15-deoxy-Δ^12,14-PGJ₂: endogenous PPARγ ligand or minor eicosanoid degradation product?

Abstract

15-Deoxy-Δ^12,14-prostaglandin J₂ (15d-PGJ₂) affects gene transcription by activating PPARγ and by covalent addition to transcription factors and signaling molecules, However, it is not known whether the high concentrations of 15d-PGJ₂ required for these responses are consistent with physiological levels. A new study suggests that in vivo 15d-PGJ₂ levels are actually several orders of magnitude below the levels required to induce many of the biological effects attributed to this molecule.

Over the past several years there has been increasing interest in the prostaglandin D₂–derived (PGD₂-derived) product 15-deoxy-Δ^12,14-PGJ₂ (15d-PGJ₂) (Figure 1). This substance was first identified in 1983 as a degradation product of PGD₂, formed after incubation for extended periods of time in the presence of albumin (1). 15d-PGJ₂ attracted relatively little attention until 1995, when Forman et al. (2) and Kliewer et al. (3) independently reported in Cell that it is capable of activating the transcription factor PPARγ. This finding was of considerable interest because the endogenous ligand for PPARγ was unknown, in spite of the availability of a number of pharmacological agonists that are used in the treatment of type 2 diabetes because of their ability to enhance sensitivity to insulin. The hypothesis that 15d-PGJ₂ fulfilled this role was thus very attractive, and 15d-PGJ₂ is often referred to as an endogenous PPARγ ligand. For this to be correct, 15d-PGJ₂ would have to act at concentrations compatible with its physiological levels. Bell-Parikh et al. now report a highly sensitive and specific assay that they have used to address this question (4).

Figure 1.

Formation of 15d-PGJ₂ and its effects on receptors and signaling molecules. The degradation of PGD₂ to 15d-PGJ₂ is enhanced by albumin. In addition to its effects on intracellular proteins, 15d-PGJ₂ also activates the DP₂ receptor. PGD₂ can also be metabolized enzymatically (Enz) to 11β-PGF_2α and 13,14-dihydro-15-keto-PGD₂ (dhk-PGD₂). The latter metabolite also activates the DP₂ receptor. Some of the effects of 15d-PGJ₂ on intracellular proteins are mediated by Michael addition to the reactive cyclopentenone ring (box). In contrast to its effect on the DP₂ receptor, 15d-PGJ₂ does not activate the DP₁ receptor.

Conversion of PGD₂ to 15d-PGJ₂

The synthesis of prostaglandins is initiated by activation of phospholipase A₂, resulting in the release of arachidonic acid, which is then converted by cyclooxygenases to PGH₂. This unstable intermediate is converted enzymatically to a series of biologically active prostanoids, each of which has its own selective receptor(s). PGD₂ is formed by the actions of hematopoietic and lipocalin-type PGD₂ synthases and is metabolized enzymatically to 11β-PGF_2α and 13,14-dihydro-15-keto-PGD₂. Alternatively, PGD₂ can undergo chemical dehydration, losing a molecule of water to form the cyclopentenone prostaglandin PGJ₂. PGJ₂ can then undergo further dehydration by loss of the 15-hydroxyl group, which, coupled with migration of the 13,14-double bond of PGJ_2, results in the formation of 15d-PGJ₂ (Figure 1). These reactions are promoted by albumin but proceed at a relatively slow rate compared to the very rapid formation of PGs from PGH₂ by prostanoid synthases. There is no evidence for the enzymatic formation of any of these cyclopentenone PGs, and this lack of evidence is one of the factors that has brought into question their physiological relevance.

