Purinergic P2X7 receptors regulate secretion of interleukin-1 receptor antagonist and beta cell function and survival
- 1k Downloads
In obesity, beta cells activate compensatory mechanisms to adapt to the higher insulin demand. Interleukin-1 receptor antagonist (IL-1Ra) prevents obesity-induced hyperglycaemia and is a potent target for the treatment of diabetes, but the mechanisms of its secretion and regulation in obesity are unknown. In the present study, we hypothesise the regulation of IL-1Ra secretion by purinergic P2X7 receptors in islets.
Production and regulation of P2X7 were studied in pancreatic sections from lean and obese diabetic patients, non-diabetic controls and in isolated islets. IL-1Ra, IL-1β and insulin secretion, glucose tolerance and beta cell mass were studied in P2x7 (also known as P2Rx7)-knockout mice.
P2X7 levels were elevated in beta cells of obese patients, but downregulated in patients with type 2 diabetes mellitus. Elevated glucose and non-esterified fatty acids rapidly activated P2X7 and IL-1Ra secretion in human islets, and this was inhibited by P2X7 blockade. In line with our results in vitro, P2x7-knockout mice had a lower capacity to secrete IL-1Ra. They exhibited severe and rapid hyperglycaemia, glucose intolerance and impaired beta cell function in response to a high-fat/high-sucrose diet, were unable to compensate by increasing their beta cell mass in response to the diet and showed increased beta cell apoptosis.
Our study shows a tight correlation of P2X7 activation, IL-1Ra secretion and regulation of beta cell mass and function. The increase in P2X7 production is one mechanism that may explain how beta cells compensate by adapting to the higher insulin demand. Disturbances within that system may result in the progression of diabetes.
KeywordsDiabetes Interleukin-1 receptor antagonist Islets Obesity Purinergic receptor P2X7
Glucose-stimulated insulin secretion
Interleukin-1 receptor antagonist
Intraperitoneal glucose tolerance test
Intraperitoneal insulin tolerance test
An alarming increase in obesity and type 2 diabetes mellitus has been recorded among the global population. The rise in type 2 diabetes is tightly connected with a dramatic increase in obesity, a major risk factor for the development of type 2 diabetes . Most obese people have become insulin resistant but compensate for the enhanced insulin demand by increasing insulin secretion . In order to explore how beta cells increase mass under physiological conditions, it is important to stimulate beta cell expansion in vivo and in vitro. The pathology of type 2 diabetes comprises initial hyperinsulinaemia with, thereafter, a continuous decline in beta cell function in association with a progressive decrease in beta cell mass due to an increasing rate of beta cell apoptosis . A previously described mechanism for glucose-induced beta cell apoptosis involves the production and secretion of interleukin-1β (IL-1β) by the beta cell itself , which is blocked by an anti-inflammatory cytokine and naturally occurring antagonist of interleukin-1α (IL-1α) and β [5, 6, 7], the interleukin-1 receptor antagonist (IL-1Ra) . Similarly to IL-1β, IL-1Ra binds to type 1 and 2 IL-1 receptors (IL-1R1, IL-1R2), but lacks a second binding domain. IL-1Ra does not recruit the IL-1 receptor accessory protein, the second component of the receptor complex, and blocks IL-1R1 signalling. Endogenous production and secretion of IL-1Ra has been shown to limit inflammation and tissue damage.
Exogenous IL-1Ra protects against IL-1β-induced beta cell damage [9, 10], counteracts low-dose-streptozotocin-induced diabetes  and autoimmune diabetes . Results from a clinical study show that IL-1Ra improves glycaemic control and beta cell function in patients with type 2 diabetes .
The secretory mechanisms and those regulating changes of IL-1Ra are unknown. An increasing body of evidence suggests that IL-1β, as well as intracellular IL-1Ra, is secreted by a leaderless pathway via activation of the P2X7 receptor (P2X7; ). P2X7, a member of the purinergic receptor family, is a G-protein-coupled ionotropic receptor for extracellular ATP that forms a pore in response to ligand stimulation, regulating cell permeability, cytokine release and apoptosis . The cytokine IL-1β, which lacks a leader signal sequence, does not follow the classic endoplasmic-reticulum-to-Golgi pathway of secretion. Its secretion is governed by P2X7-mediated pathways [16, 17]. This raises a question about the mechanisms of its release; it has been demonstrated that it is secreted upon activation of P2X7 receptors, leading to shedding of microvesicles that contain the cytokine .
