Prostaglandin and PPAR control of immune cell function

Peroxisome proliferator-activated receptors (PPARs) belong to a superfamily of intracellular ligand-activated receptors. These receptors are also transcription factors exerting their regulatory functions directly at the gene level.¹ PPARs appear to act as messengers responsible for translating a wide variety of stimuli into changes in gene expression and cellular differentiation pathways. They were originally identified as molecular targets for compounds that induce an increase in the size and number of hepatic and renal peroxisomes. Peroxisomes are single-membrane organelles present in nearly all eukaryotic cells and carry out metabolic peroxidation reactions, primarily the β-oxidation of fatty acids. Peroxisome proliferators increase the capacity of peroxisomes to metabolize fatty acids via increased expression of the enzymes required for β-oxidation.^2,3 Currently three subtypes of mammalian PPARs have been identified, termed α, δ (originally termed β) and γ, encoded by different genes and showing distinct tissue distributions. Both PPAR-α and PPAR-γ act as regulators of lipid metabolism; PPAR-α plays a crucial role in hepatic liver metabolism and PPAR-γ participates in adipocyte differentiation and function. No specific function has yet been identified for PPAR-δ, although it has a very wide distribution, can be activated by prostacyclin, may be involved in the control of cellular development and is potentially involved in the pathogenesis of tumours such as in colorectal cancer.⁴

PPARs are activated by a group of structurally diverse compounds including fatty acids, eicosanoids and drugs such as the hypolipideamic fibrates.⁵ It was also fortuitously discovered that they can be activated by the thiazolidinediones class of antidiabetic drugs (oral hypoglycaemics), thus providing a useful tool with which to characterize their distribution and the spectrum of actions that they modulate. Specific thiazolidinediones such as troglitazone and rosiglitazone have been used to elucidate the role of PPARγ-activated pathways and these compounds have been shown to bind directly to the PPARs.⁶ The discovery that PPARs were activated by direct ligand binding led to a search for the physiological ligand regulator of PPARs, especially PPAR-γ. A potential candidate, on the basis of binding studies and the ability to functionally activate PPAR-γ, was 15-deoxy-Δ^12,14 -prostaglandin J₂ (15d-PGJ₂). It is now generally accepted that the naturally occurring activating ligand for PPAR-γ is 15d-PGJ₂, which can be derived as a metabolite from prostaglandin D₂ (PGD₂), and is predominantly found in adipose tissue.⁷ PGD₂ metabolites, however, have not been identified in adipocytes but are major products of arachidonic metabolism in macrophages, and PGD₂ synthase, which is required for 15d-PGJ₂ synthesis, is predominantly expressed in macrophages and other antigen-presenting cells.⁸ Although much of the early work surrounding PPARs was directed at understanding mechanisms of lipid metabolism and adipose tissue function in obesity, this raised the possibility that PPARs, specifically PPAR-γ, were important regulators of immune function.

PPAR-γ potentially possessed an important role in regulating inflammatory responses by modulating the activity of macrophages/monocytes. Several studies on the role of PPAR-γ in macrophage functions confirmed that activation of PPAR-γ did indeed modulate their activity. Ricote et al.⁹ showed that resting murine bone-marrow-derived macrophages expressed low levels of PPAR-γ mRNA, whereas activated murine peritoneal macrophages expressed high levels of PPAR-γ mRNA and in the presence of either 15d-PGJ₂ or the thiazolidinedione BRL 49653 (rosiglitazone), a specific PPAR-γ agonist, IFNγ-activated macrophages retained the morphological characteristics of resting cells. This indicates that PPAR-γ pathways are capable of halting macrophage activation. Ricote et al.⁹ also examined two other markers of macrophage activation: (a) expression of inducible nitric oxide (iNOS) (cytotoxic for microorganisms) and (b) up-regulation of expression of gelatinase B (a matrix metalloproteinase associated with tissue damage during inflammation). 15d-PGJ₂ and synthetic PPAR-γ ligands inhibited IFNγ-stimulated nitric oxide (NO) production and the induction of iNOS mRNA. 15d-PGJ₂ also inhibited the induction of gelatinase B mRNA. These observations suggested that PPAR-γ can inhibit the expression of genes which are up-regulated during macrophage activation and differentiation.

