Nonsteroidal antiinflammatory drugs inhibiting prostanoid efflux: As easy as ABC? (original) (raw)

Nonsteroidal antiinflammatory drugs (NSAIDs) and prostanoids continue to be fascinating research targets. Humans have been using NSAIDs in one form or another, from folk remedies through to the products of modern pharmaceutical research, for thousands of years. However, we still have not characterized all of the systems through which these drugs produce their beneficial and harmful effects (1,2). We are certain, from seminal work carried out in the 1960s and early 1970s, that NSAIDs have the common property of inhibiting the activity of cyclooxygenase (COX), the enzyme that initiates the formation of prostanoids (1). An explosion of research in the last decade or so has led us to understand that COX exists in two forms, COX-1 and -2, and that drugs targeted against COX-2 still retain the antiinflammatory properties of traditional NSAIDs (1,2). Because the traditional NSAIDs generally inhibit both isoforms of COX, it is the ability of NSAIDs to inhibit COX-1 that has become associated with their deleterious side effects, particularly within the gastrointestinal tract. However, this dualist approach, COX-1 = good prostanoids and COX-2 = bad prostanoids, does not appear to supply a full understanding of the effects of NSAIDs in vivo. For that reason, researchers have continued to look for other targets at which NSAIDs may exert effects. For example, NSAIDs have been shown to inhibit factors within inflammatory pathways, such as nuclear factor κ-B (3), and to elevate the production of antiinflammatory adenosine (4). Other researchers have suggested that there could be yet more COX isoforms on which the NSAIDs act (see ref. 2). Experiments with chiral NSAIDs have also shown that enantiomers inactive on COX duplicate some of the therapeutic effects of NSAIDs. For example, whereas for ketoprofen and flurbiprofen only the S- and not the R-enantiomers inhibit COX (5) at in flammatory sites and promote gastrointestinal damage, both enantiomers are antinociceptive (6). Interestingly, although R-flurbiprofen is not a COX inhibitor, it does reduce prostaglandin production within the spinal cord, which may explain its antinociceptive properties (7). Others have found therapeutic effects of NSAID metabolites that are inactive on COX; sulindac sulfone is a clear example. The NSAID sulindac is an inactive drug metabolized to a pharmacologically active sulfide derivative, which potently inhibits COX. Sulindac is also metabolized to a sulfone derivative that is inactive against COX. However, like sulindac sulfide, sulindac sulfone both promotes cellular apoptosis and inhibits carcinogenesis. This activity suggests that the chemopreventive properties of sulindac are not necessarily associated with its ability to inhibit COX (8). So it appears clear that, although they share the ability to inhibit COX enzymes, the NSAIDs, as befits a chemically disparate group of drugs, also in fluence other systems.

Some NSAIDs could inhibit the active transport of prostanoids from their generating cells.

In this issue of PNAS, Reid et al. (9) present evidence that some NSAIDs could inhibit the active transport of prostanoids from their generating cells. This is an interesting idea. After synthesis, how do prostanoids leave cells? It would appear unlikely to be via simple diffusion, because prostanoids are largely water-soluble and cross membranes and tight cell barriers poorly, clearly suggesting an active process to export the prostaglandins. In recent years, the ATP-binding cassette (ABC) transporter family of proteins has attracted considerable interest. The human genome appears to contain 48 ABC genes, 16 of which have known functions and 14 of which are associated with a defined human disease (10). ABC transporters move a variety of substances, including lipids, bile salts, and toxic compounds, through membranes while consuming ATP. The multidrug resistance proteins (MRP) form a subfamily within the ABC transporters. MRP1 is a high-affinity leukotriene C4 transporter, and mice that harbor deletions of this gene have an altered response to inflammatory stimuli but are otherwise healthy and fertile. MRP2 is the major transporter responsible for the secretion of bilirubin glucuronides into bile (11). MRP1 and -4 may also be involved in the transport of dehydroepiandrosterone 3-sulfate (the most abundant circulating steroid in humans), and MRP4 transports conjugated steroids and bile acids (12). Through the use of model systems such as inside-out vesicles, overexpressing cells, and small interfering RNA knock-down, Reid et al. (9) show that MRP4 has particular affinity for the transport of prostaglandin (PG)E1 and PGE2 and also displays a high affinity for PGF1α, PGF2α, PGA1, and thromboxane B2.