The pore-forming ATP receptor P2X7 is also responsible for IL-1Ra release in endothelial cells and macrophages . Consequently, we hypothesised that P2X7 could be involved in IL-1Ra release in beta cells. The work presented here evaluates an important role for P2X7 in beta cells, not only for cytokine release but also for insulin release and apoptosis. We have found upregulation of P2X7 in islets from obese individuals, while receptors were almost undetectable in type 2 diabetes. We then investigated whether P2X7 is involved in the compensatory mechanisms of the beta cells in response to an increased insulin demand in insulin resistance.
P2x7 (also known as P2rx7)-knockout mice (B6.129P2-P2rx7 tm1Gab /J; P2X7-KO) backcrossed to C57BL/6 mice for ten generations and C57BL/6 J wild-type (WT) mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA) at 4 weeks of age. Beginning at 5 weeks of age and continuing for 16 weeks, knockout as well as wild-type animals were fed a normal diet (ND; P2X7-KO-ND and WT-ND, respectively; normal diet was Harlan Teklad Rodent Diet 8604, containing 12.2, 57.6 and 30.2% energy from fat, carbohydrate and protein, respectively) or a high-fat/high-sucrose diet (HFD; P2X7-KO-HFD and WT-HFD; high-fat/high-sucrose diet was Surwit diet, Research Diets, New Brunswick, NJ, USA containing 58, 26 and 16% energy from fat, carbohydrate and protein, respectively ). Further details are described in the Electronic supplementary material (ESM).
Intraperitoneal glucose and insulin tolerance tests
At the beginning of the experiment and after 4, 8, 12 and 16 weeks of diet, all animals underwent in vivo studies. For intraperitoneal glucose tolerance tests (IPGTTs), glucose at a dose of 1 g/kg body weight and for intraperitoneal insulin tolerance tests (IPITTs), recombinant human insulin (Novolin; Novo Nordisk, Bagsværd, Denmark) at a dose of 0.5 U/kg was injected intraperitoneally as described before .
Islet isolation and culture
After 12 and 16 weeks of diet and treatment, mice were killed, blood was taken by cardiac puncture and serum stored at −80°C until further analysis. Thereafter, islets from all groups were isolated as described previously . Human islets were isolated from pancreases of six healthy organ donors as described previously . Mouse islets were cultured in RPMI-1640 medium (11.1 mmol/l glucose), and human islets in CMRL-1066 medium (5.5 mmol/l glucose), as described previously .
Functional analysis of P2X7 in human islets
Freshly isolated human islets were grown on dishes coated with extracellular matrix until a monolayer formed. Cells were loaded with 10 μmol/l YO-PRO-1 (0.375 kDa; Molecular Probes/Invitrogen, Carlsbad, CA, USA) in PBS after a protocol by Narcisse et al. . Further details are described in the ESM.
Glucose-stimulated insulin secretion
Islets used to perform glucose-stimulated insulin secretion (GSIS) experiments were kept in culture medium for 24 h and GSIS was performed as described previously . Islets were extracted with 0.18 mol/l HCl in 70% (vol./vol.) ethanol for determination of insulin content. Islet insulin was determined using mouse or human insulin ELISA (ALPCO, Salem, NH, USA).
IL-1Ra and IL-1β analysis
Serum obtained from retro-orbital bleedings and cell culture supernatant fractions from mouse isolated islets were evaluated using mouse IL-1Ra or IL-1β Quantikine ELISA (R&D Systems, Minneapolis, MN, USA) (see ESM).
Beta cell mass and histochemical analyses
After 12 weeks of diet and treatment, pancreases were weighed and fixed overnight in 4% (vol./vol.) paraformaldehyde at 4°C under continuous shaking followed by paraffin embedding, orienting pancreases such that sections were cut along the head–tail axis. Beta cell mass and apoptosis were analysed as described previously  (P2X7 staining is described in the ESM). Human pancreatic tissue was obtained at autopsy from non-diabetic individuals and from people with type 2 diabetes (all with documented fasting plasma glucose >8.048 mmol/l). ESM Table 1 summarises age, BMI, sex, glucose levels and diagnosis.
