Spline-fitting with a genetic algorithm: a method for developing classification structure-activity relationships

Classification methods allow for the development of structure-activity relationship models when the target property is categorical rather than continuous. We describe a classification method which fits descriptor splines to activities, with descriptors selected using a genetic algorithm. This method, which we identify as SFGA, is compared to the well-established techniques of recursive partitioning (RP) and soft independent modeling by class analogy (SIMCA) using five series of compounds: cyclooxygenase-2 (COX-2) inhibitors, benzodiazepine receptor (BZR) ligands, estrogen receptor (ER) ligands, dihydrofolate reductase (DHFR) inhibitors, and monoamine oxidase (MAO) inhibitors. Only 1-D and 2-D descriptors were used. Approximately 40% of compounds in each series were assigned to a test set, "cherry-picked" from the complete set such that they lie outside the training set as much as possible. SFGA produced models that were more predictive for all but the DHFR set, for which SIMCA was most predictive. RP gave the least predictive models for all but the MAO set. A similar trend was observed when using training and test sets to which compounds were randomly assigned and when gradually eliminating compounds from the (designed) training set. The stability of models was examined for the random and reduced sets, where stability means that classification statistics and the selected descriptors are similar for models derived from different sets. Here, SIMCA produced the most stable models, followed by SFGA and RP. We show that a consensus approach that combines all three methods outperforms the single best model for all data sets.

Cyclooxygenases: new forms, new inhibitors, and lessons from the clinic

The beneficial actions of nonsteroidal anti-inflammatory drugs (NSAIDs) have been linked to their ability to inhibit inducible COX-2 at sites of inflammation, and their side effects (e.g., gastric damage) to inhibition of constitutive COX-1. Selective inhibitors of COX-2, such as celecoxib, etoricoxib, lumiracoxib, rofecoxib, and valdecoxib have been developed and the greatest recent growth in our knowledge in this area has been come from the clinical use of these compounds. Although clinical data indicate that COX-2 selectivity is associated with a reduction in severe gastrointestinal events, they also reveal there are roles for constitutive COX-2 within tissues such as the brain, kidney, pancreas, intestine, and blood vessels. We now better understand the roles of COX-1 and COX-2 in functions as disparate as the perception of pain and the progression of cancers. Clinical use of COX-2-selective compounds has ignited strong debates regarding potential side effects, most notably those within the cardiovascular system such as myocardial infarctions, strokes, and elevation in blood pressure. This review will discuss how the latest studies help us understand the roles of COX-1 and COX-2 and what clinically proven benefits the newer generation of COX-2-selective inhibitors offer

Prostaglandin H synthases: members of a class of quasi-linear threshold switches

Prostaglandin H synthase (PTGS or COX) enzymes catalyze rate-limiting steps in the synthesis of potent prostanoids, including various prostaglandins, thromboxane, and prostacyclin. Mechanisms that have evolved for regulating prostanoid biosynthesis reflect a tension between achieving a rapid but measured response to cellular signals while minimizing spurious activation by signal noise. We found through mathematical modeling that the PTGS enzymes can be thought of as regulatory switches with approximately linear output above an adjustable threshold. In vivo synthesis allows continuous production while signal remains above threshold. Different isoforms show specific adaptions reflecting their physiological roles as constitutive or inducible enzymes. Mathematical modeling helps explain how these adaptations enable the PTGS1 and PTGS2 enzymes to maintain their physiological roles while avoiding potentially damaging consequences.

Cyclooxygenase-3: axiom, dogma, anomaly, enigma or splice error?--Not as easy as 1, 2, 3

A continued need to develop safe and effective analgesics and anti-inflammatory drugs fuels the ongoing investigations of cyclooxygenase (COX). In addition, a long unanswered question in the biomedical arena revolves around the mechanism of action of acetaminophen leading to its analgesic and antipyretic activity. Upon the discovery of COX in 1971, alternative enzyme forms were initially suggested. With the development of molecular biology, the discovery of two major cyclooxygenase genes (COX-1 and COX-2) was heralded in 1990, which has subsequently led to the clinical use and development of selective COX-2 inhibitors. Splice variants of both COX-1 and COX-2 were first encountered in the early 1990s, as were single nucleotide polymorphisms of COX-1 and 2. There have been some recently well-publicized investigations of COX-1 and -2 enzyme variants that may assist in our eventual conceptual understanding of the mechanisms of action of acetaminophen. The term "COX-3" has been utilized by some scientists and in the media. The evidence for "COX-3" in terms of various COX-variants and possible scientific and therapeutic implications of COX variant enzymes are described.