So might it be that MRP4 is generally the transporter for prostanoids? Because prostanoids are produced by almost all tissues, we should expect to see MRP4 expressed widely. Fortunately, this appears to be the case, for, in addition to its appearance in cancerous tissues (11), MRP4 has been detected in intestinal (13), neuronal (14), prostate (15), and renal (16) tissues and within blood vessels (17). Of course, it may well be significant that MRP4 is highly expressed in kidney and prostate, because they are associated with the production and clearance of high levels of prostanoids. [Prostanoids, of course, are named after the prostate gland, being first characterized as smooth muscle contracting activity within seminal fluid (18).]

So what do these findings suggest? Clearly that there are particular processes regulating the release of prostanoids from cells, and that these processes could be targets for the NSAIDs. Maybe that is not too surprising. We know that NSAIDs have a common ability to bind to the active site of COX (19), a site for which arachidonic acid is the natural substrate. NSAIDs may also bind to other similar sites (20) and so could well associate with MRP4. However, because we know that different NSAIDs bind to the active sites of COX-1 and -2 to differing extents and indeed have little effect against lipoxygenase, which also uses arachidonic acid as a substrate (21), we should expect that different NSAIDs will bind differently to MRP4. This is indeed what Reid et al. (9) report. Their data show that diclofenac is rather inactive, failing to inhibit MRP4 activity by 50% even at a concentration of 100 μM. The COX-2-selective inhibitors celecoxib and rofecoxib are also largely without effect. However, the IC50 values for flurbiprofen, ibuprofen, indomethacin, indoprofen, and ketoprofen against MRP4 all lie between 5 and 50 μM, which are within the plasma concentrations found when the drugs are used therapeutically. These concentrations also lie within the levels of drug required to produce 50% or more inhibition of COX-1 and -2 in in vitro human whole blood assays (22). However, of course in both in vivo and whole blood experiments at matching concentrations there will be considerably less free NSAID than in the studies of Reid et al. (9), because NSAIDs are strongly bound by plasma proteins. Interestingly, with the exception of indomethacin, the drugs found to be active against MRP4 are all members of the same chemical family, being arylpropionic acids. Diclofenac is a member of the heteroaryl acetic acid family, whereas celecoxib and rofecoxib are diarylheterocycles. So these data may point toward structure–activity relationships that could be exploited to find even more potent inhibitors of MRP4.

The data of Reid et al. (9) may also offer some explanations of the other puzzling properties and actions of NSAIDs, including those touched on above. For example, it has been recognized for some time that studies with isolated COX enzymes and broken cell systems yield potency data for the NSAIDs that do not always correlate with the potencies against prostanoid formation found in intact cells and/or tissues (23). Possibly some of the effects of the NSAIDs may now be explained by their ability to inhibit prostanoid release from cells as well as the intracellular formation of prostanoids. What about other effects of NSAIDs? Regular NSAID use has been linked to reductions in the occurrences of certain cancers and, although these reductions appear most associated with the inhibition of COX-2, it may not be the whole story. As remarked above, sulindac sulfone does not inhibit COX but has anticancer properties. Might these anticancer properties be exerted via influences on MRP4? Sulindac is of the same chemical class as indomethacin, so we might predict that it could inhibit MRP4. Of course, if NSAIDs inhibit the activity of MRP4, they might also increase the effectiveness of drugs transported out of cells by this protein. These drugs might include anticancer agents, pro viding us with another potential mechanism through which NSAIDs could inhibit the development of cancers. In addition, NSAIDs could increase the effectiveness of anti-HIV drugs, because MRP4 is associated with resistance to nucleoside analogues such as azidothymidine (24). Interestingly, ketoprofen, flurbiprofen, indomethacin, naproxen, and ibuprofen (i.e., once again arylpropionic and indole/indene acetic acid drugs) all reduce the transport of adefovir, a nucleoside analog with anti-HIV activity, by the human renal organic anion transporter (25). It must be said, however, that despite NSAIDs being taken by possibly ≈50% of those receiving anti-HIV drugs, no reports appear to show that NSAIDs increase the effectiveness of nucleoside analogues (26). This may well be because NSAIDs have not been discriminated by different chemical classes, and so the ineffectiveness of some chemical classes may mask the beneficial effects of others.

