Keywords

1 Introduction

Caffeine is a nonselective adenosine receptor antagonist, and it is believed that at concentrations achieved during normal human consumption, caffeine exerts its biological effects through antagonism of adenosine receptors (Fredholm et al. 1999). Similar to caffeine, theophylline is a nonselective albeit more potent antagonist of adenosine receptors (Fredholm et al. 2001). Adenosine is an extracellular purine nucleoside signaling molecule, which regulates cell and tissue function both in health and in disease (Hasko et al. 2008). Adenosine is generated following the degradation of its precursor, ATP, a process which can occur both extra- and intracellularly. ATP, a mostly intracellular molecule, is liberated from the cell following stressful and injurious events, and is metabolized to adenosine via a cascade of ectonucleotidases, including CD39 (nucleoside triphosphate diphosphorylase) and CD73 (5′-ectonucleotidase) (Yegutkin 2008). Adenosine that accumulates intracellularly following ATP metabolism is extruded from the cell via nucleoside transporters (Volonte and D’Ambrosi 2009). Cells of the immune/inflammatory system, including neutrophils, mast cells, endothelial cells, lymphocytes, and platelets, have been appreciated as the most prodigious sources of extracellular adenosine (Eltzschig et al. 2008; Hasko et al. 2008). In addition to serving as a source for adenosine release, immune cells are also among the most widely studied cell types targeted by the regulatory influences of adenosine (Deaglio et al. 2007; Erdmann et al. 2005; Feoktistov and Biaggioni 1995; Fozard et al. 1996; Hasko et al. 2007, 2008; Holgate 2005; Sitkovsky 2009).

2 Adenosine Receptor Expression in the Immune System

Adenosine produces its biological effects by binding to and activating one or more of four membrane-spanning adenosine receptors, designated A1, A2A, A2B, and A3. All four adenosine receptors contain seven transmembrane domains and couple to intracellular GTP binding proteins (G proteins). Adenosine elicits activation of A1, A2A, and A3 receptors with EC50 values that range from 0.01 to 1 μM, and A2B receptor activation occurs at adenosine levels that exceed 10 μM (EC50 24 μM) (Fredholm et al. 2001). Because physiological adenosine concentrations are less than 1 μM, physiological levels of adenosine can activate only A1, A2A, and A3 receptors, and A2B receptor activation requires pathophysiological conditions (Fredholm 2007). In addition to adenosine concentrations at the cell surface, receptor density and the functionality of the intracellular signaling pathways coupled to adenosine receptors are also key factors in dictating the nature and magnitude of the effect of adenosine on the cell. For example, A2A receptor activation inhibits production of the T helper (Th)-1-inducing cytokine interleukin (IL)-12 more potently by human monocytes that are pretreated with the proinflammatory cytokine IL-1 or tumor necrosis factor (TNF)-α, mediators that also increase A2A receptor expression in these cells (Khoa et al. 2001). In addition, the effect of adenosine can also be affected by the polarized localization of adenosine receptors: A3 adenosine receptors accrue at the leading edge of migrating neutrophils and are instrumental in directing the movement of cells in response to chemotactic mediators (Chen et al. 2006). Finally, it is important to keep in mind that results regarding adenosine receptor function in one species cannot be readily extrapolated to another one, because sequence differences in cloned adenosine receptors have been shown to be associated with differential pharmacological responses to selective agonists and antagonists. In this regard, A3 receptors were not even discovered until they were cloned (Zhou et al. 1992), because the A3 receptor, especially in rodents, is insensitive to caffeine and theophylline, antagonists which had been pivotal in identifying adenosine-receptor-mediated effects.

