SciELO - Scientific Electronic Library Online

 
vol.39 issue11Cytotoxic activity of the dichloromethane fraction from Vernonia scorpioides (Lam.) Pers. (Asteraceae) against Ehrlich's tumor cells in miceElectrical field stimulation improves bone mineral density in ovariectomized rats author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

Share


Brazilian Journal of Medical and Biological Research

Print version ISSN 0100-879XOn-line version ISSN 1414-431X

Braz J Med Biol Res vol.39 no.11 Ribeirão Preto Nov. 2006  Epub Sep 22, 2006

 

Braz J Med Biol Res, November 2006, Volume 39(11) 1493-1499

Adenosine A1 receptor-mediated inhibition of in vitro prolactin secretion from the rat anterior pituitary

D.L.W. Picanço-Diniz1, M.M. Valença2 and J. Antunes-Rodrigues3

1Departamento de Fisiologia, Centro de Ciências Biológicas, Universidade Federal do Pará, Belém, PA, Brasil
2Departamento de Neurologia e Psiquiatria, Centro de Ciências da Saúde, Universidade Federal de Pernambuco, Recife, PE, Brasil
3Departamento de Fisiologia, Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto, SP, Brasil

Abstract
Introduction
Material and Methods
Results
Discussion
References
Acknowledgments
Correspondence and Footnotes


Abstract

In previous studies, we demonstrated biphasic purinergic effects on prolactin (PRL) secretion stimulated by an adenosine A2 agonist. In the present study, we investigated the role of the activation of adenosine A1 receptors by (R)-N6-(2-phenylisopropyl)adenosine (R-PIA) at the pituitary level in in vitro PRL secretion. Hemipituitaries (one per cuvette in five replicates) from adult male rats were incubated. Administration of R-PIA (0.001, 0.01, 0.1, 1, and 10 µM) induced a reduction of PRL secretion into the medium in a U-shaped dose-response curve. The maximal reduction was obtained with 0.1 µM R-PIA (mean ± SEM, 36.01 ± 5.53 ng/mg tissue weight (t.w.)) treatment compared to control (264.56 ± 15.46 ng/mg t.w.). R-PIA inhibition (0.01 µM = 141.97 ± 15.79 vs control = 244.77 ± 13.79 ng/mg t.w.) of PRL release was blocked by 1 µM cyclopentyltheophylline, a specific A1 receptor antagonist (1 µM = 212.360 ± 26.560 ng/mg t.w.), whereas cyclopentyltheophylline alone (0.01, 0.1, 1 µM) had no effect. R-PIA (0.001, 0.01, 0.1, 1 µM) produced inhibition of PRL secretion stimulated by both phospholipase C (0.5 IU/mL; 977.44 ± 76.17 ng/mg t.w.) and dibutyryl cAMP (1 mM; 415.93 ± 37.66 ng/mg t.w.) with nadir established at the dose of 0.1 µM (225.55 ± 71.42 and 201.9 ± 19.08 ng/mg t.w., respectively). Similarly, R-PIA (0.01 µM) decreased (242.00 ± 24.00 ng/mg t.w.) the PRL secretion stimulated by cholera toxin (0.5 mg/mL; 1050.00 ± 70.00 ng/mg t.w.). In contrast, R-PIA had no effect (468.00 ± 34.00 ng/mg t.w.) on PRL secretion stimulation by pertussis toxin (0.5 mg/mL; 430.00 ± 26.00 ng/mg t.w.). These results suggest that inhibition of PRL secretion after A1 receptor activation by R-PIA is mediated by a Gi protein-dependent mechanism.

Key words: Cyclopentyltheophylline, Dibutyryl cAMP, R-PIA, Pertussis toxin, Cholera toxin


Introduction

The modulation of the activity of adenyl cyclase is one of the main functions described for purinergic mechanisms involving type A1 and A2 outer membrane receptors (1). The ability of A1 and A2 receptors to inhibit and activate the enzyme, respectively, has also been used for some time to identify and differentiate adenosine receptors (2). An alternative approach to the classification of adenosine receptors is based on the premise that A1 receptors produce more potent responses to purine-derived analogues with substitutions at the N6 position, such as (R)-N6-(2-phenylisopropyl)adenosine (R-PIA), than to carboxamide analogues such as 5'-(N-ethylcarboxyamido) adenosine (3).

