was performed on a Waters Alliance 2690 machine coupled to a Merck-Hitachi F 1050 fluorescence detector (
ex=280 nm, em=410 nm). 0.01 ml samples were injected by an autosampler onto a Waters XTerra MS C18 column (50x4.6 mm id, 5 µm particle size, pore size 125E), and compounds were eluted with a binary gradient of tetrabutyl ammonium hydrogen sulfate (TBAS) buffer (5.7 mM TBAS and 30.5 mM KH2PO4 adjusted to pH 5.8 with 2 M KOH) and acetonitrile at a flow rate of 1.5 ml/min. The starting eluate was 100% buffer A (6% acetonitrile in TBAS buffer) followed by a linear gradient to 66% buffer A and 34% buffer B (65% acetonitrile in TBAS buffer) within 1 min. This eluate mixture was held constant for 1.4 min, restored back to 100% buffer A within 0.1 min, and held and equilibrated for 4.5 min. The retention times for l,N 6-ethenoadenosine and l,N 6-ethenoAMP were 1.5 and 1.8 min, respectively. External standards of known concentration were used for identification and quantification. The Millenium software (Waters Corp.) was used for data analysis.
Western blot analysis
GH3, MMQ, and TtT/GF cells were cultured at a density of 1 and 0.5 x 106 cells/well, respectively, in six-well multidishes. Cell monolayers were allowed to reach 80% confluence before switching to medium containing 1% FCS. Time-course experiments (10 µM NECA for 0–8 h.) and dose-response experiments (0–100 µM NECA or adenosine for 4 h) were performed. Experiments with adenosine were carried out in the presence of 10 µM iodotubericidin and 5 µM EHNA. Cells were rinsed 3 times with 1mM sodium orthovanadate in PBS and then lysed in 200 µl RIPA buffer (Santa Cruz Biotech, Santa Cruz, CA) containing 1 mM sodium orthovanadate, 0.1 mg/ml PMSF, chymostatin, leupeptin, antipain and pepstatin (all at 10 µg/ml). Lysates after centrifugation were stored at –80C. Ten microliter aliquots from each sample were mixed with an equal volume of electrophoresis buffer, boiled, and electrophoresed in 10% polyacrylamide. Proteins were transferred onto PVDF membranes and incubated either overnight at 4°C or 1 h at room temperature with, respectively, antisera to Cx43 (1:3000) and ß-actin (1:1000). Secondary antisera conjugated to horseradish peroxidase, at 1:2000, were applied for 1 h at room temperature, and proteins were visualized with enhanced chemiluminescence (ECL) Plus reagent.
Dye transfer across gap junctions
TtT/GF cells were seeded into tissue culture dishes at a density of 75 x 103 cells/cm2 and incubated overnight. Dishes were washed in DMEM 1% FCS with 20 mM HEPES and incubated for 4 h with 10 µM NECA or adenosine in the same media. Microinjection was performed using an Eppendorf Femtojet 5247 device; briefly 1 µl of Alexa Fluor 488 (10 mM) dye was loaded into the apparatus and 50 picolitres microinjected intracytoplasmically into each of 20 cells in each dish that was maintained at 37°C. The individually injected cells were sufficiently far apart on the dish such that the dye transfer into neighboring cells could only have arisen from a single injected cell. In some experiments, the gap junction blocker, octanol (1 mM) was added 30 min before the addition of the dye. 10 min after microinjection, the dye transferred into neighboring cells was visualized and cells counted using a Zeiss Axiovert 200 fluorescent microscope. The number of cells that contained dye in the untreated and treated cells was compared.
Statistics
Experiments were performed two to six times (n). Results, as appropriate, were expressed as mean ± SE and compared by ANOVA and Tukey multiple comparison test. P < 0.05 was deemed to be significant.