The effects of NSAIDs on MRP4 may also explain discrepancies within the central nervous system. Could the ability of the non-COX inhibitor R-flurbiprofen to lower prostanoids within the central nervous system be explained by inhibition of MRP4? If so, this ability might provide an approach to produce analgesic compounds that, like R-flurbiprofen, reduce neuronal prostaglandin formation but promote little gastric damage. Of course, one would first need to test which of the enantiomers of flurbiprofen produces the inhibition of MRP4 reported by Reid et al. (9). We might also wonder whether an effect on MRP4 underlies the potentially beneficial effects of NSAIDs in Alzheimer's disease (27), because these effects do not appear well associated with any of the currently understood targets of NSAIDs (28).

So, in the middle of our fourth millennium of NSAID use, we draw closer to an understanding of the processes underlying prostanoid production and release and to a fuller understanding of the multiple systems on which these drugs can act (Table 1). The findings of Reid et al. (9) may point us toward both new ways to influence prostanoid release from cells and new approaches to limit the activity of MRP4. Might drugs designed to inhibit MRP4 without concomitantly inhibiting COX drive up the intracellular concentration of prostanoids and so provide a novel route via which to activate prostanoid-sensitive nuclear receptors such as peroxisome proliferator-activated receptor γ? Could a new generation of NSAID derivatives be used to increase the efficacy of drugs used in the treatment of a number of pathologies, including HIV? Clearly, we still have more to learn about our favorite and longest-used drugs.

Table 1. Some proposed actions of NSAIDs.

NSAID
Action Target Chemical class Example
Inhibit prostanoid production COX All All
Inhibit prostanoid release MRP4 Arylpropionic acids Flurbiprofen, ibuprofen, indoprofen, ketoprofen
Indole/indene acetic acids Indomethacin
Inhibit COX expression Nuclear factor κ-B and related factors; activator protein 1 Diarylheterocycles Rofecoxib
Indole/indene acetic acids Sulindac
Salicylic acid derivatives Aspirin, sodium salicylate, sulfasalazine
Stimulate nuclear receptors PPAR-γ Anthranilic acids Flufenamic acid
Arylpropionic acids Ibuprofen, fenoprofen
Indole/indene acetic acids Indomethacin
Promote apoptosis Mitogen-activated protein kinase Indole/indene acetic acids Indomethacin
cMYC, nuclear factor κ-B, p38 Indole/indene acetic acids Sulindac
Elevate adenosine Increased intracellular breakdown of ATP Salicylic acid derivatives Aspirin, sodium salicylate
Inhibit neutrophil adhesiveness Erk Salicylic acid derivatives Aspirin, sodium salicylate
Inhibit drug transport Renal organic anion transporter Arylpropionic acids Ibuprofen, flurbiprofen, ketoprofen
Indole/indene acetic acids Indomethacin, naproxen
Salicylic acid derivatives Diflunisal
Promote neuroprotection Blockade of _N_-methyl-d-aspartate receptors Salicylic acid derivatives Sulphasalazine

See companion article on page9244.

References