The involvement of guanine nucleotide-binding proteins (GNBP) in the signal transduction following activation of A1 receptors has expanded the list of probable receptor subtypes. The classification of these subtypes has been based on their blockade by pertussis toxin (PTX) in neurons (4). From this viewpoint, activation of A1A, A1C and A1B receptors may regulate Ca2+ influx, K+ efflux and the inhibition of adenyl cyclase, respectively. Moreover, all of these mechanisms may depend on the activation of membrane GNBP, since coupling between the A1 receptor and G0 and Gi proteins has been characterized using R-PIA as the binding agonist (5). Outflow of Ca2+ from the cell may also be mediated by GNBP following the activation of A1 receptors, a mechanism that may use Ca2+/Na+ exchange and depend on a PTX-sensitive pathway (6). On the other hand, purinergic activation mechanisms have been related to inositol triphosphate synthesis because studies of isolated sympathetic ganglion have demonstrated that endogenously released adenosine can inhibit postsynaptic stimulation and turnover of myo-inositol (7).

Activation of A1 receptors can also induce responses mediated by mechanisms based on the activation of different membrane GNBP. Some studies have shown that R-PIA inhibits the release of prolactin (PRL) induced by thyrotropin releasing hormone by blocking the synthesis of phosphatidylinositol and cAMP in GH3 cell lines (8). A probable mechanism of autocrine regulation of PRL secretion by adenosine has been suggested on the basis of results obtained with GH4C1 cell lines (9). Both adenosine and adenosine deaminase were released into the medium, indicating that the levels of adenosine released may be controlled within narrow limits.

In the present study, we analyzed possible mechanisms of signal transduction involving the participation of Gs or Gi proteins, and the synthesis of membrane phosphoinositide and cAMP triggered by A1 receptor activation in preparations of isolated hemipituitaries.


Material and Methods

We used male Wistar rats weighing 200 to 220 g from the central Animal House of the Ribeirão Preto, Faculty of Medicine (University of São Paulo). The animals were kept in collective cages in an artificially controlled environment with a temperature of 22-24ºC and 14 h of light (7:00-21:00). Animals had free access to food and tap water.

Drugs and solutions

Phospholipase C, dibutyryl cyclic AMP (dcAMP), PTX, cholera toxin (CTX), bovine serum albumin, and HEPES were obtained from Sigma (St. Louis, MO, USA). The other substances - R-PIA and cyclopentyltheophylline (CPT) - were provided by Research Biochemicals Incorporated (Natick, MA, USA). The nutrient solution consisted of Earle's salt solution containing 0.1% bovine serum albumin and 15 mM HEPES at pH 7.4.

Experimental procedures

In all experiments, the animals were brought into the laboratory for a period of adaptation of approximately 1 h in order to eliminate possible interference due to stress prior to decapitation at 10:00 am. The brain was then removed, the neural lobe discarded and the anterior pituitary dissected in situ, divided longitudinally into two approximately equal parts, and immersed in nutrient solution (4ºC). Each hemipituitary was transferred in a random sequence to an individual cuvette (assayed with five replicates per group) containing 1 mL nutrient solution (37ºC) and then incubated for 60 min in a Dubnoff metabolic shaker with constant shaking (50 cycles per min, 95% O2/5% CO2) for washing and for the stabilization of basal hormone secretion levels. All tests were conducted immediately after the preincubation period except for the PTX and CTX experiments. All concentrations of the tested drugs used were based on their effects in similar experimental situations. In the experiment in which PTX and CTX were added, the preliminary incubation lasted 180 min, during which the nutrient solution was replaced at 60-min intervals. After incubation, samples were placed in plastic tubes at -20ºC and the hemipituitaries were weighed on a torsion scale. The recorded concentration of PRL released into the nutrient solution was divided by the weight (mg) of the respective hemipituitary and is reported as ng/mg tissue weight. At the end of each experiment, 56 mM KCl was added to evaluate the functional viability of cells in the preparation on the basis of the release of PRL from intracellular stores. A significant increase in basal PRL secretion was recorded (basal = 270 ± 30 vs KCl = 1350 ± 52 ng/mL, P < 0.001) demonstrating that the cells maintained their secretory response for more than 195 min of incubation, thus guaranteeing the viability of the preparation.