RESULTS
CD73 and Adenosine deaminase
mRNA encoding CD73 and AD, two enzymes involved directly in the catabolism and metabolism of extracellular adenosine, was demonstrated by RT-PCR of anterior pituitary tissue and in all anterior pituitary cell lines tested (Fig. 1
). Immunocytochemistry was used to examine the distribution of CD73 and AD protein in sections of rat pituitary and showed that 20% of cells stained for each molecule. These CD73 and AD positive cells also clustered around the sinusoids within the gland (Fig 2
A, B) and showed a similar staining pattern to a subset of prolactin cells. The staining for prolactin and AD was performed in adjacent pituitary sections; very intense prolactin staining was found in a few cells and more diffuse prolactin staining in many more cells (Fig. 2C
). Under light microscopy, it is not known whether these cell types correspond to Type 1 and Type 2 prolactin cells (32)
. CD73 and AD immunostaining seemed to correspond to the light microscopic characteristics of the abundant prolactin cells.
|
|
CD73 and AD immunostaining was also observed in acetone-fixed GH3 (prolactin/GH secreting) and MMQ (prolactin secreting) cell lines. The presence of AD (Fig. 2D
) and CD73 (Fig. 2E
) in MMQ cells indicates that CD73 and AD are likely to be colocalized together as well as with prolactin. These findings were extended to the human pituitary gland, and we demonstrated colocalization of CD73 with some prolactin- (Fig. 2F
) and GH-producing cells (Fig. 2G
). Colocalization of CD73 with prolactin or GH is indicated by the arrows. However, not all of the prolactin- or GH-secreting cells immunostained for CD73, and some cells expressed only CD73. The antiserum used for the detection of AD proved ineffective in paraffin-wax-embedded sections.
Adenosine production and CD73 enzyme activity
The presence of CD73 in cells of prolactin/GH lineages indicates that they may be a source of adenosine produced by enzymatic cleavage of ATP, via ADP and AMP. We thus investigated the production of adenosine by the GH3 and MMQ cell lines. The time-related basal accumulation of extracellular adenosine from MMQ and GH3 (and for comparison TtT/GF) cells was measured by HPLC. Both MMQ and GH3 cells secreted adenosine in a linear fashion with the former showing a 3-fold greater rate of production (3–4 µM after 1 h in MMQ cells; Fig 3
A). On the other hand, adenosine production was barely detectable (<0.3 µM) in the media from TtT/GF cells and there was little difference in values between any of the time points.
|
Extracellular adenosine can be produced directly by cell secretion or as a consequence of extracellular AMP dephosphorylation by CD73. We thus investigated whether the pituitary cell lines dephosphorylate exogenously added etheno-AMP to adenosine. Figure 3B
shows that etheno-AMP is degraded to adenosine by MMQ, GH3, and TtT/GF cells at similar rates (50–60% within 1 h and completely within 2 h). AOPCP, but not levamisole, blocked the degradation of etheno-AMP in GH3 cells, indicating that the mediating enzyme is CD73 and not alkaline phosphatase (Fig. 3B
). The degradation of AMP in MMQ and TtT/GF cells does not, however, appear to be mediated by CD73 or alkaline phosphatase, as neither AOPCP nor levamisole had any effect on the rate of production of adenosine.
Connexin 43 expression
The universal adenosine receptor agonist NECA (10 µM) stimulated Cx43 expression in TtT/GF cells (a representative Western blot analysis blot is shown in Fig. 4
A). Cx43 expression increased after 30 min of incubation and was stable from 2–8 h, and densitometric measurements from four separate experiments (Fig. 2B
) showed a 3-fold increase in expression at these time points. Two distinct bands were also visible on the gel films and these corresponded to phosphorylated and nonphosphorylated forms of Cx43 (33)
with both forms increasing after NECA exposure. Dose-response experiments were carried out at 4 h, and both NECA and adenosine showed dose-related stimulation of Cx43 expression (a representative Western blot analysis blot for the effect of adenosine is shown in Fig. 4C
). Figure 4D
shows the mean densitometric values for three separate experiments; the EC50 values for adenosine and NECA stimulation of Cx43 were estimated to be 0.5 (Fig. 4D
) and 0.1 µM (data not shown), respectively. In addition to changes in Cx43 protein expression, NECA also stimulated increases in mRNA expression. Quantitative RT-PCR showed that, relative to either ß-actin or PGK-1, NECA stimulated a 4- to 5-fold (P<0.001) and 2- to 3-fold (P<0.01) fold increase in Cx43 mRNA after, respectively, 1 and 2 or 4 h of exposure. By 8 h, the relative expression of Cx43 had returned to normal.