RNA extraction and PCR analysis
The total RNA of isolated islets was extracted after overnight culture as described previously . The primers were: P2X7 5′AGATCGTGGAGAATGGAGTG3/5′TCTCGTGGTGTAGTTGTGG3′; tubulin 5′TCCACGTTGGCCAGGCTGGTGTCCAG3′/5′CTGTGATGAGCTGCTCAGGGTGG3′.
Western blot analysis
Islet lysates were prepared from 100 islets per treatment condition as described previously . PVDF filters (Amersham Biosciences/GE Healthcare, Uppsala, Sweden) were incubated with rabbit anti-P2X7 (Almone Labs, Jerusalem, Israel), rabbit-anti-IL-1Ra (R&D Systems) and anti-actin (#4967; Cell Signaling Technology, Beverly, MA, USA), followed by incubation with horseradish-peroxidase-linked anti-rabbit IgG.
Samples were evaluated in a randomised manner by two investigators (R. Glas and K. Maedler) who were blinded to the treatment conditions. Data are presented as means ± SE and were analysed by paired Student’s t test or by analysis of variance with a Bonferroni correction for multiple group comparisons.
P2X7 is produced in beta cells, upregulated in obesity and decreased in diabetes
P2X7 is upregulated by short-term exposure to elevated glucose and NEFA in isolated human islets
As we have observed changes in P2X7 levels in obesity and diabetes, we tested whether P2X7 is activated by a diabetic milieu in culture. Islets were exposed to increased glucose concentrations (5.5–33.3 mmol/l, Fig. 2b,c) or 0.5 mmol/l palmitic acid (Fig. 2d) for 10 min after YO-PRO-1 loading. Elevated glucose concentration activated P2X7, as did palmitic acid; we observed a nine-fold activation by 33.3 mmol/l glucose, six-fold activation by 11.1 mmol/l glucose and a nine-fold activation by palmitic acid within 5 to 10 min of exposure (p < 0.001). Peak levels slightly decreased after 15 min exposure, but remained elevated compared with control conditions throughout the 1 h incubation period, similar to the finding with BzATP stimulation. In all assays, co-culture with the P2X7 inhibitors KN-62 (Fig. 2a–d) and apyrase (Fig. 2a, b, d), reduced the P2X7 activation induced by BzATP, elevated glucose or palmitic acid.
As BzATP evokes P2X7 effects very briefly, even within 1 to 30 s , we wanted to see if there is a change in P2X7 level in response to the activation. Elevated glucose concentrations increased P2X7 mRNA levels (Fig. 2e: 1.7-fold increase with 11.1 mmol/l glucose, 2.5-fold increase with 33.3 mmol/l glucose and 1.8-fold increase with palmitic acid, compared with control [untreated, with 5.5 mmol/l glucose], p < 0.05) after 10 min of exposure. This was confirmed by western blot analysis (Fig. 2f, g). Changes in mRNA level and protein production were detected within this short time, whereas values returned to basal after 24 h of exposure (not shown). Previous reports show that IL-1β induces increases in P2X7 at 6 h, with levels decreasing after 48 h of treatment . One explanation for the increase in P2X7 within this narrow timeframe would be a rapid de novo synthesis of P2X7 mRNA and P2X7 protein, given weight by its relatively small size (595 amino acids/68 kDa).
Recently, we have shown that IL-1Ra can protect from impaired beta cell function and death in vivo and in vitro [4, 8]. The mechanisms of its secretion from the beta cell are unknown. We tested whether activation of P2X7 leads to IL-1Ra secretion. Long-term glucose exposure had no effect on IL-1Ra secretion (data not shown and [28, 29]). But short-term (30 min) exposure of islets to elevated glucose, palmitic acid or BzATP induced IL-1Ra secretion, which was inhibited by KN-62 (Fig. 2h). Analysis of insulin secretion in the same supernatant fractions showed similar regulation of insulin secretion. Elevated glucose levels (11.1 and 33.3 mmol/l), 0.5 mmol/l palmitic acid and 100 μmol/l BzATP induced insulin secretion, which was prevented by co-culture with KN-62 (Fig. 2i).