Radioimmunoassay

PRL concentrations in the nutrient solution were determined by double-antibody radioimmunoassay (10). The hormones for radioiodination and specific antibodies were obtained from the National Institute of Arthritis, Diabetes and Digestive Diseases (NIDDK, Baltimore, MD, USA) Rat Pituitary Hormone Program. All the samples from one experimental group (control versus respective groups) were analyzed in a single assay. The intra- and inter-assay variations for the PRL assay were, 2.57 and 17.08%, respectively.

Statistical analysis

Data are reported as means ± SEM. The significance of differences in sequential changes in hormone levels between groups was determined by one-way ANOVA followed by the Bonferroni (for comparisons between groups with one treatment only) or Newman-Keuls (for groups with two or more treatments) post hoc test. The level of significance was set at P < 0.05.


Results

Effects of R-PIA on basal PRL secretion

As shown in Figure 1, R-PIA induced a significant reduction in PRL release at all concentrations used (0.001, 0.01, 0.1, 1, and 10 µM) in comparison with those of the basal group. The dose-effect relationship was linear up to 0.1 µM, which induced a decrease of approximately 80% in basal PRL levels. At subsequent doses (1 and 10 µM), there was a reduction in inhibitory power, pointing to a biphasic pattern in the secretory response.

Increasing doses of CPT (0.01, 0.1, and 1 µM), a specific A1 receptor antagonist, were administered 30 min before incubation with R-PIA to determine which dose would block the response to the latter. Under equimolar conditions (0.01 µM), the antagonist caused only a partial blockade of the inhibitory effects of R-PIA on PRL secretion. Total blockade of the effects of the agonist on PRL secretion was only observed at a dose 100 times higher - 1 µM. Administration of the antagonist alone had no effect on basal PRL levels at any dose used (Figure 2).

Effects of R-PIA on PRL secretion stimulated by dibutyryl cAMP

A cAMP analogue, dcAMP, that mimics the intracellular effects of cAMP, was used in an attempt to confirm that the inhibition of PRL secretion induced by the activation of A1 receptors was independent of the increase in cellular cAMP levels. Administration of dcAMP (1 mM) produced a sharp increase (1.5 times) in basal PRL levels. The administration of increasing doses of R-PIA up to 0.1 µM inhibited the PRL secretion induced by dcAMP. At the highest dose (1 µM), the expected increase in the response to the concomitant addition of dcAMP was blocked completely (Figure 3).

Effects of R-PIA on PRL secretion stimulated by phospholipase C

We used phospholipase C to determine whether the inhibitory effect of R-PIA on PRL secretion was altered by activation of the synthesis of inositol triphosphate and diacylglycerol promoted by this enzyme. The addition of phospholipase C (0.5 IU/mL) to the incubation medium induced a substantial increase in basal PRL levels. A dose-dependent decrease in phospholipase C-stimulated PRL secretion was observed after combination of the compound with increasing doses of R-PIA (Figure 4). At the highest dose (1 µM), R-PIA prevented any increase in PRL levels expected in response to the action of the enzyme.

Effects of R-PIA administration on PRL secretion stimulated by pertussis toxin and cholera toxin

The irreversible blockade of the cycle of Gi and Gs protein activation induced by PTX and CTX, respectively, was tested to determine whether the inhibitory effect of R-PIA on PRL secretion in response to A1 activation was dependent on the integrity of this pathway. A significant increase in the medium concentration of PRL was observed 180 min after administration of 0.5 mg/mL CTX or PTX (Figure 5). R-PIA (0.01 µM) inhibited the secretion of PRL stimulated by CTX, but had no effect on the secretory response of this hormone to PTX administration (Figure 5).


Figure 1. Effect of the activation of A1 receptors by different concentrations of (R)-N6-(2- phenyliso-propyl)adenosine (R-PIA) on basal prolactin (PRL) secretion. Data are reported as means ± SEM ng PRL/mg tissue weight (t.w.) for 5 replicates. *P < 0.01 compared to the untreated (0) group; +P < 0.05 and ++P < 0.001 for the comparisons indicated in the figure (ANOVA followed by the Bonferroni multiple comparison test).