|
Intercellular communication
Figure 5
A–E shows the effect of NECA and adenosine on dye transfer in TtT/GF cells. Single cells were microinjected with Alexa 488, and dye transfer into adjoining cells was monitored after 10 min. Compared to untreated cells, 10 µM NECA and adenosine stimulated dye transfer into neighboring cells by up to 8-fold, [no treatment: 4±1.7 cells, adenosine-treated 24±5.7 (P<0.001); NECA-treated 33±7.8 cells (P<0.001); Fig. 4A-C
]. In the presence of octanol, the stimulatory effects of adenosine and NECA on dye transfer were completely inhibited (Fig. 4D-E
). Octanol also inhibited dye transfer in control cells (data not shown). Figure 4F
shows the mean values of the number of cells that contained dye after each treatment in three to six experiments. For the duration of these experiments, none of the treatments (NECA, adenosine or octanol) had any effect on cell morphology when observed using phase contrast microscopy.
Coculture of TtT/GF and MMQ cells
In coculture experiments, Fig. 6
shows that on increasing the number of MMQ cells there is a cell number related increased expression of both the non- and phosphorylated forms of Cx43 in the TtT/GF cells (Fig. 6A, B
). In the same experiments, adenosine levels, measured in the conditioned media surrounding the TtT/GF cells were, however, unexpectedly low (Fig. 6C
) when compared with that observed with MMQ cells alone (Fig. 3A
).
|
DISCUSSION
Anterior pituitary cells express several adenosine receptors (7
8
9
10
, 12
13
, 34)
and their effects on mediating hormone secretion, although well-studied, remain confusing with both stimulatory and inhibitory actions having been demonstrated. For example, early reports (6
, 9)
showed that adenosine inhibited prolactin secretion but others (10
11)
showed adenosine stimulated prolactin secretion. Nevertheless, it is clear that functional adenosine receptors do exist on cells in the anterior pituitary gland. To understand more about the role of adenosine in the anterior pituitary gland, we investigated whether pituitary cells released adenosine into the extracellular matrix.
The presence of adenosine outside of the cell may be due to increased transport or enhanced degradation from ATP. The present RT-PCR data suggest that most pituitary cells have the potential to produce adenosine from ATP, but immunocytochemical data indicate that their expression is much more localized and that only 20% of cells in the rat anterior pituitary were positive for CD73 or AD protein. Although CD73 and AD colocalization experiments were not carried out, the presence of encoding mRNA and proteins for both molecules was demonstrated in several rodent pituitary cell lines including GH3 and MMQ (somatotroph/lactotroph), AtT-20 (corticotroph), and TtT/GF (FS). All CD73- and AD-positive cells in the rat anterior pituitary gland border the sinusoids or blood vessels suggesting that they are cells that secrete into the circulation. These cells also show a pattern that is characteristic of lactotrophs and evidence for colocalization of CD73 with prolactin and GH was found in normal human pituitary sections. However not all lactotrophs and somatotrophs expressed CD73 and some CD73 positive cells were neither lactotrophs nor somatotrophs. It is conceivable that there may be a specific subpopulation of cells in the anterior pituitary gland that expresses CD73 and AD, but the precise role of such cells, if they exist, is unknown.