Cyclooxygenases in the skin: pharmacological and toxicological implications

Cyclooxygenase (COX), a prostaglandin-endoperoxide synthase (PTGS), catalyzes the formation of prostaglandins from arachidonic acid. Prostaglandins are lipid signaling mediators that play a central role in a broad range of diverse physiological and pathophysiological processes, including inflammation, reproduction, nocioception, and gastrointestinal protection. Inhibition of cyclooxygenase activity is the mechanism by which nonsteroidal antiinflammatory drugs (NSAIDS) exert their analgesic, antipyretic, antiinflammatory, and antithrombotic effects. COX is currently believed to exist in three isoforms. In this review, we provide a concise state-of-the-art description of the role of COX in pharmacology and toxicology of skin including its involvement in normal physiology, cutaneous inflammation, nociception, wound healing, and tumorigenesis. COX-dependent pathways influence keratinocyte differentiation, hair follicle development, and hair growth. The critical role of COX-2 in pathophysiology of skin is also addressed. COX-2 mediates inflammatory processes in skin, including inflammatory hyperalgesia and nociception, and administration of specific COX-2 inhibitors reduces edema, vascular permeability, and other markers of cutaneous inflammation. A number of studies in animal models and in humans show that COX-2 inhibitors possess cancer chemopreventive properties. Selective COX-2 inhibitors have a more favorable side-effect profile. Topical formulations of COX-2 inhibitors are being developed as a novel pharmacologic approach for the treatment of COX-2 mediated skin diseases.

Dietary conjugated linoleic acid reduces cerebral prostaglandin E(2) in mice

Conjugated linoleic acids (CLA) are naturally occurring fatty acids that have been recognized to modify prostaglandin (PG) production in specific tissues. So far, no relationship between PG production and CLA has been reported in the brain. Thus, the effects of CLA on cerebral prostaglandin E(2) (PGE(2)) levels and cyclooxygenase (COX)-1 and COX-2 mRNA expression in mice were determined in the present study. Mice given the diet with or without 3% CLA at the age of 9 weeks were mated, reproduced and lactated. Offspring mice given each diet until 25 days and 8 weeks of age were sacrificed and their brains removed. Production of PGE(2) was reduced while COX mRNA was quantitatively increased by CLA supplementation. These results imply that brain PGE(2) production can be reduced by dietary CLA without inhibition of COX gene expression in mice.

Prostaglandin H synthase 2 variant (Val511Ala) in African Americans may reduce the risk for colorectal neoplasia

Prostaglandin H synthase 2 (also known as cyclooxygenase-2) is thought to play a role in the prevention of colon cancer by aspirin, an inhibitor of the enzyme. We used DNA heteroduplex analysis to screen the prostaglandin H synthase 2 gene, to search for naturally occurring enzyme variants that may simulate the effects of aspirin. We found among African-Americans a single-nucleotide polymorphism that changes valine to alanine at residue 511 (V511A; GTT>GCT; g.5939T>C; allele frequency 0.045). The polymorphism was also seen among Asian-Indians (allele frequency, 0.03) but not among Chinese, Filipinos, Hispanics, Japanese, Koreans, Samoans, and Caucasians. The amino acid change is predicted to open a 53 cubic angstrom cavity near the active site of the enzyme, but no change in V(max), K(m), or thermal stability was observed for the variant enzyme in COS-1 cell transfection assays. Case-control analysis of African-Americans from two different study populations showed a 0.56 odds ratio for colorectal adenomas among polymorphism carriers (95% confidence interval, 0.25-1.27; 161 cases and 219 controls). A similar analysis of African-Americans nested in the Multiethnic Cohort Study showed a 0.67 odds ratio for colorectal cancer (95% confidence interval, 0.28-1.56; 138 cases and 258 controls). Consistency of the results across all three of the studies is potentially compatible with a protective effect of the polymorphism, mimicking aspirin.

Tyrosine nitration in prostaglandin H(2) synthase

In this study, we investigated the effects of various nitrogen oxide (NO(x)) species on the extent of prostaglandin H(2) synthase-1 (PGHS-1) nitration in purified protein and in vascular smooth muscle cells. We also examined PGHS-1 activity under these conditions and found the degree of nitration to correlate inversely with enzyme activity. In addition, since NO(x) species are thought to invoke damage during the pathogenesis of atherosclerosis, we examined human atheromatous tissue for PGHS-1 nitration. Both peroxynitrite and tetranitromethane induced Tyr nitration of purified PGHS-1, whereas 1-hydroxy-2-oxo-3-(N-methyl-aminopropyl)-3-methyl-1-triazene (NOC-7; a nitric oxide-releasing compound) did not. Smooth muscle cells treated with peroxynitrite showed PGHS-1 nitration. The extent of nitration by specific NO(x) species was determined by electrospray ionization mass spectrometry. Tetranitromethane was more effective than peroxynitrite, NOC-7, and nitrogen dioxide at nitrating a tyrosine-containing peptide (12%, 5%, 1%, and <1% nitration, respectively). Nitrogen dioxide and, to a lesser extent, peroxynitrite, induced dityrosine formation. Using UV/Vis spectroscopy, it was estimated that the reaction of PGHS-1 with excess peroxynitrite yielded two nitrated tyrosines/PGHS-1 subunit. Finally, atherosclerotic tissue obtained from endarterectomy patients was shown to contain nitrated PGHS-1. Thus, prolonged exposure to elevated levels of peroxynitrite may cause oxidative damage through tyrosine nitration.

