Isoprostanes

Prostanoids That Arise Spontaneously in Membrane Phospholipids

by John A. Lawson and Kirk M. Maxey, M.D.

Isoprostanes are prostaglandin-like compounds generated by the free radical-induced oxidation of unsaturated fatty acids in membrane phospholipids. They are of interest for two main reasons; their biological activity and their ability to act as an indicator of the oxidative status of an organism.

Formation of Isoprostanes

The first report of non-enzymatic formation of prostaglandins was by Nugteren in 1967,¹ when he described the formation of PGE₁ from the autoxidation of 8,11,14-eicosatrienoic acid. Ironically, Nugteren later developed an assay for tetranor prostaglandin metabolites in human urine² with which he measured almost ten times the amount of metabolites than could be accounted for by the current knowledge of prostaglandin production. It is likely that at least part of this "extra" mass could be due to isoprostane metabolites. Interest in these non-enzymatically formed compounds was mainly limited to synthetic chemists^3-7 until isoprostanes of the F₂ series, that is those containing the five-membered PGF-type ring which are isomeric to PGF_2α, were isolated from a biological source—normal human plasma—by Morrow, et al. in 1990.⁸ They proposed a scheme in which four series of regioisomers of PGG₂ are formed, then reduced to PGF_2α isomers (Fig. 1). They also presented evidence that isoprostanes are formed in situ in phospholipids, from which they are presumably cleaved by phospholipases A₂.^9,10 It is not known how long the peroxide may persist in the membrane, or whether the reduction step is enzymatically controlled. Since PGG₂ spontaneously rearranges to PGD₂ and PGE₂ it was suspected that isoprostanes of the D and E series would also be found. They were subsequently reported by Morrow et al. in 1994.¹¹ More recently, the discovery of isoleukotrienes has been reported.¹² Thus far, such compounds have been derived from arachidonic acid, but there is no reason to believe that they will not be formed from other polyunsaturated fatty acids, as shown by mechanistic studies which have generally been performed with 18 or 20-carbon fatty acids with at least three double bonds.

Biological Activity of Isoprostanes

Most of the investigation of the biological effects of isoprostanes has been performed with 8-epi PGF_2α, for the simple reason that it was commercially available. It had been previously synthesized for other unrelated reasons and was therefore available for infusion, bioassay, receptor affinity studies, etc. By coincidence, it proved to be one of the more prevalent isoprostanes. 8-epi PGF_2α also has significant biological activity. It has been shown to be a potent vasoconstrictor in the rat lung13 and kidney, where its effects are seemingly mediated through thromboxane receptors.¹⁴ However, evidence exists for a separate isoprostane receptor that is distinct from the thromboxane receptor.^15,16 8-epi PGF_2α is a mitogen in 3T3 cells and in vascular smooth muscle cells.¹⁷ It may play a role in pulmonary oxygen toxicity.¹⁸

Isoprostanes are potential mediators of hepatorenal syndrome¹⁹ and atherosclerosis.²⁰ Another way in which isoprostanes can exert biological activity is by modification of the fluidity and integrity of membranes, a known result of oxidative damage. This occurs as a result of the distorted shape of the isoprostanes in comparison to the normal fatty acids present in the phospholipids.²¹

Measurement of Isoprostanes

Incidental oxidation of biomolecules has been implicated in a growing list of pathological conditions. A partial list includes aging, cancer, arthritis, ischemia-reperfusion damage, atherosclerosis, and alcoholic liver disease (for a review see Ref. 22). Accurate and efficient assay methods might lead to improved diagnosis and earlier, more effective treatment of these conditions. The development of antioxidative pharmaceuticals has been hampered by the lack of reliable methodology for assessing the oxidative status of an individual. The most commonly used technique involves the reaction of aldehydes, produced during the decomposition of lipid peroxides, with thiobarbituric acid (hence, TBARS for thiobarbituric acid reactive substances). This method has been criticized for its ambiguity and inaccu-racy.^23,24 Several quite sensitive assays for peroxides, (e.g. HPLC with chemiluminescence detection) are available. However, peroxides are inherently unstable and because an entire class of compounds is being measured with no true internal standard, potential errors are introduced. Assays designed to quantitate conjugated dienes suffer from interferences at the lower end of the UV spectrum and lack of sensitivity. Since isoprostanes are relatively stable compounds, a sensitive, specific assay has obvious potential for monitoring the oxidative status of an organism. Of the isoprostanes discovered so far, those of the F-series are the obvious targets for quantitation due to their chemical stability and their excellent chromatographic characteristics. Any assay for measuring isoprostanes must consider that they are produced artifactually in plasma during freeze/thaw cycles,⁸ and that 8-epi PGF_2α can be produced as a product of cyclooxygenase in platelets²⁵ and in serum.²⁶ However, this enzymatic 8-epi PGF_2α is not significant under normal circumstances.²⁶