3 Adenosine Receptor Signaling in the Immune System

Adenosine receptors in general dictate cell function through coupling to G proteins, but some G-protein-independent actions have also been reported (Fredholm et al. 2007). Adenosine receptors were initially classified as A1 [cyclic AMP (cAMP)-decreasing] or A2 (cAMP-increasing) receptors (van Calker et al. 1979). Subsequently the cAMP-increasing A2 receptors were divided into two groups: high-affinity A2A receptors and low-affinity A2B receptors (Bruns et al. 1986). More recent studies have revealed that in addition to A1 receptors, A3 receptors also decrease intracellular cAMP concentrations (Jin et al. 1997). In addition to signaling via the adenylyl cyclase–cAMP system, adenosine receptors can signal through a variety of other pathways. A2A receptors, as other Gs-protein-coupled receptors, signal chiefly via the adenylate cyclase–cAMP–protein kinase A (PKA) canonical pathway, but they can also activate exchange factor directly activated by cAMP (Epac) (Fredholm et al. 2007). Signaling downstream from PKA occurs through phosphorylation of the transcription factor CREB on serine residue 133, leading to direct CREB-mediated transcriptional activation (Nemeth et al. 2003). Activated CREB can also modulate gene expression indirectly by competing with nuclear factor (NF)-κB or other transcription factors for an important cofactor, CBP (Fredholm et al. 2007). In other cell types, adenosine A2A receptors stimulate collagen production via mitogen-activated protein (MAP) kinases (Che et al. 2007) and inhibit neutrophil superoxide production through activation of protein phosphatases (Revan et al. 1996). Furthermore, recent results implicated CEBPβ in the stimulatory effect of A2A receptor agonists on IL-10 production by Escherichia coli-challenged macrophages (Csoka et al. 2007). A2B receptor stimulation can induce both adenylyl cyclase activation via Gs and phospholipase C activation via Gq (Feoktistov and Biaggioni 1997). Interaction between these two pathways is important for upregulation of IL-4 production by mast cells upon A2B receptor activation (Ryzhov et al. 2006). Specifically, Gq-mediated activation of phospholipase Cβ causes calcium mobilization and an increase in NFATc1-dependent IL-4 transcription, whose response is further facilitated by Gs-mediated NFATc1 protein accumulation. Traditionally, A3 receptor activation is linked to Gi-mediated inhibition of adenylyl cyclase and Gq-mediated stimulation of phospholipase C (Gessi et al. 2008) and A3 receptors can activate phospholipase D, RhoA, WNT, MAP kinase and phosphatidylinositol 3-kinase pathways in governing cell function. For example, A3-receptor-mediated augmentation of histamine released in sensitized murine mast cells was blocked by inactivating Gi proteins with pertussis toxin and by using pharmacological phosphatidylinositol 3-kinase inhibitors(Zhong et al. 2003). Caffeine and theophylline block A1, A2A, and A2B receptors at pharmacologically relevant concentrations but it appears that these methylxanthines are not potent A3 receptor antagonists (Fredholm et al. 1999). Although there is controversy regarding what constitutes relevant caffeine concentrations in the immune system (Horrigan et al. 2006), the inhibitory effect of adenosine on formyl-Met-Leu-Phe (fMLP)-stimulated respiratory burst in neutrophil leukocytes was reversed by 30 μM caffeine (Fredholm et al. 1996), a concentration that occurs in plasma following consumption of caffeine-containing beverages (Fredholm et al. 1999). In the same study, theophylline proved to be an antagonist of adenosine receptors and it was more potent than caffeine. In addition to its effect on neutrophil respiratory burst, caffeine inhibited the adenosine-receptor mediated accumulation of cAMP in rat thymocytes with a K i value of approximately 20 μM (Fredholm and Sandberg 1983). Theophylline was slightly more potent and its K i value was approximately 10 μM.

4 Caffeine and Theophylline Are Phosphodiesterase Inhibitors at High Concentrations

Higher concentrations of both caffeine and theophylline inhibit cAMP phosphodiesterase (PDE) (Beavo et al. 1971). Inhibition of PDEs by caffeine or theophylline generally requires concentrations of 100–1,000 μM, which typically exceeds the concentrations observed in blood following normal oral dosing and can be associated with toxicity (Sawynok and Yaksh 1993). Nevertheless, it is plausible that some of the biological effects of these methylxanthines are due to PDE inhibition.

5 Theophylline as a Histone Deacetylase Activator

Expression of inflammatory gene expression is governed by a balance between histone acetylation and deacetylation (Barnes 2006). Inflammatory stimuli activate transcription factors such as NF-κB, which leads to histone acetylation and increased inflammatory gene expression. This process is reversed by recruitment of histone deacetylases to the promoter of inflammatory genes. Recent studies have shown that theophylline at clinically relevant concentrations is an activator of histone deacetylases, which explains its anti-inflammatory effects especially in combination with glucocorticoids (Cosio et al. 2004).

6 Effect of Methylxanthines on Inflammatory Cells

6.1 Caffeine

On the basis of the fact that caffeine plasma concentrations rarely exceed 50–60 μM during normal human consumption (Fredholm et al. 1999), we will first consider studies that showed caffeine being efficacious at altering inflammatory cell function at these low concentrations. Sullivan et al. (1995) demonstrated that caffeine increased chemiluminescence and myeloperoxidase release by lipopolysaccharide (LPS)-primed, fMLP-stimulated mixed leukocyte cultures, and increases were already evident at 10 μM caffeine. Consistent with the fact that caffeine antagonizes adenosine receptors at this concentration, addition of the adenosine-degrading enzyme adenosine deaminase abolished the effect of caffeine, indicating that caffeine acted by altering an endogenous adenosine response. In addition, the possibility of PDE inhibition as a mechanism for the enhancing effect of caffeine on leukocyte activation was excluded by a lack of effect of caffeine on leukocyte cAMP levels. In a more recent study, 50 μM caffeine decreased TNF-α production by LPS-activated cord blood (neonatal) monocytes by 20% (Chavez-Valdez et al. 2009). In the same study, however, caffeine failed to affect TNF-α production by adult monocytes.