Regulation of IL-1Ra production by glucose and palmitic acid was confirmed by western blot analysis. Short-term exposure of human islets to elevated glucose and palmitic acid (30 min–8 h) induced IL-1Ra secretion, whereas no differences were observed at 24 h treatment. These data indicate a parallel induction of IL-1Ra and P2X7 in pancreatic islets.
P2X7-deficient mice have impaired glucose tolerance, beta cell mass and function and fail to adapt IL-1Ra secretion to a high-fat diet
After 12 weeks of the diet, high-fat feeding increased glucose levels and impaired glucose tolerance. The highest glucose levels were found in the P2X7-KO mice fed an HFD, with differences significant at all time points during an IPGTT compared with all other groups (WT-ND, WT-HFD, P2X7-KO-ND; Fig. 3b, p < 0.01). Glucose tolerance further deteriorated after 16 weeks of HFD diet (not shown).
Figure 3c shows serum insulin levels during an IPGTT after 12 weeks of diet and treatment. As previously described , high-fat feeding resulted in hyperglycaemia (Fig. 3b) as well as hyperinsulinaemia (Fig. 3c, p < 0.001) compared with control diet mice. Compared with mice consuming an ND, fasted insulin levels were 3.4-fold higher in HFD-fed WT mice and 5.7-fold higher in the P2X7-KO mice (p < 0.01). The HFD-fed WT mice had a much lower capacity to increase their glucose-stimulated insulin release (Fig. 3d, p < 0.01), whereas the P2X7-KO mice on the HFD completely failed to further increase their insulin levels in response to a glucose challenge (Fig. 3d, p < 0.001). P2X7-KO-ND mice had a 1.3-fold lower stimulatory index than the WT control mice (p < 0.05). These data indicate that P2X7-KO mice have an impaired beta cell function, which deteriorates under the challenge of an HFD treatment.
To determine insulin sensitivity, we performed insulin tolerance tests by measuring glucose concentrations after intraperitoneal insulin injection of 0.5 U/kg body weight. Glucose levels from all mice were significantly higher at all time points during the IPITT compared with control wild-type mice on an ND (unchanged in P2X7-KO-ND at 120 min, Fig. 3e). To investigate whether the impaired glucose tolerance and insulin secretion was due to a defect in beta cell survival together with beta cell function, we analysed beta cell mass from the four groups after 12 and 16 weeks of treatment. As reported previously , after 12 weeks of diet, islet beta cell mass was increased by high-fat feeding in the WT mice compared with those on a normal diet (1.6-fold increase, p < 0.01; Fig. 3f,g). In contrast, P2X7-KO mice had no such compensatory capacity in response to the HFD. There was no significant change in beta cell mass in the P2X7-KO mice fed the HFD compared with the P2X7-KO-ND mice. Moreover, beta cell mass was significantly decreased compared with the wild-type mice, independent of the diet (1.5-fold with ND and 1.5-fold with HFD, p < 0.0.01, Fig. 3f, g). This was confirmed after 16 weeks of diet. We have previously reported the tendency to beta cell apoptosis in mice after 12 weeks of HFD despite an increased beta cell mass at that time point . Here, we analysed beta cell mass after 12 and 16 weeks of diet. In contrast to 12 weeks, there was no increase in beta cell mass in wild-type mice fed an HFD compared with ND-fed animals after 16 weeks. To investigate whether the changes in beta cell mass correlate with the increased rate of apoptosis, we analysed TUNEL-positive beta cells in pancreatic sections. After 16 weeks of diet, we observed a clear increase in beta cell apoptosis in the control HFD group. This was further increased in the P2X7-KO-HFD mice (1.5-fold raise in apoptosis in the WT-HFD and 11.2-fold in the P2X7-KO-HFD compared with ND controls, p < 0.01, Fig. 3h). As P2X7 regulates IL-1Ra secretion, we tested whether IL-1Ra levels are altered in P2X7-KO mice. The serum concentration of IL-1Ra was measured in WT and P2X7-KO mice during the 16 weeks of HFD feeding. In WT-ND mice, IL-1Ra levels were unchanged during the experiment. IL-1Ra levels were significantly lower in P2X7-KO-ND mice at 8, 12 and 16 weeks. Interestingly, after 4 weeks of diet, IL-1Ra increased 2.2-fold in the WT-HFD mice; these high levels were maintained at 8 weeks (1.9-fold compared with WT) and time-dependently decreased again after 12 and 16 weeks of diet (1.5-fold and 2.4-fold decrease from 8 weeks, respectively). After 16 weeks, IL-1Ra levels with HFD were lower than with the ND (2.6-fold decrease), which correlates with the changes in beta cell mass. P2X7-KO mice, which had beta cells unable to compensate via increasing beta cell mass, had no adaptive increase in IL-1Ra levels (Fig. 3i). We have also measured IL-1β serum levels together with IL-1Ra in the mouse studies. Unfortunately, the levels were very low and mostly under the lower level of the ELISA kit. Positive IL-1β readings were found for some but not all samples from the HFD-fed control mice at 12 and 16 weeks of diet, but none were found in the ND or P2X7-KO mice, making statistical data analysis not possible.