[View larger version of this image (29 K JPG file)]


Figure 2. Effect of prior (30 min) administration of cyclopentyltheophylline (CPT) associated or not with 0.01 µM (R)-N6-(2-phenylisopropyl)adenosine (R-PIA) on prolactin (PRL) secretion. Data are reported as means ± SEM ng PRL/mg tissue weight (t.w.) for 5 replicates. *P < 0.05 compared to the untreated group; +P < 0.05, ++P < 0.01 and +++P < 0.001 compared to CPT-treated groups (ANOVA followed by the Newman-Keuls multiple comparison test).

[View larger version of this image (32 K JPG file)]


Figure 3. Effect of the administration of (R)-N6-(2-phenylisopropyl)adenosine (R-PIA) associated or not with 1000 µM dibutyryl cAMP (dcAMP) on prolactin (PRL) secretion. Data are reported as means ± SEM ng PRL/mg tissue weight (t.w.) for 5 replicates. *P < 0.001 compared to untreated group. +P < 0.001 (dcAMP vs dcAMP + R-PIA group). Newman-Keuls multiple comparison test was used after ANOVA.

[View larger version of this image (29 K JPG file)]


Figure 4. Effect of the administration of (R)-N6-(2-phenyliso-propyl)adenosine (R-PIA) associated or not with 0.5 IU/mL phospholipase C (PLC) on prolactin (PRL) secretion. Data are reported as means ± SEM ng PRL/mg tissue weight (t.w.) for 5 replicates. *P < 0.001 compared to untreated group. +P < 0.01 and ++P < 0.001 (PLC vs PLC + R-PIA group) (ANOVA followed by the Newman-Keuls multiple comparison test).

[View larger version of this image (26 K JPG file)]


Figure 5. Effect of the administration of 0.01 µM (R)-N6-(2-phenylisopropyl)adenosine (R-PIA) on the prolactin (PRL) secretion induced by 0.3 µg/mL pertussis toxin (PTX) or 6.0 µg/mL cholera toxin (CTX) (pre-incubated for 180 min). Data are reported as means ± SEM ng PRL/mg tissue weight (t.w.) for 5 replicates. *P < 0.01 and **P < 0.001 compared to the untreated group; +P < 0.001 compared to the treated group (ANOVA followed by the Newman-Keuls multiple comparison test).

[View larger version of this image (31 K JPG file)]


Discussion

We demonstrated the in vitro inhibition of PRL secretion by the anterior pituitary gland in response to graded doses of R-PIA. The doses of 1 and 10 µM caused a decrease that was lower than the maximum inhibition produced by the 0.1-µM dose (Figure 2), indicating a biphasic pattern of response (U-shaped dose-response curve). These results compare well with our earlier report of biphasic purinergic effects on PRL secretion stimulated by the adenosine agonist (A2 > A1) 5-(N-methylcarboxyamido)adenosine (MECA) (11). We showed that the 10 µM dose had only a residual stimulatory effect on PRL secretion in hemipituitaries incubated in vitro compared with 1 µM (maximum stimulatory response). This suggests the co-existence of A2 and A1 adenosine receptors in the PRL-secreting cells.

This biphasic pattern was confirmed in the present study by the addition of R-PIA, an A1 > A2 agonist, given that the lower concentrations brought about a dose-related reduction in PRL secretion, whereas higher concentrations were relatively less potent. Other investigators have shown that R-PIA influences basal adenyl cyclase activity, exhibiting inhibitory effects at low concentrations, but stimulatory effects at higher ones. A similar tendency was found in PRL secretion, since low concentrations of R-PIA reduced both basal and thyrotropin releasing hormone-stimulated PRL release, whereas higher concentrations restored PRL release in perifused pituitaries (12).

The use of CPT, a specific A1 receptor antagonist, proved to be effective in blocking the R-PIA-induced inhibition of PRL secretion only at a 100-times higher MECA dose. This antagonist alone did not produce any variation in basal PRL secretion at any of the concentrations used. These data indicate that low R-PIA concentrations have specific effects that may be mediated by the activation of A1 adenosine receptors. In addition, the fact that CPT caused no change in basal PRL levels suggests that the adenosine released at this time did not produce perceptible autocrine regulatory effects in this type of preparation (Figure 2).