Analysis by GC/Mass Spectroscopy

GC/MS assays utilizing a single deuterated internal standard for the measurement of all F-isoprostanes are extremely sensitive, but may have errors introduced by variable recovery of the myriad analytes through the purification scheme, or by impurities coeluting with the isoprostane zone on the GC. An assay directed at a single compound, e.g. 8-epi PGF_2α, may be more reliable due to the ability to introduce a true stable isotope-labeled internal standard (Fig. 2). Three internal standards for GC/MS measurement of 8-epi PGF_2α have been used in the past; these are [¹⁸O]-8-epi PGF_2α, 3,3,4,4-d₄-8-epi PGF_2α, and the corresponding tetradeuterated PGF_2α compound with the natural configuration (i.e., α-configuration) at C-8. The [¹⁸O] standard shows almost identical chromatographic retention compared to the natural compound; the d4 standards, particularly PGF_2α-d4, have distinctly different retention times. The main advantage of the d4 standards is chemical stability, since [¹⁸O] can be exchanged due to esterase activity in the sample. Derivatization of isoprostane mixtures containing 8-epi PGF_2α in the traditional manner (pentafluorobenzyl ester, trimethylsilyl ether formation) gives derivatives with identical GC retention times for several isoprostane isomers leading to an overestimation of the true 8-epi PGF_2α content. Cleaner separation and better quantitation of 8-epi PGF_2α is obtained when a tert-butyl-dimethylsilyl group is used to derivatize the alcohol groups instead. Recently, urinary metabolites of isoprostanes have been measured by GC/MS.²⁷

Immunoassay Measurement

Enzyme immunoassay (EIA) and radioimmunoassay (RIA) techniques have been used to characterize 8-epi PGF_2α excretion in man.²⁶ These assays have the practical advantage of being easier to run using less expensive equipment than GC/MS. On the other hand, all immunoassays suffer from immunologic interference, which often makes extensive sample preparation necessary. One recent study in which both RIA and EIA techniques were validated in human urine determined that a reverse phase extraction followed by thin layer chromatography were both necessary and sufficient to give immunoassay results that were consistent with GC/MS measurements of the same samples.²⁶

Results of Selected Studies

Normal plasma from healthy volunteers shows low but detectable levels of 8-epi PGF_2α that range between 40 - 100 pg/ml and rise with the age of the test subject.⁹ Rats show similar plasma levels which increase dramatically (by 10-100 fold) in the face of an oxidative insult such as diquat or carbon tetrachloride ingestion. Normal human urinary levels range between 50 - 100 ng/mmole creatinine, which is an order of magnitude larger than the urinary levels of most enzymatically derived prostanoids. Samples of both plasma and urine from heavy smokers show significant elevations in 8-epi PGF_2α content. Thus, there is a strong basis for the use of 8-epi PGF_2α measurements both as markers of clinical conditions and as a general indicator of oxidative stress. Simpler assay methodology and sample preparation may be necessary before 8-epi PGF_2α measurement replaces some of the older, less precise methodologies.