Isolated islets from P2X7-deficient mice display a defect in insulin, IL-1Ra and IL-1β secretion
The present study shows that P2X7 is produced in human islets in the beta cell and is regulated during the progression of type 2 diabetes. In human obesity, P2X7 is produced in large amounts, but is downregulated in pancreases obtained from patients with type 2 diabetes, in lean as well as in obese individuals. These data suggest that P2X7 plays an important role in beta cell compensation. This is confirmed in our subsequent studies using P2X7-KO mice. These mice show impaired glucose tolerance and glucose-stimulated insulin secretion together with an increased number of apoptotic beta cells, which leads to a decrease in beta cell mass. When P2X7-KO mice were challenged with a diet enriched with fat and sucrose, they became severely glucose intolerant and were completely unable to upregulate insulin secretion in response to glucose during an IPGTT in vivo or a GSIS assay in vitro. Although they had increased basal insulin levels, which indicate an attempt to adapt for the higher insulin demand, they were unable to compensate by increasing beta cell mass during the diet. Beta cell mass adaptation is an important mechanism to maintain normoglycaemia in obese insulin-resistant individuals, in whom beta cell mass is almost doubled compared with lean individuals . Similar increases in beta cell mass in diet-induced obesity were measured in mice using a diet enriched with fat in a previous  or the present study. Interestingly, the high-fat/high-sucrose ‘Surwit diet’ induces an early reduction in beta cell proliferation at a time point when beta cell mass is still increased , indicating a development towards fewer cells. This tendency is substantiated in the present study, where we show a decrease in beta cell mass with time together with an induction of beta cell apoptosis after 16 weeks of diet. In comparison, P2X7-KO mice show a higher apoptosis rate under the ND as well as under the HFD. The increase in basal insulin levels in response to the diet together with a defect in the compensation of beta cell mass indicates that, despite the lower beta cell mass, high amounts of insulin can be secreted. However, the functional failure is apparent in the abolished response to glucose stimulation. A complete loss of P2X7 was also observed in the NOD mouse with manifest diabetes .
Little is known about the purinergic receptor family in the beta cell and the effects on insulin release are controversial. A low concentration of ATP stimulates insulin secretion via the metabotropic purinergic receptor subtype P2Y rather than via P2X receptors (the ionotropic selective cation channel receptors) . P2X7 was found in rat alpha but not beta cells , its production in human islets has not been studied before. Here, we show clearly the production and regulation of P2X7 in the beta cell.
Fibroblasts from donors with type 2 diabetes have increased P2X7-mediated responses compared with non-diabetic donors, but levels of P2X7 were unchanged . In contrast, we report here a decrease in the level of P2X7 in beta cells in type 2 diabetes. Consistent with the results from fibroblasts, P2X7 activation was enhanced in isolated islets by a diabetic milieu (high glucose and palmitic acid), an effect prevented by P2X7 inhibition. High glucose induces caspase 3 activation in beta cells as well as in fibroblasts. It is possible that beta cell P2X7 receptors are activated and produced in high numbers at a time when IL-1Ra is highly secreted (e.g. in obesity) and when diabetes develops, levels are lowered in specific cells.
Our in vitro studies on isolated islets show that P2X7 is directly involved in insulin secretion. Activation of P2X7 by BzATP (which specifically targets P2X7 receptors [27, 35, 36]) or acute stimulation of insulin secretion by glucose or palmitic acid for 1 h resulted in insulin release that was partially inhibited by two P2X7 inhibitors.