As to a possible mechanism, Ricote et al.⁹ provided data from transcriptional assays in cultured macrophage-like cell lines which demonstrated that activation of PPAR-γ by either BRL 49653 or 15d-PGJ₂ antagonizes the effects of three other different classes of transcription factor, AP-1, STAT 1 and NF-κB, which play general roles in mediating inflammatory responses in many cell types including macrophages. A major primary activation response of monocytes/macrophages is the release of proinflammatory cytokines in response to a wide range of stimuli, including bacterial lipopolysaccharide and phorbol esters. PPAR-γ agonists have been shown to inhibit the production of monocyte (Mø) inflammatory cytokines at concentrations similar to those found to promote adipogenesis. Using freshly isolated human peripheral blood mononuclear cells, Jiang et al.¹⁰ demonstrated that PMA-induced tumour necrosis factor-α (TNF-α), interleukin-1β (IL-1β) and IL-6 synthesis could be potently inhibited by PGJ₂ and 15d-PGJ₂ whereas A, B and E series prostaglandins had little or no effect at similar concentrations. The structurally dissimilar PPAR-γ agonist troglitazone also inhibited PMA-induced TNF-α production whereas the PPAR-α ligands leukotriene B₄, eicosatetraynoic acid and 8[S]-hydroxyeicosatetraenoic acid had no effect. Using RNA blot analysis, the study by Jiang et al.¹⁰ clearly showed that PMA-induced TNF-α mRNA accumulation in Mø was inhibited by 15d-PGJ₂ and troglitazone, indicating that PPAR-γ agonists inhibit cytokine synthesis pretranslationally. Interestingly, several non-steroidal anti-inflammatory drugs are known to have PPAR-γ agonist activity at high concentrations,¹¹ and Jiang et al.¹⁰ also showed that at relatively high concentrations indomethacin, ibuprofen and fenoprofen suppressed TNF-α mRNA accumulation.

More recently, the potential immunoregulatory control by PPARs and 15d-PGJ₂ has been investigated in lymphocytes. PPAR-γ ligands and 15d-PGJ₂ were shown to display dose-dependent antiproliferative and cytotoxic effects on normal and malignant murine B cells.¹² The cells were shown to express PPAR-γ mRNA and the 67-kDa PPAR-γ protein and the cytotoxic actions of PPAR-γ activation on the B cells occurred as a result of apoptosis.¹² Similar observations have been reported for murine T lymphocytes by Harris & Phipps.¹³ A dramatic inhibition of both naive and antigen-activated T-cell proliferation was induced by troglitazone and 15d-PGJ₂, which was accompanied by apoptosis of the cells and expression of PPAR-γ mRNA and protein.¹³ This clearly indicates that lymphocytes can also be controlled via PPAR activation and also revisits the question of prostaglandins as endogenous, but not necessarily physiological, possibly pathophysiological regulators of immune function via PPARs. Classically, prostaglandins have always been associated with immune/inflammatory responses as they are the end mediators which culminate in the symptomatic manifestations of inflammation. In addition, anti-inflammatory therapy has traditionally been directed towards the inhibition of prostaglandin biosynthesis with the use of non-steroidal anti-inflammatory agents such as aspirin, indomethacin and ibuprofen.