References

1. Nugteren, D.H. et al. Recueil 86, 1237-1245 (1967).
2. Nugteren, D.H. J. Biol. Chem. 250, 2808-2812 (1975).
3. Porter, N.A., Funk, M.O. J. Org. Chem. 40, 3614-3615 (1975).
4. Pryor, W.A., Stanley, J.P. J.Org Chem. 40, 3616-3617 (1975).
5. Mihelich, E.D. J. Am. Chem. Soc. 102, 7141-7143 (1980).
6. O'Connor, D.E., Mihelich, E.D., Coleman, M.C. J. Am. Chem. Soc. 103, 223-224 (1980).
7. O'Connor, D.E., et al. J. Am. Chem. Soc. 184, 1-10 (1984).
8. Morrow, J.D., Harris, T.M., Roberts, L.J. II Anal. Biochem. 184, 1-10 (1989).
9. Morrow, J.D., et al. Proc. Natl. Acad. Sci. USA 89, 10721-10725 (1992).
10. Kayganich-Harrison, K.A., et al. J. Lipid Res. 34, 1229-1235 (1993).
11. Morrow, J.D., et al. J. Biol. Chem. 269, 4317-4326 (1993).
12. Harrison K.A., Murphy, R.C. J. Biol. Chem., 270, 17273-17278 (1995).
13. Banerjee, M., Am. J. Physiol. 263, (Heart Circ. Physiol. 32): H660-H663 (1992).
14. Takabashi, K., et al. J. Clin. Invest. 90, 136-141 (1992).
15. Fukunaga, M., et al. Am. J. Physiol. 264, (Cell Physiol. 33): C1619-C1624 (1993).
16. Yura, T., et al. Advances in Prostaglandin, Thromboxane, & Leukotriene Research 23, 237-239 (1995).
17. Hanasaki, K., et al. Biochem. Pharmacol. 40, 2535-2542 (1990).
18. Vacchiano, C.A. et al. J. Applied Phys. 77, 2912-2917 (1994).
19. Morrow, J.D., et al. J. Lipid Mediators S(1-3), 417-420 (1993).
20. Gopaul, N.K., et al. Biochemical & Biophysical Research Communications 200, 338-343 (1994).
21. Morrow, J.D., et al. Proc. Natl. Acad. Sci. USA 89, 10721-10725, (1992).
22. Cross, C.E., et al. Annals of Int. Med. 107, 526-545 (1987).
23. Gutteridge, J.M.C., Halliwell, B. Trends Biochem. Sci. 15, 129-135 (1990).
24. Gutteridge, J.M.C., Free Radical Res. Commun. 1, 173-184 (1986).
25. Pratico, D., et al. J. Biol Chem. 270, 9800-9808 (1995).
26. Wang. Z., et al. J. Pharmacol. and Exper. Ther. 275, 94-100 (1995).
27. Awad, J.A. et al. J. Biol. Chem. 268, 4161-4169 (1992).

Prostaglandins in the Nucleus: The Second Wave

The potent biological activity of prostaglandins was evident soon after their characterization by Bergström in 1957.¹ Nearly 25 years passed before the transduction of this activity through rhodopsin-type, seven-transmembrane G-protein-coupled receptors was fully formalized.² Eight different receptors (EP₁, EP₂, EP₃, EP₄, FP, IP, TP, and DP) have now been characterized pharmacologically and cloned in at least one species.³ Although this paradigm fully explains a broad range of prostaglandin actions in many tissues, some subtle inconsistencies remain. The 5-lipoxygenase and inducible cyclooxygenase (PGHS-2) enzymes were recently shown to occur primarily in the nuclear envelope.^4,5 This implies that the eicosanoids synthesized through their activation are contained within the nucleus, separated by two layers of membrane from the binding sites of their classical receptors. Antimitotic and antiproliferative activity in the A and J-series prostaglandins has been known for several years,10 but the signal transduction pathway associated with this activity has remained obscure.