We identified CD73 in the pituitary gland and associated cell lines and showed that the enzyme was active by monitoring the breakdown of exogenously added fluorescently labeled AMP. In GH3 cells, CD73 was indeed shown to be responsible for the degradation of ethenoAMP, but in MMQ and TtT/GF cells the degradation of ethenoAMP was not inhibited by either AOPCP or levamisole, indicating that the mediating enzyme was neither CD73 nor alkaline phosphatase (35
36)
. Although MMQ and TtT/GF cells do express CD73, the mechanism of ethenoAMP breakdown in these cells remains to be determined. It is possible that MMQ and TtT/GF cells do not express the dimeric form of CD73 that is required for functional activity (37)
. On the other hand, the conversion of ethenoAMP to ethenoadenosine may be due to activities of nonspecific pyrophosphatases. Non-AOPCP inhibitable nucleotide phosphatases have been described previously (38)
.
Our data suggest that lactotroph- or somatotroph-like cells may be an important source of adenosine, although this does not preclude the fact that other cell types are also likely to produce adenosine, especially in times of cell stress, hypoxia, or inflammation. It is thus likely that pituitary cells respond to adenosine produced in either an autocrine or paracrine fashion. FS cells, which form a cellular network allowing propagation of Ca2+ waves between other FS cells and endocrine cells (39
40)
, are noted for their paracrine roles in the regulation of pituitary gland function. Intercellular communication between FS cells and lactotrophs has been particularly well documented and shown to be bidirectional. Lactotrophs activate FS cells by the production of TGFß3 and conversely FS cells release basic fibroblast growth factor (bFGF) that induces growth of lactotrophs (41
42)
. Our data from the coculture experiments also indicate that lactotrophs can influence the activity of FS cells (Fig. 6)
. The levels of adenosine in these experiments (24 h), however, were considerably lower than that observed for MMQ cells alone (1–3 h; Fig. 3A
). The reason for this is not clear; a possible explanation is that there were other enzymes (hydrolases, kinases, etc.) present in the tissue culture sera that may have modified the structure of the adenosine molecule. Additionally adenosine may also be taken up by the TtT/GF cells themselves. Nevertheless, the observed increased Cx43 expression in TtT/GF cells correlates with the number of MMQ cells plated out.
A likely route for communication of small molecules with signaling potential between neighboring cells is via intercellular gap junctions. This is particularly pertinent for the passage of small molecules such as adenosine, ATP, glutamate, ions, and amino acids (15)
. The association of ATP and gap junctions is well recognized and, for example, it has been shown that ATP decreases gap junction communication in astrocytes (27)
and down-regulation of Cx43 leads to decreased expression of the P2Y1 and increased expression of the P2Y4 receptors (43)
. A further paracrine route by which ATP is released by cells under stress via connexin hemichannels should also be considered. Before hydrolysis, ATP can activate purinergic receptors in neighboring cells in a paracrine fashion, and this underpins the propagation of calcium wave signaling between cells, also in a connexin-dependent manner (44)
. Although genome wide expression profiling in neuroblastoma cells shows that chloroadenosine can up-regulate Cx26 (45)
and that adenosine, in C6 glioma cells, can pass more efficiently through channels composed of Cx32 molecules (46)
, there is no information on a link between adenosine and gap junctions in the pituitary gland. Our findings, as far as we aware, are the first report of adenosine having a stimulatory action on Cx43 expression and intercellular communication. We also showed however that adenosine stimulates the expression of phosphorylated and nonphosphorylated forms of Cx43; the functional consequences of this, in the context of gap junction channel gating, is unclear (47)
. These findings are consistent with the reports showing that inhibition of cellular metabolism or ATP depletion induces gap junction communication (20
, 48)
. We concentrated on studying Cx43 rather than other connexin molecules as Cx43 appears to be the major and perhaps only form of connexin the pituitary gland (49
, 50)
In summary, our data show that the pituitary gland possesses the machinery to generate adenosine from ATP and that adenosine influences gap junction communication. Adenosine, however, also influences the function of pituitary cells (6
7
8
9
10
11
12
13)
including its contrasting effects on the growth of lactotrophs and FS cells (12)
, suggesting that other molecules may also play a pivotal roles in the function of these cells. On the other hand, it is possible that gap junction communication between FS cells and between FS cells and lactotrophs responds differently to adenosine.
|
Received for publication March 27, 2006. Accepted for publication August 21, 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