The major arachidonic acid metabolite produced by activated immune cells, especially monocytes/macrophages, is PGE₂. PGE₂ is not only produced in response to cytokine stimulation, primarily IL-1 and TNF-α, but also provides feedback regulation by inhibiting the further release of these cytokines, thereby limiting the response.¹⁴ Thus, PGE₂ is generally an inhibitory end signal for both monocytes/macrophages and T lymphocytes and appears to mediate this effect via EP2/EP4 receptors which are G-protein coupled to the activation of adenylate cyclase. PGD₂ is also reportedly an inhibitory prostanoid and acts via its respective receptor, the DP receptor, which is also a cyclic AMP coupled receptor. In T cells, PGE₂ can inhibit mitogen and cytokine-amplified proliferation¹⁵ in the nanomolar concentration range and this appears to occur via a cyclic AMP-dependent mechanism,¹⁶ indicating a role for EP2/EP4 receptors in this response. In this respect, it is highly unlikely that either PGD₂ or PGE₂ acts via PPAR-γ as they can stimulate elevations in cyclic AMP levels at lower concentrations than those required for ligand activation of PPAR-γ. Also, it is much more likely that in this context the responses would be mediated via cyclic AMP response elements (CREs) which can act independently of PPARs; however, it is thought that PPARs may modulate the action of cyclic AMP via binding to the CRE binding protein.¹⁷

It has now been shown by Harris & Phipps (in this issue of Immunology) that PPAR-γ agonists and 15d-PGJ₂ can also induce apoptosis in human transformed T cells, as they previously showed for murine T cells.¹³ However, they have also demonstrated other particularly novel effects on human T cells, primarily that PGD₂ can also induce identical responses, albeit at 5-fold higher concentrations. In addition, they showed that this was restricted to transformed T cells, with normal anti-CD3/anti-CD28-activated T cells being unaffected. The EC₅₀ concentration of 15d-PGJ₂ is less than 1 µm (in the 0·5–1·0 µm range), but the EC₅₀ for PGD₂ is 5- to 10-fold greater, possibly supporting the view that PGD₂ requires to be converted to its 15d-PGJ₂ metabolite either within cells or extracellularly in order to exert its action.

Harris & Phipps (in this issue of Immunology) also examined the possibility that PGD₂ may be acting through the classical PGD₂ (DP) receptor, as they were able to show that the transformed cells, which responded to PGD₂, also expressed mRNA for the DP receptor. However, they also demonstrated that a potent DP receptor agonist was not able to mimic the actions of PGD₂, and they also showed that a DP receptor antagonist was unable to inhibit the actions of PGD₂. The study did not show whether authentic PGD₂ itself was binding directly to PPAR-γ or whether in their cell system the PGD₂ could be converted to 15d-PGJ₂; however, this has been confirmed in other systems and would most likely have occurred under the similar incubation conditions in the study by Harris & Phipps. It has been shown that mast cells in vitro can rapidly convert both exogenous added and endogenously produced PGD₂ to PGJ2.¹⁸ It has also been observed that human T cells can synthesize PGD_2·¹⁹

The study of Harris & Phipps would tend to suggest that the regulatory effects of PGD₂ on transformed human T cells do not occur through the known DP receptor, although from a pharmacological perspective it is possible that the data point to a new form of the DP receptor. Similar observations on the inability of various PGE₂ agonists and antagonists to modulate PGE₂ functions led to the discovery that there are four forms of the prostaglandin E receptor (EP1–EP4).

Cautionary Observations

The development of new PPAR-γ ligands, including novel prostaglandin analogues, may be of potential therapeutic use in a variety of diseases, such as chronic inflammatory disorders where activated macrophages play a pivotal role or in the control of T-cell leukaemia as suggested by the study of Harris & Phipps (see above). Highly selective ligands for the different PPAR forms have recently been described.²⁰ However, although many studies have employed PPAR ligands such as the thiazolidinediones to elucidate the potential role of PPARs in specific functions, it would appear that some of the actions of thiazolidinediones are not always mediated via PPARs. It has been shown that the inhibition of cell proliferation and tumour growth by thiazolidinediones is independent of PPARs. Instead they appear to inhibit translation initiation by rendering eIF2 inactive.²¹ Similarly, macrophage inflammatory actions can be suppressed by 15d-PGJ₂ and thiazolidinediones that are independent of PPAR-γ even although PPAR-γ is required for the control of lipid metabolism in macrophages.²²