The hypothesis that some prostaglandin synthesis is compartmentalized within the nucleus and functions in controlling gene expression received support from a pair of recent articles in Cell.^6,11 These papers present evidence that 15-deoxy-Δ^12,14-PGJ₂ is the natural ligand for the γ-isoform of the Peroxisome Proliferator Activated Receptor (PPAR, an orphan receptor which until this time had no known natural ligand). PPARs are part of the nuclear receptor superfamily that includes receptors for steroids, thyroxine, and retinoids. These soluble receptors act by binding to specific control sequences (HREs, or Hormone Response Elements) in the promoter region of their target gene. The receptor is inactive and may actually block transcription in the absence of its activating hormone ligand. In the case of PPARγ, a second nuclear receptor, the 9-cis retinoic acid receptor, must bind to its HRE alongside PPAR, forming a heterodimer. Transcription is activated when the ligand binding domain of each nuclear receptor binds its respective ligand (See Diagram 1). Precisely which genes contain a PPARγ HRE dimer in their promoter region is still unknown. Such genes are likely to include regulatory genes that code for other transcription factors. In fibroblasts, the cascade of activated genes and their expressed protein products leads to differentiation into an adipocyte phenotype.

This adipogenic effect of PPARγ activators has been known for some time.⁷ The antidiabetic activity of a series of thiazolidinediones, exemplified by BRL 49653, can be explained by their ability to induce adipocyte differentiation through the activation of the PPARγ expression cascade. In addition to these synthetic pharmaceutical compounds, other members of the PPAR receptor family (PPARα, β, and δ) are activated by naturally occurring lipids including oleic and linoleic acids. An intriguing observation has been that ω-3 fatty acids such as EPA and DHA lower serum lipid levels. It is possible that they preferentially activate a PPAR isotype.⁸ This suggests that modified gene expression could provide an explanation for some of the well documented health effects of diets rich in these marine polyunsaturated fatty acids.

Evidence for an endogenous signal transduction pathway involving 15-deoxy-Δ^12,14 -PGJ₂ is incomplete. Its immediate precursors in the prostaglandin metabolic pathway, Δ¹²-PGJ₂ and PGJ₂, have both been observed in human urine, but 15-deoxy-Δ^12,14-PGJ₂ itself has not.⁹ One of the most consistent findings from the many studies of classical prostaglandin receptors was the absolute requirement for a 15(S)-hydroxyl for full agonist activity. Yet, the antimitotic and antiproliferative prostanoids such as PGJ₂ seemed to violate this rule. A more complete understanding of the metabolic fate of prostaglandins synthesized within the nucleus is required before a complete signalling pathway can be constructed. Still, the findings of Forman and Lehmann open the way for an active search for additional nuclear prostaglandin receptors, with the expectation that they may have strikingly different structure-activity relationships when compared to their cytoplasmic membrane counterparts. The additional possibilities for new therapeutic target molecules should make this an area of vigorous research. Nuclear receptors may be the catalyst that sparks a second wave of discovery in the otherwise mature field of eicosanoid research.

References

1. Bergström, S. and Sjövall, J. Acta. Chem. Scand. 11, 1086 (1957).
2. Coleman, R.A., Humphrey, P.P.A., Kennedy, I. and Lumley, P. Trends. Pharmacol. Sci. 5, 303 (1984).
3. Coleman, R.A., Smith, W.L., Narumiya, S. Pharm. Rev. 46, 205 (1994).
4. Woods, J.W., Evans, J.F., Ethier, D., et al. J. Exp. Med. 178, 1935-1946 (1993).
5. Morita, I., Schindler, M., Regier, M.K., et al. J. Biol. Chem. 270, 10902-10908 (1995).
6. Forman, B.M., Tontonoz, P., Chen, J., et al. Cell 83, 803-812 (1995).
7. Tontonoz, P., Hu, E. Spiegelman, B.M. Cell 79, 1147 (1994).
8. Wahli, W., Braissant, O. Desvergne, B. Chemistry and Biology 2, 261 (1995).
9. Hirata, Y., Hayashi, H., Ito, S., Kikawa, Y., et al. J. Biol. Chem. 263, 16619 (1988).
10. Marini, S., Palamara, A.T., Garaci, E. et al. Br. J. Cancer 61, 394 (1990).
11. Kliewer, S.A., Lenhard, J.M., Willson, T.M. et al. Cell 83, 813-819 (1996).

John Lawson is currently Director of the Mass Spectrometry Laboratory of The Center for Experimental Therapeutics at The University of Pennsylvania, Philadelphia, PA.