15d-PGJ₂ as a PPARγ ligand

The two papers published in Cell in 1995 sparked considerable research on the biological effects of 15d-PGJ₂. As it has been considered to be an endogenous ligand for PPARγ, 15d-PGJ₂ has been employed in many studies on the role of this transcription factor and in most cases shown to have effects similar to those of synthetic PPARγ ligands. This raised the possibility that 15d-PGJ₂ could be a physiological regulator of processes affected by PPARγ, including adipocyte differentiation and regulation of glucose levels (5). In addition, there has recently been considerable interest in the role of PPARγ in regulating the inflammatory response, as 15d-PGJ₂ and other PPARγ agonists inhibit the expression of a variety of proteins with proinflammatory properties, including cyclooxygenase-2, iNOS, and a variety of cytokines (6). These effects appear to be mediated by inhibitory effects on various transcription factors, including NF-κB, activator protein–1 (AP-1), and signal transducer and activator of transcription–1 (STAT-1). This has led to the proposal that 15d-PGJ₂ may be an anti-inflammatory mediator that plays an important role in the resolution of inflammation (7). There is also evidence that PPARγ activators may have antitumorigenic effects due to their inhibitory effects on tumor cell proliferation and angiogenesis.

Although it is clear that 15d-PGJ₂ can stimulate PPARγ, the physiological and pharmacological relevance of this may be debatable for several reasons. First, the concentrations required to activate PPARγ are generally reported to be in the μM range, which is very high for prostaglandins, since they usually act at low nM concentrations and are normally formed only in small amounts. However, 15d-PGJ₂ acts intracellularly, in contrast to other PGs, which act principally as extracellular mediators, interacting with receptors on the cell surface. Thus it may be argued that it is difficult to know the intracellular concentration of 15d-PGJ₂. These key issues are addressed by Bell-Parikh et al. (4). Secondly, there is growing evidence that 15d-PGJ₂ can induce a variety of responses independently of PPARγ, some of which would oppose those expected by activation of PPARγ. This lack of selectivity could limit the potential pharmacological usefulness of this substance.

Other effects of 15d-PGJ₂

The expression of PPARγ is more restricted than is the expression of other members of the PPAR family (8), and 15d-PGJ₂ has been shown to induce responses in cells devoid of this receptor (9). Some of these effects may be mediated through the covalent binding of 15d-PGJ₂ to proteins. This is due to the reactive cyclopentenone ring of 15d-PGJ₂, which readily reacts with substances containing nucleophilic groups such as cysteinyl thiol groups of proteins. Such reactions are termed Michael addition reactions and also occur with 2-cyclohexenone itself. Thus 15d-PGJ₂, in addition to inhibiting NF-κB activation via PPARγ, has been shown to covalently bind to IκB kinase, inhibiting its activity and causing NF-κB to remain in an inactive state. 15d-PGJ₂ also inhibits the binding of NF-κB subunits to DNA by a similar mechanism (10). Similarly, 15d-PGJ₂ was been found to form an adduct with cysteine-184 of H-Ras in 3T3 cells, but not with either N-Ras or K-Ras, which do not contain a cysteine residue in this position (11). However, in contrast to its inhibitory effect on NF-κB, 15d-PGD₂ activates H-Ras, resulting in the stimulation of MAPK and PI-3K, which appear to be responsible for the positive effect of 15d-PGJ₂ on cell proliferation.

In addition to the intracellular effects of 15d-PGJ₂ on transcription factors and signaling molecules, 15d-PGJ₂ can also act on at least one cell membrane receptor. The parent compound, PGD₂, acts through two such receptors: the DP₁ receptor and the DP₂ receptor, which is also known as the chemoattractant receptor–homologous molecule expressed on TH2 cells (CRTH2). We have shown that this substance activates the DP₂ receptor on eosinophils with a potency (concentration of agonist that elicits 50% of the maximal response [EC₅₀] ∼10 nM) nearly equal to that of PGD₂, the main ligand for this receptor (12). This effect thus occurs at a much lower concentration than that required for activation of PPARγ and would be pro-inflammatory rather than anti-inflammatory.

Is 15d-PGJ₂ biologically relevant?