In a similar study, we showed that caffeine, another xanthinic antagonist, had no effect on basal PRL secretion in the hemipituitary in vitro, but efficiently blocked adenosine-induced secretion (13). This suggests that adenosine has a regulatory action on PRL release via the A2 receptor, but not under basal conditions. This evidence is reinforced by the fact that two selective A1 adenosine receptor antagonists and a highly specific A2 receptor antagonist had no effect on basal PRL release (14).

Indeed, studies of GH4C1 cell line cultures have shown the release of adenosine accompanied by the enzyme adenosine deaminase under basal experimental conditions (9). The addition of this enzyme alone increased PRL secretion, suggesting that the adenosine released may have a tonic-inhibitory autocrine action (9,15). These findings support the hypothesis that the activation of A1 adenosine receptors may result in a reduction of PRL secretion, and that this effect may depend on the levels of adenosine deaminase in the medium.

The increase in PRL secretion observed after the addition of both dcAMP (Figure 3) and phospholipase C (Figure 4) supports the evidence from other studies which have demonstrated the contribution of cAMP and membrane phosphoinositides to the regulation of basal or induced PRL secretion (16,17). The R-PIA inhibition of both the dcAMP- and phospholipase C-induced PRL secretion suggests that the A1 receptors may inhibit PRL secretion through a mechanism independent of the corresponding second messengers (i.e., cAMP and Ca2+ mobilization).

Our data suggest that the inhibition induced by activation of A1 receptors depends on the activation of Gi protein, but not Gs protein, given that R-PIA inhibited PRL secretion stimulated by CTX, but had no effect under stimulation by PTX (Figure 5). Several studies have shown a similar role of PTX in preventing the dopamine-induced inhibition of PRL secretion in adenohypophyseal cell cultures (18). An alternative interpretation would be the blockade of Ca2+ influx or an increase in Ca2+ efflux from the cell mediated at the level of membrane GNBP. Some studies have shown that activation of A1 receptors decreased basal or protein kinase C-stimulated intracellular Ca2+ release in a GH3B6 cell line (19), suggesting that this inhibition may depend on the triggering of a mechanism that precedes activation of this kinase. Ca2+ outflow from the cell may also be mediated by GNBP after A1 receptor activation, a mechanism that would involve exchange with Na+ and be dependent on a pathway sensitive to PTX (6).

Our results show that the activation of A1 receptors by R-PIA inhibits pituitary PRL secretion in a dose-dependent manner through a mechanism that is dependent on the blockade of Gi protein. By contrast, this mechanism is not dependent on the blockade of the secretory effect via GNBP caused by either cAMP or membrane phosphoinositides.


References

1. Sattin A, Rall TW. The effect of adenosine and adenine nucleotides on the cyclic adenosine 3',5'-phosphate content of guinea pig cerebral cortex slices. Mol Pharmacol 1970; 6: 13-23.         [ Links ]

2. Burnstock G, Brown CM. An introduction to purinergic receptors. In: Burnstock G (Editor), Purinergic receptors. New York: Chapman and Hall; 1981.         [ Links ]

3. Stone TW. Summary of a Symposium: Discussion on Purine Receptors Nomenclature. In: Stone TW (Editor), Pharmacology and physiological roles. Deerfield Beach: VHC Publishers; 1985.         [ Links ]

4. Fredholm BB, Dunwiddie TV. How does adenosine inhibit transmitter release? Trends Pharmacol Sci 1988; 9: 130-134.         [ Links ]

5. Freissmuth M, Selzer E, Schutz W. Interactions of purified bovine brain A1-adenosine receptors with G-proteins. Reciprocal modulation of agonist and antagonist binding. Biochem J 1991; 275 (Pt 3): 651-656.         [ Links ]

6. Brechler V, Pavoine C, Lotersztajn S, Garbarz E, Pecker F. Activation of Na+/Ca2+ exchange by adenosine in ewe heart sarcolemma is mediated by a pertussis toxin-sensitive G protein. J Biol Chem 1990; 265: 16851-16855.         [ Links ]

7. Rubio R, Bencherif M, Berne RM. Inositol phospholipid metabolism during and following synaptic activation: role of adenosine. J Neurochem 1989; 52: 797-806.         [ Links ]