Although secretory pathways of insulin and IL-1Ra are distinct, one possible common step could be the calcium exchange; activation of P2X7 increases intracellular calcium levels by calcium influx from the extracellular space. This leads to activation of the leaderless IL-1Ra-releasing pathway through pore formation. Increased intracellular calcium levels would also be a trigger for exocytosis of insulin-containing secretory granules.
High glucose can also activate P2X7 in fibroblasts . In parallel, BzATP, glucose and palmitic acid induced short-term secretion of IL-1Ra, which was also inhibited by KN-62. Apyrase, a less specific P2X7 inhibitor that affects the ligands of all P2X7 receptors, showed similarities to the specific P2X7 inhibitor KN-62 in blocking the effect of BzATP on islets, suggesting that P2X7 is mainly activated by glucose and palmitic acid in islets.
Intracellular IL-1Ra has no signal peptide. A possible mechanism for release of intracellular IL-1Ra from the cells to bind to IL-1R1 is through vesicle shedding, which occurs through activation of the P2X7 by ATP . Our data indicate that IL-1Ra secretion is an ATP-dependent process that requires P2X7 activation. This may be how intracellular IL-1Ra, as well as IL-1β, is released from beta cells. Similar results are obtained from macrophages and endothelial cells, demonstrating that intracellular IL-1Ra and IL-1β are secreted by an extracellular ATP-dependent mechanism involving P2X7 activation . We hypothesised that the balance of IL-1β and IL-1Ra is crucial for the onset of diabetes. P2X7 receptors have a dual role in this system: they mediate the release of both pro-inflammatory IL-1β and anti-inflammatory IL-1Ra. From in vitro experiments we know that an excess of 10- to 1,000-fold IL-1Ra is needed to inhibit the IL-1β effects, dependent on the exposure time [4, 9]. Βeta cell dysfunction and increased apoptosis may occur rapidly when IL-1Ra levels are decreased.
Corresponding with the levels of P2X7, increased IL-1Ra serum levels correlate with obesity and insulin resistance [39, 40, 41, 42], whereas IL-1Ra is decreased in type 2 diabetes . Recent results from the Whitehall II study show that IL-1Ra levels are increased before the onset of type 2 diabetes , which is consistent with our findings in mice fed with an HFD. The increase in IL-1Ra could be an attempt by the body to counteract the deleterious effects of IL-1β and to preserve beta cell survival, insulin secretion and insulin sensitivity. In contrast, it has been speculated that increased IL-1Ra could have an additional metabolic effect that leads to insulin resistance. This is not supported by our results, as daily IL-1Ra injections over 12 weeks did not impair insulin sensitivity in C57Bl/6 mice in a previous study , and increased IL-1Ra levels did not lead to changes in insulin sensitivity in the present study. As shown by the bell-shaped curve in Fig. 3i, decreased IL-1Ra levels occurred together with the increase in beta cell apoptosis and the development of diabetes in our study. It will now be important to determine whether long-standing type 2 diabetes leads to decreased IL-1Ra levels in humans in a follow-up study.
This work was supported by the American Diabetes Association (Junior Faculty Grant 706JF41), the German Research Foundation (DFG, Emmy Noether Program, MA 4172/1-1) and the European Foundation for the Study of Diabetes (EFSD)/Merck Sharp & Dohme. R. Glas is a recipient of the German Research Foundation Fellowship award (Schr 1166/1-1), N. S. Sauter and F. T. Schulthess are recipients of the Swiss National Foundation Fellowship award. Human islets were provided through the Islet Cell Resource Consortium, administered by the Administrative Bioinformatics Coordinating Center (ABCC) and supported by the National Center for Research Resources (NCRR), the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) and the Juvenile Diabetes Research Foundation (JDRF). We thank J. Bergemann (Uni Bremen), H. Gerber, M. Kennedy and L. N. Le (UCLA) for excellent technical assistance, C. Lyon (UCLA) for help with the luminex assays and P. U. Heitz, P. Komminoth, P. Saremaslani (University Hospital Zurich) and the National Disease Research Interchange for providing human pancreatic sections.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.