Caffeine at concentrations of 100 μM or higher appears to downregulate inflammatory cell function. Horrigan et al. (2004) found that caffeine suppressed TNF-α release by LPS-stimulated human whole blood at 100 μM by approximately 40% in a cAMP-PKA dependent fashion. In addition, the production of IL-1β, IL-12, and IL-10 was not affected by 100 μM caffeine (Horrigan et al. 2004). In another study, caffeine was shown to decrease colony formation in cultures of murine bone-marrow-derived macrophages, and again, 100 μM was the lowest effective concentration (Inouye and Wharton 1986).

6.2 Theophylline

Theophylline has been used as a bronchodilator in the therapy of asthma and chronic obstructive pulmonary disease (COPD) for several decades. Its therapeutic concentration is between 55 and 110 μM and it causes unacceptable side effects above 110 μM (Ito et al. 2002). It has long been suggested that theophylline may exert some of its beneficial effects in asthma and COPD by downregulating inflammation. One widely held view is that the beneficial effect of theophylline is secondary to its ability to block A2B receptors (Hasko et al. 2009). This is based, in part, on recent in vivo evidence (Mustafa et al. 2007) documenting that selective antagonists of adenosine A2B receptors inhibit airway inflammation and airway reactivity induced by allergen or AMP in a murine asthma model. The contribution of A2B receptor blockade to the anti-inflammatory effects of theophylline is also underlined by the observation that at therapeutic concentrations, theophylline is a relatively potent A2B receptor antagonist. Although A2B receptors on mast cells seem to be major players in triggering the lung inflammatory response in asthma and COPD (Polosa and Holgate 2006), proinflammatory effects of A2B receptor stimulation have also been observed with human bronchial smooth-muscle cells (Zhong et al. 2004), human bronchial epithelial cells (Zhong et al. 2006), and human lung fibroblasts (Zhong et al. 2005), which produce increased levels of IL-6 (Zhong et al. 2004, 2005) and IL-19 (Zhong et al. 2006) following A2B receptor activation.

In addition to A2B receptors, theophylline blocks A1 and A2A receptors at therapeutically relevant concentrations and thus it can reverse the effects of endogenously released adenosine, resulting in both pro- and anti-inflammatory effects depending on which adenosine receptors are expressed. For example, 10 μM theophylline enhanced TNF-α production by human monocytes stimulated with advanced glycation end products, an effect that was postulated to be secondary to blockade of anti-inflammatory adenosine (A2A or A3) receptors on monocytes (Meiners et al. 2004). Similar to TNF-α production by monocytes, theophylline at therapeutically relevant concentrations (maximal efficacy at 50 μM) augmented superoxide production by fMLP-activated neutrophils (Yasui et al. 2000b). Because 8-sulfophenyltheophylline, a nonselective adenosine receptor antagonist, reproduced the stimulatory effect of theophylline on superoxide production, it was concluded that theophylline acts by antagonizing adenosine receptors. Finally, 50 μM theophylline was shown to accelerate human granulocyte apoptosis (Yasui et al. 1997, 2000a), an observation consistent with an adenosine-receptor-mediated effect, because adenosine can delay the apoptosis of granulocytes (Walker et al. 1997) and other adenosine receptor antagonists can mimic the effect of theophylline (Yasui et al. 2000a).

In addition to blockade of adenosine receptors, recent studies have revealed a further mechanism that can explain the anti-inflammatory effects of theophylline in asthma. This new evidence is based on the idea that inflammatory gene expression is regulated by a balance between histone acetylation and deacetylation (Barnes 2006). It appears that theophylline is able to tip this balance in favor of histone deacetylation, resulting in decreased inflammatory gene expression (Ito et al. 2002).

Lastly, theophylline has anti-inflammatory effects in various in vitro cellular systems when used at therapeutically irrelevant, high concentrations. Theophylline inhibited TNF-α production by LPS-stimulated human mononuclear cells at 200–1,000 μM (Endres et al. 1991); this effect was mediated by inhibition of PDE. In addition, theophylline decreased arachidonate (Hichami et al. 1995), thromboxane B2, (Baker and Fuller 1992), and lysosomal enzyme (Hichami et al. 1995) release by stimulated mononuclear cells at supratherapeutic concentrations.