8. Delahunty TM, Cronin MJ, Linden J. Regulation of GH3-cell function via adenosine A1 receptors. Inhibition of prolactin release, cyclic AMP production and inositol phosphate generation. Biochem J 1988; 255: 69-77.         [ Links ]

9. Dorflinger LJ, Schonbrunn A. Adenosine inhibits prolactin and growth hormone secretion in a clonal pituitary cell line. Endocrinology 1985; 117: 2330-2338.         [ Links ]

10. Niswender GD, Chen CL, Midgley AR Jr, Meites J, Ellis S. Radioimmunoassay for rat prolactin. Proc Soc Exp Biol Med 1969; 130: 793-797.         [ Links ]

11. Picanço-Diniz DLW, Valença MM, Favaretto ALV, Antunes-Rodrigues J. Stimulatory effects of adenosine on prolactin secretion in the pituitary gland of the rat. Braz J Med Biol Res 2002; 35: 855-860.         [ Links ]

12. Schettini G, Landolfi E, Meucci O, Florio T, Grimaldi M, Ventra C, et al. Adenosine and its analogue (-)-N6-R-phenyl-isopropyladenosine modulate anterior pituitary adenylate cyclase activity and prolactin secretion in the rat. J Mol Endocrinol 1990; 5: 69-76.         [ Links ]

13. Picanço-Diniz DLW, López-Jiménez MA, Valença MM, Favaretto ALV, Antunes-Rodrigues J. Effect of adenosine on gonadotropin and prolactin secretion by hemipituitaries in vitro. Braz J Med Biol Res 1989; 22: 783-785.         [ Links ]

14. Yu WH, Kimura M, Walczewska A, Porter JC, McCann SM. Adenosine acts by A1 receptors to stimulate release of prolactin from anterior-pituitaries in vitro. Proc Natl Acad Sci U S A 1998; 95: 7795-7798.         [ Links ]

15. Kumari M, Buckingham JC, Poyser RH, Cover PO. Roles for adenosine A1- and A2-receptors in the control of thyrotrophin and prolactin release from the anterior pituitary gland. Regul Pept 1999; 79: 41-46.         [ Links ]

16. Dannies PS, Gautvik KM, Tashjian AH Jr. A possible role of cyclic AMP in mediating the effects of thyrotropin-releasing hormone on prolactin release and on prolactin and growth hormone synthesis in pituitary cells in culture. Endocrinology 1976; 98: 1147-1159.         [ Links ]

17. Enjalbert A, Sladeczek F, Guillon G, Bertrand P, Shu C, Epelbaum J, et al. Angiotensin II and dopamine modulate both cAMP and inositol phosphate productions in anterior pituitary cells. Involvement in prolactin secretion. J Biol Chem 1986; 261: 4071-4075.         [ Links ]

18. Senogles SE, Benovic JL, Amlaiky N, Unson C, Milligan G, Vinitsky R, et al. The D2-dopamine receptor of anterior pituitary is functionally associated with a pertussis toxin-sensitive guanine nucleotide binding protein. J Biol Chem 1987; 262: 4860-4867.         [ Links ]

19. Mollard P, Guerineau N, Chiavaroli C, Schlegel W, Cooper DM. Adenosine A1 receptor-induced inhibition of Ca2+ transients linked to action potentials in clonal pituitary cells. Eur J Pharmacol 1991; 206: 271-277.         [ Links ]

Acknowledgments

We thank Rubens Fernando de Melo, Leonardo Fidelis Filho, Gilberto Lopes, Manoel Corrêa de Lima, Marina Holanda, and Maria Valci Aparecida dos Santos Silva for technical assistance. We also thank Gareth Cuttle (Department of Physiological Sciences, CCB, Federal University of Santa Catarina) for the English revision.


Correspondence and Footnotes

Address for correspondence: J. Antunes-Rodrigues, Departamento de Fisiologia, FMRP, USP, Avenida Bandeirantes, 3900, 14049-900 Ribeirão Preto, SP, Brasil. Fax: +55-16-3602-0017. E-mail: antunes@fmrp.usp.br

Research supported by FAPESP (Nos. 91/0567-0 and 94/3805-7) and CNPq (Nos. 50167/91-7 and 521593/94-8). Received November 3, 2005. Accepted June 21, 2006.

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License