[New nonsteroidal anti-inflammatory agents: nitric oxide donors and selective cyclooxygenase-2 inhibitors]

The use of non steroidal anti-inflammatory drugs as analgesic or anti-inflammatory agents is primarily limited by their toxicity to the gastrointestinal tract. Two strategies have been developed recently in order to improve the safety of these drugs. The first approach is the linking of a nitric oxide-releasing moiety to the available compounds. The rationale is that nitric oxide may prevent non steroidal anti-inflammatory drugs-induced ulcerations by preventing mucosal ischemia. The second approach is based on the discovery of two isoforms (COX-1 and COX-2) of the cyclo-oxygenase enzyme. It was hypothesized that the constitutively expressed COX-1 isoenzyme leads to the synthesis of prostaglandins with homeostatic functions whereas COX-2 is merely responsible for the production of prostaglandins mediating pain, fever and inflammation. Accordingly, selective COX-2 inhibitors have been developed. Clinical trials indicate that these compounds are roughly as effective as the available non steroidal anti-inflammatory agents without causing acute gastrointestinal damage. There is some evidence that both COX-1 and COX-2 isoforms are involved in the production of prostaglandins associated with inflammation and homeostatic functions. Finally, the true benefit/risk ratio of these new non steroidal anti-inflammatory drugs remains to be assessed.

Prostaglandin endoperoxide H synthase-1 and -2 mRNA levels and enzyme activity in human decidua at term labor

OBJECTIVE

To determine the labor-related changes of prostaglandin endoperoxide H synthase (PGHS) activity and PGHS-1 and -2 abundance in term decidua and to assess the contribution of the PGHS isoforms to the total PGHS activity present in the tissue.

METHODS

Decidua was collected after elective cesarean delivery (CD) or spontaneous labor (SL) at term. Prostaglandin endoperoxide H synthase activity was determined in microsomal fractions, and PGHS-1 and -2 mRNA levels were measured by ribonuclease protection assays. Prostaglandin endoperoxide H synthase-1 and -2 mRNAs were localized in tissue sections by in situ hybridization.

RESULTS

Prostaglandin endoperoxide H synthase specific activity in decidua microsomes at CD was 111 +/- 3 pg prostaglandin-E2/minute/microgram protein (mean +/- standard error, N = 10 patients), not different from enzyme activity measured after SL (110 +/- 27 N = 10 patients, P = .97, Wilcoxon's rank sum test). Prostaglandin endoperoxide H synthase-1 mRNA abundance in CD tissues was 0.283 +/- 0.047 relative densitometric units (mean +/- standard error, n = 26 patients), which did not change with labor (SL: 0.329 +/- 0.073, n = 20 patients, P = .68). Prostaglandin endoperoxide H synthase-2 mRNA abundance was also unaffected by labor (CD: 0.933 +/- 0.255, n = 27 patients; SL: 0.714 +/- 0.179, n = 23 patients, mean +/- standard error, P = .66). Prostaglandin endoperoxide H synthase specific activity was positively and significantly (P < .05) correlated with both PGHS-1 and -2 mRNA levels. In situ hybridization showed the pervasive presence of both PGHS mRNAs in decidua cells with no detectable changes associated with labor.

CONCLUSION

Both isoforms of PGHS are present in term decidua and contribute to enzyme activity and prostaglandin production. Mechanisms regulating decidual prostanoid biosynthesis at labor do not involve changing the levels of expression of the two PGHS isoforms.

NSAID gastroenteropathy: past, present and future

The toxicity of nonsteroidal anti-inflammatory drugs (NSAIDs) in the gastrointestinal tract continues to be a major limitation to their use in the treatment of inflammatory disorders. Better understanding of the pathogenesis of NSAID enteropathy has facilitated the development of novel NSAIDs that spare the gastrointestinal tract. In particular, identification and characterization of the inducible form of prostaglandin synthase has led to the design of novel NSAIDs that specifically target that enzyme. The pathogenesis of NSAID gastroenteropathy is reviewed, as are the strategies that have been used in the past and are used now to develop NSAIDs that spare the gastrointestinal tract. Also reviewed are the strategies being employed to achieve this goal in the future.