7 Enprofylline is Anti-inflammatory by Blocking A2B Receptors

On the basis of the efficacy of theophylline as an antiasthma drug (Holgate 2005), efforts to develop a similar xanthine-based compound with a better safety profile led to the development of enprofylline (3-propyl xanthine). It was subsequently demonstrated that similar to theophylline, enprofylline can weakly block A2B receptors on mast cells, which might explain its efficacy as an antiasthma agent. The advantage of enprofylline over theophylline is that enprofylline is a selective A2B antagonist, whereas theophylline blocks other adenosine receptors as well (Fozard and Hannon 1999).

8 Effects of Methylxanthines in Inflammatory States

8.1 Caffeine Exacerbates Acute Inflammatory Liver Injury by Blocking the Physiological Anti-inflammatory Effect of Endogenous Adenosine

Ohta and Sitkovsky (2001) demonstrated that endogenous adenosine by engaging A2A receptors has a nonredundant role in the prevention of inflammatory liver damage induced by concanavalin A. They showed that mice deficient in A2A receptors exhibited increased and protracted production of proinflammatory cytokines, including TNF-α and interferon-γ, which was paralleled by augmented biochemical and histological signs of liver injury. These results raised the possibility that caffeine acting as an A2A receptor antagonist might interfere with the endogenous protective mechanism rendered by adenosine–A2A receptor interaction. To address this possibility, Ohta et al. (2007) injected mice with 20 mg/kg caffeine, which corresponds with caffeine amounts ingested during normal human consumption (Fredholm et al. 1999), before inducing liver injury using concanavalin A. The results showed that caffeine exacerbated liver injury as determined by biochemical and histological analysis, and the increased liver injury coincided with increased production of harmful proinflammatory cytokines.

To investigate the role of A2A-receptor-mediated signaling in the exacerbation of liver injury following caffeine administration, caffeine (10 and 20 mg/kg) was injected into A2A receptor knockout and wild-type mice and liver injury was induced using concanavalin A (Ohta et al. 2007). While caffeine produced an exacerbated liver injury in wild-type mice, knockout mice failed to respond to caffeine with increased acute liver injury, confirming that caffeine enhances liver injury through blocking the action of endogenous adenosine. Importantly, when caffeine was administered at the supratherapeutic dose of 100 mg/kg, it no longer enhanced inflammation; it actually suppressed it. This anti-inflammatory effect is consistent with caffeine being a PDE inhibitor at high concentrations, because other PDE inhibitors have similar anti-inflammatory effects in general acute inflammation (Hasko et al. 1998; Nemeth et al. 1997).

8.2 Caffeine and Theophylline Reverse the Anti-inflammatory Effects of Methotrexate

Low-dose orally administered methotrexate is currently the gold standard therapy for the treatment of rheumatoid arthritis. Recent studies have demonstrated that most of the anti-inflammatory effects of methotrexate are mediated by enhanced adenosine release and activation of anti-inflammatory adenosine receptors (Cronstein 2005). Indeed, most of the genetic markers associated with response or lack of response to methotrexate are in enzymes inhibited by methotrexate polyglutamates that are involved in adenosine generation (Dervieux 2009; Dervieux et al. 2004, 2005, 2009; Weisman et al. 2006; Wessels et al. 2006a, b).

Because it is difficult to sample adenosine levels at inflamed sites in patients with rheumatoid arthritis and it is, therefore, difficult to validate the hypothesis that adenosine mediates the anti-inflammatory effects of methotrexate therapy, experiments were carried out with the adjuvant arthritis model of rheumatoid arthritis using adenosine receptor antagonists. In these studies, both theophylline and caffeine reversed the anti-inflammatory effects of methotrexate in rats with adjuvant arthritis (Montesinos et al. 2000) without affecting the course of the arthritis itself. This finding suggested that such commonly used adenosine receptor antagonists as caffeine might reverse the therapeutic effects of methotrexate therapy in patients as well. Indeed, a small prospective study by Nesher et al. (2003) demonstrated that patients started on methotrexate but told to abstain from caffeine responded significantly better than those who did not. Similar observations were made by Silke et al. (2001) of methotrexate in patients as well. In contrast, a larger retrospective study (Benito-Garcia et al. 2006) did not confirm the effect of caffeine consumption on the response to methotrexate; however, since all of the patients in this study had been taking methotrexate for longer than 2 years, it is likely that any patients who did not respond at all had already stopped methotrexate therapy by the time that the survey was done.

9 Conclusion

Both caffeine and theophylline play a variety of roles in regulating inflammatory responses. Although many of the effects of these commonly used methylxanthines are attributable to adenosine receptor blockade, these methylxanthines regulate inflammation by multiple mechanisms.