15d-PGJ₂ clearly induces or inhibits a variety of PPARγ-dependent and PPARγ-independent responses and is active both in vitro and in vivo. However, the high concentrations required to induce of these effects brings into question their physiological relevance, as PGs normally exhibit EC₅₀ values in the low nM range. Nevertheless, the number of publications on 15d-PGJ₂ in high-profile journals continues to increase rapidly despite the paucity of information on its concentrations in biological fluids. Measurement of 15d-PGJ₂ in a physiologically relevant setting has been hampered by the lack of reliable immunoassay methods. To raise antibodies to 15d-PGJ₂ and other prostaglandins, these small molecules must first be conjugated to proteins. As 15d-PGJ₂ spontaneously forms such conjugates in biological milieu, it is difficult to know how much of the immunoreactive material is free 15d-PGJ₂ and how much is 15d-PGJ₂ or related lipids bound to proteins or glutathione. Because of such uncertainties, one commercially available immunoassay (Cayman Chemical Co., Ann Arbor, Michigan, USA) has been withdrawn from the market. Another antibody has been reported to detect relatively high levels (∼80 nM) of 15d-PGJ₂ in medium from RAW264.7 macrophage-like cells stimulated with LPS (13). However, additional validation of this assay would be desirable. The gold standard for measurement of eicosanoids in biological fluids is clearly mass spectrometry, and the availability of reliable mass spectrometric methods to quantitate 15d-PGJ₂ is essential to accurately estimate its biological levels as well as to validate immunological assays. In this issue of the JCI, Bell-Parikh et al. (4) report a highly sensitive liquid chromatography/tandem mass spectrometry assay for this substance. The authors show that although 3T3-L1 fibroblasts can produce a very small amount of 15d-PGJ₂ (∼5 pM), its production does not increase during differentiation of these cells to adipocytes. Furthermore, the amounts of 15d-PGJ₂ released into the medium are more than five orders of magnitude lower than the amounts required for PPARγ-dependent differentiation of these cells into adipocytes, and the concentration of cell-associated 15d-PGJ₂ (∼1 nM) is over three orders of magnitude lower. The levels of 15d-PGJ₂ excreted in the urine were also found to be very low compared to the levels of other PGs and were not altered in diabetes or following administration of LPS. 15d-PGJ₂ was barely detectable in synovial fluid. These results provide strong evidence that, although 15d-PGJ₂ can be produced in vivo, its levels are far too low to be compatible with a role for this substance as an endogenous activator of PPARγ. Furthermore, because of its lack of selectivity for PPARγ, 15d-PGJ₂ may be of limited usefulness in defining the role of this receptor.

Conclusions

The study by Bell-Parikh et al. (4), using a highly sensitive and selective assay, suggests that the amounts of 15d-PGJ₂ generated in vivo are insufficient for it to play a physiological role in modulating most of the responses, including activation of PPARγ, that it has been shown to induce at μM concentrations in vitro. There are only limited instances in which 15d-PGJ₂ has been demonstrated to act at low (1–10) nM levels, such as activation of eosinophils though the DP₂ receptor (12) and stimulation of tumor cell proliferation (14). Further information on the ability of certain cell types such as macrophages (13) to produce 15d-PGJ₂, using mass spectrometric techniques such as those described in the current issue, would seem important in determining whether it could possibly affect such processes. The usefulness of 15d-PGJ₂ at a pharmacological level is also rather questionable, as it has been shown to elicit a variety of responses, many of them in opposition to one another. For example, it has been reported to be antiangiogenic (15), yet also to induce VEGF synthesis (16), and it both induces and prevents proliferation of tumor cells (17). Because of its reactivity, 15d-PGJ₂, like cyclopentenone itself, can react with proteins and either inhibit or enhance their activity. Therefore, if used at sufficiently high concentrations, it will certainly induce or inhibit responses. While this in itself may be of interest, it does not necessarily add to our understanding of physiologically relevant regulatory mechanisms.