Multiple second messenger pathways regulate IL-1 beta-induced expression of PGHS-2 mRNA in normal human skin fibroblasts

Very little is known about the specific regulation of PGHS-2 mRNA compared with PGHS-1 mRNA. Using normal human fibroblasts, we show that at baseline there is constitutive expression of PGHS-1 mRNA and barely detectable amounts of PGHS-2 mRNA. There was a marked increase in PGHS-2 mRNA transcription following exposure to IL-1 beta. Maximal expression of PGHS-2 mRNA occurred with concentrations of IL-1 beta > or = 1 ng/ml at 3 hours after stimulation. Downregulation of protein kinase C (PKC) activity by pretreating fibroblast cultures with PMA inhibited IL-1-induced PGHS-2 mRNA expression without affecting the constitutive expression of PGHS-1 mRNA. The addition of various PKC inhibitors also blocked the IL-1 beta induction of PGHS-2 mRNA but did not alter PGHS-1 mRNA expression; inhibitors of protein kinase A (PKA) or tyrosine kinase (TK) had only a limited effect on IL-1 beta-induced PGHS-2 mRNA expression. These findings show that IL-1 beta increases PGHS-2 mRNA, at least in part, via activation of PKC. Activation of PKA or TK appears to have a more limited role in this process.

Prostaglandin endoperoxide-H synthase-1 and -2 messenger ribonucleic acid levels in human amnion with spontaneous labor onset

Increased prostaglandin (PG) production within the uterine compartment has a pivotal role in the processes leading to labor onset in women. Two PG endoperoxide-H synthase (PGHS) isoenzymes have been identified in a number of cell types. PGHS-1 is constitutively expressed in most cases, whereas PGHS-2 expression is rapidly induced by several agonists. The aims of this study were to determine the levels of PGHS-1 and PGHS-2 expression before and after spontaneous labor (SL) onset in the amnion and to assess the contribution of PGHS-1 and PGHS-2 to enzyme activity. We established and validated ribonuclease protection assays to quantify PGHS-1 and PGHS-2 messenger ribonucleic acid (mRNA) levels in the amnion. PGHS enzyme activity was measured with an established assay. The antisense RNA probes used in the protection assays were generated using human PGHS-1 and PGHS-2 complementary DNAs. These probes specifically detected the 2.8-kilobase mRNA of PGHS-1 and the 4.8-kilobase mRNA of PGHS-2 in amnion RNA samples on Northern blots. We measured mRNA levels in amnion from patients after SL at term and from patients not in labor undergoing elective cesarean section (CS) at term. PGHS-2 mRNA levels were markedly higher after SL compared to levels in CS amnion [5.18 +/- 1.08 (n = 16) and 2.27 +/- 0.50 (n = 15), densitometric units, respectively; P < 0.02], whereas there was no difference in PGHS-1 mRNA levels after labor compared with CS samples. PGHS-2 mRNA levels were also positively correlated with PGHS enzyme activity in 4 separate assays with a total of 25 patients (r = 0.65-0.88; P < 0.05). There was no correlation between PGHS-1 mRNA levels and enzyme activity. We conclude that PGHS-2 mRNA is present in human amnion; its levels are elevated after SL onset, and they are correlated with enzyme activity. The stimulation of PGHS activity at labor onset probably involves increased expression of PGHS-2. The expression of PGHS-1 does not change in association with labor in human amnion.

Expression of cyclooxygenase types 1 and 2 in human fetal membranes at term

OBJECTIVE

Our purpose was to compare the level of expression of both type 1 and type 2 cyclooxygenase genes before and after labor and to localize their expression within the fetal membranes.

STUDY DESIGN

The sites of type 2 and type 1 cyclooxygenase messenger ribonucleic acid synthesis were identified with in situ hybridization. Expression of both types 1 and 2 cyclooxygenase was studied by reverse transcriptase polymerase chain reaction.

RESULTS

Cyclooxygenase type 2 and type 1 expression was localized within the amniotic epithelium and amniotic mesoderm. Type 1 but not type 2 enzyme was also expressed in the chorionic mesoderm. Expression of the type 2 enzyme was significantly increased with the onset of labor. Type 1 enzyme expression did not significantly change with labor.

CONCLUSION

It is most likely that it is the inducible type 2 cyclooxygenase enzyme that mediates the increase in prostaglandin synthesis in amnion with the onset of labor.